CUDA Available:  True
cuDNN Version:  92000
Processing file: 331_UCR_id_29_Facility_tr_50000_1st_837400.csv, Data shape: (900000, 1)
Label shape: (900000,)
Train data shape: (50000, 1), Test data shape: (850000, 1), Label test shape: (850000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 331_UCR_id_29_Facility_tr_50000_1st_837400.csv...
Using CUDA device: 2
Data shape: (850000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/3 [00:00<?, ?it/s]batched_time_series shape: torch.Size([64, 5000, 1])
batched_mask shape: torch.Size([64, 5000])
batched_time_series shape: torch.Size([64, 5000, 1])
batched_mask shape: torch.Size([64, 5000])
batched_time_series shape: torch.Size([42, 5000, 1])
batched_mask shape: torch.Size([42, 5000])

100%|██████████| 3/3 [00:00<00:00,  8.42it/s]
Original shapes - Output: (850000,), Label: (850000,)
Aligned shapes - Output: (850000,), Label: (850000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   4%|▎         | 7/200 [00:01<00:31,  6.13it/s]

Calculating Affiliation Metrics:   7%|▋         | 14/200 [00:02<00:31,  5.92it/s]

Calculating Affiliation Metrics:  10%|█         | 20/200 [00:03<00:31,  5.69it/s]

Calculating Affiliation Metrics:  13%|█▎        | 26/200 [00:04<00:31,  5.46it/s]

Calculating Affiliation Metrics:  16%|█▌        | 32/200 [00:05<00:33,  5.08it/s]

Calculating Affiliation Metrics:  19%|█▉        | 38/200 [00:07<00:32,  4.98it/s]

Calculating Affiliation Metrics:  22%|██▏       | 44/200 [00:08<00:31,  4.92it/s]

Calculating Affiliation Metrics:  24%|██▍       | 49/200 [00:09<00:31,  4.87it/s]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:10<00:30,  4.84it/s]

Calculating Affiliation Metrics:  30%|██▉       | 59/200 [00:11<00:29,  4.74it/s]

Calculating Affiliation Metrics:  32%|███▏      | 64/200 [00:12<00:28,  4.74it/s]

Calculating Affiliation Metrics:  34%|███▍      | 69/200 [00:13<00:27,  4.73it/s]

Calculating Affiliation Metrics:  37%|███▋      | 74/200 [00:14<00:27,  4.62it/s]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [00:16<00:26,  4.65it/s]

Calculating Affiliation Metrics:  42%|████▏     | 84/200 [00:17<00:24,  4.68it/s]

Calculating Affiliation Metrics:  44%|████▍     | 89/200 [00:18<00:23,  4.63it/s]

Calculating Affiliation Metrics:  47%|████▋     | 94/200 [00:19<00:22,  4.68it/s]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [00:20<00:21,  4.71it/s]

Calculating Affiliation Metrics:  52%|█████▎    | 105/200 [00:21<00:20,  4.73it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 111/200 [00:22<00:18,  4.88it/s]

Calculating Affiliation Metrics:  58%|█████▊    | 117/200 [00:23<00:16,  5.03it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 123/200 [00:24<00:15,  5.12it/s]

Calculating Affiliation Metrics:  64%|██████▍   | 129/200 [00:25<00:13,  5.31it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 136/200 [00:27<00:11,  5.53it/s]

Calculating Affiliation Metrics:  72%|███████▏  | 143/200 [00:28<00:10,  5.69it/s]

Calculating Affiliation Metrics:  75%|███████▌  | 150/200 [00:29<00:08,  5.95it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 157/200 [00:30<00:06,  6.22it/s]

Calculating Affiliation Metrics:  82%|████████▎ | 165/200 [00:31<00:05,  6.56it/s]

Calculating Affiliation Metrics:  87%|████████▋ | 174/200 [00:32<00:03,  7.02it/s]

Calculating Affiliation Metrics:  92%|█████████▏| 184/200 [00:33<00:02,  7.65it/s]

Calculating Affiliation Metrics:  98%|█████████▊| 195/200 [00:34<00:00,  8.55it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:35<00:00,  5.71it/s]
Table1-only metrics completed in 62.53s
Evaluation Result:  {'AUC-PR': 0.0013464454203579027, 'AUC-ROC': 0.32809648015754694, 'VUS-PR': 0.00137445124563997, 'VUS-ROC': 0.33120542178553397, 'Standard-F1': 0.0041339114641747235, 'Standard-Precision': 0.00207124194008768, 'Standard-Recall': 0.9988235294117642, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6706459120544042, 'Affiliation-P': 0.5062973135602555, 'Affiliation-R': 0.9929741176470589, 'F1_T': 0.004132462573116624, 'Precision_T': 0.002070511961197704, 'Recall_T': 0.9994117647058818}
Results for 331_UCR_id_29_Facility_tr_50000_1st_837400.csv: {'filename': '331_UCR_id_29_Facility_tr_50000_1st_837400.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '50000', 'data_shape': '900000x1', 'output_length': 850000, 'label_length': 850000, 'aligned_length': 850000, 'AUC-PR': 0.0013464454203579027, 'AUC-ROC': 0.32809648015754694, 'VUS-PR': 0.00137445124563997, 'VUS-ROC': 0.33120542178553397, 'Standard-F1': 0.0041339114641747235, 'Standard-Precision': 0.00207124194008768, 'Standard-Recall': 0.9988235294117642, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6706459120544042, 'Affiliation-P': 0.5062973135602555, 'Affiliation-R': 0.9929741176470589, 'F1_T': 0.004132462573116624, 'Precision_T': 0.002070511961197704, 'Recall_T': 0.9994117647058818}
Logits metrics skipped for 331_UCR_id_29_Facility_tr_50000_1st_837400.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/331_UCR_id_29_Facility_tr_50000_1st_837400_results.pkl
Skipping file: 223_LTDB_id_8_Medical_tr_4456_1st_4556.csv due to filter criteria.
Processing file: 405_UCR_id_103_Sensor_tr_2827_1st_5988.csv, Data shape: (11308, 1)
Label shape: (11308,)
Train data shape: (2827, 1), Test data shape: (8481, 1), Label test shape: (8481,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 405_UCR_id_103_Sensor_tr_2827_1st_5988.csv...
Using CUDA device: 2
Data shape: (8481, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 8481, window_size 5000, adding 1519 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 46.53it/s]
Original shapes - Output: (10000,), Label: (8481,)
Aligned shapes - Output: (8481,), Label: (8481,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 455.89it/s]
Table1-only metrics completed in 0.87s
Evaluation Result:  {'AUC-PR': 0.7259997031680985, 'AUC-ROC': 0.9969074624222869, 'VUS-PR': 0.6874091576759801, 'VUS-ROC': 0.997172420262111, 'Standard-F1': 0.7218045112777266, 'Standard-Precision': 0.5680473372781032, 'Standard-Recall': 0.9896907216494744, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9764494164893107, 'Affiliation-P': 0.9543301991044276, 'Affiliation-R': 0.9996183099396224, 'F1_T': 0.7232142857137881, 'Precision_T': 0.6377952755905462, 'Recall_T': 0.8350515463917441}
Results for 405_UCR_id_103_Sensor_tr_2827_1st_5988.csv: {'filename': '405_UCR_id_103_Sensor_tr_2827_1st_5988.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 98, 'train_index': '2827', 'data_shape': '11308x1', 'output_length': 10000, 'label_length': 8481, 'aligned_length': 8481, 'AUC-PR': 0.7259997031680985, 'AUC-ROC': 0.9969074624222869, 'VUS-PR': 0.6874091576759801, 'VUS-ROC': 0.997172420262111, 'Standard-F1': 0.7218045112777266, 'Standard-Precision': 0.5680473372781032, 'Standard-Recall': 0.9896907216494744, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9764494164893107, 'Affiliation-P': 0.9543301991044276, 'Affiliation-R': 0.9996183099396224, 'F1_T': 0.7232142857137881, 'Precision_T': 0.6377952755905462, 'Recall_T': 0.8350515463917441}
Logits metrics skipped for 405_UCR_id_103_Sensor_tr_2827_1st_5988.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/405_UCR_id_103_Sensor_tr_2827_1st_5988_results.pkl
Processing file: 774_YAHOO_id_224_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 774_YAHOO_id_224_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 149.64it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1647.55it/s]
Table1-only metrics completed in 0.37s
Evaluation Result:  {'AUC-PR': 0.928030303030303, 'AUC-ROC': 0.9991471734892787, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9969368810069911, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.9997412008281573, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Results for 774_YAHOO_id_224_Synthetic_tr_500_1st_623.csv: {'filename': '774_YAHOO_id_224_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 35, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.928030303030303, 'AUC-ROC': 0.9991471734892787, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9969368810069911, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.9997412008281573, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 774_YAHOO_id_224_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/774_YAHOO_id_224_Synthetic_tr_500_1st_623_results.pkl
Processing file: 003_NAB_id_3_WebService_tr_1362_1st_1462.csv, Data shape: (15852, 1)
Label shape: (15852,)
Train data shape: (1362, 1), Test data shape: (14490, 1), Label test shape: (14490,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 003_NAB_id_3_WebService_tr_1362_1st_1462.csv...
Using CUDA device: 2
Data shape: (14490, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 14490, window_size 5000, adding 510 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.57it/s]
Original shapes - Output: (15000,), Label: (14490,)
Aligned shapes - Output: (14490,), Label: (14490,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  87%|████████▋ | 174/200 [00:01<00:00, 173.94it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 186.17it/s]
Table1-only metrics completed in 1.35s
Evaluation Result:  {'AUC-PR': 0.12649351467730846, 'AUC-ROC': 0.40084894525681414, 'VUS-PR': 0.12673334966150882, 'VUS-ROC': 0.4093645544956059, 'Standard-F1': 0.1894378623643336, 'Standard-Precision': 0.10477518299058904, 'Standard-Recall': 0.9868680236375569, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8120946871508055, 'Affiliation-P': 0.7008753540496135, 'Affiliation-R': 0.9652698033391557, 'F1_T': 0.1894378623643336, 'Precision_T': 0.10477518299058904, 'Recall_T': 0.9868680236375569}
Results for 003_NAB_id_3_WebService_tr_1362_1st_1462.csv: {'filename': '003_NAB_id_3_WebService_tr_1362_1st_1462.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 21, 'train_index': '1362', 'data_shape': '15852x1', 'output_length': 15000, 'label_length': 14490, 'aligned_length': 14490, 'AUC-PR': 0.12649351467730846, 'AUC-ROC': 0.40084894525681414, 'VUS-PR': 0.12673334966150882, 'VUS-ROC': 0.4093645544956059, 'Standard-F1': 0.1894378623643336, 'Standard-Precision': 0.10477518299058904, 'Standard-Recall': 0.9868680236375569, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8120946871508055, 'Affiliation-P': 0.7008753540496135, 'Affiliation-R': 0.9652698033391557, 'F1_T': 0.1894378623643336, 'Precision_T': 0.10477518299058904, 'Recall_T': 0.9868680236375569}
Logits metrics skipped for 003_NAB_id_3_WebService_tr_1362_1st_1462.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/003_NAB_id_3_WebService_tr_1362_1st_1462_results.pkl
Skipping file: 175_MITDB_id_6_Medical_tr_27274_1st_27374.csv due to filter criteria.
Skipping file: 186_SMD_id_9_Facility_tr_5925_1st_17580.csv due to filter criteria.
Processing file: 524_UCR_id_222_Medical_tr_1875_1st_5865.csv, Data shape: (7501, 1)
Label shape: (7501,)
Train data shape: (1875, 1), Test data shape: (5626, 1), Label test shape: (5626,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 524_UCR_id_222_Medical_tr_1875_1st_5865.csv...
Using CUDA device: 2
Data shape: (5626, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5626, window_size 5000, adding 4374 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 164.19it/s]
Original shapes - Output: (10000,), Label: (5626,)
Aligned shapes - Output: (5626,), Label: (5626,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 677.28it/s]
Table1-only metrics completed in 0.92s
Evaluation Result:  {'AUC-PR': 0.057432531037086655, 'AUC-ROC': 0.8438487876505147, 'VUS-PR': 0.14262315431116515, 'VUS-ROC': 0.906255973551499, 'Standard-F1': 0.13393665158353232, 'Standard-Precision': 0.07429718875502, 'Standard-Recall': 0.6788990825688012, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7531832502028889, 'Affiliation-P': 0.6177691015850056, 'Affiliation-R': 0.9646285104870246, 'F1_T': 0.13357400722003898, 'Precision_T': 0.074074074074074, 'Recall_T': 0.6788990825688012}
Results for 524_UCR_id_222_Medical_tr_1875_1st_5865.csv: {'filename': '524_UCR_id_222_Medical_tr_1875_1st_5865.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 165, 'train_index': '1875', 'data_shape': '7501x1', 'output_length': 10000, 'label_length': 5626, 'aligned_length': 5626, 'AUC-PR': 0.057432531037086655, 'AUC-ROC': 0.8438487876505147, 'VUS-PR': 0.14262315431116515, 'VUS-ROC': 0.906255973551499, 'Standard-F1': 0.13393665158353232, 'Standard-Precision': 0.07429718875502, 'Standard-Recall': 0.6788990825688012, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7531832502028889, 'Affiliation-P': 0.6177691015850056, 'Affiliation-R': 0.9646285104870246, 'F1_T': 0.13357400722003898, 'Precision_T': 0.074074074074074, 'Recall_T': 0.6788990825688012}
Logits metrics skipped for 524_UCR_id_222_Medical_tr_1875_1st_5865.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/524_UCR_id_222_Medical_tr_1875_1st_5865_results.pkl
Processing file: 136_WSD_id_108_WebService_tr_3414_1st_3514.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (3414, 1), Test data shape: (14822, 1), Label test shape: (14822,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 136_WSD_id_108_WebService_tr_3414_1st_3514.csv...
Using CUDA device: 2
Data shape: (14822, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 14822, window_size 5000, adding 178 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.11it/s]
Original shapes - Output: (15000,), Label: (14822,)
Aligned shapes - Output: (14822,), Label: (14822,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  86%|████████▌ | 172/200 [00:01<00:00, 171.44it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 186.15it/s]
Table1-only metrics completed in 3.40s
Evaluation Result:  {'AUC-PR': 0.846913133392995, 'AUC-ROC': 0.9768885942446575, 'VUS-PR': 0.855728679642281, 'VUS-ROC': 0.9966606261901367, 'Standard-F1': 0.8192771084332353, 'Standard-Precision': 0.9127516778523429, 'Standard-Recall': 0.7431693989070998, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9888646502480822, 'Affiliation-P': 0.9807043770916153, 'Affiliation-R': 0.9971618633288555, 'F1_T': 0.8192771084332353, 'Precision_T': 0.9127516778523429, 'Recall_T': 0.7431693989070998}
Results for 136_WSD_id_108_WebService_tr_3414_1st_3514.csv: {'filename': '136_WSD_id_108_WebService_tr_3414_1st_3514.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '3414', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 14822, 'aligned_length': 14822, 'AUC-PR': 0.846913133392995, 'AUC-ROC': 0.9768885942446575, 'VUS-PR': 0.855728679642281, 'VUS-ROC': 0.9966606261901367, 'Standard-F1': 0.8192771084332353, 'Standard-Precision': 0.9127516778523429, 'Standard-Recall': 0.7431693989070998, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9888646502480822, 'Affiliation-P': 0.9807043770916153, 'Affiliation-R': 0.9971618633288555, 'F1_T': 0.8192771084332353, 'Precision_T': 0.9127516778523429, 'Recall_T': 0.7431693989070998}
Logits metrics skipped for 136_WSD_id_108_WebService_tr_3414_1st_3514.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/136_WSD_id_108_WebService_tr_3414_1st_3514_results.pkl
Processing file: 262_IOPS_id_3_WebService_tr_7028_1st_7128.csv, Data shape: (125135, 1)
Label shape: (125135,)
Train data shape: (7028, 1), Test data shape: (118107, 1), Label test shape: (118107,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 262_IOPS_id_3_WebService_tr_7028_1st_7128.csv...
Using CUDA device: 2
Data shape: (118107, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 118107, window_size 5000, adding 1893 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([24, 5000, 1])
batched_mask shape: torch.Size([24, 5000])

100%|██████████| 1/1 [00:00<00:00, 54.61it/s]
Original shapes - Output: (120000,), Label: (118107,)
Aligned shapes - Output: (118107,), Label: (118107,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  10%|▉         | 19/200 [00:01<00:09, 18.20it/s]

Calculating Affiliation Metrics:  19%|█▉        | 38/200 [00:03<00:13, 11.76it/s]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:04<00:14,  9.89it/s]

Calculating Affiliation Metrics:  32%|███▏      | 63/200 [00:06<00:15,  9.05it/s]

Calculating Affiliation Metrics:  36%|███▋      | 73/200 [00:07<00:14,  8.52it/s]

Calculating Affiliation Metrics:  41%|████      | 82/200 [00:08<00:14,  8.23it/s]

Calculating Affiliation Metrics:  46%|████▌     | 91/200 [00:09<00:13,  8.18it/s]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [00:11<00:11,  8.46it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [00:12<00:09,  9.02it/s]

Calculating Affiliation Metrics:  62%|██████▎   | 125/200 [00:13<00:07, 10.03it/s]

Calculating Affiliation Metrics:  70%|██████▉   | 139/200 [00:14<00:05, 11.08it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 155/200 [00:15<00:03, 12.44it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 175/200 [00:16<00:01, 14.45it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:17<00:00, 11.74it/s]
Table1-only metrics completed in 18.39s
Evaluation Result:  {'AUC-PR': 0.6110534823895907, 'AUC-ROC': 0.9706060856739391, 'VUS-PR': 0.5487932542193538, 'VUS-ROC': 0.9543736935564373, 'Standard-F1': 0.672354329278861, 'Standard-Precision': 0.6602495543672012, 'Standard-Recall': 0.6849112426035501, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9208674235873449, 'Affiliation-P': 0.8899677139091703, 'Affiliation-R': 0.9539899888566801, 'F1_T': 0.6682060390758771, 'Precision_T': 0.6428571428571427, 'Recall_T': 0.695636094674556}
Results for 262_IOPS_id_3_WebService_tr_7028_1st_7128.csv: {'filename': '262_IOPS_id_3_WebService_tr_7028_1st_7128.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 6, 'train_index': '7028', 'data_shape': '125135x1', 'output_length': 120000, 'label_length': 118107, 'aligned_length': 118107, 'AUC-PR': 0.6110534823895907, 'AUC-ROC': 0.9706060856739391, 'VUS-PR': 0.5487932542193538, 'VUS-ROC': 0.9543736935564373, 'Standard-F1': 0.672354329278861, 'Standard-Precision': 0.6602495543672012, 'Standard-Recall': 0.6849112426035501, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9208674235873449, 'Affiliation-P': 0.8899677139091703, 'Affiliation-R': 0.9539899888566801, 'F1_T': 0.6682060390758771, 'Precision_T': 0.6428571428571427, 'Recall_T': 0.695636094674556}
Logits metrics skipped for 262_IOPS_id_3_WebService_tr_7028_1st_7128.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/262_IOPS_id_3_WebService_tr_7028_1st_7128_results.pkl
Processing file: 336_UCR_id_34_Facility_tr_32191_1st_124159.csv, Data shape: (128767, 1)
Label shape: (128767,)
Train data shape: (32191, 1), Test data shape: (96576, 1), Label test shape: (96576,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 336_UCR_id_34_Facility_tr_32191_1st_124159.csv...
Using CUDA device: 2
Data shape: (96576, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 96576, window_size 5000, adding 3424 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([20, 5000, 1])
batched_mask shape: torch.Size([20, 5000])

100%|██████████| 1/1 [00:00<00:00, 61.15it/s]
Original shapes - Output: (100000,), Label: (96576,)
Aligned shapes - Output: (96576,), Label: (96576,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  20%|█▉        | 39/200 [00:01<00:04, 38.32it/s]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [00:02<00:03, 33.61it/s]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [00:03<00:02, 34.84it/s]

Calculating Affiliation Metrics:  80%|████████  | 160/200 [00:04<00:01, 38.52it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:04<00:00, 40.74it/s]
Table1-only metrics completed in 10.96s
Evaluation Result:  {'AUC-PR': 0.18622977108939354, 'AUC-ROC': 0.9497561560004805, 'VUS-PR': 0.1568415346620275, 'VUS-ROC': 0.9511883515240346, 'Standard-F1': 0.26841574167458926, 'Standard-Precision': 0.2301038062283735, 'Standard-Recall': 0.32203389830508433, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7484902290678958, 'Affiliation-P': 0.5983020345008141, 'Affiliation-R': 0.9993514748875592, 'F1_T': 0.2678571428566456, 'Precision_T': 0.24844720496894385, 'Recall_T': 0.2905569007263919}
Results for 336_UCR_id_34_Facility_tr_32191_1st_124159.csv: {'filename': '336_UCR_id_34_Facility_tr_32191_1st_124159.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 243, 'train_index': '32191', 'data_shape': '128767x1', 'output_length': 100000, 'label_length': 96576, 'aligned_length': 96576, 'AUC-PR': 0.18622977108939354, 'AUC-ROC': 0.9497561560004805, 'VUS-PR': 0.1568415346620275, 'VUS-ROC': 0.9511883515240346, 'Standard-F1': 0.26841574167458926, 'Standard-Precision': 0.2301038062283735, 'Standard-Recall': 0.32203389830508433, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7484902290678958, 'Affiliation-P': 0.5983020345008141, 'Affiliation-R': 0.9993514748875592, 'F1_T': 0.2678571428566456, 'Precision_T': 0.24844720496894385, 'Recall_T': 0.2905569007263919}
Logits metrics skipped for 336_UCR_id_34_Facility_tr_32191_1st_124159.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/336_UCR_id_34_Facility_tr_32191_1st_124159_results.pkl
Processing file: 715_YAHOO_id_165_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 715_YAHOO_id_165_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 156.32it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1455.56it/s]
Table1-only metrics completed in 1.22s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 715_YAHOO_id_165_Synthetic_tr_500_1st_658.csv: {'filename': '715_YAHOO_id_165_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 235, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 715_YAHOO_id_165_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/715_YAHOO_id_165_Synthetic_tr_500_1st_658_results.pkl
Processing file: 439_UCR_id_137_HumanActivity_tr_48750_1st_110800.csv, Data shape: (195000, 1)
Label shape: (195000,)
Train data shape: (48750, 1), Test data shape: (146250, 1), Label test shape: (146250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 439_UCR_id_137_HumanActivity_tr_48750_1st_110800.csv...
Using CUDA device: 2
Data shape: (146250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 146250, window_size 5000, adding 3750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([30, 5000, 1])
batched_mask shape: torch.Size([30, 5000])

100%|██████████| 1/1 [00:00<00:00, 46.25it/s]
Original shapes - Output: (150000,), Label: (146250,)
Aligned shapes - Output: (146250,), Label: (146250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  14%|█▎        | 27/200 [00:01<00:06, 25.87it/s]

Calculating Affiliation Metrics:  26%|██▋       | 53/200 [00:02<00:06, 23.12it/s]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [00:03<00:05, 22.71it/s]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [00:04<00:04, 23.02it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [00:05<00:03, 24.02it/s]

Calculating Affiliation Metrics:  80%|████████  | 160/200 [00:06<00:01, 26.94it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:07<00:00, 27.70it/s]
Table1-only metrics completed in 12.44s
Evaluation Result:  {'AUC-PR': 0.00022158209616662973, 'AUC-ROC': 0.9691485069983384, 'VUS-PR': 0.637072717815489, 'VUS-ROC': 0.9937160263079845, 'Standard-F1': 0.0004326195111395196, 'Standard-Precision': 0.00021635655560363476, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7293836441980652, 'Affiliation-P': 0.574041490784348, 'Affiliation-R': 0.9999931623931624, 'F1_T': 0.0003948667324773938, 'Precision_T': 0.0001974723538704581, 'Recall_T': 0.9999999999989999}
Results for 439_UCR_id_137_HumanActivity_tr_48750_1st_110800.csv: {'filename': '439_UCR_id_137_HumanActivity_tr_48750_1st_110800.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '48750', 'data_shape': '195000x1', 'output_length': 150000, 'label_length': 146250, 'aligned_length': 146250, 'AUC-PR': 0.00022158209616662973, 'AUC-ROC': 0.9691485069983384, 'VUS-PR': 0.637072717815489, 'VUS-ROC': 0.9937160263079845, 'Standard-F1': 0.0004326195111395196, 'Standard-Precision': 0.00021635655560363476, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7293836441980652, 'Affiliation-P': 0.574041490784348, 'Affiliation-R': 0.9999931623931624, 'F1_T': 0.0003948667324773938, 'Precision_T': 0.0001974723538704581, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 439_UCR_id_137_HumanActivity_tr_48750_1st_110800.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/439_UCR_id_137_HumanActivity_tr_48750_1st_110800_results.pkl
Processing file: 751_YAHOO_id_201_WebService_tr_500_1st_142.csv, Data shape: (1439, 1)
Label shape: (1439,)
Train data shape: (500, 1), Test data shape: (939, 1), Label test shape: (939,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 751_YAHOO_id_201_WebService_tr_500_1st_142.csv...
Using CUDA device: 2
Data shape: (939, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 939, 1])
batched_mask shape: torch.Size([1, 939])

100%|██████████| 1/1 [00:00<00:00, 149.05it/s]
Original shapes - Output: (939,), Label: (939,)
Aligned shapes - Output: (939,), Label: (939,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1904.05it/s]
Table1-only metrics completed in 0.25s
Evaluation Result:  {'AUC-PR': 0.7099837191667783, 'AUC-ROC': 0.958542413381123, 'VUS-PR': 0.884774738265331, 'VUS-ROC': 0.9938685799256395, 'Standard-F1': 0.631578947367856, 'Standard-Precision': 0.59999999999994, 'Standard-Recall': 0.6666666666665926, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.997339010111761, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9946921443736729, 'F1_T': 0.631578947367856, 'Precision_T': 0.59999999999994, 'Recall_T': 0.6666666666665926}
Results for 751_YAHOO_id_201_WebService_tr_500_1st_142.csv: {'filename': '751_YAHOO_id_201_WebService_tr_500_1st_142.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1439x1', 'output_length': 939, 'label_length': 939, 'aligned_length': 939, 'AUC-PR': 0.7099837191667783, 'AUC-ROC': 0.958542413381123, 'VUS-PR': 0.884774738265331, 'VUS-ROC': 0.9938685799256395, 'Standard-F1': 0.631578947367856, 'Standard-Precision': 0.59999999999994, 'Standard-Recall': 0.6666666666665926, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.997339010111761, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9946921443736729, 'F1_T': 0.631578947367856, 'Precision_T': 0.59999999999994, 'Recall_T': 0.6666666666665926}
Logits metrics skipped for 751_YAHOO_id_201_WebService_tr_500_1st_142.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/751_YAHOO_id_201_WebService_tr_500_1st_142_results.pkl
Processing file: 069_WSD_id_41_WebService_tr_4548_1st_13988.csv, Data shape: (18192, 1)
Label shape: (18192,)
Train data shape: (4548, 1), Test data shape: (13644, 1), Label test shape: (13644,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 069_WSD_id_41_WebService_tr_4548_1st_13988.csv...
Using CUDA device: 2
Data shape: (13644, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13644, window_size 5000, adding 1356 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.21it/s]
Original shapes - Output: (15000,), Label: (13644,)
Aligned shapes - Output: (13644,), Label: (13644,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 316.99it/s]
Table1-only metrics completed in 0.81s
Evaluation Result:  {'AUC-PR': 0.2265209892172869, 'AUC-ROC': 0.8027347166808688, 'VUS-PR': 0.0953276469485786, 'VUS-ROC': 0.8031713790281173, 'Standard-F1': 0.05333333333317431, 'Standard-Precision': 0.02919708029197059, 'Standard-Recall': 0.307692307692284, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666397743178708, 'Affiliation-P': 0.49996974671779154, 'Affiliation-R': 1.0, 'F1_T': 0.05333333333317431, 'Precision_T': 0.02919708029197059, 'Recall_T': 0.307692307692284}
Results for 069_WSD_id_41_WebService_tr_4548_1st_13988.csv: {'filename': '069_WSD_id_41_WebService_tr_4548_1st_13988.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '4548', 'data_shape': '18192x1', 'output_length': 15000, 'label_length': 13644, 'aligned_length': 13644, 'AUC-PR': 0.2265209892172869, 'AUC-ROC': 0.8027347166808688, 'VUS-PR': 0.0953276469485786, 'VUS-ROC': 0.8031713790281173, 'Standard-F1': 0.05333333333317431, 'Standard-Precision': 0.02919708029197059, 'Standard-Recall': 0.307692307692284, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666397743178708, 'Affiliation-P': 0.49996974671779154, 'Affiliation-R': 1.0, 'F1_T': 0.05333333333317431, 'Precision_T': 0.02919708029197059, 'Recall_T': 0.307692307692284}
Logits metrics skipped for 069_WSD_id_41_WebService_tr_4548_1st_13988.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/069_WSD_id_41_WebService_tr_4548_1st_13988_results.pkl
Skipping file: 187_SMD_id_10_Facility_tr_5000_1st_15849.csv due to filter criteria.
Processing file: 100_WSD_id_72_WebService_tr_4559_1st_10254.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 100_WSD_id_72_WebService_tr_4559_1st_10254.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 167.11it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 273.92it/s]
Table1-only metrics completed in 0.90s
Evaluation Result:  {'AUC-PR': 0.1393880912648796, 'AUC-ROC': 0.9660644309092242, 'VUS-PR': 0.064586618593873, 'VUS-ROC': 0.9667285737279848, 'Standard-F1': 0.11180124223590805, 'Standard-Precision': 0.05960264900662243, 'Standard-Recall': 0.8999999999999819, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9321588471604005, 'Affiliation-P': 0.8729479916112842, 'Affiliation-R': 0.9999865526958029, 'F1_T': 0.10917537746795074, 'Precision_T': 0.05795314426633778, 'Recall_T': 0.9399999999999812}
Results for 100_WSD_id_72_WebService_tr_4559_1st_10254.csv: {'filename': '100_WSD_id_72_WebService_tr_4559_1st_10254.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 11, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.1393880912648796, 'AUC-ROC': 0.9660644309092242, 'VUS-PR': 0.064586618593873, 'VUS-ROC': 0.9667285737279848, 'Standard-F1': 0.11180124223590805, 'Standard-Precision': 0.05960264900662243, 'Standard-Recall': 0.8999999999999819, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9321588471604005, 'Affiliation-P': 0.8729479916112842, 'Affiliation-R': 0.9999865526958029, 'F1_T': 0.10917537746795074, 'Precision_T': 0.05795314426633778, 'Recall_T': 0.9399999999999812}
Logits metrics skipped for 100_WSD_id_72_WebService_tr_4559_1st_10254.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/100_WSD_id_72_WebService_tr_4559_1st_10254_results.pkl
Processing file: 569_YAHOO_id_19_WebService_tr_500_1st_30.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 569_YAHOO_id_19_WebService_tr_500_1st_30.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 159.19it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1553.85it/s]
Table1-only metrics completed in 0.38s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.9999444925344662, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9912683954748697, 'Affiliation-P': 0.9826879528985507, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Results for 569_YAHOO_id_19_WebService_tr_500_1st_30.csv: {'filename': '569_YAHOO_id_19_WebService_tr_500_1st_30.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.9999444925344662, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9912683954748697, 'Affiliation-P': 0.9826879528985507, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 569_YAHOO_id_19_WebService_tr_500_1st_30.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/569_YAHOO_id_19_WebService_tr_500_1st_30_results.pkl
Processing file: 440_UCR_id_138_Medical_tr_16700_1st_16800.csv, Data shape: (200000, 1)
Label shape: (200000,)
Train data shape: (16700, 1), Test data shape: (183300, 1), Label test shape: (183300,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 440_UCR_id_138_Medical_tr_16700_1st_16800.csv...
Using CUDA device: 2
Data shape: (183300, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 183300, window_size 5000, adding 1700 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([37, 5000, 1])
batched_mask shape: torch.Size([37, 5000])

100%|██████████| 1/1 [00:00<00:00, 40.07it/s]
Original shapes - Output: (185000,), Label: (183300,)
Aligned shapes - Output: (183300,), Label: (183300,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  11%|█         | 22/200 [00:01<00:08, 21.38it/s]

Calculating Affiliation Metrics:  22%|██▏       | 44/200 [00:02<00:08, 18.71it/s]

Calculating Affiliation Metrics:  32%|███▏      | 63/200 [00:03<00:07, 18.05it/s]

Calculating Affiliation Metrics:  41%|████      | 82/200 [00:04<00:06, 17.87it/s]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [00:05<00:05, 18.15it/s]

Calculating Affiliation Metrics:  60%|██████    | 121/200 [00:06<00:04, 18.66it/s]

Calculating Affiliation Metrics:  72%|███████▏  | 144/200 [00:07<00:02, 19.97it/s]

Calculating Affiliation Metrics:  86%|████████▋ | 173/200 [00:08<00:01, 22.59it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:09<00:00, 21.69it/s]
Table1-only metrics completed in 15.71s
Evaluation Result:  {'AUC-PR': 0.004793120222495215, 'AUC-ROC': 0.8355131967213114, 'VUS-PR': 0.005323753321349645, 'VUS-ROC': 0.8494921384727966, 'Standard-F1': 0.011364754501608886, 'Standard-Precision': 0.0057246233148295, 'Standard-Recall': 0.7699999999999974, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6948036008216953, 'Affiliation-P': 0.532336447381676, 'Affiliation-R': 0.999999963629751, 'F1_T': 0.011356386372298806, 'Precision_T': 0.0057889944061402365, 'Recall_T': 0.29666666666666563}
Results for 440_UCR_id_138_Medical_tr_16700_1st_16800.csv: {'filename': '440_UCR_id_138_Medical_tr_16700_1st_16800.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '16700', 'data_shape': '200000x1', 'output_length': 185000, 'label_length': 183300, 'aligned_length': 183300, 'AUC-PR': 0.004793120222495215, 'AUC-ROC': 0.8355131967213114, 'VUS-PR': 0.005323753321349645, 'VUS-ROC': 0.8494921384727966, 'Standard-F1': 0.011364754501608886, 'Standard-Precision': 0.0057246233148295, 'Standard-Recall': 0.7699999999999974, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6948036008216953, 'Affiliation-P': 0.532336447381676, 'Affiliation-R': 0.999999963629751, 'F1_T': 0.011356386372298806, 'Precision_T': 0.0057889944061402365, 'Recall_T': 0.29666666666666563}
Logits metrics skipped for 440_UCR_id_138_Medical_tr_16700_1st_16800.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/440_UCR_id_138_Medical_tr_16700_1st_16800_results.pkl
Processing file: 011_NAB_id_11_Facility_tr_1007_1st_1526.csv, Data shape: (4031, 1)
Label shape: (4031,)
Train data shape: (1007, 1), Test data shape: (3024, 1), Label test shape: (3024,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 011_NAB_id_11_Facility_tr_1007_1st_1526.csv...
Using CUDA device: 2
Data shape: (3024, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 3024, 1])
batched_mask shape: torch.Size([1, 3024])

100%|██████████| 1/1 [00:00<00:00, 147.33it/s]
Original shapes - Output: (3024,), Label: (3024,)
Aligned shapes - Output: (3024,), Label: (3024,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1182.56it/s]
Table1-only metrics completed in 0.97s
Evaluation Result:  {'AUC-PR': 0.4600058910525907, 'AUC-ROC': 0.6523372218085546, 'VUS-PR': 0.4706830882292029, 'VUS-ROC': 0.6784018555528253, 'Standard-F1': 0.4751203852323793, 'Standard-Precision': 0.9932885906040202, 'Standard-Recall': 0.3122362869198306, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.990080122426828, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9803551196615542, 'F1_T': 0.47435897435860763, 'Precision_T': 0.9866666666666601, 'Recall_T': 0.3122362869198306}
Results for 011_NAB_id_11_Facility_tr_1007_1st_1526.csv: {'filename': '011_NAB_id_11_Facility_tr_1007_1st_1526.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 204, 'train_index': '1007', 'data_shape': '4031x1', 'output_length': 3024, 'label_length': 3024, 'aligned_length': 3024, 'AUC-PR': 0.4600058910525907, 'AUC-ROC': 0.6523372218085546, 'VUS-PR': 0.4706830882292029, 'VUS-ROC': 0.6784018555528253, 'Standard-F1': 0.4751203852323793, 'Standard-Precision': 0.9932885906040202, 'Standard-Recall': 0.3122362869198306, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.990080122426828, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9803551196615542, 'F1_T': 0.47435897435860763, 'Precision_T': 0.9866666666666601, 'Recall_T': 0.3122362869198306}
Logits metrics skipped for 011_NAB_id_11_Facility_tr_1007_1st_1526.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/011_NAB_id_11_Facility_tr_1007_1st_1526_results.pkl
Processing file: 400_UCR_id_98_Environment_tr_2046_1st_5391.csv, Data shape: (8184, 1)
Label shape: (8184,)
Train data shape: (2046, 1), Test data shape: (6138, 1), Label test shape: (6138,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 400_UCR_id_98_Environment_tr_2046_1st_5391.csv...
Using CUDA device: 2
Data shape: (6138, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6138, window_size 5000, adding 3862 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 164.06it/s]
Original shapes - Output: (10000,), Label: (6138,)
Aligned shapes - Output: (6138,), Label: (6138,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 496.04it/s]
Table1-only metrics completed in 0.53s
Evaluation Result:  {'AUC-PR': 0.004424778761061947, 'AUC-ROC': 0.9633371354081799, 'VUS-PR': 0.025034744998869644, 'VUS-ROC': 0.9836585359548589, 'Standard-F1': 0.008620689655163757, 'Standard-Precision': 0.004329004329004311, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7305006032955564, 'Affiliation-P': 0.5754241437151526, 'Affiliation-R': 1.0, 'F1_T': 0.008196721311467177, 'Precision_T': 0.004115226337448543, 'Recall_T': 0.9999999999989999}
Results for 400_UCR_id_98_Environment_tr_2046_1st_5391.csv: {'filename': '400_UCR_id_98_Environment_tr_2046_1st_5391.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '2046', 'data_shape': '8184x1', 'output_length': 10000, 'label_length': 6138, 'aligned_length': 6138, 'AUC-PR': 0.004424778761061947, 'AUC-ROC': 0.9633371354081799, 'VUS-PR': 0.025034744998869644, 'VUS-ROC': 0.9836585359548589, 'Standard-F1': 0.008620689655163757, 'Standard-Precision': 0.004329004329004311, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7305006032955564, 'Affiliation-P': 0.5754241437151526, 'Affiliation-R': 1.0, 'F1_T': 0.008196721311467177, 'Precision_T': 0.004115226337448543, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 400_UCR_id_98_Environment_tr_2046_1st_5391.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/400_UCR_id_98_Environment_tr_2046_1st_5391_results.pkl
Skipping file: 830_Exathlon_id_21_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 111_WSD_id_83_WebService_tr_3934_1st_7531.csv, Data shape: (15736, 1)
Label shape: (15736,)
Train data shape: (3934, 1), Test data shape: (11802, 1), Label test shape: (11802,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 111_WSD_id_83_WebService_tr_3934_1st_7531.csv...
Using CUDA device: 2
Data shape: (11802, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 11802, window_size 5000, adding 3198 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.05it/s]
Original shapes - Output: (15000,), Label: (11802,)
Aligned shapes - Output: (11802,), Label: (11802,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 360.33it/s]
Table1-only metrics completed in 0.73s
Evaluation Result:  {'AUC-PR': 0.33411726533185493, 'AUC-ROC': 0.8505805576743791, 'VUS-PR': 0.3340995052338893, 'VUS-ROC': 0.8501285419668329, 'Standard-F1': 0.016393442622902585, 'Standard-Precision': 0.008403361344537745, 'Standard-Recall': 0.33333333333322224, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9192533693788234, 'Affiliation-P': 0.8506133770234072, 'Affiliation-R': 0.9999435123990285, 'F1_T': 0.016393442622902585, 'Precision_T': 0.008403361344537745, 'Recall_T': 0.33333333333322224}
Results for 111_WSD_id_83_WebService_tr_3934_1st_7531.csv: {'filename': '111_WSD_id_83_WebService_tr_3934_1st_7531.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 29, 'train_index': '3934', 'data_shape': '15736x1', 'output_length': 15000, 'label_length': 11802, 'aligned_length': 11802, 'AUC-PR': 0.33411726533185493, 'AUC-ROC': 0.8505805576743791, 'VUS-PR': 0.3340995052338893, 'VUS-ROC': 0.8501285419668329, 'Standard-F1': 0.016393442622902585, 'Standard-Precision': 0.008403361344537745, 'Standard-Recall': 0.33333333333322224, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9192533693788234, 'Affiliation-P': 0.8506133770234072, 'Affiliation-R': 0.9999435123990285, 'F1_T': 0.016393442622902585, 'Precision_T': 0.008403361344537745, 'Recall_T': 0.33333333333322224}
Logits metrics skipped for 111_WSD_id_83_WebService_tr_3934_1st_7531.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/111_WSD_id_83_WebService_tr_3934_1st_7531_results.pkl
Processing file: 326_UCR_id_24_Medical_tr_1875_1st_4187.csv, Data shape: (7501, 1)
Label shape: (7501,)
Train data shape: (1875, 1), Test data shape: (5626, 1), Label test shape: (5626,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 326_UCR_id_24_Medical_tr_1875_1st_4187.csv...
Using CUDA device: 2
Data shape: (5626, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5626, window_size 5000, adding 4374 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 164.77it/s]
Original shapes - Output: (10000,), Label: (5626,)
Aligned shapes - Output: (5626,), Label: (5626,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 654.18it/s]
Table1-only metrics completed in 0.99s
Evaluation Result:  {'AUC-PR': 0.8674674862174863, 'AUC-ROC': 0.9997921862011637, 'VUS-PR': 0.6385720793692566, 'VUS-ROC': 0.9992595059180761, 'Standard-F1': 0.3478260869562243, 'Standard-Precision': 0.21052631578946998, 'Standard-Recall': 0.9999999999999166, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7730060426936796, 'Affiliation-P': 0.6299998774164286, 'Affiliation-R': 1.0, 'F1_T': 0.3478260869562243, 'Precision_T': 0.21052631578946998, 'Recall_T': 0.9999999999999166}
Results for 326_UCR_id_24_Medical_tr_1875_1st_4187.csv: {'filename': '326_UCR_id_24_Medical_tr_1875_1st_4187.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 183, 'train_index': '1875', 'data_shape': '7501x1', 'output_length': 10000, 'label_length': 5626, 'aligned_length': 5626, 'AUC-PR': 0.8674674862174863, 'AUC-ROC': 0.9997921862011637, 'VUS-PR': 0.6385720793692566, 'VUS-ROC': 0.9992595059180761, 'Standard-F1': 0.3478260869562243, 'Standard-Precision': 0.21052631578946998, 'Standard-Recall': 0.9999999999999166, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7730060426936796, 'Affiliation-P': 0.6299998774164286, 'Affiliation-R': 1.0, 'F1_T': 0.3478260869562243, 'Precision_T': 0.21052631578946998, 'Recall_T': 0.9999999999999166}
Logits metrics skipped for 326_UCR_id_24_Medical_tr_1875_1st_4187.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/326_UCR_id_24_Medical_tr_1875_1st_4187_results.pkl
Processing file: 769_YAHOO_id_219_WebService_tr_500_1st_205.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 769_YAHOO_id_219_WebService_tr_500_1st_205.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 165.65it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1048.10it/s]
Table1-only metrics completed in 0.54s
Evaluation Result:  {'AUC-PR': 0.9999999999999998, 'AUC-ROC': 1.0, 'VUS-PR': 0.9403737455954969, 'VUS-ROC': 0.9997457084024357, 'Standard-F1': 0.7368421052626148, 'Standard-Precision': 0.5833333333332847, 'Standard-Recall': 0.9999999999998571, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9959318464630988, 'Affiliation-P': 0.9918966585633252, 'Affiliation-R': 1.0, 'F1_T': 0.7368421052626148, 'Precision_T': 0.5833333333332847, 'Recall_T': 0.9999999999998571}
Results for 769_YAHOO_id_219_WebService_tr_500_1st_205.csv: {'filename': '769_YAHOO_id_219_WebService_tr_500_1st_205.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9999999999999998, 'AUC-ROC': 1.0, 'VUS-PR': 0.9403737455954969, 'VUS-ROC': 0.9997457084024357, 'Standard-F1': 0.7368421052626148, 'Standard-Precision': 0.5833333333332847, 'Standard-Recall': 0.9999999999998571, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9959318464630988, 'Affiliation-P': 0.9918966585633252, 'Affiliation-R': 1.0, 'F1_T': 0.7368421052626148, 'Precision_T': 0.5833333333332847, 'Recall_T': 0.9999999999998571}
Logits metrics skipped for 769_YAHOO_id_219_WebService_tr_500_1st_205.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/769_YAHOO_id_219_WebService_tr_500_1st_205_results.pkl
Processing file: 369_UCR_id_67_Sensor_tr_2750_1st_5400.csv, Data shape: (11000, 1)
Label shape: (11000,)
Train data shape: (2750, 1), Test data shape: (8250, 1), Label test shape: (8250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 369_UCR_id_67_Sensor_tr_2750_1st_5400.csv...
Using CUDA device: 2
Data shape: (8250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 8250, window_size 5000, adding 1750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 168.16it/s]
Original shapes - Output: (10000,), Label: (8250,)
Aligned shapes - Output: (8250,), Label: (8250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 368.03it/s]
Table1-only metrics completed in 0.90s
Evaluation Result:  {'AUC-PR': 0.03650230745905811, 'AUC-ROC': 0.6639819875776397, 'VUS-PR': 0.07793451841775985, 'VUS-ROC': 0.7072709092544094, 'Standard-F1': 0.07837582625109035, 'Standard-Precision': 0.04112983151635281, 'Standard-Recall': 0.8299999999999959, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8134166547148286, 'Affiliation-P': 0.6971428571428572, 'Affiliation-R': 0.97624, 'F1_T': 0.07827616534732133, 'Precision_T': 0.04093836246550137, 'Recall_T': 0.8899999999999956}
Results for 369_UCR_id_67_Sensor_tr_2750_1st_5400.csv: {'filename': '369_UCR_id_67_Sensor_tr_2750_1st_5400.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 81, 'train_index': '2750', 'data_shape': '11000x1', 'output_length': 10000, 'label_length': 8250, 'aligned_length': 8250, 'AUC-PR': 0.03650230745905811, 'AUC-ROC': 0.6639819875776397, 'VUS-PR': 0.07793451841775985, 'VUS-ROC': 0.7072709092544094, 'Standard-F1': 0.07837582625109035, 'Standard-Precision': 0.04112983151635281, 'Standard-Recall': 0.8299999999999959, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8134166547148286, 'Affiliation-P': 0.6971428571428572, 'Affiliation-R': 0.97624, 'F1_T': 0.07827616534732133, 'Precision_T': 0.04093836246550137, 'Recall_T': 0.8899999999999956}
Logits metrics skipped for 369_UCR_id_67_Sensor_tr_2750_1st_5400.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/369_UCR_id_67_Sensor_tr_2750_1st_5400_results.pkl
Skipping file: 549_SMAP_id_19_Sensor_tr_1998_1st_2098.csv due to filter criteria.
Skipping file: 544_SMAP_id_14_Sensor_tr_1978_1st_4450.csv due to filter criteria.
Processing file: 737_YAHOO_id_187_WebService_tr_500_1st_230.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 737_YAHOO_id_187_WebService_tr_500_1st_230.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 160.18it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1256.10it/s]
Table1-only metrics completed in 0.41s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.9999444925344662, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990571103505619, 'Affiliation-P': 0.9981159971079513, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Results for 737_YAHOO_id_187_WebService_tr_500_1st_230.csv: {'filename': '737_YAHOO_id_187_WebService_tr_500_1st_230.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.9999444925344662, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990571103505619, 'Affiliation-P': 0.9981159971079513, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 737_YAHOO_id_187_WebService_tr_500_1st_230.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/737_YAHOO_id_187_WebService_tr_500_1st_230_results.pkl
Processing file: 107_WSD_id_79_WebService_tr_1095_1st_1195.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (1095, 1), Test data shape: (17141, 1), Label test shape: (17141,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 107_WSD_id_79_WebService_tr_1095_1st_1195.csv...
Using CUDA device: 2
Data shape: (17141, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 17141, window_size 5000, adding 2859 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.33it/s]
Original shapes - Output: (20000,), Label: (17141,)
Aligned shapes - Output: (17141,), Label: (17141,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  93%|█████████▎| 186/200 [00:01<00:00, 185.84it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 191.91it/s]
Table1-only metrics completed in 3.08s
Evaluation Result:  {'AUC-PR': 0.5398232514438462, 'AUC-ROC': 0.9585979871171116, 'VUS-PR': 0.5312479007392606, 'VUS-ROC': 0.9775014530185913, 'Standard-F1': 0.5967302452311121, 'Standard-Precision': 0.5367647058823516, 'Standard-Recall': 0.6717791411042924, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9420930955812463, 'Affiliation-P': 0.8910786873878426, 'Affiliation-R': 0.9993033864178669, 'F1_T': 0.5938748335547668, 'Precision_T': 0.5247058823529399, 'Recall_T': 0.684049079754599}
Results for 107_WSD_id_79_WebService_tr_1095_1st_1195.csv: {'filename': '107_WSD_id_79_WebService_tr_1095_1st_1195.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '1095', 'data_shape': '18236x1', 'output_length': 20000, 'label_length': 17141, 'aligned_length': 17141, 'AUC-PR': 0.5398232514438462, 'AUC-ROC': 0.9585979871171116, 'VUS-PR': 0.5312479007392606, 'VUS-ROC': 0.9775014530185913, 'Standard-F1': 0.5967302452311121, 'Standard-Precision': 0.5367647058823516, 'Standard-Recall': 0.6717791411042924, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9420930955812463, 'Affiliation-P': 0.8910786873878426, 'Affiliation-R': 0.9993033864178669, 'F1_T': 0.5938748335547668, 'Precision_T': 0.5247058823529399, 'Recall_T': 0.684049079754599}
Logits metrics skipped for 107_WSD_id_79_WebService_tr_1095_1st_1195.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/107_WSD_id_79_WebService_tr_1095_1st_1195_results.pkl
Processing file: 728_YAHOO_id_178_WebService_tr_500_1st_68.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 728_YAHOO_id_178_WebService_tr_500_1st_68.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 163.66it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1278.62it/s]
Table1-only metrics completed in 0.48s
Evaluation Result:  {'AUC-PR': 0.6610037227949598, 'AUC-ROC': 0.9808347529812607, 'VUS-PR': 0.44276093907268355, 'VUS-ROC': 0.9755226981718771, 'Standard-F1': 0.4444444444439506, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.6666666666665555, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9185821680501621, 'Affiliation-P': 0.8494239145233293, 'Affiliation-R': 1.0, 'F1_T': 0.4444444444439506, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.6666666666665555}
Results for 728_YAHOO_id_178_WebService_tr_500_1st_68.csv: {'filename': '728_YAHOO_id_178_WebService_tr_500_1st_68.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.6610037227949598, 'AUC-ROC': 0.9808347529812607, 'VUS-PR': 0.44276093907268355, 'VUS-ROC': 0.9755226981718771, 'Standard-F1': 0.4444444444439506, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.6666666666665555, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9185821680501621, 'Affiliation-P': 0.8494239145233293, 'Affiliation-R': 1.0, 'F1_T': 0.4444444444439506, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.6666666666665555}
Logits metrics skipped for 728_YAHOO_id_178_WebService_tr_500_1st_68.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/728_YAHOO_id_178_WebService_tr_500_1st_68_results.pkl
Processing file: 619_YAHOO_id_69_WebService_tr_500_1st_8.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 619_YAHOO_id_69_WebService_tr_500_1st_8.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.45it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1292.89it/s]
Table1-only metrics completed in 0.36s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9504941880689357, 'VUS-ROC': 0.9998948142040058, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9977802441731404, 'Affiliation-P': 0.9955703211517165, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Results for 619_YAHOO_id_69_WebService_tr_500_1st_8.csv: {'filename': '619_YAHOO_id_69_WebService_tr_500_1st_8.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9504941880689357, 'VUS-ROC': 0.9998948142040058, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9977802441731404, 'Affiliation-P': 0.9955703211517165, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Logits metrics skipped for 619_YAHOO_id_69_WebService_tr_500_1st_8.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/619_YAHOO_id_69_WebService_tr_500_1st_8_results.pkl
Processing file: 324_UCR_id_22_Medical_tr_1875_1st_5865.csv, Data shape: (7501, 1)
Label shape: (7501,)
Train data shape: (1875, 1), Test data shape: (5626, 1), Label test shape: (5626,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 324_UCR_id_22_Medical_tr_1875_1st_5865.csv...
Using CUDA device: 2
Data shape: (5626, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5626, window_size 5000, adding 4374 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 166.05it/s]
Original shapes - Output: (10000,), Label: (5626,)
Aligned shapes - Output: (5626,), Label: (5626,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 709.15it/s]
Table1-only metrics completed in 0.90s
Evaluation Result:  {'AUC-PR': 0.7484204518469434, 'AUC-ROC': 0.9919930556594879, 'VUS-PR': 0.9085577134159801, 'VUS-ROC': 0.9977181003096759, 'Standard-F1': 0.6489361702122718, 'Standard-Precision': 0.7721518987341675, 'Standard-Recall': 0.5596330275229306, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9961317528578041, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9922933170698297, 'F1_T': 0.6321243523311081, 'Precision_T': 0.7261904761904676, 'Recall_T': 0.5596330275229306}
Results for 324_UCR_id_22_Medical_tr_1875_1st_5865.csv: {'filename': '324_UCR_id_22_Medical_tr_1875_1st_5865.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 165, 'train_index': '1875', 'data_shape': '7501x1', 'output_length': 10000, 'label_length': 5626, 'aligned_length': 5626, 'AUC-PR': 0.7484204518469434, 'AUC-ROC': 0.9919930556594879, 'VUS-PR': 0.9085577134159801, 'VUS-ROC': 0.9977181003096759, 'Standard-F1': 0.6489361702122718, 'Standard-Precision': 0.7721518987341675, 'Standard-Recall': 0.5596330275229306, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9961317528578041, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9922933170698297, 'F1_T': 0.6321243523311081, 'Precision_T': 0.7261904761904676, 'Recall_T': 0.5596330275229306}
Logits metrics skipped for 324_UCR_id_22_Medical_tr_1875_1st_5865.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/324_UCR_id_22_Medical_tr_1875_1st_5865_results.pkl
Processing file: 676_YAHOO_id_126_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 676_YAHOO_id_126_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 160.35it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1954.82it/s]
Table1-only metrics completed in 0.71s
Evaluation Result:  {'AUC-PR': 0.9328843995510663, 'AUC-ROC': 0.9992690058479533, 'VUS-PR': 0.9930845166427625, 'VUS-ROC': 0.9999361563702144, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 676_YAHOO_id_126_Synthetic_tr_500_1st_623.csv: {'filename': '676_YAHOO_id_126_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 103, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9328843995510663, 'AUC-ROC': 0.9992690058479533, 'VUS-PR': 0.9930845166427625, 'VUS-ROC': 0.9999361563702144, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 676_YAHOO_id_126_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/676_YAHOO_id_126_Synthetic_tr_500_1st_623_results.pkl
Skipping file: 860_OPPORTUNITY_id_19_HumanActivity_tr_808_1st_908.csv due to filter criteria.
Processing file: 505_UCR_id_203_HumanActivity_tr_16000_1st_33070.csv, Data shape: (64000, 1)
Label shape: (64000,)
Train data shape: (16000, 1), Test data shape: (48000, 1), Label test shape: (48000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 505_UCR_id_203_HumanActivity_tr_16000_1st_33070.csv...
Using CUDA device: 2
Data shape: (48000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 48000, window_size 5000, adding 2000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([10, 5000, 1])
batched_mask shape: torch.Size([10, 5000])

100%|██████████| 1/1 [00:00<00:00, 97.40it/s]
Original shapes - Output: (50000,), Label: (48000,)
Aligned shapes - Output: (48000,), Label: (48000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  41%|████      | 82/200 [00:01<00:01, 81.55it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 176/200 [00:02<00:00, 88.43it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 92.13it/s]
Table1-only metrics completed in 3.29s
Evaluation Result:  {'AUC-PR': 0.00610469828094513, 'AUC-ROC': 0.8178638546669451, 'VUS-PR': 0.007724288580911773, 'VUS-ROC': 0.8404232985240885, 'Standard-F1': 0.016036655211692657, 'Standard-Precision': 0.009174311926605493, 'Standard-Recall': 0.06363636363636306, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6743638411651324, 'Affiliation-P': 0.5087096007986438, 'Affiliation-R': 1.0, 'F1_T': 0.016928657799043803, 'Precision_T': 0.009762900976290083, 'Recall_T': 0.06363636363636306}
Results for 505_UCR_id_203_HumanActivity_tr_16000_1st_33070.csv: {'filename': '505_UCR_id_203_HumanActivity_tr_16000_1st_33070.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '16000', 'data_shape': '64000x1', 'output_length': 50000, 'label_length': 48000, 'aligned_length': 48000, 'AUC-PR': 0.00610469828094513, 'AUC-ROC': 0.8178638546669451, 'VUS-PR': 0.007724288580911773, 'VUS-ROC': 0.8404232985240885, 'Standard-F1': 0.016036655211692657, 'Standard-Precision': 0.009174311926605493, 'Standard-Recall': 0.06363636363636306, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6743638411651324, 'Affiliation-P': 0.5087096007986438, 'Affiliation-R': 1.0, 'F1_T': 0.016928657799043803, 'Precision_T': 0.009762900976290083, 'Recall_T': 0.06363636363636306}
Logits metrics skipped for 505_UCR_id_203_HumanActivity_tr_16000_1st_33070.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/505_UCR_id_203_HumanActivity_tr_16000_1st_33070_results.pkl
Processing file: 632_YAHOO_id_82_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 632_YAHOO_id_82_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 159.72it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2114.83it/s]
Table1-only metrics completed in 0.71s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 632_YAHOO_id_82_Synthetic_tr_500_1st_658.csv: {'filename': '632_YAHOO_id_82_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 113, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 632_YAHOO_id_82_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/632_YAHOO_id_82_Synthetic_tr_500_1st_658_results.pkl
Processing file: 434_UCR_id_132_Medical_tr_14750_1st_57530.csv, Data shape: (59000, 1)
Label shape: (59000,)
Train data shape: (14750, 1), Test data shape: (44250, 1), Label test shape: (44250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 434_UCR_id_132_Medical_tr_14750_1st_57530.csv...
Using CUDA device: 2
Data shape: (44250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 44250, window_size 5000, adding 750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([9, 5000, 1])
batched_mask shape: torch.Size([9, 5000])

100%|██████████| 1/1 [00:00<00:00, 112.18it/s]
Original shapes - Output: (45000,), Label: (44250,)
Aligned shapes - Output: (44250,), Label: (44250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  43%|████▎     | 86/200 [00:01<00:01, 85.48it/s]

Calculating Affiliation Metrics:  94%|█████████▍| 188/200 [00:02<00:00, 95.10it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 96.58it/s]
Table1-only metrics completed in 2.28s
Evaluation Result:  {'AUC-PR': 0.027234060264518575, 'AUC-ROC': 0.9032124871037124, 'VUS-PR': 0.02779183107778027, 'VUS-ROC': 0.9032189171703939, 'Standard-F1': 0.06547359231769236, 'Standard-Precision': 0.03402389233328292, 'Standard-Recall': 0.865384615384612, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8290246641740977, 'Affiliation-P': 0.7089176273363595, 'Affiliation-R': 0.9981311603650588, 'F1_T': 0.06474190726150543, 'Precision_T': 0.03391384051329056, 'Recall_T': 0.7115384615384588}
Results for 434_UCR_id_132_Medical_tr_14750_1st_57530.csv: {'filename': '434_UCR_id_132_Medical_tr_14750_1st_57530.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 6, 'train_index': '14750', 'data_shape': '59000x1', 'output_length': 45000, 'label_length': 44250, 'aligned_length': 44250, 'AUC-PR': 0.027234060264518575, 'AUC-ROC': 0.9032124871037124, 'VUS-PR': 0.02779183107778027, 'VUS-ROC': 0.9032189171703939, 'Standard-F1': 0.06547359231769236, 'Standard-Precision': 0.03402389233328292, 'Standard-Recall': 0.865384615384612, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8290246641740977, 'Affiliation-P': 0.7089176273363595, 'Affiliation-R': 0.9981311603650588, 'F1_T': 0.06474190726150543, 'Precision_T': 0.03391384051329056, 'Recall_T': 0.7115384615384588}
Logits metrics skipped for 434_UCR_id_132_Medical_tr_14750_1st_57530.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/434_UCR_id_132_Medical_tr_14750_1st_57530_results.pkl
Processing file: 510_UCR_id_208_Facility_tr_50000_1st_126920.csv, Data shape: (300000, 1)
Label shape: (300000,)
Train data shape: (50000, 1), Test data shape: (250000, 1), Label test shape: (250000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 510_UCR_id_208_Facility_tr_50000_1st_126920.csv...
Using CUDA device: 2
Data shape: (250000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([50, 5000, 1])
batched_mask shape: torch.Size([50, 5000])

100%|██████████| 1/1 [00:00<00:00, 28.29it/s]
Original shapes - Output: (250000,), Label: (250000,)
Aligned shapes - Output: (250000,), Label: (250000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  10%|▉         | 19/200 [00:01<00:09, 18.13it/s]

Calculating Affiliation Metrics:  19%|█▉        | 38/200 [00:02<00:10, 15.42it/s]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:03<00:10, 14.22it/s]

Calculating Affiliation Metrics:  34%|███▍      | 69/200 [00:04<00:09, 13.73it/s]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [00:05<00:08, 13.49it/s]

Calculating Affiliation Metrics:  48%|████▊     | 97/200 [00:06<00:07, 13.56it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 111/200 [00:07<00:06, 13.67it/s]

Calculating Affiliation Metrics:  64%|██████▍   | 128/200 [00:08<00:04, 14.52it/s]

Calculating Affiliation Metrics:  74%|███████▎  | 147/200 [00:09<00:03, 15.69it/s]

Calculating Affiliation Metrics:  84%|████████▍ | 169/200 [00:11<00:01, 17.40it/s]

Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:12<00:00, 21.41it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:12<00:00, 16.63it/s]
Table1-only metrics completed in 20.59s
Evaluation Result:  {'AUC-PR': 0.9230053723544528, 'AUC-ROC': 0.9996182372715332, 'VUS-PR': 0.43396973349848905, 'VUS-ROC': 0.9986804805450961, 'Standard-F1': 0.3010169491522836, 'Standard-Precision': 0.17759999999999992, 'Standard-Recall': 0.9866666666666645, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7949814219017882, 'Affiliation-P': 0.6597254988066826, 'Affiliation-R': 0.9999998844444443, 'F1_T': 0.3010169491522836, 'Precision_T': 0.17759999999999992, 'Recall_T': 0.9866666666666645}
Results for 510_UCR_id_208_Facility_tr_50000_1st_126920.csv: {'filename': '510_UCR_id_208_Facility_tr_50000_1st_126920.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 124, 'train_index': '50000', 'data_shape': '300000x1', 'output_length': 250000, 'label_length': 250000, 'aligned_length': 250000, 'AUC-PR': 0.9230053723544528, 'AUC-ROC': 0.9996182372715332, 'VUS-PR': 0.43396973349848905, 'VUS-ROC': 0.9986804805450961, 'Standard-F1': 0.3010169491522836, 'Standard-Precision': 0.17759999999999992, 'Standard-Recall': 0.9866666666666645, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7949814219017882, 'Affiliation-P': 0.6597254988066826, 'Affiliation-R': 0.9999998844444443, 'F1_T': 0.3010169491522836, 'Precision_T': 0.17759999999999992, 'Recall_T': 0.9866666666666645}
Logits metrics skipped for 510_UCR_id_208_Facility_tr_50000_1st_126920.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/510_UCR_id_208_Facility_tr_50000_1st_126920_results.pkl
Processing file: 082_WSD_id_54_WebService_tr_4437_1st_4805.csv, Data shape: (17748, 1)
Label shape: (17748,)
Train data shape: (4437, 1), Test data shape: (13311, 1), Label test shape: (13311,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 082_WSD_id_54_WebService_tr_4437_1st_4805.csv...
Using CUDA device: 2
Data shape: (13311, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13311, window_size 5000, adding 1689 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 152.69it/s]
Original shapes - Output: (15000,), Label: (13311,)
Aligned shapes - Output: (13311,), Label: (13311,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 255.63it/s]
Table1-only metrics completed in 0.97s
Evaluation Result:  {'AUC-PR': 0.9444258727050526, 'AUC-ROC': 0.9955685875377631, 'VUS-PR': 0.9397357949661279, 'VUS-ROC': 0.9982008120675218, 'Standard-F1': 0.9163636363631301, 'Standard-Precision': 0.9402985074626796, 'Standard-Recall': 0.8936170212765895, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9993111223006812, 'Affiliation-P': 0.9989413775178477, 'Affiliation-R': 0.999681140897031, 'F1_T': 0.9163636363631301, 'Precision_T': 0.9402985074626796, 'Recall_T': 0.8936170212765895}
Results for 082_WSD_id_54_WebService_tr_4437_1st_4805.csv: {'filename': '082_WSD_id_54_WebService_tr_4437_1st_4805.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 9, 'train_index': '4437', 'data_shape': '17748x1', 'output_length': 15000, 'label_length': 13311, 'aligned_length': 13311, 'AUC-PR': 0.9444258727050526, 'AUC-ROC': 0.9955685875377631, 'VUS-PR': 0.9397357949661279, 'VUS-ROC': 0.9982008120675218, 'Standard-F1': 0.9163636363631301, 'Standard-Precision': 0.9402985074626796, 'Standard-Recall': 0.8936170212765895, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9993111223006812, 'Affiliation-P': 0.9989413775178477, 'Affiliation-R': 0.999681140897031, 'F1_T': 0.9163636363631301, 'Precision_T': 0.9402985074626796, 'Recall_T': 0.8936170212765895}
Logits metrics skipped for 082_WSD_id_54_WebService_tr_4437_1st_4805.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/082_WSD_id_54_WebService_tr_4437_1st_4805_results.pkl
Processing file: 770_YAHOO_id_220_WebService_tr_500_1st_29.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 770_YAHOO_id_220_WebService_tr_500_1st_29.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 160.86it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1036.50it/s]
Table1-only metrics completed in 0.58s
Evaluation Result:  {'AUC-PR': 0.9999999999999998, 'AUC-ROC': 1.0, 'VUS-PR': 0.9797147347885332, 'VUS-ROC': 0.9999134919304835, 'Standard-F1': 0.7368421052626148, 'Standard-Precision': 0.5833333333332847, 'Standard-Recall': 0.9999999999998571, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9966827377453832, 'Affiliation-P': 0.9933874111820586, 'Affiliation-R': 1.0, 'F1_T': 0.7368421052626148, 'Precision_T': 0.5833333333332847, 'Recall_T': 0.9999999999998571}
Results for 770_YAHOO_id_220_WebService_tr_500_1st_29.csv: {'filename': '770_YAHOO_id_220_WebService_tr_500_1st_29.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9999999999999998, 'AUC-ROC': 1.0, 'VUS-PR': 0.9797147347885332, 'VUS-ROC': 0.9999134919304835, 'Standard-F1': 0.7368421052626148, 'Standard-Precision': 0.5833333333332847, 'Standard-Recall': 0.9999999999998571, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9966827377453832, 'Affiliation-P': 0.9933874111820586, 'Affiliation-R': 1.0, 'F1_T': 0.7368421052626148, 'Precision_T': 0.5833333333332847, 'Recall_T': 0.9999999999998571}
Logits metrics skipped for 770_YAHOO_id_220_WebService_tr_500_1st_29.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/770_YAHOO_id_220_WebService_tr_500_1st_29_results.pkl
Processing file: 670_YAHOO_id_120_WebService_tr_500_1st_221.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 670_YAHOO_id_120_WebService_tr_500_1st_221.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 163.96it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 797.74it/s]
Table1-only metrics completed in 0.82s
Evaluation Result:  {'AUC-PR': 0.9757575757575758, 'AUC-ROC': 0.9996889338206704, 'VUS-PR': 0.9938383891815036, 'VUS-ROC': 0.999960445171973, 'Standard-F1': 0.8695652173907297, 'Standard-Precision': 0.8333333333332639, 'Standard-Recall': 0.9090909090908265, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9935367124529142, 'Affiliation-P': 0.9871564365495391, 'Affiliation-R': 1.0, 'F1_T': 0.8695652173907297, 'Precision_T': 0.8333333333332639, 'Recall_T': 0.9090909090908265}
Results for 670_YAHOO_id_120_WebService_tr_500_1st_221.csv: {'filename': '670_YAHOO_id_120_WebService_tr_500_1st_221.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9757575757575758, 'AUC-ROC': 0.9996889338206704, 'VUS-PR': 0.9938383891815036, 'VUS-ROC': 0.999960445171973, 'Standard-F1': 0.8695652173907297, 'Standard-Precision': 0.8333333333332639, 'Standard-Recall': 0.9090909090908265, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9935367124529142, 'Affiliation-P': 0.9871564365495391, 'Affiliation-R': 1.0, 'F1_T': 0.8695652173907297, 'Precision_T': 0.8333333333332639, 'Recall_T': 0.9090909090908265}
Logits metrics skipped for 670_YAHOO_id_120_WebService_tr_500_1st_221.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/670_YAHOO_id_120_WebService_tr_500_1st_221_results.pkl
Processing file: 748_YAHOO_id_198_WebService_tr_500_1st_170.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 748_YAHOO_id_198_WebService_tr_500_1st_170.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.59it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1614.04it/s]
Table1-only metrics completed in 0.28s
Evaluation Result:  {'AUC-PR': 0.75, 'AUC-ROC': 0.999151103565365, 'VUS-PR': 0.5500000000000003, 'VUS-ROC': 0.9990449915110355, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7773009671427651, 'Affiliation-P': 0.6357255107386066, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Results for 748_YAHOO_id_198_WebService_tr_500_1st_170.csv: {'filename': '748_YAHOO_id_198_WebService_tr_500_1st_170.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.75, 'AUC-ROC': 0.999151103565365, 'VUS-PR': 0.5500000000000003, 'VUS-ROC': 0.9990449915110355, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7773009671427651, 'Affiliation-P': 0.6357255107386066, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Logits metrics skipped for 748_YAHOO_id_198_WebService_tr_500_1st_170.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/748_YAHOO_id_198_WebService_tr_500_1st_170_results.pkl
Skipping file: 831_Exathlon_id_22_Facility_tr_11665_1st_13484.csv due to filter criteria.
Processing file: 653_YAHOO_id_103_WebService_tr_500_1st_443.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 653_YAHOO_id_103_WebService_tr_500_1st_443.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 166.54it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1634.52it/s]
Table1-only metrics completed in 0.28s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9216866243901459, 'VUS-ROC': 0.9998337154555766, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9983874489163446, 'Affiliation-P': 0.9967800901018857, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Results for 653_YAHOO_id_103_WebService_tr_500_1st_443.csv: {'filename': '653_YAHOO_id_103_WebService_tr_500_1st_443.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9216866243901459, 'VUS-ROC': 0.9998337154555766, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9983874489163446, 'Affiliation-P': 0.9967800901018857, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Logits metrics skipped for 653_YAHOO_id_103_WebService_tr_500_1st_443.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/653_YAHOO_id_103_WebService_tr_500_1st_443_results.pkl
Skipping file: 205_SMD_id_28_Facility_tr_5925_1st_17580.csv due to filter criteria.
Processing file: 481_UCR_id_179_Medical_tr_1900_1st_5684.csv, Data shape: (7601, 1)
Label shape: (7601,)
Train data shape: (1900, 1), Test data shape: (5701, 1), Label test shape: (5701,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 481_UCR_id_179_Medical_tr_1900_1st_5684.csv...
Using CUDA device: 2
Data shape: (5701, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5701, window_size 5000, adding 4299 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 168.28it/s]
Original shapes - Output: (10000,), Label: (5701,)
Aligned shapes - Output: (5701,), Label: (5701,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 737.47it/s]
Table1-only metrics completed in 0.93s
Evaluation Result:  {'AUC-PR': 0.9160588847272813, 'AUC-ROC': 0.9939847065204674, 'VUS-PR': 0.9775231413173227, 'VUS-ROC': 0.9990138128127685, 'Standard-F1': 0.8328075709774153, 'Standard-Precision': 0.8979591836734633, 'Standard-Recall': 0.7764705882352896, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990808848476945, 'Affiliation-P': 0.9994264926824432, 'Affiliation-R': 0.998735515956953, 'F1_T': 0.8269230769225756, 'Precision_T': 0.9084507042253458, 'Recall_T': 0.7588235294117602}
Results for 481_UCR_id_179_Medical_tr_1900_1st_5684.csv: {'filename': '481_UCR_id_179_Medical_tr_1900_1st_5684.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 173, 'train_index': '1900', 'data_shape': '7601x1', 'output_length': 10000, 'label_length': 5701, 'aligned_length': 5701, 'AUC-PR': 0.9160588847272813, 'AUC-ROC': 0.9939847065204674, 'VUS-PR': 0.9775231413173227, 'VUS-ROC': 0.9990138128127685, 'Standard-F1': 0.8328075709774153, 'Standard-Precision': 0.8979591836734633, 'Standard-Recall': 0.7764705882352896, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990808848476945, 'Affiliation-P': 0.9994264926824432, 'Affiliation-R': 0.998735515956953, 'F1_T': 0.8269230769225756, 'Precision_T': 0.9084507042253458, 'Recall_T': 0.7588235294117602}
Logits metrics skipped for 481_UCR_id_179_Medical_tr_1900_1st_5684.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/481_UCR_id_179_Medical_tr_1900_1st_5684_results.pkl
Processing file: 780_YAHOO_id_230_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 780_YAHOO_id_230_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 159.62it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2098.79it/s]
Table1-only metrics completed in 0.60s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8252551020408159, 'Affiliation-P': 0.7024972855591748, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 780_YAHOO_id_230_Synthetic_tr_500_1st_893.csv: {'filename': '780_YAHOO_id_230_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 154, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8252551020408159, 'Affiliation-P': 0.7024972855591748, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 780_YAHOO_id_230_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/780_YAHOO_id_230_Synthetic_tr_500_1st_893_results.pkl
Processing file: 592_YAHOO_id_42_WebService_tr_500_1st_28.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 592_YAHOO_id_42_WebService_tr_500_1st_28.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.84it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1270.27it/s]
Table1-only metrics completed in 0.31s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9076368799663692, 'VUS-ROC': 0.9998038765704814, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9972000525347038, 'Affiliation-P': 0.9944157407020754, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Results for 592_YAHOO_id_42_WebService_tr_500_1st_28.csv: {'filename': '592_YAHOO_id_42_WebService_tr_500_1st_28.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9076368799663692, 'VUS-ROC': 0.9998038765704814, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9972000525347038, 'Affiliation-P': 0.9944157407020754, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Logits metrics skipped for 592_YAHOO_id_42_WebService_tr_500_1st_28.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/592_YAHOO_id_42_WebService_tr_500_1st_28_results.pkl
Processing file: 330_UCR_id_28_Facility_tr_7482_1st_23405.csv, Data shape: (29931, 1)
Label shape: (29931,)
Train data shape: (7482, 1), Test data shape: (22449, 1), Label test shape: (22449,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 330_UCR_id_28_Facility_tr_7482_1st_23405.csv...
Using CUDA device: 2
Data shape: (22449, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22449, window_size 5000, adding 2551 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 150.28it/s]
Original shapes - Output: (25000,), Label: (22449,)
Aligned shapes - Output: (22449,), Label: (22449,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  60%|██████    | 121/200 [00:01<00:00, 120.83it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 145.57it/s]
Table1-only metrics completed in 2.28s
Evaluation Result:  {'AUC-PR': 0.018130097088486674, 'AUC-ROC': 0.917759678836901, 'VUS-PR': 0.018881245897680723, 'VUS-ROC': 0.9179774230427842, 'Standard-F1': 0.04575645756447495, 'Standard-Precision': 0.02416212003117691, 'Standard-Recall': 0.4305555555555496, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6800431224560596, 'Affiliation-P': 0.5152010145372512, 'Affiliation-R': 1.0, 'F1_T': 0.0449438202246284, 'Precision_T': 0.023594725884802204, 'Recall_T': 0.4722222222222157}
Results for 330_UCR_id_28_Facility_tr_7482_1st_23405.csv: {'filename': '330_UCR_id_28_Facility_tr_7482_1st_23405.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 168, 'train_index': '7482', 'data_shape': '29931x1', 'output_length': 25000, 'label_length': 22449, 'aligned_length': 22449, 'AUC-PR': 0.018130097088486674, 'AUC-ROC': 0.917759678836901, 'VUS-PR': 0.018881245897680723, 'VUS-ROC': 0.9179774230427842, 'Standard-F1': 0.04575645756447495, 'Standard-Precision': 0.02416212003117691, 'Standard-Recall': 0.4305555555555496, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6800431224560596, 'Affiliation-P': 0.5152010145372512, 'Affiliation-R': 1.0, 'F1_T': 0.0449438202246284, 'Precision_T': 0.023594725884802204, 'Recall_T': 0.4722222222222157}
Logits metrics skipped for 330_UCR_id_28_Facility_tr_7482_1st_23405.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/330_UCR_id_28_Facility_tr_7482_1st_23405_results.pkl
Processing file: 605_YAHOO_id_55_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 605_YAHOO_id_55_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 160.27it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2549.64it/s]
Table1-only metrics completed in 0.93s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7607642626480081, 'Affiliation-P': 0.6138979370249729, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 605_YAHOO_id_55_Synthetic_tr_500_1st_893.csv: {'filename': '605_YAHOO_id_55_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 234, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7607642626480081, 'Affiliation-P': 0.6138979370249729, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 605_YAHOO_id_55_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/605_YAHOO_id_55_Synthetic_tr_500_1st_893_results.pkl
Processing file: 386_UCR_id_84_Environment_tr_2046_1st_5703.csv, Data shape: (8184, 1)
Label shape: (8184,)
Train data shape: (2046, 1), Test data shape: (6138, 1), Label test shape: (6138,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 386_UCR_id_84_Environment_tr_2046_1st_5703.csv...
Using CUDA device: 2
Data shape: (6138, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6138, window_size 5000, adding 3862 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 156.22it/s]
Original shapes - Output: (10000,), Label: (6138,)
Aligned shapes - Output: (6138,), Label: (6138,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 501.49it/s]
Table1-only metrics completed in 0.52s
Evaluation Result:  {'AUC-PR': 0.05672799642594676, 'AUC-ROC': 0.9567999672881911, 'VUS-PR': 0.05822463173037553, 'VUS-ROC': 0.9636730501415957, 'Standard-F1': 0.13168724279817481, 'Standard-Precision': 0.07305936073059327, 'Standard-Recall': 0.666666666666639, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7604640726261629, 'Affiliation-P': 0.6135070842499365, 'Affiliation-R': 1.0, 'F1_T': 0.1279461279459785, 'Precision_T': 0.06959706959706934, 'Recall_T': 0.7916666666666338}
Results for 386_UCR_id_84_Environment_tr_2046_1st_5703.csv: {'filename': '386_UCR_id_84_Environment_tr_2046_1st_5703.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '2046', 'data_shape': '8184x1', 'output_length': 10000, 'label_length': 6138, 'aligned_length': 6138, 'AUC-PR': 0.05672799642594676, 'AUC-ROC': 0.9567999672881911, 'VUS-PR': 0.05822463173037553, 'VUS-ROC': 0.9636730501415957, 'Standard-F1': 0.13168724279817481, 'Standard-Precision': 0.07305936073059327, 'Standard-Recall': 0.666666666666639, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7604640726261629, 'Affiliation-P': 0.6135070842499365, 'Affiliation-R': 1.0, 'F1_T': 0.1279461279459785, 'Precision_T': 0.06959706959706934, 'Recall_T': 0.7916666666666338}
Logits metrics skipped for 386_UCR_id_84_Environment_tr_2046_1st_5703.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/386_UCR_id_84_Environment_tr_2046_1st_5703_results.pkl
Processing file: 332_UCR_id_30_Facility_tr_13750_1st_52600.csv, Data shape: (55000, 1)
Label shape: (55000,)
Train data shape: (13750, 1), Test data shape: (41250, 1), Label test shape: (41250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 332_UCR_id_30_Facility_tr_13750_1st_52600.csv...
Using CUDA device: 2
Data shape: (41250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 41250, window_size 5000, adding 3750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([9, 5000, 1])
batched_mask shape: torch.Size([9, 5000])

100%|██████████| 1/1 [00:00<00:00, 112.51it/s]
Original shapes - Output: (45000,), Label: (41250,)
Aligned shapes - Output: (41250,), Label: (41250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  45%|████▌     | 90/200 [00:01<00:01, 89.96it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 105.55it/s]
Table1-only metrics completed in 3.61s
Evaluation Result:  {'AUC-PR': 0.01724414323425273, 'AUC-ROC': 0.8689833130328867, 'VUS-PR': 0.024613597300065997, 'VUS-ROC': 0.8936619409250298, 'Standard-F1': 0.041882296305164234, 'Standard-Precision': 0.02145763965963744, 'Standard-Recall': 0.8699999999999957, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7076118175181529, 'Affiliation-P': 0.5475226616195811, 'Affiliation-R': 1.0, 'F1_T': 0.0420067210753252, 'Precision_T': 0.021519921298573535, 'Recall_T': 0.8749999999999957}
Results for 332_UCR_id_30_Facility_tr_13750_1st_52600.csv: {'filename': '332_UCR_id_30_Facility_tr_13750_1st_52600.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 233, 'train_index': '13750', 'data_shape': '55000x1', 'output_length': 45000, 'label_length': 41250, 'aligned_length': 41250, 'AUC-PR': 0.01724414323425273, 'AUC-ROC': 0.8689833130328867, 'VUS-PR': 0.024613597300065997, 'VUS-ROC': 0.8936619409250298, 'Standard-F1': 0.041882296305164234, 'Standard-Precision': 0.02145763965963744, 'Standard-Recall': 0.8699999999999957, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7076118175181529, 'Affiliation-P': 0.5475226616195811, 'Affiliation-R': 1.0, 'F1_T': 0.0420067210753252, 'Precision_T': 0.021519921298573535, 'Recall_T': 0.8749999999999957}
Logits metrics skipped for 332_UCR_id_30_Facility_tr_13750_1st_52600.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/332_UCR_id_30_Facility_tr_13750_1st_52600_results.pkl
Skipping file: 144_MSL_id_5_Sensor_tr_1525_1st_4590.csv due to filter criteria.
Processing file: 480_UCR_id_178_Facility_tr_50000_1st_160720.csv, Data shape: (299890, 1)
Label shape: (299890,)
Train data shape: (50000, 1), Test data shape: (249890, 1), Label test shape: (249890,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 480_UCR_id_178_Facility_tr_50000_1st_160720.csv...
Using CUDA device: 2
Data shape: (249890, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 249890, window_size 5000, adding 110 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([50, 5000, 1])
batched_mask shape: torch.Size([50, 5000])

100%|██████████| 1/1 [00:00<00:00, 28.38it/s]
Original shapes - Output: (250000,), Label: (249890,)
Aligned shapes - Output: (249890,), Label: (249890,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   9%|▉         | 18/200 [00:01<00:10, 17.84it/s]

Calculating Affiliation Metrics:  18%|█▊        | 36/200 [00:02<00:10, 15.95it/s]

Calculating Affiliation Metrics:  26%|██▋       | 53/200 [00:03<00:10, 14.64it/s]

Calculating Affiliation Metrics:  34%|███▍      | 68/200 [00:04<00:09, 14.05it/s]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [00:05<00:08, 13.92it/s]

Calculating Affiliation Metrics:  48%|████▊     | 97/200 [00:06<00:07, 13.83it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [00:07<00:06, 13.96it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [00:08<00:05, 14.18it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [00:09<00:03, 15.32it/s]

Calculating Affiliation Metrics:  83%|████████▎ | 166/200 [00:10<00:02, 16.88it/s]

Calculating Affiliation Metrics:  97%|█████████▋| 194/200 [00:11<00:00, 20.05it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:12<00:00, 16.60it/s]
Table1-only metrics completed in 18.04s
Evaluation Result:  {'AUC-PR': 0.3085181689089266, 'AUC-ROC': 0.8595625686702961, 'VUS-PR': 0.11778661458260628, 'VUS-ROC': 0.8608115187272047, 'Standard-F1': 0.18228008891678882, 'Standard-Precision': 0.11484593837535009, 'Standard-Recall': 0.4415384615384608, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6778872984200043, 'Affiliation-P': 0.5127303425872041, 'Affiliation-R': 1.0, 'F1_T': 0.18228008891678882, 'Precision_T': 0.11484593837535009, 'Recall_T': 0.4415384615384608}
Results for 480_UCR_id_178_Facility_tr_50000_1st_160720.csv: {'filename': '480_UCR_id_178_Facility_tr_50000_1st_160720.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 81, 'train_index': '50000', 'data_shape': '299890x1', 'output_length': 250000, 'label_length': 249890, 'aligned_length': 249890, 'AUC-PR': 0.3085181689089266, 'AUC-ROC': 0.8595625686702961, 'VUS-PR': 0.11778661458260628, 'VUS-ROC': 0.8608115187272047, 'Standard-F1': 0.18228008891678882, 'Standard-Precision': 0.11484593837535009, 'Standard-Recall': 0.4415384615384608, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6778872984200043, 'Affiliation-P': 0.5127303425872041, 'Affiliation-R': 1.0, 'F1_T': 0.18228008891678882, 'Precision_T': 0.11484593837535009, 'Recall_T': 0.4415384615384608}
Logits metrics skipped for 480_UCR_id_178_Facility_tr_50000_1st_160720.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/480_UCR_id_178_Facility_tr_50000_1st_160720_results.pkl
Processing file: 138_WSD_id_110_WebService_tr_4584_1st_6039.csv, Data shape: (18336, 1)
Label shape: (18336,)
Train data shape: (4584, 1), Test data shape: (13752, 1), Label test shape: (13752,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 138_WSD_id_110_WebService_tr_4584_1st_6039.csv...
Using CUDA device: 2
Data shape: (13752, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13752, window_size 5000, adding 1248 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.16it/s]
Original shapes - Output: (15000,), Label: (13752,)
Aligned shapes - Output: (13752,), Label: (13752,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 287.41it/s]
Table1-only metrics completed in 0.89s
Evaluation Result:  {'AUC-PR': 0.0023892930163070323, 'AUC-ROC': 0.3218966781817062, 'VUS-PR': 0.00193842340727028, 'VUS-ROC': 0.3522983321505811, 'Standard-F1': 0.013986013985927108, 'Standard-Precision': 0.007326007326007298, 'Standard-Recall': 0.153846153846142, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7525580054561236, 'Affiliation-P': 0.6034317086345883, 'Affiliation-R': 0.9995860741934041, 'F1_T': 0.013986013985927108, 'Precision_T': 0.007326007326007298, 'Recall_T': 0.153846153846142}
Results for 138_WSD_id_110_WebService_tr_4584_1st_6039.csv: {'filename': '138_WSD_id_110_WebService_tr_4584_1st_6039.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '4584', 'data_shape': '18336x1', 'output_length': 15000, 'label_length': 13752, 'aligned_length': 13752, 'AUC-PR': 0.0023892930163070323, 'AUC-ROC': 0.3218966781817062, 'VUS-PR': 0.00193842340727028, 'VUS-ROC': 0.3522983321505811, 'Standard-F1': 0.013986013985927108, 'Standard-Precision': 0.007326007326007298, 'Standard-Recall': 0.153846153846142, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7525580054561236, 'Affiliation-P': 0.6034317086345883, 'Affiliation-R': 0.9995860741934041, 'F1_T': 0.013986013985927108, 'Precision_T': 0.007326007326007298, 'Recall_T': 0.153846153846142}
Logits metrics skipped for 138_WSD_id_110_WebService_tr_4584_1st_6039.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/138_WSD_id_110_WebService_tr_4584_1st_6039_results.pkl
Processing file: 159_Stock_id_11_Finance_tr_500_1st_5.csv, Data shape: (20000, 1)
Label shape: (20000,)
Train data shape: (500, 1), Test data shape: (19500, 1), Label test shape: (19500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 159_Stock_id_11_Finance_tr_500_1st_5.csv...
Using CUDA device: 2
Data shape: (19500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 19500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.44it/s]
Original shapes - Output: (20000,), Label: (19500,)
Aligned shapes - Output: (19500,), Label: (19500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   4%|▎         | 7/200 [00:01<00:31,  6.16it/s]

Calculating Affiliation Metrics:   7%|▋         | 14/200 [00:04<01:08,  2.72it/s]

Calculating Affiliation Metrics:   9%|▉         | 18/200 [00:07<01:24,  2.16it/s]

Calculating Affiliation Metrics:  10%|█         | 21/200 [00:09<01:35,  1.88it/s]

Calculating Affiliation Metrics:  12%|█▏        | 24/200 [00:11<01:46,  1.65it/s]

Calculating Affiliation Metrics:  13%|█▎        | 26/200 [00:13<01:55,  1.50it/s]

Calculating Affiliation Metrics:  14%|█▍        | 28/200 [00:15<02:02,  1.41it/s]

Calculating Affiliation Metrics:  15%|█▌        | 30/200 [00:17<02:07,  1.33it/s]

Calculating Affiliation Metrics:  16%|█▌        | 32/200 [00:19<02:11,  1.27it/s]

Calculating Affiliation Metrics:  17%|█▋        | 34/200 [00:20<02:14,  1.24it/s]

Calculating Affiliation Metrics:  18%|█▊        | 36/200 [00:22<02:15,  1.21it/s]

Calculating Affiliation Metrics:  19%|█▉        | 38/200 [00:24<02:19,  1.16it/s]

Calculating Affiliation Metrics:  20%|██        | 40/200 [00:26<02:22,  1.12it/s]

Calculating Affiliation Metrics:  21%|██        | 42/200 [00:28<02:26,  1.08it/s]

Calculating Affiliation Metrics:  22%|██▏       | 44/200 [00:30<02:32,  1.02it/s]

Calculating Affiliation Metrics:  23%|██▎       | 46/200 [00:32<02:37,  1.02s/it]

Calculating Affiliation Metrics:  24%|██▎       | 47/200 [00:33<02:38,  1.04s/it]

Calculating Affiliation Metrics:  24%|██▍       | 48/200 [00:35<02:40,  1.06s/it]

Calculating Affiliation Metrics:  24%|██▍       | 49/200 [00:36<02:42,  1.08s/it]

Calculating Affiliation Metrics:  25%|██▌       | 50/200 [00:37<02:44,  1.10s/it]

Calculating Affiliation Metrics:  26%|██▌       | 51/200 [00:38<02:45,  1.11s/it]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:39<02:45,  1.12s/it]

Calculating Affiliation Metrics:  26%|██▋       | 53/200 [00:40<02:46,  1.13s/it]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:41<02:44,  1.13s/it]

Calculating Affiliation Metrics:  28%|██▊       | 55/200 [00:43<02:43,  1.13s/it]

Calculating Affiliation Metrics:  28%|██▊       | 56/200 [00:44<02:42,  1.13s/it]

Calculating Affiliation Metrics:  28%|██▊       | 57/200 [00:45<02:43,  1.14s/it]

Calculating Affiliation Metrics:  29%|██▉       | 58/200 [00:46<02:43,  1.15s/it]

Calculating Affiliation Metrics:  30%|██▉       | 59/200 [00:47<02:44,  1.16s/it]

Calculating Affiliation Metrics:  30%|███       | 60/200 [00:48<02:43,  1.17s/it]

Calculating Affiliation Metrics:  30%|███       | 61/200 [00:50<02:43,  1.17s/it]

Calculating Affiliation Metrics:  31%|███       | 62/200 [00:51<02:42,  1.18s/it]

Calculating Affiliation Metrics:  32%|███▏      | 63/200 [00:52<02:41,  1.18s/it]

Calculating Affiliation Metrics:  32%|███▏      | 64/200 [00:53<02:41,  1.19s/it]

Calculating Affiliation Metrics:  32%|███▎      | 65/200 [00:54<02:41,  1.19s/it]

Calculating Affiliation Metrics:  33%|███▎      | 66/200 [00:56<02:41,  1.21s/it]

Calculating Affiliation Metrics:  34%|███▎      | 67/200 [00:57<02:41,  1.22s/it]

Calculating Affiliation Metrics:  34%|███▍      | 68/200 [00:58<02:41,  1.22s/it]

Calculating Affiliation Metrics:  34%|███▍      | 69/200 [00:59<02:41,  1.23s/it]

Calculating Affiliation Metrics:  35%|███▌      | 70/200 [01:01<02:44,  1.26s/it]

Calculating Affiliation Metrics:  36%|███▌      | 71/200 [01:02<02:43,  1.27s/it]

Calculating Affiliation Metrics:  36%|███▌      | 72/200 [01:03<02:41,  1.26s/it]

Calculating Affiliation Metrics:  36%|███▋      | 73/200 [01:05<02:40,  1.26s/it]

Calculating Affiliation Metrics:  37%|███▋      | 74/200 [01:06<02:38,  1.26s/it]

Calculating Affiliation Metrics:  38%|███▊      | 75/200 [01:07<02:38,  1.27s/it]

Calculating Affiliation Metrics:  38%|███▊      | 76/200 [01:08<02:39,  1.29s/it]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [01:10<02:36,  1.28s/it]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [01:11<02:35,  1.27s/it]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [01:12<02:33,  1.27s/it]

Calculating Affiliation Metrics:  40%|████      | 80/200 [01:13<02:32,  1.27s/it]

Calculating Affiliation Metrics:  40%|████      | 81/200 [01:15<02:30,  1.26s/it]

Calculating Affiliation Metrics:  41%|████      | 82/200 [01:16<02:27,  1.25s/it]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [01:17<02:25,  1.25s/it]

Calculating Affiliation Metrics:  42%|████▏     | 84/200 [01:18<02:23,  1.24s/it]

Calculating Affiliation Metrics:  42%|████▎     | 85/200 [01:19<02:19,  1.21s/it]

Calculating Affiliation Metrics:  43%|████▎     | 86/200 [01:21<02:16,  1.19s/it]

Calculating Affiliation Metrics:  44%|████▎     | 87/200 [01:22<02:13,  1.18s/it]

Calculating Affiliation Metrics:  44%|████▍     | 88/200 [01:23<02:10,  1.16s/it]

Calculating Affiliation Metrics:  44%|████▍     | 89/200 [01:24<02:06,  1.14s/it]

Calculating Affiliation Metrics:  45%|████▌     | 90/200 [01:25<02:05,  1.14s/it]

Calculating Affiliation Metrics:  46%|████▌     | 91/200 [01:26<02:03,  1.13s/it]

Calculating Affiliation Metrics:  46%|████▌     | 92/200 [01:27<02:02,  1.13s/it]

Calculating Affiliation Metrics:  46%|████▋     | 93/200 [01:28<02:00,  1.13s/it]

Calculating Affiliation Metrics:  47%|████▋     | 94/200 [01:30<01:59,  1.12s/it]

Calculating Affiliation Metrics:  48%|████▊     | 95/200 [01:31<01:58,  1.13s/it]

Calculating Affiliation Metrics:  48%|████▊     | 96/200 [01:32<01:58,  1.14s/it]

Calculating Affiliation Metrics:  48%|████▊     | 97/200 [01:33<01:58,  1.15s/it]

Calculating Affiliation Metrics:  49%|████▉     | 98/200 [01:34<01:56,  1.14s/it]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [01:35<01:55,  1.15s/it]

Calculating Affiliation Metrics:  50%|█████     | 100/200 [01:37<01:55,  1.15s/it]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [01:38<01:53,  1.15s/it]

Calculating Affiliation Metrics:  51%|█████     | 102/200 [01:39<01:52,  1.15s/it]

Calculating Affiliation Metrics:  52%|█████▏    | 103/200 [01:40<01:50,  1.14s/it]

Calculating Affiliation Metrics:  52%|█████▏    | 104/200 [01:41<01:50,  1.15s/it]

Calculating Affiliation Metrics:  52%|█████▎    | 105/200 [01:42<01:49,  1.15s/it]

Calculating Affiliation Metrics:  53%|█████▎    | 106/200 [01:43<01:48,  1.16s/it]

Calculating Affiliation Metrics:  54%|█████▎    | 107/200 [01:45<01:47,  1.15s/it]

Calculating Affiliation Metrics:  54%|█████▍    | 108/200 [01:46<01:45,  1.15s/it]

Calculating Affiliation Metrics:  55%|█████▍    | 109/200 [01:47<01:43,  1.14s/it]

Calculating Affiliation Metrics:  55%|█████▌    | 110/200 [01:48<01:42,  1.14s/it]

Calculating Affiliation Metrics:  56%|█████▌    | 111/200 [01:49<01:41,  1.14s/it]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [01:50<01:41,  1.15s/it]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [01:51<01:39,  1.15s/it]

Calculating Affiliation Metrics:  57%|█████▋    | 114/200 [01:53<01:38,  1.14s/it]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [01:54<01:35,  1.13s/it]

Calculating Affiliation Metrics:  58%|█████▊    | 116/200 [01:55<01:34,  1.12s/it]

Calculating Affiliation Metrics:  58%|█████▊    | 117/200 [01:56<01:32,  1.11s/it]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [01:57<01:30,  1.11s/it]

Calculating Affiliation Metrics:  60%|█████▉    | 119/200 [01:58<01:29,  1.10s/it]

Calculating Affiliation Metrics:  60%|██████    | 120/200 [01:59<01:27,  1.10s/it]

Calculating Affiliation Metrics:  60%|██████    | 121/200 [02:00<01:26,  1.09s/it]

Calculating Affiliation Metrics:  61%|██████    | 122/200 [02:01<01:24,  1.08s/it]

Calculating Affiliation Metrics:  62%|██████▏   | 123/200 [02:02<01:22,  1.08s/it]

Calculating Affiliation Metrics:  62%|██████▏   | 124/200 [02:03<01:21,  1.08s/it]

Calculating Affiliation Metrics:  62%|██████▎   | 125/200 [02:04<01:20,  1.07s/it]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [02:05<01:18,  1.05s/it]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [02:07<01:16,  1.05s/it]

Calculating Affiliation Metrics:  64%|██████▍   | 128/200 [02:08<01:15,  1.05s/it]

Calculating Affiliation Metrics:  64%|██████▍   | 129/200 [02:09<01:14,  1.04s/it]

Calculating Affiliation Metrics:  65%|██████▌   | 130/200 [02:10<01:12,  1.04s/it]

Calculating Affiliation Metrics:  66%|██████▌   | 131/200 [02:11<01:11,  1.04s/it]

Calculating Affiliation Metrics:  66%|██████▋   | 133/200 [02:13<01:08,  1.02s/it]

Calculating Affiliation Metrics:  67%|██████▋   | 134/200 [02:14<01:06,  1.01s/it]

Calculating Affiliation Metrics:  68%|██████▊   | 135/200 [02:15<01:06,  1.02s/it]

Calculating Affiliation Metrics:  68%|██████▊   | 136/200 [02:16<01:04,  1.01s/it]

Calculating Affiliation Metrics:  69%|██████▉   | 138/200 [02:18<01:01,  1.01it/s]

Calculating Affiliation Metrics:  70%|███████   | 140/200 [02:20<00:59,  1.00it/s]

Calculating Affiliation Metrics:  71%|███████   | 142/200 [02:22<00:57,  1.01it/s]

Calculating Affiliation Metrics:  72%|███████▏  | 144/200 [02:23<00:54,  1.03it/s]

Calculating Affiliation Metrics:  73%|███████▎  | 146/200 [02:25<00:51,  1.04it/s]

Calculating Affiliation Metrics:  74%|███████▍  | 148/200 [02:27<00:48,  1.07it/s]

Calculating Affiliation Metrics:  75%|███████▌  | 150/200 [02:29<00:45,  1.10it/s]

Calculating Affiliation Metrics:  76%|███████▌  | 152/200 [02:31<00:42,  1.12it/s]

Calculating Affiliation Metrics:  77%|███████▋  | 154/200 [02:32<00:39,  1.15it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 156/200 [02:34<00:36,  1.19it/s]

Calculating Affiliation Metrics:  79%|███████▉  | 158/200 [02:35<00:34,  1.23it/s]

Calculating Affiliation Metrics:  80%|████████  | 160/200 [02:37<00:31,  1.27it/s]

Calculating Affiliation Metrics:  81%|████████  | 162/200 [02:38<00:29,  1.31it/s]

Calculating Affiliation Metrics:  82%|████████▏ | 164/200 [02:39<00:26,  1.35it/s]

Calculating Affiliation Metrics:  83%|████████▎ | 166/200 [02:41<00:24,  1.40it/s]

Calculating Affiliation Metrics:  84%|████████▍ | 168/200 [02:42<00:21,  1.46it/s]

Calculating Affiliation Metrics:  85%|████████▌ | 170/200 [02:43<00:19,  1.53it/s]

Calculating Affiliation Metrics:  86%|████████▌ | 172/200 [02:44<00:17,  1.60it/s]

Calculating Affiliation Metrics:  87%|████████▋ | 174/200 [02:45<00:15,  1.70it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 177/200 [02:47<00:12,  1.83it/s]

Calculating Affiliation Metrics:  90%|█████████ | 180/200 [02:48<00:10,  1.99it/s]

Calculating Affiliation Metrics:  92%|█████████▏| 183/200 [02:49<00:07,  2.16it/s]

Calculating Affiliation Metrics:  93%|█████████▎| 186/200 [02:50<00:05,  2.37it/s]

Calculating Affiliation Metrics:  95%|█████████▌| 190/200 [02:51<00:03,  2.73it/s]

Calculating Affiliation Metrics:  98%|█████████▊| 196/200 [02:52<00:01,  3.54it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [02:53<00:00,  1.16it/s]
Table1-only metrics completed in 255.70s
Evaluation Result:  {'AUC-PR': 0.19096698683642827, 'AUC-ROC': 0.5532055366912365, 'VUS-PR': 0.9680838405448553, 'VUS-ROC': 0.9794029524161394, 'Standard-F1': 0.2476482820078857, 'Standard-Precision': 0.1431924882629108, 'Standard-Recall': 0.9154465592972179, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6871249477693895, 'Affiliation-P': 0.5233742134119665, 'Affiliation-R': 1.0, 'F1_T': 0.2474767464870015, 'Precision_T': 0.1430778032036613, 'Recall_T': 0.9154465592972179}
Results for 159_Stock_id_11_Finance_tr_500_1st_5.csv: {'filename': '159_Stock_id_11_Finance_tr_500_1st_5.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 215, 'train_index': '500', 'data_shape': '20000x1', 'output_length': 20000, 'label_length': 19500, 'aligned_length': 19500, 'AUC-PR': 0.19096698683642827, 'AUC-ROC': 0.5532055366912365, 'VUS-PR': 0.9680838405448553, 'VUS-ROC': 0.9794029524161394, 'Standard-F1': 0.2476482820078857, 'Standard-Precision': 0.1431924882629108, 'Standard-Recall': 0.9154465592972179, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6871249477693895, 'Affiliation-P': 0.5233742134119665, 'Affiliation-R': 1.0, 'F1_T': 0.2474767464870015, 'Precision_T': 0.1430778032036613, 'Recall_T': 0.9154465592972179}
Logits metrics skipped for 159_Stock_id_11_Finance_tr_500_1st_5.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/159_Stock_id_11_Finance_tr_500_1st_5_results.pkl
Processing file: 558_YAHOO_id_8_WebService_tr_500_1st_1125.csv, Data shape: (1427, 1)
Label shape: (1427,)
Train data shape: (500, 1), Test data shape: (927, 1), Label test shape: (927,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 558_YAHOO_id_8_WebService_tr_500_1st_1125.csv...
Using CUDA device: 2
Data shape: (927, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 927, 1])
batched_mask shape: torch.Size([1, 927])

100%|██████████| 1/1 [00:00<00:00, 143.54it/s]
Original shapes - Output: (927,), Label: (927,)
Aligned shapes - Output: (927,), Label: (927,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1818.22it/s]
Table1-only metrics completed in 0.26s
Evaluation Result:  {'AUC-PR': 0.07687521343432402, 'AUC-ROC': 0.4673848878394332, 'VUS-PR': 0.09359110990044331, 'VUS-ROC': 0.5375284825719658, 'Standard-F1': 0.19412515964221705, 'Standard-Precision': 0.10810810810810795, 'Standard-Recall': 0.9499999999999882, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6678972378553133, 'Affiliation-P': 0.5013856714627623, 'Affiliation-R': 1.0, 'F1_T': 0.19412515964221705, 'Precision_T': 0.10810810810810795, 'Recall_T': 0.9499999999999882}
Results for 558_YAHOO_id_8_WebService_tr_500_1st_1125.csv: {'filename': '558_YAHOO_id_8_WebService_tr_500_1st_1125.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1427x1', 'output_length': 927, 'label_length': 927, 'aligned_length': 927, 'AUC-PR': 0.07687521343432402, 'AUC-ROC': 0.4673848878394332, 'VUS-PR': 0.09359110990044331, 'VUS-ROC': 0.5375284825719658, 'Standard-F1': 0.19412515964221705, 'Standard-Precision': 0.10810810810810795, 'Standard-Recall': 0.9499999999999882, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6678972378553133, 'Affiliation-P': 0.5013856714627623, 'Affiliation-R': 1.0, 'F1_T': 0.19412515964221705, 'Precision_T': 0.10810810810810795, 'Recall_T': 0.9499999999999882}
Logits metrics skipped for 558_YAHOO_id_8_WebService_tr_500_1st_1125.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/558_YAHOO_id_8_WebService_tr_500_1st_1125_results.pkl
Processing file: 661_YAHOO_id_111_WebService_tr_500_1st_279.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 661_YAHOO_id_111_WebService_tr_500_1st_279.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 163.32it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1230.08it/s]
Table1-only metrics completed in 0.47s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9882803785638185, 'VUS-ROC': 0.9999500743451507, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Results for 661_YAHOO_id_111_WebService_tr_500_1st_279.csv: {'filename': '661_YAHOO_id_111_WebService_tr_500_1st_279.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9882803785638185, 'VUS-ROC': 0.9999500743451507, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Logits metrics skipped for 661_YAHOO_id_111_WebService_tr_500_1st_279.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/661_YAHOO_id_111_WebService_tr_500_1st_279_results.pkl
Processing file: 001_NAB_id_1_Facility_tr_1007_1st_2014.csv, Data shape: (4031, 1)
Label shape: (4031,)
Train data shape: (1007, 1), Test data shape: (3024, 1), Label test shape: (3024,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 001_NAB_id_1_Facility_tr_1007_1st_2014.csv...
Using CUDA device: 2
Data shape: (3024, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 3024, 1])
batched_mask shape: torch.Size([1, 3024])

100%|██████████| 1/1 [00:00<00:00, 165.20it/s]
Original shapes - Output: (3024,), Label: (3024,)
Aligned shapes - Output: (3024,), Label: (3024,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 707.68it/s]
Table1-only metrics completed in 0.36s
Evaluation Result:  {'AUC-PR': 0.13395314164720817, 'AUC-ROC': 0.3342460658798607, 'VUS-PR': 0.13127986217699725, 'VUS-ROC': 0.3367944946820482, 'Standard-F1': 0.20323741007175783, 'Standard-Precision': 0.11326428332776475, 'Standard-Recall': 0.9883381924198221, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8908308684087175, 'Affiliation-P': 0.8719911419218178, 'Affiliation-R': 0.9105026518732657, 'F1_T': 0.20323741007175783, 'Precision_T': 0.11326428332776475, 'Recall_T': 0.9883381924198221}
Results for 001_NAB_id_1_Facility_tr_1007_1st_2014.csv: {'filename': '001_NAB_id_1_Facility_tr_1007_1st_2014.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 6, 'train_index': '1007', 'data_shape': '4031x1', 'output_length': 3024, 'label_length': 3024, 'aligned_length': 3024, 'AUC-PR': 0.13395314164720817, 'AUC-ROC': 0.3342460658798607, 'VUS-PR': 0.13127986217699725, 'VUS-ROC': 0.3367944946820482, 'Standard-F1': 0.20323741007175783, 'Standard-Precision': 0.11326428332776475, 'Standard-Recall': 0.9883381924198221, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8908308684087175, 'Affiliation-P': 0.8719911419218178, 'Affiliation-R': 0.9105026518732657, 'F1_T': 0.20323741007175783, 'Precision_T': 0.11326428332776475, 'Recall_T': 0.9883381924198221}
Logits metrics skipped for 001_NAB_id_1_Facility_tr_1007_1st_2014.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/001_NAB_id_1_Facility_tr_1007_1st_2014_results.pkl
Processing file: 041_WSD_id_13_WebService_tr_4296_1st_5196.csv, Data shape: (17186, 1)
Label shape: (17186,)
Train data shape: (4296, 1), Test data shape: (12890, 1), Label test shape: (12890,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 041_WSD_id_13_WebService_tr_4296_1st_5196.csv...
Using CUDA device: 2
Data shape: (12890, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 12890, window_size 5000, adding 2110 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.80it/s]
Original shapes - Output: (15000,), Label: (12890,)
Aligned shapes - Output: (12890,), Label: (12890,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 247.61it/s]
Table1-only metrics completed in 2.63s
Evaluation Result:  {'AUC-PR': 0.8653928923161337, 'AUC-ROC': 0.9913551646510654, 'VUS-PR': 0.9412298399542525, 'VUS-ROC': 0.9991766794808781, 'Standard-F1': 0.8488745980702341, 'Standard-Precision': 0.851612903225801, 'Standard-Recall': 0.8461538461538407, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984593407593902, 'Affiliation-P': 0.9998232695139911, 'Affiliation-R': 0.9970991281962297, 'F1_T': 0.814035087718797, 'Precision_T': 0.8992248062015434, 'Recall_T': 0.7435897435897388}
Results for 041_WSD_id_13_WebService_tr_4296_1st_5196.csv: {'filename': '041_WSD_id_13_WebService_tr_4296_1st_5196.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4296', 'data_shape': '17186x1', 'output_length': 15000, 'label_length': 12890, 'aligned_length': 12890, 'AUC-PR': 0.8653928923161337, 'AUC-ROC': 0.9913551646510654, 'VUS-PR': 0.9412298399542525, 'VUS-ROC': 0.9991766794808781, 'Standard-F1': 0.8488745980702341, 'Standard-Precision': 0.851612903225801, 'Standard-Recall': 0.8461538461538407, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984593407593902, 'Affiliation-P': 0.9998232695139911, 'Affiliation-R': 0.9970991281962297, 'F1_T': 0.814035087718797, 'Precision_T': 0.8992248062015434, 'Recall_T': 0.7435897435897388}
Logits metrics skipped for 041_WSD_id_13_WebService_tr_4296_1st_5196.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/041_WSD_id_13_WebService_tr_4296_1st_5196_results.pkl
Processing file: 052_WSD_id_24_WebService_tr_4559_1st_9102.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 052_WSD_id_24_WebService_tr_4559_1st_9102.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.77it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 308.51it/s]
Table1-only metrics completed in 0.75s
Evaluation Result:  {'AUC-PR': 0.9219004657802979, 'AUC-ROC': 0.9956236984275496, 'VUS-PR': 0.5663531151062139, 'VUS-ROC': 0.9969503376930688, 'Standard-F1': 0.3742690058476302, 'Standard-Precision': 0.23357664233576472, 'Standard-Recall': 0.9411764705882075, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.810481470123972, 'Affiliation-P': 0.6813717286799251, 'Affiliation-R': 0.9999586665839999, 'F1_T': 0.3742690058476302, 'Precision_T': 0.23357664233576472, 'Recall_T': 0.9411764705882075}
Results for 052_WSD_id_24_WebService_tr_4559_1st_9102.csv: {'filename': '052_WSD_id_24_WebService_tr_4559_1st_9102.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 6, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.9219004657802979, 'AUC-ROC': 0.9956236984275496, 'VUS-PR': 0.5663531151062139, 'VUS-ROC': 0.9969503376930688, 'Standard-F1': 0.3742690058476302, 'Standard-Precision': 0.23357664233576472, 'Standard-Recall': 0.9411764705882075, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.810481470123972, 'Affiliation-P': 0.6813717286799251, 'Affiliation-R': 0.9999586665839999, 'F1_T': 0.3742690058476302, 'Precision_T': 0.23357664233576472, 'Recall_T': 0.9411764705882075}
Logits metrics skipped for 052_WSD_id_24_WebService_tr_4559_1st_9102.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/052_WSD_id_24_WebService_tr_4559_1st_9102_results.pkl
Skipping file: 818_Exathlon_id_9_Facility_tr_11665_1st_13484.csv due to filter criteria.
Processing file: 617_YAHOO_id_67_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 617_YAHOO_id_67_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 161.52it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1555.80it/s]
Table1-only metrics completed in 0.65s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984775989560672, 'Affiliation-P': 0.9969598262757872, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 617_YAHOO_id_67_Synthetic_tr_500_1st_893.csv: {'filename': '617_YAHOO_id_67_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 158, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984775989560672, 'Affiliation-P': 0.9969598262757872, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 617_YAHOO_id_67_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/617_YAHOO_id_67_Synthetic_tr_500_1st_893_results.pkl
Processing file: 094_WSD_id_66_WebService_tr_3309_1st_3914.csv, Data shape: (13237, 1)
Label shape: (13237,)
Train data shape: (3309, 1), Test data shape: (9928, 1), Label test shape: (9928,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 094_WSD_id_66_WebService_tr_3309_1st_3914.csv...
Using CUDA device: 2
Data shape: (9928, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9928, window_size 5000, adding 72 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.75it/s]
Original shapes - Output: (10000,), Label: (9928,)
Aligned shapes - Output: (9928,), Label: (9928,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 354.06it/s]
Table1-only metrics completed in 1.67s
Evaluation Result:  {'AUC-PR': 0.32542296337718807, 'AUC-ROC': 0.7591717087117875, 'VUS-PR': 0.23566663412164843, 'VUS-ROC': 0.964435171920038, 'Standard-F1': 0.19259259259220568, 'Standard-Precision': 0.1299999999999987, 'Standard-Recall': 0.3714285714285608, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8558036065658936, 'Affiliation-P': 0.7482343984601886, 'Affiliation-R': 0.9994950660649873, 'F1_T': 0.19259259259220568, 'Precision_T': 0.1299999999999987, 'Recall_T': 0.3714285714285608}
Results for 094_WSD_id_66_WebService_tr_3309_1st_3914.csv: {'filename': '094_WSD_id_66_WebService_tr_3309_1st_3914.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '3309', 'data_shape': '13237x1', 'output_length': 10000, 'label_length': 9928, 'aligned_length': 9928, 'AUC-PR': 0.32542296337718807, 'AUC-ROC': 0.7591717087117875, 'VUS-PR': 0.23566663412164843, 'VUS-ROC': 0.964435171920038, 'Standard-F1': 0.19259259259220568, 'Standard-Precision': 0.1299999999999987, 'Standard-Recall': 0.3714285714285608, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8558036065658936, 'Affiliation-P': 0.7482343984601886, 'Affiliation-R': 0.9994950660649873, 'F1_T': 0.19259259259220568, 'Precision_T': 0.1299999999999987, 'Recall_T': 0.3714285714285608}
Logits metrics skipped for 094_WSD_id_66_WebService_tr_3309_1st_3914.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/094_WSD_id_66_WebService_tr_3309_1st_3914_results.pkl
Processing file: 740_YAHOO_id_190_WebService_tr_500_1st_0.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 740_YAHOO_id_190_WebService_tr_500_1st_0.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 167.46it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1261.29it/s]
Table1-only metrics completed in 0.46s
Evaluation Result:  {'AUC-PR': 0.8016638935108154, 'AUC-ROC': 0.8985531914893617, 'VUS-PR': 0.6090115121836138, 'VUS-ROC': 0.8967268597474326, 'Standard-F1': 0.4705882352936471, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.79999999999984, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9404089885643532, 'Affiliation-P': 0.9390635959784897, 'Affiliation-R': 0.9417582417582417, 'F1_T': 0.4705882352936471, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.79999999999984}
Results for 740_YAHOO_id_190_WebService_tr_500_1st_0.csv: {'filename': '740_YAHOO_id_190_WebService_tr_500_1st_0.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.8016638935108154, 'AUC-ROC': 0.8985531914893617, 'VUS-PR': 0.6090115121836138, 'VUS-ROC': 0.8967268597474326, 'Standard-F1': 0.4705882352936471, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.79999999999984, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9404089885643532, 'Affiliation-P': 0.9390635959784897, 'Affiliation-R': 0.9417582417582417, 'F1_T': 0.4705882352936471, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.79999999999984}
Logits metrics skipped for 740_YAHOO_id_190_WebService_tr_500_1st_0.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/740_YAHOO_id_190_WebService_tr_500_1st_0_results.pkl
Processing file: 669_YAHOO_id_119_WebService_tr_500_1st_21.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 669_YAHOO_id_119_WebService_tr_500_1st_21.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 168.04it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 981.50it/s]
Table1-only metrics completed in 0.62s
Evaluation Result:  {'AUC-PR': 0.9999999999999998, 'AUC-ROC': 1.0, 'VUS-PR': 0.9795409654774682, 'VUS-ROC': 0.9999127503007479, 'Standard-F1': 0.7368421052626148, 'Standard-Precision': 0.5833333333332847, 'Standard-Recall': 0.9999999999998571, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9962221276701799, 'Affiliation-P': 0.9924726925469061, 'Affiliation-R': 1.0, 'F1_T': 0.7368421052626148, 'Precision_T': 0.5833333333332847, 'Recall_T': 0.9999999999998571}
Results for 669_YAHOO_id_119_WebService_tr_500_1st_21.csv: {'filename': '669_YAHOO_id_119_WebService_tr_500_1st_21.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9999999999999998, 'AUC-ROC': 1.0, 'VUS-PR': 0.9795409654774682, 'VUS-ROC': 0.9999127503007479, 'Standard-F1': 0.7368421052626148, 'Standard-Precision': 0.5833333333332847, 'Standard-Recall': 0.9999999999998571, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9962221276701799, 'Affiliation-P': 0.9924726925469061, 'Affiliation-R': 1.0, 'F1_T': 0.7368421052626148, 'Precision_T': 0.5833333333332847, 'Recall_T': 0.9999999999998571}
Logits metrics skipped for 669_YAHOO_id_119_WebService_tr_500_1st_21.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/669_YAHOO_id_119_WebService_tr_500_1st_21_results.pkl
Processing file: 430_UCR_id_128_HumanActivity_tr_12500_1st_46500.csv, Data shape: (50000, 1)
Label shape: (50000,)
Train data shape: (12500, 1), Test data shape: (37500, 1), Label test shape: (37500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 430_UCR_id_128_HumanActivity_tr_12500_1st_46500.csv...
Using CUDA device: 2
Data shape: (37500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 37500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([8, 5000, 1])
batched_mask shape: torch.Size([8, 5000])

100%|██████████| 1/1 [00:00<00:00, 126.97it/s]
Original shapes - Output: (40000,), Label: (37500,)
Aligned shapes - Output: (37500,), Label: (37500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  51%|█████     | 102/200 [00:01<00:00, 101.05it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 112.09it/s]
Table1-only metrics completed in 2.70s
Evaluation Result:  {'AUC-PR': 0.010474022617717457, 'AUC-ROC': 0.6332654569892474, 'VUS-PR': 0.012726214722600063, 'VUS-ROC': 0.6769882593345992, 'Standard-F1': 0.026397232841745524, 'Standard-Precision': 0.013569155904922326, 'Standard-Recall': 0.48333333333333167, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6818128420333902, 'Affiliation-P': 0.5175095051901504, 'Affiliation-R': 0.9989763999999999, 'F1_T': 0.026322955432460403, 'Precision_T': 0.013529905757208171, 'Recall_T': 0.48333333333333167}
Results for 430_UCR_id_128_HumanActivity_tr_12500_1st_46500.csv: {'filename': '430_UCR_id_128_HumanActivity_tr_12500_1st_46500.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '12500', 'data_shape': '50000x1', 'output_length': 40000, 'label_length': 37500, 'aligned_length': 37500, 'AUC-PR': 0.010474022617717457, 'AUC-ROC': 0.6332654569892474, 'VUS-PR': 0.012726214722600063, 'VUS-ROC': 0.6769882593345992, 'Standard-F1': 0.026397232841745524, 'Standard-Precision': 0.013569155904922326, 'Standard-Recall': 0.48333333333333167, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6818128420333902, 'Affiliation-P': 0.5175095051901504, 'Affiliation-R': 0.9989763999999999, 'F1_T': 0.026322955432460403, 'Precision_T': 0.013529905757208171, 'Recall_T': 0.48333333333333167}
Logits metrics skipped for 430_UCR_id_128_HumanActivity_tr_12500_1st_46500.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/430_UCR_id_128_HumanActivity_tr_12500_1st_46500_results.pkl
Processing file: 760_YAHOO_id_210_WebService_tr_500_1st_23.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 760_YAHOO_id_210_WebService_tr_500_1st_23.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.37it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1344.39it/s]
Table1-only metrics completed in 0.35s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9504941880689357, 'VUS-ROC': 0.9998948142040058, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9972665652268263, 'Affiliation-P': 0.9945480330496977, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Results for 760_YAHOO_id_210_WebService_tr_500_1st_23.csv: {'filename': '760_YAHOO_id_210_WebService_tr_500_1st_23.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9504941880689357, 'VUS-ROC': 0.9998948142040058, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9972665652268263, 'Affiliation-P': 0.9945480330496977, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Logits metrics skipped for 760_YAHOO_id_210_WebService_tr_500_1st_23.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/760_YAHOO_id_210_WebService_tr_500_1st_23_results.pkl
Processing file: 270_IOPS_id_11_WebService_tr_5638_1st_5738.csv, Data shape: (60000, 1)
Label shape: (60000,)
Train data shape: (5638, 1), Test data shape: (54362, 1), Label test shape: (54362,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 270_IOPS_id_11_WebService_tr_5638_1st_5738.csv...
Using CUDA device: 2
Data shape: (54362, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 54362, window_size 5000, adding 638 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([11, 5000, 1])
batched_mask shape: torch.Size([11, 5000])

100%|██████████| 1/1 [00:00<00:00, 96.59it/s]
Original shapes - Output: (55000,), Label: (54362,)
Aligned shapes - Output: (54362,), Label: (54362,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:01<00:02, 53.17it/s]

Calculating Affiliation Metrics:  54%|█████▍    | 108/200 [00:02<00:01, 46.76it/s]

Calculating Affiliation Metrics:  90%|████████▉ | 179/200 [00:03<00:00, 56.76it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:03<00:00, 58.24it/s]
Table1-only metrics completed in 7.22s
Evaluation Result:  {'AUC-PR': 0.0883568091279113, 'AUC-ROC': 0.7790572060776503, 'VUS-PR': 0.034653348973682614, 'VUS-ROC': 0.8829399620725982, 'Standard-F1': 0.07194244604282513, 'Standard-Precision': 0.04595588235294109, 'Standard-Recall': 0.16556291390728367, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7950275156551694, 'Affiliation-P': 0.7518898322408846, 'Affiliation-R': 0.8434162864444479, 'F1_T': 0.07194244604282513, 'Precision_T': 0.04595588235294109, 'Recall_T': 0.16556291390728367}
Results for 270_IOPS_id_11_WebService_tr_5638_1st_5738.csv: {'filename': '270_IOPS_id_11_WebService_tr_5638_1st_5738.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '5638', 'data_shape': '60000x1', 'output_length': 55000, 'label_length': 54362, 'aligned_length': 54362, 'AUC-PR': 0.0883568091279113, 'AUC-ROC': 0.7790572060776503, 'VUS-PR': 0.034653348973682614, 'VUS-ROC': 0.8829399620725982, 'Standard-F1': 0.07194244604282513, 'Standard-Precision': 0.04595588235294109, 'Standard-Recall': 0.16556291390728367, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7950275156551694, 'Affiliation-P': 0.7518898322408846, 'Affiliation-R': 0.8434162864444479, 'F1_T': 0.07194244604282513, 'Precision_T': 0.04595588235294109, 'Recall_T': 0.16556291390728367}
Logits metrics skipped for 270_IOPS_id_11_WebService_tr_5638_1st_5738.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/270_IOPS_id_11_WebService_tr_5638_1st_5738_results.pkl
Processing file: 096_WSD_id_68_WebService_tr_1341_1st_1441.csv, Data shape: (10744, 1)
Label shape: (10744,)
Train data shape: (1341, 1), Test data shape: (9403, 1), Label test shape: (9403,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 096_WSD_id_68_WebService_tr_1341_1st_1441.csv...
Using CUDA device: 2
Data shape: (9403, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9403, window_size 5000, adding 597 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.67it/s]
Original shapes - Output: (10000,), Label: (9403,)
Aligned shapes - Output: (9403,), Label: (9403,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 357.27it/s]
Table1-only metrics completed in 1.64s
Evaluation Result:  {'AUC-PR': 0.4232052592072345, 'AUC-ROC': 0.948239060832444, 'VUS-PR': 0.6259425189093919, 'VUS-ROC': 0.9959260173841196, 'Standard-F1': 0.2465753424654002, 'Standard-Precision': 0.1592920353982287, 'Standard-Recall': 0.5454545454545289, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975463394866229, 'Affiliation-P': 0.9966214012705619, 'Affiliation-R': 0.9984729961193617, 'F1_T': 0.22988505747095442, 'Precision_T': 0.14184397163120468, 'Recall_T': 0.6060606060605876}
Results for 096_WSD_id_68_WebService_tr_1341_1st_1441.csv: {'filename': '096_WSD_id_68_WebService_tr_1341_1st_1441.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '1341', 'data_shape': '10744x1', 'output_length': 10000, 'label_length': 9403, 'aligned_length': 9403, 'AUC-PR': 0.4232052592072345, 'AUC-ROC': 0.948239060832444, 'VUS-PR': 0.6259425189093919, 'VUS-ROC': 0.9959260173841196, 'Standard-F1': 0.2465753424654002, 'Standard-Precision': 0.1592920353982287, 'Standard-Recall': 0.5454545454545289, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975463394866229, 'Affiliation-P': 0.9966214012705619, 'Affiliation-R': 0.9984729961193617, 'F1_T': 0.22988505747095442, 'Precision_T': 0.14184397163120468, 'Recall_T': 0.6060606060605876}
Logits metrics skipped for 096_WSD_id_68_WebService_tr_1341_1st_1441.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/096_WSD_id_68_WebService_tr_1341_1st_1441_results.pkl
Processing file: 792_YAHOO_id_242_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 792_YAHOO_id_242_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 161.80it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2433.31it/s]
Table1-only metrics completed in 0.50s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9536468984321739, 'Affiliation-P': 0.911400651465798, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 792_YAHOO_id_242_Synthetic_tr_500_1st_893.csv: {'filename': '792_YAHOO_id_242_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9536468984321739, 'Affiliation-P': 0.911400651465798, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 792_YAHOO_id_242_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/792_YAHOO_id_242_Synthetic_tr_500_1st_893_results.pkl
Skipping file: 847_OPPORTUNITY_id_6_HumanActivity_tr_808_1st_908.csv due to filter criteria.
Processing file: 233_MGAB_id_9_Synthetic_tr_25000_1st_41300.csv, Data shape: (100000, 1)
Label shape: (100000,)
Train data shape: (25000, 1), Test data shape: (75000, 1), Label test shape: (75000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 233_MGAB_id_9_Synthetic_tr_25000_1st_41300.csv...
Using CUDA device: 2
Data shape: (75000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([15, 5000, 1])
batched_mask shape: torch.Size([15, 5000])

100%|██████████| 1/1 [00:00<00:00, 78.61it/s]
Original shapes - Output: (75000,), Label: (75000,)
Aligned shapes - Output: (75000,), Label: (75000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  20%|██        | 41/200 [00:01<00:03, 39.86it/s]

Calculating Affiliation Metrics:  40%|████      | 81/200 [00:02<00:04, 28.76it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [00:04<00:03, 24.72it/s]

Calculating Affiliation Metrics:  69%|██████▉   | 138/200 [00:05<00:02, 23.50it/s]

Calculating Affiliation Metrics:  82%|████████▎ | 165/200 [00:06<00:01, 24.38it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:07<00:00, 27.93it/s]
Table1-only metrics completed in 10.89s
Evaluation Result:  {'AUC-PR': 0.05845942216652192, 'AUC-ROC': 0.824506584224599, 'VUS-PR': 0.09263304883273679, 'VUS-ROC': 0.8897122680226318, 'Standard-F1': 0.09894736842072001, 'Standard-Precision': 0.06266666666666658, 'Standard-Recall': 0.23499999999999882, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6855916064016115, 'Affiliation-P': 0.5224885364393292, 'Affiliation-R': 0.9967396804754923, 'F1_T': 0.09894736842072001, 'Precision_T': 0.06266666666666658, 'Recall_T': 0.23499999999999882}
Results for 233_MGAB_id_9_Synthetic_tr_25000_1st_41300.csv: {'filename': '233_MGAB_id_9_Synthetic_tr_25000_1st_41300.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 103, 'train_index': '25000', 'data_shape': '100000x1', 'output_length': 75000, 'label_length': 75000, 'aligned_length': 75000, 'AUC-PR': 0.05845942216652192, 'AUC-ROC': 0.824506584224599, 'VUS-PR': 0.09263304883273679, 'VUS-ROC': 0.8897122680226318, 'Standard-F1': 0.09894736842072001, 'Standard-Precision': 0.06266666666666658, 'Standard-Recall': 0.23499999999999882, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6855916064016115, 'Affiliation-P': 0.5224885364393292, 'Affiliation-R': 0.9967396804754923, 'F1_T': 0.09894736842072001, 'Precision_T': 0.06266666666666658, 'Recall_T': 0.23499999999999882}
Logits metrics skipped for 233_MGAB_id_9_Synthetic_tr_25000_1st_41300.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/233_MGAB_id_9_Synthetic_tr_25000_1st_41300_results.pkl
Processing file: 530_UCR_id_228_HumanActivity_tr_16250_1st_38500.csv, Data shape: (65000, 1)
Label shape: (65000,)
Train data shape: (16250, 1), Test data shape: (48750, 1), Label test shape: (48750,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 530_UCR_id_228_HumanActivity_tr_16250_1st_38500.csv...
Using CUDA device: 2
Data shape: (48750, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 48750, window_size 5000, adding 1250 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([10, 5000, 1])
batched_mask shape: torch.Size([10, 5000])

100%|██████████| 1/1 [00:00<00:00, 93.48it/s]
Original shapes - Output: (50000,), Label: (48750,)
Aligned shapes - Output: (48750,), Label: (48750,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  38%|███▊      | 75/200 [00:01<00:01, 74.50it/s]

Calculating Affiliation Metrics:  98%|█████████▊| 196/200 [00:02<00:00, 101.55it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 98.55it/s] 
Table1-only metrics completed in 3.09s
Evaluation Result:  {'AUC-PR': 0.1141377451888472, 'AUC-ROC': 0.8795427244582043, 'VUS-PR': 0.09182973547881192, 'VUS-ROC': 0.8990889497750443, 'Standard-F1': 0.18424242424212653, 'Standard-Precision': 0.11259259259259252, 'Standard-Recall': 0.5066666666666649, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6676272258400483, 'Affiliation-P': 0.5010814081374348, 'Affiliation-R': 1.0, 'F1_T': 0.18037135278482694, 'Precision_T': 0.11258278145695355, 'Recall_T': 0.4533333333333318}
Results for 530_UCR_id_228_HumanActivity_tr_16250_1st_38500.csv: {'filename': '530_UCR_id_228_HumanActivity_tr_16250_1st_38500.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '16250', 'data_shape': '65000x1', 'output_length': 50000, 'label_length': 48750, 'aligned_length': 48750, 'AUC-PR': 0.1141377451888472, 'AUC-ROC': 0.8795427244582043, 'VUS-PR': 0.09182973547881192, 'VUS-ROC': 0.8990889497750443, 'Standard-F1': 0.18424242424212653, 'Standard-Precision': 0.11259259259259252, 'Standard-Recall': 0.5066666666666649, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6676272258400483, 'Affiliation-P': 0.5010814081374348, 'Affiliation-R': 1.0, 'F1_T': 0.18037135278482694, 'Precision_T': 0.11258278145695355, 'Recall_T': 0.4533333333333318}
Logits metrics skipped for 530_UCR_id_228_HumanActivity_tr_16250_1st_38500.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/530_UCR_id_228_HumanActivity_tr_16250_1st_38500_results.pkl
Processing file: 763_YAHOO_id_213_WebService_tr_500_1st_43.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 763_YAHOO_id_213_WebService_tr_500_1st_43.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 160.66it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 978.95it/s]
Table1-only metrics completed in 0.66s
Evaluation Result:  {'AUC-PR': 0.7681023641851107, 'AUC-ROC': 0.9896544368600683, 'VUS-PR': 0.6783750729122453, 'VUS-ROC': 0.9918555800577924, 'Standard-F1': 0.59999999999946, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.7499999999999063, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9787280328493019, 'Affiliation-P': 0.9593382807668522, 'Affiliation-R': 0.9989177489177489, 'F1_T': 0.59999999999946, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.7499999999999063}
Results for 763_YAHOO_id_213_WebService_tr_500_1st_43.csv: {'filename': '763_YAHOO_id_213_WebService_tr_500_1st_43.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.7681023641851107, 'AUC-ROC': 0.9896544368600683, 'VUS-PR': 0.6783750729122453, 'VUS-ROC': 0.9918555800577924, 'Standard-F1': 0.59999999999946, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.7499999999999063, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9787280328493019, 'Affiliation-P': 0.9593382807668522, 'Affiliation-R': 0.9989177489177489, 'F1_T': 0.59999999999946, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.7499999999999063}
Logits metrics skipped for 763_YAHOO_id_213_WebService_tr_500_1st_43.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/763_YAHOO_id_213_WebService_tr_500_1st_43_results.pkl
Processing file: 310_UCR_id_8_HumanActivity_tr_2631_1st_8285.csv, Data shape: (10524, 1)
Label shape: (10524,)
Train data shape: (2631, 1), Test data shape: (7893, 1), Label test shape: (7893,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 310_UCR_id_8_HumanActivity_tr_2631_1st_8285.csv...
Using CUDA device: 2
Data shape: (7893, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 7893, window_size 5000, adding 2107 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 164.12it/s]
Original shapes - Output: (10000,), Label: (7893,)
Aligned shapes - Output: (7893,), Label: (7893,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 394.18it/s]
Table1-only metrics completed in 0.64s
Evaluation Result:  {'AUC-PR': 0.11041425303940344, 'AUC-ROC': 0.953898003306626, 'VUS-PR': 0.0713058285387828, 'VUS-ROC': 0.9596963256524761, 'Standard-F1': 0.15999999999963266, 'Standard-Precision': 0.10526315789473574, 'Standard-Recall': 0.33333333333332227, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9098715556607074, 'Affiliation-P': 0.8356106676802231, 'Affiliation-R': 0.9986190295198277, 'F1_T': 0.14678899082528643, 'Precision_T': 0.10126582278480885, 'Recall_T': 0.2666666666666578}
Results for 310_UCR_id_8_HumanActivity_tr_2631_1st_8285.csv: {'filename': '310_UCR_id_8_HumanActivity_tr_2631_1st_8285.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '2631', 'data_shape': '10524x1', 'output_length': 10000, 'label_length': 7893, 'aligned_length': 7893, 'AUC-PR': 0.11041425303940344, 'AUC-ROC': 0.953898003306626, 'VUS-PR': 0.0713058285387828, 'VUS-ROC': 0.9596963256524761, 'Standard-F1': 0.15999999999963266, 'Standard-Precision': 0.10526315789473574, 'Standard-Recall': 0.33333333333332227, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9098715556607074, 'Affiliation-P': 0.8356106676802231, 'Affiliation-R': 0.9986190295198277, 'F1_T': 0.14678899082528643, 'Precision_T': 0.10126582278480885, 'Recall_T': 0.2666666666666578}
Logits metrics skipped for 310_UCR_id_8_HumanActivity_tr_2631_1st_8285.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/310_UCR_id_8_HumanActivity_tr_2631_1st_8285_results.pkl
Processing file: 598_YAHOO_id_48_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 598_YAHOO_id_48_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 164.61it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1990.81it/s]
Table1-only metrics completed in 0.67s
Evaluation Result:  {'AUC-PR': 0.95, 'AUC-ROC': 0.9997273718647764, 'VUS-PR': 0.989484647286154, 'VUS-ROC': 0.9999651591048541, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 598_YAHOO_id_48_Synthetic_tr_500_1st_658.csv: {'filename': '598_YAHOO_id_48_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 105, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.95, 'AUC-ROC': 0.9997273718647764, 'VUS-PR': 0.989484647286154, 'VUS-ROC': 0.9999651591048541, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 598_YAHOO_id_48_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/598_YAHOO_id_48_Synthetic_tr_500_1st_658_results.pkl
Processing file: 582_YAHOO_id_32_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 582_YAHOO_id_32_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 160.30it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2032.10it/s]
Table1-only metrics completed in 0.61s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 582_YAHOO_id_32_Synthetic_tr_500_1st_658.csv: {'filename': '582_YAHOO_id_32_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 92, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 582_YAHOO_id_32_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/582_YAHOO_id_32_Synthetic_tr_500_1st_658_results.pkl
Processing file: 381_UCR_id_79_HumanActivity_tr_16000_1st_31200.csv, Data shape: (64000, 1)
Label shape: (64000,)
Train data shape: (16000, 1), Test data shape: (48000, 1), Label test shape: (48000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 381_UCR_id_79_HumanActivity_tr_16000_1st_31200.csv...
Using CUDA device: 2
Data shape: (48000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 48000, window_size 5000, adding 2000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([10, 5000, 1])
batched_mask shape: torch.Size([10, 5000])

100%|██████████| 1/1 [00:00<00:00, 97.16it/s]
Original shapes - Output: (50000,), Label: (48000,)
Aligned shapes - Output: (48000,), Label: (48000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  44%|████▎     | 87/200 [00:01<00:01, 86.69it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 106.18it/s]
Table1-only metrics completed in 3.01s
Evaluation Result:  {'AUC-PR': 0.10453615213314463, 'AUC-ROC': 0.938081731784583, 'VUS-PR': 0.10990122039361366, 'VUS-ROC': 0.9409579743385393, 'Standard-F1': 0.19174852652237045, 'Standard-Precision': 0.10990990990990988, 'Standard-Recall': 0.7507692307692296, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6832956597716028, 'Affiliation-P': 0.5189438804865474, 'Affiliation-R': 1.0, 'F1_T': 0.19244170463653876, 'Precision_T': 0.11652061019149623, 'Recall_T': 0.5523076923076914}
Results for 381_UCR_id_79_HumanActivity_tr_16000_1st_31200.csv: {'filename': '381_UCR_id_79_HumanActivity_tr_16000_1st_31200.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '16000', 'data_shape': '64000x1', 'output_length': 50000, 'label_length': 48000, 'aligned_length': 48000, 'AUC-PR': 0.10453615213314463, 'AUC-ROC': 0.938081731784583, 'VUS-PR': 0.10990122039361366, 'VUS-ROC': 0.9409579743385393, 'Standard-F1': 0.19174852652237045, 'Standard-Precision': 0.10990990990990988, 'Standard-Recall': 0.7507692307692296, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6832956597716028, 'Affiliation-P': 0.5189438804865474, 'Affiliation-R': 1.0, 'F1_T': 0.19244170463653876, 'Precision_T': 0.11652061019149623, 'Recall_T': 0.5523076923076914}
Logits metrics skipped for 381_UCR_id_79_HumanActivity_tr_16000_1st_31200.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/381_UCR_id_79_HumanActivity_tr_16000_1st_31200_results.pkl
Processing file: 108_WSD_id_80_WebService_tr_2368_1st_2468.csv, Data shape: (18220, 1)
Label shape: (18220,)
Train data shape: (2368, 1), Test data shape: (15852, 1), Label test shape: (15852,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 108_WSD_id_80_WebService_tr_2368_1st_2468.csv...
Using CUDA device: 2
Data shape: (15852, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 15852, window_size 5000, adding 4148 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.24it/s]
Original shapes - Output: (20000,), Label: (15852,)
Aligned shapes - Output: (15852,), Label: (15852,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [00:01<00:00, 114.69it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 145.25it/s]
Table1-only metrics completed in 5.32s
Evaluation Result:  {'AUC-PR': 0.22614153860969946, 'AUC-ROC': 0.8057565597363764, 'VUS-PR': 0.30512776807768627, 'VUS-ROC': 0.9415202542614508, 'Standard-F1': 0.21590909090859436, 'Standard-Precision': 0.2389937106918224, 'Standard-Recall': 0.19689119170984354, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8341688989078436, 'Affiliation-P': 0.7973288024495514, 'Affiliation-R': 0.8745782599623476, 'F1_T': 0.21590909090859436, 'Precision_T': 0.2389937106918224, 'Recall_T': 0.19689119170984354}
Results for 108_WSD_id_80_WebService_tr_2368_1st_2468.csv: {'filename': '108_WSD_id_80_WebService_tr_2368_1st_2468.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '2368', 'data_shape': '18220x1', 'output_length': 20000, 'label_length': 15852, 'aligned_length': 15852, 'AUC-PR': 0.22614153860969946, 'AUC-ROC': 0.8057565597363764, 'VUS-PR': 0.30512776807768627, 'VUS-ROC': 0.9415202542614508, 'Standard-F1': 0.21590909090859436, 'Standard-Precision': 0.2389937106918224, 'Standard-Recall': 0.19689119170984354, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8341688989078436, 'Affiliation-P': 0.7973288024495514, 'Affiliation-R': 0.8745782599623476, 'F1_T': 0.21590909090859436, 'Precision_T': 0.2389937106918224, 'Recall_T': 0.19689119170984354}
Logits metrics skipped for 108_WSD_id_80_WebService_tr_2368_1st_2468.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/108_WSD_id_80_WebService_tr_2368_1st_2468_results.pkl
Processing file: 393_UCR_id_91_Environment_tr_2046_1st_4852.csv, Data shape: (8184, 1)
Label shape: (8184,)
Train data shape: (2046, 1), Test data shape: (6138, 1), Label test shape: (6138,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 393_UCR_id_91_Environment_tr_2046_1st_4852.csv...
Using CUDA device: 2
Data shape: (6138, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6138, window_size 5000, adding 3862 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 153.55it/s]
Original shapes - Output: (10000,), Label: (6138,)
Aligned shapes - Output: (6138,), Label: (6138,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 460.40it/s]
Table1-only metrics completed in 0.56s
Evaluation Result:  {'AUC-PR': 0.939098899835069, 'AUC-ROC': 0.9995347564313082, 'VUS-PR': 0.9089476983791417, 'VUS-ROC': 0.9995392905552873, 'Standard-F1': 0.8545454545449471, 'Standard-Precision': 0.75806451612902, 'Standard-Recall': 0.9791666666666462, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.996564765774566, 'Affiliation-P': 0.9935567958460778, 'Affiliation-R': 0.9995910041272945, 'F1_T': 0.8545454545449471, 'Precision_T': 0.75806451612902, 'Recall_T': 0.9791666666666462}
Results for 393_UCR_id_91_Environment_tr_2046_1st_4852.csv: {'filename': '393_UCR_id_91_Environment_tr_2046_1st_4852.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '2046', 'data_shape': '8184x1', 'output_length': 10000, 'label_length': 6138, 'aligned_length': 6138, 'AUC-PR': 0.939098899835069, 'AUC-ROC': 0.9995347564313082, 'VUS-PR': 0.9089476983791417, 'VUS-ROC': 0.9995392905552873, 'Standard-F1': 0.8545454545449471, 'Standard-Precision': 0.75806451612902, 'Standard-Recall': 0.9791666666666462, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.996564765774566, 'Affiliation-P': 0.9935567958460778, 'Affiliation-R': 0.9995910041272945, 'F1_T': 0.8545454545449471, 'Precision_T': 0.75806451612902, 'Recall_T': 0.9791666666666462}
Logits metrics skipped for 393_UCR_id_91_Environment_tr_2046_1st_4852.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/393_UCR_id_91_Environment_tr_2046_1st_4852_results.pkl
Processing file: 462_UCR_id_160_Medical_tr_1853_1st_6200.csv, Data shape: (7415, 1)
Label shape: (7415,)
Train data shape: (1853, 1), Test data shape: (5562, 1), Label test shape: (5562,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 462_UCR_id_160_Medical_tr_1853_1st_6200.csv...
Using CUDA device: 2
Data shape: (5562, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5562, window_size 5000, adding 4438 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 151.63it/s]
Original shapes - Output: (10000,), Label: (5562,)
Aligned shapes - Output: (5562,), Label: (5562,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 727.11it/s]
Table1-only metrics completed in 0.96s
Evaluation Result:  {'AUC-PR': 0.404434773583591, 'AUC-ROC': 0.9309543550357828, 'VUS-PR': 0.43128693275827346, 'VUS-ROC': 0.9536332035510583, 'Standard-F1': 0.42194092826963303, 'Standard-Precision': 0.7462686567164069, 'Standard-Recall': 0.2941176470588218, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9899984565382749, 'Affiliation-P': 0.9908434787075564, 'Affiliation-R': 0.9891548744632698, 'F1_T': 0.4189723320153661, 'Precision_T': 0.6385542168674623, 'Recall_T': 0.3117647058823511}
Results for 462_UCR_id_160_Medical_tr_1853_1st_6200.csv: {'filename': '462_UCR_id_160_Medical_tr_1853_1st_6200.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 173, 'train_index': '1853', 'data_shape': '7415x1', 'output_length': 10000, 'label_length': 5562, 'aligned_length': 5562, 'AUC-PR': 0.404434773583591, 'AUC-ROC': 0.9309543550357828, 'VUS-PR': 0.43128693275827346, 'VUS-ROC': 0.9536332035510583, 'Standard-F1': 0.42194092826963303, 'Standard-Precision': 0.7462686567164069, 'Standard-Recall': 0.2941176470588218, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9899984565382749, 'Affiliation-P': 0.9908434787075564, 'Affiliation-R': 0.9891548744632698, 'F1_T': 0.4189723320153661, 'Precision_T': 0.6385542168674623, 'Recall_T': 0.3117647058823511}
Logits metrics skipped for 462_UCR_id_160_Medical_tr_1853_1st_6200.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/462_UCR_id_160_Medical_tr_1853_1st_6200_results.pkl
Processing file: 701_YAHOO_id_151_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 701_YAHOO_id_151_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 70.63it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2197.44it/s]
Table1-only metrics completed in 0.61s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8735322896281795, 'Affiliation-P': 0.7754614549402823, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 701_YAHOO_id_151_Synthetic_tr_500_1st_893.csv: {'filename': '701_YAHOO_id_151_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 156, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8735322896281795, 'Affiliation-P': 0.7754614549402823, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 701_YAHOO_id_151_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/701_YAHOO_id_151_Synthetic_tr_500_1st_893_results.pkl
Processing file: 687_YAHOO_id_137_WebService_tr_500_1st_89.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 687_YAHOO_id_137_WebService_tr_500_1st_89.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 167.46it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1008.87it/s]
Table1-only metrics completed in 0.46s
Evaluation Result:  {'AUC-PR': 1.0000000000000002, 'AUC-ROC': 1.0, 'VUS-PR': 0.9833456090279421, 'VUS-ROC': 0.9999288788929532, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9979955677272886, 'Affiliation-P': 0.9959991548776596, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Results for 687_YAHOO_id_137_WebService_tr_500_1st_89.csv: {'filename': '687_YAHOO_id_137_WebService_tr_500_1st_89.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 12, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0000000000000002, 'AUC-ROC': 1.0, 'VUS-PR': 0.9833456090279421, 'VUS-ROC': 0.9999288788929532, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9979955677272886, 'Affiliation-P': 0.9959991548776596, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 687_YAHOO_id_137_WebService_tr_500_1st_89.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/687_YAHOO_id_137_WebService_tr_500_1st_89_results.pkl
Processing file: 509_UCR_id_207_Medical_tr_1875_1st_6599.csv, Data shape: (7501, 1)
Label shape: (7501,)
Train data shape: (1875, 1), Test data shape: (5626, 1), Label test shape: (5626,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 509_UCR_id_207_Medical_tr_1875_1st_6599.csv...
Using CUDA device: 2
Data shape: (5626, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5626, window_size 5000, adding 4374 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.62it/s]
Original shapes - Output: (10000,), Label: (5626,)
Aligned shapes - Output: (5626,), Label: (5626,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 744.22it/s]
Table1-only metrics completed in 0.93s
Evaluation Result:  {'AUC-PR': 0.007592255409503436, 'AUC-ROC': 0.04372778305705134, 'VUS-PR': 0.010072924497622522, 'VUS-ROC': 0.06771991948618607, 'Standard-F1': 0.026897894177992977, 'Standard-Precision': 0.013646974322140418, 'Standard-Recall': 0.9268292682926716, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7371878004111552, 'Affiliation-P': 0.6193431440280639, 'Affiliation-R': 0.9104159260575897, 'F1_T': 0.026897894177992977, 'Precision_T': 0.013646974322140418, 'Recall_T': 0.9268292682926716}
Results for 509_UCR_id_207_Medical_tr_1875_1st_6599.csv: {'filename': '509_UCR_id_207_Medical_tr_1875_1st_6599.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 175, 'train_index': '1875', 'data_shape': '7501x1', 'output_length': 10000, 'label_length': 5626, 'aligned_length': 5626, 'AUC-PR': 0.007592255409503436, 'AUC-ROC': 0.04372778305705134, 'VUS-PR': 0.010072924497622522, 'VUS-ROC': 0.06771991948618607, 'Standard-F1': 0.026897894177992977, 'Standard-Precision': 0.013646974322140418, 'Standard-Recall': 0.9268292682926716, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7371878004111552, 'Affiliation-P': 0.6193431440280639, 'Affiliation-R': 0.9104159260575897, 'F1_T': 0.026897894177992977, 'Precision_T': 0.013646974322140418, 'Recall_T': 0.9268292682926716}
Logits metrics skipped for 509_UCR_id_207_Medical_tr_1875_1st_6599.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/509_UCR_id_207_Medical_tr_1875_1st_6599_results.pkl
Processing file: 008_NAB_id_8_Synthetic_tr_1007_1st_2734.csv, Data shape: (4031, 1)
Label shape: (4031,)
Train data shape: (1007, 1), Test data shape: (3024, 1), Label test shape: (3024,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 008_NAB_id_8_Synthetic_tr_1007_1st_2734.csv...
Using CUDA device: 2
Data shape: (3024, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 3024, 1])
batched_mask shape: torch.Size([1, 3024])

100%|██████████| 1/1 [00:00<00:00, 143.44it/s]
Original shapes - Output: (3024,), Label: (3024,)
Aligned shapes - Output: (3024,), Label: (3024,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 720.45it/s]
Table1-only metrics completed in 0.56s
Evaluation Result:  {'AUC-PR': 0.7381484546010738, 'AUC-ROC': 0.81268618767338, 'VUS-PR': 0.7696972372800132, 'VUS-ROC': 0.8406271461188041, 'Standard-F1': 0.7780821917803248, 'Standard-Precision': 0.8658536585365827, 'Standard-Recall': 0.7064676616915405, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9874417925565432, 'Affiliation-P': 0.9906639739229025, 'Affiliation-R': 0.984240503830056, 'F1_T': 0.7780821917803248, 'Precision_T': 0.8658536585365827, 'Recall_T': 0.7064676616915405}
Results for 008_NAB_id_8_Synthetic_tr_1007_1st_2734.csv: {'filename': '008_NAB_id_8_Synthetic_tr_1007_1st_2734.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 71, 'train_index': '1007', 'data_shape': '4031x1', 'output_length': 3024, 'label_length': 3024, 'aligned_length': 3024, 'AUC-PR': 0.7381484546010738, 'AUC-ROC': 0.81268618767338, 'VUS-PR': 0.7696972372800132, 'VUS-ROC': 0.8406271461188041, 'Standard-F1': 0.7780821917803248, 'Standard-Precision': 0.8658536585365827, 'Standard-Recall': 0.7064676616915405, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9874417925565432, 'Affiliation-P': 0.9906639739229025, 'Affiliation-R': 0.984240503830056, 'F1_T': 0.7780821917803248, 'Precision_T': 0.8658536585365827, 'Recall_T': 0.7064676616915405}
Logits metrics skipped for 008_NAB_id_8_Synthetic_tr_1007_1st_2734.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/008_NAB_id_8_Synthetic_tr_1007_1st_2734_results.pkl
Processing file: 630_YAHOO_id_80_WebService_tr_500_1st_631.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 630_YAHOO_id_80_WebService_tr_500_1st_631.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 163.72it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1303.18it/s]
Table1-only metrics completed in 0.43s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5882352941171626, 'Standard-Precision': 0.41666666666663194, 'Standard-Recall': 0.9999999999997999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9991518235793039, 'Affiliation-P': 0.9983050847457626, 'Affiliation-R': 1.0, 'F1_T': 0.5882352941171626, 'Precision_T': 0.41666666666663194, 'Recall_T': 0.9999999999997999}
Results for 630_YAHOO_id_80_WebService_tr_500_1st_631.csv: {'filename': '630_YAHOO_id_80_WebService_tr_500_1st_631.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5882352941171626, 'Standard-Precision': 0.41666666666663194, 'Standard-Recall': 0.9999999999997999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9991518235793039, 'Affiliation-P': 0.9983050847457626, 'Affiliation-R': 1.0, 'F1_T': 0.5882352941171626, 'Precision_T': 0.41666666666663194, 'Recall_T': 0.9999999999997999}
Logits metrics skipped for 630_YAHOO_id_80_WebService_tr_500_1st_631.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/630_YAHOO_id_80_WebService_tr_500_1st_631_results.pkl
Processing file: 529_UCR_id_227_Facility_tr_19948_1st_52000.csv, Data shape: (79795, 1)
Label shape: (79795,)
Train data shape: (19948, 1), Test data shape: (59847, 1), Label test shape: (59847,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 529_UCR_id_227_Facility_tr_19948_1st_52000.csv...
Using CUDA device: 2
Data shape: (59847, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 59847, window_size 5000, adding 153 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([12, 5000, 1])
batched_mask shape: torch.Size([12, 5000])

100%|██████████| 1/1 [00:00<00:00, 92.94it/s]
Original shapes - Output: (60000,), Label: (59847,)
Aligned shapes - Output: (59847,), Label: (59847,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  34%|███▍      | 69/200 [00:01<00:01, 68.71it/s]

Calculating Affiliation Metrics:  76%|███████▋  | 153/200 [00:02<00:00, 77.66it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 84.14it/s]
Table1-only metrics completed in 4.54s
Evaluation Result:  {'AUC-PR': 0.04027768853259675, 'AUC-ROC': 0.7205242732036445, 'VUS-PR': 0.04281466170923027, 'VUS-ROC': 0.7515056191592049, 'Standard-F1': 0.12851937727705978, 'Standard-Precision': 0.07161314138058322, 'Standard-Recall': 0.6258064516129022, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6834974034967796, 'Affiliation-P': 0.519997479803318, 'Affiliation-R': 0.9969687622792961, 'F1_T': 0.12650505699131212, 'Precision_T': 0.07024425031199856, 'Recall_T': 0.6354838709677408}
Results for 529_UCR_id_227_Facility_tr_19948_1st_52000.csv: {'filename': '529_UCR_id_227_Facility_tr_19948_1st_52000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 208, 'train_index': '19948', 'data_shape': '79795x1', 'output_length': 60000, 'label_length': 59847, 'aligned_length': 59847, 'AUC-PR': 0.04027768853259675, 'AUC-ROC': 0.7205242732036445, 'VUS-PR': 0.04281466170923027, 'VUS-ROC': 0.7515056191592049, 'Standard-F1': 0.12851937727705978, 'Standard-Precision': 0.07161314138058322, 'Standard-Recall': 0.6258064516129022, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6834974034967796, 'Affiliation-P': 0.519997479803318, 'Affiliation-R': 0.9969687622792961, 'F1_T': 0.12650505699131212, 'Precision_T': 0.07024425031199856, 'Recall_T': 0.6354838709677408}
Logits metrics skipped for 529_UCR_id_227_Facility_tr_19948_1st_52000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/529_UCR_id_227_Facility_tr_19948_1st_52000_results.pkl
Processing file: 319_UCR_id_17_Environment_tr_6166_1st_19280.csv, Data shape: (24667, 1)
Label shape: (24667,)
Train data shape: (6166, 1), Test data shape: (18501, 1), Label test shape: (18501,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 319_UCR_id_17_Environment_tr_6166_1st_19280.csv...
Using CUDA device: 2
Data shape: (18501, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 18501, window_size 5000, adding 1499 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.25it/s]
Original shapes - Output: (20000,), Label: (18501,)
Aligned shapes - Output: (18501,), Label: (18501,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 240.45it/s]
Table1-only metrics completed in 1.13s
Evaluation Result:  {'AUC-PR': 0.8493793898198435, 'AUC-ROC': 0.9932078403576685, 'VUS-PR': 0.809822891702075, 'VUS-ROC': 0.9945312285657936, 'Standard-F1': 0.7456647398838916, 'Standard-Precision': 0.6935483870967705, 'Standard-Recall': 0.806249999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9896208486671252, 'Affiliation-P': 0.9798941411478486, 'Affiliation-R': 0.9995425922923085, 'F1_T': 0.7456647398838916, 'Precision_T': 0.6935483870967705, 'Recall_T': 0.806249999999995}
Results for 319_UCR_id_17_Environment_tr_6166_1st_19280.csv: {'filename': '319_UCR_id_17_Environment_tr_6166_1st_19280.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 46, 'train_index': '6166', 'data_shape': '24667x1', 'output_length': 20000, 'label_length': 18501, 'aligned_length': 18501, 'AUC-PR': 0.8493793898198435, 'AUC-ROC': 0.9932078403576685, 'VUS-PR': 0.809822891702075, 'VUS-ROC': 0.9945312285657936, 'Standard-F1': 0.7456647398838916, 'Standard-Precision': 0.6935483870967705, 'Standard-Recall': 0.806249999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9896208486671252, 'Affiliation-P': 0.9798941411478486, 'Affiliation-R': 0.9995425922923085, 'F1_T': 0.7456647398838916, 'Precision_T': 0.6935483870967705, 'Recall_T': 0.806249999999995}
Logits metrics skipped for 319_UCR_id_17_Environment_tr_6166_1st_19280.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/319_UCR_id_17_Environment_tr_6166_1st_19280_results.pkl
Processing file: 363_UCR_id_61_Medical_tr_7500_1st_16000.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 363_UCR_id_61_Medical_tr_7500_1st_16000.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 151.02it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 212.01it/s]
Table1-only metrics completed in 1.88s
Evaluation Result:  {'AUC-PR': 0.012609363382004489, 'AUC-ROC': 0.8407178571428571, 'VUS-PR': 0.016383462379185556, 'VUS-ROC': 0.8618435898597944, 'Standard-F1': 0.03334641035696428, 'Standard-Precision': 0.017006802721088433, 'Standard-Recall': 0.8499999999999915, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.712967080776677, 'Affiliation-P': 0.5539618063591778, 'Affiliation-R': 1.0, 'F1_T': 0.03339882121803612, 'Precision_T': 0.01703406813627254, 'Recall_T': 0.8499999999999915}
Results for 363_UCR_id_61_Medical_tr_7500_1st_16000.csv: {'filename': '363_UCR_id_61_Medical_tr_7500_1st_16000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 166, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.012609363382004489, 'AUC-ROC': 0.8407178571428571, 'VUS-PR': 0.016383462379185556, 'VUS-ROC': 0.8618435898597944, 'Standard-F1': 0.03334641035696428, 'Standard-Precision': 0.017006802721088433, 'Standard-Recall': 0.8499999999999915, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.712967080776677, 'Affiliation-P': 0.5539618063591778, 'Affiliation-R': 1.0, 'F1_T': 0.03339882121803612, 'Precision_T': 0.01703406813627254, 'Recall_T': 0.8499999999999915}
Logits metrics skipped for 363_UCR_id_61_Medical_tr_7500_1st_16000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/363_UCR_id_61_Medical_tr_7500_1st_16000_results.pkl
Processing file: 322_UCR_id_20_Facility_tr_7482_1st_24005.csv, Data shape: (29931, 1)
Label shape: (29931,)
Train data shape: (7482, 1), Test data shape: (22449, 1), Label test shape: (22449,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 322_UCR_id_20_Facility_tr_7482_1st_24005.csv...
Using CUDA device: 2
Data shape: (22449, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22449, window_size 5000, adding 2551 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 152.85it/s]
Original shapes - Output: (25000,), Label: (22449,)
Aligned shapes - Output: (22449,), Label: (22449,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  58%|█████▊    | 116/200 [00:01<00:00, 115.35it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 138.99it/s]
Table1-only metrics completed in 2.32s
Evaluation Result:  {'AUC-PR': 0.06565818942242788, 'AUC-ROC': 0.9783725104646139, 'VUS-PR': 0.09557978711830414, 'VUS-ROC': 0.9832497322426674, 'Standard-F1': 0.14578005115072756, 'Standard-Precision': 0.08028169014084495, 'Standard-Recall': 0.7916666666666557, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666822222785804, 'Affiliation-P': 0.5000175002675741, 'Affiliation-R': 1.0, 'F1_T': 0.1399046104926426, 'Precision_T': 0.07899461400359052, 'Recall_T': 0.6111111111111027}
Results for 322_UCR_id_20_Facility_tr_7482_1st_24005.csv: {'filename': '322_UCR_id_20_Facility_tr_7482_1st_24005.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 168, 'train_index': '7482', 'data_shape': '29931x1', 'output_length': 25000, 'label_length': 22449, 'aligned_length': 22449, 'AUC-PR': 0.06565818942242788, 'AUC-ROC': 0.9783725104646139, 'VUS-PR': 0.09557978711830414, 'VUS-ROC': 0.9832497322426674, 'Standard-F1': 0.14578005115072756, 'Standard-Precision': 0.08028169014084495, 'Standard-Recall': 0.7916666666666557, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666822222785804, 'Affiliation-P': 0.5000175002675741, 'Affiliation-R': 1.0, 'F1_T': 0.1399046104926426, 'Precision_T': 0.07899461400359052, 'Recall_T': 0.6111111111111027}
Logits metrics skipped for 322_UCR_id_20_Facility_tr_7482_1st_24005.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/322_UCR_id_20_Facility_tr_7482_1st_24005_results.pkl
Processing file: 089_WSD_id_61_WebService_tr_1010_1st_1110.csv, Data shape: (17429, 1)
Label shape: (17429,)
Train data shape: (1010, 1), Test data shape: (16419, 1), Label test shape: (16419,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 089_WSD_id_61_WebService_tr_1010_1st_1110.csv...
Using CUDA device: 2
Data shape: (16419, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 16419, window_size 5000, adding 3581 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.22it/s]
Original shapes - Output: (20000,), Label: (16419,)
Aligned shapes - Output: (16419,), Label: (16419,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  62%|██████▎   | 125/200 [00:01<00:00, 124.77it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 146.64it/s]
Table1-only metrics completed in 6.15s
Evaluation Result:  {'AUC-PR': 0.25103408660047277, 'AUC-ROC': 0.890635764259542, 'VUS-PR': 0.2748066174666184, 'VUS-ROC': 0.9083625334941219, 'Standard-F1': 0.31489361702077556, 'Standard-Precision': 0.32314410480349204, 'Standard-Recall': 0.30705394190871244, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8091648848902994, 'Affiliation-P': 0.7271447870502079, 'Affiliation-R': 0.9120407849100762, 'F1_T': 0.3203285420939546, 'Precision_T': 0.31707317073170604, 'Recall_T': 0.32365145228215636}
Results for 089_WSD_id_61_WebService_tr_1010_1st_1110.csv: {'filename': '089_WSD_id_61_WebService_tr_1010_1st_1110.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '1010', 'data_shape': '17429x1', 'output_length': 20000, 'label_length': 16419, 'aligned_length': 16419, 'AUC-PR': 0.25103408660047277, 'AUC-ROC': 0.890635764259542, 'VUS-PR': 0.2748066174666184, 'VUS-ROC': 0.9083625334941219, 'Standard-F1': 0.31489361702077556, 'Standard-Precision': 0.32314410480349204, 'Standard-Recall': 0.30705394190871244, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8091648848902994, 'Affiliation-P': 0.7271447870502079, 'Affiliation-R': 0.9120407849100762, 'F1_T': 0.3203285420939546, 'Precision_T': 0.31707317073170604, 'Recall_T': 0.32365145228215636}
Logits metrics skipped for 089_WSD_id_61_WebService_tr_1010_1st_1110.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/089_WSD_id_61_WebService_tr_1010_1st_1110_results.pkl
Processing file: 389_UCR_id_87_Facility_tr_7482_1st_18485.csv, Data shape: (29931, 1)
Label shape: (29931,)
Train data shape: (7482, 1), Test data shape: (22449, 1), Label test shape: (22449,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 389_UCR_id_87_Facility_tr_7482_1st_18485.csv...
Using CUDA device: 2
Data shape: (22449, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22449, window_size 5000, adding 2551 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 149.84it/s]
Original shapes - Output: (25000,), Label: (22449,)
Aligned shapes - Output: (22449,), Label: (22449,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  60%|██████    | 120/200 [00:01<00:00, 119.43it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 139.54it/s]
Table1-only metrics completed in 2.39s
Evaluation Result:  {'AUC-PR': 0.026808937174728813, 'AUC-ROC': 0.7435252749406189, 'VUS-PR': 0.028528744616528327, 'VUS-ROC': 0.753470404039549, 'Standard-F1': 0.07574654042233951, 'Standard-Precision': 0.03982233113799969, 'Standard-Recall': 0.7738095238095215, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6984838682853424, 'Affiliation-P': 0.5678901675656562, 'Affiliation-R': 0.9070782662924851, 'F1_T': 0.07577965607685234, 'Precision_T': 0.039840637450199196, 'Recall_T': 0.7738095238095215}
Results for 389_UCR_id_87_Facility_tr_7482_1st_18485.csv: {'filename': '389_UCR_id_87_Facility_tr_7482_1st_18485.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 168, 'train_index': '7482', 'data_shape': '29931x1', 'output_length': 25000, 'label_length': 22449, 'aligned_length': 22449, 'AUC-PR': 0.026808937174728813, 'AUC-ROC': 0.7435252749406189, 'VUS-PR': 0.028528744616528327, 'VUS-ROC': 0.753470404039549, 'Standard-F1': 0.07574654042233951, 'Standard-Precision': 0.03982233113799969, 'Standard-Recall': 0.7738095238095215, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6984838682853424, 'Affiliation-P': 0.5678901675656562, 'Affiliation-R': 0.9070782662924851, 'F1_T': 0.07577965607685234, 'Precision_T': 0.039840637450199196, 'Recall_T': 0.7738095238095215}
Logits metrics skipped for 389_UCR_id_87_Facility_tr_7482_1st_18485.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/389_UCR_id_87_Facility_tr_7482_1st_18485_results.pkl
Processing file: 024_NAB_id_24_Synthetic_tr_1007_1st_2787.csv, Data shape: (4031, 1)
Label shape: (4031,)
Train data shape: (1007, 1), Test data shape: (3024, 1), Label test shape: (3024,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 024_NAB_id_24_Synthetic_tr_1007_1st_2787.csv...
Using CUDA device: 2
Data shape: (3024, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 3024, 1])
batched_mask shape: torch.Size([1, 3024])

100%|██████████| 1/1 [00:00<00:00, 163.93it/s]
Original shapes - Output: (3024,), Label: (3024,)
Aligned shapes - Output: (3024,), Label: (3024,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1131.74it/s]
Table1-only metrics completed in 1.30s
Evaluation Result:  {'AUC-PR': 0.8031405020707514, 'AUC-ROC': 0.9433044312626782, 'VUS-PR': 0.8961273146050446, 'VUS-ROC': 0.9722011841656736, 'Standard-F1': 0.6732869910620608, 'Standard-Precision': 0.6862348178137637, 'Standard-Recall': 0.6608187134502911, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9911932495723914, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9825402626937715, 'F1_T': 0.6726368159198968, 'Precision_T': 0.6869918699186978, 'Recall_T': 0.6588693957114996}
Results for 024_NAB_id_24_Synthetic_tr_1007_1st_2787.csv: {'filename': '024_NAB_id_24_Synthetic_tr_1007_1st_2787.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 288, 'train_index': '1007', 'data_shape': '4031x1', 'output_length': 3024, 'label_length': 3024, 'aligned_length': 3024, 'AUC-PR': 0.8031405020707514, 'AUC-ROC': 0.9433044312626782, 'VUS-PR': 0.8961273146050446, 'VUS-ROC': 0.9722011841656736, 'Standard-F1': 0.6732869910620608, 'Standard-Precision': 0.6862348178137637, 'Standard-Recall': 0.6608187134502911, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9911932495723914, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9825402626937715, 'F1_T': 0.6726368159198968, 'Precision_T': 0.6869918699186978, 'Recall_T': 0.6588693957114996}
Logits metrics skipped for 024_NAB_id_24_Synthetic_tr_1007_1st_2787.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/024_NAB_id_24_Synthetic_tr_1007_1st_2787_results.pkl
Processing file: 735_YAHOO_id_185_WebService_tr_500_1st_20.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 735_YAHOO_id_185_WebService_tr_500_1st_20.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 163.85it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1455.15it/s]
Table1-only metrics completed in 0.41s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.8125, 'VUS-ROC': 0.999601403061224, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9775199419869466, 'Affiliation-P': 0.9560283687943262, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Results for 735_YAHOO_id_185_WebService_tr_500_1st_20.csv: {'filename': '735_YAHOO_id_185_WebService_tr_500_1st_20.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.8125, 'VUS-ROC': 0.999601403061224, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9775199419869466, 'Affiliation-P': 0.9560283687943262, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 735_YAHOO_id_185_WebService_tr_500_1st_20.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/735_YAHOO_id_185_WebService_tr_500_1st_20_results.pkl
Processing file: 721_YAHOO_id_171_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 721_YAHOO_id_171_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 162.49it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1879.37it/s]
Table1-only metrics completed in 0.54s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9901315789473678, 'Affiliation-P': 0.9804560260586319, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 721_YAHOO_id_171_Synthetic_tr_500_1st_893.csv: {'filename': '721_YAHOO_id_171_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9901315789473678, 'Affiliation-P': 0.9804560260586319, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 721_YAHOO_id_171_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/721_YAHOO_id_171_Synthetic_tr_500_1st_893_results.pkl
Processing file: 423_UCR_id_121_Facility_tr_13750_1st_52600.csv, Data shape: (55000, 1)
Label shape: (55000,)
Train data shape: (13750, 1), Test data shape: (41250, 1), Label test shape: (41250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 423_UCR_id_121_Facility_tr_13750_1st_52600.csv...
Using CUDA device: 2
Data shape: (41250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 41250, window_size 5000, adding 3750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([9, 5000, 1])
batched_mask shape: torch.Size([9, 5000])

100%|██████████| 1/1 [00:00<00:00, 114.64it/s]
Original shapes - Output: (45000,), Label: (41250,)
Aligned shapes - Output: (41250,), Label: (41250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  44%|████▍     | 89/200 [00:01<00:01, 88.76it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 100.92it/s]
Table1-only metrics completed in 3.63s
Evaluation Result:  {'AUC-PR': 0.010315883460500719, 'AUC-ROC': 0.77198209500609, 'VUS-PR': 0.01415136228408195, 'VUS-ROC': 0.8183172813950262, 'Standard-F1': 0.023867403314888097, 'Standard-Precision': 0.01211214953271028, 'Standard-Recall': 0.809999999999996, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6890141379346408, 'Affiliation-P': 0.5262674071908798, 'Affiliation-R': 0.9974829696969698, 'F1_T': 0.023799030409843235, 'Precision_T': 0.012076934545996718, 'Recall_T': 0.809999999999996}
Results for 423_UCR_id_121_Facility_tr_13750_1st_52600.csv: {'filename': '423_UCR_id_121_Facility_tr_13750_1st_52600.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 233, 'train_index': '13750', 'data_shape': '55000x1', 'output_length': 45000, 'label_length': 41250, 'aligned_length': 41250, 'AUC-PR': 0.010315883460500719, 'AUC-ROC': 0.77198209500609, 'VUS-PR': 0.01415136228408195, 'VUS-ROC': 0.8183172813950262, 'Standard-F1': 0.023867403314888097, 'Standard-Precision': 0.01211214953271028, 'Standard-Recall': 0.809999999999996, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6890141379346408, 'Affiliation-P': 0.5262674071908798, 'Affiliation-R': 0.9974829696969698, 'F1_T': 0.023799030409843235, 'Precision_T': 0.012076934545996718, 'Recall_T': 0.809999999999996}
Logits metrics skipped for 423_UCR_id_121_Facility_tr_13750_1st_52600.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/423_UCR_id_121_Facility_tr_13750_1st_52600_results.pkl
Processing file: 722_YAHOO_id_172_WebService_tr_500_1st_353.csv, Data shape: (1461, 1)
Label shape: (1461,)
Train data shape: (500, 1), Test data shape: (961, 1), Label test shape: (961,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 722_YAHOO_id_172_WebService_tr_500_1st_353.csv...
Using CUDA device: 2
Data shape: (961, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 961, 1])
batched_mask shape: torch.Size([1, 961])

100%|██████████| 1/1 [00:00<00:00, 154.21it/s]
Original shapes - Output: (961,), Label: (961,)
Aligned shapes - Output: (961,), Label: (961,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2110.25it/s]
Table1-only metrics completed in 0.25s
Evaluation Result:  {'AUC-PR': 0.7104228288051817, 'AUC-ROC': 0.9966878083157154, 'VUS-PR': 0.9577565624037476, 'VUS-ROC': 0.9995819048009222, 'Standard-F1': 0.8484848484843012, 'Standard-Precision': 0.7777777777777346, 'Standard-Recall': 0.9333333333332711, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9941222294262854, 'Affiliation-P': 0.9898166776988108, 'Affiliation-R': 0.9984654017857143, 'F1_T': 0.7999999999994646, 'Precision_T': 0.6999999999999651, 'Recall_T': 0.9333333333332711}
Results for 722_YAHOO_id_172_WebService_tr_500_1st_353.csv: {'filename': '722_YAHOO_id_172_WebService_tr_500_1st_353.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1461x1', 'output_length': 961, 'label_length': 961, 'aligned_length': 961, 'AUC-PR': 0.7104228288051817, 'AUC-ROC': 0.9966878083157154, 'VUS-PR': 0.9577565624037476, 'VUS-ROC': 0.9995819048009222, 'Standard-F1': 0.8484848484843012, 'Standard-Precision': 0.7777777777777346, 'Standard-Recall': 0.9333333333332711, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9941222294262854, 'Affiliation-P': 0.9898166776988108, 'Affiliation-R': 0.9984654017857143, 'F1_T': 0.7999999999994646, 'Precision_T': 0.6999999999999651, 'Recall_T': 0.9333333333332711}
Logits metrics skipped for 722_YAHOO_id_172_WebService_tr_500_1st_353.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/722_YAHOO_id_172_WebService_tr_500_1st_353_results.pkl
Processing file: 287_TODS_id_1_Synthetic_tr_500_1st_11.csv, Data shape: (5000, 1)
Label shape: (5000,)
Train data shape: (500, 1), Test data shape: (4500, 1), Label test shape: (4500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 287_TODS_id_1_Synthetic_tr_500_1st_11.csv...
Using CUDA device: 2
Data shape: (4500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 4500, 1])
batched_mask shape: torch.Size([1, 4500])

100%|██████████| 1/1 [00:00<00:00, 161.77it/s]
Original shapes - Output: (4500,), Label: (4500,)
Aligned shapes - Output: (4500,), Label: (4500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  16%|█▌        | 32/200 [00:01<00:05, 30.81it/s]

Calculating Affiliation Metrics:  32%|███▏      | 63/200 [00:02<00:05, 23.65it/s]

Calculating Affiliation Metrics:  44%|████▍     | 88/200 [00:03<00:05, 21.73it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 111/200 [00:05<00:04, 21.14it/s]

Calculating Affiliation Metrics:  66%|██████▋   | 133/200 [00:06<00:03, 20.72it/s]

Calculating Affiliation Metrics:  81%|████████  | 162/200 [00:07<00:01, 23.08it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:07<00:00, 25.34it/s]
Table1-only metrics completed in 15.80s
Evaluation Result:  {'AUC-PR': 0.4612743689039724, 'AUC-ROC': 0.7680796188670774, 'VUS-PR': 0.9313095976370981, 'VUS-ROC': 0.9636930847370455, 'Standard-F1': 0.49122807017500575, 'Standard-Precision': 0.7850467289719553, 'Standard-Recall': 0.35744680851063676, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6897467631197983, 'Affiliation-P': 0.5459529672161644, 'Affiliation-R': 0.9363687783719739, 'F1_T': 0.49567723342895476, 'Precision_T': 0.767857142857136, 'Recall_T': 0.3659574468085091}
Results for 287_TODS_id_1_Synthetic_tr_500_1st_11.csv: {'filename': '287_TODS_id_1_Synthetic_tr_500_1st_11.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '500', 'data_shape': '5000x1', 'output_length': 4500, 'label_length': 4500, 'aligned_length': 4500, 'AUC-PR': 0.4612743689039724, 'AUC-ROC': 0.7680796188670774, 'VUS-PR': 0.9313095976370981, 'VUS-ROC': 0.9636930847370455, 'Standard-F1': 0.49122807017500575, 'Standard-Precision': 0.7850467289719553, 'Standard-Recall': 0.35744680851063676, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6897467631197983, 'Affiliation-P': 0.5459529672161644, 'Affiliation-R': 0.9363687783719739, 'F1_T': 0.49567723342895476, 'Precision_T': 0.767857142857136, 'Recall_T': 0.3659574468085091}
Logits metrics skipped for 287_TODS_id_1_Synthetic_tr_500_1st_11.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/287_TODS_id_1_Synthetic_tr_500_1st_11_results.pkl
Processing file: 263_IOPS_id_4_WebService_tr_3482_1st_3582.csv, Data shape: (149123, 1)
Label shape: (149123,)
Train data shape: (3482, 1), Test data shape: (145641, 1), Label test shape: (145641,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 263_IOPS_id_4_WebService_tr_3482_1st_3582.csv...
Using CUDA device: 2
Data shape: (145641, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 145641, window_size 5000, adding 4359 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([30, 5000, 1])
batched_mask shape: torch.Size([30, 5000])

100%|██████████| 1/1 [00:00<00:00, 46.97it/s]
Original shapes - Output: (150000,), Label: (145641,)
Aligned shapes - Output: (145641,), Label: (145641,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   8%|▊         | 15/200 [00:01<00:13, 13.76it/s]

Calculating Affiliation Metrics:  14%|█▍        | 29/200 [00:02<00:18,  9.36it/s]

Calculating Affiliation Metrics:  20%|██        | 40/200 [00:04<00:19,  8.19it/s]

Calculating Affiliation Metrics:  24%|██▍       | 49/200 [00:05<00:19,  7.60it/s]

Calculating Affiliation Metrics:  28%|██▊       | 57/200 [00:07<00:19,  7.18it/s]

Calculating Affiliation Metrics:  32%|███▎      | 65/200 [00:08<00:19,  6.86it/s]

Calculating Affiliation Metrics:  36%|███▌      | 72/200 [00:09<00:18,  6.76it/s]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [00:10<00:17,  6.73it/s]

Calculating Affiliation Metrics:  43%|████▎     | 86/200 [00:11<00:16,  6.77it/s]

Calculating Affiliation Metrics:  47%|████▋     | 94/200 [00:12<00:15,  6.97it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 103/200 [00:13<00:13,  7.36it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [00:14<00:11,  7.74it/s]

Calculating Affiliation Metrics:  61%|██████    | 122/200 [00:15<00:09,  8.34it/s]

Calculating Affiliation Metrics:  66%|██████▋   | 133/200 [00:16<00:07,  9.07it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [00:17<00:05,  9.84it/s]

Calculating Affiliation Metrics:  79%|███████▉  | 158/200 [00:18<00:03, 10.53it/s]

Calculating Affiliation Metrics:  86%|████████▋ | 173/200 [00:19<00:02, 11.74it/s]

Calculating Affiliation Metrics:  98%|█████████▊| 195/200 [00:21<00:00, 14.67it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:21<00:00,  9.46it/s]
Table1-only metrics completed in 37.52s
Evaluation Result:  {'AUC-PR': 0.5164196275354365, 'AUC-ROC': 0.9734648760415439, 'VUS-PR': 0.437520468267254, 'VUS-ROC': 0.9906411605534823, 'Standard-F1': 0.40231419670626384, 'Standard-Precision': 0.3102264927934109, 'Standard-Recall': 0.5721518987341765, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9450724114495695, 'Affiliation-P': 0.9479544224231115, 'Affiliation-R': 0.9422078713826113, 'F1_T': 0.40231419670626384, 'Precision_T': 0.3102264927934109, 'Recall_T': 0.5721518987341765}
Results for 263_IOPS_id_4_WebService_tr_3482_1st_3582.csv: {'filename': '263_IOPS_id_4_WebService_tr_3482_1st_3582.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '3482', 'data_shape': '149123x1', 'output_length': 150000, 'label_length': 145641, 'aligned_length': 145641, 'AUC-PR': 0.5164196275354365, 'AUC-ROC': 0.9734648760415439, 'VUS-PR': 0.437520468267254, 'VUS-ROC': 0.9906411605534823, 'Standard-F1': 0.40231419670626384, 'Standard-Precision': 0.3102264927934109, 'Standard-Recall': 0.5721518987341765, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9450724114495695, 'Affiliation-P': 0.9479544224231115, 'Affiliation-R': 0.9422078713826113, 'F1_T': 0.40231419670626384, 'Precision_T': 0.3102264927934109, 'Recall_T': 0.5721518987341765}
Logits metrics skipped for 263_IOPS_id_4_WebService_tr_3482_1st_3582.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/263_IOPS_id_4_WebService_tr_3482_1st_3582_results.pkl
Processing file: 131_WSD_id_103_WebService_tr_4573_1st_7021.csv, Data shape: (18294, 1)
Label shape: (18294,)
Train data shape: (4573, 1), Test data shape: (13721, 1), Label test shape: (13721,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 131_WSD_id_103_WebService_tr_4573_1st_7021.csv...
Using CUDA device: 2
Data shape: (13721, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13721, window_size 5000, adding 1279 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 142.63it/s]
Original shapes - Output: (15000,), Label: (13721,)
Aligned shapes - Output: (13721,), Label: (13721,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 255.69it/s]
Table1-only metrics completed in 0.96s
Evaluation Result:  {'AUC-PR': 0.2839041087455404, 'AUC-ROC': 0.9730533545823887, 'VUS-PR': 0.2606126712979332, 'VUS-ROC': 0.9934852763849089, 'Standard-F1': 0.1324503311256687, 'Standard-Precision': 0.0724637681159415, 'Standard-Recall': 0.7692307692307101, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7594807476206555, 'Affiliation-P': 0.6122470158405238, 'Affiliation-R': 0.9999495439332186, 'F1_T': 0.1324503311256687, 'Precision_T': 0.0724637681159415, 'Recall_T': 0.7692307692307101}
Results for 131_WSD_id_103_WebService_tr_4573_1st_7021.csv: {'filename': '131_WSD_id_103_WebService_tr_4573_1st_7021.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '4573', 'data_shape': '18294x1', 'output_length': 15000, 'label_length': 13721, 'aligned_length': 13721, 'AUC-PR': 0.2839041087455404, 'AUC-ROC': 0.9730533545823887, 'VUS-PR': 0.2606126712979332, 'VUS-ROC': 0.9934852763849089, 'Standard-F1': 0.1324503311256687, 'Standard-Precision': 0.0724637681159415, 'Standard-Recall': 0.7692307692307101, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7594807476206555, 'Affiliation-P': 0.6122470158405238, 'Affiliation-R': 0.9999495439332186, 'F1_T': 0.1324503311256687, 'Precision_T': 0.0724637681159415, 'Recall_T': 0.7692307692307101}
Logits metrics skipped for 131_WSD_id_103_WebService_tr_4573_1st_7021.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/131_WSD_id_103_WebService_tr_4573_1st_7021_results.pkl
Processing file: 415_UCR_id_113_Medical_tr_3000_1st_8612.csv, Data shape: (12000, 1)
Label shape: (12000,)
Train data shape: (3000, 1), Test data shape: (9000, 1), Label test shape: (9000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 415_UCR_id_113_Medical_tr_3000_1st_8612.csv...
Using CUDA device: 2
Data shape: (9000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9000, window_size 5000, adding 1000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.02it/s]
Original shapes - Output: (10000,), Label: (9000,)
Aligned shapes - Output: (9000,), Label: (9000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 438.59it/s]
Table1-only metrics completed in 1.34s
Evaluation Result:  {'AUC-PR': 0.16558673340414176, 'AUC-ROC': 0.97166185320974, 'VUS-PR': 0.20803195020387782, 'VUS-ROC': 0.9773665804039561, 'Standard-F1': 0.3349056603769867, 'Standard-Precision': 0.2218749999999993, 'Standard-Recall': 0.6826923076923012, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666914810333767, 'Affiliation-P': 0.5000279166821004, 'Affiliation-R': 1.0, 'F1_T': 0.3347826086953008, 'Precision_T': 0.21629213483146006, 'Recall_T': 0.7403846153846083}
Results for 415_UCR_id_113_Medical_tr_3000_1st_8612.csv: {'filename': '415_UCR_id_113_Medical_tr_3000_1st_8612.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 218, 'train_index': '3000', 'data_shape': '12000x1', 'output_length': 10000, 'label_length': 9000, 'aligned_length': 9000, 'AUC-PR': 0.16558673340414176, 'AUC-ROC': 0.97166185320974, 'VUS-PR': 0.20803195020387782, 'VUS-ROC': 0.9773665804039561, 'Standard-F1': 0.3349056603769867, 'Standard-Precision': 0.2218749999999993, 'Standard-Recall': 0.6826923076923012, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666914810333767, 'Affiliation-P': 0.5000279166821004, 'Affiliation-R': 1.0, 'F1_T': 0.3347826086953008, 'Precision_T': 0.21629213483146006, 'Recall_T': 0.7403846153846083}
Logits metrics skipped for 415_UCR_id_113_Medical_tr_3000_1st_8612.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/415_UCR_id_113_Medical_tr_3000_1st_8612_results.pkl
Processing file: 499_UCR_id_197_Environment_tr_2046_1st_5703.csv, Data shape: (8184, 1)
Label shape: (8184,)
Train data shape: (2046, 1), Test data shape: (6138, 1), Label test shape: (6138,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 499_UCR_id_197_Environment_tr_2046_1st_5703.csv...
Using CUDA device: 2
Data shape: (6138, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6138, window_size 5000, adding 3862 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.75it/s]
Original shapes - Output: (10000,), Label: (6138,)
Aligned shapes - Output: (6138,), Label: (6138,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 500.87it/s]
Table1-only metrics completed in 0.52s
Evaluation Result:  {'AUC-PR': 0.07159797240743856, 'AUC-ROC': 0.9566841129647803, 'VUS-PR': 0.06804198096353431, 'VUS-ROC': 0.9625316770293384, 'Standard-F1': 0.15533980582503537, 'Standard-Precision': 0.08791208791208743, 'Standard-Recall': 0.666666666666639, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7537083864425881, 'Affiliation-P': 0.604760858729688, 'Affiliation-R': 1.0, 'F1_T': 0.15458937198046985, 'Precision_T': 0.08743169398907057, 'Recall_T': 0.666666666666639}
Results for 499_UCR_id_197_Environment_tr_2046_1st_5703.csv: {'filename': '499_UCR_id_197_Environment_tr_2046_1st_5703.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '2046', 'data_shape': '8184x1', 'output_length': 10000, 'label_length': 6138, 'aligned_length': 6138, 'AUC-PR': 0.07159797240743856, 'AUC-ROC': 0.9566841129647803, 'VUS-PR': 0.06804198096353431, 'VUS-ROC': 0.9625316770293384, 'Standard-F1': 0.15533980582503537, 'Standard-Precision': 0.08791208791208743, 'Standard-Recall': 0.666666666666639, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7537083864425881, 'Affiliation-P': 0.604760858729688, 'Affiliation-R': 1.0, 'F1_T': 0.15458937198046985, 'Precision_T': 0.08743169398907057, 'Recall_T': 0.666666666666639}
Logits metrics skipped for 499_UCR_id_197_Environment_tr_2046_1st_5703.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/499_UCR_id_197_Environment_tr_2046_1st_5703_results.pkl
Processing file: 044_WSD_id_16_WebService_tr_4559_1st_6975.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 044_WSD_id_16_WebService_tr_4559_1st_6975.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.48it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 301.73it/s]
Table1-only metrics completed in 1.24s
Evaluation Result:  {'AUC-PR': 0.017235464550425218, 'AUC-ROC': 0.4215566599486992, 'VUS-PR': 0.013281850076992006, 'VUS-ROC': 0.46024667241713996, 'Standard-F1': 0.035502958579574244, 'Standard-Precision': 0.021897810218977944, 'Standard-Recall': 0.09374999999999707, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6682851911191563, 'Affiliation-P': 0.5018233406951693, 'Affiliation-R': 0.9999988575711047, 'F1_T': 0.035502958579574244, 'Precision_T': 0.021897810218977944, 'Recall_T': 0.09374999999999707}
Results for 044_WSD_id_16_WebService_tr_4559_1st_6975.csv: {'filename': '044_WSD_id_16_WebService_tr_4559_1st_6975.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.017235464550425218, 'AUC-ROC': 0.4215566599486992, 'VUS-PR': 0.013281850076992006, 'VUS-ROC': 0.46024667241713996, 'Standard-F1': 0.035502958579574244, 'Standard-Precision': 0.021897810218977944, 'Standard-Recall': 0.09374999999999707, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6682851911191563, 'Affiliation-P': 0.5018233406951693, 'Affiliation-R': 0.9999988575711047, 'F1_T': 0.035502958579574244, 'Precision_T': 0.021897810218977944, 'Recall_T': 0.09374999999999707}
Logits metrics skipped for 044_WSD_id_16_WebService_tr_4559_1st_6975.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/044_WSD_id_16_WebService_tr_4559_1st_6975_results.pkl
Processing file: 640_YAHOO_id_90_WebService_tr_500_1st_18.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 640_YAHOO_id_90_WebService_tr_500_1st_18.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 165.40it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 529.60it/s]
Table1-only metrics completed in 0.68s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9880560957740645, 'VUS-ROC': 0.9999233318860189, 'Standard-F1': 0.9565217391298525, 'Standard-Precision': 0.9166666666665902, 'Standard-Recall': 0.9999999999999091, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9995497523637996, 'Affiliation-P': 0.9990999099909992, 'Affiliation-R': 1.0, 'F1_T': 0.9565217391298525, 'Precision_T': 0.9166666666665902, 'Recall_T': 0.9999999999999091}
Results for 640_YAHOO_id_90_WebService_tr_500_1st_18.csv: {'filename': '640_YAHOO_id_90_WebService_tr_500_1st_18.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 11, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9880560957740645, 'VUS-ROC': 0.9999233318860189, 'Standard-F1': 0.9565217391298525, 'Standard-Precision': 0.9166666666665902, 'Standard-Recall': 0.9999999999999091, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9995497523637996, 'Affiliation-P': 0.9990999099909992, 'Affiliation-R': 1.0, 'F1_T': 0.9565217391298525, 'Precision_T': 0.9166666666665902, 'Recall_T': 0.9999999999999091}
Logits metrics skipped for 640_YAHOO_id_90_WebService_tr_500_1st_18.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/640_YAHOO_id_90_WebService_tr_500_1st_18_results.pkl
Processing file: 523_UCR_id_221_Facility_tr_3000_1st_6527.csv, Data shape: (12000, 1)
Label shape: (12000,)
Train data shape: (3000, 1), Test data shape: (9000, 1), Label test shape: (9000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 523_UCR_id_221_Facility_tr_3000_1st_6527.csv...
Using CUDA device: 2
Data shape: (9000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9000, window_size 5000, adding 1000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.86it/s]
Original shapes - Output: (10000,), Label: (9000,)
Aligned shapes - Output: (9000,), Label: (9000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 527.18it/s]
Table1-only metrics completed in 1.26s
Evaluation Result:  {'AUC-PR': 0.6100758922851612, 'AUC-ROC': 0.8933903648208718, 'VUS-PR': 0.5722787337322855, 'VUS-ROC': 0.9405529925325052, 'Standard-F1': 0.5961538461533494, 'Standard-Precision': 0.6888888888888812, 'Standard-Recall': 0.5254237288135549, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9983877060107579, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9967806026365348, 'F1_T': 0.5961538461533494, 'Precision_T': 0.6888888888888812, 'Recall_T': 0.5254237288135549}
Results for 523_UCR_id_221_Facility_tr_3000_1st_6527.csv: {'filename': '523_UCR_id_221_Facility_tr_3000_1st_6527.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 218, 'train_index': '3000', 'data_shape': '12000x1', 'output_length': 10000, 'label_length': 9000, 'aligned_length': 9000, 'AUC-PR': 0.6100758922851612, 'AUC-ROC': 0.8933903648208718, 'VUS-PR': 0.5722787337322855, 'VUS-ROC': 0.9405529925325052, 'Standard-F1': 0.5961538461533494, 'Standard-Precision': 0.6888888888888812, 'Standard-Recall': 0.5254237288135549, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9983877060107579, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9967806026365348, 'F1_T': 0.5961538461533494, 'Precision_T': 0.6888888888888812, 'Recall_T': 0.5254237288135549}
Logits metrics skipped for 523_UCR_id_221_Facility_tr_3000_1st_6527.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/523_UCR_id_221_Facility_tr_3000_1st_6527_results.pkl
Skipping file: 202_SMD_id_25_Facility_tr_6250_1st_21230.csv due to filter criteria.
Processing file: 290_TODS_id_4_Synthetic_tr_500_1st_26.csv, Data shape: (5000, 1)
Label shape: (5000,)
Train data shape: (500, 1), Test data shape: (4500, 1), Label test shape: (4500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 290_TODS_id_4_Synthetic_tr_500_1st_26.csv...
Using CUDA device: 2
Data shape: (4500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 4500, 1])
batched_mask shape: torch.Size([1, 4500])

100%|██████████| 1/1 [00:00<00:00, 159.35it/s]
Original shapes - Output: (4500,), Label: (4500,)
Aligned shapes - Output: (4500,), Label: (4500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  25%|██▌       | 50/200 [00:01<00:03, 48.69it/s]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [00:02<00:03, 32.69it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 135/200 [00:04<00:02, 31.18it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:04<00:00, 41.58it/s]
Table1-only metrics completed in 10.93s
Evaluation Result:  {'AUC-PR': 0.9161108154645464, 'AUC-ROC': 0.964924460137988, 'VUS-PR': 0.9249709833129275, 'VUS-ROC': 0.9712978365041872, 'Standard-F1': 0.8911651728548136, 'Standard-Precision': 0.9560439560439533, 'Standard-Recall': 0.8345323741007173, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8024873052305797, 'Affiliation-P': 0.789756792099377, 'Affiliation-R': 0.8156349624635222, 'F1_T': 0.8830188679240272, 'Precision_T': 0.928571428571426, 'Recall_T': 0.841726618705034}
Results for 290_TODS_id_4_Synthetic_tr_500_1st_26.csv: {'filename': '290_TODS_id_4_Synthetic_tr_500_1st_26.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '500', 'data_shape': '5000x1', 'output_length': 4500, 'label_length': 4500, 'aligned_length': 4500, 'AUC-PR': 0.9161108154645464, 'AUC-ROC': 0.964924460137988, 'VUS-PR': 0.9249709833129275, 'VUS-ROC': 0.9712978365041872, 'Standard-F1': 0.8911651728548136, 'Standard-Precision': 0.9560439560439533, 'Standard-Recall': 0.8345323741007173, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8024873052305797, 'Affiliation-P': 0.789756792099377, 'Affiliation-R': 0.8156349624635222, 'F1_T': 0.8830188679240272, 'Precision_T': 0.928571428571426, 'Recall_T': 0.841726618705034}
Logits metrics skipped for 290_TODS_id_4_Synthetic_tr_500_1st_26.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/290_TODS_id_4_Synthetic_tr_500_1st_26_results.pkl
Skipping file: 188_SMD_id_11_Facility_tr_500_1st_392.csv due to filter criteria.
Processing file: 603_YAHOO_id_53_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 603_YAHOO_id_53_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 75.38it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2212.17it/s]
Table1-only metrics completed in 0.50s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7479608482871121, 'Affiliation-P': 0.5973941368078176, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 603_YAHOO_id_53_Synthetic_tr_500_1st_893.csv: {'filename': '603_YAHOO_id_53_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7479608482871121, 'Affiliation-P': 0.5973941368078176, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 603_YAHOO_id_53_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/603_YAHOO_id_53_Synthetic_tr_500_1st_893_results.pkl
Processing file: 124_WSD_id_96_WebService_tr_4559_1st_14350.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 124_WSD_id_96_WebService_tr_4559_1st_14350.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.23it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 297.05it/s]
Table1-only metrics completed in 0.77s
Evaluation Result:  {'AUC-PR': 0.8364335664335665, 'AUC-ROC': 0.9959244896465941, 'VUS-PR': 0.26042938939260607, 'VUS-ROC': 0.9969379646971143, 'Standard-F1': 0.12244897959170827, 'Standard-Precision': 0.06569343065693382, 'Standard-Recall': 0.89999999999991, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7163551868218907, 'Affiliation-P': 0.5580656755272388, 'Affiliation-R': 0.9999926884550705, 'F1_T': 0.12244897959170827, 'Precision_T': 0.06569343065693382, 'Recall_T': 0.89999999999991}
Results for 124_WSD_id_96_WebService_tr_4559_1st_14350.csv: {'filename': '124_WSD_id_96_WebService_tr_4559_1st_14350.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 7, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.8364335664335665, 'AUC-ROC': 0.9959244896465941, 'VUS-PR': 0.26042938939260607, 'VUS-ROC': 0.9969379646971143, 'Standard-F1': 0.12244897959170827, 'Standard-Precision': 0.06569343065693382, 'Standard-Recall': 0.89999999999991, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7163551868218907, 'Affiliation-P': 0.5580656755272388, 'Affiliation-R': 0.9999926884550705, 'F1_T': 0.12244897959170827, 'Precision_T': 0.06569343065693382, 'Recall_T': 0.89999999999991}
Logits metrics skipped for 124_WSD_id_96_WebService_tr_4559_1st_14350.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/124_WSD_id_96_WebService_tr_4559_1st_14350_results.pkl
Skipping file: 146_MSL_id_7_Sensor_tr_554_1st_1172.csv due to filter criteria.
Skipping file: 211_SMD_id_34_Facility_tr_5925_1st_17580.csv due to filter criteria.
Processing file: 753_YAHOO_id_203_WebService_tr_500_1st_6.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 753_YAHOO_id_203_WebService_tr_500_1st_6.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.95it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 880.78it/s]
Table1-only metrics completed in 0.76s
Evaluation Result:  {'AUC-PR': 0.9999999999999999, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.9090909090903305, 'Standard-Precision': 0.8333333333332639, 'Standard-Recall': 0.9999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9978020307462322, 'Affiliation-P': 0.995613702439641, 'Affiliation-R': 1.0, 'F1_T': 0.9090909090903305, 'Precision_T': 0.8333333333332639, 'Recall_T': 0.9999999999999}
Results for 753_YAHOO_id_203_WebService_tr_500_1st_6.csv: {'filename': '753_YAHOO_id_203_WebService_tr_500_1st_6.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9999999999999999, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.9090909090903305, 'Standard-Precision': 0.8333333333332639, 'Standard-Recall': 0.9999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9978020307462322, 'Affiliation-P': 0.995613702439641, 'Affiliation-R': 1.0, 'F1_T': 0.9090909090903305, 'Precision_T': 0.8333333333332639, 'Recall_T': 0.9999999999999}
Logits metrics skipped for 753_YAHOO_id_203_WebService_tr_500_1st_6.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/753_YAHOO_id_203_WebService_tr_500_1st_6_results.pkl
Processing file: 047_WSD_id_19_WebService_tr_4559_1st_16362.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 047_WSD_id_19_WebService_tr_4559_1st_16362.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 154.66it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 308.80it/s]
Table1-only metrics completed in 0.77s
Evaluation Result:  {'AUC-PR': 0.7273905891027739, 'AUC-ROC': 0.9995869972083102, 'VUS-PR': 0.3212470264957523, 'VUS-ROC': 0.9986338103238861, 'Standard-F1': 0.18543046357598822, 'Standard-Precision': 0.10218978102189707, 'Standard-Recall': 0.9999999999999286, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7213523330378806, 'Affiliation-P': 0.5641525430939937, 'Affiliation-R': 1.0, 'F1_T': 0.18543046357598822, 'Precision_T': 0.10218978102189707, 'Recall_T': 0.9999999999999286}
Results for 047_WSD_id_19_WebService_tr_4559_1st_16362.csv: {'filename': '047_WSD_id_19_WebService_tr_4559_1st_16362.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 15, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.7273905891027739, 'AUC-ROC': 0.9995869972083102, 'VUS-PR': 0.3212470264957523, 'VUS-ROC': 0.9986338103238861, 'Standard-F1': 0.18543046357598822, 'Standard-Precision': 0.10218978102189707, 'Standard-Recall': 0.9999999999999286, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7213523330378806, 'Affiliation-P': 0.5641525430939937, 'Affiliation-R': 1.0, 'F1_T': 0.18543046357598822, 'Precision_T': 0.10218978102189707, 'Recall_T': 0.9999999999999286}
Logits metrics skipped for 047_WSD_id_19_WebService_tr_4559_1st_16362.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/047_WSD_id_19_WebService_tr_4559_1st_16362_results.pkl
Skipping file: 811_Exathlon_id_2_Facility_tr_10766_1st_12590.csv due to filter criteria.
Skipping file: 208_SMD_id_31_Facility_tr_5000_1st_15144.csv due to filter criteria.
Skipping file: 176_MITDB_id_7_Medical_tr_6232_1st_6332.csv due to filter criteria.
Processing file: 039_WSD_id_11_WebService_tr_1746_1st_1846.csv, Data shape: (18219, 1)
Label shape: (18219,)
Train data shape: (1746, 1), Test data shape: (16473, 1), Label test shape: (16473,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 039_WSD_id_11_WebService_tr_1746_1st_1846.csv...
Using CUDA device: 2
Data shape: (16473, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 16473, window_size 5000, adding 3527 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.63it/s]
Original shapes - Output: (20000,), Label: (16473,)
Aligned shapes - Output: (16473,), Label: (16473,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 206.31it/s]
Table1-only metrics completed in 2.70s
Evaluation Result:  {'AUC-PR': 0.6844181213614187, 'AUC-ROC': 0.9692876071157694, 'VUS-PR': 0.6911373681919069, 'VUS-ROC': 0.9882028214643037, 'Standard-F1': 0.5791245791240813, 'Standard-Precision': 0.521212121212118, 'Standard-Recall': 0.6515151515151466, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.963107421700855, 'Affiliation-P': 0.9397437807239684, 'Affiliation-R': 0.9876624017214882, 'F1_T': 0.5791245791240813, 'Precision_T': 0.521212121212118, 'Recall_T': 0.6515151515151466}
Results for 039_WSD_id_11_WebService_tr_1746_1st_1846.csv: {'filename': '039_WSD_id_11_WebService_tr_1746_1st_1846.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '1746', 'data_shape': '18219x1', 'output_length': 20000, 'label_length': 16473, 'aligned_length': 16473, 'AUC-PR': 0.6844181213614187, 'AUC-ROC': 0.9692876071157694, 'VUS-PR': 0.6911373681919069, 'VUS-ROC': 0.9882028214643037, 'Standard-F1': 0.5791245791240813, 'Standard-Precision': 0.521212121212118, 'Standard-Recall': 0.6515151515151466, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.963107421700855, 'Affiliation-P': 0.9397437807239684, 'Affiliation-R': 0.9876624017214882, 'F1_T': 0.5791245791240813, 'Precision_T': 0.521212121212118, 'Recall_T': 0.6515151515151466}
Logits metrics skipped for 039_WSD_id_11_WebService_tr_1746_1st_1846.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/039_WSD_id_11_WebService_tr_1746_1st_1846_results.pkl
Processing file: 351_UCR_id_49_Medical_tr_7464_1st_12190.csv, Data shape: (29859, 1)
Label shape: (29859,)
Train data shape: (7464, 1), Test data shape: (22395, 1), Label test shape: (22395,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 351_UCR_id_49_Medical_tr_7464_1st_12190.csv...
Using CUDA device: 2
Data shape: (22395, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22395, window_size 5000, adding 2605 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 151.14it/s]
Original shapes - Output: (25000,), Label: (22395,)
Aligned shapes - Output: (22395,), Label: (22395,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 111/200 [00:01<00:00, 110.84it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 125.30it/s]
Table1-only metrics completed in 1.86s
Evaluation Result:  {'AUC-PR': 0.23559912799917782, 'AUC-ROC': 0.9767546758991359, 'VUS-PR': 0.2510986781665996, 'VUS-ROC': 0.978122585701894, 'Standard-F1': 0.43026706231408923, 'Standard-Precision': 0.32657657657657585, 'Standard-Recall': 0.6304347826086929, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.813787870814358, 'Affiliation-P': 0.6862546527797374, 'Affiliation-R': 0.9995423085510159, 'F1_T': 0.42962962962917906, 'Precision_T': 0.3258426966292127, 'Recall_T': 0.6304347826086929}
Results for 351_UCR_id_49_Medical_tr_7464_1st_12190.csv: {'filename': '351_UCR_id_49_Medical_tr_7464_1st_12190.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 33, 'train_index': '7464', 'data_shape': '29859x1', 'output_length': 25000, 'label_length': 22395, 'aligned_length': 22395, 'AUC-PR': 0.23559912799917782, 'AUC-ROC': 0.9767546758991359, 'VUS-PR': 0.2510986781665996, 'VUS-ROC': 0.978122585701894, 'Standard-F1': 0.43026706231408923, 'Standard-Precision': 0.32657657657657585, 'Standard-Recall': 0.6304347826086929, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.813787870814358, 'Affiliation-P': 0.6862546527797374, 'Affiliation-R': 0.9995423085510159, 'F1_T': 0.42962962962917906, 'Precision_T': 0.3258426966292127, 'Recall_T': 0.6304347826086929}
Logits metrics skipped for 351_UCR_id_49_Medical_tr_7464_1st_12190.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/351_UCR_id_49_Medical_tr_7464_1st_12190_results.pkl
Skipping file: 545_SMAP_id_15_Sensor_tr_1173_1st_2750.csv due to filter criteria.
Processing file: 301_TODS_id_15_Synthetic_tr_500_1st_245.csv, Data shape: (5000, 1)
Label shape: (5000,)
Train data shape: (500, 1), Test data shape: (4500, 1), Label test shape: (4500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 301_TODS_id_15_Synthetic_tr_500_1st_245.csv...
Using CUDA device: 2
Data shape: (4500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 4500, 1])
batched_mask shape: torch.Size([1, 4500])

100%|██████████| 1/1 [00:00<00:00, 159.44it/s]
Original shapes - Output: (4500,), Label: (4500,)
Aligned shapes - Output: (4500,), Label: (4500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 465.34it/s]
Table1-only metrics completed in 2.82s
Evaluation Result:  {'AUC-PR': 0.9364818544650781, 'AUC-ROC': 0.9875792682926829, 'VUS-PR': 0.9400171188451172, 'VUS-ROC': 0.9831519494291815, 'Standard-F1': 0.8732394366192162, 'Standard-Precision': 0.8950131233595776, 'Standard-Recall': 0.8524999999999978, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9303894273886284, 'Affiliation-P': 0.9481518804924228, 'Affiliation-R': 0.9132802511424001, 'F1_T': 0.8714652956293181, 'Precision_T': 0.8968253968253944, 'Recall_T': 0.8474999999999978}
Results for 301_TODS_id_15_Synthetic_tr_500_1st_245.csv: {'filename': '301_TODS_id_15_Synthetic_tr_500_1st_245.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 99, 'train_index': '500', 'data_shape': '5000x1', 'output_length': 4500, 'label_length': 4500, 'aligned_length': 4500, 'AUC-PR': 0.9364818544650781, 'AUC-ROC': 0.9875792682926829, 'VUS-PR': 0.9400171188451172, 'VUS-ROC': 0.9831519494291815, 'Standard-F1': 0.8732394366192162, 'Standard-Precision': 0.8950131233595776, 'Standard-Recall': 0.8524999999999978, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9303894273886284, 'Affiliation-P': 0.9481518804924228, 'Affiliation-R': 0.9132802511424001, 'F1_T': 0.8714652956293181, 'Precision_T': 0.8968253968253944, 'Recall_T': 0.8474999999999978}
Logits metrics skipped for 301_TODS_id_15_Synthetic_tr_500_1st_245.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/301_TODS_id_15_Synthetic_tr_500_1st_245_results.pkl
Processing file: 437_UCR_id_135_Medical_tr_2499_1st_8000.csv, Data shape: (9998, 1)
Label shape: (9998,)
Train data shape: (2499, 1), Test data shape: (7499, 1), Label test shape: (7499,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 437_UCR_id_135_Medical_tr_2499_1st_8000.csv...
Using CUDA device: 2
Data shape: (7499, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 7499, window_size 5000, adding 2501 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 156.53it/s]
Original shapes - Output: (10000,), Label: (7499,)
Aligned shapes - Output: (7499,), Label: (7499,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 485.24it/s]
Table1-only metrics completed in 0.55s
Evaluation Result:  {'AUC-PR': 0.2727096907351155, 'AUC-ROC': 0.9915975381321916, 'VUS-PR': 0.3730974552257258, 'VUS-ROC': 0.9958597700773866, 'Standard-F1': 0.4199999999996166, 'Standard-Precision': 0.2799999999999963, 'Standard-Recall': 0.8399999999999664, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9349564448037156, 'Affiliation-P': 0.878096026838666, 'Affiliation-R': 0.9996906254167223, 'F1_T': 0.4199999999996166, 'Precision_T': 0.2799999999999963, 'Recall_T': 0.8399999999999664}
Results for 437_UCR_id_135_Medical_tr_2499_1st_8000.csv: {'filename': '437_UCR_id_135_Medical_tr_2499_1st_8000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 26, 'train_index': '2499', 'data_shape': '9998x1', 'output_length': 10000, 'label_length': 7499, 'aligned_length': 7499, 'AUC-PR': 0.2727096907351155, 'AUC-ROC': 0.9915975381321916, 'VUS-PR': 0.3730974552257258, 'VUS-ROC': 0.9958597700773866, 'Standard-F1': 0.4199999999996166, 'Standard-Precision': 0.2799999999999963, 'Standard-Recall': 0.8399999999999664, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9349564448037156, 'Affiliation-P': 0.878096026838666, 'Affiliation-R': 0.9996906254167223, 'F1_T': 0.4199999999996166, 'Precision_T': 0.2799999999999963, 'Recall_T': 0.8399999999999664}
Logits metrics skipped for 437_UCR_id_135_Medical_tr_2499_1st_8000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/437_UCR_id_135_Medical_tr_2499_1st_8000_results.pkl
Processing file: 054_WSD_id_26_WebService_tr_2409_1st_2509.csv, Data shape: (18391, 1)
Label shape: (18391,)
Train data shape: (2409, 1), Test data shape: (15982, 1), Label test shape: (15982,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 054_WSD_id_26_WebService_tr_2409_1st_2509.csv...
Using CUDA device: 2
Data shape: (15982, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 15982, window_size 5000, adding 4018 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.18it/s]
Original shapes - Output: (20000,), Label: (15982,)
Aligned shapes - Output: (15982,), Label: (15982,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 218.23it/s]
Table1-only metrics completed in 2.62s
Evaluation Result:  {'AUC-PR': 0.14003499083397755, 'AUC-ROC': 0.6941699604743083, 'VUS-PR': 0.07056321244346997, 'VUS-ROC': 0.9344618943584871, 'Standard-F1': 0.08849557522082466, 'Standard-Precision': 0.06249999999999961, 'Standard-Recall': 0.15151515151514924, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9805347348042278, 'Affiliation-P': 0.9649102783080478, 'Affiliation-R': 0.9966735224643045, 'F1_T': 0.08849557522082466, 'Precision_T': 0.06249999999999961, 'Recall_T': 0.15151515151514924}
Results for 054_WSD_id_26_WebService_tr_2409_1st_2509.csv: {'filename': '054_WSD_id_26_WebService_tr_2409_1st_2509.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '2409', 'data_shape': '18391x1', 'output_length': 20000, 'label_length': 15982, 'aligned_length': 15982, 'AUC-PR': 0.14003499083397755, 'AUC-ROC': 0.6941699604743083, 'VUS-PR': 0.07056321244346997, 'VUS-ROC': 0.9344618943584871, 'Standard-F1': 0.08849557522082466, 'Standard-Precision': 0.06249999999999961, 'Standard-Recall': 0.15151515151514924, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9805347348042278, 'Affiliation-P': 0.9649102783080478, 'Affiliation-R': 0.9966735224643045, 'F1_T': 0.08849557522082466, 'Precision_T': 0.06249999999999961, 'Recall_T': 0.15151515151514924}
Logits metrics skipped for 054_WSD_id_26_WebService_tr_2409_1st_2509.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/054_WSD_id_26_WebService_tr_2409_1st_2509_results.pkl
Processing file: 315_UCR_id_13_HumanActivity_tr_16000_1st_33070.csv, Data shape: (64000, 1)
Label shape: (64000,)
Train data shape: (16000, 1), Test data shape: (48000, 1), Label test shape: (48000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 315_UCR_id_13_HumanActivity_tr_16000_1st_33070.csv...
Using CUDA device: 2
Data shape: (48000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 48000, window_size 5000, adding 2000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([10, 5000, 1])
batched_mask shape: torch.Size([10, 5000])

100%|██████████| 1/1 [00:00<00:00, 96.22it/s]
Original shapes - Output: (50000,), Label: (48000,)
Aligned shapes - Output: (48000,), Label: (48000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  45%|████▌     | 90/200 [00:01<00:01, 89.20it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 107.31it/s]
Table1-only metrics completed in 2.90s
Evaluation Result:  {'AUC-PR': 0.01284012652531941, 'AUC-ROC': 0.910479413048843, 'VUS-PR': 0.013949454711426897, 'VUS-ROC': 0.9195492608627903, 'Standard-F1': 0.03419927655369546, 'Standard-Precision': 0.017741385192766967, 'Standard-Recall': 0.47272727272726844, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6840208312775057, 'Affiliation-P': 0.5197808958796282, 'Affiliation-R': 1.0, 'F1_T': 0.03384518959567143, 'Precision_T': 0.0175267770204479, 'Recall_T': 0.49090909090908647}
Results for 315_UCR_id_13_HumanActivity_tr_16000_1st_33070.csv: {'filename': '315_UCR_id_13_HumanActivity_tr_16000_1st_33070.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '16000', 'data_shape': '64000x1', 'output_length': 50000, 'label_length': 48000, 'aligned_length': 48000, 'AUC-PR': 0.01284012652531941, 'AUC-ROC': 0.910479413048843, 'VUS-PR': 0.013949454711426897, 'VUS-ROC': 0.9195492608627903, 'Standard-F1': 0.03419927655369546, 'Standard-Precision': 0.017741385192766967, 'Standard-Recall': 0.47272727272726844, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6840208312775057, 'Affiliation-P': 0.5197808958796282, 'Affiliation-R': 1.0, 'F1_T': 0.03384518959567143, 'Precision_T': 0.0175267770204479, 'Recall_T': 0.49090909090908647}
Logits metrics skipped for 315_UCR_id_13_HumanActivity_tr_16000_1st_33070.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/315_UCR_id_13_HumanActivity_tr_16000_1st_33070_results.pkl
Processing file: 587_YAHOO_id_37_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 587_YAHOO_id_37_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 159.63it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1702.81it/s]
Table1-only metrics completed in 0.36s
Evaluation Result:  {'AUC-PR': 0.9459876543209877, 'AUC-ROC': 0.9993908382066277, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984359113984375, 'Affiliation-P': 0.9969624510347107, 'Affiliation-R': 0.9999137336093858, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 587_YAHOO_id_37_Synthetic_tr_500_1st_623.csv: {'filename': '587_YAHOO_id_37_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 35, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9459876543209877, 'AUC-ROC': 0.9993908382066277, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984359113984375, 'Affiliation-P': 0.9969624510347107, 'Affiliation-R': 0.9999137336093858, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 587_YAHOO_id_37_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/587_YAHOO_id_37_Synthetic_tr_500_1st_623_results.pkl
Processing file: 606_YAHOO_id_56_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 606_YAHOO_id_56_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 162.51it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1601.30it/s]
Table1-only metrics completed in 1.32s
Evaluation Result:  {'AUC-PR': 0.9085858585858586, 'AUC-ROC': 0.9989035087719298, 'VUS-PR': 0.996865867206443, 'VUS-ROC': 0.999971294636075, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Results for 606_YAHOO_id_56_Synthetic_tr_500_1st_623.csv: {'filename': '606_YAHOO_id_56_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 235, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9085858585858586, 'AUC-ROC': 0.9989035087719298, 'VUS-PR': 0.996865867206443, 'VUS-ROC': 0.999971294636075, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 606_YAHOO_id_56_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/606_YAHOO_id_56_Synthetic_tr_500_1st_623_results.pkl
Processing file: 692_YAHOO_id_142_WebService_tr_500_1st_294.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 692_YAHOO_id_142_WebService_tr_500_1st_294.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 162.37it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 795.78it/s]
Table1-only metrics completed in 0.81s
Evaluation Result:  {'AUC-PR': 0.9999999999999999, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.9090909090903305, 'Standard-Precision': 0.8333333333332639, 'Standard-Recall': 0.9999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990321148913486, 'Affiliation-P': 0.9980661015741863, 'Affiliation-R': 1.0, 'F1_T': 0.9090909090903305, 'Precision_T': 0.8333333333332639, 'Recall_T': 0.9999999999999}
Results for 692_YAHOO_id_142_WebService_tr_500_1st_294.csv: {'filename': '692_YAHOO_id_142_WebService_tr_500_1st_294.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9999999999999999, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.9090909090903305, 'Standard-Precision': 0.8333333333332639, 'Standard-Recall': 0.9999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990321148913486, 'Affiliation-P': 0.9980661015741863, 'Affiliation-R': 1.0, 'F1_T': 0.9090909090903305, 'Precision_T': 0.8333333333332639, 'Recall_T': 0.9999999999999}
Logits metrics skipped for 692_YAHOO_id_142_WebService_tr_500_1st_294.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/692_YAHOO_id_142_WebService_tr_500_1st_294_results.pkl
Skipping file: 192_SMD_id_15_Facility_tr_5925_1st_17580.csv due to filter criteria.
Processing file: 153_Stock_id_5_Finance_tr_500_1st_1.csv, Data shape: (10000, 1)
Label shape: (10000,)
Train data shape: (500, 1), Test data shape: (9500, 1), Label test shape: (9500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 153_Stock_id_5_Finance_tr_500_1st_1.csv...
Using CUDA device: 2
Data shape: (9500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 166.77it/s]
Original shapes - Output: (10000,), Label: (9500,)
Aligned shapes - Output: (9500,), Label: (9500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  12%|█▏        | 23/200 [00:01<00:08, 22.12it/s]

Calculating Affiliation Metrics:  23%|██▎       | 46/200 [00:03<00:11, 13.64it/s]

Calculating Affiliation Metrics:  31%|███       | 62/200 [00:05<00:12, 10.87it/s]

Calculating Affiliation Metrics:  38%|███▊      | 75/200 [00:06<00:13,  9.51it/s]

Calculating Affiliation Metrics:  43%|████▎     | 86/200 [00:08<00:13,  8.31it/s]

Calculating Affiliation Metrics:  48%|████▊     | 95/200 [00:10<00:13,  7.52it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 103/200 [00:11<00:13,  7.05it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 111/200 [00:12<00:13,  6.78it/s]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [00:13<00:12,  6.78it/s]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [00:15<00:10,  7.00it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 135/200 [00:16<00:08,  7.42it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [00:17<00:06,  8.02it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 157/200 [00:18<00:04,  9.10it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 176/200 [00:19<00:02, 11.81it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:19<00:00, 10.21it/s]
Table1-only metrics completed in 36.41s
Evaluation Result:  {'AUC-PR': 0.07157413553897266, 'AUC-ROC': 0.5963062001227747, 'VUS-PR': 0.8338200081876661, 'VUS-ROC': 0.9343145010887977, 'Standard-F1': 0.1304915180510117, 'Standard-Precision': 0.08112493239588962, 'Standard-Recall': 0.3333333333333326, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6748127029888804, 'Affiliation-P': 0.5108775552977954, 'Affiliation-R': 0.9936711581757911, 'F1_T': 0.130677290836359, 'Precision_T': 0.07961165048543686, 'Recall_T': 0.3644444444444436}
Results for 153_Stock_id_5_Finance_tr_500_1st_1.csv: {'filename': '153_Stock_id_5_Finance_tr_500_1st_1.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '10000x1', 'output_length': 10000, 'label_length': 9500, 'aligned_length': 9500, 'AUC-PR': 0.07157413553897266, 'AUC-ROC': 0.5963062001227747, 'VUS-PR': 0.8338200081876661, 'VUS-ROC': 0.9343145010887977, 'Standard-F1': 0.1304915180510117, 'Standard-Precision': 0.08112493239588962, 'Standard-Recall': 0.3333333333333326, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6748127029888804, 'Affiliation-P': 0.5108775552977954, 'Affiliation-R': 0.9936711581757911, 'F1_T': 0.130677290836359, 'Precision_T': 0.07961165048543686, 'Recall_T': 0.3644444444444436}
Logits metrics skipped for 153_Stock_id_5_Finance_tr_500_1st_1.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/153_Stock_id_5_Finance_tr_500_1st_1_results.pkl
Skipping file: 859_OPPORTUNITY_id_18_HumanActivity_tr_895_1st_995.csv due to filter criteria.
Processing file: 065_WSD_id_37_WebService_tr_3451_1st_3551.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (3451, 1), Test data shape: (14785, 1), Label test shape: (14785,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 065_WSD_id_37_WebService_tr_3451_1st_3551.csv...
Using CUDA device: 2
Data shape: (14785, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 14785, window_size 5000, adding 215 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 157.54it/s]
Original shapes - Output: (15000,), Label: (14785,)
Aligned shapes - Output: (14785,), Label: (14785,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 273.95it/s]
Table1-only metrics completed in 1.64s
Evaluation Result:  {'AUC-PR': 0.6270996456667458, 'AUC-ROC': 0.9279187656367675, 'VUS-PR': 0.8441336606381015, 'VUS-ROC': 0.9914502396559992, 'Standard-F1': 0.6911447084228329, 'Standard-Precision': 0.6060606060606037, 'Standard-Recall': 0.8040201005025085, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9957523181359074, 'Affiliation-P': 0.9935798208177902, 'Affiliation-R': 0.997934336756694, 'F1_T': 0.6815415821496171, 'Precision_T': 0.5714285714285694, 'Recall_T': 0.844221105527634}
Results for 065_WSD_id_37_WebService_tr_3451_1st_3551.csv: {'filename': '065_WSD_id_37_WebService_tr_3451_1st_3551.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '3451', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 14785, 'aligned_length': 14785, 'AUC-PR': 0.6270996456667458, 'AUC-ROC': 0.9279187656367675, 'VUS-PR': 0.8441336606381015, 'VUS-ROC': 0.9914502396559992, 'Standard-F1': 0.6911447084228329, 'Standard-Precision': 0.6060606060606037, 'Standard-Recall': 0.8040201005025085, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9957523181359074, 'Affiliation-P': 0.9935798208177902, 'Affiliation-R': 0.997934336756694, 'F1_T': 0.6815415821496171, 'Precision_T': 0.5714285714285694, 'Recall_T': 0.844221105527634}
Logits metrics skipped for 065_WSD_id_37_WebService_tr_3451_1st_3551.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/065_WSD_id_37_WebService_tr_3451_1st_3551_results.pkl
Processing file: 599_YAHOO_id_49_WebService_tr_500_1st_653.csv, Data shape: (1461, 1)
Label shape: (1461,)
Train data shape: (500, 1), Test data shape: (961, 1), Label test shape: (961,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 599_YAHOO_id_49_WebService_tr_500_1st_653.csv...
Using CUDA device: 2
Data shape: (961, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 961, 1])
batched_mask shape: torch.Size([1, 961])

100%|██████████| 1/1 [00:00<00:00, 162.02it/s]
Original shapes - Output: (961,), Label: (961,)
Aligned shapes - Output: (961,), Label: (961,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1640.34it/s]
Table1-only metrics completed in 1.42s
Evaluation Result:  {'AUC-PR': 0.8956845238095238, 'AUC-ROC': 0.9992130115424973, 'VUS-PR': 0.995012948481207, 'VUS-ROC': 0.9999685386067009, 'Standard-F1': 0.8888888888882962, 'Standard-Precision': 0.79999999999992, 'Standard-Recall': 0.999999999999875, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990904484089338, 'Affiliation-P': 0.9981825498825149, 'Affiliation-R': 1.0, 'F1_T': 0.8888888888882962, 'Precision_T': 0.79999999999992, 'Recall_T': 0.999999999999875}
Results for 599_YAHOO_id_49_WebService_tr_500_1st_653.csv: {'filename': '599_YAHOO_id_49_WebService_tr_500_1st_653.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 144, 'train_index': '500', 'data_shape': '1461x1', 'output_length': 961, 'label_length': 961, 'aligned_length': 961, 'AUC-PR': 0.8956845238095238, 'AUC-ROC': 0.9992130115424973, 'VUS-PR': 0.995012948481207, 'VUS-ROC': 0.9999685386067009, 'Standard-F1': 0.8888888888882962, 'Standard-Precision': 0.79999999999992, 'Standard-Recall': 0.999999999999875, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990904484089338, 'Affiliation-P': 0.9981825498825149, 'Affiliation-R': 1.0, 'F1_T': 0.8888888888882962, 'Precision_T': 0.79999999999992, 'Recall_T': 0.999999999999875}
Logits metrics skipped for 599_YAHOO_id_49_WebService_tr_500_1st_653.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/599_YAHOO_id_49_WebService_tr_500_1st_653_results.pkl
Processing file: 079_WSD_id_51_WebService_tr_4559_1st_10672.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 079_WSD_id_51_WebService_tr_4559_1st_10672.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.34it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 298.06it/s]
Table1-only metrics completed in 0.80s
Evaluation Result:  {'AUC-PR': 0.15006220166452103, 'AUC-ROC': 0.9833672500339167, 'VUS-PR': 0.12512162460186407, 'VUS-ROC': 0.987165611323333, 'Standard-F1': 0.20731707317045414, 'Standard-Precision': 0.12408759124087501, 'Standard-Recall': 0.6296296296296063, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9435979418512875, 'Affiliation-P': 0.8934360215201265, 'Affiliation-R': 0.9997275973417818, 'F1_T': 0.20731707317045414, 'Precision_T': 0.12408759124087501, 'Recall_T': 0.6296296296296063}
Results for 079_WSD_id_51_WebService_tr_4559_1st_10672.csv: {'filename': '079_WSD_id_51_WebService_tr_4559_1st_10672.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 9, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.15006220166452103, 'AUC-ROC': 0.9833672500339167, 'VUS-PR': 0.12512162460186407, 'VUS-ROC': 0.987165611323333, 'Standard-F1': 0.20731707317045414, 'Standard-Precision': 0.12408759124087501, 'Standard-Recall': 0.6296296296296063, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9435979418512875, 'Affiliation-P': 0.8934360215201265, 'Affiliation-R': 0.9997275973417818, 'F1_T': 0.20731707317045414, 'Precision_T': 0.12408759124087501, 'Recall_T': 0.6296296296296063}
Logits metrics skipped for 079_WSD_id_51_WebService_tr_4559_1st_10672.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/079_WSD_id_51_WebService_tr_4559_1st_10672_results.pkl
Processing file: 557_YAHOO_id_7_WebService_tr_500_1st_770.csv, Data shape: (1424, 1)
Label shape: (1424,)
Train data shape: (500, 1), Test data shape: (924, 1), Label test shape: (924,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 557_YAHOO_id_7_WebService_tr_500_1st_770.csv...
Using CUDA device: 2
Data shape: (924, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 924, 1])
batched_mask shape: torch.Size([1, 924])

100%|██████████| 1/1 [00:00<00:00, 161.64it/s]
Original shapes - Output: (924,), Label: (924,)
Aligned shapes - Output: (924,), Label: (924,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2260.32it/s]
Table1-only metrics completed in 0.28s
Evaluation Result:  {'AUC-PR': 0.9465453057471778, 'AUC-ROC': 0.9849385977663871, 'VUS-PR': 0.9463159467409944, 'VUS-ROC': 0.9859167837336505, 'Standard-F1': 0.9246435845208838, 'Standard-Precision': 0.8598484848484815, 'Standard-Recall': 0.9999999999999957, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9892896147639515, 'Affiliation-P': 0.9809783481237985, 'Affiliation-R': 0.9977429179847174, 'F1_T': 0.9246435845208838, 'Precision_T': 0.8598484848484815, 'Recall_T': 0.9999999999999957}
Results for 557_YAHOO_id_7_WebService_tr_500_1st_770.csv: {'filename': '557_YAHOO_id_7_WebService_tr_500_1st_770.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 23, 'train_index': '500', 'data_shape': '1424x1', 'output_length': 924, 'label_length': 924, 'aligned_length': 924, 'AUC-PR': 0.9465453057471778, 'AUC-ROC': 0.9849385977663871, 'VUS-PR': 0.9463159467409944, 'VUS-ROC': 0.9859167837336505, 'Standard-F1': 0.9246435845208838, 'Standard-Precision': 0.8598484848484815, 'Standard-Recall': 0.9999999999999957, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9892896147639515, 'Affiliation-P': 0.9809783481237985, 'Affiliation-R': 0.9977429179847174, 'F1_T': 0.9246435845208838, 'Precision_T': 0.8598484848484815, 'Recall_T': 0.9999999999999957}
Logits metrics skipped for 557_YAHOO_id_7_WebService_tr_500_1st_770.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/557_YAHOO_id_7_WebService_tr_500_1st_770_results.pkl
Processing file: 122_WSD_id_94_WebService_tr_4559_1st_8618.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 122_WSD_id_94_WebService_tr_4559_1st_8618.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.65it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 282.46it/s]
Table1-only metrics completed in 1.15s
Evaluation Result:  {'AUC-PR': 0.0398356304249305, 'AUC-ROC': 0.45452531280480774, 'VUS-PR': 0.018198588715688566, 'VUS-ROC': 0.6584406468620837, 'Standard-F1': 0.02429149797535767, 'Standard-Precision': 0.015706806282722433, 'Standard-Recall': 0.05357142857142761, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7927811928305804, 'Affiliation-P': 0.6578395786327906, 'Affiliation-R': 0.9973701970769464, 'F1_T': 0.022988505746789217, 'Precision_T': 0.014634146341463344, 'Recall_T': 0.05357142857142761}
Results for 122_WSD_id_94_WebService_tr_4559_1st_8618.csv: {'filename': '122_WSD_id_94_WebService_tr_4559_1st_8618.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 60, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.0398356304249305, 'AUC-ROC': 0.45452531280480774, 'VUS-PR': 0.018198588715688566, 'VUS-ROC': 0.6584406468620837, 'Standard-F1': 0.02429149797535767, 'Standard-Precision': 0.015706806282722433, 'Standard-Recall': 0.05357142857142761, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7927811928305804, 'Affiliation-P': 0.6578395786327906, 'Affiliation-R': 0.9973701970769464, 'F1_T': 0.022988505746789217, 'Precision_T': 0.014634146341463344, 'Recall_T': 0.05357142857142761}
Logits metrics skipped for 122_WSD_id_94_WebService_tr_4559_1st_8618.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/122_WSD_id_94_WebService_tr_4559_1st_8618_results.pkl
Processing file: 372_UCR_id_70_Medical_tr_7500_1st_17000.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 372_UCR_id_70_Medical_tr_7500_1st_17000.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 153.02it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 208.54it/s]
Table1-only metrics completed in 1.87s
Evaluation Result:  {'AUC-PR': 0.0199303701638569, 'AUC-ROC': 0.8609285714285715, 'VUS-PR': 0.018800643712042844, 'VUS-ROC': 0.8683531807785843, 'Standard-F1': 0.04576475311112322, 'Standard-Precision': 0.02383939774153073, 'Standard-Recall': 0.5699999999999943, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7023047275388644, 'Affiliation-P': 0.5411938707358575, 'Affiliation-R': 1.0, 'F1_T': 0.04632645674983973, 'Precision_T': 0.024033045437476523, 'Recall_T': 0.6399999999999937}
Results for 372_UCR_id_70_Medical_tr_7500_1st_17000.csv: {'filename': '372_UCR_id_70_Medical_tr_7500_1st_17000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 166, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.0199303701638569, 'AUC-ROC': 0.8609285714285715, 'VUS-PR': 0.018800643712042844, 'VUS-ROC': 0.8683531807785843, 'Standard-F1': 0.04576475311112322, 'Standard-Precision': 0.02383939774153073, 'Standard-Recall': 0.5699999999999943, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7023047275388644, 'Affiliation-P': 0.5411938707358575, 'Affiliation-R': 1.0, 'F1_T': 0.04632645674983973, 'Precision_T': 0.024033045437476523, 'Recall_T': 0.6399999999999937}
Logits metrics skipped for 372_UCR_id_70_Medical_tr_7500_1st_17000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/372_UCR_id_70_Medical_tr_7500_1st_17000_results.pkl
Processing file: 484_UCR_id_182_Medical_tr_1875_1st_5759.csv, Data shape: (7500, 1)
Label shape: (7500,)
Train data shape: (1875, 1), Test data shape: (5625, 1), Label test shape: (5625,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 484_UCR_id_182_Medical_tr_1875_1st_5759.csv...
Using CUDA device: 2
Data shape: (5625, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5625, window_size 5000, adding 4375 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.08it/s]
Original shapes - Output: (10000,), Label: (5625,)
Aligned shapes - Output: (5625,), Label: (5625,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 663.45it/s]
Table1-only metrics completed in 1.00s
Evaluation Result:  {'AUC-PR': 0.821215983855942, 'AUC-ROC': 0.9790645013723697, 'VUS-PR': 0.9313748348622737, 'VUS-ROC': 0.9944257798807601, 'Standard-F1': 0.7867647058818628, 'Standard-Precision': 0.9553571428571344, 'Standard-Recall': 0.6687499999999958, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9980905275303844, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9961883333333332, 'F1_T': 0.7867647058818628, 'Precision_T': 0.9553571428571344, 'Recall_T': 0.6687499999999958}
Results for 484_UCR_id_182_Medical_tr_1875_1st_5759.csv: {'filename': '484_UCR_id_182_Medical_tr_1875_1st_5759.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 183, 'train_index': '1875', 'data_shape': '7500x1', 'output_length': 10000, 'label_length': 5625, 'aligned_length': 5625, 'AUC-PR': 0.821215983855942, 'AUC-ROC': 0.9790645013723697, 'VUS-PR': 0.9313748348622737, 'VUS-ROC': 0.9944257798807601, 'Standard-F1': 0.7867647058818628, 'Standard-Precision': 0.9553571428571344, 'Standard-Recall': 0.6687499999999958, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9980905275303844, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9961883333333332, 'F1_T': 0.7867647058818628, 'Precision_T': 0.9553571428571344, 'Recall_T': 0.6687499999999958}
Logits metrics skipped for 484_UCR_id_182_Medical_tr_1875_1st_5759.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/484_UCR_id_182_Medical_tr_1875_1st_5759_results.pkl
Processing file: 643_YAHOO_id_93_WebService_tr_500_1st_1038.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 643_YAHOO_id_93_WebService_tr_500_1st_1038.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 154.67it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2115.18it/s]
Table1-only metrics completed in 0.19s
Evaluation Result:  {'AUC-PR': 0.004484304932735426, 'AUC-ROC': 0.811704834605598, 'VUS-PR': 0.004484304932735426, 'VUS-ROC': 0.8134011874469887, 'Standard-F1': 0.008928571428562458, 'Standard-Precision': 0.004484304932735406, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6952357074437646, 'Affiliation-P': 0.5328439101300749, 'Affiliation-R': 1.0, 'F1_T': 0.008771929824552592, 'Precision_T': 0.004405286343612315, 'Recall_T': 0.9999999999989999}
Results for 643_YAHOO_id_93_WebService_tr_500_1st_1038.csv: {'filename': '643_YAHOO_id_93_WebService_tr_500_1st_1038.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.004484304932735426, 'AUC-ROC': 0.811704834605598, 'VUS-PR': 0.004484304932735426, 'VUS-ROC': 0.8134011874469887, 'Standard-F1': 0.008928571428562458, 'Standard-Precision': 0.004484304932735406, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6952357074437646, 'Affiliation-P': 0.5328439101300749, 'Affiliation-R': 1.0, 'F1_T': 0.008771929824552592, 'Precision_T': 0.004405286343612315, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 643_YAHOO_id_93_WebService_tr_500_1st_1038.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/643_YAHOO_id_93_WebService_tr_500_1st_1038_results.pkl
Skipping file: 251_SVDB_id_15_Medical_tr_5421_1st_5521.csv due to filter criteria.
Processing file: 747_YAHOO_id_197_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 747_YAHOO_id_197_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 163.26it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2389.91it/s]
Table1-only metrics completed in 0.33s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8576744186046505, 'Affiliation-P': 0.7508143322475569, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 747_YAHOO_id_197_Synthetic_tr_500_1st_893.csv: {'filename': '747_YAHOO_id_197_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 70, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8576744186046505, 'Affiliation-P': 0.7508143322475569, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 747_YAHOO_id_197_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/747_YAHOO_id_197_Synthetic_tr_500_1st_893_results.pkl
Processing file: 806_YAHOO_id_256_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 806_YAHOO_id_256_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 164.79it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2337.44it/s]
Table1-only metrics completed in 0.48s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9982597346095274, 'Affiliation-P': 0.9965255157437568, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 806_YAHOO_id_256_Synthetic_tr_500_1st_893.csv: {'filename': '806_YAHOO_id_256_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 118, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9982597346095274, 'Affiliation-P': 0.9965255157437568, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 806_YAHOO_id_256_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/806_YAHOO_id_256_Synthetic_tr_500_1st_893_results.pkl
Skipping file: 217_LTDB_id_2_Medical_tr_500_1st_266.csv due to filter criteria.
Processing file: 743_YAHOO_id_193_WebService_tr_500_1st_133.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 743_YAHOO_id_193_WebService_tr_500_1st_133.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 152.86it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1251.34it/s]
Table1-only metrics completed in 0.69s
Evaluation Result:  {'AUC-PR': 0.4898249405425028, 'AUC-ROC': 0.8274029794098112, 'VUS-PR': 0.3438598492364695, 'VUS-ROC': 0.8235310312935944, 'Standard-F1': 0.41666666666616325, 'Standard-Precision': 0.3333333333333111, 'Standard-Recall': 0.5555555555554939, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7912352212657102, 'Affiliation-P': 0.6545816317499106, 'Affiliation-R': 1.0, 'F1_T': 0.3809523809518549, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.44444444444439507}
Results for 743_YAHOO_id_193_WebService_tr_500_1st_133.csv: {'filename': '743_YAHOO_id_193_WebService_tr_500_1st_133.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.4898249405425028, 'AUC-ROC': 0.8274029794098112, 'VUS-PR': 0.3438598492364695, 'VUS-ROC': 0.8235310312935944, 'Standard-F1': 0.41666666666616325, 'Standard-Precision': 0.3333333333333111, 'Standard-Recall': 0.5555555555554939, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7912352212657102, 'Affiliation-P': 0.6545816317499106, 'Affiliation-R': 1.0, 'F1_T': 0.3809523809518549, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.44444444444439507}
Logits metrics skipped for 743_YAHOO_id_193_WebService_tr_500_1st_133.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/743_YAHOO_id_193_WebService_tr_500_1st_133_results.pkl
Processing file: 797_YAHOO_id_247_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 797_YAHOO_id_247_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 162.93it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2042.93it/s]
Table1-only metrics completed in 0.63s
Evaluation Result:  {'AUC-PR': 0.021739130434782608, 'AUC-ROC': 0.9510869565217391, 'VUS-PR': 0.19700939938526488, 'VUS-ROC': 0.9682318037436857, 'Standard-F1': 0.04255319148931825, 'Standard-Precision': 0.021739130434782136, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9263604345453639, 'Affiliation-P': 0.8628225564257167, 'Affiliation-R': 1.0, 'F1_T': 0.04255319148931825, 'Precision_T': 0.021739130434782136, 'Recall_T': 0.9999999999989999}
Results for 797_YAHOO_id_247_Synthetic_tr_500_1st_893.csv: {'filename': '797_YAHOO_id_247_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 158, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.021739130434782608, 'AUC-ROC': 0.9510869565217391, 'VUS-PR': 0.19700939938526488, 'VUS-ROC': 0.9682318037436857, 'Standard-F1': 0.04255319148931825, 'Standard-Precision': 0.021739130434782136, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9263604345453639, 'Affiliation-P': 0.8628225564257167, 'Affiliation-R': 1.0, 'F1_T': 0.04255319148931825, 'Precision_T': 0.021739130434782136, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 797_YAHOO_id_247_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/797_YAHOO_id_247_Synthetic_tr_500_1st_893_results.pkl
Skipping file: 823_Exathlon_id_14_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 396_UCR_id_94_Medical_tr_1875_1st_5607.csv, Data shape: (7501, 1)
Label shape: (7501,)
Train data shape: (1875, 1), Test data shape: (5626, 1), Label test shape: (5626,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 396_UCR_id_94_Medical_tr_1875_1st_5607.csv...
Using CUDA device: 2
Data shape: (5626, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5626, window_size 5000, adding 4374 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 153.01it/s]
Original shapes - Output: (10000,), Label: (5626,)
Aligned shapes - Output: (5626,), Label: (5626,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 681.96it/s]
Table1-only metrics completed in 0.89s
Evaluation Result:  {'AUC-PR': 0.536838046748622, 'AUC-ROC': 0.9958524339663829, 'VUS-PR': 0.39268187337991534, 'VUS-ROC': 0.9957134213731713, 'Standard-F1': 0.5476190476185983, 'Standard-Precision': 0.4035087719298175, 'Standard-Recall': 0.8518518518518203, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.749176001908703, 'Affiliation-P': 0.5989459772533258, 'Affiliation-R': 1.0, 'F1_T': 0.5476190476185983, 'Precision_T': 0.4035087719298175, 'Recall_T': 0.8518518518518203}
Results for 396_UCR_id_94_Medical_tr_1875_1st_5607.csv: {'filename': '396_UCR_id_94_Medical_tr_1875_1st_5607.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 154, 'train_index': '1875', 'data_shape': '7501x1', 'output_length': 10000, 'label_length': 5626, 'aligned_length': 5626, 'AUC-PR': 0.536838046748622, 'AUC-ROC': 0.9958524339663829, 'VUS-PR': 0.39268187337991534, 'VUS-ROC': 0.9957134213731713, 'Standard-F1': 0.5476190476185983, 'Standard-Precision': 0.4035087719298175, 'Standard-Recall': 0.8518518518518203, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.749176001908703, 'Affiliation-P': 0.5989459772533258, 'Affiliation-R': 1.0, 'F1_T': 0.5476190476185983, 'Precision_T': 0.4035087719298175, 'Recall_T': 0.8518518518518203}
Logits metrics skipped for 396_UCR_id_94_Medical_tr_1875_1st_5607.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/396_UCR_id_94_Medical_tr_1875_1st_5607_results.pkl
Processing file: 288_TODS_id_2_Synthetic_tr_500_1st_65.csv, Data shape: (5000, 1)
Label shape: (5000,)
Train data shape: (500, 1), Test data shape: (4500, 1), Label test shape: (4500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 288_TODS_id_2_Synthetic_tr_500_1st_65.csv...
Using CUDA device: 2
Data shape: (4500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 4500, 1])
batched_mask shape: torch.Size([1, 4500])

100%|██████████| 1/1 [00:00<00:00, 161.36it/s]
Original shapes - Output: (4500,), Label: (4500,)
Aligned shapes - Output: (4500,), Label: (4500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  22%|██▏       | 43/200 [00:01<00:03, 41.56it/s]

Calculating Affiliation Metrics:  42%|████▎     | 85/200 [00:03<00:04, 26.60it/s]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [00:04<00:03, 22.91it/s]

Calculating Affiliation Metrics:  70%|███████   | 140/200 [00:05<00:02, 22.01it/s]

Calculating Affiliation Metrics:  92%|█████████▎| 185/200 [00:06<00:00, 28.41it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:06<00:00, 28.81it/s]
Table1-only metrics completed in 13.61s
Evaluation Result:  {'AUC-PR': 0.8599109320203452, 'AUC-ROC': 0.9425765896242952, 'VUS-PR': 0.9181471915352272, 'VUS-ROC': 0.9737385346354482, 'Standard-F1': 0.8319185059417767, 'Standard-Precision': 0.9210526315789438, 'Standard-Recall': 0.7585139318885424, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7946154253876733, 'Affiliation-P': 0.821609975647109, 'Affiliation-R': 0.7693382964274729, 'F1_T': 0.8305084745757728, 'Precision_T': 0.9176029962546781, 'Recall_T': 0.7585139318885424}
Results for 288_TODS_id_2_Synthetic_tr_500_1st_65.csv: {'filename': '288_TODS_id_2_Synthetic_tr_500_1st_65.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '500', 'data_shape': '5000x1', 'output_length': 4500, 'label_length': 4500, 'aligned_length': 4500, 'AUC-PR': 0.8599109320203452, 'AUC-ROC': 0.9425765896242952, 'VUS-PR': 0.9181471915352272, 'VUS-ROC': 0.9737385346354482, 'Standard-F1': 0.8319185059417767, 'Standard-Precision': 0.9210526315789438, 'Standard-Recall': 0.7585139318885424, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7946154253876733, 'Affiliation-P': 0.821609975647109, 'Affiliation-R': 0.7693382964274729, 'F1_T': 0.8305084745757728, 'Precision_T': 0.9176029962546781, 'Recall_T': 0.7585139318885424}
Logits metrics skipped for 288_TODS_id_2_Synthetic_tr_500_1st_65.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/288_TODS_id_2_Synthetic_tr_500_1st_65_results.pkl
Processing file: 384_UCR_id_82_Facility_tr_50000_1st_884100.csv, Data shape: (900000, 1)
Label shape: (900000,)
Train data shape: (50000, 1), Test data shape: (850000, 1), Label test shape: (850000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 384_UCR_id_82_Facility_tr_50000_1st_884100.csv...
Using CUDA device: 2
Data shape: (850000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/3 [00:00<?, ?it/s]batched_time_series shape: torch.Size([64, 5000, 1])
batched_mask shape: torch.Size([64, 5000])
batched_time_series shape: torch.Size([64, 5000, 1])
batched_mask shape: torch.Size([64, 5000])
batched_time_series shape: torch.Size([42, 5000, 1])
batched_mask shape: torch.Size([42, 5000])

100%|██████████| 3/3 [00:00<00:00, 25.59it/s]
Original shapes - Output: (850000,), Label: (850000,)
Aligned shapes - Output: (850000,), Label: (850000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   4%|▎         | 7/200 [00:01<00:31,  6.14it/s]

Calculating Affiliation Metrics:   7%|▋         | 14/200 [00:02<00:32,  5.81it/s]

Calculating Affiliation Metrics:  10%|█         | 20/200 [00:03<00:32,  5.56it/s]

Calculating Affiliation Metrics:  13%|█▎        | 26/200 [00:04<00:32,  5.34it/s]

Calculating Affiliation Metrics:  16%|█▌        | 32/200 [00:06<00:33,  5.01it/s]

Calculating Affiliation Metrics:  19%|█▉        | 38/200 [00:07<00:32,  4.93it/s]

Calculating Affiliation Metrics:  22%|██▏       | 43/200 [00:08<00:32,  4.89it/s]

Calculating Affiliation Metrics:  24%|██▍       | 48/200 [00:09<00:31,  4.85it/s]

Calculating Affiliation Metrics:  26%|██▋       | 53/200 [00:10<00:30,  4.80it/s]

Calculating Affiliation Metrics:  29%|██▉       | 58/200 [00:11<00:30,  4.71it/s]

Calculating Affiliation Metrics:  32%|███▏      | 63/200 [00:12<00:28,  4.73it/s]

Calculating Affiliation Metrics:  34%|███▍      | 68/200 [00:13<00:28,  4.67it/s]

Calculating Affiliation Metrics:  36%|███▋      | 73/200 [00:14<00:26,  4.74it/s]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [00:15<00:25,  4.80it/s]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [00:16<00:24,  4.82it/s]

Calculating Affiliation Metrics:  44%|████▍     | 88/200 [00:17<00:23,  4.71it/s]

Calculating Affiliation Metrics:  46%|████▋     | 93/200 [00:18<00:22,  4.75it/s]

Calculating Affiliation Metrics:  49%|████▉     | 98/200 [00:19<00:21,  4.82it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 103/200 [00:21<00:20,  4.80it/s]

Calculating Affiliation Metrics:  55%|█████▍    | 109/200 [00:22<00:18,  4.89it/s]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [00:23<00:16,  5.00it/s]

Calculating Affiliation Metrics:  60%|██████    | 121/200 [00:24<00:15,  5.08it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [00:25<00:13,  5.28it/s]

Calculating Affiliation Metrics:  67%|██████▋   | 134/200 [00:26<00:11,  5.53it/s]

Calculating Affiliation Metrics:  70%|███████   | 141/200 [00:27<00:10,  5.81it/s]

Calculating Affiliation Metrics:  74%|███████▍  | 148/200 [00:28<00:08,  5.98it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 156/200 [00:29<00:06,  6.35it/s]

Calculating Affiliation Metrics:  82%|████████▏ | 164/200 [00:30<00:05,  6.75it/s]

Calculating Affiliation Metrics:  86%|████████▋ | 173/200 [00:32<00:03,  7.27it/s]

Calculating Affiliation Metrics:  92%|█████████▏| 183/200 [00:33<00:02,  7.94it/s]

Calculating Affiliation Metrics:  98%|█████████▊| 195/200 [00:34<00:00,  8.90it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:34<00:00,  5.79it/s]
Table1-only metrics completed in 61.96s
Evaluation Result:  {'AUC-PR': 0.0005580316088002027, 'AUC-ROC': 0.9029614895870102, 'VUS-PR': 0.0007273009741812736, 'VUS-ROC': 0.9118743723832635, 'Standard-F1': 0.0017453662423384132, 'Standard-Precision': 0.0008735117069558117, 'Standard-Recall': 0.9199999999999908, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6748736326769119, 'Affiliation-P': 0.5108381275034429, 'Affiliation-R': 0.9940847058823529, 'F1_T': 0.001705436307453492, 'Precision_T': 0.0008535007296054624, 'Recall_T': 0.9299999999999907}
Results for 384_UCR_id_82_Facility_tr_50000_1st_884100.csv: {'filename': '384_UCR_id_82_Facility_tr_50000_1st_884100.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '50000', 'data_shape': '900000x1', 'output_length': 850000, 'label_length': 850000, 'aligned_length': 850000, 'AUC-PR': 0.0005580316088002027, 'AUC-ROC': 0.9029614895870102, 'VUS-PR': 0.0007273009741812736, 'VUS-ROC': 0.9118743723832635, 'Standard-F1': 0.0017453662423384132, 'Standard-Precision': 0.0008735117069558117, 'Standard-Recall': 0.9199999999999908, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6748736326769119, 'Affiliation-P': 0.5108381275034429, 'Affiliation-R': 0.9940847058823529, 'F1_T': 0.001705436307453492, 'Precision_T': 0.0008535007296054624, 'Recall_T': 0.9299999999999907}
Logits metrics skipped for 384_UCR_id_82_Facility_tr_50000_1st_884100.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/384_UCR_id_82_Facility_tr_50000_1st_884100_results.pkl
Processing file: 526_UCR_id_224_Sensor_tr_2827_1st_4711.csv, Data shape: (11308, 1)
Label shape: (11308,)
Train data shape: (2827, 1), Test data shape: (8481, 1), Label test shape: (8481,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 526_UCR_id_224_Sensor_tr_2827_1st_4711.csv...
Using CUDA device: 2
Data shape: (8481, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 8481, window_size 5000, adding 1519 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 147.51it/s]
Original shapes - Output: (10000,), Label: (8481,)
Aligned shapes - Output: (8481,), Label: (8481,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 452.00it/s]
Table1-only metrics completed in 0.85s
Evaluation Result:  {'AUC-PR': 0.992059993814742, 'AUC-ROC': 0.9998466283805661, 'VUS-PR': 0.9861573387116974, 'VUS-ROC': 0.9999029100858927, 'Standard-F1': 0.9599999999994906, 'Standard-Precision': 0.9411764705882261, 'Standard-Recall': 0.979591836734684, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999605946418935, 'Affiliation-P': 1.0, 'Affiliation-R': 0.99992119238923, 'F1_T': 0.9289617486333722, 'Precision_T': 0.9999999999999883, 'Recall_T': 0.8673469387755014}
Results for 526_UCR_id_224_Sensor_tr_2827_1st_4711.csv: {'filename': '526_UCR_id_224_Sensor_tr_2827_1st_4711.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 98, 'train_index': '2827', 'data_shape': '11308x1', 'output_length': 10000, 'label_length': 8481, 'aligned_length': 8481, 'AUC-PR': 0.992059993814742, 'AUC-ROC': 0.9998466283805661, 'VUS-PR': 0.9861573387116974, 'VUS-ROC': 0.9999029100858927, 'Standard-F1': 0.9599999999994906, 'Standard-Precision': 0.9411764705882261, 'Standard-Recall': 0.979591836734684, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999605946418935, 'Affiliation-P': 1.0, 'Affiliation-R': 0.99992119238923, 'F1_T': 0.9289617486333722, 'Precision_T': 0.9999999999999883, 'Recall_T': 0.8673469387755014}
Logits metrics skipped for 526_UCR_id_224_Sensor_tr_2827_1st_4711.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/526_UCR_id_224_Sensor_tr_2827_1st_4711_results.pkl
Processing file: 613_YAHOO_id_63_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 613_YAHOO_id_63_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 160.08it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1826.06it/s]
Table1-only metrics completed in 0.35s
Evaluation Result:  {'AUC-PR': 0.9459876543209877, 'AUC-ROC': 0.9993908382066277, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.998374868670618, 'Affiliation-P': 0.9968407330726171, 'Affiliation-R': 0.9999137336093858, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 613_YAHOO_id_63_Synthetic_tr_500_1st_623.csv: {'filename': '613_YAHOO_id_63_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 35, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9459876543209877, 'AUC-ROC': 0.9993908382066277, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.998374868670618, 'Affiliation-P': 0.9968407330726171, 'Affiliation-R': 0.9999137336093858, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 613_YAHOO_id_63_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/613_YAHOO_id_63_Synthetic_tr_500_1st_623_results.pkl
Skipping file: 852_OPPORTUNITY_id_11_HumanActivity_tr_500_1st_566.csv due to filter criteria.
Processing file: 298_TODS_id_12_Synthetic_tr_500_1st_0.csv, Data shape: (5000, 1)
Label shape: (5000,)
Train data shape: (500, 1), Test data shape: (4500, 1), Label test shape: (4500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 298_TODS_id_12_Synthetic_tr_500_1st_0.csv...
Using CUDA device: 2
Data shape: (4500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 4500, 1])
batched_mask shape: torch.Size([1, 4500])

100%|██████████| 1/1 [00:00<00:00, 162.55it/s]
Original shapes - Output: (4500,), Label: (4500,)
Aligned shapes - Output: (4500,), Label: (4500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  32%|███▎      | 65/200 [00:01<00:02, 64.74it/s]

Calculating Affiliation Metrics:  65%|██████▌   | 130/200 [00:02<00:01, 57.98it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 72.22it/s]
Table1-only metrics completed in 8.46s
Evaluation Result:  {'AUC-PR': 0.870923229609403, 'AUC-ROC': 0.9777351595100822, 'VUS-PR': 0.9731304760018066, 'VUS-ROC': 0.9976519339223047, 'Standard-F1': 0.8072072072067067, 'Standard-Precision': 0.8682170542635624, 'Standard-Recall': 0.7542087542087517, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9543797931025695, 'Affiliation-P': 0.9478919079164384, 'Affiliation-R': 0.9609571035581336, 'F1_T': 0.8049645390065906, 'Precision_T': 0.8501872659175997, 'Recall_T': 0.7643097643097617}
Results for 298_TODS_id_12_Synthetic_tr_500_1st_0.csv: {'filename': '298_TODS_id_12_Synthetic_tr_500_1st_0.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '500', 'data_shape': '5000x1', 'output_length': 4500, 'label_length': 4500, 'aligned_length': 4500, 'AUC-PR': 0.870923229609403, 'AUC-ROC': 0.9777351595100822, 'VUS-PR': 0.9731304760018066, 'VUS-ROC': 0.9976519339223047, 'Standard-F1': 0.8072072072067067, 'Standard-Precision': 0.8682170542635624, 'Standard-Recall': 0.7542087542087517, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9543797931025695, 'Affiliation-P': 0.9478919079164384, 'Affiliation-R': 0.9609571035581336, 'F1_T': 0.8049645390065906, 'Precision_T': 0.8501872659175997, 'Recall_T': 0.7643097643097617}
Logits metrics skipped for 298_TODS_id_12_Synthetic_tr_500_1st_0.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/298_TODS_id_12_Synthetic_tr_500_1st_0_results.pkl
Processing file: 091_WSD_id_63_WebService_tr_500_1st_363.csv, Data shape: (18230, 1)
Label shape: (18230,)
Train data shape: (500, 1), Test data shape: (17730, 1), Label test shape: (17730,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 091_WSD_id_63_WebService_tr_500_1st_363.csv...
Using CUDA device: 2
Data shape: (17730, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 17730, window_size 5000, adding 2270 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 149.45it/s]
Original shapes - Output: (20000,), Label: (17730,)
Aligned shapes - Output: (17730,), Label: (17730,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 175/200 [00:01<00:00, 174.27it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 184.34it/s]
Table1-only metrics completed in 1.39s
Evaluation Result:  {'AUC-PR': 0.16614279315247177, 'AUC-ROC': 0.8158908681175773, 'VUS-PR': 0.08774388976973212, 'VUS-ROC': 0.8631905064579765, 'Standard-F1': 0.12742382271419694, 'Standard-Precision': 0.10798122065727649, 'Standard-Recall': 0.15540540540540437, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9814149819320188, 'Affiliation-P': 0.9678119653867733, 'Affiliation-R': 0.9954058423753026, 'F1_T': 0.12444444444410172, 'Precision_T': 0.07969639468690687, 'Recall_T': 0.2837837837837819}
Results for 091_WSD_id_63_WebService_tr_500_1st_363.csv: {'filename': '091_WSD_id_63_WebService_tr_500_1st_363.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 10, 'train_index': '500', 'data_shape': '18230x1', 'output_length': 20000, 'label_length': 17730, 'aligned_length': 17730, 'AUC-PR': 0.16614279315247177, 'AUC-ROC': 0.8158908681175773, 'VUS-PR': 0.08774388976973212, 'VUS-ROC': 0.8631905064579765, 'Standard-F1': 0.12742382271419694, 'Standard-Precision': 0.10798122065727649, 'Standard-Recall': 0.15540540540540437, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9814149819320188, 'Affiliation-P': 0.9678119653867733, 'Affiliation-R': 0.9954058423753026, 'F1_T': 0.12444444444410172, 'Precision_T': 0.07969639468690687, 'Recall_T': 0.2837837837837819}
Logits metrics skipped for 091_WSD_id_63_WebService_tr_500_1st_363.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/091_WSD_id_63_WebService_tr_500_1st_363_results.pkl
Processing file: 109_WSD_id_81_WebService_tr_3175_1st_3275.csv, Data shape: (18058, 1)
Label shape: (18058,)
Train data shape: (3175, 1), Test data shape: (14883, 1), Label test shape: (14883,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 109_WSD_id_81_WebService_tr_3175_1st_3275.csv...
Using CUDA device: 2
Data shape: (14883, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 14883, window_size 5000, adding 117 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.81it/s]
Original shapes - Output: (15000,), Label: (14883,)
Aligned shapes - Output: (14883,), Label: (14883,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [00:01<00:00, 144.62it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 172.33it/s]
Table1-only metrics completed in 1.58s
Evaluation Result:  {'AUC-PR': 0.680749003428847, 'AUC-ROC': 0.9668324387179834, 'VUS-PR': 0.5267735488844782, 'VUS-ROC': 0.9800199659961613, 'Standard-F1': 0.44859813084069433, 'Standard-Precision': 0.3221476510067093, 'Standard-Recall': 0.7384615384615272, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8657601874448838, 'Affiliation-P': 0.7645648149556945, 'Affiliation-R': 0.9978297689574808, 'F1_T': 0.44859813084069433, 'Precision_T': 0.3221476510067093, 'Recall_T': 0.7384615384615272}
Results for 109_WSD_id_81_WebService_tr_3175_1st_3275.csv: {'filename': '109_WSD_id_81_WebService_tr_3175_1st_3275.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 23, 'train_index': '3175', 'data_shape': '18058x1', 'output_length': 15000, 'label_length': 14883, 'aligned_length': 14883, 'AUC-PR': 0.680749003428847, 'AUC-ROC': 0.9668324387179834, 'VUS-PR': 0.5267735488844782, 'VUS-ROC': 0.9800199659961613, 'Standard-F1': 0.44859813084069433, 'Standard-Precision': 0.3221476510067093, 'Standard-Recall': 0.7384615384615272, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8657601874448838, 'Affiliation-P': 0.7645648149556945, 'Affiliation-R': 0.9978297689574808, 'F1_T': 0.44859813084069433, 'Precision_T': 0.3221476510067093, 'Recall_T': 0.7384615384615272}
Logits metrics skipped for 109_WSD_id_81_WebService_tr_3175_1st_3275.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/109_WSD_id_81_WebService_tr_3175_1st_3275_results.pkl
Processing file: 744_YAHOO_id_194_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 744_YAHOO_id_194_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 160.54it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2197.04it/s]
Table1-only metrics completed in 0.82s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 744_YAHOO_id_194_Synthetic_tr_500_1st_658.csv: {'filename': '744_YAHOO_id_194_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 132, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 744_YAHOO_id_194_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/744_YAHOO_id_194_Synthetic_tr_500_1st_658_results.pkl
Processing file: 459_UCR_id_157_HumanActivity_tr_50000_1st_110260.csv, Data shape: (200000, 1)
Label shape: (200000,)
Train data shape: (50000, 1), Test data shape: (150000, 1), Label test shape: (150000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 459_UCR_id_157_HumanActivity_tr_50000_1st_110260.csv...
Using CUDA device: 2
Data shape: (150000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([30, 5000, 1])
batched_mask shape: torch.Size([30, 5000])

100%|██████████| 1/1 [00:00<00:00, 46.88it/s]
Original shapes - Output: (150000,), Label: (150000,)
Aligned shapes - Output: (150000,), Label: (150000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  13%|█▎        | 26/200 [00:01<00:06, 25.15it/s]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:02<00:05, 24.71it/s]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [00:03<00:04, 25.33it/s]

Calculating Affiliation Metrics:  55%|█████▍    | 109/200 [00:04<00:03, 27.10it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [00:05<00:01, 30.10it/s]

Calculating Affiliation Metrics:  96%|█████████▋| 193/200 [00:06<00:00, 35.92it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:06<00:00, 31.89it/s]
Table1-only metrics completed in 11.81s
Evaluation Result:  {'AUC-PR': 0.02623655263131347, 'AUC-ROC': 0.9808028934232302, 'VUS-PR': 0.03164978562602504, 'VUS-ROC': 0.9809578549500813, 'Standard-F1': 0.07135287884129712, 'Standard-Precision': 0.03770063456513623, 'Standard-Recall': 0.664473684210522, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7098604041761226, 'Affiliation-P': 0.55021984169303, 'Affiliation-R': 1.0, 'F1_T': 0.0702875399360098, 'Precision_T': 0.03693754197447951, 'Recall_T': 0.723684210526311}
Results for 459_UCR_id_157_HumanActivity_tr_50000_1st_110260.csv: {'filename': '459_UCR_id_157_HumanActivity_tr_50000_1st_110260.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '50000', 'data_shape': '200000x1', 'output_length': 150000, 'label_length': 150000, 'aligned_length': 150000, 'AUC-PR': 0.02623655263131347, 'AUC-ROC': 0.9808028934232302, 'VUS-PR': 0.03164978562602504, 'VUS-ROC': 0.9809578549500813, 'Standard-F1': 0.07135287884129712, 'Standard-Precision': 0.03770063456513623, 'Standard-Recall': 0.664473684210522, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7098604041761226, 'Affiliation-P': 0.55021984169303, 'Affiliation-R': 1.0, 'F1_T': 0.0702875399360098, 'Precision_T': 0.03693754197447951, 'Recall_T': 0.723684210526311}
Logits metrics skipped for 459_UCR_id_157_HumanActivity_tr_50000_1st_110260.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/459_UCR_id_157_HumanActivity_tr_50000_1st_110260_results.pkl
Processing file: 429_UCR_id_127_Medical_tr_14825_1st_29000.csv, Data shape: (59302, 1)
Label shape: (59302,)
Train data shape: (14825, 1), Test data shape: (44477, 1), Label test shape: (44477,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 429_UCR_id_127_Medical_tr_14825_1st_29000.csv...
Using CUDA device: 2
Data shape: (44477, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 44477, window_size 5000, adding 523 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([9, 5000, 1])
batched_mask shape: torch.Size([9, 5000])

100%|██████████| 1/1 [00:00<00:00, 109.90it/s]
Original shapes - Output: (45000,), Label: (44477,)
Aligned shapes - Output: (44477,), Label: (44477,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  43%|████▎     | 86/200 [00:01<00:01, 85.70it/s]

Calculating Affiliation Metrics:  95%|█████████▌| 190/200 [00:02<00:00, 96.22it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 97.08it/s]
Table1-only metrics completed in 3.84s
Evaluation Result:  {'AUC-PR': 0.004473529268976305, 'AUC-ROC': 0.7701061360614732, 'VUS-PR': 0.006400680210490531, 'VUS-ROC': 0.8131342184191905, 'Standard-F1': 0.011321822813461708, 'Standard-Precision': 0.00569313976658127, 'Standard-Recall': 0.99999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8176687553082637, 'Affiliation-P': 0.7034739878151361, 'Affiliation-R': 0.976122490275873, 'F1_T': 0.011205737337505663, 'Precision_T': 0.005634437683119224, 'Recall_T': 0.99999999999999}
Results for 429_UCR_id_127_Medical_tr_14825_1st_29000.csv: {'filename': '429_UCR_id_127_Medical_tr_14825_1st_29000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 232, 'train_index': '14825', 'data_shape': '59302x1', 'output_length': 45000, 'label_length': 44477, 'aligned_length': 44477, 'AUC-PR': 0.004473529268976305, 'AUC-ROC': 0.7701061360614732, 'VUS-PR': 0.006400680210490531, 'VUS-ROC': 0.8131342184191905, 'Standard-F1': 0.011321822813461708, 'Standard-Precision': 0.00569313976658127, 'Standard-Recall': 0.99999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8176687553082637, 'Affiliation-P': 0.7034739878151361, 'Affiliation-R': 0.976122490275873, 'F1_T': 0.011205737337505663, 'Precision_T': 0.005634437683119224, 'Recall_T': 0.99999999999999}
Logits metrics skipped for 429_UCR_id_127_Medical_tr_14825_1st_29000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/429_UCR_id_127_Medical_tr_14825_1st_29000_results.pkl
Processing file: 028_NAB_id_28_Facility_tr_1007_1st_3447.csv, Data shape: (4031, 1)
Label shape: (4031,)
Train data shape: (1007, 1), Test data shape: (3024, 1), Label test shape: (3024,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 028_NAB_id_28_Facility_tr_1007_1st_3447.csv...
Using CUDA device: 2
Data shape: (3024, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 3024, 1])
batched_mask shape: torch.Size([1, 3024])

100%|██████████| 1/1 [00:00<00:00, 160.14it/s]
Original shapes - Output: (3024,), Label: (3024,)
Aligned shapes - Output: (3024,), Label: (3024,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 506.25it/s]
Table1-only metrics completed in 1.50s
Evaluation Result:  {'AUC-PR': 0.5242416809923691, 'AUC-ROC': 0.8924449982966057, 'VUS-PR': 0.6003342551940815, 'VUS-ROC': 0.92389391279495, 'Standard-F1': 0.556783919597498, 'Standard-Precision': 0.49026548672566284, 'Standard-Recall': 0.6441860465116264, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9201193964299019, 'Affiliation-P': 0.8616874969094596, 'Affiliation-R': 0.9870524178663712, 'F1_T': 0.5556670010025173, 'Precision_T': 0.488536155202821, 'Recall_T': 0.6441860465116264}
Results for 028_NAB_id_28_Facility_tr_1007_1st_3447.csv: {'filename': '028_NAB_id_28_Facility_tr_1007_1st_3447.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 289, 'train_index': '1007', 'data_shape': '4031x1', 'output_length': 3024, 'label_length': 3024, 'aligned_length': 3024, 'AUC-PR': 0.5242416809923691, 'AUC-ROC': 0.8924449982966057, 'VUS-PR': 0.6003342551940815, 'VUS-ROC': 0.92389391279495, 'Standard-F1': 0.556783919597498, 'Standard-Precision': 0.49026548672566284, 'Standard-Recall': 0.6441860465116264, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9201193964299019, 'Affiliation-P': 0.8616874969094596, 'Affiliation-R': 0.9870524178663712, 'F1_T': 0.5556670010025173, 'Precision_T': 0.488536155202821, 'Recall_T': 0.6441860465116264}
Logits metrics skipped for 028_NAB_id_28_Facility_tr_1007_1st_3447.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/028_NAB_id_28_Facility_tr_1007_1st_3447_results.pkl
Processing file: 629_YAHOO_id_79_WebService_tr_500_1st_222.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 629_YAHOO_id_79_WebService_tr_500_1st_222.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.29it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 874.85it/s]
Table1-only metrics completed in 0.74s
Evaluation Result:  {'AUC-PR': 1.0000000000000002, 'AUC-ROC': 1.0, 'VUS-PR': 0.9308641975308646, 'VUS-ROC': 0.9997048001602519, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9925072690673221, 'Affiliation-P': 0.9851259851259851, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Results for 629_YAHOO_id_79_WebService_tr_500_1st_222.csv: {'filename': '629_YAHOO_id_79_WebService_tr_500_1st_222.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0000000000000002, 'AUC-ROC': 1.0, 'VUS-PR': 0.9308641975308646, 'VUS-ROC': 0.9997048001602519, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9925072690673221, 'Affiliation-P': 0.9851259851259851, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 629_YAHOO_id_79_WebService_tr_500_1st_222.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/629_YAHOO_id_79_WebService_tr_500_1st_222_results.pkl
Processing file: 518_UCR_id_216_Medical_tr_9625_1st_20950.csv, Data shape: (38501, 1)
Label shape: (38501,)
Train data shape: (9625, 1), Test data shape: (28876, 1), Label test shape: (28876,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 518_UCR_id_216_Medical_tr_9625_1st_20950.csv...
Using CUDA device: 2
Data shape: (28876, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 28876, window_size 5000, adding 1124 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 138.75it/s]
Original shapes - Output: (30000,), Label: (28876,)
Aligned shapes - Output: (28876,), Label: (28876,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [00:01<00:00, 125.33it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 143.14it/s]
Table1-only metrics completed in 1.73s
Evaluation Result:  {'AUC-PR': 0.42747055785330035, 'AUC-ROC': 0.9759288913643853, 'VUS-PR': 0.35310577394418374, 'VUS-ROC': 0.9797275747581747, 'Standard-F1': 0.43145161290280215, 'Standard-Precision': 0.3092485549132939, 'Standard-Recall': 0.7133333333333286, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8773763166142678, 'Affiliation-P': 0.7816732196429892, 'Affiliation-R': 0.999783441843284, 'F1_T': 0.4282460136669741, 'Precision_T': 0.3252595155709331, 'Recall_T': 0.6266666666666625}
Results for 518_UCR_id_216_Medical_tr_9625_1st_20950.csv: {'filename': '518_UCR_id_216_Medical_tr_9625_1st_20950.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 40, 'train_index': '9625', 'data_shape': '38501x1', 'output_length': 30000, 'label_length': 28876, 'aligned_length': 28876, 'AUC-PR': 0.42747055785330035, 'AUC-ROC': 0.9759288913643853, 'VUS-PR': 0.35310577394418374, 'VUS-ROC': 0.9797275747581747, 'Standard-F1': 0.43145161290280215, 'Standard-Precision': 0.3092485549132939, 'Standard-Recall': 0.7133333333333286, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8773763166142678, 'Affiliation-P': 0.7816732196429892, 'Affiliation-R': 0.999783441843284, 'F1_T': 0.4282460136669741, 'Precision_T': 0.3252595155709331, 'Recall_T': 0.6266666666666625}
Logits metrics skipped for 518_UCR_id_216_Medical_tr_9625_1st_20950.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/518_UCR_id_216_Medical_tr_9625_1st_20950_results.pkl
Processing file: 485_UCR_id_183_Medical_tr_7516_1st_17390.csv, Data shape: (30066, 1)
Label shape: (30066,)
Train data shape: (7516, 1), Test data shape: (22550, 1), Label test shape: (22550,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 485_UCR_id_183_Medical_tr_7516_1st_17390.csv...
Using CUDA device: 2
Data shape: (22550, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22550, window_size 5000, adding 2450 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 151.33it/s]
Original shapes - Output: (25000,), Label: (22550,)
Aligned shapes - Output: (22550,), Label: (22550,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  54%|█████▍    | 108/200 [00:01<00:00, 107.11it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 119.83it/s]
Table1-only metrics completed in 1.93s
Evaluation Result:  {'AUC-PR': 0.01190165920114515, 'AUC-ROC': 0.6145762711864406, 'VUS-PR': 0.01268545995068669, 'VUS-ROC': 0.6405718647032386, 'Standard-F1': 0.043545878693441835, 'Standard-Precision': 0.024221453287197214, 'Standard-Recall': 0.21538461538461373, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6667017657743896, 'Affiliation-P': 0.5000394875356718, 'Affiliation-R': 1.0, 'F1_T': 0.043408360128430135, 'Precision_T': 0.024236983842010753, 'Recall_T': 0.2076923076923061}
Results for 485_UCR_id_183_Medical_tr_7516_1st_17390.csv: {'filename': '485_UCR_id_183_Medical_tr_7516_1st_17390.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 34, 'train_index': '7516', 'data_shape': '30066x1', 'output_length': 25000, 'label_length': 22550, 'aligned_length': 22550, 'AUC-PR': 0.01190165920114515, 'AUC-ROC': 0.6145762711864406, 'VUS-PR': 0.01268545995068669, 'VUS-ROC': 0.6405718647032386, 'Standard-F1': 0.043545878693441835, 'Standard-Precision': 0.024221453287197214, 'Standard-Recall': 0.21538461538461373, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6667017657743896, 'Affiliation-P': 0.5000394875356718, 'Affiliation-R': 1.0, 'F1_T': 0.043408360128430135, 'Precision_T': 0.024236983842010753, 'Recall_T': 0.2076923076923061}
Logits metrics skipped for 485_UCR_id_183_Medical_tr_7516_1st_17390.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/485_UCR_id_183_Medical_tr_7516_1st_17390_results.pkl
Processing file: 050_WSD_id_22_WebService_tr_4559_1st_12816.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 050_WSD_id_22_WebService_tr_4559_1st_12816.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.12it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 307.97it/s]
Table1-only metrics completed in 0.77s
Evaluation Result:  {'AUC-PR': 0.09748548738731204, 'AUC-ROC': 0.6525560275377179, 'VUS-PR': 0.0633936730339713, 'VUS-ROC': 0.7289363330654103, 'Standard-F1': 0.06716417910432021, 'Standard-Precision': 0.03673469387755087, 'Standard-Recall': 0.39130434782607, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9289656695885791, 'Affiliation-P': 0.8679789300348729, 'Affiliation-R': 0.999170298597137, 'F1_T': 0.061016949152398185, 'Precision_T': 0.03308823529411752, 'Recall_T': 0.39130434782607}
Results for 050_WSD_id_22_WebService_tr_4559_1st_12816.csv: {'filename': '050_WSD_id_22_WebService_tr_4559_1st_12816.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 14, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.09748548738731204, 'AUC-ROC': 0.6525560275377179, 'VUS-PR': 0.0633936730339713, 'VUS-ROC': 0.7289363330654103, 'Standard-F1': 0.06716417910432021, 'Standard-Precision': 0.03673469387755087, 'Standard-Recall': 0.39130434782607, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9289656695885791, 'Affiliation-P': 0.8679789300348729, 'Affiliation-R': 0.999170298597137, 'F1_T': 0.061016949152398185, 'Precision_T': 0.03308823529411752, 'Recall_T': 0.39130434782607}
Logits metrics skipped for 050_WSD_id_22_WebService_tr_4559_1st_12816.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/050_WSD_id_22_WebService_tr_4559_1st_12816_results.pkl
Processing file: 489_UCR_id_187_Medical_tr_1875_1st_4485.csv, Data shape: (7500, 1)
Label shape: (7500,)
Train data shape: (1875, 1), Test data shape: (5625, 1), Label test shape: (5625,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 489_UCR_id_187_Medical_tr_1875_1st_4485.csv...
Using CUDA device: 2
Data shape: (5625, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5625, window_size 5000, adding 4375 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 155.99it/s]
Original shapes - Output: (10000,), Label: (5625,)
Aligned shapes - Output: (5625,), Label: (5625,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 592.39it/s]
Table1-only metrics completed in 1.06s
Evaluation Result:  {'AUC-PR': 0.9988917306052856, 'AUC-ROC': 0.9999769235958007, 'VUS-PR': 0.9887527422717804, 'VUS-ROC': 0.9998774091353424, 'Standard-F1': 0.995073891625106, 'Standard-Precision': 0.9999999999999901, 'Standard-Recall': 0.9901960784313629, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.99987362237494, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9997472766884532, 'F1_T': 0.9439252336443521, 'Precision_T': 0.9017857142857063, 'Recall_T': 0.9901960784313629}
Results for 489_UCR_id_187_Medical_tr_1875_1st_4485.csv: {'filename': '489_UCR_id_187_Medical_tr_1875_1st_4485.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 183, 'train_index': '1875', 'data_shape': '7500x1', 'output_length': 10000, 'label_length': 5625, 'aligned_length': 5625, 'AUC-PR': 0.9988917306052856, 'AUC-ROC': 0.9999769235958007, 'VUS-PR': 0.9887527422717804, 'VUS-ROC': 0.9998774091353424, 'Standard-F1': 0.995073891625106, 'Standard-Precision': 0.9999999999999901, 'Standard-Recall': 0.9901960784313629, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.99987362237494, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9997472766884532, 'F1_T': 0.9439252336443521, 'Precision_T': 0.9017857142857063, 'Recall_T': 0.9901960784313629}
Logits metrics skipped for 489_UCR_id_187_Medical_tr_1875_1st_4485.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/489_UCR_id_187_Medical_tr_1875_1st_4485_results.pkl
Processing file: 706_YAHOO_id_156_WebService_tr_500_1st_201.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 706_YAHOO_id_156_WebService_tr_500_1st_201.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.75it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 775.17it/s]
Table1-only metrics completed in 0.74s
Evaluation Result:  {'AUC-PR': 0.9999999999999999, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.9090909090903305, 'Standard-Precision': 0.8333333333332639, 'Standard-Recall': 0.9999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.998933901918976, 'Affiliation-P': 0.9978700745473909, 'Affiliation-R': 1.0, 'F1_T': 0.9090909090903305, 'Precision_T': 0.8333333333332639, 'Recall_T': 0.9999999999999}
Results for 706_YAHOO_id_156_WebService_tr_500_1st_201.csv: {'filename': '706_YAHOO_id_156_WebService_tr_500_1st_201.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9999999999999999, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.9090909090903305, 'Standard-Precision': 0.8333333333332639, 'Standard-Recall': 0.9999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.998933901918976, 'Affiliation-P': 0.9978700745473909, 'Affiliation-R': 1.0, 'F1_T': 0.9090909090903305, 'Precision_T': 0.8333333333332639, 'Recall_T': 0.9999999999999}
Logits metrics skipped for 706_YAHOO_id_156_WebService_tr_500_1st_201.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/706_YAHOO_id_156_WebService_tr_500_1st_201_results.pkl
Processing file: 487_UCR_id_185_Medical_tr_49938_1st_191200.csv, Data shape: (199753, 1)
Label shape: (199753,)
Train data shape: (49938, 1), Test data shape: (149815, 1), Label test shape: (149815,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 487_UCR_id_185_Medical_tr_49938_1st_191200.csv...
Using CUDA device: 2
Data shape: (149815, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 149815, window_size 5000, adding 185 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([30, 5000, 1])
batched_mask shape: torch.Size([30, 5000])

100%|██████████| 1/1 [00:00<00:00, 47.25it/s]
Original shapes - Output: (150000,), Label: (149815,)
Aligned shapes - Output: (149815,), Label: (149815,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  14%|█▍        | 29/200 [00:01<00:06, 28.26it/s]

Calculating Affiliation Metrics:  29%|██▉       | 58/200 [00:02<00:05, 25.10it/s]

Calculating Affiliation Metrics:  42%|████▎     | 85/200 [00:03<00:04, 25.62it/s]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [00:04<00:03, 27.25it/s]

Calculating Affiliation Metrics:  75%|███████▌  | 150/200 [00:05<00:01, 29.69it/s]

Calculating Affiliation Metrics:  98%|█████████▊| 195/200 [00:06<00:00, 34.69it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:06<00:00, 31.24it/s]
Table1-only metrics completed in 10.92s
Evaluation Result:  {'AUC-PR': 0.005679568194232504, 'AUC-ROC': 0.6062828784783846, 'VUS-PR': 0.006187993178947254, 'VUS-ROC': 0.6282899728966381, 'Standard-F1': 0.019975468722537564, 'Standard-Precision': 0.010445299615173171, 'Standard-Recall': 0.22799999999999954, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6701148370149912, 'Affiliation-P': 0.5097970253738476, 'Affiliation-R': 0.977518940026032, 'F1_T': 0.020017256255310017, 'Precision_T': 0.010459873760144272, 'Recall_T': 0.2319999999999995}
Results for 487_UCR_id_185_Medical_tr_49938_1st_191200.csv: {'filename': '487_UCR_id_185_Medical_tr_49938_1st_191200.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '49938', 'data_shape': '199753x1', 'output_length': 150000, 'label_length': 149815, 'aligned_length': 149815, 'AUC-PR': 0.005679568194232504, 'AUC-ROC': 0.6062828784783846, 'VUS-PR': 0.006187993178947254, 'VUS-ROC': 0.6282899728966381, 'Standard-F1': 0.019975468722537564, 'Standard-Precision': 0.010445299615173171, 'Standard-Recall': 0.22799999999999954, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6701148370149912, 'Affiliation-P': 0.5097970253738476, 'Affiliation-R': 0.977518940026032, 'F1_T': 0.020017256255310017, 'Precision_T': 0.010459873760144272, 'Recall_T': 0.2319999999999995}
Logits metrics skipped for 487_UCR_id_185_Medical_tr_49938_1st_191200.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/487_UCR_id_185_Medical_tr_49938_1st_191200_results.pkl
Processing file: 786_YAHOO_id_236_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 786_YAHOO_id_236_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 147.92it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1669.99it/s]
Table1-only metrics completed in 0.82s
Evaluation Result:  {'AUC-PR': 0.8691718775052109, 'AUC-ROC': 0.9990253411306044, 'VUS-PR': 0.9943053101472179, 'VUS-ROC': 0.9999613014421813, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 786_YAHOO_id_236_Synthetic_tr_500_1st_623.csv: {'filename': '786_YAHOO_id_236_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 118, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.8691718775052109, 'AUC-ROC': 0.9990253411306044, 'VUS-PR': 0.9943053101472179, 'VUS-ROC': 0.9999613014421813, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 786_YAHOO_id_236_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/786_YAHOO_id_236_Synthetic_tr_500_1st_623_results.pkl
Processing file: 075_WSD_id_47_WebService_tr_3296_1st_3468.csv, Data shape: (13186, 1)
Label shape: (13186,)
Train data shape: (3296, 1), Test data shape: (9890, 1), Label test shape: (9890,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 075_WSD_id_47_WebService_tr_3296_1st_3468.csv...
Using CUDA device: 2
Data shape: (9890, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9890, window_size 5000, adding 110 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 156.19it/s]
Original shapes - Output: (10000,), Label: (9890,)
Aligned shapes - Output: (9890,), Label: (9890,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 344.44it/s]
Table1-only metrics completed in 1.81s
Evaluation Result:  {'AUC-PR': 0.15070917081118065, 'AUC-ROC': 0.8204218308327897, 'VUS-PR': 0.0639449281890801, 'VUS-ROC': 0.9421536188802568, 'Standard-F1': 0.07462686567125473, 'Standard-Precision': 0.05050505050505, 'Standard-Recall': 0.14285714285713877, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.953312620128151, 'Affiliation-P': 0.9138095923417254, 'Affiliation-R': 0.9963853093298246, 'F1_T': 0.07462686567125473, 'Precision_T': 0.05050505050505, 'Recall_T': 0.14285714285713877}
Results for 075_WSD_id_47_WebService_tr_3296_1st_3468.csv: {'filename': '075_WSD_id_47_WebService_tr_3296_1st_3468.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '3296', 'data_shape': '13186x1', 'output_length': 10000, 'label_length': 9890, 'aligned_length': 9890, 'AUC-PR': 0.15070917081118065, 'AUC-ROC': 0.8204218308327897, 'VUS-PR': 0.0639449281890801, 'VUS-ROC': 0.9421536188802568, 'Standard-F1': 0.07462686567125473, 'Standard-Precision': 0.05050505050505, 'Standard-Recall': 0.14285714285713877, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.953312620128151, 'Affiliation-P': 0.9138095923417254, 'Affiliation-R': 0.9963853093298246, 'F1_T': 0.07462686567125473, 'Precision_T': 0.05050505050505, 'Recall_T': 0.14285714285713877}
Logits metrics skipped for 075_WSD_id_47_WebService_tr_3296_1st_3468.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/075_WSD_id_47_WebService_tr_3296_1st_3468_results.pkl
Processing file: 807_YAHOO_id_257_WebService_tr_500_1st_1210.csv, Data shape: (1424, 1)
Label shape: (1424,)
Train data shape: (500, 1), Test data shape: (924, 1), Label test shape: (924,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 807_YAHOO_id_257_WebService_tr_500_1st_1210.csv...
Using CUDA device: 2
Data shape: (924, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 924, 1])
batched_mask shape: torch.Size([1, 924])

100%|██████████| 1/1 [00:00<00:00, 135.84it/s]
Original shapes - Output: (924,), Label: (924,)
Aligned shapes - Output: (924,), Label: (924,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1712.43it/s]
Table1-only metrics completed in 0.29s
Evaluation Result:  {'AUC-PR': 0.3383270679297945, 'AUC-ROC': 0.9663402295868734, 'VUS-PR': 0.6332401501807817, 'VUS-ROC': 0.9873088321276714, 'Standard-F1': 0.41176470588196573, 'Standard-Precision': 0.2745098039215632, 'Standard-Recall': 0.8235294117646575, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8911557706012364, 'Affiliation-P': 0.8782050274941083, 'Affiliation-R': 0.9044941956882256, 'F1_T': 0.41176470588196573, 'Precision_T': 0.2745098039215632, 'Recall_T': 0.8235294117646575}
Results for 807_YAHOO_id_257_WebService_tr_500_1st_1210.csv: {'filename': '807_YAHOO_id_257_WebService_tr_500_1st_1210.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1424x1', 'output_length': 924, 'label_length': 924, 'aligned_length': 924, 'AUC-PR': 0.3383270679297945, 'AUC-ROC': 0.9663402295868734, 'VUS-PR': 0.6332401501807817, 'VUS-ROC': 0.9873088321276714, 'Standard-F1': 0.41176470588196573, 'Standard-Precision': 0.2745098039215632, 'Standard-Recall': 0.8235294117646575, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8911557706012364, 'Affiliation-P': 0.8782050274941083, 'Affiliation-R': 0.9044941956882256, 'F1_T': 0.41176470588196573, 'Precision_T': 0.2745098039215632, 'Recall_T': 0.8235294117646575}
Logits metrics skipped for 807_YAHOO_id_257_WebService_tr_500_1st_1210.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/807_YAHOO_id_257_WebService_tr_500_1st_1210_results.pkl
Processing file: 750_YAHOO_id_200_WebService_tr_500_1st_591.csv, Data shape: (1427, 1)
Label shape: (1427,)
Train data shape: (500, 1), Test data shape: (927, 1), Label test shape: (927,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 750_YAHOO_id_200_WebService_tr_500_1st_591.csv...
Using CUDA device: 2
Data shape: (927, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 927, 1])
batched_mask shape: torch.Size([1, 927])

100%|██████████| 1/1 [00:00<00:00, 131.56it/s]
Original shapes - Output: (927,), Label: (927,)
Aligned shapes - Output: (927,), Label: (927,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1831.60it/s]
Table1-only metrics completed in 0.34s
Evaluation Result:  {'AUC-PR': 0.4078490960502905, 'AUC-ROC': 0.6274809160305344, 'VUS-PR': 0.5503533158125634, 'VUS-ROC': 0.9009542927297729, 'Standard-F1': 0.39999999999946, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.39999999999996, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8219688835835831, 'Affiliation-P': 0.7018203211421781, 'Affiliation-R': 0.9917525773195877, 'F1_T': 0.39999999999946, 'Precision_T': 0.39999999999996, 'Recall_T': 0.39999999999996}
Results for 750_YAHOO_id_200_WebService_tr_500_1st_591.csv: {'filename': '750_YAHOO_id_200_WebService_tr_500_1st_591.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1427x1', 'output_length': 927, 'label_length': 927, 'aligned_length': 927, 'AUC-PR': 0.4078490960502905, 'AUC-ROC': 0.6274809160305344, 'VUS-PR': 0.5503533158125634, 'VUS-ROC': 0.9009542927297729, 'Standard-F1': 0.39999999999946, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.39999999999996, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8219688835835831, 'Affiliation-P': 0.7018203211421781, 'Affiliation-R': 0.9917525773195877, 'F1_T': 0.39999999999946, 'Precision_T': 0.39999999999996, 'Recall_T': 0.39999999999996}
Logits metrics skipped for 750_YAHOO_id_200_WebService_tr_500_1st_591.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/750_YAHOO_id_200_WebService_tr_500_1st_591_results.pkl
Processing file: 403_UCR_id_101_Medical_tr_1875_1st_4187.csv, Data shape: (7501, 1)
Label shape: (7501,)
Train data shape: (1875, 1), Test data shape: (5626, 1), Label test shape: (5626,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 403_UCR_id_101_Medical_tr_1875_1st_4187.csv...
Using CUDA device: 2
Data shape: (5626, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5626, window_size 5000, adding 4374 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.52it/s]
Original shapes - Output: (10000,), Label: (5626,)
Aligned shapes - Output: (5626,), Label: (5626,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 633.10it/s]
Table1-only metrics completed in 0.99s
Evaluation Result:  {'AUC-PR': 0.8237846089729651, 'AUC-ROC': 0.9989163994774967, 'VUS-PR': 0.6864998768187986, 'VUS-ROC': 0.9993399606827884, 'Standard-F1': 0.3188405797098483, 'Standard-Precision': 0.19298245614034748, 'Standard-Recall': 0.9166666666665902, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9928132886139225, 'Affiliation-P': 0.985743530652022, 'Affiliation-R': 0.9999851878184619, 'F1_T': 0.3188405797098483, 'Precision_T': 0.19298245614034748, 'Recall_T': 0.9166666666665902}
Results for 403_UCR_id_101_Medical_tr_1875_1st_4187.csv: {'filename': '403_UCR_id_101_Medical_tr_1875_1st_4187.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 183, 'train_index': '1875', 'data_shape': '7501x1', 'output_length': 10000, 'label_length': 5626, 'aligned_length': 5626, 'AUC-PR': 0.8237846089729651, 'AUC-ROC': 0.9989163994774967, 'VUS-PR': 0.6864998768187986, 'VUS-ROC': 0.9993399606827884, 'Standard-F1': 0.3188405797098483, 'Standard-Precision': 0.19298245614034748, 'Standard-Recall': 0.9166666666665902, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9928132886139225, 'Affiliation-P': 0.985743530652022, 'Affiliation-R': 0.9999851878184619, 'F1_T': 0.3188405797098483, 'Precision_T': 0.19298245614034748, 'Recall_T': 0.9166666666665902}
Logits metrics skipped for 403_UCR_id_101_Medical_tr_1875_1st_4187.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/403_UCR_id_101_Medical_tr_1875_1st_4187_results.pkl
Processing file: 628_YAHOO_id_78_WebService_tr_500_1st_147.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 628_YAHOO_id_78_WebService_tr_500_1st_147.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.37it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1104.22it/s]
Table1-only metrics completed in 0.54s
Evaluation Result:  {'AUC-PR': 0.976190476190476, 'AUC-ROC': 0.9998580352072687, 'VUS-PR': 0.9538803429243552, 'VUS-ROC': 0.9998034791218806, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9985494106980956, 'Affiliation-P': 0.9971030237189934, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Results for 628_YAHOO_id_78_WebService_tr_500_1st_147.csv: {'filename': '628_YAHOO_id_78_WebService_tr_500_1st_147.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.976190476190476, 'AUC-ROC': 0.9998580352072687, 'VUS-PR': 0.9538803429243552, 'VUS-ROC': 0.9998034791218806, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9985494106980956, 'Affiliation-P': 0.9971030237189934, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Logits metrics skipped for 628_YAHOO_id_78_WebService_tr_500_1st_147.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/628_YAHOO_id_78_WebService_tr_500_1st_147_results.pkl
Processing file: 357_UCR_id_55_Facility_tr_12500_1st_46600.csv, Data shape: (50000, 1)
Label shape: (50000,)
Train data shape: (12500, 1), Test data shape: (37500, 1), Label test shape: (37500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 357_UCR_id_55_Facility_tr_12500_1st_46600.csv...
Using CUDA device: 2
Data shape: (37500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 37500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([8, 5000, 1])
batched_mask shape: torch.Size([8, 5000])

100%|██████████| 1/1 [00:00<00:00, 128.56it/s]
Original shapes - Output: (40000,), Label: (37500,)
Aligned shapes - Output: (37500,), Label: (37500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  54%|█████▎    | 107/200 [00:01<00:00, 106.71it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 126.40it/s]
Table1-only metrics completed in 3.03s
Evaluation Result:  {'AUC-PR': 0.8910255980635616, 'AUC-ROC': 0.9972550627240143, 'VUS-PR': 0.8389855116540438, 'VUS-ROC': 0.9973655699638693, 'Standard-F1': 0.8474074074069109, 'Standard-Precision': 0.7626666666666646, 'Standard-Recall': 0.95333333333333, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.989793002220548, 'Affiliation-P': 0.9798689483065953, 'Affiliation-R': 0.9999201333333334, 'F1_T': 0.8474074074069109, 'Precision_T': 0.7626666666666646, 'Recall_T': 0.95333333333333}
Results for 357_UCR_id_55_Facility_tr_12500_1st_46600.csv: {'filename': '357_UCR_id_55_Facility_tr_12500_1st_46600.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 208, 'train_index': '12500', 'data_shape': '50000x1', 'output_length': 40000, 'label_length': 37500, 'aligned_length': 37500, 'AUC-PR': 0.8910255980635616, 'AUC-ROC': 0.9972550627240143, 'VUS-PR': 0.8389855116540438, 'VUS-ROC': 0.9973655699638693, 'Standard-F1': 0.8474074074069109, 'Standard-Precision': 0.7626666666666646, 'Standard-Recall': 0.95333333333333, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.989793002220548, 'Affiliation-P': 0.9798689483065953, 'Affiliation-R': 0.9999201333333334, 'F1_T': 0.8474074074069109, 'Precision_T': 0.7626666666666646, 'Recall_T': 0.95333333333333}
Logits metrics skipped for 357_UCR_id_55_Facility_tr_12500_1st_46600.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/357_UCR_id_55_Facility_tr_12500_1st_46600_results.pkl
Processing file: 407_UCR_id_105_Medical_tr_7500_1st_16900.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 407_UCR_id_105_Medical_tr_7500_1st_16900.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 146.82it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 200.23it/s]
Table1-only metrics completed in 1.92s
Evaluation Result:  {'AUC-PR': 0.5834004096712747, 'AUC-ROC': 0.9909013452914798, 'VUS-PR': 0.505436180138896, 'VUS-ROC': 0.9909014951572467, 'Standard-F1': 0.4767025089601118, 'Standard-Precision': 0.37150837988826707, 'Standard-Recall': 0.6649999999999967, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.816967400713032, 'Affiliation-P': 0.6912109161793372, 'Affiliation-R': 0.9986601111111111, 'F1_T': 0.4776119402980378, 'Precision_T': 0.38095238095237977, 'Recall_T': 0.6399999999999968}
Results for 407_UCR_id_105_Medical_tr_7500_1st_16900.csv: {'filename': '407_UCR_id_105_Medical_tr_7500_1st_16900.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 166, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.5834004096712747, 'AUC-ROC': 0.9909013452914798, 'VUS-PR': 0.505436180138896, 'VUS-ROC': 0.9909014951572467, 'Standard-F1': 0.4767025089601118, 'Standard-Precision': 0.37150837988826707, 'Standard-Recall': 0.6649999999999967, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.816967400713032, 'Affiliation-P': 0.6912109161793372, 'Affiliation-R': 0.9986601111111111, 'F1_T': 0.4776119402980378, 'Precision_T': 0.38095238095237977, 'Recall_T': 0.6399999999999968}
Logits metrics skipped for 407_UCR_id_105_Medical_tr_7500_1st_16900.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/407_UCR_id_105_Medical_tr_7500_1st_16900_results.pkl
Processing file: 665_YAHOO_id_115_WebService_tr_500_1st_28.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 665_YAHOO_id_115_WebService_tr_500_1st_28.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.11it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 598.06it/s]
Table1-only metrics completed in 0.94s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9690984850934948, 'VUS-ROC': 0.9998014903784371, 'Standard-F1': 0.9999999999994167, 'Standard-Precision': 0.9999999999999166, 'Standard-Recall': 0.9999999999999166, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.9999999999994167, 'Precision_T': 0.9999999999999166, 'Recall_T': 0.9999999999999166}
Results for 665_YAHOO_id_115_WebService_tr_500_1st_28.csv: {'filename': '665_YAHOO_id_115_WebService_tr_500_1st_28.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9690984850934948, 'VUS-ROC': 0.9998014903784371, 'Standard-F1': 0.9999999999994167, 'Standard-Precision': 0.9999999999999166, 'Standard-Recall': 0.9999999999999166, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.9999999999994167, 'Precision_T': 0.9999999999999166, 'Recall_T': 0.9999999999999166}
Logits metrics skipped for 665_YAHOO_id_115_WebService_tr_500_1st_28.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/665_YAHOO_id_115_WebService_tr_500_1st_28_results.pkl
Skipping file: 207_SMD_id_30_Facility_tr_7174_1st_21230.csv due to filter criteria.
Processing file: 625_YAHOO_id_75_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 625_YAHOO_id_75_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 158.49it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1933.96it/s]
Table1-only metrics completed in 1.16s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 625_YAHOO_id_75_Synthetic_tr_500_1st_658.csv: {'filename': '625_YAHOO_id_75_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 235, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 625_YAHOO_id_75_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/625_YAHOO_id_75_Synthetic_tr_500_1st_658_results.pkl
Processing file: 699_YAHOO_id_149_WebService_tr_500_1st_99.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 699_YAHOO_id_149_WebService_tr_500_1st_99.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.13it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1906.16it/s]
Table1-only metrics completed in 0.20s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8799385080463143, 'Affiliation-P': 0.7856162490788504, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Results for 699_YAHOO_id_149_WebService_tr_500_1st_99.csv: {'filename': '699_YAHOO_id_149_WebService_tr_500_1st_99.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8799385080463143, 'Affiliation-P': 0.7856162490788504, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 699_YAHOO_id_149_WebService_tr_500_1st_99.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/699_YAHOO_id_149_WebService_tr_500_1st_99_results.pkl
Processing file: 342_UCR_id_40_Medical_tr_7500_1st_17000.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 342_UCR_id_40_Medical_tr_7500_1st_17000.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 152.80it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 209.38it/s]
Table1-only metrics completed in 1.86s
Evaluation Result:  {'AUC-PR': 0.015355494252803564, 'AUC-ROC': 0.841820982142857, 'VUS-PR': 0.016410809845367944, 'VUS-ROC': 0.8655092396962168, 'Standard-F1': 0.045011756802084864, 'Standard-Precision': 0.02328814737573861, 'Standard-Recall': 0.6699999999999934, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7050878242954721, 'Affiliation-P': 0.5445075415543575, 'Affiliation-R': 0.9999957777777779, 'F1_T': 0.04489112227799217, 'Precision_T': 0.02322357019064124, 'Recall_T': 0.6699999999999934}
Results for 342_UCR_id_40_Medical_tr_7500_1st_17000.csv: {'filename': '342_UCR_id_40_Medical_tr_7500_1st_17000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 166, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.015355494252803564, 'AUC-ROC': 0.841820982142857, 'VUS-PR': 0.016410809845367944, 'VUS-ROC': 0.8655092396962168, 'Standard-F1': 0.045011756802084864, 'Standard-Precision': 0.02328814737573861, 'Standard-Recall': 0.6699999999999934, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7050878242954721, 'Affiliation-P': 0.5445075415543575, 'Affiliation-R': 0.9999957777777779, 'F1_T': 0.04489112227799217, 'Precision_T': 0.02322357019064124, 'Recall_T': 0.6699999999999934}
Logits metrics skipped for 342_UCR_id_40_Medical_tr_7500_1st_17000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/342_UCR_id_40_Medical_tr_7500_1st_17000_results.pkl
Processing file: 359_UCR_id_57_Medical_tr_1913_1st_3474.csv, Data shape: (7654, 1)
Label shape: (7654,)
Train data shape: (1913, 1), Test data shape: (5741, 1), Label test shape: (5741,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 359_UCR_id_57_Medical_tr_1913_1st_3474.csv...
Using CUDA device: 2
Data shape: (5741, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5741, window_size 5000, adding 4259 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 62.86it/s]
Original shapes - Output: (10000,), Label: (5741,)
Aligned shapes - Output: (5741,), Label: (5741,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 714.20it/s]
Table1-only metrics completed in 0.87s
Evaluation Result:  {'AUC-PR': 0.5834423571584036, 'AUC-ROC': 0.945149740711225, 'VUS-PR': 0.6902213129863122, 'VUS-ROC': 0.9729331390536279, 'Standard-F1': 0.5637982195840695, 'Standard-Precision': 0.5219780219780191, 'Standard-Recall': 0.6129032258064476, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9939568218439644, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9879862449500199, 'F1_T': 0.5536723163836853, 'Precision_T': 0.49246231155778647, 'Recall_T': 0.632258064516125}
Results for 359_UCR_id_57_Medical_tr_1913_1st_3474.csv: {'filename': '359_UCR_id_57_Medical_tr_1913_1st_3474.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 154, 'train_index': '1913', 'data_shape': '7654x1', 'output_length': 10000, 'label_length': 5741, 'aligned_length': 5741, 'AUC-PR': 0.5834423571584036, 'AUC-ROC': 0.945149740711225, 'VUS-PR': 0.6902213129863122, 'VUS-ROC': 0.9729331390536279, 'Standard-F1': 0.5637982195840695, 'Standard-Precision': 0.5219780219780191, 'Standard-Recall': 0.6129032258064476, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9939568218439644, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9879862449500199, 'F1_T': 0.5536723163836853, 'Precision_T': 0.49246231155778647, 'Recall_T': 0.632258064516125}
Logits metrics skipped for 359_UCR_id_57_Medical_tr_1913_1st_3474.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/359_UCR_id_57_Medical_tr_1913_1st_3474_results.pkl
Processing file: 804_YAHOO_id_254_WebService_tr_500_1st_420.csv, Data shape: (1461, 1)
Label shape: (1461,)
Train data shape: (500, 1), Test data shape: (961, 1), Label test shape: (961,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 804_YAHOO_id_254_WebService_tr_500_1st_420.csv...
Using CUDA device: 2
Data shape: (961, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 961, 1])
batched_mask shape: torch.Size([1, 961])

100%|██████████| 1/1 [00:00<00:00, 161.69it/s]
Original shapes - Output: (961,), Label: (961,)
Aligned shapes - Output: (961,), Label: (961,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2075.68it/s]
Table1-only metrics completed in 0.24s
Evaluation Result:  {'AUC-PR': 0.9166666666666665, 'AUC-ROC': 0.999652052887961, 'VUS-PR': 0.7916666666666667, 'VUS-ROC': 0.9995650661099517, 'Standard-F1': 0.46153846153803546, 'Standard-Precision': 0.29999999999997, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9193710789249471, 'Affiliation-P': 0.8507740825688073, 'Affiliation-R': 1.0, 'F1_T': 0.46153846153803546, 'Precision_T': 0.29999999999997, 'Recall_T': 0.9999999999996666}
Results for 804_YAHOO_id_254_WebService_tr_500_1st_420.csv: {'filename': '804_YAHOO_id_254_WebService_tr_500_1st_420.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1461x1', 'output_length': 961, 'label_length': 961, 'aligned_length': 961, 'AUC-PR': 0.9166666666666665, 'AUC-ROC': 0.999652052887961, 'VUS-PR': 0.7916666666666667, 'VUS-ROC': 0.9995650661099517, 'Standard-F1': 0.46153846153803546, 'Standard-Precision': 0.29999999999997, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9193710789249471, 'Affiliation-P': 0.8507740825688073, 'Affiliation-R': 1.0, 'F1_T': 0.46153846153803546, 'Precision_T': 0.29999999999997, 'Recall_T': 0.9999999999996666}
Logits metrics skipped for 804_YAHOO_id_254_WebService_tr_500_1st_420.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/804_YAHOO_id_254_WebService_tr_500_1st_420_results.pkl
Processing file: 327_UCR_id_25_Medical_tr_15000_1st_51370.csv, Data shape: (60000, 1)
Label shape: (60000,)
Train data shape: (15000, 1), Test data shape: (45000, 1), Label test shape: (45000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 327_UCR_id_25_Medical_tr_15000_1st_51370.csv...
Using CUDA device: 2
Data shape: (45000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([9, 5000, 1])
batched_mask shape: torch.Size([9, 5000])

100%|██████████| 1/1 [00:00<00:00, 112.07it/s]
Original shapes - Output: (45000,), Label: (45000,)
Aligned shapes - Output: (45000,), Label: (45000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  42%|████▎     | 85/200 [00:01<00:01, 84.02it/s]

Calculating Affiliation Metrics:  90%|█████████ | 181/200 [00:02<00:00, 90.84it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 94.02it/s]
Table1-only metrics completed in 3.45s
Evaluation Result:  {'AUC-PR': 0.07690708279609626, 'AUC-ROC': 0.9397715147367847, 'VUS-PR': 0.09777129086897221, 'VUS-ROC': 0.9536184881825629, 'Standard-F1': 0.18089085345098702, 'Standard-Precision': 0.10346168805912093, 'Standard-Recall': 0.7189189189189169, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6899773905749758, 'Affiliation-P': 0.5267586845466156, 'Affiliation-R': 0.9997568468468468, 'F1_T': 0.18046181172268452, 'Precision_T': 0.1038854805725971, 'Recall_T': 0.6864864864864846}
Results for 327_UCR_id_25_Medical_tr_15000_1st_51370.csv: {'filename': '327_UCR_id_25_Medical_tr_15000_1st_51370.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 172, 'train_index': '15000', 'data_shape': '60000x1', 'output_length': 45000, 'label_length': 45000, 'aligned_length': 45000, 'AUC-PR': 0.07690708279609626, 'AUC-ROC': 0.9397715147367847, 'VUS-PR': 0.09777129086897221, 'VUS-ROC': 0.9536184881825629, 'Standard-F1': 0.18089085345098702, 'Standard-Precision': 0.10346168805912093, 'Standard-Recall': 0.7189189189189169, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6899773905749758, 'Affiliation-P': 0.5267586845466156, 'Affiliation-R': 0.9997568468468468, 'F1_T': 0.18046181172268452, 'Precision_T': 0.1038854805725971, 'Recall_T': 0.6864864864864846}
Logits metrics skipped for 327_UCR_id_25_Medical_tr_15000_1st_51370.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/327_UCR_id_25_Medical_tr_15000_1st_51370_results.pkl
Processing file: 341_UCR_id_39_HumanActivity_tr_48812_1st_128430.csv, Data shape: (195250, 1)
Label shape: (195250,)
Train data shape: (48812, 1), Test data shape: (146438, 1), Label test shape: (146438,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 341_UCR_id_39_HumanActivity_tr_48812_1st_128430.csv...
Using CUDA device: 2
Data shape: (146438, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 146438, window_size 5000, adding 3562 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([30, 5000, 1])
batched_mask shape: torch.Size([30, 5000])

100%|██████████| 1/1 [00:00<00:00, 46.85it/s]
Original shapes - Output: (150000,), Label: (146438,)
Aligned shapes - Output: (146438,), Label: (146438,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  13%|█▎        | 26/200 [00:01<00:06, 25.82it/s]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:02<00:06, 22.09it/s]

Calculating Affiliation Metrics:  38%|███▊      | 75/200 [00:03<00:05, 21.63it/s]

Calculating Affiliation Metrics:  49%|████▉     | 98/200 [00:04<00:04, 22.05it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 124/200 [00:05<00:03, 23.13it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 155/200 [00:06<00:01, 25.69it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:07<00:00, 27.00it/s]
Table1-only metrics completed in 12.70s
Evaluation Result:  {'AUC-PR': 8.251914444151043e-06, 'AUC-ROC': 0.17245641470393414, 'VUS-PR': 1.3570465860655272e-05, 'VUS-ROC': 0.21023642632755693, 'Standard-F1': 1.6475550283362992e-05, 'Standard-Precision': 8.237843003188046e-06, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8470199415569087, 'Affiliation-P': 0.7445102573830028, 'Affiliation-R': 0.9822655321706114, 'F1_T': 1.6407294683199748e-05, 'Precision_T': 8.203714641989893e-06, 'Recall_T': 0.9999999999989999}
Results for 341_UCR_id_39_HumanActivity_tr_48812_1st_128430.csv: {'filename': '341_UCR_id_39_HumanActivity_tr_48812_1st_128430.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '48812', 'data_shape': '195250x1', 'output_length': 150000, 'label_length': 146438, 'aligned_length': 146438, 'AUC-PR': 8.251914444151043e-06, 'AUC-ROC': 0.17245641470393414, 'VUS-PR': 1.3570465860655272e-05, 'VUS-ROC': 0.21023642632755693, 'Standard-F1': 1.6475550283362992e-05, 'Standard-Precision': 8.237843003188046e-06, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8470199415569087, 'Affiliation-P': 0.7445102573830028, 'Affiliation-R': 0.9822655321706114, 'F1_T': 1.6407294683199748e-05, 'Precision_T': 8.203714641989893e-06, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 341_UCR_id_39_HumanActivity_tr_48812_1st_128430.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/341_UCR_id_39_HumanActivity_tr_48812_1st_128430_results.pkl
Processing file: 486_UCR_id_184_Medical_tr_1875_1st_5607.csv, Data shape: (7501, 1)
Label shape: (7501,)
Train data shape: (1875, 1), Test data shape: (5626, 1), Label test shape: (5626,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 486_UCR_id_184_Medical_tr_1875_1st_5607.csv...
Using CUDA device: 2
Data shape: (5626, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5626, window_size 5000, adding 4374 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.59it/s]
Original shapes - Output: (10000,), Label: (5626,)
Aligned shapes - Output: (5626,), Label: (5626,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 690.01it/s]
Table1-only metrics completed in 0.87s
Evaluation Result:  {'AUC-PR': 0.9468230465813473, 'AUC-ROC': 0.9988291560000794, 'VUS-PR': 0.927409977884103, 'VUS-ROC': 0.9996165728985014, 'Standard-F1': 0.6190476190471681, 'Standard-Precision': 0.45614035087718496, 'Standard-Recall': 0.9629629629629274, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9670367070469184, 'Affiliation-P': 0.9361829927553109, 'Affiliation-R': 0.9999934168082052, 'F1_T': 0.6190476190471681, 'Precision_T': 0.45614035087718496, 'Recall_T': 0.9629629629629274}
Results for 486_UCR_id_184_Medical_tr_1875_1st_5607.csv: {'filename': '486_UCR_id_184_Medical_tr_1875_1st_5607.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 154, 'train_index': '1875', 'data_shape': '7501x1', 'output_length': 10000, 'label_length': 5626, 'aligned_length': 5626, 'AUC-PR': 0.9468230465813473, 'AUC-ROC': 0.9988291560000794, 'VUS-PR': 0.927409977884103, 'VUS-ROC': 0.9996165728985014, 'Standard-F1': 0.6190476190471681, 'Standard-Precision': 0.45614035087718496, 'Standard-Recall': 0.9629629629629274, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9670367070469184, 'Affiliation-P': 0.9361829927553109, 'Affiliation-R': 0.9999934168082052, 'F1_T': 0.6190476190471681, 'Precision_T': 0.45614035087718496, 'Recall_T': 0.9629629629629274}
Logits metrics skipped for 486_UCR_id_184_Medical_tr_1875_1st_5607.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/486_UCR_id_184_Medical_tr_1875_1st_5607_results.pkl
Processing file: 070_WSD_id_42_WebService_tr_2102_1st_2202.csv, Data shape: (11686, 1)
Label shape: (11686,)
Train data shape: (2102, 1), Test data shape: (9584, 1), Label test shape: (9584,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 070_WSD_id_42_WebService_tr_2102_1st_2202.csv...
Using CUDA device: 2
Data shape: (9584, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9584, window_size 5000, adding 416 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 164.83it/s]
Original shapes - Output: (10000,), Label: (9584,)
Aligned shapes - Output: (9584,), Label: (9584,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 320.61it/s]
Table1-only metrics completed in 2.52s
Evaluation Result:  {'AUC-PR': 0.1835689349178055, 'AUC-ROC': 0.7243409531292658, 'VUS-PR': 0.14116583261289625, 'VUS-ROC': 0.9355220759237522, 'Standard-F1': 0.14857142857093164, 'Standard-Precision': 0.13541666666666527, 'Standard-Recall': 0.1645569620253144, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9383151295976013, 'Affiliation-P': 0.8884862245120727, 'Affiliation-R': 0.9940652130345474, 'F1_T': 0.14857142857093164, 'Precision_T': 0.13541666666666527, 'Recall_T': 0.1645569620253144}
Results for 070_WSD_id_42_WebService_tr_2102_1st_2202.csv: {'filename': '070_WSD_id_42_WebService_tr_2102_1st_2202.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '2102', 'data_shape': '11686x1', 'output_length': 10000, 'label_length': 9584, 'aligned_length': 9584, 'AUC-PR': 0.1835689349178055, 'AUC-ROC': 0.7243409531292658, 'VUS-PR': 0.14116583261289625, 'VUS-ROC': 0.9355220759237522, 'Standard-F1': 0.14857142857093164, 'Standard-Precision': 0.13541666666666527, 'Standard-Recall': 0.1645569620253144, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9383151295976013, 'Affiliation-P': 0.8884862245120727, 'Affiliation-R': 0.9940652130345474, 'F1_T': 0.14857142857093164, 'Precision_T': 0.13541666666666527, 'Recall_T': 0.1645569620253144}
Logits metrics skipped for 070_WSD_id_42_WebService_tr_2102_1st_2202.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/070_WSD_id_42_WebService_tr_2102_1st_2202_results.pkl
Processing file: 700_YAHOO_id_150_WebService_tr_500_1st_82.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 700_YAHOO_id_150_WebService_tr_500_1st_82.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 160.40it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 629.20it/s]
Table1-only metrics completed in 0.89s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.977955362133306, 'VUS-ROC': 0.999858519900786, 'Standard-F1': 0.9565217391298525, 'Standard-Precision': 0.9166666666665902, 'Standard-Recall': 0.9999999999999091, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9939024390243898, 'Affiliation-P': 0.9878787878787879, 'Affiliation-R': 1.0, 'F1_T': 0.9565217391298525, 'Precision_T': 0.9166666666665902, 'Recall_T': 0.9999999999999091}
Results for 700_YAHOO_id_150_WebService_tr_500_1st_82.csv: {'filename': '700_YAHOO_id_150_WebService_tr_500_1st_82.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.977955362133306, 'VUS-ROC': 0.999858519900786, 'Standard-F1': 0.9565217391298525, 'Standard-Precision': 0.9166666666665902, 'Standard-Recall': 0.9999999999999091, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9939024390243898, 'Affiliation-P': 0.9878787878787879, 'Affiliation-R': 1.0, 'F1_T': 0.9565217391298525, 'Precision_T': 0.9166666666665902, 'Recall_T': 0.9999999999999091}
Logits metrics skipped for 700_YAHOO_id_150_WebService_tr_500_1st_82.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/700_YAHOO_id_150_WebService_tr_500_1st_82_results.pkl
Skipping file: 841_Exathlon_id_32_Facility_tr_11663_1st_13482.csv due to filter criteria.
Skipping file: 206_SMD_id_29_Facility_tr_6250_1st_21230.csv due to filter criteria.
Processing file: 378_UCR_id_76_Medical_tr_9594_1st_11111.csv, Data shape: (38379, 1)
Label shape: (38379,)
Train data shape: (9594, 1), Test data shape: (28785, 1), Label test shape: (28785,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 378_UCR_id_76_Medical_tr_9594_1st_11111.csv...
Using CUDA device: 2
Data shape: (28785, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 28785, window_size 5000, adding 1215 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 140.09it/s]
Original shapes - Output: (30000,), Label: (28785,)
Aligned shapes - Output: (28785,), Label: (28785,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 135/200 [00:01<00:00, 134.92it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 148.46it/s]
Table1-only metrics completed in 1.67s
Evaluation Result:  {'AUC-PR': 0.9875925353129187, 'AUC-ROC': 0.9997587589332404, 'VUS-PR': 0.848950085142614, 'VUS-ROC': 0.9994701063273489, 'Standard-F1': 0.5103092783501302, 'Standard-Precision': 0.3437499999999988, 'Standard-Recall': 0.9899999999999901, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9500605998258406, 'Affiliation-P': 0.9048722694411002, 'Affiliation-R': 0.9999994788952579, 'F1_T': 0.5103092783501302, 'Precision_T': 0.3437499999999988, 'Recall_T': 0.9899999999999901}
Results for 378_UCR_id_76_Medical_tr_9594_1st_11111.csv: {'filename': '378_UCR_id_76_Medical_tr_9594_1st_11111.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 40, 'train_index': '9594', 'data_shape': '38379x1', 'output_length': 30000, 'label_length': 28785, 'aligned_length': 28785, 'AUC-PR': 0.9875925353129187, 'AUC-ROC': 0.9997587589332404, 'VUS-PR': 0.848950085142614, 'VUS-ROC': 0.9994701063273489, 'Standard-F1': 0.5103092783501302, 'Standard-Precision': 0.3437499999999988, 'Standard-Recall': 0.9899999999999901, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9500605998258406, 'Affiliation-P': 0.9048722694411002, 'Affiliation-R': 0.9999994788952579, 'F1_T': 0.5103092783501302, 'Precision_T': 0.3437499999999988, 'Recall_T': 0.9899999999999901}
Logits metrics skipped for 378_UCR_id_76_Medical_tr_9594_1st_11111.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/378_UCR_id_76_Medical_tr_9594_1st_11111_results.pkl
Processing file: 554_YAHOO_id_4_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 554_YAHOO_id_4_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 160.90it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1777.70it/s]
Table1-only metrics completed in 0.97s
Evaluation Result:  {'AUC-PR': 0.8875, 'AUC-ROC': 0.9994547437295529, 'VUS-PR': 0.992632013705543, 'VUS-ROC': 0.9999757092987542, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 554_YAHOO_id_4_Synthetic_tr_500_1st_658.csv: {'filename': '554_YAHOO_id_4_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 157, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.8875, 'AUC-ROC': 0.9994547437295529, 'VUS-PR': 0.992632013705543, 'VUS-ROC': 0.9999757092987542, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 554_YAHOO_id_4_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/554_YAHOO_id_4_Synthetic_tr_500_1st_658_results.pkl
Processing file: 057_WSD_id_29_WebService_tr_2229_1st_2329.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (2229, 1), Test data shape: (16007, 1), Label test shape: (16007,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 057_WSD_id_29_WebService_tr_2229_1st_2329.csv...
Using CUDA device: 2
Data shape: (16007, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 16007, window_size 5000, adding 3993 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.09it/s]
Original shapes - Output: (20000,), Label: (16007,)
Aligned shapes - Output: (16007,), Label: (16007,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 269.45it/s]
Table1-only metrics completed in 0.85s
Evaluation Result:  {'AUC-PR': 0.17527295426460288, 'AUC-ROC': 0.7969514116392798, 'VUS-PR': 0.06537302392500946, 'VUS-ROC': 0.8275480081201554, 'Standard-F1': 0.07368421052605635, 'Standard-Precision': 0.043478260869564946, 'Standard-Recall': 0.24137931034481927, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7416780765989034, 'Affiliation-P': 0.5895453146655746, 'Affiliation-R': 0.9996348580254759, 'F1_T': 0.07368421052605635, 'Precision_T': 0.043478260869564946, 'Recall_T': 0.24137931034481927}
Results for 057_WSD_id_29_WebService_tr_2229_1st_2329.csv: {'filename': '057_WSD_id_29_WebService_tr_2229_1st_2329.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 9, 'train_index': '2229', 'data_shape': '18236x1', 'output_length': 20000, 'label_length': 16007, 'aligned_length': 16007, 'AUC-PR': 0.17527295426460288, 'AUC-ROC': 0.7969514116392798, 'VUS-PR': 0.06537302392500946, 'VUS-ROC': 0.8275480081201554, 'Standard-F1': 0.07368421052605635, 'Standard-Precision': 0.043478260869564946, 'Standard-Recall': 0.24137931034481927, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7416780765989034, 'Affiliation-P': 0.5895453146655746, 'Affiliation-R': 0.9996348580254759, 'F1_T': 0.07368421052605635, 'Precision_T': 0.043478260869564946, 'Recall_T': 0.24137931034481927}
Logits metrics skipped for 057_WSD_id_29_WebService_tr_2229_1st_2329.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/057_WSD_id_29_WebService_tr_2229_1st_2329_results.pkl
Processing file: 116_WSD_id_88_WebService_tr_3934_1st_13861.csv, Data shape: (15736, 1)
Label shape: (15736,)
Train data shape: (3934, 1), Test data shape: (11802, 1), Label test shape: (11802,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 116_WSD_id_88_WebService_tr_3934_1st_13861.csv...
Using CUDA device: 2
Data shape: (11802, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 11802, window_size 5000, adding 3198 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.68it/s]
Original shapes - Output: (15000,), Label: (11802,)
Aligned shapes - Output: (11802,), Label: (11802,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 347.91it/s]
Table1-only metrics completed in 1.11s
Evaluation Result:  {'AUC-PR': 0.6567460317460317, 'AUC-ROC': 0.9990336526235484, 'VUS-PR': 0.3134368048048491, 'VUS-ROC': 0.998582510888358, 'Standard-F1': 0.08064516129024389, 'Standard-Precision': 0.042016806722688725, 'Standard-Recall': 0.9999999999997999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666303347817328, 'Affiliation-P': 0.49995912774317464, 'Affiliation-R': 1.0, 'F1_T': 0.08064516129024389, 'Precision_T': 0.042016806722688725, 'Recall_T': 0.9999999999997999}
Results for 116_WSD_id_88_WebService_tr_3934_1st_13861.csv: {'filename': '116_WSD_id_88_WebService_tr_3934_1st_13861.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '3934', 'data_shape': '15736x1', 'output_length': 15000, 'label_length': 11802, 'aligned_length': 11802, 'AUC-PR': 0.6567460317460317, 'AUC-ROC': 0.9990336526235484, 'VUS-PR': 0.3134368048048491, 'VUS-ROC': 0.998582510888358, 'Standard-F1': 0.08064516129024389, 'Standard-Precision': 0.042016806722688725, 'Standard-Recall': 0.9999999999997999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666303347817328, 'Affiliation-P': 0.49995912774317464, 'Affiliation-R': 1.0, 'F1_T': 0.08064516129024389, 'Precision_T': 0.042016806722688725, 'Recall_T': 0.9999999999997999}
Logits metrics skipped for 116_WSD_id_88_WebService_tr_3934_1st_13861.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/116_WSD_id_88_WebService_tr_3934_1st_13861_results.pkl
Processing file: 515_UCR_id_213_Medical_tr_10499_1st_15000.csv, Data shape: (41999, 1)
Label shape: (41999,)
Train data shape: (10499, 1), Test data shape: (31500, 1), Label test shape: (31500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 515_UCR_id_213_Medical_tr_10499_1st_15000.csv...
Using CUDA device: 2
Data shape: (31500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 31500, window_size 5000, adding 3500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([7, 5000, 1])
batched_mask shape: torch.Size([7, 5000])

100%|██████████| 1/1 [00:00<00:00, 132.84it/s]
Original shapes - Output: (35000,), Label: (31500,)
Aligned shapes - Output: (31500,), Label: (31500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  53%|█████▎    | 106/200 [00:01<00:00, 105.14it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 124.51it/s]
Table1-only metrics completed in 1.92s
Evaluation Result:  {'AUC-PR': 0.3067347554060075, 'AUC-ROC': 0.9961433571201309, 'VUS-PR': 0.2546521108623803, 'VUS-ROC': 0.9963768619386525, 'Standard-F1': 0.36363636363606416, 'Standard-Precision': 0.2222222222222215, 'Standard-Recall': 0.9999999999999858, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6843385040685759, 'Affiliation-P': 0.520147854280784, 'Affiliation-R': 1.0, 'F1_T': 0.36363636363606416, 'Precision_T': 0.2222222222222215, 'Recall_T': 0.9999999999999858}
Results for 515_UCR_id_213_Medical_tr_10499_1st_15000.csv: {'filename': '515_UCR_id_213_Medical_tr_10499_1st_15000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 33, 'train_index': '10499', 'data_shape': '41999x1', 'output_length': 35000, 'label_length': 31500, 'aligned_length': 31500, 'AUC-PR': 0.3067347554060075, 'AUC-ROC': 0.9961433571201309, 'VUS-PR': 0.2546521108623803, 'VUS-ROC': 0.9963768619386525, 'Standard-F1': 0.36363636363606416, 'Standard-Precision': 0.2222222222222215, 'Standard-Recall': 0.9999999999999858, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6843385040685759, 'Affiliation-P': 0.520147854280784, 'Affiliation-R': 1.0, 'F1_T': 0.36363636363606416, 'Precision_T': 0.2222222222222215, 'Recall_T': 0.9999999999999858}
Logits metrics skipped for 515_UCR_id_213_Medical_tr_10499_1st_15000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/515_UCR_id_213_Medical_tr_10499_1st_15000_results.pkl
Processing file: 424_UCR_id_122_Medical_tr_7464_1st_12190.csv, Data shape: (29859, 1)
Label shape: (29859,)
Train data shape: (7464, 1), Test data shape: (22395, 1), Label test shape: (22395,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 424_UCR_id_122_Medical_tr_7464_1st_12190.csv...
Using CUDA device: 2
Data shape: (22395, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22395, window_size 5000, adding 2605 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 150.19it/s]
Original shapes - Output: (25000,), Label: (22395,)
Aligned shapes - Output: (22395,), Label: (22395,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  55%|█████▌    | 110/200 [00:01<00:00, 109.94it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 122.69it/s]
Table1-only metrics completed in 1.88s
Evaluation Result:  {'AUC-PR': 0.44065155579021903, 'AUC-ROC': 0.9905840582979433, 'VUS-PR': 0.46749604160843566, 'VUS-ROC': 0.9912029441226762, 'Standard-F1': 0.6051660516600257, 'Standard-Precision': 0.5256410256410239, 'Standard-Recall': 0.7130434782608664, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8345050994995344, 'Affiliation-P': 0.7162864261341522, 'Affiliation-R': 0.9994597978974344, 'F1_T': 0.6053097345127892, 'Precision_T': 0.5104477611940282, 'Recall_T': 0.743478260869562}
Results for 424_UCR_id_122_Medical_tr_7464_1st_12190.csv: {'filename': '424_UCR_id_122_Medical_tr_7464_1st_12190.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 33, 'train_index': '7464', 'data_shape': '29859x1', 'output_length': 25000, 'label_length': 22395, 'aligned_length': 22395, 'AUC-PR': 0.44065155579021903, 'AUC-ROC': 0.9905840582979433, 'VUS-PR': 0.46749604160843566, 'VUS-ROC': 0.9912029441226762, 'Standard-F1': 0.6051660516600257, 'Standard-Precision': 0.5256410256410239, 'Standard-Recall': 0.7130434782608664, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8345050994995344, 'Affiliation-P': 0.7162864261341522, 'Affiliation-R': 0.9994597978974344, 'F1_T': 0.6053097345127892, 'Precision_T': 0.5104477611940282, 'Recall_T': 0.743478260869562}
Logits metrics skipped for 424_UCR_id_122_Medical_tr_7464_1st_12190.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/424_UCR_id_122_Medical_tr_7464_1st_12190_results.pkl
Processing file: 313_UCR_id_11_Sensor_tr_2827_1st_4711.csv, Data shape: (11308, 1)
Label shape: (11308,)
Train data shape: (2827, 1), Test data shape: (8481, 1), Label test shape: (8481,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 313_UCR_id_11_Sensor_tr_2827_1st_4711.csv...
Using CUDA device: 2
Data shape: (8481, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 8481, window_size 5000, adding 1519 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.66it/s]
Original shapes - Output: (10000,), Label: (8481,)
Aligned shapes - Output: (8481,), Label: (8481,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 437.87it/s]
Table1-only metrics completed in 0.86s
Evaluation Result:  {'AUC-PR': 0.9664055969720364, 'AUC-ROC': 0.9993938169327138, 'VUS-PR': 0.9333408847204642, 'VUS-ROC': 0.9992529111899846, 'Standard-F1': 0.9289617486333722, 'Standard-Precision': 0.9999999999999883, 'Standard-Recall': 0.8673469387755014, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996434339030261, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9992871219941815, 'F1_T': 0.9289617486333722, 'Precision_T': 0.9999999999999883, 'Recall_T': 0.8673469387755014}
Results for 313_UCR_id_11_Sensor_tr_2827_1st_4711.csv: {'filename': '313_UCR_id_11_Sensor_tr_2827_1st_4711.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 98, 'train_index': '2827', 'data_shape': '11308x1', 'output_length': 10000, 'label_length': 8481, 'aligned_length': 8481, 'AUC-PR': 0.9664055969720364, 'AUC-ROC': 0.9993938169327138, 'VUS-PR': 0.9333408847204642, 'VUS-ROC': 0.9992529111899846, 'Standard-F1': 0.9289617486333722, 'Standard-Precision': 0.9999999999999883, 'Standard-Recall': 0.8673469387755014, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996434339030261, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9992871219941815, 'F1_T': 0.9289617486333722, 'Precision_T': 0.9999999999999883, 'Recall_T': 0.8673469387755014}
Logits metrics skipped for 313_UCR_id_11_Sensor_tr_2827_1st_4711.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/313_UCR_id_11_Sensor_tr_2827_1st_4711_results.pkl
Processing file: 051_WSD_id_23_WebService_tr_3919_1st_5154.csv, Data shape: (15676, 1)
Label shape: (15676,)
Train data shape: (3919, 1), Test data shape: (11757, 1), Label test shape: (11757,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 051_WSD_id_23_WebService_tr_3919_1st_5154.csv...
Using CUDA device: 2
Data shape: (11757, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 11757, window_size 5000, adding 3243 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 164.31it/s]
Original shapes - Output: (15000,), Label: (11757,)
Aligned shapes - Output: (11757,), Label: (11757,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 268.42it/s]
Table1-only metrics completed in 1.12s
Evaluation Result:  {'AUC-PR': 0.46284120532844086, 'AUC-ROC': 0.8627311846342367, 'VUS-PR': 0.3096584505558355, 'VUS-ROC': 0.9359952972934975, 'Standard-F1': 0.3146067415725833, 'Standard-Precision': 0.23728813559321835, 'Standard-Recall': 0.46666666666665885, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.891335171096, 'Affiliation-P': 0.8047699460697222, 'Affiliation-R': 0.9987678116774519, 'F1_T': 0.3146067415725833, 'Precision_T': 0.23728813559321835, 'Recall_T': 0.46666666666665885}
Results for 051_WSD_id_23_WebService_tr_3919_1st_5154.csv: {'filename': '051_WSD_id_23_WebService_tr_3919_1st_5154.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '3919', 'data_shape': '15676x1', 'output_length': 15000, 'label_length': 11757, 'aligned_length': 11757, 'AUC-PR': 0.46284120532844086, 'AUC-ROC': 0.8627311846342367, 'VUS-PR': 0.3096584505558355, 'VUS-ROC': 0.9359952972934975, 'Standard-F1': 0.3146067415725833, 'Standard-Precision': 0.23728813559321835, 'Standard-Recall': 0.46666666666665885, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.891335171096, 'Affiliation-P': 0.8047699460697222, 'Affiliation-R': 0.9987678116774519, 'F1_T': 0.3146067415725833, 'Precision_T': 0.23728813559321835, 'Recall_T': 0.46666666666665885}
Logits metrics skipped for 051_WSD_id_23_WebService_tr_3919_1st_5154.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/051_WSD_id_23_WebService_tr_3919_1st_5154_results.pkl
Processing file: 267_IOPS_id_8_WebService_tr_1534_1st_1882.csv, Data shape: (6138, 1)
Label shape: (6138,)
Train data shape: (1534, 1), Test data shape: (4604, 1), Label test shape: (4604,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 267_IOPS_id_8_WebService_tr_1534_1st_1882.csv...
Using CUDA device: 2
Data shape: (4604, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 4604, 1])
batched_mask shape: torch.Size([1, 4604])

100%|██████████| 1/1 [00:00<00:00, 166.07it/s]
Original shapes - Output: (4604,), Label: (4604,)
Aligned shapes - Output: (4604,), Label: (4604,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 689.84it/s]
Table1-only metrics completed in 3.86s
Evaluation Result:  {'AUC-PR': 0.4031089192558346, 'AUC-ROC': 0.9501830408411172, 'VUS-PR': 0.6736941359370042, 'VUS-ROC': 0.9692216372198583, 'Standard-F1': 0.39370078740108233, 'Standard-Precision': 0.3378378378378333, 'Standard-Recall': 0.47169811320753824, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.928044846326477, 'Affiliation-P': 0.9016151054301537, 'Affiliation-R': 0.9560708918050175, 'F1_T': 0.39344262295032184, 'Precision_T': 0.34782608695651673, 'Recall_T': 0.45283018867923674}
Results for 267_IOPS_id_8_WebService_tr_1534_1st_1882.csv: {'filename': '267_IOPS_id_8_WebService_tr_1534_1st_1882.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 282, 'train_index': '1534', 'data_shape': '6138x1', 'output_length': 4604, 'label_length': 4604, 'aligned_length': 4604, 'AUC-PR': 0.4031089192558346, 'AUC-ROC': 0.9501830408411172, 'VUS-PR': 0.6736941359370042, 'VUS-ROC': 0.9692216372198583, 'Standard-F1': 0.39370078740108233, 'Standard-Precision': 0.3378378378378333, 'Standard-Recall': 0.47169811320753824, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.928044846326477, 'Affiliation-P': 0.9016151054301537, 'Affiliation-R': 0.9560708918050175, 'F1_T': 0.39344262295032184, 'Precision_T': 0.34782608695651673, 'Recall_T': 0.45283018867923674}
Logits metrics skipped for 267_IOPS_id_8_WebService_tr_1534_1st_1882.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/267_IOPS_id_8_WebService_tr_1534_1st_1882_results.pkl
Processing file: 621_YAHOO_id_71_WebService_tr_500_1st_721.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 621_YAHOO_id_71_WebService_tr_500_1st_721.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 160.03it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1349.94it/s]
Table1-only metrics completed in 0.30s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.6791147355915822, 'VUS-ROC': 0.999318764404511, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9778043632988299, 'Affiliation-P': 0.9565726248396065, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Results for 621_YAHOO_id_71_WebService_tr_500_1st_721.csv: {'filename': '621_YAHOO_id_71_WebService_tr_500_1st_721.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.6791147355915822, 'VUS-ROC': 0.999318764404511, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9778043632988299, 'Affiliation-P': 0.9565726248396065, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Logits metrics skipped for 621_YAHOO_id_71_WebService_tr_500_1st_721.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/621_YAHOO_id_71_WebService_tr_500_1st_721_results.pkl
Processing file: 720_YAHOO_id_170_WebService_tr_500_1st_3.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 720_YAHOO_id_170_WebService_tr_500_1st_3.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.97it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 825.08it/s]
Table1-only metrics completed in 0.83s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9928762646208401, 'VUS-ROC': 0.9999542400458173, 'Standard-F1': 0.9999999999994167, 'Standard-Precision': 0.9999999999999166, 'Standard-Recall': 0.9999999999999166, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.9999999999994167, 'Precision_T': 0.9999999999999166, 'Recall_T': 0.9999999999999166}
Results for 720_YAHOO_id_170_WebService_tr_500_1st_3.csv: {'filename': '720_YAHOO_id_170_WebService_tr_500_1st_3.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9928762646208401, 'VUS-ROC': 0.9999542400458173, 'Standard-F1': 0.9999999999994167, 'Standard-Precision': 0.9999999999999166, 'Standard-Recall': 0.9999999999999166, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.9999999999994167, 'Precision_T': 0.9999999999999166, 'Recall_T': 0.9999999999999166}
Logits metrics skipped for 720_YAHOO_id_170_WebService_tr_500_1st_3.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/720_YAHOO_id_170_WebService_tr_500_1st_3_results.pkl
Processing file: 387_UCR_id_85_Environment_tr_2046_1st_5549.csv, Data shape: (8184, 1)
Label shape: (8184,)
Train data shape: (2046, 1), Test data shape: (6138, 1), Label test shape: (6138,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 387_UCR_id_85_Environment_tr_2046_1st_5549.csv...
Using CUDA device: 2
Data shape: (6138, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6138, window_size 5000, adding 3862 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 167.06it/s]
Original shapes - Output: (10000,), Label: (6138,)
Aligned shapes - Output: (6138,), Label: (6138,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 488.10it/s]
Table1-only metrics completed in 0.53s
Evaluation Result:  {'AUC-PR': 0.8775498014527123, 'AUC-ROC': 0.9984024356869186, 'VUS-PR': 0.8450104599745977, 'VUS-ROC': 0.998642383402066, 'Standard-F1': 0.7272727272722221, 'Standard-Precision': 0.6451612903225702, 'Standard-Recall': 0.8333333333333159, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9994850894598114, 'Affiliation-P': 0.9997424820525757, 'Affiliation-R': 0.9992278293689584, 'F1_T': 0.7272727272722221, 'Precision_T': 0.6451612903225702, 'Recall_T': 0.8333333333333159}
Results for 387_UCR_id_85_Environment_tr_2046_1st_5549.csv: {'filename': '387_UCR_id_85_Environment_tr_2046_1st_5549.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '2046', 'data_shape': '8184x1', 'output_length': 10000, 'label_length': 6138, 'aligned_length': 6138, 'AUC-PR': 0.8775498014527123, 'AUC-ROC': 0.9984024356869186, 'VUS-PR': 0.8450104599745977, 'VUS-ROC': 0.998642383402066, 'Standard-F1': 0.7272727272722221, 'Standard-Precision': 0.6451612903225702, 'Standard-Recall': 0.8333333333333159, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9994850894598114, 'Affiliation-P': 0.9997424820525757, 'Affiliation-R': 0.9992278293689584, 'F1_T': 0.7272727272722221, 'Precision_T': 0.6451612903225702, 'Recall_T': 0.8333333333333159}
Logits metrics skipped for 387_UCR_id_85_Environment_tr_2046_1st_5549.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/387_UCR_id_85_Environment_tr_2046_1st_5549_results.pkl
Processing file: 616_YAHOO_id_66_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 616_YAHOO_id_66_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 159.84it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1959.46it/s]
Table1-only metrics completed in 1.22s
Evaluation Result:  {'AUC-PR': 0.8741943241943242, 'AUC-ROC': 0.9986598440545809, 'VUS-PR': 0.996865867206443, 'VUS-ROC': 0.999971294636075, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Results for 616_YAHOO_id_66_Synthetic_tr_500_1st_623.csv: {'filename': '616_YAHOO_id_66_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 235, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.8741943241943242, 'AUC-ROC': 0.9986598440545809, 'VUS-PR': 0.996865867206443, 'VUS-ROC': 0.999971294636075, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 616_YAHOO_id_66_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/616_YAHOO_id_66_Synthetic_tr_500_1st_623_results.pkl
Processing file: 580_YAHOO_id_30_WebService_tr_500_1st_236.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 580_YAHOO_id_30_WebService_tr_500_1st_236.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 166.94it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1910.55it/s]
Table1-only metrics completed in 0.21s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9737705116048722, 'Affiliation-P': 0.9488818267419963, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Results for 580_YAHOO_id_30_WebService_tr_500_1st_236.csv: {'filename': '580_YAHOO_id_30_WebService_tr_500_1st_236.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9737705116048722, 'Affiliation-P': 0.9488818267419963, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 580_YAHOO_id_30_WebService_tr_500_1st_236.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/580_YAHOO_id_30_WebService_tr_500_1st_236_results.pkl
Processing file: 442_UCR_id_140_Medical_tr_1875_1st_4187.csv, Data shape: (7501, 1)
Label shape: (7501,)
Train data shape: (1875, 1), Test data shape: (5626, 1), Label test shape: (5626,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 442_UCR_id_140_Medical_tr_1875_1st_4187.csv...
Using CUDA device: 2
Data shape: (5626, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5626, window_size 5000, adding 4374 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.14it/s]
Original shapes - Output: (10000,), Label: (5626,)
Aligned shapes - Output: (5626,), Label: (5626,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 658.64it/s]
Table1-only metrics completed in 0.99s
Evaluation Result:  {'AUC-PR': 0.11921224923065409, 'AUC-ROC': 0.9650427502671892, 'VUS-PR': 0.4987196987027938, 'VUS-ROC': 0.9933790731252187, 'Standard-F1': 0.07843137254881047, 'Standard-Precision': 0.044444444444443953, 'Standard-Recall': 0.33333333333330556, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9911983413766178, 'Affiliation-P': 0.9839844564031529, 'Affiliation-R': 0.9985187818461903, 'F1_T': 0.06451612903208222, 'Precision_T': 0.0357142857142854, 'Recall_T': 0.33333333333330556}
Results for 442_UCR_id_140_Medical_tr_1875_1st_4187.csv: {'filename': '442_UCR_id_140_Medical_tr_1875_1st_4187.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 183, 'train_index': '1875', 'data_shape': '7501x1', 'output_length': 10000, 'label_length': 5626, 'aligned_length': 5626, 'AUC-PR': 0.11921224923065409, 'AUC-ROC': 0.9650427502671892, 'VUS-PR': 0.4987196987027938, 'VUS-ROC': 0.9933790731252187, 'Standard-F1': 0.07843137254881047, 'Standard-Precision': 0.044444444444443953, 'Standard-Recall': 0.33333333333330556, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9911983413766178, 'Affiliation-P': 0.9839844564031529, 'Affiliation-R': 0.9985187818461903, 'F1_T': 0.06451612903208222, 'Precision_T': 0.0357142857142854, 'Recall_T': 0.33333333333330556}
Logits metrics skipped for 442_UCR_id_140_Medical_tr_1875_1st_4187.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/442_UCR_id_140_Medical_tr_1875_1st_4187_results.pkl
Skipping file: 224_LTDB_id_9_Medical_tr_4456_1st_4556.csv due to filter criteria.
Processing file: 081_WSD_id_53_WebService_tr_4559_1st_8448.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 081_WSD_id_53_WebService_tr_4559_1st_8448.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 164.90it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 290.06it/s]
Table1-only metrics completed in 1.62s
Evaluation Result:  {'AUC-PR': 0.5497691072924706, 'AUC-ROC': 0.9508659657732312, 'VUS-PR': 0.6097390366245151, 'VUS-ROC': 0.9833212476573019, 'Standard-F1': 0.5923753665684307, 'Standard-Precision': 0.7372262773722574, 'Standard-Recall': 0.49509803921568385, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9204285744084415, 'Affiliation-P': 0.8610392405265767, 'Affiliation-R': 0.9886174926218126, 'F1_T': 0.5923753665684307, 'Precision_T': 0.7372262773722574, 'Recall_T': 0.49509803921568385}
Results for 081_WSD_id_53_WebService_tr_4559_1st_8448.csv: {'filename': '081_WSD_id_53_WebService_tr_4559_1st_8448.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.5497691072924706, 'AUC-ROC': 0.9508659657732312, 'VUS-PR': 0.6097390366245151, 'VUS-ROC': 0.9833212476573019, 'Standard-F1': 0.5923753665684307, 'Standard-Precision': 0.7372262773722574, 'Standard-Recall': 0.49509803921568385, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9204285744084415, 'Affiliation-P': 0.8610392405265767, 'Affiliation-R': 0.9886174926218126, 'F1_T': 0.5923753665684307, 'Precision_T': 0.7372262773722574, 'Recall_T': 0.49509803921568385}
Logits metrics skipped for 081_WSD_id_53_WebService_tr_4559_1st_8448.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/081_WSD_id_53_WebService_tr_4559_1st_8448_results.pkl
Processing file: 799_YAHOO_id_249_WebService_tr_500_1st_852.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 799_YAHOO_id_249_WebService_tr_500_1st_852.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 168.14it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1102.57it/s]
Table1-only metrics completed in 0.54s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9686253650432952, 'VUS-ROC': 0.9998662757066167, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Results for 799_YAHOO_id_249_WebService_tr_500_1st_852.csv: {'filename': '799_YAHOO_id_249_WebService_tr_500_1st_852.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9686253650432952, 'VUS-ROC': 0.9998662757066167, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Logits metrics skipped for 799_YAHOO_id_249_WebService_tr_500_1st_852.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/799_YAHOO_id_249_WebService_tr_500_1st_852_results.pkl
Processing file: 291_TODS_id_5_Synthetic_tr_500_1st_11.csv, Data shape: (5000, 1)
Label shape: (5000,)
Train data shape: (500, 1), Test data shape: (4500, 1), Label test shape: (4500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 291_TODS_id_5_Synthetic_tr_500_1st_11.csv...
Using CUDA device: 2
Data shape: (4500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 4500, 1])
batched_mask shape: torch.Size([1, 4500])

100%|██████████| 1/1 [00:00<00:00, 162.21it/s]
Original shapes - Output: (4500,), Label: (4500,)
Aligned shapes - Output: (4500,), Label: (4500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  14%|█▍        | 28/200 [00:01<00:06, 27.02it/s]

Calculating Affiliation Metrics:  28%|██▊       | 56/200 [00:02<00:07, 18.17it/s]

Calculating Affiliation Metrics:  38%|███▊      | 76/200 [00:04<00:07, 15.51it/s]

Calculating Affiliation Metrics:  46%|████▋     | 93/200 [00:05<00:07, 14.44it/s]

Calculating Affiliation Metrics:  54%|█████▍    | 108/200 [00:07<00:06, 14.15it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 123/200 [00:08<00:05, 14.27it/s]

Calculating Affiliation Metrics:  70%|███████   | 141/200 [00:09<00:03, 15.22it/s]

Calculating Affiliation Metrics:  82%|████████▏ | 163/200 [00:10<00:02, 17.10it/s]

Calculating Affiliation Metrics:  96%|█████████▋| 193/200 [00:11<00:00, 20.76it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:11<00:00, 17.91it/s]
Table1-only metrics completed in 19.04s
Evaluation Result:  {'AUC-PR': 0.5717415079360385, 'AUC-ROC': 0.8212117432840288, 'VUS-PR': 0.9502119572029407, 'VUS-ROC': 0.9767370828533808, 'Standard-F1': 0.5978835978831243, 'Standard-Precision': 0.7902097902097848, 'Standard-Recall': 0.4808510638297852, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.756722686853597, 'Affiliation-P': 0.7278448630976857, 'Affiliation-R': 0.7879866852298832, 'F1_T': 0.5962059620591548, 'Precision_T': 0.8208955223880536, 'Recall_T': 0.46808510638297673}
Results for 291_TODS_id_5_Synthetic_tr_500_1st_11.csv: {'filename': '291_TODS_id_5_Synthetic_tr_500_1st_11.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '500', 'data_shape': '5000x1', 'output_length': 4500, 'label_length': 4500, 'aligned_length': 4500, 'AUC-PR': 0.5717415079360385, 'AUC-ROC': 0.8212117432840288, 'VUS-PR': 0.9502119572029407, 'VUS-ROC': 0.9767370828533808, 'Standard-F1': 0.5978835978831243, 'Standard-Precision': 0.7902097902097848, 'Standard-Recall': 0.4808510638297852, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.756722686853597, 'Affiliation-P': 0.7278448630976857, 'Affiliation-R': 0.7879866852298832, 'F1_T': 0.5962059620591548, 'Precision_T': 0.8208955223880536, 'Recall_T': 0.46808510638297673}
Logits metrics skipped for 291_TODS_id_5_Synthetic_tr_500_1st_11.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/291_TODS_id_5_Synthetic_tr_500_1st_11_results.pkl
Processing file: 027_NAB_id_27_Facility_tr_757_1st_2582.csv, Data shape: (3031, 1)
Label shape: (3031,)
Train data shape: (757, 1), Test data shape: (2274, 1), Label test shape: (2274,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 027_NAB_id_27_Facility_tr_757_1st_2582.csv...
Using CUDA device: 2
Data shape: (2274, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 2274, 1])
batched_mask shape: torch.Size([1, 2274])

100%|██████████| 1/1 [00:00<00:00, 147.89it/s]
Original shapes - Output: (2274,), Label: (2274,)
Aligned shapes - Output: (2274,), Label: (2274,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1157.29it/s]
Table1-only metrics completed in 0.63s
Evaluation Result:  {'AUC-PR': 0.1685352174308083, 'AUC-ROC': 0.5227844744455159, 'VUS-PR': 0.16973416564309435, 'VUS-ROC': 0.5602961557349269, 'Standard-F1': 0.21445221445171575, 'Standard-Precision': 0.20087336244541398, 'Standard-Recall': 0.22999999999999884, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8750424674418112, 'Affiliation-P': 0.8042729404866608, 'Affiliation-R': 0.9594678979771328, 'F1_T': 0.21176470588185367, 'Precision_T': 0.19999999999999912, 'Recall_T': 0.22499999999999887}
Results for 027_NAB_id_27_Facility_tr_757_1st_2582.csv: {'filename': '027_NAB_id_27_Facility_tr_757_1st_2582.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '757', 'data_shape': '3031x1', 'output_length': 2274, 'label_length': 2274, 'aligned_length': 2274, 'AUC-PR': 0.1685352174308083, 'AUC-ROC': 0.5227844744455159, 'VUS-PR': 0.16973416564309435, 'VUS-ROC': 0.5602961557349269, 'Standard-F1': 0.21445221445171575, 'Standard-Precision': 0.20087336244541398, 'Standard-Recall': 0.22999999999999884, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8750424674418112, 'Affiliation-P': 0.8042729404866608, 'Affiliation-R': 0.9594678979771328, 'F1_T': 0.21176470588185367, 'Precision_T': 0.19999999999999912, 'Recall_T': 0.22499999999999887}
Logits metrics skipped for 027_NAB_id_27_Facility_tr_757_1st_2582.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/027_NAB_id_27_Facility_tr_757_1st_2582_results.pkl
Processing file: 562_YAHOO_id_12_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 562_YAHOO_id_12_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 161.89it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2374.84it/s]
Table1-only metrics completed in 0.50s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.748981255093724, 'Affiliation-P': 0.5986970684039088, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 562_YAHOO_id_12_Synthetic_tr_500_1st_893.csv: {'filename': '562_YAHOO_id_12_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 126, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.748981255093724, 'Affiliation-P': 0.5986970684039088, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 562_YAHOO_id_12_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/562_YAHOO_id_12_Synthetic_tr_500_1st_893_results.pkl
Skipping file: 193_SMD_id_16_Facility_tr_5925_1st_17580.csv due to filter criteria.
Skipping file: 244_SVDB_id_8_Medical_tr_17803_1st_17903.csv due to filter criteria.
Skipping file: 216_LTDB_id_1_Medical_tr_500_1st_266.csv due to filter criteria.
Processing file: 046_WSD_id_18_WebService_tr_2357_1st_2457.csv, Data shape: (18220, 1)
Label shape: (18220,)
Train data shape: (2357, 1), Test data shape: (15863, 1), Label test shape: (15863,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 046_WSD_id_18_WebService_tr_2357_1st_2457.csv...
Using CUDA device: 2
Data shape: (15863, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 15863, window_size 5000, adding 4137 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.34it/s]
Original shapes - Output: (20000,), Label: (15863,)
Aligned shapes - Output: (15863,), Label: (15863,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  71%|███████   | 142/200 [00:01<00:00, 141.45it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 163.89it/s]
Table1-only metrics completed in 4.41s
Evaluation Result:  {'AUC-PR': 0.2613701905502321, 'AUC-ROC': 0.7523431884964968, 'VUS-PR': 0.4090179842678898, 'VUS-ROC': 0.9585770623944228, 'Standard-F1': 0.24913494809638914, 'Standard-Precision': 0.22641509433962123, 'Standard-Recall': 0.27692307692307483, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9902736779410951, 'Affiliation-P': 0.9845493173754897, 'Affiliation-R': 0.9960649928761363, 'F1_T': 0.24913494809638914, 'Precision_T': 0.22641509433962123, 'Recall_T': 0.27692307692307483}
Results for 046_WSD_id_18_WebService_tr_2357_1st_2457.csv: {'filename': '046_WSD_id_18_WebService_tr_2357_1st_2457.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '2357', 'data_shape': '18220x1', 'output_length': 20000, 'label_length': 15863, 'aligned_length': 15863, 'AUC-PR': 0.2613701905502321, 'AUC-ROC': 0.7523431884964968, 'VUS-PR': 0.4090179842678898, 'VUS-ROC': 0.9585770623944228, 'Standard-F1': 0.24913494809638914, 'Standard-Precision': 0.22641509433962123, 'Standard-Recall': 0.27692307692307483, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9902736779410951, 'Affiliation-P': 0.9845493173754897, 'Affiliation-R': 0.9960649928761363, 'F1_T': 0.24913494809638914, 'Precision_T': 0.22641509433962123, 'Recall_T': 0.27692307692307483}
Logits metrics skipped for 046_WSD_id_18_WebService_tr_2357_1st_2457.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/046_WSD_id_18_WebService_tr_2357_1st_2457_results.pkl
Skipping file: 239_SVDB_id_3_Medical_tr_50000_1st_106634.csv due to filter criteria.
Processing file: 791_YAHOO_id_241_WebService_tr_500_1st_593.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 791_YAHOO_id_241_WebService_tr_500_1st_593.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 159.30it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1325.03it/s]
Table1-only metrics completed in 0.30s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9216866243901459, 'VUS-ROC': 0.9998337154555766, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975358611596298, 'Affiliation-P': 0.9950838364288619, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Results for 791_YAHOO_id_241_WebService_tr_500_1st_593.csv: {'filename': '791_YAHOO_id_241_WebService_tr_500_1st_593.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9216866243901459, 'VUS-ROC': 0.9998337154555766, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975358611596298, 'Affiliation-P': 0.9950838364288619, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Logits metrics skipped for 791_YAHOO_id_241_WebService_tr_500_1st_593.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/791_YAHOO_id_241_WebService_tr_500_1st_593_results.pkl
Processing file: 020_NAB_id_20_Synthetic_tr_1007_1st_2679.csv, Data shape: (4031, 1)
Label shape: (4031,)
Train data shape: (1007, 1), Test data shape: (3024, 1), Label test shape: (3024,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 020_NAB_id_20_Synthetic_tr_1007_1st_2679.csv...
Using CUDA device: 2
Data shape: (3024, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 3024, 1])
batched_mask shape: torch.Size([1, 3024])

100%|██████████| 1/1 [00:00<00:00, 167.49it/s]
Original shapes - Output: (3024,), Label: (3024,)
Aligned shapes - Output: (3024,), Label: (3024,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 795.19it/s]
Table1-only metrics completed in 1.38s
Evaluation Result:  {'AUC-PR': 0.8025743041334196, 'AUC-ROC': 0.9127821302277147, 'VUS-PR': 0.8068723923219892, 'VUS-ROC': 0.9216058235607457, 'Standard-F1': 0.6843853820593, 'Standard-Precision': 0.7066895368782149, 'Standard-Recall': 0.6634460547504015, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9852716424953327, 'Affiliation-P': 0.9981507588415482, 'Affiliation-R': 0.9727206502568823, 'F1_T': 0.6832917705730653, 'Precision_T': 0.706185567010308, 'Recall_T': 0.6618357487922695}
Results for 020_NAB_id_20_Synthetic_tr_1007_1st_2679.csv: {'filename': '020_NAB_id_20_Synthetic_tr_1007_1st_2679.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 288, 'train_index': '1007', 'data_shape': '4031x1', 'output_length': 3024, 'label_length': 3024, 'aligned_length': 3024, 'AUC-PR': 0.8025743041334196, 'AUC-ROC': 0.9127821302277147, 'VUS-PR': 0.8068723923219892, 'VUS-ROC': 0.9216058235607457, 'Standard-F1': 0.6843853820593, 'Standard-Precision': 0.7066895368782149, 'Standard-Recall': 0.6634460547504015, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9852716424953327, 'Affiliation-P': 0.9981507588415482, 'Affiliation-R': 0.9727206502568823, 'F1_T': 0.6832917705730653, 'Precision_T': 0.706185567010308, 'Recall_T': 0.6618357487922695}
Logits metrics skipped for 020_NAB_id_20_Synthetic_tr_1007_1st_2679.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/020_NAB_id_20_Synthetic_tr_1007_1st_2679_results.pkl
Processing file: 410_UCR_id_108_Medical_tr_1875_1st_4187.csv, Data shape: (7500, 1)
Label shape: (7500,)
Train data shape: (1875, 1), Test data shape: (5625, 1), Label test shape: (5625,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 410_UCR_id_108_Medical_tr_1875_1st_4187.csv...
Using CUDA device: 2
Data shape: (5625, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5625, window_size 5000, adding 4375 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.20it/s]
Original shapes - Output: (10000,), Label: (5625,)
Aligned shapes - Output: (5625,), Label: (5625,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 611.68it/s]
Table1-only metrics completed in 1.01s
Evaluation Result:  {'AUC-PR': 0.867989417989418, 'AUC-ROC': 0.9995369545859305, 'VUS-PR': 0.4369034488864955, 'VUS-ROC': 0.9985660956771606, 'Standard-F1': 0.2985074626863043, 'Standard-Precision': 0.17543859649122498, 'Standard-Recall': 0.9999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6833834497052472, 'Affiliation-P': 0.5190451612903225, 'Affiliation-R': 1.0, 'F1_T': 0.2985074626863043, 'Precision_T': 0.17543859649122498, 'Recall_T': 0.9999999999999}
Results for 410_UCR_id_108_Medical_tr_1875_1st_4187.csv: {'filename': '410_UCR_id_108_Medical_tr_1875_1st_4187.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 183, 'train_index': '1875', 'data_shape': '7500x1', 'output_length': 10000, 'label_length': 5625, 'aligned_length': 5625, 'AUC-PR': 0.867989417989418, 'AUC-ROC': 0.9995369545859305, 'VUS-PR': 0.4369034488864955, 'VUS-ROC': 0.9985660956771606, 'Standard-F1': 0.2985074626863043, 'Standard-Precision': 0.17543859649122498, 'Standard-Recall': 0.9999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6833834497052472, 'Affiliation-P': 0.5190451612903225, 'Affiliation-R': 1.0, 'F1_T': 0.2985074626863043, 'Precision_T': 0.17543859649122498, 'Recall_T': 0.9999999999999}
Logits metrics skipped for 410_UCR_id_108_Medical_tr_1875_1st_4187.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/410_UCR_id_108_Medical_tr_1875_1st_4187_results.pkl
Processing file: 380_UCR_id_78_Medical_tr_1830_1st_4675.csv, Data shape: (7321, 1)
Label shape: (7321,)
Train data shape: (1830, 1), Test data shape: (5491, 1), Label test shape: (5491,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 380_UCR_id_78_Medical_tr_1830_1st_4675.csv...
Using CUDA device: 2
Data shape: (5491, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5491, window_size 5000, adding 4509 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.74it/s]
Original shapes - Output: (10000,), Label: (5491,)
Aligned shapes - Output: (5491,), Label: (5491,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 686.19it/s]
Table1-only metrics completed in 0.99s
Evaluation Result:  {'AUC-PR': 0.5313844707586088, 'AUC-ROC': 0.9290171929469409, 'VUS-PR': 0.5279442286387755, 'VUS-ROC': 0.9314784300428529, 'Standard-F1': 0.6280991735532173, 'Standard-Precision': 0.6195652173913027, 'Standard-Recall': 0.6368715083798865, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9162807806523979, 'Affiliation-P': 0.848544423551061, 'Affiliation-R': 0.9957696138628065, 'F1_T': 0.6274509803916573, 'Precision_T': 0.5896805896805882, 'Recall_T': 0.670391061452512}
Results for 380_UCR_id_78_Medical_tr_1830_1st_4675.csv: {'filename': '380_UCR_id_78_Medical_tr_1830_1st_4675.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 178, 'train_index': '1830', 'data_shape': '7321x1', 'output_length': 10000, 'label_length': 5491, 'aligned_length': 5491, 'AUC-PR': 0.5313844707586088, 'AUC-ROC': 0.9290171929469409, 'VUS-PR': 0.5279442286387755, 'VUS-ROC': 0.9314784300428529, 'Standard-F1': 0.6280991735532173, 'Standard-Precision': 0.6195652173913027, 'Standard-Recall': 0.6368715083798865, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9162807806523979, 'Affiliation-P': 0.848544423551061, 'Affiliation-R': 0.9957696138628065, 'F1_T': 0.6274509803916573, 'Precision_T': 0.5896805896805882, 'Recall_T': 0.670391061452512}
Logits metrics skipped for 380_UCR_id_78_Medical_tr_1830_1st_4675.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/380_UCR_id_78_Medical_tr_1830_1st_4675_results.pkl
Processing file: 655_YAHOO_id_105_WebService_tr_500_1st_17.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 655_YAHOO_id_105_WebService_tr_500_1st_17.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 163.90it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 832.14it/s]
Table1-only metrics completed in 0.63s
Evaluation Result:  {'AUC-PR': 1.0000000000000002, 'AUC-ROC': 1.0, 'VUS-PR': 0.9917870404463361, 'VUS-ROC': 0.9999649278022338, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9943394579699197, 'Affiliation-P': 0.9887426387065896, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Results for 655_YAHOO_id_105_WebService_tr_500_1st_17.csv: {'filename': '655_YAHOO_id_105_WebService_tr_500_1st_17.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0000000000000002, 'AUC-ROC': 1.0, 'VUS-PR': 0.9917870404463361, 'VUS-ROC': 0.9999649278022338, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9943394579699197, 'Affiliation-P': 0.9887426387065896, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 655_YAHOO_id_105_WebService_tr_500_1st_17.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/655_YAHOO_id_105_WebService_tr_500_1st_17_results.pkl
Processing file: 479_UCR_id_177_Medical_tr_2000_1st_7000.csv, Data shape: (8000, 1)
Label shape: (8000,)
Train data shape: (2000, 1), Test data shape: (6000, 1), Label test shape: (6000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 479_UCR_id_177_Medical_tr_2000_1st_7000.csv...
Using CUDA device: 2
Data shape: (6000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6000, window_size 5000, adding 4000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.77it/s]
Original shapes - Output: (10000,), Label: (6000,)
Aligned shapes - Output: (6000,), Label: (6000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 532.97it/s]
Table1-only metrics completed in 0.50s
Evaluation Result:  {'AUC-PR': 0.23434302476829394, 'AUC-ROC': 0.9726924369747898, 'VUS-PR': 0.23731740399409265, 'VUS-ROC': 0.9725092436974795, 'Standard-F1': 0.40983606557328006, 'Standard-Precision': 0.34722222222221744, 'Standard-Recall': 0.49999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9132315871286788, 'Affiliation-P': 0.8410956284153005, 'Affiliation-R': 0.9989016666666666, 'F1_T': 0.385714285713821, 'Precision_T': 0.29999999999999666, 'Recall_T': 0.5399999999999892}
Results for 479_UCR_id_177_Medical_tr_2000_1st_7000.csv: {'filename': '479_UCR_id_177_Medical_tr_2000_1st_7000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 26, 'train_index': '2000', 'data_shape': '8000x1', 'output_length': 10000, 'label_length': 6000, 'aligned_length': 6000, 'AUC-PR': 0.23434302476829394, 'AUC-ROC': 0.9726924369747898, 'VUS-PR': 0.23731740399409265, 'VUS-ROC': 0.9725092436974795, 'Standard-F1': 0.40983606557328006, 'Standard-Precision': 0.34722222222221744, 'Standard-Recall': 0.49999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9132315871286788, 'Affiliation-P': 0.8410956284153005, 'Affiliation-R': 0.9989016666666666, 'F1_T': 0.385714285713821, 'Precision_T': 0.29999999999999666, 'Recall_T': 0.5399999999999892}
Logits metrics skipped for 479_UCR_id_177_Medical_tr_2000_1st_7000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/479_UCR_id_177_Medical_tr_2000_1st_7000_results.pkl
Processing file: 072_WSD_id_44_WebService_tr_1691_1st_1791.csv, Data shape: (12500, 1)
Label shape: (12500,)
Train data shape: (1691, 1), Test data shape: (10809, 1), Label test shape: (10809,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 072_WSD_id_44_WebService_tr_1691_1st_1791.csv...
Using CUDA device: 2
Data shape: (10809, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 10809, window_size 5000, adding 4191 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.63it/s]
Original shapes - Output: (15000,), Label: (10809,)
Aligned shapes - Output: (10809,), Label: (10809,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 302.38it/s]
Table1-only metrics completed in 1.72s
Evaluation Result:  {'AUC-PR': 0.24580735748690125, 'AUC-ROC': 0.8666824580605079, 'VUS-PR': 0.1900087015947708, 'VUS-ROC': 0.8974380872662361, 'Standard-F1': 0.28378378378331665, 'Standard-Precision': 0.38532110091742766, 'Standard-Recall': 0.22459893048128224, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7690168590147024, 'Affiliation-P': 0.6522152246935516, 'Affiliation-R': 0.9367795523205656, 'F1_T': 0.28378378378331665, 'Precision_T': 0.38532110091742766, 'Recall_T': 0.22459893048128224}
Results for 072_WSD_id_44_WebService_tr_1691_1st_1791.csv: {'filename': '072_WSD_id_44_WebService_tr_1691_1st_1791.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '1691', 'data_shape': '12500x1', 'output_length': 15000, 'label_length': 10809, 'aligned_length': 10809, 'AUC-PR': 0.24580735748690125, 'AUC-ROC': 0.8666824580605079, 'VUS-PR': 0.1900087015947708, 'VUS-ROC': 0.8974380872662361, 'Standard-F1': 0.28378378378331665, 'Standard-Precision': 0.38532110091742766, 'Standard-Recall': 0.22459893048128224, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7690168590147024, 'Affiliation-P': 0.6522152246935516, 'Affiliation-R': 0.9367795523205656, 'F1_T': 0.28378378378331665, 'Precision_T': 0.38532110091742766, 'Recall_T': 0.22459893048128224}
Logits metrics skipped for 072_WSD_id_44_WebService_tr_1691_1st_1791.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/072_WSD_id_44_WebService_tr_1691_1st_1791_results.pkl
Skipping file: 191_SMD_id_14_Facility_tr_4596_1st_4696.csv due to filter criteria.
Processing file: 741_YAHOO_id_191_WebService_tr_500_1st_1114.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 741_YAHOO_id_191_WebService_tr_500_1st_1114.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 156.93it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1362.05it/s]
Table1-only metrics completed in 0.29s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9373442067747467, 'VUS-ROC': 0.9998669662346468, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9953151772532757, 'Affiliation-P': 0.9906740449528921, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Results for 741_YAHOO_id_191_WebService_tr_500_1st_1114.csv: {'filename': '741_YAHOO_id_191_WebService_tr_500_1st_1114.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9373442067747467, 'VUS-ROC': 0.9998669662346468, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9953151772532757, 'Affiliation-P': 0.9906740449528921, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Logits metrics skipped for 741_YAHOO_id_191_WebService_tr_500_1st_1114.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/741_YAHOO_id_191_WebService_tr_500_1st_1114_results.pkl
Processing file: 412_UCR_id_110_Medical_tr_7500_1st_16900.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 412_UCR_id_110_Medical_tr_7500_1st_16900.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 151.70it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 207.96it/s]
Table1-only metrics completed in 1.87s
Evaluation Result:  {'AUC-PR': 0.3985245779831696, 'AUC-ROC': 0.9516668161434977, 'VUS-PR': 0.3027053666706489, 'VUS-ROC': 0.9520033633525564, 'Standard-F1': 0.38117647058773524, 'Standard-Precision': 0.35999999999999843, 'Standard-Recall': 0.404999999999998, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7951486710494541, 'Affiliation-P': 0.6605937621832358, 'Affiliation-R': 0.9985388888888889, 'F1_T': 0.38117647058773524, 'Precision_T': 0.35999999999999843, 'Recall_T': 0.404999999999998}
Results for 412_UCR_id_110_Medical_tr_7500_1st_16900.csv: {'filename': '412_UCR_id_110_Medical_tr_7500_1st_16900.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 166, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.3985245779831696, 'AUC-ROC': 0.9516668161434977, 'VUS-PR': 0.3027053666706489, 'VUS-ROC': 0.9520033633525564, 'Standard-F1': 0.38117647058773524, 'Standard-Precision': 0.35999999999999843, 'Standard-Recall': 0.404999999999998, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7951486710494541, 'Affiliation-P': 0.6605937621832358, 'Affiliation-R': 0.9985388888888889, 'F1_T': 0.38117647058773524, 'Precision_T': 0.35999999999999843, 'Recall_T': 0.404999999999998}
Logits metrics skipped for 412_UCR_id_110_Medical_tr_7500_1st_16900.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/412_UCR_id_110_Medical_tr_7500_1st_16900_results.pkl
Processing file: 746_YAHOO_id_196_WebService_tr_500_1st_413.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 746_YAHOO_id_196_WebService_tr_500_1st_413.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 160.38it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1383.04it/s]
Table1-only metrics completed in 0.31s
Evaluation Result:  {'AUC-PR': 0.7666666666666666, 'AUC-ROC': 0.9980175587652224, 'VUS-PR': 0.8050372846179066, 'VUS-ROC': 0.9991379871769176, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9944170215958227, 'Affiliation-P': 0.9888960363807335, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Results for 746_YAHOO_id_196_WebService_tr_500_1st_413.csv: {'filename': '746_YAHOO_id_196_WebService_tr_500_1st_413.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.7666666666666666, 'AUC-ROC': 0.9980175587652224, 'VUS-PR': 0.8050372846179066, 'VUS-ROC': 0.9991379871769176, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9944170215958227, 'Affiliation-P': 0.9888960363807335, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Logits metrics skipped for 746_YAHOO_id_196_WebService_tr_500_1st_413.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/746_YAHOO_id_196_WebService_tr_500_1st_413_results.pkl
Processing file: 615_YAHOO_id_65_WebService_tr_500_1st_348.csv, Data shape: (1439, 1)
Label shape: (1439,)
Train data shape: (500, 1), Test data shape: (939, 1), Label test shape: (939,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 615_YAHOO_id_65_WebService_tr_500_1st_348.csv...
Using CUDA device: 2
Data shape: (939, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 939, 1])
batched_mask shape: torch.Size([1, 939])

100%|██████████| 1/1 [00:00<00:00, 160.03it/s]
Original shapes - Output: (939,), Label: (939,)
Aligned shapes - Output: (939,), Label: (939,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Table1-only metrics completed in 0.01s
Evaluation Result:  {'AUC-PR': 0.0, 'AUC-ROC': nan, 'VUS-PR': 0.0, 'VUS-ROC': 0.0, 'Standard-F1': 0.0, 'Standard-Precision': 0.0, 'Standard-Recall': 0.0, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.0, 'Affiliation-P': 0.0, 'Affiliation-R': 0.0, 'F1_T': 0.0, 'Precision_T': 0.0, 'Recall_T': 0.0}
Results for 615_YAHOO_id_65_WebService_tr_500_1st_348.csv: {'filename': '615_YAHOO_id_65_WebService_tr_500_1st_348.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 25, 'train_index': '500', 'data_shape': '1439x1', 'output_length': 939, 'label_length': 939, 'aligned_length': 939, 'AUC-PR': 0.0, 'AUC-ROC': nan, 'VUS-PR': 0.0, 'VUS-ROC': 0.0, 'Standard-F1': 0.0, 'Standard-Precision': 0.0, 'Standard-Recall': 0.0, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.0, 'Affiliation-P': 0.0, 'Affiliation-R': 0.0, 'F1_T': 0.0, 'Precision_T': 0.0, 'Recall_T': 0.0}
Logits metrics skipped for 615_YAHOO_id_65_WebService_tr_500_1st_348.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/615_YAHOO_id_65_WebService_tr_500_1st_348_results.pkl
Processing file: 016_NAB_id_16_Environment_tr_1816_1st_3540.csv, Data shape: (7266, 1)
Label shape: (7266,)
Train data shape: (1816, 1), Test data shape: (5450, 1), Label test shape: (5450,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 016_NAB_id_16_Environment_tr_1816_1st_3540.csv...
Using CUDA device: 2
Data shape: (5450, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5450, window_size 5000, adding 4550 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 167.97it/s]
Original shapes - Output: (10000,), Label: (5450,)
Aligned shapes - Output: (5450,), Label: (5450,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 645.12it/s]
Table1-only metrics completed in 0.49s
Evaluation Result:  {'AUC-PR': 0.5195230846621263, 'AUC-ROC': 0.8849537529547373, 'VUS-PR': 0.5085693199651634, 'VUS-ROC': 0.8901432020292339, 'Standard-F1': 0.6228936664724926, 'Standard-Precision': 0.5376128385155461, 'Standard-Recall': 0.7403314917127062, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8922084289950754, 'Affiliation-P': 0.8835139213950407, 'Affiliation-R': 0.9010757596057007, 'F1_T': 0.6239813736898492, 'Precision_T': 0.5392354124748485, 'Recall_T': 0.7403314917127062}
Results for 016_NAB_id_16_Environment_tr_1816_1st_3540.csv: {'filename': '016_NAB_id_16_Environment_tr_1816_1st_3540.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 23, 'train_index': '1816', 'data_shape': '7266x1', 'output_length': 10000, 'label_length': 5450, 'aligned_length': 5450, 'AUC-PR': 0.5195230846621263, 'AUC-ROC': 0.8849537529547373, 'VUS-PR': 0.5085693199651634, 'VUS-ROC': 0.8901432020292339, 'Standard-F1': 0.6228936664724926, 'Standard-Precision': 0.5376128385155461, 'Standard-Recall': 0.7403314917127062, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8922084289950754, 'Affiliation-P': 0.8835139213950407, 'Affiliation-R': 0.9010757596057007, 'F1_T': 0.6239813736898492, 'Precision_T': 0.5392354124748485, 'Recall_T': 0.7403314917127062}
Logits metrics skipped for 016_NAB_id_16_Environment_tr_1816_1st_3540.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/016_NAB_id_16_Environment_tr_1816_1st_3540_results.pkl
Skipping file: 248_SVDB_id_12_Medical_tr_18462_1st_18562.csv due to filter criteria.
Processing file: 677_YAHOO_id_127_WebService_tr_500_1st_46.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 677_YAHOO_id_127_WebService_tr_500_1st_46.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 163.06it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1029.11it/s]
Table1-only metrics completed in 0.36s
Evaluation Result:  {'AUC-PR': 0.7175438596491227, 'AUC-ROC': 0.9899204997160704, 'VUS-PR': 0.6837678090406686, 'VUS-ROC': 0.991731969866847, 'Standard-F1': 0.4444444444439506, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.6666666666665555, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.953427216567122, 'Affiliation-P': 0.9109994370766771, 'Affiliation-R': 1.0, 'F1_T': 0.4444444444439506, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.6666666666665555}
Results for 677_YAHOO_id_127_WebService_tr_500_1st_46.csv: {'filename': '677_YAHOO_id_127_WebService_tr_500_1st_46.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 11, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.7175438596491227, 'AUC-ROC': 0.9899204997160704, 'VUS-PR': 0.6837678090406686, 'VUS-ROC': 0.991731969866847, 'Standard-F1': 0.4444444444439506, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.6666666666665555, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.953427216567122, 'Affiliation-P': 0.9109994370766771, 'Affiliation-R': 1.0, 'F1_T': 0.4444444444439506, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.6666666666665555}
Logits metrics skipped for 677_YAHOO_id_127_WebService_tr_500_1st_46.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/677_YAHOO_id_127_WebService_tr_500_1st_46_results.pkl
Processing file: 278_NEK_id_2_WebService_tr_500_1st_64.csv, Data shape: (1073, 1)
Label shape: (1073,)
Train data shape: (500, 1), Test data shape: (573, 1), Label test shape: (573,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 278_NEK_id_2_WebService_tr_500_1st_64.csv...
Using CUDA device: 2
Data shape: (573, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 573, 1])
batched_mask shape: torch.Size([1, 573])

100%|██████████| 1/1 [00:00<00:00, 156.80it/s]
Original shapes - Output: (573,), Label: (573,)
Aligned shapes - Output: (573,), Label: (573,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2427.08it/s]
Table1-only metrics completed in 0.24s
Evaluation Result:  {'AUC-PR': 0.5238807184369183, 'AUC-ROC': 0.6656665027053615, 'VUS-PR': 0.6365151590081561, 'VUS-ROC': 0.7648226937018819, 'Standard-F1': 0.6129032258059574, 'Standard-Precision': 0.7916666666666338, 'Standard-Recall': 0.49999999999998684, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7900047151595575, 'Affiliation-P': 1.0, 'Affiliation-R': 0.6528990030434154, 'F1_T': 0.5937499999994991, 'Precision_T': 0.7307692307692027, 'Recall_T': 0.49999999999998684}
Results for 278_NEK_id_2_WebService_tr_500_1st_64.csv: {'filename': '278_NEK_id_2_WebService_tr_500_1st_64.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1073x1', 'output_length': 573, 'label_length': 573, 'aligned_length': 573, 'AUC-PR': 0.5238807184369183, 'AUC-ROC': 0.6656665027053615, 'VUS-PR': 0.6365151590081561, 'VUS-ROC': 0.7648226937018819, 'Standard-F1': 0.6129032258059574, 'Standard-Precision': 0.7916666666666338, 'Standard-Recall': 0.49999999999998684, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7900047151595575, 'Affiliation-P': 1.0, 'Affiliation-R': 0.6528990030434154, 'F1_T': 0.5937499999994991, 'Precision_T': 0.7307692307692027, 'Recall_T': 0.49999999999998684}
Logits metrics skipped for 278_NEK_id_2_WebService_tr_500_1st_64.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/278_NEK_id_2_WebService_tr_500_1st_64_results.pkl
Processing file: 062_WSD_id_34_WebService_tr_3371_1st_3471.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (3371, 1), Test data shape: (14865, 1), Label test shape: (14865,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 062_WSD_id_34_WebService_tr_3371_1st_3471.csv...
Using CUDA device: 2
Data shape: (14865, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 14865, window_size 5000, adding 135 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.80it/s]
Original shapes - Output: (15000,), Label: (14865,)
Aligned shapes - Output: (14865,), Label: (14865,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 156/200 [00:01<00:00, 155.91it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 170.44it/s]
Table1-only metrics completed in 1.67s
Evaluation Result:  {'AUC-PR': 0.3099993885775265, 'AUC-ROC': 0.8284397313838832, 'VUS-PR': 0.1458266223079579, 'VUS-ROC': 0.8325515521563446, 'Standard-F1': 0.3136094674551209, 'Standard-Precision': 0.2977528089887624, 'Standard-Recall': 0.33124999999999793, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7902608423044507, 'Affiliation-P': 0.6931790923588396, 'Affiliation-R': 0.9189644952578732, 'F1_T': 0.3106796116499841, 'Precision_T': 0.3221476510067093, 'Recall_T': 0.29999999999999816}
Results for 062_WSD_id_34_WebService_tr_3371_1st_3471.csv: {'filename': '062_WSD_id_34_WebService_tr_3371_1st_3471.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 11, 'train_index': '3371', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 14865, 'aligned_length': 14865, 'AUC-PR': 0.3099993885775265, 'AUC-ROC': 0.8284397313838832, 'VUS-PR': 0.1458266223079579, 'VUS-ROC': 0.8325515521563446, 'Standard-F1': 0.3136094674551209, 'Standard-Precision': 0.2977528089887624, 'Standard-Recall': 0.33124999999999793, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7902608423044507, 'Affiliation-P': 0.6931790923588396, 'Affiliation-R': 0.9189644952578732, 'F1_T': 0.3106796116499841, 'Precision_T': 0.3221476510067093, 'Recall_T': 0.29999999999999816}
Logits metrics skipped for 062_WSD_id_34_WebService_tr_3371_1st_3471.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/062_WSD_id_34_WebService_tr_3371_1st_3471_results.pkl
Processing file: 398_UCR_id_96_Medical_tr_1913_1st_3474.csv, Data shape: (7654, 1)
Label shape: (7654,)
Train data shape: (1913, 1), Test data shape: (5741, 1), Label test shape: (5741,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 398_UCR_id_96_Medical_tr_1913_1st_3474.csv...
Using CUDA device: 2
Data shape: (5741, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5741, window_size 5000, adding 4259 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 156.32it/s]
Original shapes - Output: (10000,), Label: (5741,)
Aligned shapes - Output: (5741,), Label: (5741,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 716.02it/s]
Table1-only metrics completed in 0.87s
Evaluation Result:  {'AUC-PR': 0.5413052346142503, 'AUC-ROC': 0.9597784784542, 'VUS-PR': 0.6874947584565027, 'VUS-ROC': 0.9765293283518428, 'Standard-F1': 0.5348189415036846, 'Standard-Precision': 0.47058823529411536, 'Standard-Recall': 0.6193548387096734, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9919584256695997, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9840451534238726, 'F1_T': 0.5423728813554368, 'Precision_T': 0.48241206030150513, 'Recall_T': 0.6193548387096734}
Results for 398_UCR_id_96_Medical_tr_1913_1st_3474.csv: {'filename': '398_UCR_id_96_Medical_tr_1913_1st_3474.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 154, 'train_index': '1913', 'data_shape': '7654x1', 'output_length': 10000, 'label_length': 5741, 'aligned_length': 5741, 'AUC-PR': 0.5413052346142503, 'AUC-ROC': 0.9597784784542, 'VUS-PR': 0.6874947584565027, 'VUS-ROC': 0.9765293283518428, 'Standard-F1': 0.5348189415036846, 'Standard-Precision': 0.47058823529411536, 'Standard-Recall': 0.6193548387096734, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9919584256695997, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9840451534238726, 'F1_T': 0.5423728813554368, 'Precision_T': 0.48241206030150513, 'Recall_T': 0.6193548387096734}
Logits metrics skipped for 398_UCR_id_96_Medical_tr_1913_1st_3474.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/398_UCR_id_96_Medical_tr_1913_1st_3474_results.pkl
Processing file: 377_UCR_id_75_Medical_tr_1875_1st_6599.csv, Data shape: (7501, 1)
Label shape: (7501,)
Train data shape: (1875, 1), Test data shape: (5626, 1), Label test shape: (5626,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 377_UCR_id_75_Medical_tr_1875_1st_6599.csv...
Using CUDA device: 2
Data shape: (5626, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5626, window_size 5000, adding 4374 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.13it/s]
Original shapes - Output: (10000,), Label: (5626,)
Aligned shapes - Output: (5626,), Label: (5626,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 675.77it/s]
Table1-only metrics completed in 0.95s
Evaluation Result:  {'AUC-PR': 0.0083837704022319, 'AUC-ROC': 0.1624850420582128, 'VUS-PR': 0.011780733459405108, 'VUS-ROC': 0.2702712307250088, 'Standard-F1': 0.029959810010931384, 'Standard-Precision': 0.015207715133531155, 'Standard-Recall': 0.9999999999999879, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8157787590680821, 'Affiliation-P': 0.7630857962415877, 'Affiliation-R': 0.8762886597938145, 'F1_T': 0.029899726526862052, 'Precision_T': 0.01517675365537664, 'Recall_T': 0.9999999999999879}
Results for 377_UCR_id_75_Medical_tr_1875_1st_6599.csv: {'filename': '377_UCR_id_75_Medical_tr_1875_1st_6599.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 175, 'train_index': '1875', 'data_shape': '7501x1', 'output_length': 10000, 'label_length': 5626, 'aligned_length': 5626, 'AUC-PR': 0.0083837704022319, 'AUC-ROC': 0.1624850420582128, 'VUS-PR': 0.011780733459405108, 'VUS-ROC': 0.2702712307250088, 'Standard-F1': 0.029959810010931384, 'Standard-Precision': 0.015207715133531155, 'Standard-Recall': 0.9999999999999879, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8157787590680821, 'Affiliation-P': 0.7630857962415877, 'Affiliation-R': 0.8762886597938145, 'F1_T': 0.029899726526862052, 'Precision_T': 0.01517675365537664, 'Recall_T': 0.9999999999999879}
Logits metrics skipped for 377_UCR_id_75_Medical_tr_1875_1st_6599.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/377_UCR_id_75_Medical_tr_1875_1st_6599_results.pkl
Processing file: 038_WSD_id_10_WebService_tr_4042_1st_4142.csv, Data shape: (18084, 1)
Label shape: (18084,)
Train data shape: (4042, 1), Test data shape: (14042, 1), Label test shape: (14042,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 038_WSD_id_10_WebService_tr_4042_1st_4142.csv...
Using CUDA device: 2
Data shape: (14042, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 14042, window_size 5000, adding 958 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.74it/s]
Original shapes - Output: (15000,), Label: (14042,)
Aligned shapes - Output: (14042,), Label: (14042,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 242.38it/s]
Table1-only metrics completed in 2.51s
Evaluation Result:  {'AUC-PR': 0.8847210266862365, 'AUC-ROC': 0.9936601273178263, 'VUS-PR': 0.8739745855468275, 'VUS-ROC': 0.9992066215511823, 'Standard-F1': 0.5700483091783041, 'Standard-Precision': 0.4184397163120538, 'Standard-Recall': 0.8939393939393805, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9933419968808124, 'Affiliation-P': 0.987450696760842, 'Affiliation-R': 0.9993040159348033, 'F1_T': 0.5700483091783041, 'Precision_T': 0.4184397163120538, 'Recall_T': 0.8939393939393805}
Results for 038_WSD_id_10_WebService_tr_4042_1st_4142.csv: {'filename': '038_WSD_id_10_WebService_tr_4042_1st_4142.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4042', 'data_shape': '18084x1', 'output_length': 15000, 'label_length': 14042, 'aligned_length': 14042, 'AUC-PR': 0.8847210266862365, 'AUC-ROC': 0.9936601273178263, 'VUS-PR': 0.8739745855468275, 'VUS-ROC': 0.9992066215511823, 'Standard-F1': 0.5700483091783041, 'Standard-Precision': 0.4184397163120538, 'Standard-Recall': 0.8939393939393805, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9933419968808124, 'Affiliation-P': 0.987450696760842, 'Affiliation-R': 0.9993040159348033, 'F1_T': 0.5700483091783041, 'Precision_T': 0.4184397163120538, 'Recall_T': 0.8939393939393805}
Logits metrics skipped for 038_WSD_id_10_WebService_tr_4042_1st_4142.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/038_WSD_id_10_WebService_tr_4042_1st_4142_results.pkl
Processing file: 456_UCR_id_154_Facility_tr_22500_1st_72150.csv, Data shape: (90000, 1)
Label shape: (90000,)
Train data shape: (22500, 1), Test data shape: (67500, 1), Label test shape: (67500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 456_UCR_id_154_Facility_tr_22500_1st_72150.csv...
Using CUDA device: 2
Data shape: (67500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 67500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([14, 5000, 1])
batched_mask shape: torch.Size([14, 5000])

100%|██████████| 1/1 [00:00<00:00, 74.79it/s]
Original shapes - Output: (70000,), Label: (67500,)
Aligned shapes - Output: (67500,), Label: (67500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  30%|██▉       | 59/200 [00:01<00:02, 58.34it/s]

Calculating Affiliation Metrics:  60%|█████▉    | 119/200 [00:02<00:01, 58.83it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:03<00:00, 66.56it/s]
Table1-only metrics completed in 4.81s
Evaluation Result:  {'AUC-PR': 0.004457700535518897, 'AUC-ROC': 0.48194652431806584, 'VUS-PR': 0.0049387352922352, 'VUS-ROC': 0.5083354139958273, 'Standard-F1': 0.012627432808142467, 'Standard-Precision': 0.006356055746690769, 'Standard-Recall': 0.9478260869565189, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.718645553489306, 'Affiliation-P': 0.5635735587002096, 'Affiliation-R': 0.9914518518518518, 'F1_T': 0.012606256866927169, 'Precision_T': 0.0063453254162300615, 'Recall_T': 0.9478260869565189}
Results for 456_UCR_id_154_Facility_tr_22500_1st_72150.csv: {'filename': '456_UCR_id_154_Facility_tr_22500_1st_72150.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '22500', 'data_shape': '90000x1', 'output_length': 70000, 'label_length': 67500, 'aligned_length': 67500, 'AUC-PR': 0.004457700535518897, 'AUC-ROC': 0.48194652431806584, 'VUS-PR': 0.0049387352922352, 'VUS-ROC': 0.5083354139958273, 'Standard-F1': 0.012627432808142467, 'Standard-Precision': 0.006356055746690769, 'Standard-Recall': 0.9478260869565189, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.718645553489306, 'Affiliation-P': 0.5635735587002096, 'Affiliation-R': 0.9914518518518518, 'F1_T': 0.012606256866927169, 'Precision_T': 0.0063453254162300615, 'Recall_T': 0.9478260869565189}
Logits metrics skipped for 456_UCR_id_154_Facility_tr_22500_1st_72150.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/456_UCR_id_154_Facility_tr_22500_1st_72150_results.pkl
Processing file: 137_WSD_id_109_WebService_tr_4559_1st_14404.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 137_WSD_id_109_WebService_tr_4559_1st_14404.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.14it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 264.02it/s]
Table1-only metrics completed in 1.01s
Evaluation Result:  {'AUC-PR': 0.09082497912794345, 'AUC-ROC': 0.5206195190313916, 'VUS-PR': 0.033065167125894655, 'VUS-ROC': 0.6127563957446683, 'Standard-F1': 0.08163265306080285, 'Standard-Precision': 0.05839416058394118, 'Standard-Recall': 0.13559322033898075, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8938576339755983, 'Affiliation-P': 0.8095284747515625, 'Affiliation-R': 0.9977990761766533, 'F1_T': 0.08163265306080285, 'Precision_T': 0.05839416058394118, 'Recall_T': 0.13559322033898075}
Results for 137_WSD_id_109_WebService_tr_4559_1st_14404.csv: {'filename': '137_WSD_id_109_WebService_tr_4559_1st_14404.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.09082497912794345, 'AUC-ROC': 0.5206195190313916, 'VUS-PR': 0.033065167125894655, 'VUS-ROC': 0.6127563957446683, 'Standard-F1': 0.08163265306080285, 'Standard-Precision': 0.05839416058394118, 'Standard-Recall': 0.13559322033898075, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8938576339755983, 'Affiliation-P': 0.8095284747515625, 'Affiliation-R': 0.9977990761766533, 'F1_T': 0.08163265306080285, 'Precision_T': 0.05839416058394118, 'Recall_T': 0.13559322033898075}
Logits metrics skipped for 137_WSD_id_109_WebService_tr_4559_1st_14404.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/137_WSD_id_109_WebService_tr_4559_1st_14404_results.pkl
Skipping file: 838_Exathlon_id_29_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 654_YAHOO_id_104_WebService_tr_500_1st_20.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 654_YAHOO_id_104_WebService_tr_500_1st_20.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 165.72it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1313.97it/s]
Table1-only metrics completed in 0.45s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5882352941171626, 'Standard-Precision': 0.41666666666663194, 'Standard-Recall': 0.9999999999997999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9986329460013664, 'Affiliation-P': 0.9972696245733788, 'Affiliation-R': 1.0, 'F1_T': 0.5882352941171626, 'Precision_T': 0.41666666666663194, 'Recall_T': 0.9999999999997999}
Results for 654_YAHOO_id_104_WebService_tr_500_1st_20.csv: {'filename': '654_YAHOO_id_104_WebService_tr_500_1st_20.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5882352941171626, 'Standard-Precision': 0.41666666666663194, 'Standard-Recall': 0.9999999999997999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9986329460013664, 'Affiliation-P': 0.9972696245733788, 'Affiliation-R': 1.0, 'F1_T': 0.5882352941171626, 'Precision_T': 0.41666666666663194, 'Recall_T': 0.9999999999997999}
Logits metrics skipped for 654_YAHOO_id_104_WebService_tr_500_1st_20.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/654_YAHOO_id_104_WebService_tr_500_1st_20_results.pkl
Processing file: 450_UCR_id_148_Medical_tr_7500_1st_16900.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 450_UCR_id_148_Medical_tr_7500_1st_16900.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 153.09it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  89%|████████▉ | 178/200 [00:01<00:00, 177.57it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 186.30it/s]
Table1-only metrics completed in 1.99s
Evaluation Result:  {'AUC-PR': 0.42680147588493766, 'AUC-ROC': 0.926220403587444, 'VUS-PR': 0.3448899866492446, 'VUS-ROC': 0.9300760639311537, 'Standard-F1': 0.40941176470538215, 'Standard-Precision': 0.38666666666666494, 'Standard-Recall': 0.43499999999999783, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9037739197429868, 'Affiliation-P': 0.8253820662768032, 'Affiliation-R': 0.9986192222222222, 'F1_T': 0.40941176470538215, 'Precision_T': 0.38666666666666494, 'Recall_T': 0.43499999999999783}
Results for 450_UCR_id_148_Medical_tr_7500_1st_16900.csv: {'filename': '450_UCR_id_148_Medical_tr_7500_1st_16900.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 166, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.42680147588493766, 'AUC-ROC': 0.926220403587444, 'VUS-PR': 0.3448899866492446, 'VUS-ROC': 0.9300760639311537, 'Standard-F1': 0.40941176470538215, 'Standard-Precision': 0.38666666666666494, 'Standard-Recall': 0.43499999999999783, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9037739197429868, 'Affiliation-P': 0.8253820662768032, 'Affiliation-R': 0.9986192222222222, 'F1_T': 0.40941176470538215, 'Precision_T': 0.38666666666666494, 'Recall_T': 0.43499999999999783}
Logits metrics skipped for 450_UCR_id_148_Medical_tr_7500_1st_16900.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/450_UCR_id_148_Medical_tr_7500_1st_16900_results.pkl
Skipping file: 222_LTDB_id_7_Medical_tr_3618_1st_3718.csv due to filter criteria.
Processing file: 793_YAHOO_id_243_WebService_tr_500_1st_28.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 793_YAHOO_id_243_WebService_tr_500_1st_28.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 167.95it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1301.50it/s]
Table1-only metrics completed in 0.35s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9504941880689357, 'VUS-ROC': 0.9998948142040058, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9970625102682831, 'Affiliation-P': 0.9941422276826004, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Results for 793_YAHOO_id_243_WebService_tr_500_1st_28.csv: {'filename': '793_YAHOO_id_243_WebService_tr_500_1st_28.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9504941880689357, 'VUS-ROC': 0.9998948142040058, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9970625102682831, 'Affiliation-P': 0.9941422276826004, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Logits metrics skipped for 793_YAHOO_id_243_WebService_tr_500_1st_28.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/793_YAHOO_id_243_WebService_tr_500_1st_28_results.pkl
Processing file: 635_YAHOO_id_85_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 635_YAHOO_id_85_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 164.99it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1826.37it/s]
Table1-only metrics completed in 0.75s
Evaluation Result:  {'AUC-PR': 0.8875, 'AUC-ROC': 0.9994547437295529, 'VUS-PR': 0.9899053693003445, 'VUS-ROC': 0.9999661054096812, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 635_YAHOO_id_85_Synthetic_tr_500_1st_658.csv: {'filename': '635_YAHOO_id_85_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 116, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.8875, 'AUC-ROC': 0.9994547437295529, 'VUS-PR': 0.9899053693003445, 'VUS-ROC': 0.9999661054096812, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 635_YAHOO_id_85_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/635_YAHOO_id_85_Synthetic_tr_500_1st_658_results.pkl
Processing file: 311_UCR_id_9_HumanActivity_tr_16000_1st_38400.csv, Data shape: (64000, 1)
Label shape: (64000,)
Train data shape: (16000, 1), Test data shape: (48000, 1), Label test shape: (48000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 311_UCR_id_9_HumanActivity_tr_16000_1st_38400.csv...
Using CUDA device: 2
Data shape: (48000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 48000, window_size 5000, adding 2000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([10, 5000, 1])
batched_mask shape: torch.Size([10, 5000])

100%|██████████| 1/1 [00:00<00:00, 96.36it/s]
Original shapes - Output: (50000,), Label: (48000,)
Aligned shapes - Output: (48000,), Label: (48000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  45%|████▌     | 90/200 [00:01<00:01, 89.45it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 107.50it/s]
Table1-only metrics completed in 3.00s
Evaluation Result:  {'AUC-PR': 0.07250330866175672, 'AUC-ROC': 0.8970958951271186, 'VUS-PR': 0.0759515657580949, 'VUS-ROC': 0.9011735478146738, 'Standard-F1': 0.1762016899845653, 'Standard-Precision': 0.099080853715532, 'Standard-Recall': 0.7949999999999989, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6843037499476237, 'Affiliation-P': 0.5201076995700057, 'Affiliation-R': 1.0, 'F1_T': 0.17579972183568537, 'Precision_T': 0.09890453834115805, 'Recall_T': 0.789999999999999}
Results for 311_UCR_id_9_HumanActivity_tr_16000_1st_38400.csv: {'filename': '311_UCR_id_9_HumanActivity_tr_16000_1st_38400.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '16000', 'data_shape': '64000x1', 'output_length': 50000, 'label_length': 48000, 'aligned_length': 48000, 'AUC-PR': 0.07250330866175672, 'AUC-ROC': 0.8970958951271186, 'VUS-PR': 0.0759515657580949, 'VUS-ROC': 0.9011735478146738, 'Standard-F1': 0.1762016899845653, 'Standard-Precision': 0.099080853715532, 'Standard-Recall': 0.7949999999999989, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6843037499476237, 'Affiliation-P': 0.5201076995700057, 'Affiliation-R': 1.0, 'F1_T': 0.17579972183568537, 'Precision_T': 0.09890453834115805, 'Recall_T': 0.789999999999999}
Logits metrics skipped for 311_UCR_id_9_HumanActivity_tr_16000_1st_38400.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/311_UCR_id_9_HumanActivity_tr_16000_1st_38400_results.pkl
Skipping file: 213_SMD_id_36_Facility_tr_5925_1st_17580.csv due to filter criteria.
Processing file: 493_UCR_id_191_Sensor_tr_2847_1st_9330.csv, Data shape: (11388, 1)
Label shape: (11388,)
Train data shape: (2847, 1), Test data shape: (8541, 1), Label test shape: (8541,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 493_UCR_id_191_Sensor_tr_2847_1st_9330.csv...
Using CUDA device: 2
Data shape: (8541, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 8541, window_size 5000, adding 1459 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 152.57it/s]
Original shapes - Output: (10000,), Label: (8541,)
Aligned shapes - Output: (8541,), Label: (8541,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 423.52it/s]
Table1-only metrics completed in 0.87s
Evaluation Result:  {'AUC-PR': 0.0342571954842503, 'AUC-ROC': 0.9829445551517992, 'VUS-PR': 0.09490501326661312, 'VUS-ROC': 0.9836191688755228, 'Standard-F1': 0.09999999999989395, 'Standard-Precision': 0.05294117647058792, 'Standard-Recall': 0.89999999999991, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7448993267840003, 'Affiliation-P': 0.593497671286649, 'Affiliation-R': 1.0, 'F1_T': 0.09999999999989395, 'Precision_T': 0.05294117647058792, 'Recall_T': 0.89999999999991}
Results for 493_UCR_id_191_Sensor_tr_2847_1st_9330.csv: {'filename': '493_UCR_id_191_Sensor_tr_2847_1st_9330.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 98, 'train_index': '2847', 'data_shape': '11388x1', 'output_length': 10000, 'label_length': 8541, 'aligned_length': 8541, 'AUC-PR': 0.0342571954842503, 'AUC-ROC': 0.9829445551517992, 'VUS-PR': 0.09490501326661312, 'VUS-ROC': 0.9836191688755228, 'Standard-F1': 0.09999999999989395, 'Standard-Precision': 0.05294117647058792, 'Standard-Recall': 0.89999999999991, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7448993267840003, 'Affiliation-P': 0.593497671286649, 'Affiliation-R': 1.0, 'F1_T': 0.09999999999989395, 'Precision_T': 0.05294117647058792, 'Recall_T': 0.89999999999991}
Logits metrics skipped for 493_UCR_id_191_Sensor_tr_2847_1st_9330.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/493_UCR_id_191_Sensor_tr_2847_1st_9330_results.pkl
Processing file: 449_UCR_id_147_Sensor_tr_2847_1st_9330.csv, Data shape: (11388, 1)
Label shape: (11388,)
Train data shape: (2847, 1), Test data shape: (8541, 1), Label test shape: (8541,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 449_UCR_id_147_Sensor_tr_2847_1st_9330.csv...
Using CUDA device: 2
Data shape: (8541, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 8541, window_size 5000, adding 1459 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 153.20it/s]
Original shapes - Output: (10000,), Label: (8541,)
Aligned shapes - Output: (8541,), Label: (8541,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 460.33it/s]
Table1-only metrics completed in 0.84s
Evaluation Result:  {'AUC-PR': 0.005176730576509256, 'AUC-ROC': 0.8381901301137029, 'VUS-PR': 0.005754812857910034, 'VUS-ROC': 0.8533447938552481, 'Standard-F1': 0.020833333333146267, 'Standard-Precision': 0.011627906976744052, 'Standard-Recall': 0.09999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8955989207061429, 'Affiliation-P': 0.8549649706068948, 'Affiliation-R': 0.9402880224798033, 'F1_T': 0.020833333333146267, 'Precision_T': 0.011627906976744052, 'Recall_T': 0.09999999999999}
Results for 449_UCR_id_147_Sensor_tr_2847_1st_9330.csv: {'filename': '449_UCR_id_147_Sensor_tr_2847_1st_9330.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 98, 'train_index': '2847', 'data_shape': '11388x1', 'output_length': 10000, 'label_length': 8541, 'aligned_length': 8541, 'AUC-PR': 0.005176730576509256, 'AUC-ROC': 0.8381901301137029, 'VUS-PR': 0.005754812857910034, 'VUS-ROC': 0.8533447938552481, 'Standard-F1': 0.020833333333146267, 'Standard-Precision': 0.011627906976744052, 'Standard-Recall': 0.09999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8955989207061429, 'Affiliation-P': 0.8549649706068948, 'Affiliation-R': 0.9402880224798033, 'F1_T': 0.020833333333146267, 'Precision_T': 0.011627906976744052, 'Recall_T': 0.09999999999999}
Logits metrics skipped for 449_UCR_id_147_Sensor_tr_2847_1st_9330.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/449_UCR_id_147_Sensor_tr_2847_1st_9330_results.pkl
Processing file: 134_WSD_id_106_WebService_tr_4559_1st_7990.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 134_WSD_id_106_WebService_tr_4559_1st_7990.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.03it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 249.51it/s]
Table1-only metrics completed in 1.78s
Evaluation Result:  {'AUC-PR': 0.27259550773722885, 'AUC-ROC': 0.8422663198301594, 'VUS-PR': 0.19381768303113536, 'VUS-ROC': 0.8928512527840168, 'Standard-F1': 0.3045267489706991, 'Standard-Precision': 0.270072992700728, 'Standard-Recall': 0.3490566037735816, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9238751097504012, 'Affiliation-P': 0.8608071616783771, 'Affiliation-R': 0.9969151539816066, 'F1_T': 0.3045267489706991, 'Precision_T': 0.270072992700728, 'Recall_T': 0.3490566037735816}
Results for 134_WSD_id_106_WebService_tr_4559_1st_7990.csv: {'filename': '134_WSD_id_106_WebService_tr_4559_1st_7990.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.27259550773722885, 'AUC-ROC': 0.8422663198301594, 'VUS-PR': 0.19381768303113536, 'VUS-ROC': 0.8928512527840168, 'Standard-F1': 0.3045267489706991, 'Standard-Precision': 0.270072992700728, 'Standard-Recall': 0.3490566037735816, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9238751097504012, 'Affiliation-P': 0.8608071616783771, 'Affiliation-R': 0.9969151539816066, 'F1_T': 0.3045267489706991, 'Precision_T': 0.270072992700728, 'Recall_T': 0.3490566037735816}
Logits metrics skipped for 134_WSD_id_106_WebService_tr_4559_1st_7990.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/134_WSD_id_106_WebService_tr_4559_1st_7990_results.pkl
Processing file: 667_YAHOO_id_117_WebService_tr_500_1st_421.csv, Data shape: (1461, 1)
Label shape: (1461,)
Train data shape: (500, 1), Test data shape: (961, 1), Label test shape: (961,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 667_YAHOO_id_117_WebService_tr_500_1st_421.csv...
Using CUDA device: 2
Data shape: (961, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 961, 1])
batched_mask shape: torch.Size([1, 961])

100%|██████████| 1/1 [00:00<00:00, 159.81it/s]
Original shapes - Output: (961,), Label: (961,)
Aligned shapes - Output: (961,), Label: (961,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Table1-only metrics completed in 0.01s
Evaluation Result:  {'AUC-PR': 0.0, 'AUC-ROC': nan, 'VUS-PR': 0.0, 'VUS-ROC': 0.0, 'Standard-F1': 0.0, 'Standard-Precision': 0.0, 'Standard-Recall': 0.0, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.0, 'Affiliation-P': 0.0, 'Affiliation-R': 0.0, 'F1_T': 0.0, 'Precision_T': 0.0, 'Recall_T': 0.0}
Results for 667_YAHOO_id_117_WebService_tr_500_1st_421.csv: {'filename': '667_YAHOO_id_117_WebService_tr_500_1st_421.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1461x1', 'output_length': 961, 'label_length': 961, 'aligned_length': 961, 'AUC-PR': 0.0, 'AUC-ROC': nan, 'VUS-PR': 0.0, 'VUS-ROC': 0.0, 'Standard-F1': 0.0, 'Standard-Precision': 0.0, 'Standard-Recall': 0.0, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.0, 'Affiliation-P': 0.0, 'Affiliation-R': 0.0, 'F1_T': 0.0, 'Precision_T': 0.0, 'Recall_T': 0.0}
Logits metrics skipped for 667_YAHOO_id_117_WebService_tr_500_1st_421.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/667_YAHOO_id_117_WebService_tr_500_1st_421_results.pkl
Processing file: 320_UCR_id_18_Medical_tr_2500_1st_7000.csv, Data shape: (10001, 1)
Label shape: (10001,)
Train data shape: (2500, 1), Test data shape: (7501, 1), Label test shape: (7501,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 320_UCR_id_18_Medical_tr_2500_1st_7000.csv...
Using CUDA device: 2
Data shape: (7501, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 7501, window_size 5000, adding 2499 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 167.55it/s]
Original shapes - Output: (10000,), Label: (7501,)
Aligned shapes - Output: (7501,), Label: (7501,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 464.18it/s]
Table1-only metrics completed in 0.57s
Evaluation Result:  {'AUC-PR': 0.003739455054698939, 'AUC-ROC': 0.49592647124436706, 'VUS-PR': 0.004166277103926644, 'VUS-ROC': 0.5181390316373504, 'Standard-F1': 0.010196905766515524, 'Standard-Precision': 0.00512548603746907, 'Standard-Recall': 0.9666666666666345, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.80152499168213, 'Affiliation-P': 0.7923083939685235, 'Affiliation-R': 0.8109585388614851, 'F1_T': 0.010139860139849702, 'Precision_T': 0.0050966608084358515, 'Recall_T': 0.9666666666666345}
Results for 320_UCR_id_18_Medical_tr_2500_1st_7000.csv: {'filename': '320_UCR_id_18_Medical_tr_2500_1st_7000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 26, 'train_index': '2500', 'data_shape': '10001x1', 'output_length': 10000, 'label_length': 7501, 'aligned_length': 7501, 'AUC-PR': 0.003739455054698939, 'AUC-ROC': 0.49592647124436706, 'VUS-PR': 0.004166277103926644, 'VUS-ROC': 0.5181390316373504, 'Standard-F1': 0.010196905766515524, 'Standard-Precision': 0.00512548603746907, 'Standard-Recall': 0.9666666666666345, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.80152499168213, 'Affiliation-P': 0.7923083939685235, 'Affiliation-R': 0.8109585388614851, 'F1_T': 0.010139860139849702, 'Precision_T': 0.0050966608084358515, 'Recall_T': 0.9666666666666345}
Logits metrics skipped for 320_UCR_id_18_Medical_tr_2500_1st_7000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/320_UCR_id_18_Medical_tr_2500_1st_7000_results.pkl
Processing file: 719_YAHOO_id_169_WebService_tr_500_1st_852.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 719_YAHOO_id_169_WebService_tr_500_1st_852.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.18it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1136.23it/s]
Table1-only metrics completed in 0.53s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9686253650432952, 'VUS-ROC': 0.9998662757066167, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Results for 719_YAHOO_id_169_WebService_tr_500_1st_852.csv: {'filename': '719_YAHOO_id_169_WebService_tr_500_1st_852.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9686253650432952, 'VUS-ROC': 0.9998662757066167, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Logits metrics skipped for 719_YAHOO_id_169_WebService_tr_500_1st_852.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/719_YAHOO_id_169_WebService_tr_500_1st_852_results.pkl
Processing file: 634_YAHOO_id_84_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 634_YAHOO_id_84_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 164.29it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2055.36it/s]
Table1-only metrics completed in 1.15s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 634_YAHOO_id_84_Synthetic_tr_500_1st_658.csv: {'filename': '634_YAHOO_id_84_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 235, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 634_YAHOO_id_84_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/634_YAHOO_id_84_Synthetic_tr_500_1st_658_results.pkl
Processing file: 680_YAHOO_id_130_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 680_YAHOO_id_130_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 159.06it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2324.04it/s]
Table1-only metrics completed in 0.48s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9583804569102007, 'Affiliation-P': 0.920086862106406, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 680_YAHOO_id_130_Synthetic_tr_500_1st_893.csv: {'filename': '680_YAHOO_id_130_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 118, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9583804569102007, 'Affiliation-P': 0.920086862106406, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 680_YAHOO_id_130_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/680_YAHOO_id_130_Synthetic_tr_500_1st_893_results.pkl
Processing file: 125_WSD_id_97_WebService_tr_2217_1st_2317.csv, Data shape: (18250, 1)
Label shape: (18250,)
Train data shape: (2217, 1), Test data shape: (16033, 1), Label test shape: (16033,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 125_WSD_id_97_WebService_tr_2217_1st_2317.csv...
Using CUDA device: 2
Data shape: (16033, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 16033, window_size 5000, adding 3967 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.79it/s]
Original shapes - Output: (20000,), Label: (16033,)
Aligned shapes - Output: (16033,), Label: (16033,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 131/200 [00:01<00:00, 130.81it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 158.79it/s]
Table1-only metrics completed in 2.04s
Evaluation Result:  {'AUC-PR': 0.14990760046810914, 'AUC-ROC': 0.796843143523672, 'VUS-PR': 0.09121291208318523, 'VUS-ROC': 0.900113749954962, 'Standard-F1': 0.13003095975182122, 'Standard-Precision': 0.13043478260869484, 'Standard-Recall': 0.12962962962962885, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.900045984623181, 'Affiliation-P': 0.8218502175872675, 'Affiliation-R': 0.9946864269455853, 'F1_T': 0.13003095975182122, 'Precision_T': 0.13043478260869484, 'Recall_T': 0.12962962962962885}
Results for 125_WSD_id_97_WebService_tr_2217_1st_2317.csv: {'filename': '125_WSD_id_97_WebService_tr_2217_1st_2317.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '2217', 'data_shape': '18250x1', 'output_length': 20000, 'label_length': 16033, 'aligned_length': 16033, 'AUC-PR': 0.14990760046810914, 'AUC-ROC': 0.796843143523672, 'VUS-PR': 0.09121291208318523, 'VUS-ROC': 0.900113749954962, 'Standard-F1': 0.13003095975182122, 'Standard-Precision': 0.13043478260869484, 'Standard-Recall': 0.12962962962962885, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.900045984623181, 'Affiliation-P': 0.8218502175872675, 'Affiliation-R': 0.9946864269455853, 'F1_T': 0.13003095975182122, 'Precision_T': 0.13043478260869484, 'Recall_T': 0.12962962962962885}
Logits metrics skipped for 125_WSD_id_97_WebService_tr_2217_1st_2317.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/125_WSD_id_97_WebService_tr_2217_1st_2317_results.pkl
Processing file: 325_UCR_id_23_Medical_tr_3000_1st_6168.csv, Data shape: (12000, 1)
Label shape: (12000,)
Train data shape: (3000, 1), Test data shape: (9000, 1), Label test shape: (9000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 325_UCR_id_23_Medical_tr_3000_1st_6168.csv...
Using CUDA device: 2
Data shape: (9000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9000, window_size 5000, adding 1000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.89it/s]
Original shapes - Output: (10000,), Label: (9000,)
Aligned shapes - Output: (9000,), Label: (9000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 512.05it/s]
Table1-only metrics completed in 1.27s
Evaluation Result:  {'AUC-PR': 0.03245642063669648, 'AUC-ROC': 0.9316836858987373, 'VUS-PR': 0.03546755833001213, 'VUS-ROC': 0.933247261915515, 'Standard-F1': 0.07894736842096063, 'Standard-Precision': 0.04147465437788014, 'Standard-Recall': 0.8181818181817996, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7303273329516933, 'Affiliation-P': 0.6393140096618357, 'Affiliation-R': 0.8515555555555555, 'F1_T': 0.0781990521326024, 'Precision_T': 0.041249999999999946, 'Recall_T': 0.7499999999999829}
Results for 325_UCR_id_23_Medical_tr_3000_1st_6168.csv: {'filename': '325_UCR_id_23_Medical_tr_3000_1st_6168.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 218, 'train_index': '3000', 'data_shape': '12000x1', 'output_length': 10000, 'label_length': 9000, 'aligned_length': 9000, 'AUC-PR': 0.03245642063669648, 'AUC-ROC': 0.9316836858987373, 'VUS-PR': 0.03546755833001213, 'VUS-ROC': 0.933247261915515, 'Standard-F1': 0.07894736842096063, 'Standard-Precision': 0.04147465437788014, 'Standard-Recall': 0.8181818181817996, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7303273329516933, 'Affiliation-P': 0.6393140096618357, 'Affiliation-R': 0.8515555555555555, 'F1_T': 0.0781990521326024, 'Precision_T': 0.041249999999999946, 'Recall_T': 0.7499999999999829}
Logits metrics skipped for 325_UCR_id_23_Medical_tr_3000_1st_6168.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/325_UCR_id_23_Medical_tr_3000_1st_6168_results.pkl
Processing file: 102_WSD_id_74_WebService_tr_3306_1st_3639.csv, Data shape: (13227, 1)
Label shape: (13227,)
Train data shape: (3306, 1), Test data shape: (9921, 1), Label test shape: (9921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 102_WSD_id_74_WebService_tr_3306_1st_3639.csv...
Using CUDA device: 2
Data shape: (9921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9921, window_size 5000, adding 79 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 168.43it/s]
Original shapes - Output: (10000,), Label: (9921,)
Aligned shapes - Output: (9921,), Label: (9921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 363.53it/s]
Table1-only metrics completed in 0.69s
Evaluation Result:  {'AUC-PR': 0.322127576697516, 'AUC-ROC': 0.8308037146004434, 'VUS-PR': 0.09599647350663493, 'VUS-ROC': 0.8991147158427194, 'Standard-F1': 0.13953488372057957, 'Standard-Precision': 0.08999999999999911, 'Standard-Recall': 0.3103448275861962, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8406479709031267, 'Affiliation-P': 0.7262320551357097, 'Affiliation-R': 0.9978577946390841, 'F1_T': 0.13953488372057957, 'Precision_T': 0.08999999999999911, 'Recall_T': 0.3103448275861962}
Results for 102_WSD_id_74_WebService_tr_3306_1st_3639.csv: {'filename': '102_WSD_id_74_WebService_tr_3306_1st_3639.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 10, 'train_index': '3306', 'data_shape': '13227x1', 'output_length': 10000, 'label_length': 9921, 'aligned_length': 9921, 'AUC-PR': 0.322127576697516, 'AUC-ROC': 0.8308037146004434, 'VUS-PR': 0.09599647350663493, 'VUS-ROC': 0.8991147158427194, 'Standard-F1': 0.13953488372057957, 'Standard-Precision': 0.08999999999999911, 'Standard-Recall': 0.3103448275861962, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8406479709031267, 'Affiliation-P': 0.7262320551357097, 'Affiliation-R': 0.9978577946390841, 'F1_T': 0.13953488372057957, 'Precision_T': 0.08999999999999911, 'Recall_T': 0.3103448275861962}
Logits metrics skipped for 102_WSD_id_74_WebService_tr_3306_1st_3639.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/102_WSD_id_74_WebService_tr_3306_1st_3639_results.pkl
Processing file: 447_UCR_id_145_Medical_tr_9625_1st_12240.csv, Data shape: (38501, 1)
Label shape: (38501,)
Train data shape: (9625, 1), Test data shape: (28876, 1), Label test shape: (28876,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 447_UCR_id_145_Medical_tr_9625_1st_12240.csv...
Using CUDA device: 2
Data shape: (28876, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 28876, window_size 5000, adding 1124 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 142.66it/s]
Original shapes - Output: (30000,), Label: (28876,)
Aligned shapes - Output: (28876,), Label: (28876,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [00:01<00:00, 126.77it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 143.63it/s]
Table1-only metrics completed in 1.72s
Evaluation Result:  {'AUC-PR': 0.06623338258289913, 'AUC-ROC': 0.8572332848718494, 'VUS-PR': 0.06506272163801313, 'VUS-ROC': 0.8830469803444365, 'Standard-F1': 0.11764705882330422, 'Standard-Precision': 0.06753812636165563, 'Standard-Recall': 0.4558823529411698, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7728784632913293, 'Affiliation-P': 0.6300256220741688, 'Affiliation-R': 0.9995082930664995, 'F1_T': 0.12424849699375207, 'Precision_T': 0.07192575406032466, 'Recall_T': 0.4558823529411698}
Results for 447_UCR_id_145_Medical_tr_9625_1st_12240.csv: {'filename': '447_UCR_id_145_Medical_tr_9625_1st_12240.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 40, 'train_index': '9625', 'data_shape': '38501x1', 'output_length': 30000, 'label_length': 28876, 'aligned_length': 28876, 'AUC-PR': 0.06623338258289913, 'AUC-ROC': 0.8572332848718494, 'VUS-PR': 0.06506272163801313, 'VUS-ROC': 0.8830469803444365, 'Standard-F1': 0.11764705882330422, 'Standard-Precision': 0.06753812636165563, 'Standard-Recall': 0.4558823529411698, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7728784632913293, 'Affiliation-P': 0.6300256220741688, 'Affiliation-R': 0.9995082930664995, 'F1_T': 0.12424849699375207, 'Precision_T': 0.07192575406032466, 'Recall_T': 0.4558823529411698}
Logits metrics skipped for 447_UCR_id_145_Medical_tr_9625_1st_12240.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/447_UCR_id_145_Medical_tr_9625_1st_12240_results.pkl
Processing file: 618_YAHOO_id_68_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 618_YAHOO_id_68_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 162.11it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2406.12it/s]
Table1-only metrics completed in 0.52s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9711796246648787, 'Affiliation-P': 0.9439739413680781, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 618_YAHOO_id_68_Synthetic_tr_500_1st_893.csv: {'filename': '618_YAHOO_id_68_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 134, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9711796246648787, 'Affiliation-P': 0.9439739413680781, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 618_YAHOO_id_68_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/618_YAHOO_id_68_Synthetic_tr_500_1st_893_results.pkl
Processing file: 358_UCR_id_56_Facility_tr_10500_1st_29080.csv, Data shape: (42000, 1)
Label shape: (42000,)
Train data shape: (10500, 1), Test data shape: (31500, 1), Label test shape: (31500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 358_UCR_id_56_Facility_tr_10500_1st_29080.csv...
Using CUDA device: 2
Data shape: (31500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 31500, window_size 5000, adding 3500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([7, 5000, 1])
batched_mask shape: torch.Size([7, 5000])

100%|██████████| 1/1 [00:00<00:00, 136.04it/s]
Original shapes - Output: (35000,), Label: (31500,)
Aligned shapes - Output: (31500,), Label: (31500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  58%|█████▊    | 116/200 [00:01<00:00, 115.69it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 137.31it/s]
Table1-only metrics completed in 1.79s
Evaluation Result:  {'AUC-PR': 0.007554849781827621, 'AUC-ROC': 0.8819126378286684, 'VUS-PR': 0.009082346336455056, 'VUS-ROC': 0.8873396668425101, 'Standard-F1': 0.022420672620156058, 'Standard-Precision': 0.011339611762444742, 'Standard-Recall': 0.983333333333317, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7877929104363248, 'Affiliation-P': 0.6502558297713913, 'Affiliation-R': 0.9991187830687831, 'F1_T': 0.022376543209853654, 'Precision_T': 0.011319281811085088, 'Recall_T': 0.9666666666666506}
Results for 358_UCR_id_56_Facility_tr_10500_1st_29080.csv: {'filename': '358_UCR_id_56_Facility_tr_10500_1st_29080.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 37, 'train_index': '10500', 'data_shape': '42000x1', 'output_length': 35000, 'label_length': 31500, 'aligned_length': 31500, 'AUC-PR': 0.007554849781827621, 'AUC-ROC': 0.8819126378286684, 'VUS-PR': 0.009082346336455056, 'VUS-ROC': 0.8873396668425101, 'Standard-F1': 0.022420672620156058, 'Standard-Precision': 0.011339611762444742, 'Standard-Recall': 0.983333333333317, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7877929104363248, 'Affiliation-P': 0.6502558297713913, 'Affiliation-R': 0.9991187830687831, 'F1_T': 0.022376543209853654, 'Precision_T': 0.011319281811085088, 'Recall_T': 0.9666666666666506}
Logits metrics skipped for 358_UCR_id_56_Facility_tr_10500_1st_29080.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/358_UCR_id_56_Facility_tr_10500_1st_29080_results.pkl
Skipping file: 532_SMAP_id_2_Sensor_tr_2075_1st_5550.csv due to filter criteria.
Processing file: 034_WSD_id_6_WebService_tr_4559_1st_8461.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 034_WSD_id_6_WebService_tr_4559_1st_8461.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.47it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 264.99it/s]
Table1-only metrics completed in 1.70s
Evaluation Result:  {'AUC-PR': 0.6729935366036764, 'AUC-ROC': 0.9763708175140342, 'VUS-PR': 0.7645828969610553, 'VUS-ROC': 0.9921737057320009, 'Standard-F1': 0.7102272727267709, 'Standard-Precision': 0.7621951219512149, 'Standard-Recall': 0.664893617021273, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9607606190126613, 'Affiliation-P': 0.9272400298694734, 'Affiliation-R': 0.9967957106838323, 'F1_T': 0.7015384615379696, 'Precision_T': 0.8321167883211619, 'Recall_T': 0.6063829787234011}
Results for 034_WSD_id_6_WebService_tr_4559_1st_8461.csv: {'filename': '034_WSD_id_6_WebService_tr_4559_1st_8461.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.6729935366036764, 'AUC-ROC': 0.9763708175140342, 'VUS-PR': 0.7645828969610553, 'VUS-ROC': 0.9921737057320009, 'Standard-F1': 0.7102272727267709, 'Standard-Precision': 0.7621951219512149, 'Standard-Recall': 0.664893617021273, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9607606190126613, 'Affiliation-P': 0.9272400298694734, 'Affiliation-R': 0.9967957106838323, 'F1_T': 0.7015384615379696, 'Precision_T': 0.8321167883211619, 'Recall_T': 0.6063829787234011}
Logits metrics skipped for 034_WSD_id_6_WebService_tr_4559_1st_8461.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/034_WSD_id_6_WebService_tr_4559_1st_8461_results.pkl
Processing file: 228_MGAB_id_4_Synthetic_tr_25000_1st_38948.csv, Data shape: (100000, 1)
Label shape: (100000,)
Train data shape: (25000, 1), Test data shape: (75000, 1), Label test shape: (75000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 228_MGAB_id_4_Synthetic_tr_25000_1st_38948.csv...
Using CUDA device: 2
Data shape: (75000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([15, 5000, 1])
batched_mask shape: torch.Size([15, 5000])

100%|██████████| 1/1 [00:00<00:00, 79.20it/s]
Original shapes - Output: (75000,), Label: (75000,)
Aligned shapes - Output: (75000,), Label: (75000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  18%|█▊        | 35/200 [00:01<00:04, 34.08it/s]

Calculating Affiliation Metrics:  35%|███▌      | 70/200 [00:02<00:04, 26.70it/s]

Calculating Affiliation Metrics:  49%|████▉     | 98/200 [00:03<00:04, 24.52it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 124/200 [00:04<00:03, 24.23it/s]

Calculating Affiliation Metrics:  76%|███████▋  | 153/200 [00:05<00:01, 25.52it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:06<00:00, 28.88it/s]
Table1-only metrics completed in 8.81s
Evaluation Result:  {'AUC-PR': 0.016579868048315433, 'AUC-ROC': 0.7808413435828878, 'VUS-PR': 0.020246448377575, 'VUS-ROC': 0.8080564669582884, 'Standard-F1': 0.06526315789440429, 'Standard-Precision': 0.04133333333333328, 'Standard-Recall': 0.15499999999999922, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7912669919775897, 'Affiliation-P': 0.6986293176037403, 'Affiliation-R': 0.9122276269376803, 'F1_T': 0.06526315789440429, 'Precision_T': 0.04133333333333328, 'Recall_T': 0.15499999999999922}
Results for 228_MGAB_id_4_Synthetic_tr_25000_1st_38948.csv: {'filename': '228_MGAB_id_4_Synthetic_tr_25000_1st_38948.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 48, 'train_index': '25000', 'data_shape': '100000x1', 'output_length': 75000, 'label_length': 75000, 'aligned_length': 75000, 'AUC-PR': 0.016579868048315433, 'AUC-ROC': 0.7808413435828878, 'VUS-PR': 0.020246448377575, 'VUS-ROC': 0.8080564669582884, 'Standard-F1': 0.06526315789440429, 'Standard-Precision': 0.04133333333333328, 'Standard-Recall': 0.15499999999999922, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7912669919775897, 'Affiliation-P': 0.6986293176037403, 'Affiliation-R': 0.9122276269376803, 'F1_T': 0.06526315789440429, 'Precision_T': 0.04133333333333328, 'Recall_T': 0.15499999999999922}
Logits metrics skipped for 228_MGAB_id_4_Synthetic_tr_25000_1st_38948.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/228_MGAB_id_4_Synthetic_tr_25000_1st_38948_results.pkl
Processing file: 476_UCR_id_174_Facility_tr_50000_1st_64632.csv, Data shape: (299840, 1)
Label shape: (299840,)
Train data shape: (50000, 1), Test data shape: (249840, 1), Label test shape: (249840,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 476_UCR_id_174_Facility_tr_50000_1st_64632.csv...
Using CUDA device: 2
Data shape: (249840, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 249840, window_size 5000, adding 160 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([50, 5000, 1])
batched_mask shape: torch.Size([50, 5000])

100%|██████████| 1/1 [00:00<00:00, 28.41it/s]
Original shapes - Output: (250000,), Label: (249840,)
Aligned shapes - Output: (249840,), Label: (249840,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   9%|▉         | 18/200 [00:01<00:10, 17.90it/s]

Calculating Affiliation Metrics:  18%|█▊        | 36/200 [00:02<00:10, 15.60it/s]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:03<00:10, 14.40it/s]

Calculating Affiliation Metrics:  34%|███▎      | 67/200 [00:04<00:09, 13.79it/s]

Calculating Affiliation Metrics:  40%|████      | 81/200 [00:05<00:09, 13.16it/s]

Calculating Affiliation Metrics:  48%|████▊     | 95/200 [00:06<00:07, 13.14it/s]

Calculating Affiliation Metrics:  55%|█████▌    | 110/200 [00:07<00:06, 13.46it/s]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [00:08<00:05, 14.18it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [00:10<00:03, 15.39it/s]

Calculating Affiliation Metrics:  82%|████████▎ | 165/200 [00:11<00:02, 16.72it/s]

Calculating Affiliation Metrics:  97%|█████████▋| 194/200 [00:12<00:00, 20.23it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:12<00:00, 16.43it/s]
Table1-only metrics completed in 20.81s
Evaluation Result:  {'AUC-PR': 0.1485653621213903, 'AUC-ROC': 0.9896659832035604, 'VUS-PR': 0.0882170558394711, 'VUS-ROC': 0.9893442615460126, 'Standard-F1': 0.1162191982344963, 'Standard-Precision': 0.0632252901160464, 'Standard-Recall': 0.7181818181818149, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7184207664486352, 'Affiliation-P': 0.5605875132696391, 'Affiliation-R': 0.9999589100660786, 'F1_T': 0.1162191982344963, 'Precision_T': 0.0632252901160464, 'Recall_T': 0.7181818181818149}
Results for 476_UCR_id_174_Facility_tr_50000_1st_64632.csv: {'filename': '476_UCR_id_174_Facility_tr_50000_1st_64632.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 124, 'train_index': '50000', 'data_shape': '299840x1', 'output_length': 250000, 'label_length': 249840, 'aligned_length': 249840, 'AUC-PR': 0.1485653621213903, 'AUC-ROC': 0.9896659832035604, 'VUS-PR': 0.0882170558394711, 'VUS-ROC': 0.9893442615460126, 'Standard-F1': 0.1162191982344963, 'Standard-Precision': 0.0632252901160464, 'Standard-Recall': 0.7181818181818149, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7184207664486352, 'Affiliation-P': 0.5605875132696391, 'Affiliation-R': 0.9999589100660786, 'F1_T': 0.1162191982344963, 'Precision_T': 0.0632252901160464, 'Recall_T': 0.7181818181818149}
Logits metrics skipped for 476_UCR_id_174_Facility_tr_50000_1st_64632.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/476_UCR_id_174_Facility_tr_50000_1st_64632_results.pkl
Processing file: 002_NAB_id_2_WebService_tr_1500_1st_4106.csv, Data shape: (6000, 1)
Label shape: (6000,)
Train data shape: (1500, 1), Test data shape: (4500, 1), Label test shape: (4500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 002_NAB_id_2_WebService_tr_1500_1st_4106.csv...
Using CUDA device: 2
Data shape: (4500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 4500, 1])
batched_mask shape: torch.Size([1, 4500])

100%|██████████| 1/1 [00:00<00:00, 59.73it/s]
Original shapes - Output: (4500,), Label: (4500,)
Aligned shapes - Output: (4500,), Label: (4500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 676.44it/s]
Table1-only metrics completed in 0.82s
Evaluation Result:  {'AUC-PR': 0.2242600498059015, 'AUC-ROC': 0.49259264828215676, 'VUS-PR': 0.237448505148488, 'VUS-ROC': 0.5288321955932875, 'Standard-F1': 0.26135831381707625, 'Standard-Precision': 0.15359207266721714, 'Standard-Recall': 0.875981161695446, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9478149838075468, 'Affiliation-P': 0.927463768115942, 'Affiliation-R': 0.9690793650793651, 'F1_T': 0.2612359550559259, 'Precision_T': 0.15350756533700133, 'Recall_T': 0.875981161695446}
Results for 002_NAB_id_2_WebService_tr_1500_1st_4106.csv: {'filename': '002_NAB_id_2_WebService_tr_1500_1st_4106.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '1500', 'data_shape': '6000x1', 'output_length': 4500, 'label_length': 4500, 'aligned_length': 4500, 'AUC-PR': 0.2242600498059015, 'AUC-ROC': 0.49259264828215676, 'VUS-PR': 0.237448505148488, 'VUS-ROC': 0.5288321955932875, 'Standard-F1': 0.26135831381707625, 'Standard-Precision': 0.15359207266721714, 'Standard-Recall': 0.875981161695446, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9478149838075468, 'Affiliation-P': 0.927463768115942, 'Affiliation-R': 0.9690793650793651, 'F1_T': 0.2612359550559259, 'Precision_T': 0.15350756533700133, 'Recall_T': 0.875981161695446}
Logits metrics skipped for 002_NAB_id_2_WebService_tr_1500_1st_4106.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/002_NAB_id_2_WebService_tr_1500_1st_4106_results.pkl
Skipping file: 832_Exathlon_id_23_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 418_UCR_id_116_Facility_tr_10000_1st_35774.csv, Data shape: (40000, 1)
Label shape: (40000,)
Train data shape: (10000, 1), Test data shape: (30000, 1), Label test shape: (30000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 418_UCR_id_116_Facility_tr_10000_1st_35774.csv...
Using CUDA device: 2
Data shape: (30000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 143.38it/s]
Original shapes - Output: (30000,), Label: (30000,)
Aligned shapes - Output: (30000,), Label: (30000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  60%|██████    | 121/200 [00:01<00:00, 120.66it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 133.69it/s]
Table1-only metrics completed in 2.21s
Evaluation Result:  {'AUC-PR': 0.02006501852496715, 'AUC-ROC': 0.7010327759197325, 'VUS-PR': 0.027570372671711196, 'VUS-ROC': 0.7981701902341318, 'Standard-F1': 0.0825688073392463, 'Standard-Precision': 0.04663212435233154, 'Standard-Recall': 0.35999999999999643, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6863929303210116, 'Affiliation-P': 0.522671037414966, 'Affiliation-R': 0.9994666666666666, 'F1_T': 0.08072653884946521, 'Precision_T': 0.04489337822671151, 'Recall_T': 0.399999999999996}
Results for 418_UCR_id_116_Facility_tr_10000_1st_35774.csv: {'filename': '418_UCR_id_116_Facility_tr_10000_1st_35774.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 108, 'train_index': '10000', 'data_shape': '40000x1', 'output_length': 30000, 'label_length': 30000, 'aligned_length': 30000, 'AUC-PR': 0.02006501852496715, 'AUC-ROC': 0.7010327759197325, 'VUS-PR': 0.027570372671711196, 'VUS-ROC': 0.7981701902341318, 'Standard-F1': 0.0825688073392463, 'Standard-Precision': 0.04663212435233154, 'Standard-Recall': 0.35999999999999643, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6863929303210116, 'Affiliation-P': 0.522671037414966, 'Affiliation-R': 0.9994666666666666, 'F1_T': 0.08072653884946521, 'Precision_T': 0.04489337822671151, 'Recall_T': 0.399999999999996}
Logits metrics skipped for 418_UCR_id_116_Facility_tr_10000_1st_35774.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/418_UCR_id_116_Facility_tr_10000_1st_35774_results.pkl
Skipping file: 212_SMD_id_35_Facility_tr_5925_1st_17580.csv due to filter criteria.
Processing file: 101_WSD_id_73_WebService_tr_1513_1st_1613.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (1513, 1), Test data shape: (16723, 1), Label test shape: (16723,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 101_WSD_id_73_WebService_tr_1513_1st_1613.csv...
Using CUDA device: 2
Data shape: (16723, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 16723, window_size 5000, adding 3277 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.89it/s]
Original shapes - Output: (20000,), Label: (16723,)
Aligned shapes - Output: (16723,), Label: (16723,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 262.46it/s]
Table1-only metrics completed in 1.36s
Evaluation Result:  {'AUC-PR': 0.11268039824329651, 'AUC-ROC': 0.9699422816021762, 'VUS-PR': 0.091181541301687, 'VUS-ROC': 0.97432436035014, 'Standard-F1': 0.17015706806265818, 'Standard-Precision': 0.09379509379509365, 'Standard-Recall': 0.915492957746466, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.763642438370015, 'Affiliation-P': 0.6176550069894403, 'Affiliation-R': 1.0, 'F1_T': 0.16916780354689143, 'Precision_T': 0.09365558912386693, 'Recall_T': 0.873239436619706}
Results for 101_WSD_id_73_WebService_tr_1513_1st_1613.csv: {'filename': '101_WSD_id_73_WebService_tr_1513_1st_1613.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '1513', 'data_shape': '18236x1', 'output_length': 20000, 'label_length': 16723, 'aligned_length': 16723, 'AUC-PR': 0.11268039824329651, 'AUC-ROC': 0.9699422816021762, 'VUS-PR': 0.091181541301687, 'VUS-ROC': 0.97432436035014, 'Standard-F1': 0.17015706806265818, 'Standard-Precision': 0.09379509379509365, 'Standard-Recall': 0.915492957746466, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.763642438370015, 'Affiliation-P': 0.6176550069894403, 'Affiliation-R': 1.0, 'F1_T': 0.16916780354689143, 'Precision_T': 0.09365558912386693, 'Recall_T': 0.873239436619706}
Logits metrics skipped for 101_WSD_id_73_WebService_tr_1513_1st_1613.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/101_WSD_id_73_WebService_tr_1513_1st_1613_results.pkl
Processing file: 431_UCR_id_129_Facility_tr_32500_1st_114283.csv, Data shape: (130001, 1)
Label shape: (130001,)
Train data shape: (32500, 1), Test data shape: (97501, 1), Label test shape: (97501,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 431_UCR_id_129_Facility_tr_32500_1st_114283.csv...
Using CUDA device: 2
Data shape: (97501, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 97501, window_size 5000, adding 2499 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([20, 5000, 1])
batched_mask shape: torch.Size([20, 5000])

100%|██████████| 1/1 [00:00<00:00, 64.60it/s]
Original shapes - Output: (100000,), Label: (97501,)
Aligned shapes - Output: (97501,), Label: (97501,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  20%|█▉        | 39/200 [00:01<00:04, 38.21it/s]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [00:02<00:03, 37.73it/s]

Calculating Affiliation Metrics:  61%|██████    | 122/200 [00:03<00:01, 40.51it/s]

Calculating Affiliation Metrics:  92%|█████████▎| 185/200 [00:04<00:00, 49.07it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:04<00:00, 47.23it/s]
Table1-only metrics completed in 10.37s
Evaluation Result:  {'AUC-PR': 0.8664883333531394, 'AUC-ROC': 0.9991121429615271, 'VUS-PR': 0.2247543195984253, 'VUS-ROC': 0.9984117031676192, 'Standard-F1': 0.12464046021080956, 'Standard-Precision': 0.06659836065573764, 'Standard-Recall': 0.9701492537313289, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7143021524101337, 'Affiliation-P': 0.5555766251690982, 'Affiliation-R': 0.9999961730204986, 'F1_T': 0.12464046021080956, 'Precision_T': 0.06659836065573764, 'Recall_T': 0.9701492537313289}
Results for 431_UCR_id_129_Facility_tr_32500_1st_114283.csv: {'filename': '431_UCR_id_129_Facility_tr_32500_1st_114283.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 243, 'train_index': '32500', 'data_shape': '130001x1', 'output_length': 100000, 'label_length': 97501, 'aligned_length': 97501, 'AUC-PR': 0.8664883333531394, 'AUC-ROC': 0.9991121429615271, 'VUS-PR': 0.2247543195984253, 'VUS-ROC': 0.9984117031676192, 'Standard-F1': 0.12464046021080956, 'Standard-Precision': 0.06659836065573764, 'Standard-Recall': 0.9701492537313289, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7143021524101337, 'Affiliation-P': 0.5555766251690982, 'Affiliation-R': 0.9999961730204986, 'F1_T': 0.12464046021080956, 'Precision_T': 0.06659836065573764, 'Recall_T': 0.9701492537313289}
Logits metrics skipped for 431_UCR_id_129_Facility_tr_32500_1st_114283.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/431_UCR_id_129_Facility_tr_32500_1st_114283_results.pkl
Processing file: 809_YAHOO_id_259_WebService_tr_500_1st_123.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 809_YAHOO_id_259_WebService_tr_500_1st_123.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 159.06it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 901.31it/s]
Table1-only metrics completed in 0.68s
Evaluation Result:  {'AUC-PR': 0.9035714285714286, 'AUC-ROC': 0.9971203071672355, 'VUS-PR': 0.8201147298075928, 'VUS-ROC': 0.9980679140163987, 'Standard-F1': 0.6999999999994501, 'Standard-Precision': 0.5833333333332847, 'Standard-Recall': 0.8749999999998906, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9874351471636867, 'Affiliation-P': 0.975182127246236, 'Affiliation-R': 1.0, 'F1_T': 0.6999999999994501, 'Precision_T': 0.5833333333332847, 'Recall_T': 0.8749999999998906}
Results for 809_YAHOO_id_259_WebService_tr_500_1st_123.csv: {'filename': '809_YAHOO_id_259_WebService_tr_500_1st_123.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9035714285714286, 'AUC-ROC': 0.9971203071672355, 'VUS-PR': 0.8201147298075928, 'VUS-ROC': 0.9980679140163987, 'Standard-F1': 0.6999999999994501, 'Standard-Precision': 0.5833333333332847, 'Standard-Recall': 0.8749999999998906, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9874351471636867, 'Affiliation-P': 0.975182127246236, 'Affiliation-R': 1.0, 'F1_T': 0.6999999999994501, 'Precision_T': 0.5833333333332847, 'Recall_T': 0.8749999999998906}
Logits metrics skipped for 809_YAHOO_id_259_WebService_tr_500_1st_123.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/809_YAHOO_id_259_WebService_tr_500_1st_123_results.pkl
Processing file: 637_YAHOO_id_87_WebService_tr_500_1st_338.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 637_YAHOO_id_87_WebService_tr_500_1st_338.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 163.98it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1206.01it/s]
Table1-only metrics completed in 0.53s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9737413771279045, 'VUS-ROC': 0.9998880879317802, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Results for 637_YAHOO_id_87_WebService_tr_500_1st_338.csv: {'filename': '637_YAHOO_id_87_WebService_tr_500_1st_338.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9737413771279045, 'VUS-ROC': 0.9998880879317802, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Logits metrics skipped for 637_YAHOO_id_87_WebService_tr_500_1st_338.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/637_YAHOO_id_87_WebService_tr_500_1st_338_results.pkl
Processing file: 300_TODS_id_14_Synthetic_tr_1250_1st_2555.csv, Data shape: (5000, 1)
Label shape: (5000,)
Train data shape: (1250, 1), Test data shape: (3750, 1), Label test shape: (3750,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 300_TODS_id_14_Synthetic_tr_1250_1st_2555.csv...
Using CUDA device: 2
Data shape: (3750, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 3750, 1])
batched_mask shape: torch.Size([1, 3750])

100%|██████████| 1/1 [00:00<00:00, 117.61it/s]
Original shapes - Output: (3750,), Label: (3750,)
Aligned shapes - Output: (3750,), Label: (3750,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 334.09it/s]
Table1-only metrics completed in 1.44s
Evaluation Result:  {'AUC-PR': 0.9990696859794775, 'AUC-ROC': 0.9998828358208955, 'VUS-PR': 0.999909179274038, 'VUS-ROC': 0.9999927517801622, 'Standard-F1': 0.9837296620770944, 'Standard-Precision': 0.984962406015035, 'Standard-Recall': 0.9824999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9998779395178787, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9997559088296445, 'F1_T': 0.9814126394047019, 'Precision_T': 0.9729729729729705, 'Recall_T': 0.9899999999999974}
Results for 300_TODS_id_14_Synthetic_tr_1250_1st_2555.csv: {'filename': '300_TODS_id_14_Synthetic_tr_1250_1st_2555.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '1250', 'data_shape': '5000x1', 'output_length': 3750, 'label_length': 3750, 'aligned_length': 3750, 'AUC-PR': 0.9990696859794775, 'AUC-ROC': 0.9998828358208955, 'VUS-PR': 0.999909179274038, 'VUS-ROC': 0.9999927517801622, 'Standard-F1': 0.9837296620770944, 'Standard-Precision': 0.984962406015035, 'Standard-Recall': 0.9824999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9998779395178787, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9997559088296445, 'F1_T': 0.9814126394047019, 'Precision_T': 0.9729729729729705, 'Recall_T': 0.9899999999999974}
Logits metrics skipped for 300_TODS_id_14_Synthetic_tr_1250_1st_2555.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/300_TODS_id_14_Synthetic_tr_1250_1st_2555_results.pkl
Processing file: 097_WSD_id_69_WebService_tr_4520_1st_6465.csv, Data shape: (18083, 1)
Label shape: (18083,)
Train data shape: (4520, 1), Test data shape: (13563, 1), Label test shape: (13563,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 097_WSD_id_69_WebService_tr_4520_1st_6465.csv...
Using CUDA device: 2
Data shape: (13563, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13563, window_size 5000, adding 1437 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.95it/s]
Original shapes - Output: (15000,), Label: (13563,)
Aligned shapes - Output: (13563,), Label: (13563,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 286.93it/s]
Table1-only metrics completed in 2.11s
Evaluation Result:  {'AUC-PR': 0.360486891954663, 'AUC-ROC': 0.7925360939822069, 'VUS-PR': 0.5630055153901761, 'VUS-ROC': 0.9808066674254134, 'Standard-F1': 0.2111111111107394, 'Standard-Precision': 0.13970588235294015, 'Standard-Recall': 0.431818181818172, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9877801180860433, 'Affiliation-P': 0.9771961256015171, 'Affiliation-R': 0.9985958910191591, 'F1_T': 0.2111111111107394, 'Precision_T': 0.13970588235294015, 'Recall_T': 0.431818181818172}
Results for 097_WSD_id_69_WebService_tr_4520_1st_6465.csv: {'filename': '097_WSD_id_69_WebService_tr_4520_1st_6465.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4520', 'data_shape': '18083x1', 'output_length': 15000, 'label_length': 13563, 'aligned_length': 13563, 'AUC-PR': 0.360486891954663, 'AUC-ROC': 0.7925360939822069, 'VUS-PR': 0.5630055153901761, 'VUS-ROC': 0.9808066674254134, 'Standard-F1': 0.2111111111107394, 'Standard-Precision': 0.13970588235294015, 'Standard-Recall': 0.431818181818172, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9877801180860433, 'Affiliation-P': 0.9771961256015171, 'Affiliation-R': 0.9985958910191591, 'F1_T': 0.2111111111107394, 'Precision_T': 0.13970588235294015, 'Recall_T': 0.431818181818172}
Logits metrics skipped for 097_WSD_id_69_WebService_tr_4520_1st_6465.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/097_WSD_id_69_WebService_tr_4520_1st_6465_results.pkl
Processing file: 077_WSD_id_49_WebService_tr_3310_1st_3716.csv, Data shape: (13243, 1)
Label shape: (13243,)
Train data shape: (3310, 1), Test data shape: (9933, 1), Label test shape: (9933,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 077_WSD_id_49_WebService_tr_3310_1st_3716.csv...
Using CUDA device: 2
Data shape: (9933, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9933, window_size 5000, adding 67 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.12it/s]
Original shapes - Output: (10000,), Label: (9933,)
Aligned shapes - Output: (9933,), Label: (9933,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 334.23it/s]
Table1-only metrics completed in 2.50s
Evaluation Result:  {'AUC-PR': 0.3267197845982554, 'AUC-ROC': 0.914819951730791, 'VUS-PR': 0.24917009166634296, 'VUS-ROC': 0.977818367736918, 'Standard-F1': 0.23899371069135417, 'Standard-Precision': 0.18999999999999811, 'Standard-Recall': 0.3220338983050793, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9192758849172783, 'Affiliation-P': 0.8524909411107705, 'Affiliation-R': 0.9974142579569646, 'F1_T': 0.23899371069135417, 'Precision_T': 0.18999999999999811, 'Recall_T': 0.3220338983050793}
Results for 077_WSD_id_49_WebService_tr_3310_1st_3716.csv: {'filename': '077_WSD_id_49_WebService_tr_3310_1st_3716.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '3310', 'data_shape': '13243x1', 'output_length': 10000, 'label_length': 9933, 'aligned_length': 9933, 'AUC-PR': 0.3267197845982554, 'AUC-ROC': 0.914819951730791, 'VUS-PR': 0.24917009166634296, 'VUS-ROC': 0.977818367736918, 'Standard-F1': 0.23899371069135417, 'Standard-Precision': 0.18999999999999811, 'Standard-Recall': 0.3220338983050793, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9192758849172783, 'Affiliation-P': 0.8524909411107705, 'Affiliation-R': 0.9974142579569646, 'F1_T': 0.23899371069135417, 'Precision_T': 0.18999999999999811, 'Recall_T': 0.3220338983050793}
Logits metrics skipped for 077_WSD_id_49_WebService_tr_3310_1st_3716.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/077_WSD_id_49_WebService_tr_3310_1st_3716_results.pkl
Processing file: 623_YAHOO_id_73_WebService_tr_500_1st_117.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 623_YAHOO_id_73_WebService_tr_500_1st_117.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 155.34it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 930.47it/s]
Table1-only metrics completed in 0.78s
Evaluation Result:  {'AUC-PR': 0.5507436180999238, 'AUC-ROC': 0.7043315965471654, 'VUS-PR': 0.509220739350414, 'VUS-ROC': 0.7408454129547166, 'Standard-F1': 0.5217391304342381, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.5454545454544959, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7131175048953955, 'Affiliation-P': 0.5541434494665579, 'Affiliation-R': 1.0, 'F1_T': 0.5217391304342381, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.5454545454544959}
Results for 623_YAHOO_id_73_WebService_tr_500_1st_117.csv: {'filename': '623_YAHOO_id_73_WebService_tr_500_1st_117.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.5507436180999238, 'AUC-ROC': 0.7043315965471654, 'VUS-PR': 0.509220739350414, 'VUS-ROC': 0.7408454129547166, 'Standard-F1': 0.5217391304342381, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.5454545454544959, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7131175048953955, 'Affiliation-P': 0.5541434494665579, 'Affiliation-R': 1.0, 'F1_T': 0.5217391304342381, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.5454545454544959}
Logits metrics skipped for 623_YAHOO_id_73_WebService_tr_500_1st_117.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/623_YAHOO_id_73_WebService_tr_500_1st_117_results.pkl
Processing file: 478_UCR_id_176_Medical_tr_7500_1st_17210.csv, Data shape: (30001, 1)
Label shape: (30001,)
Train data shape: (7500, 1), Test data shape: (22501, 1), Label test shape: (22501,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 478_UCR_id_176_Medical_tr_7500_1st_17210.csv...
Using CUDA device: 2
Data shape: (22501, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22501, window_size 5000, adding 2499 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 152.57it/s]
Original shapes - Output: (25000,), Label: (22501,)
Aligned shapes - Output: (22501,), Label: (22501,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [00:01<00:00, 114.01it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 126.62it/s]
Table1-only metrics completed in 1.83s
Evaluation Result:  {'AUC-PR': 0.4863582210594262, 'AUC-ROC': 0.9311558505189078, 'VUS-PR': 0.2513940483047216, 'VUS-ROC': 0.9615275729913212, 'Standard-F1': 0.2536231884054986, 'Standard-Precision': 0.15486725663716747, 'Standard-Recall': 0.699999999999986, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7104507103264331, 'Affiliation-P': 0.5509524073516876, 'Affiliation-R': 0.9999244478023199, 'F1_T': 0.2536231884054986, 'Precision_T': 0.15486725663716747, 'Recall_T': 0.699999999999986}
Results for 478_UCR_id_176_Medical_tr_7500_1st_17210.csv: {'filename': '478_UCR_id_176_Medical_tr_7500_1st_17210.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 33, 'train_index': '7500', 'data_shape': '30001x1', 'output_length': 25000, 'label_length': 22501, 'aligned_length': 22501, 'AUC-PR': 0.4863582210594262, 'AUC-ROC': 0.9311558505189078, 'VUS-PR': 0.2513940483047216, 'VUS-ROC': 0.9615275729913212, 'Standard-F1': 0.2536231884054986, 'Standard-Precision': 0.15486725663716747, 'Standard-Recall': 0.699999999999986, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7104507103264331, 'Affiliation-P': 0.5509524073516876, 'Affiliation-R': 0.9999244478023199, 'F1_T': 0.2536231884054986, 'Precision_T': 0.15486725663716747, 'Recall_T': 0.699999999999986}
Logits metrics skipped for 478_UCR_id_176_Medical_tr_7500_1st_17210.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/478_UCR_id_176_Medical_tr_7500_1st_17210_results.pkl
Skipping file: 242_SVDB_id_6_Medical_tr_10726_1st_10826.csv due to filter criteria.
Processing file: 350_UCR_id_48_Facility_tr_7482_1st_23405.csv, Data shape: (29931, 1)
Label shape: (29931,)
Train data shape: (7482, 1), Test data shape: (22449, 1), Label test shape: (22449,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 350_UCR_id_48_Facility_tr_7482_1st_23405.csv...
Using CUDA device: 2
Data shape: (22449, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22449, window_size 5000, adding 2551 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 152.09it/s]
Original shapes - Output: (25000,), Label: (22449,)
Aligned shapes - Output: (22449,), Label: (22449,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [00:01<00:00, 126.77it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 147.20it/s]
Table1-only metrics completed in 2.27s
Evaluation Result:  {'AUC-PR': 0.02899585870922633, 'AUC-ROC': 0.950302393826995, 'VUS-PR': 0.031237381031770633, 'VUS-ROC': 0.951216591156893, 'Standard-F1': 0.07446113651199388, 'Standard-Precision': 0.039067854694996546, 'Standard-Recall': 0.7916666666666557, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666822222785804, 'Affiliation-P': 0.5000175002675741, 'Affiliation-R': 1.0, 'F1_T': 0.07534699272958541, 'Precision_T': 0.039555863983344874, 'Recall_T': 0.7916666666666557}
Results for 350_UCR_id_48_Facility_tr_7482_1st_23405.csv: {'filename': '350_UCR_id_48_Facility_tr_7482_1st_23405.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 168, 'train_index': '7482', 'data_shape': '29931x1', 'output_length': 25000, 'label_length': 22449, 'aligned_length': 22449, 'AUC-PR': 0.02899585870922633, 'AUC-ROC': 0.950302393826995, 'VUS-PR': 0.031237381031770633, 'VUS-ROC': 0.951216591156893, 'Standard-F1': 0.07446113651199388, 'Standard-Precision': 0.039067854694996546, 'Standard-Recall': 0.7916666666666557, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666822222785804, 'Affiliation-P': 0.5000175002675741, 'Affiliation-R': 1.0, 'F1_T': 0.07534699272958541, 'Precision_T': 0.039555863983344874, 'Recall_T': 0.7916666666666557}
Logits metrics skipped for 350_UCR_id_48_Facility_tr_7482_1st_23405.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/350_UCR_id_48_Facility_tr_7482_1st_23405_results.pkl
Processing file: 604_YAHOO_id_54_WebService_tr_500_1st_116.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 604_YAHOO_id_54_WebService_tr_500_1st_116.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 165.30it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1234.44it/s]
Table1-only metrics completed in 0.48s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9767262940224073, 'VUS-ROC': 0.9999008051910141, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Results for 604_YAHOO_id_54_WebService_tr_500_1st_116.csv: {'filename': '604_YAHOO_id_54_WebService_tr_500_1st_116.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9767262940224073, 'VUS-ROC': 0.9999008051910141, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Logits metrics skipped for 604_YAHOO_id_54_WebService_tr_500_1st_116.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/604_YAHOO_id_54_WebService_tr_500_1st_116_results.pkl
Processing file: 133_WSD_id_105_WebService_tr_3934_1st_5937.csv, Data shape: (15736, 1)
Label shape: (15736,)
Train data shape: (3934, 1), Test data shape: (11802, 1), Label test shape: (11802,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 133_WSD_id_105_WebService_tr_3934_1st_5937.csv...
Using CUDA device: 2
Data shape: (11802, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 11802, window_size 5000, adding 3198 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.58it/s]
Original shapes - Output: (15000,), Label: (11802,)
Aligned shapes - Output: (11802,), Label: (11802,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 331.62it/s]
Table1-only metrics completed in 0.70s
Evaluation Result:  {'AUC-PR': 0.4388400687486346, 'AUC-ROC': 0.9960464918978535, 'VUS-PR': 0.19097745174764388, 'VUS-ROC': 0.9957783234499079, 'Standard-F1': 0.19402985074606696, 'Standard-Precision': 0.10924369747899068, 'Standard-Recall': 0.8666666666666089, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9942290005554503, 'Affiliation-P': 0.9885463082029966, 'Affiliation-R': 0.9999774049596114, 'F1_T': 0.19402985074606696, 'Precision_T': 0.10924369747899068, 'Recall_T': 0.8666666666666089}
Results for 133_WSD_id_105_WebService_tr_3934_1st_5937.csv: {'filename': '133_WSD_id_105_WebService_tr_3934_1st_5937.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 11, 'train_index': '3934', 'data_shape': '15736x1', 'output_length': 15000, 'label_length': 11802, 'aligned_length': 11802, 'AUC-PR': 0.4388400687486346, 'AUC-ROC': 0.9960464918978535, 'VUS-PR': 0.19097745174764388, 'VUS-ROC': 0.9957783234499079, 'Standard-F1': 0.19402985074606696, 'Standard-Precision': 0.10924369747899068, 'Standard-Recall': 0.8666666666666089, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9942290005554503, 'Affiliation-P': 0.9885463082029966, 'Affiliation-R': 0.9999774049596114, 'F1_T': 0.19402985074606696, 'Precision_T': 0.10924369747899068, 'Recall_T': 0.8666666666666089}
Logits metrics skipped for 133_WSD_id_105_WebService_tr_3934_1st_5937.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/133_WSD_id_105_WebService_tr_3934_1st_5937_results.pkl
Processing file: 235_SED_id_2_Medical_tr_2499_1st_3840.csv, Data shape: (9999, 1)
Label shape: (9999,)
Train data shape: (2499, 1), Test data shape: (7500, 1), Label test shape: (7500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 235_SED_id_2_Medical_tr_2499_1st_3840.csv...
Using CUDA device: 2
Data shape: (7500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 7500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 166.49it/s]
Original shapes - Output: (10000,), Label: (7500,)
Aligned shapes - Output: (7500,), Label: (7500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 401.67it/s]
Table1-only metrics completed in 2.19s
Evaluation Result:  {'AUC-PR': 0.6264585841089553, 'AUC-ROC': 0.9807080874086566, 'VUS-PR': 0.8476378149110302, 'VUS-ROC': 0.9930260175094193, 'Standard-F1': 0.7391304347821095, 'Standard-Precision': 0.6891891891891876, 'Standard-Recall': 0.7968749999999979, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9679202442192637, 'Affiliation-P': 0.9444951462548684, 'Affiliation-R': 0.9925368589758087, 'F1_T': 0.7382388419777879, 'Precision_T': 0.6876404494382007, 'Recall_T': 0.7968749999999979}
Results for 235_SED_id_2_Medical_tr_2499_1st_3840.csv: {'filename': '235_SED_id_2_Medical_tr_2499_1st_3840.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '2499', 'data_shape': '9999x1', 'output_length': 10000, 'label_length': 7500, 'aligned_length': 7500, 'AUC-PR': 0.6264585841089553, 'AUC-ROC': 0.9807080874086566, 'VUS-PR': 0.8476378149110302, 'VUS-ROC': 0.9930260175094193, 'Standard-F1': 0.7391304347821095, 'Standard-Precision': 0.6891891891891876, 'Standard-Recall': 0.7968749999999979, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9679202442192637, 'Affiliation-P': 0.9444951462548684, 'Affiliation-R': 0.9925368589758087, 'F1_T': 0.7382388419777879, 'Precision_T': 0.6876404494382007, 'Recall_T': 0.7968749999999979}
Logits metrics skipped for 235_SED_id_2_Medical_tr_2499_1st_3840.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/235_SED_id_2_Medical_tr_2499_1st_3840_results.pkl
Processing file: 662_YAHOO_id_112_WebService_tr_500_1st_361.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 662_YAHOO_id_112_WebService_tr_500_1st_361.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 159.13it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1260.62it/s]
Table1-only metrics completed in 0.38s
Evaluation Result:  {'AUC-PR': 0.8211111111111111, 'AUC-ROC': 0.9989787234042553, 'VUS-PR': 0.9568041313492677, 'VUS-ROC': 0.9998303918348225, 'Standard-F1': 0.5882352941171626, 'Standard-Precision': 0.41666666666663194, 'Standard-Recall': 0.9999999999997999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9960872375487181, 'Affiliation-P': 0.992204975177868, 'Affiliation-R': 1.0, 'F1_T': 0.5882352941171626, 'Precision_T': 0.41666666666663194, 'Recall_T': 0.9999999999997999}
Results for 662_YAHOO_id_112_WebService_tr_500_1st_361.csv: {'filename': '662_YAHOO_id_112_WebService_tr_500_1st_361.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.8211111111111111, 'AUC-ROC': 0.9989787234042553, 'VUS-PR': 0.9568041313492677, 'VUS-ROC': 0.9998303918348225, 'Standard-F1': 0.5882352941171626, 'Standard-Precision': 0.41666666666663194, 'Standard-Recall': 0.9999999999997999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9960872375487181, 'Affiliation-P': 0.992204975177868, 'Affiliation-R': 1.0, 'F1_T': 0.5882352941171626, 'Precision_T': 0.41666666666663194, 'Recall_T': 0.9999999999997999}
Logits metrics skipped for 662_YAHOO_id_112_WebService_tr_500_1st_361.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/662_YAHOO_id_112_WebService_tr_500_1st_361_results.pkl
Skipping file: 547_SMAP_id_17_Sensor_tr_2133_1st_5400.csv due to filter criteria.
Skipping file: 833_Exathlon_id_24_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 785_YAHOO_id_235_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 785_YAHOO_id_235_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 155.56it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1711.76it/s]
Table1-only metrics completed in 1.16s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 785_YAHOO_id_235_Synthetic_tr_500_1st_658.csv: {'filename': '785_YAHOO_id_235_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 235, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 785_YAHOO_id_235_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/785_YAHOO_id_235_Synthetic_tr_500_1st_658_results.pkl
Skipping file: 214_SMD_id_37_Facility_tr_5925_1st_17580.csv due to filter criteria.
Processing file: 609_YAHOO_id_59_WebService_tr_500_1st_17.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 609_YAHOO_id_59_WebService_tr_500_1st_17.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 149.75it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 999.53it/s]
Table1-only metrics completed in 0.69s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9907345751318977, 'VUS-ROC': 0.9999604320906595, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9959916626583288, 'Affiliation-P': 0.9920153305653147, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Results for 609_YAHOO_id_59_WebService_tr_500_1st_17.csv: {'filename': '609_YAHOO_id_59_WebService_tr_500_1st_17.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9907345751318977, 'VUS-ROC': 0.9999604320906595, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9959916626583288, 'Affiliation-P': 0.9920153305653147, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 609_YAHOO_id_59_WebService_tr_500_1st_17.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/609_YAHOO_id_59_WebService_tr_500_1st_17_results.pkl
Processing file: 734_YAHOO_id_184_WebService_tr_500_1st_768.csv, Data shape: (1424, 1)
Label shape: (1424,)
Train data shape: (500, 1), Test data shape: (924, 1), Label test shape: (924,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 734_YAHOO_id_184_WebService_tr_500_1st_768.csv...
Using CUDA device: 2
Data shape: (924, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 924, 1])
batched_mask shape: torch.Size([1, 924])

100%|██████████| 1/1 [00:00<00:00, 157.95it/s]
Original shapes - Output: (924,), Label: (924,)
Aligned shapes - Output: (924,), Label: (924,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2269.23it/s]
Table1-only metrics completed in 0.28s
Evaluation Result:  {'AUC-PR': 0.8746632013794312, 'AUC-ROC': 0.9372957735796585, 'VUS-PR': 0.8928157254802837, 'VUS-ROC': 0.9486774801435846, 'Standard-F1': 0.7937915742788758, 'Standard-Precision': 0.7991071428571394, 'Standard-Recall': 0.7885462555066045, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9502577970538257, 'Affiliation-P': 0.9977858813245151, 'Affiliation-R': 0.9070517038189224, 'F1_T': 0.7955555555550521, 'Precision_T': 0.8026905829596377, 'Recall_T': 0.7885462555066045}
Results for 734_YAHOO_id_184_WebService_tr_500_1st_768.csv: {'filename': '734_YAHOO_id_184_WebService_tr_500_1st_768.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 23, 'train_index': '500', 'data_shape': '1424x1', 'output_length': 924, 'label_length': 924, 'aligned_length': 924, 'AUC-PR': 0.8746632013794312, 'AUC-ROC': 0.9372957735796585, 'VUS-PR': 0.8928157254802837, 'VUS-ROC': 0.9486774801435846, 'Standard-F1': 0.7937915742788758, 'Standard-Precision': 0.7991071428571394, 'Standard-Recall': 0.7885462555066045, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9502577970538257, 'Affiliation-P': 0.9977858813245151, 'Affiliation-R': 0.9070517038189224, 'F1_T': 0.7955555555550521, 'Precision_T': 0.8026905829596377, 'Recall_T': 0.7885462555066045}
Logits metrics skipped for 734_YAHOO_id_184_WebService_tr_500_1st_768.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/734_YAHOO_id_184_WebService_tr_500_1st_768_results.pkl
Skipping file: 142_MSL_id_3_Sensor_tr_1525_1st_4575.csv due to filter criteria.
Skipping file: 194_SMD_id_17_Facility_tr_5925_1st_17580.csv due to filter criteria.
Processing file: 685_YAHOO_id_135_WebService_tr_500_1st_121.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 685_YAHOO_id_135_WebService_tr_500_1st_121.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 155.01it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 511.38it/s]
Table1-only metrics completed in 1.06s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9957486072801093, 'VUS-ROC': 0.9999726501180933, 'Standard-F1': 0.9655172413787443, 'Standard-Precision': 0.9333333333332711, 'Standard-Recall': 0.9999999999999286, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9935030990964074, 'Affiliation-P': 0.9870900727110847, 'Affiliation-R': 1.0, 'F1_T': 0.923076923076355, 'Precision_T': 0.9999999999999166, 'Recall_T': 0.8571428571427959}
Results for 685_YAHOO_id_135_WebService_tr_500_1st_121.csv: {'filename': '685_YAHOO_id_135_WebService_tr_500_1st_121.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9957486072801093, 'VUS-ROC': 0.9999726501180933, 'Standard-F1': 0.9655172413787443, 'Standard-Precision': 0.9333333333332711, 'Standard-Recall': 0.9999999999999286, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9935030990964074, 'Affiliation-P': 0.9870900727110847, 'Affiliation-R': 1.0, 'F1_T': 0.923076923076355, 'Precision_T': 0.9999999999999166, 'Recall_T': 0.8571428571427959}
Logits metrics skipped for 685_YAHOO_id_135_WebService_tr_500_1st_121.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/685_YAHOO_id_135_WebService_tr_500_1st_121_results.pkl
Processing file: 285_NEK_id_9_WebService_tr_500_1st_780.csv, Data shape: (1073, 1)
Label shape: (1073,)
Train data shape: (500, 1), Test data shape: (573, 1), Label test shape: (573,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 285_NEK_id_9_WebService_tr_500_1st_780.csv...
Using CUDA device: 2
Data shape: (573, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 573, 1])
batched_mask shape: torch.Size([1, 573])

100%|██████████| 1/1 [00:00<00:00, 155.50it/s]
Original shapes - Output: (573,), Label: (573,)
Aligned shapes - Output: (573,), Label: (573,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 3044.98it/s]
Table1-only metrics completed in 0.21s
Evaluation Result:  {'AUC-PR': 0.9538331746522406, 'AUC-ROC': 0.9842094782954998, 'VUS-PR': 0.9685778863083828, 'VUS-ROC': 0.9910757757145906, 'Standard-F1': 0.8955223880591955, 'Standard-Precision': 0.9677419354838607, 'Standard-Recall': 0.8333333333333257, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9945234977143139, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9891066528690902, 'F1_T': 0.8910891089103847, 'Precision_T': 0.9574468085106281, 'Recall_T': 0.8333333333333257}
Results for 285_NEK_id_9_WebService_tr_500_1st_780.csv: {'filename': '285_NEK_id_9_WebService_tr_500_1st_780.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1073x1', 'output_length': 573, 'label_length': 573, 'aligned_length': 573, 'AUC-PR': 0.9538331746522406, 'AUC-ROC': 0.9842094782954998, 'VUS-PR': 0.9685778863083828, 'VUS-ROC': 0.9910757757145906, 'Standard-F1': 0.8955223880591955, 'Standard-Precision': 0.9677419354838607, 'Standard-Recall': 0.8333333333333257, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9945234977143139, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9891066528690902, 'F1_T': 0.8910891089103847, 'Precision_T': 0.9574468085106281, 'Recall_T': 0.8333333333333257}
Logits metrics skipped for 285_NEK_id_9_WebService_tr_500_1st_780.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/285_NEK_id_9_WebService_tr_500_1st_780_results.pkl
Processing file: 362_UCR_id_60_HumanActivity_tr_16000_1st_31200.csv, Data shape: (64000, 1)
Label shape: (64000,)
Train data shape: (16000, 1), Test data shape: (48000, 1), Label test shape: (48000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 362_UCR_id_60_HumanActivity_tr_16000_1st_31200.csv...
Using CUDA device: 2
Data shape: (48000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 48000, window_size 5000, adding 2000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([10, 5000, 1])
batched_mask shape: torch.Size([10, 5000])

100%|██████████| 1/1 [00:00<00:00, 96.64it/s]
Original shapes - Output: (50000,), Label: (48000,)
Aligned shapes - Output: (48000,), Label: (48000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  40%|████      | 80/200 [00:01<00:01, 79.42it/s]

Calculating Affiliation Metrics:  84%|████████▍ | 168/200 [00:02<00:00, 84.10it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 88.94it/s]
Table1-only metrics completed in 3.37s
Evaluation Result:  {'AUC-PR': 0.02906785610856038, 'AUC-ROC': 0.7295543010315978, 'VUS-PR': 0.03097033709119882, 'VUS-ROC': 0.7493084905745501, 'Standard-F1': 0.07326932794313123, 'Standard-Precision': 0.043833131801692855, 'Standard-Recall': 0.22307692307692273, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.671647501863372, 'Affiliation-P': 0.5056244504418362, 'Affiliation-R': 1.0, 'F1_T': 0.07310310057950918, 'Precision_T': 0.04371419957793186, 'Recall_T': 0.22307692307692273}
Results for 362_UCR_id_60_HumanActivity_tr_16000_1st_31200.csv: {'filename': '362_UCR_id_60_HumanActivity_tr_16000_1st_31200.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '16000', 'data_shape': '64000x1', 'output_length': 50000, 'label_length': 48000, 'aligned_length': 48000, 'AUC-PR': 0.02906785610856038, 'AUC-ROC': 0.7295543010315978, 'VUS-PR': 0.03097033709119882, 'VUS-ROC': 0.7493084905745501, 'Standard-F1': 0.07326932794313123, 'Standard-Precision': 0.043833131801692855, 'Standard-Recall': 0.22307692307692273, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.671647501863372, 'Affiliation-P': 0.5056244504418362, 'Affiliation-R': 1.0, 'F1_T': 0.07310310057950918, 'Precision_T': 0.04371419957793186, 'Recall_T': 0.22307692307692273}
Logits metrics skipped for 362_UCR_id_60_HumanActivity_tr_16000_1st_31200.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/362_UCR_id_60_HumanActivity_tr_16000_1st_31200_results.pkl
Processing file: 446_UCR_id_144_Medical_tr_7500_1st_17000.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 446_UCR_id_144_Medical_tr_7500_1st_17000.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 148.23it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  84%|████████▍ | 168/200 [00:01<00:00, 167.52it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 179.23it/s]
Table1-only metrics completed in 1.99s
Evaluation Result:  {'AUC-PR': 0.6184526252663375, 'AUC-ROC': 0.9412625, 'VUS-PR': 0.41263523806354174, 'VUS-ROC': 0.9422500376287732, 'Standard-F1': 0.40615384615341765, 'Standard-Precision': 0.29333333333333206, 'Standard-Recall': 0.6599999999999935, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8312295293772197, 'Affiliation-P': 0.7113463937621832, 'Affiliation-R': 0.9997106666666667, 'F1_T': 0.40615384615341765, 'Precision_T': 0.29333333333333206, 'Recall_T': 0.6599999999999935}
Results for 446_UCR_id_144_Medical_tr_7500_1st_17000.csv: {'filename': '446_UCR_id_144_Medical_tr_7500_1st_17000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 166, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.6184526252663375, 'AUC-ROC': 0.9412625, 'VUS-PR': 0.41263523806354174, 'VUS-ROC': 0.9422500376287732, 'Standard-F1': 0.40615384615341765, 'Standard-Precision': 0.29333333333333206, 'Standard-Recall': 0.6599999999999935, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8312295293772197, 'Affiliation-P': 0.7113463937621832, 'Affiliation-R': 0.9997106666666667, 'F1_T': 0.40615384615341765, 'Precision_T': 0.29333333333333206, 'Recall_T': 0.6599999999999935}
Logits metrics skipped for 446_UCR_id_144_Medical_tr_7500_1st_17000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/446_UCR_id_144_Medical_tr_7500_1st_17000_results.pkl
Skipping file: 246_SVDB_id_10_Medical_tr_20100_1st_68460.csv due to filter criteria.
Processing file: 691_YAHOO_id_141_WebService_tr_500_1st_57.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 691_YAHOO_id_141_WebService_tr_500_1st_57.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 165.20it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 911.88it/s]
Table1-only metrics completed in 0.78s
Evaluation Result:  {'AUC-PR': 0.9733333333333333, 'AUC-ROC': 0.9997435897435898, 'VUS-PR': 0.9682700357753978, 'VUS-ROC': 0.999803407130727, 'Standard-F1': 0.9090909090903305, 'Standard-Precision': 0.8333333333332639, 'Standard-Recall': 0.9999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.998997432448954, 'Affiliation-P': 0.9979968731678717, 'Affiliation-R': 1.0, 'F1_T': 0.9090909090903305, 'Precision_T': 0.8333333333332639, 'Recall_T': 0.9999999999999}
Results for 691_YAHOO_id_141_WebService_tr_500_1st_57.csv: {'filename': '691_YAHOO_id_141_WebService_tr_500_1st_57.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9733333333333333, 'AUC-ROC': 0.9997435897435898, 'VUS-PR': 0.9682700357753978, 'VUS-ROC': 0.999803407130727, 'Standard-F1': 0.9090909090903305, 'Standard-Precision': 0.8333333333332639, 'Standard-Recall': 0.9999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.998997432448954, 'Affiliation-P': 0.9979968731678717, 'Affiliation-R': 1.0, 'F1_T': 0.9090909090903305, 'Precision_T': 0.8333333333332639, 'Recall_T': 0.9999999999999}
Logits metrics skipped for 691_YAHOO_id_141_WebService_tr_500_1st_57.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/691_YAHOO_id_141_WebService_tr_500_1st_57_results.pkl
Processing file: 805_YAHOO_id_255_WebService_tr_500_1st_95.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 805_YAHOO_id_255_WebService_tr_500_1st_95.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 165.88it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1196.80it/s]
Table1-only metrics completed in 0.54s
Evaluation Result:  {'AUC-PR': 0.982142857142857, 'AUC-ROC': 0.999878212154427, 'VUS-PR': 0.9754834027371, 'VUS-ROC': 0.9998954457777802, 'Standard-F1': 0.7368421052626148, 'Standard-Precision': 0.5833333333332847, 'Standard-Recall': 0.9999999999998571, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9974291659669343, 'Affiliation-P': 0.9948715164140106, 'Affiliation-R': 1.0, 'F1_T': 0.7368421052626148, 'Precision_T': 0.5833333333332847, 'Recall_T': 0.9999999999998571}
Results for 805_YAHOO_id_255_WebService_tr_500_1st_95.csv: {'filename': '805_YAHOO_id_255_WebService_tr_500_1st_95.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.982142857142857, 'AUC-ROC': 0.999878212154427, 'VUS-PR': 0.9754834027371, 'VUS-ROC': 0.9998954457777802, 'Standard-F1': 0.7368421052626148, 'Standard-Precision': 0.5833333333332847, 'Standard-Recall': 0.9999999999998571, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9974291659669343, 'Affiliation-P': 0.9948715164140106, 'Affiliation-R': 1.0, 'F1_T': 0.7368421052626148, 'Precision_T': 0.5833333333332847, 'Recall_T': 0.9999999999998571}
Logits metrics skipped for 805_YAHOO_id_255_WebService_tr_500_1st_95.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/805_YAHOO_id_255_WebService_tr_500_1st_95_results.pkl
Processing file: 390_UCR_id_88_Medical_tr_7500_1st_17000.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 390_UCR_id_88_Medical_tr_7500_1st_17000.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 153.57it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  90%|█████████ | 180/200 [00:01<00:00, 179.71it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 186.60it/s]
Table1-only metrics completed in 1.97s
Evaluation Result:  {'AUC-PR': 0.0076659762738920095, 'AUC-ROC': 0.7190598214285714, 'VUS-PR': 0.008595042987553812, 'VUS-ROC': 0.7557380163944248, 'Standard-F1': 0.0230642504118292, 'Standard-Precision': 0.011725293132328306, 'Standard-Recall': 0.6999999999999931, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.686458735125367, 'Affiliation-P': 0.5226190568677095, 'Affiliation-R': 0.999936, 'F1_T': 0.022999835715426853, 'Precision_T': 0.011691999331885751, 'Recall_T': 0.6999999999999931}
Results for 390_UCR_id_88_Medical_tr_7500_1st_17000.csv: {'filename': '390_UCR_id_88_Medical_tr_7500_1st_17000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 166, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.0076659762738920095, 'AUC-ROC': 0.7190598214285714, 'VUS-PR': 0.008595042987553812, 'VUS-ROC': 0.7557380163944248, 'Standard-F1': 0.0230642504118292, 'Standard-Precision': 0.011725293132328306, 'Standard-Recall': 0.6999999999999931, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.686458735125367, 'Affiliation-P': 0.5226190568677095, 'Affiliation-R': 0.999936, 'F1_T': 0.022999835715426853, 'Precision_T': 0.011691999331885751, 'Recall_T': 0.6999999999999931}
Logits metrics skipped for 390_UCR_id_88_Medical_tr_7500_1st_17000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/390_UCR_id_88_Medical_tr_7500_1st_17000_results.pkl
Skipping file: 834_Exathlon_id_25_Facility_tr_11665_1st_13484.csv due to filter criteria.
Skipping file: 245_SVDB_id_9_Medical_tr_12607_1st_12707.csv due to filter criteria.
Processing file: 477_UCR_id_175_Medical_tr_7500_1st_17000.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 477_UCR_id_175_Medical_tr_7500_1st_17000.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 150.28it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 208.89it/s]
Table1-only metrics completed in 1.87s
Evaluation Result:  {'AUC-PR': 0.6105133357491531, 'AUC-ROC': 0.9632854910714286, 'VUS-PR': 0.4284879020333237, 'VUS-ROC': 0.9628917410714274, 'Standard-F1': 0.4184615384611099, 'Standard-Precision': 0.3022222222222209, 'Standard-Recall': 0.6799999999999933, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7617745927837198, 'Affiliation-P': 0.6153218323586744, 'Affiliation-R': 0.9997173333333333, 'F1_T': 0.4184615384611099, 'Precision_T': 0.3022222222222209, 'Recall_T': 0.6799999999999933}
Results for 477_UCR_id_175_Medical_tr_7500_1st_17000.csv: {'filename': '477_UCR_id_175_Medical_tr_7500_1st_17000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 166, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.6105133357491531, 'AUC-ROC': 0.9632854910714286, 'VUS-PR': 0.4284879020333237, 'VUS-ROC': 0.9628917410714274, 'Standard-F1': 0.4184615384611099, 'Standard-Precision': 0.3022222222222209, 'Standard-Recall': 0.6799999999999933, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7617745927837198, 'Affiliation-P': 0.6153218323586744, 'Affiliation-R': 0.9997173333333333, 'F1_T': 0.4184615384611099, 'Precision_T': 0.3022222222222209, 'Recall_T': 0.6799999999999933}
Logits metrics skipped for 477_UCR_id_175_Medical_tr_7500_1st_17000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/477_UCR_id_175_Medical_tr_7500_1st_17000_results.pkl
Processing file: 658_YAHOO_id_108_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 658_YAHOO_id_108_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 162.96it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2313.53it/s]
Table1-only metrics completed in 0.49s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7785145888594159, 'Affiliation-P': 0.6373507057546146, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 658_YAHOO_id_108_Synthetic_tr_500_1st_893.csv: {'filename': '658_YAHOO_id_108_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7785145888594159, 'Affiliation-P': 0.6373507057546146, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 658_YAHOO_id_108_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/658_YAHOO_id_108_Synthetic_tr_500_1st_893_results.pkl
Processing file: 730_YAHOO_id_180_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 730_YAHOO_id_180_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 161.77it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1458.44it/s]
Table1-only metrics completed in 1.24s
Evaluation Result:  {'AUC-PR': 0.8741943241943242, 'AUC-ROC': 0.9986598440545809, 'VUS-PR': 0.99613495713891, 'VUS-ROC': 0.9999645286392265, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9967223768363775, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.9993098688750863, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Results for 730_YAHOO_id_180_Synthetic_tr_500_1st_623.csv: {'filename': '730_YAHOO_id_180_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 189, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.8741943241943242, 'AUC-ROC': 0.9986598440545809, 'VUS-PR': 0.99613495713891, 'VUS-ROC': 0.9999645286392265, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9967223768363775, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.9993098688750863, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 730_YAHOO_id_180_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/730_YAHOO_id_180_Synthetic_tr_500_1st_623_results.pkl
Processing file: 572_YAHOO_id_22_WebService_tr_500_1st_377.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 572_YAHOO_id_22_WebService_tr_500_1st_377.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 167.55it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 999.89it/s]
Table1-only metrics completed in 0.72s
Evaluation Result:  {'AUC-PR': 1.0000000000000002, 'AUC-ROC': 1.0, 'VUS-PR': 0.9903738699022929, 'VUS-ROC': 0.9999588912898956, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9946892402798332, 'Affiliation-P': 0.9894345909086967, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Results for 572_YAHOO_id_22_WebService_tr_500_1st_377.csv: {'filename': '572_YAHOO_id_22_WebService_tr_500_1st_377.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0000000000000002, 'AUC-ROC': 1.0, 'VUS-PR': 0.9903738699022929, 'VUS-ROC': 0.9999588912898956, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9946892402798332, 'Affiliation-P': 0.9894345909086967, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 572_YAHOO_id_22_WebService_tr_500_1st_377.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/572_YAHOO_id_22_WebService_tr_500_1st_377_results.pkl
Processing file: 727_YAHOO_id_177_WebService_tr_500_1st_82.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 727_YAHOO_id_177_WebService_tr_500_1st_82.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 160.65it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 600.69it/s]
Table1-only metrics completed in 1.05s
Evaluation Result:  {'AUC-PR': 0.9999999999999999, 'AUC-ROC': 1.0, 'VUS-PR': 0.9955065475625936, 'VUS-ROC': 0.9999710924778696, 'Standard-F1': 0.9655172413787443, 'Standard-Precision': 0.9333333333332711, 'Standard-Recall': 0.9999999999999286, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9962198494153707, 'Affiliation-P': 0.9924681702812571, 'Affiliation-R': 1.0, 'F1_T': 0.923076923076355, 'Precision_T': 0.9999999999999166, 'Recall_T': 0.8571428571427959}
Results for 727_YAHOO_id_177_WebService_tr_500_1st_82.csv: {'filename': '727_YAHOO_id_177_WebService_tr_500_1st_82.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9999999999999999, 'AUC-ROC': 1.0, 'VUS-PR': 0.9955065475625936, 'VUS-ROC': 0.9999710924778696, 'Standard-F1': 0.9655172413787443, 'Standard-Precision': 0.9333333333332711, 'Standard-Recall': 0.9999999999999286, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9962198494153707, 'Affiliation-P': 0.9924681702812571, 'Affiliation-R': 1.0, 'F1_T': 0.923076923076355, 'Precision_T': 0.9999999999999166, 'Recall_T': 0.8571428571427959}
Logits metrics skipped for 727_YAHOO_id_177_WebService_tr_500_1st_82.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/727_YAHOO_id_177_WebService_tr_500_1st_82_results.pkl
Processing file: 436_UCR_id_134_Medical_tr_1875_1st_4526.csv, Data shape: (7501, 1)
Label shape: (7501,)
Train data shape: (1875, 1), Test data shape: (5626, 1), Label test shape: (5626,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 436_UCR_id_134_Medical_tr_1875_1st_4526.csv...
Using CUDA device: 2
Data shape: (5626, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5626, window_size 5000, adding 4374 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 166.63it/s]
Original shapes - Output: (10000,), Label: (5626,)
Aligned shapes - Output: (5626,), Label: (5626,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 731.25it/s]
Table1-only metrics completed in 0.85s
Evaluation Result:  {'AUC-PR': 0.10848643570252828, 'AUC-ROC': 0.979288777698356, 'VUS-PR': 0.162750493301126, 'VUS-ROC': 0.9815931699637045, 'Standard-F1': 0.2403846153843662, 'Standard-Precision': 0.1404494382022464, 'Standard-Recall': 0.8333333333333056, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7775500300578699, 'Affiliation-P': 0.6360587747363432, 'Affiliation-R': 1.0, 'F1_T': 0.2424242424239829, 'Precision_T': 0.14285714285714202, 'Recall_T': 0.7999999999999734}
Results for 436_UCR_id_134_Medical_tr_1875_1st_4526.csv: {'filename': '436_UCR_id_134_Medical_tr_1875_1st_4526.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 154, 'train_index': '1875', 'data_shape': '7501x1', 'output_length': 10000, 'label_length': 5626, 'aligned_length': 5626, 'AUC-PR': 0.10848643570252828, 'AUC-ROC': 0.979288777698356, 'VUS-PR': 0.162750493301126, 'VUS-ROC': 0.9815931699637045, 'Standard-F1': 0.2403846153843662, 'Standard-Precision': 0.1404494382022464, 'Standard-Recall': 0.8333333333333056, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7775500300578699, 'Affiliation-P': 0.6360587747363432, 'Affiliation-R': 1.0, 'F1_T': 0.2424242424239829, 'Precision_T': 0.14285714285714202, 'Recall_T': 0.7999999999999734}
Logits metrics skipped for 436_UCR_id_134_Medical_tr_1875_1st_4526.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/436_UCR_id_134_Medical_tr_1875_1st_4526_results.pkl
Processing file: 559_YAHOO_id_9_WebService_tr_500_1st_538.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 559_YAHOO_id_9_WebService_tr_500_1st_538.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.72it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1616.47it/s]
Table1-only metrics completed in 0.27s
Evaluation Result:  {'AUC-PR': 0.2661290322580645, 'AUC-ROC': 0.9741086587436333, 'VUS-PR': 0.07419354838709678, 'VUS-ROC': 0.962542444821732, 'Standard-F1': 0.14285714285687756, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.49999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9436788713958472, 'Affiliation-P': 0.8933636238468952, 'Affiliation-R': 1.0, 'F1_T': 0.14285714285687756, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.49999999999975}
Results for 559_YAHOO_id_9_WebService_tr_500_1st_538.csv: {'filename': '559_YAHOO_id_9_WebService_tr_500_1st_538.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.2661290322580645, 'AUC-ROC': 0.9741086587436333, 'VUS-PR': 0.07419354838709678, 'VUS-ROC': 0.962542444821732, 'Standard-F1': 0.14285714285687756, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.49999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9436788713958472, 'Affiliation-P': 0.8933636238468952, 'Affiliation-R': 1.0, 'F1_T': 0.14285714285687756, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.49999999999975}
Logits metrics skipped for 559_YAHOO_id_9_WebService_tr_500_1st_538.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/559_YAHOO_id_9_WebService_tr_500_1st_538_results.pkl
Processing file: 095_WSD_id_67_WebService_tr_2182_1st_2282.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (2182, 1), Test data shape: (16054, 1), Label test shape: (16054,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 095_WSD_id_67_WebService_tr_2182_1st_2282.csv...
Using CUDA device: 2
Data shape: (16054, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 16054, window_size 5000, adding 3946 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.83it/s]
Original shapes - Output: (20000,), Label: (16054,)
Aligned shapes - Output: (16054,), Label: (16054,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 287.34it/s]
Table1-only metrics completed in 0.80s
Evaluation Result:  {'AUC-PR': 0.8508927840494989, 'AUC-ROC': 0.9779702529390809, 'VUS-PR': 0.30136635149291036, 'VUS-ROC': 0.9906037618582568, 'Standard-F1': 0.14857142857127967, 'Standard-Precision': 0.08074534161490633, 'Standard-Recall': 0.9285714285713622, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7863759723253367, 'Affiliation-P': 0.6479586935330374, 'Affiliation-R': 0.99999555073057, 'F1_T': 0.14857142857127967, 'Precision_T': 0.08074534161490633, 'Recall_T': 0.9285714285713622}
Results for 095_WSD_id_67_WebService_tr_2182_1st_2282.csv: {'filename': '095_WSD_id_67_WebService_tr_2182_1st_2282.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 8, 'train_index': '2182', 'data_shape': '18236x1', 'output_length': 20000, 'label_length': 16054, 'aligned_length': 16054, 'AUC-PR': 0.8508927840494989, 'AUC-ROC': 0.9779702529390809, 'VUS-PR': 0.30136635149291036, 'VUS-ROC': 0.9906037618582568, 'Standard-F1': 0.14857142857127967, 'Standard-Precision': 0.08074534161490633, 'Standard-Recall': 0.9285714285713622, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7863759723253367, 'Affiliation-P': 0.6479586935330374, 'Affiliation-R': 0.99999555073057, 'F1_T': 0.14857142857127967, 'Precision_T': 0.08074534161490633, 'Recall_T': 0.9285714285713622}
Logits metrics skipped for 095_WSD_id_67_WebService_tr_2182_1st_2282.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/095_WSD_id_67_WebService_tr_2182_1st_2282_results.pkl
Processing file: 030_WSD_id_2_WebService_tr_4499_1st_4599.csv, Data shape: (18084, 1)
Label shape: (18084,)
Train data shape: (4499, 1), Test data shape: (13585, 1), Label test shape: (13585,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 030_WSD_id_2_WebService_tr_4499_1st_4599.csv...
Using CUDA device: 2
Data shape: (13585, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13585, window_size 5000, adding 1415 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.98it/s]
Original shapes - Output: (15000,), Label: (13585,)
Aligned shapes - Output: (13585,), Label: (13585,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 231.29it/s]
Table1-only metrics completed in 2.91s
Evaluation Result:  {'AUC-PR': 0.39864437227099553, 'AUC-ROC': 0.8479055951827679, 'VUS-PR': 0.6024481339090916, 'VUS-ROC': 0.9885001023487588, 'Standard-F1': 0.33928571428523424, 'Standard-Precision': 0.2794117647058803, 'Standard-Recall': 0.43181818181817694, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9938539125200544, 'Affiliation-P': 0.9948681204283977, 'Affiliation-R': 0.9928417703531197, 'F1_T': 0.33928571428523424, 'Precision_T': 0.2794117647058803, 'Recall_T': 0.43181818181817694}
Results for 030_WSD_id_2_WebService_tr_4499_1st_4599.csv: {'filename': '030_WSD_id_2_WebService_tr_4499_1st_4599.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4499', 'data_shape': '18084x1', 'output_length': 15000, 'label_length': 13585, 'aligned_length': 13585, 'AUC-PR': 0.39864437227099553, 'AUC-ROC': 0.8479055951827679, 'VUS-PR': 0.6024481339090916, 'VUS-ROC': 0.9885001023487588, 'Standard-F1': 0.33928571428523424, 'Standard-Precision': 0.2794117647058803, 'Standard-Recall': 0.43181818181817694, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9938539125200544, 'Affiliation-P': 0.9948681204283977, 'Affiliation-R': 0.9928417703531197, 'F1_T': 0.33928571428523424, 'Precision_T': 0.2794117647058803, 'Recall_T': 0.43181818181817694}
Logits metrics skipped for 030_WSD_id_2_WebService_tr_4499_1st_4599.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/030_WSD_id_2_WebService_tr_4499_1st_4599_results.pkl
Processing file: 583_YAHOO_id_33_WebService_tr_500_1st_1616.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 583_YAHOO_id_33_WebService_tr_500_1st_1616.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.39it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1037.90it/s]
Table1-only metrics completed in 0.29s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9983022071307294, 'Affiliation-P': 0.9966101694915253, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Results for 583_YAHOO_id_33_WebService_tr_500_1st_1616.csv: {'filename': '583_YAHOO_id_33_WebService_tr_500_1st_1616.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9983022071307294, 'Affiliation-P': 0.9966101694915253, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 583_YAHOO_id_33_WebService_tr_500_1st_1616.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/583_YAHOO_id_33_WebService_tr_500_1st_1616_results.pkl
Skipping file: 825_Exathlon_id_16_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 620_YAHOO_id_70_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 620_YAHOO_id_70_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 158.07it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1827.70it/s]
Table1-only metrics completed in 1.15s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 620_YAHOO_id_70_Synthetic_tr_500_1st_658.csv: {'filename': '620_YAHOO_id_70_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 235, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 620_YAHOO_id_70_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/620_YAHOO_id_70_Synthetic_tr_500_1st_658_results.pkl
Processing file: 071_WSD_id_43_WebService_tr_3387_1st_3487.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (3387, 1), Test data shape: (14849, 1), Label test shape: (14849,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 071_WSD_id_43_WebService_tr_3387_1st_3487.csv...
Using CUDA device: 2
Data shape: (14849, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 14849, window_size 5000, adding 151 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.61it/s]
Original shapes - Output: (15000,), Label: (14849,)
Aligned shapes - Output: (14849,), Label: (14849,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 225.32it/s]
Table1-only metrics completed in 1.47s
Evaluation Result:  {'AUC-PR': 0.7478727073373905, 'AUC-ROC': 0.8949843616846491, 'VUS-PR': 0.7018780111431914, 'VUS-ROC': 0.8953708888153561, 'Standard-F1': 0.7427293064871964, 'Standard-Precision': 0.7748962655601657, 'Standard-Recall': 0.7131264916467778, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8701127695710865, 'Affiliation-P': 0.7930135146603084, 'Affiliation-R': 0.9638181893314861, 'F1_T': 0.7411852963235818, 'Precision_T': 0.7783613445378148, 'Recall_T': 0.7073985680190927}
Results for 071_WSD_id_43_WebService_tr_3387_1st_3487.csv: {'filename': '071_WSD_id_43_WebService_tr_3387_1st_3487.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 29, 'train_index': '3387', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 14849, 'aligned_length': 14849, 'AUC-PR': 0.7478727073373905, 'AUC-ROC': 0.8949843616846491, 'VUS-PR': 0.7018780111431914, 'VUS-ROC': 0.8953708888153561, 'Standard-F1': 0.7427293064871964, 'Standard-Precision': 0.7748962655601657, 'Standard-Recall': 0.7131264916467778, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8701127695710865, 'Affiliation-P': 0.7930135146603084, 'Affiliation-R': 0.9638181893314861, 'F1_T': 0.7411852963235818, 'Precision_T': 0.7783613445378148, 'Recall_T': 0.7073985680190927}
Logits metrics skipped for 071_WSD_id_43_WebService_tr_3387_1st_3487.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/071_WSD_id_43_WebService_tr_3387_1st_3487_results.pkl
Processing file: 265_IOPS_id_6_WebService_tr_6453_1st_6553.csv, Data shape: (128787, 1)
Label shape: (128787,)
Train data shape: (6453, 1), Test data shape: (122334, 1), Label test shape: (122334,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 265_IOPS_id_6_WebService_tr_6453_1st_6553.csv...
Using CUDA device: 2
Data shape: (122334, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 122334, window_size 5000, adding 2666 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([25, 5000, 1])
batched_mask shape: torch.Size([25, 5000])

100%|██████████| 1/1 [00:00<00:00, 54.39it/s]
Original shapes - Output: (125000,), Label: (122334,)
Aligned shapes - Output: (122334,), Label: (122334,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  10%|▉         | 19/200 [00:01<00:10, 17.73it/s]

Calculating Affiliation Metrics:  18%|█▊        | 37/200 [00:03<00:15, 10.65it/s]

Calculating Affiliation Metrics:  25%|██▌       | 50/200 [00:05<00:17,  8.73it/s]

Calculating Affiliation Metrics:  30%|███       | 60/200 [00:06<00:18,  7.68it/s]

Calculating Affiliation Metrics:  34%|███▍      | 69/200 [00:08<00:18,  6.95it/s]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [00:10<00:19,  6.42it/s]

Calculating Affiliation Metrics:  42%|████▏     | 84/200 [00:11<00:19,  5.95it/s]

Calculating Affiliation Metrics:  45%|████▌     | 90/200 [00:12<00:19,  5.75it/s]

Calculating Affiliation Metrics:  48%|████▊     | 96/200 [00:13<00:18,  5.63it/s]

Calculating Affiliation Metrics:  51%|█████     | 102/200 [00:14<00:17,  5.52it/s]

Calculating Affiliation Metrics:  54%|█████▍    | 108/200 [00:16<00:16,  5.44it/s]

Calculating Affiliation Metrics:  57%|█████▋    | 114/200 [00:17<00:15,  5.38it/s]

Calculating Affiliation Metrics:  60%|██████    | 120/200 [00:18<00:15,  5.26it/s]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [00:19<00:13,  5.30it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 132/200 [00:20<00:12,  5.36it/s]

Calculating Affiliation Metrics:  69%|██████▉   | 138/200 [00:21<00:11,  5.50it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [00:22<00:09,  5.79it/s]

Calculating Affiliation Metrics:  76%|███████▋  | 153/200 [00:23<00:07,  6.38it/s]

Calculating Affiliation Metrics:  82%|████████▏ | 164/200 [00:24<00:04,  7.57it/s]

Calculating Affiliation Metrics:  94%|█████████▍| 189/200 [00:25<00:00, 12.52it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:26<00:00,  7.69it/s]
Table1-only metrics completed in 27.13s
Evaluation Result:  {'AUC-PR': 0.540401873692394, 'AUC-ROC': 0.9938342899685952, 'VUS-PR': 0.4340397496688461, 'VUS-ROC': 0.9956791717544848, 'Standard-F1': 0.12490594431889117, 'Standard-Precision': 0.06781045751633982, 'Standard-Recall': 0.790476190476183, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8187268506827043, 'Affiliation-P': 0.693411880523393, 'Affiliation-R': 0.9993276055718265, 'F1_T': 0.12490594431889117, 'Precision_T': 0.06781045751633982, 'Recall_T': 0.790476190476183}
Results for 265_IOPS_id_6_WebService_tr_6453_1st_6553.csv: {'filename': '265_IOPS_id_6_WebService_tr_6453_1st_6553.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 12, 'train_index': '6453', 'data_shape': '128787x1', 'output_length': 125000, 'label_length': 122334, 'aligned_length': 122334, 'AUC-PR': 0.540401873692394, 'AUC-ROC': 0.9938342899685952, 'VUS-PR': 0.4340397496688461, 'VUS-ROC': 0.9956791717544848, 'Standard-F1': 0.12490594431889117, 'Standard-Precision': 0.06781045751633982, 'Standard-Recall': 0.790476190476183, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8187268506827043, 'Affiliation-P': 0.693411880523393, 'Affiliation-R': 0.9993276055718265, 'F1_T': 0.12490594431889117, 'Precision_T': 0.06781045751633982, 'Recall_T': 0.790476190476183}
Logits metrics skipped for 265_IOPS_id_6_WebService_tr_6453_1st_6553.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/265_IOPS_id_6_WebService_tr_6453_1st_6553_results.pkl
Processing file: 463_UCR_id_161_Environment_tr_2046_1st_6006.csv, Data shape: (8184, 1)
Label shape: (8184,)
Train data shape: (2046, 1), Test data shape: (6138, 1), Label test shape: (6138,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 463_UCR_id_161_Environment_tr_2046_1st_6006.csv...
Using CUDA device: 2
Data shape: (6138, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6138, window_size 5000, adding 3862 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 152.76it/s]
Original shapes - Output: (10000,), Label: (6138,)
Aligned shapes - Output: (6138,), Label: (6138,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 503.43it/s]
Table1-only metrics completed in 0.52s
Evaluation Result:  {'AUC-PR': 0.1888948052348449, 'AUC-ROC': 0.9702107279693487, 'VUS-PR': 0.1690276349126178, 'VUS-ROC': 0.9739444116294065, 'Standard-F1': 0.2471910112356583, 'Standard-Precision': 0.15068493150684864, 'Standard-Recall': 0.6874999999999857, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.762572227511092, 'Affiliation-P': 0.6162559500158045, 'Affiliation-R': 1.0, 'F1_T': 0.24521072796904664, 'Precision_T': 0.15023474178403687, 'Recall_T': 0.6666666666666528}
Results for 463_UCR_id_161_Environment_tr_2046_1st_6006.csv: {'filename': '463_UCR_id_161_Environment_tr_2046_1st_6006.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '2046', 'data_shape': '8184x1', 'output_length': 10000, 'label_length': 6138, 'aligned_length': 6138, 'AUC-PR': 0.1888948052348449, 'AUC-ROC': 0.9702107279693487, 'VUS-PR': 0.1690276349126178, 'VUS-ROC': 0.9739444116294065, 'Standard-F1': 0.2471910112356583, 'Standard-Precision': 0.15068493150684864, 'Standard-Recall': 0.6874999999999857, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.762572227511092, 'Affiliation-P': 0.6162559500158045, 'Affiliation-R': 1.0, 'F1_T': 0.24521072796904664, 'Precision_T': 0.15023474178403687, 'Recall_T': 0.6666666666666528}
Logits metrics skipped for 463_UCR_id_161_Environment_tr_2046_1st_6006.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/463_UCR_id_161_Environment_tr_2046_1st_6006_results.pkl
Processing file: 032_WSD_id_4_WebService_tr_4559_1st_11822.csv, Data shape: (10686, 1)
Label shape: (10686,)
Train data shape: (4559, 1), Test data shape: (6127, 1), Label test shape: (6127,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 032_WSD_id_4_WebService_tr_4559_1st_11822.csv...
Using CUDA device: 2
Data shape: (6127, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6127, window_size 5000, adding 3873 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 166.07it/s]
Original shapes - Output: (10000,), Label: (6127,)
Aligned shapes - Output: (6127,), Label: (6127,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 612.07it/s]
Table1-only metrics completed in 0.47s
Evaluation Result:  {'AUC-PR': 0.12199578842746031, 'AUC-ROC': 0.7474864909120681, 'VUS-PR': 0.08389976391933064, 'VUS-ROC': 0.8762817897543944, 'Standard-F1': 0.07310704960825572, 'Standard-Precision': 0.03856749311294755, 'Standard-Recall': 0.6999999999999651, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9092892028007128, 'Affiliation-P': 0.8345591783942394, 'Affiliation-R': 0.9987187857026278, 'F1_T': 0.07310704960825572, 'Precision_T': 0.03856749311294755, 'Recall_T': 0.6999999999999651}
Results for 032_WSD_id_4_WebService_tr_4559_1st_11822.csv: {'filename': '032_WSD_id_4_WebService_tr_4559_1st_11822.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '4559', 'data_shape': '10686x1', 'output_length': 10000, 'label_length': 6127, 'aligned_length': 6127, 'AUC-PR': 0.12199578842746031, 'AUC-ROC': 0.7474864909120681, 'VUS-PR': 0.08389976391933064, 'VUS-ROC': 0.8762817897543944, 'Standard-F1': 0.07310704960825572, 'Standard-Precision': 0.03856749311294755, 'Standard-Recall': 0.6999999999999651, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9092892028007128, 'Affiliation-P': 0.8345591783942394, 'Affiliation-R': 0.9987187857026278, 'F1_T': 0.07310704960825572, 'Precision_T': 0.03856749311294755, 'Recall_T': 0.6999999999999651}
Logits metrics skipped for 032_WSD_id_4_WebService_tr_4559_1st_11822.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/032_WSD_id_4_WebService_tr_4559_1st_11822_results.pkl
Processing file: 074_WSD_id_46_WebService_tr_990_1st_1090.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (990, 1), Test data shape: (17246, 1), Label test shape: (17246,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 074_WSD_id_46_WebService_tr_990_1st_1090.csv...
Using CUDA device: 2
Data shape: (17246, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 17246, window_size 5000, adding 2754 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.70it/s]
Original shapes - Output: (20000,), Label: (17246,)
Aligned shapes - Output: (17246,), Label: (17246,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [00:01<00:00, 111.44it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 131.06it/s]
Table1-only metrics completed in 2.38s
Evaluation Result:  {'AUC-PR': 0.09827392290231741, 'AUC-ROC': 0.8130032294539213, 'VUS-PR': 0.1091586469641901, 'VUS-ROC': 0.8303985011812495, 'Standard-F1': 0.2473958333328626, 'Standard-Precision': 0.19874476987447656, 'Standard-Recall': 0.32758620689655055, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8181887995908007, 'Affiliation-P': 0.9035980946785056, 'Affiliation-R': 0.7475311590667761, 'F1_T': 0.2465753424652914, 'Precision_T': 0.1929824561403505, 'Recall_T': 0.3413793103448264}
Results for 074_WSD_id_46_WebService_tr_990_1st_1090.csv: {'filename': '074_WSD_id_46_WebService_tr_990_1st_1090.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 16, 'train_index': '990', 'data_shape': '18236x1', 'output_length': 20000, 'label_length': 17246, 'aligned_length': 17246, 'AUC-PR': 0.09827392290231741, 'AUC-ROC': 0.8130032294539213, 'VUS-PR': 0.1091586469641901, 'VUS-ROC': 0.8303985011812495, 'Standard-F1': 0.2473958333328626, 'Standard-Precision': 0.19874476987447656, 'Standard-Recall': 0.32758620689655055, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8181887995908007, 'Affiliation-P': 0.9035980946785056, 'Affiliation-R': 0.7475311590667761, 'F1_T': 0.2465753424652914, 'Precision_T': 0.1929824561403505, 'Recall_T': 0.3413793103448264}
Logits metrics skipped for 074_WSD_id_46_WebService_tr_990_1st_1090.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/074_WSD_id_46_WebService_tr_990_1st_1090_results.pkl
Processing file: 274_IOPS_id_15_WebService_tr_4591_1st_4691.csv, Data shape: (149132, 1)
Label shape: (149132,)
Train data shape: (4591, 1), Test data shape: (144541, 1), Label test shape: (144541,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 274_IOPS_id_15_WebService_tr_4591_1st_4691.csv...
Using CUDA device: 2
Data shape: (144541, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 144541, window_size 5000, adding 459 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([29, 5000, 1])
batched_mask shape: torch.Size([29, 5000])

100%|██████████| 1/1 [00:00<00:00, 47.59it/s]
Original shapes - Output: (145000,), Label: (144541,)
Aligned shapes - Output: (144541,), Label: (144541,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   6%|▋         | 13/200 [00:01<00:15, 12.04it/s]

Calculating Affiliation Metrics:  13%|█▎        | 26/200 [00:03<00:21,  8.02it/s]

Calculating Affiliation Metrics:  18%|█▊        | 35/200 [00:04<00:23,  6.99it/s]

Calculating Affiliation Metrics:  22%|██▏       | 43/200 [00:06<00:24,  6.35it/s]

Calculating Affiliation Metrics:  25%|██▌       | 50/200 [00:07<00:24,  6.02it/s]

Calculating Affiliation Metrics:  28%|██▊       | 57/200 [00:08<00:24,  5.78it/s]

Calculating Affiliation Metrics:  32%|███▏      | 63/200 [00:09<00:24,  5.70it/s]

Calculating Affiliation Metrics:  34%|███▍      | 69/200 [00:10<00:22,  5.72it/s]

Calculating Affiliation Metrics:  38%|███▊      | 75/200 [00:11<00:21,  5.76it/s]

Calculating Affiliation Metrics:  40%|████      | 81/200 [00:13<00:21,  5.65it/s]

Calculating Affiliation Metrics:  44%|████▎     | 87/200 [00:14<00:20,  5.64it/s]

Calculating Affiliation Metrics:  46%|████▋     | 93/200 [00:15<00:18,  5.66it/s]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [00:16<00:17,  5.73it/s]

Calculating Affiliation Metrics:  53%|█████▎    | 106/200 [00:17<00:16,  5.83it/s]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [00:18<00:14,  5.94it/s]

Calculating Affiliation Metrics:  60%|██████    | 120/200 [00:19<00:13,  6.12it/s]

Calculating Affiliation Metrics:  64%|██████▍   | 128/200 [00:20<00:11,  6.48it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 136/200 [00:21<00:09,  6.80it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [00:22<00:07,  7.30it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 155/200 [00:23<00:05,  8.01it/s]

Calculating Affiliation Metrics:  84%|████████▎ | 167/200 [00:24<00:03,  9.07it/s]

Calculating Affiliation Metrics:  90%|████████▉ | 179/200 [00:25<00:02,  9.75it/s]

Calculating Affiliation Metrics:  98%|█████████▊| 197/200 [00:26<00:00, 12.12it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:26<00:00,  7.42it/s]
Table1-only metrics completed in 53.92s
Evaluation Result:  {'AUC-PR': 0.7907150466365063, 'AUC-ROC': 0.9463167469769297, 'VUS-PR': 0.7840306609238265, 'VUS-ROC': 0.9713420763115916, 'Standard-F1': 0.7343962823486743, 'Standard-Precision': 0.8034712482468442, 'Standard-Recall': 0.6762579312380108, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9003462916058881, 'Affiliation-P': 0.9398712842719783, 'Affiliation-R': 0.8640114786087243, 'F1_T': 0.7345926044496396, 'Precision_T': 0.8026937204827705, 'Recall_T': 0.6771432787369042}
Results for 274_IOPS_id_15_WebService_tr_4591_1st_4691.csv: {'filename': '274_IOPS_id_15_WebService_tr_4591_1st_4691.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4591', 'data_shape': '149132x1', 'output_length': 145000, 'label_length': 144541, 'aligned_length': 144541, 'AUC-PR': 0.7907150466365063, 'AUC-ROC': 0.9463167469769297, 'VUS-PR': 0.7840306609238265, 'VUS-ROC': 0.9713420763115916, 'Standard-F1': 0.7343962823486743, 'Standard-Precision': 0.8034712482468442, 'Standard-Recall': 0.6762579312380108, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9003462916058881, 'Affiliation-P': 0.9398712842719783, 'Affiliation-R': 0.8640114786087243, 'F1_T': 0.7345926044496396, 'Precision_T': 0.8026937204827705, 'Recall_T': 0.6771432787369042}
Logits metrics skipped for 274_IOPS_id_15_WebService_tr_4591_1st_4691.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/274_IOPS_id_15_WebService_tr_4591_1st_4691_results.pkl
Skipping file: 201_SMD_id_24_Facility_tr_7174_1st_21230.csv due to filter criteria.
Processing file: 451_UCR_id_149_Medical_tr_3000_1st_7175.csv, Data shape: (12000, 1)
Label shape: (12000,)
Train data shape: (3000, 1), Test data shape: (9000, 1), Label test shape: (9000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 451_UCR_id_149_Medical_tr_3000_1st_7175.csv...
Using CUDA device: 2
Data shape: (9000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9000, window_size 5000, adding 1000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 150.52it/s]
Original shapes - Output: (10000,), Label: (9000,)
Aligned shapes - Output: (9000,), Label: (9000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 457.68it/s]
Table1-only metrics completed in 1.35s
Evaluation Result:  {'AUC-PR': 0.05432968256343258, 'AUC-ROC': 0.7903085597534985, 'VUS-PR': 0.0556120514275712, 'VUS-ROC': 0.7931669487952586, 'Standard-F1': 0.11774325429243496, 'Standard-Precision': 0.07128712871287121, 'Standard-Recall': 0.33802816901408295, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6726853401521017, 'Affiliation-P': 0.5068017106276728, 'Affiliation-R': 1.0, 'F1_T': 0.11831442463504643, 'Precision_T': 0.07149853085210571, 'Recall_T': 0.3427230046948341}
Results for 451_UCR_id_149_Medical_tr_3000_1st_7175.csv: {'filename': '451_UCR_id_149_Medical_tr_3000_1st_7175.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 218, 'train_index': '3000', 'data_shape': '12000x1', 'output_length': 10000, 'label_length': 9000, 'aligned_length': 9000, 'AUC-PR': 0.05432968256343258, 'AUC-ROC': 0.7903085597534985, 'VUS-PR': 0.0556120514275712, 'VUS-ROC': 0.7931669487952586, 'Standard-F1': 0.11774325429243496, 'Standard-Precision': 0.07128712871287121, 'Standard-Recall': 0.33802816901408295, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6726853401521017, 'Affiliation-P': 0.5068017106276728, 'Affiliation-R': 1.0, 'F1_T': 0.11831442463504643, 'Precision_T': 0.07149853085210571, 'Recall_T': 0.3427230046948341}
Logits metrics skipped for 451_UCR_id_149_Medical_tr_3000_1st_7175.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/451_UCR_id_149_Medical_tr_3000_1st_7175_results.pkl
Processing file: 404_UCR_id_102_Facility_tr_13750_1st_52600.csv, Data shape: (55000, 1)
Label shape: (55000,)
Train data shape: (13750, 1), Test data shape: (41250, 1), Label test shape: (41250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 404_UCR_id_102_Facility_tr_13750_1st_52600.csv...
Using CUDA device: 2
Data shape: (41250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 41250, window_size 5000, adding 3750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([9, 5000, 1])
batched_mask shape: torch.Size([9, 5000])

100%|██████████| 1/1 [00:00<00:00, 108.57it/s]
Original shapes - Output: (45000,), Label: (41250,)
Aligned shapes - Output: (41250,), Label: (41250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  47%|████▋     | 94/200 [00:01<00:01, 93.19it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 105.86it/s]
Table1-only metrics completed in 3.60s
Evaluation Result:  {'AUC-PR': 0.2936742889303751, 'AUC-ROC': 0.9369989037758829, 'VUS-PR': 0.1776064026396815, 'VUS-ROC': 0.954788765708688, 'Standard-F1': 0.3034257748772103, 'Standard-Precision': 0.22518159806295343, 'Standard-Recall': 0.4649999999999977, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7177942282240819, 'Affiliation-P': 0.5601093822843823, 'Affiliation-R': 0.9990526060606061, 'F1_T': 0.3034257748772103, 'Precision_T': 0.22518159806295343, 'Recall_T': 0.4649999999999977}
Results for 404_UCR_id_102_Facility_tr_13750_1st_52600.csv: {'filename': '404_UCR_id_102_Facility_tr_13750_1st_52600.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 233, 'train_index': '13750', 'data_shape': '55000x1', 'output_length': 45000, 'label_length': 41250, 'aligned_length': 41250, 'AUC-PR': 0.2936742889303751, 'AUC-ROC': 0.9369989037758829, 'VUS-PR': 0.1776064026396815, 'VUS-ROC': 0.954788765708688, 'Standard-F1': 0.3034257748772103, 'Standard-Precision': 0.22518159806295343, 'Standard-Recall': 0.4649999999999977, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7177942282240819, 'Affiliation-P': 0.5601093822843823, 'Affiliation-R': 0.9990526060606061, 'F1_T': 0.3034257748772103, 'Precision_T': 0.22518159806295343, 'Recall_T': 0.4649999999999977}
Logits metrics skipped for 404_UCR_id_102_Facility_tr_13750_1st_52600.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/404_UCR_id_102_Facility_tr_13750_1st_52600_results.pkl
Processing file: 018_NAB_id_18_Facility_tr_500_1st_669.csv, Data shape: (1881, 1)
Label shape: (1881,)
Train data shape: (500, 1), Test data shape: (1381, 1), Label test shape: (1381,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 018_NAB_id_18_Facility_tr_500_1st_669.csv...
Using CUDA device: 2
Data shape: (1381, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1381, 1])
batched_mask shape: torch.Size([1, 1381])

100%|██████████| 1/1 [00:00<00:00, 155.84it/s]
Original shapes - Output: (1381,), Label: (1381,)
Aligned shapes - Output: (1381,), Label: (1381,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1336.40it/s]
Table1-only metrics completed in 0.86s
Evaluation Result:  {'AUC-PR': 0.2208616259867744, 'AUC-ROC': 0.4917559879438569, 'VUS-PR': 0.37711565865746033, 'VUS-ROC': 0.6674647833794651, 'Standard-F1': 0.28484848484799286, 'Standard-Precision': 0.3309859154929554, 'Standard-Recall': 0.24999999999999867, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8205024439430871, 'Affiliation-P': 0.737728121755906, 'Affiliation-R': 0.9241991225409325, 'F1_T': 0.2839879154073625, 'Precision_T': 0.32867132867132637, 'Recall_T': 0.24999999999999867}
Results for 018_NAB_id_18_Facility_tr_500_1st_669.csv: {'filename': '018_NAB_id_18_Facility_tr_500_1st_669.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '1881x1', 'output_length': 1381, 'label_length': 1381, 'aligned_length': 1381, 'AUC-PR': 0.2208616259867744, 'AUC-ROC': 0.4917559879438569, 'VUS-PR': 0.37711565865746033, 'VUS-ROC': 0.6674647833794651, 'Standard-F1': 0.28484848484799286, 'Standard-Precision': 0.3309859154929554, 'Standard-Recall': 0.24999999999999867, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8205024439430871, 'Affiliation-P': 0.737728121755906, 'Affiliation-R': 0.9241991225409325, 'F1_T': 0.2839879154073625, 'Precision_T': 0.32867132867132637, 'Recall_T': 0.24999999999999867}
Logits metrics skipped for 018_NAB_id_18_Facility_tr_500_1st_669.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/018_NAB_id_18_Facility_tr_500_1st_669_results.pkl
Processing file: 058_WSD_id_30_WebService_tr_1136_1st_1236.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (1136, 1), Test data shape: (17100, 1), Label test shape: (17100,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 058_WSD_id_30_WebService_tr_1136_1st_1236.csv...
Using CUDA device: 2
Data shape: (17100, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 17100, window_size 5000, adding 2900 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.13it/s]
Original shapes - Output: (20000,), Label: (17100,)
Aligned shapes - Output: (17100,), Label: (17100,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 267.36it/s]
Table1-only metrics completed in 0.86s
Evaluation Result:  {'AUC-PR': 0.5038812735865213, 'AUC-ROC': 0.963462438567751, 'VUS-PR': 0.15094907524679763, 'VUS-ROC': 0.9688345709840087, 'Standard-F1': 0.04469273743008171, 'Standard-Precision': 0.02339181286549694, 'Standard-Recall': 0.4999999999999375, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8788024175801632, 'Affiliation-P': 0.7838518291853664, 'Affiliation-R': 0.9999269005847954, 'F1_T': 0.04469273743008171, 'Precision_T': 0.02339181286549694, 'Recall_T': 0.4999999999999375}
Results for 058_WSD_id_30_WebService_tr_1136_1st_1236.csv: {'filename': '058_WSD_id_30_WebService_tr_1136_1st_1236.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 9, 'train_index': '1136', 'data_shape': '18236x1', 'output_length': 20000, 'label_length': 17100, 'aligned_length': 17100, 'AUC-PR': 0.5038812735865213, 'AUC-ROC': 0.963462438567751, 'VUS-PR': 0.15094907524679763, 'VUS-ROC': 0.9688345709840087, 'Standard-F1': 0.04469273743008171, 'Standard-Precision': 0.02339181286549694, 'Standard-Recall': 0.4999999999999375, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8788024175801632, 'Affiliation-P': 0.7838518291853664, 'Affiliation-R': 0.9999269005847954, 'F1_T': 0.04469273743008171, 'Precision_T': 0.02339181286549694, 'Recall_T': 0.4999999999999375}
Logits metrics skipped for 058_WSD_id_30_WebService_tr_1136_1st_1236.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/058_WSD_id_30_WebService_tr_1136_1st_1236_results.pkl
Processing file: 110_WSD_id_82_WebService_tr_500_1st_375.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (500, 1), Test data shape: (17736, 1), Label test shape: (17736,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 110_WSD_id_82_WebService_tr_500_1st_375.csv...
Using CUDA device: 2
Data shape: (17736, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 17736, window_size 5000, adding 2264 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.00it/s]
Original shapes - Output: (20000,), Label: (17736,)
Aligned shapes - Output: (17736,), Label: (17736,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 229.24it/s]
Table1-only metrics completed in 2.06s
Evaluation Result:  {'AUC-PR': 0.32192526427079154, 'AUC-ROC': 0.8499696835277855, 'VUS-PR': 0.11317270525320551, 'VUS-ROC': 0.9508507388603517, 'Standard-F1': 0.15254237288098824, 'Standard-Precision': 0.10055865921787653, 'Standard-Recall': 0.31578947368420496, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8250846292850588, 'Affiliation-P': 0.7031451848241536, 'Affiliation-R': 0.998190905326607, 'F1_T': 0.15254237288098824, 'Precision_T': 0.10055865921787653, 'Recall_T': 0.31578947368420496}
Results for 110_WSD_id_82_WebService_tr_500_1st_375.csv: {'filename': '110_WSD_id_82_WebService_tr_500_1st_375.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '18236x1', 'output_length': 20000, 'label_length': 17736, 'aligned_length': 17736, 'AUC-PR': 0.32192526427079154, 'AUC-ROC': 0.8499696835277855, 'VUS-PR': 0.11317270525320551, 'VUS-ROC': 0.9508507388603517, 'Standard-F1': 0.15254237288098824, 'Standard-Precision': 0.10055865921787653, 'Standard-Recall': 0.31578947368420496, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8250846292850588, 'Affiliation-P': 0.7031451848241536, 'Affiliation-R': 0.998190905326607, 'F1_T': 0.15254237288098824, 'Precision_T': 0.10055865921787653, 'Recall_T': 0.31578947368420496}
Logits metrics skipped for 110_WSD_id_82_WebService_tr_500_1st_375.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/110_WSD_id_82_WebService_tr_500_1st_375_results.pkl
Skipping file: 257_TAO_id_1_Environment_tr_500_1st_3.csv due to filter criteria.
Skipping file: 810_Exathlon_id_1_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 711_YAHOO_id_161_WebService_tr_500_1st_137.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 711_YAHOO_id_161_WebService_tr_500_1st_137.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 162.50it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1414.61it/s]
Table1-only metrics completed in 0.29s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.8082294711831648, 'VUS-ROC': 0.9995927558914194, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9430250786873685, 'Affiliation-P': 0.8921924822173166, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Results for 711_YAHOO_id_161_WebService_tr_500_1st_137.csv: {'filename': '711_YAHOO_id_161_WebService_tr_500_1st_137.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.8082294711831648, 'VUS-ROC': 0.9995927558914194, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9430250786873685, 'Affiliation-P': 0.8921924822173166, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Logits metrics skipped for 711_YAHOO_id_161_WebService_tr_500_1st_137.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/711_YAHOO_id_161_WebService_tr_500_1st_137_results.pkl
Processing file: 339_UCR_id_37_Sensor_tr_2750_1st_5400.csv, Data shape: (11000, 1)
Label shape: (11000,)
Train data shape: (2750, 1), Test data shape: (8250, 1), Label test shape: (8250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 339_UCR_id_37_Sensor_tr_2750_1st_5400.csv...
Using CUDA device: 2
Data shape: (8250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 8250, window_size 5000, adding 1750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.90it/s]
Original shapes - Output: (10000,), Label: (8250,)
Aligned shapes - Output: (8250,), Label: (8250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 453.88it/s]
Table1-only metrics completed in 0.79s
Evaluation Result:  {'AUC-PR': 0.06358880985994252, 'AUC-ROC': 0.8106347826086956, 'VUS-PR': 0.12096279292610607, 'VUS-ROC': 0.8381266669514833, 'Standard-F1': 0.14452214452188178, 'Standard-Precision': 0.08555657773689045, 'Standard-Recall': 0.4649999999999977, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9160209379650416, 'Affiliation-P': 0.8627994227994227, 'Affiliation-R': 0.97624, 'F1_T': 0.14339058999227744, 'Precision_T': 0.08428446005267771, 'Recall_T': 0.4799999999999976}
Results for 339_UCR_id_37_Sensor_tr_2750_1st_5400.csv: {'filename': '339_UCR_id_37_Sensor_tr_2750_1st_5400.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 81, 'train_index': '2750', 'data_shape': '11000x1', 'output_length': 10000, 'label_length': 8250, 'aligned_length': 8250, 'AUC-PR': 0.06358880985994252, 'AUC-ROC': 0.8106347826086956, 'VUS-PR': 0.12096279292610607, 'VUS-ROC': 0.8381266669514833, 'Standard-F1': 0.14452214452188178, 'Standard-Precision': 0.08555657773689045, 'Standard-Recall': 0.4649999999999977, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9160209379650416, 'Affiliation-P': 0.8627994227994227, 'Affiliation-R': 0.97624, 'F1_T': 0.14339058999227744, 'Precision_T': 0.08428446005267771, 'Recall_T': 0.4799999999999976}
Logits metrics skipped for 339_UCR_id_37_Sensor_tr_2750_1st_5400.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/339_UCR_id_37_Sensor_tr_2750_1st_5400_results.pkl
Processing file: 663_YAHOO_id_113_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 663_YAHOO_id_113_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 161.42it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1400.07it/s]
Table1-only metrics completed in 0.78s
Evaluation Result:  {'AUC-PR': 0.8875, 'AUC-ROC': 0.9994547437295529, 'VUS-PR': 0.9900563438934199, 'VUS-ROC': 0.9999666189733111, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 663_YAHOO_id_113_Synthetic_tr_500_1st_658.csv: {'filename': '663_YAHOO_id_113_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 118, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.8875, 'AUC-ROC': 0.9994547437295529, 'VUS-PR': 0.9900563438934199, 'VUS-ROC': 0.9999666189733111, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 663_YAHOO_id_113_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/663_YAHOO_id_113_Synthetic_tr_500_1st_658_results.pkl
Processing file: 316_UCR_id_14_Sensor_tr_2833_1st_5988.csv, Data shape: (11334, 1)
Label shape: (11334,)
Train data shape: (2833, 1), Test data shape: (8501, 1), Label test shape: (8501,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 316_UCR_id_14_Sensor_tr_2833_1st_5988.csv...
Using CUDA device: 2
Data shape: (8501, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 8501, window_size 5000, adding 1499 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.80it/s]
Original shapes - Output: (10000,), Label: (8501,)
Aligned shapes - Output: (8501,), Label: (8501,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 491.71it/s]
Table1-only metrics completed in 0.82s
Evaluation Result:  {'AUC-PR': 0.11182298123731653, 'AUC-ROC': 0.9582844693494017, 'VUS-PR': 0.15660536513692797, 'VUS-ROC': 0.9665024834774444, 'Standard-F1': 0.2675814751283666, 'Standard-Precision': 0.16049382716049349, 'Standard-Recall': 0.8041237113401979, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9587165510081175, 'Affiliation-P': 0.9209354594077793, 'Affiliation-R': 0.9997301712230339, 'F1_T': 0.2662229617301777, 'Precision_T': 0.15873015873015842, 'Recall_T': 0.8247422680412286}
Results for 316_UCR_id_14_Sensor_tr_2833_1st_5988.csv: {'filename': '316_UCR_id_14_Sensor_tr_2833_1st_5988.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 98, 'train_index': '2833', 'data_shape': '11334x1', 'output_length': 10000, 'label_length': 8501, 'aligned_length': 8501, 'AUC-PR': 0.11182298123731653, 'AUC-ROC': 0.9582844693494017, 'VUS-PR': 0.15660536513692797, 'VUS-ROC': 0.9665024834774444, 'Standard-F1': 0.2675814751283666, 'Standard-Precision': 0.16049382716049349, 'Standard-Recall': 0.8041237113401979, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9587165510081175, 'Affiliation-P': 0.9209354594077793, 'Affiliation-R': 0.9997301712230339, 'F1_T': 0.2662229617301777, 'Precision_T': 0.15873015873015842, 'Recall_T': 0.8247422680412286}
Logits metrics skipped for 316_UCR_id_14_Sensor_tr_2833_1st_5988.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/316_UCR_id_14_Sensor_tr_2833_1st_5988_results.pkl
Processing file: 766_YAHOO_id_216_WebService_tr_500_1st_114.csv, Data shape: (1440, 1)
Label shape: (1440,)
Train data shape: (500, 1), Test data shape: (940, 1), Label test shape: (940,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 766_YAHOO_id_216_WebService_tr_500_1st_114.csv...
Using CUDA device: 2
Data shape: (940, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 940, 1])
batched_mask shape: torch.Size([1, 940])

100%|██████████| 1/1 [00:00<00:00, 159.07it/s]
Original shapes - Output: (940,), Label: (940,)
Aligned shapes - Output: (940,), Label: (940,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2327.67it/s]
Table1-only metrics completed in 0.23s
Evaluation Result:  {'AUC-PR': 0.13534531962953764, 'AUC-ROC': 0.4796559977282869, 'VUS-PR': 0.09462671774669583, 'VUS-ROC': 0.5867697857280038, 'Standard-F1': 0.16216216216175897, 'Standard-Precision': 0.29999999999997, 'Standard-Recall': 0.111111111111107, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8256814776115591, 'Affiliation-P': 0.7182291511582459, 'Affiliation-R': 0.9709414024975984, 'F1_T': 0.16216216216175897, 'Precision_T': 0.29999999999997, 'Recall_T': 0.111111111111107}
Results for 766_YAHOO_id_216_WebService_tr_500_1st_114.csv: {'filename': '766_YAHOO_id_216_WebService_tr_500_1st_114.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 23, 'train_index': '500', 'data_shape': '1440x1', 'output_length': 940, 'label_length': 940, 'aligned_length': 940, 'AUC-PR': 0.13534531962953764, 'AUC-ROC': 0.4796559977282869, 'VUS-PR': 0.09462671774669583, 'VUS-ROC': 0.5867697857280038, 'Standard-F1': 0.16216216216175897, 'Standard-Precision': 0.29999999999997, 'Standard-Recall': 0.111111111111107, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8256814776115591, 'Affiliation-P': 0.7182291511582459, 'Affiliation-R': 0.9709414024975984, 'F1_T': 0.16216216216175897, 'Precision_T': 0.29999999999997, 'Recall_T': 0.111111111111107}
Logits metrics skipped for 766_YAHOO_id_216_WebService_tr_500_1st_114.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/766_YAHOO_id_216_WebService_tr_500_1st_114_results.pkl
Processing file: 444_UCR_id_142_Medical_tr_10336_1st_17001.csv, Data shape: (41347, 1)
Label shape: (41347,)
Train data shape: (10336, 1), Test data shape: (31011, 1), Label test shape: (31011,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 444_UCR_id_142_Medical_tr_10336_1st_17001.csv...
Using CUDA device: 2
Data shape: (31011, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 31011, window_size 5000, adding 3989 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([7, 5000, 1])
batched_mask shape: torch.Size([7, 5000])

100%|██████████| 1/1 [00:00<00:00, 138.04it/s]
Original shapes - Output: (35000,), Label: (31011,)
Aligned shapes - Output: (31011,), Label: (31011,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [00:01<00:00, 117.41it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 135.34it/s]
Table1-only metrics completed in 1.79s
Evaluation Result:  {'AUC-PR': 0.009834807913218349, 'AUC-ROC': 0.9754333892545275, 'VUS-PR': 0.011733402192406256, 'VUS-ROC': 0.9754486250694405, 'Standard-F1': 0.03012048192768111, 'Standard-Precision': 0.015290519877675825, 'Standard-Recall': 0.9999999999999333, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7717271771586298, 'Affiliation-P': 0.62840886587803, 'Affiliation-R': 0.999731278148614, 'F1_T': 0.029914529914498313, 'Precision_T': 0.015200868621064043, 'Recall_T': 0.9333333333332711}
Results for 444_UCR_id_142_Medical_tr_10336_1st_17001.csv: {'filename': '444_UCR_id_142_Medical_tr_10336_1st_17001.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 33, 'train_index': '10336', 'data_shape': '41347x1', 'output_length': 35000, 'label_length': 31011, 'aligned_length': 31011, 'AUC-PR': 0.009834807913218349, 'AUC-ROC': 0.9754333892545275, 'VUS-PR': 0.011733402192406256, 'VUS-ROC': 0.9754486250694405, 'Standard-F1': 0.03012048192768111, 'Standard-Precision': 0.015290519877675825, 'Standard-Recall': 0.9999999999999333, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7717271771586298, 'Affiliation-P': 0.62840886587803, 'Affiliation-R': 0.999731278148614, 'F1_T': 0.029914529914498313, 'Precision_T': 0.015200868621064043, 'Recall_T': 0.9333333333332711}
Logits metrics skipped for 444_UCR_id_142_Medical_tr_10336_1st_17001.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/444_UCR_id_142_Medical_tr_10336_1st_17001_results.pkl
Processing file: 593_YAHOO_id_43_WebService_tr_500_1st_75.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 593_YAHOO_id_43_WebService_tr_500_1st_75.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 163.11it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1467.71it/s]
Table1-only metrics completed in 0.34s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9504941880689357, 'VUS-ROC': 0.9998948142040056, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984878600530774, 'Affiliation-P': 0.9969802863358158, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Results for 593_YAHOO_id_43_WebService_tr_500_1st_75.csv: {'filename': '593_YAHOO_id_43_WebService_tr_500_1st_75.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9504941880689357, 'VUS-ROC': 0.9998948142040056, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984878600530774, 'Affiliation-P': 0.9969802863358158, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Logits metrics skipped for 593_YAHOO_id_43_WebService_tr_500_1st_75.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/593_YAHOO_id_43_WebService_tr_500_1st_75_results.pkl
Processing file: 681_YAHOO_id_131_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 681_YAHOO_id_131_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 161.70it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2023.49it/s]
Table1-only metrics completed in 0.66s
Evaluation Result:  {'AUC-PR': 0.8875, 'AUC-ROC': 0.9994547437295529, 'VUS-PR': 0.9893051815003151, 'VUS-ROC': 0.9999645556378419, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 681_YAHOO_id_131_Synthetic_tr_500_1st_658.csv: {'filename': '681_YAHOO_id_131_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 103, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.8875, 'AUC-ROC': 0.9994547437295529, 'VUS-PR': 0.9893051815003151, 'VUS-ROC': 0.9999645556378419, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 681_YAHOO_id_131_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/681_YAHOO_id_131_Synthetic_tr_500_1st_658_results.pkl
Skipping file: 190_SMD_id_13_Facility_tr_7119_1st_15849.csv due to filter criteria.
Processing file: 724_YAHOO_id_174_WebService_tr_500_1st_1030.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 724_YAHOO_id_174_WebService_tr_500_1st_1030.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 166.03it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1004.93it/s]
Table1-only metrics completed in 0.51s
Evaluation Result:  {'AUC-PR': 0.8204081632653062, 'AUC-ROC': 0.9925106382978723, 'VUS-PR': 0.8270865948735105, 'VUS-ROC': 0.9945907881305708, 'Standard-F1': 0.4705882352936471, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.79999999999984, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9564132606559509, 'Affiliation-P': 0.9164674335140649, 'Affiliation-R': 1.0, 'F1_T': 0.4705882352936471, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.79999999999984}
Results for 724_YAHOO_id_174_WebService_tr_500_1st_1030.csv: {'filename': '724_YAHOO_id_174_WebService_tr_500_1st_1030.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.8204081632653062, 'AUC-ROC': 0.9925106382978723, 'VUS-PR': 0.8270865948735105, 'VUS-ROC': 0.9945907881305708, 'Standard-F1': 0.4705882352936471, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.79999999999984, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9564132606559509, 'Affiliation-P': 0.9164674335140649, 'Affiliation-R': 1.0, 'F1_T': 0.4705882352936471, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.79999999999984}
Logits metrics skipped for 724_YAHOO_id_174_WebService_tr_500_1st_1030.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/724_YAHOO_id_174_WebService_tr_500_1st_1030_results.pkl
Processing file: 297_TODS_id_11_Synthetic_tr_500_1st_7.csv, Data shape: (5000, 1)
Label shape: (5000,)
Train data shape: (500, 1), Test data shape: (4500, 1), Label test shape: (4500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 297_TODS_id_11_Synthetic_tr_500_1st_7.csv...
Using CUDA device: 2
Data shape: (4500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 4500, 1])
batched_mask shape: torch.Size([1, 4500])

100%|██████████| 1/1 [00:00<00:00, 161.08it/s]
Original shapes - Output: (4500,), Label: (4500,)
Aligned shapes - Output: (4500,), Label: (4500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 406.05it/s]
Table1-only metrics completed in 3.99s
Evaluation Result:  {'AUC-PR': 0.2605940393618474, 'AUC-ROC': 0.9032782908344449, 'VUS-PR': 0.6730803365738892, 'VUS-ROC': 0.9634834899741348, 'Standard-F1': 0.4374999999996057, 'Standard-Precision': 0.29923664122137356, 'Standard-Recall': 0.8132780082987519, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8993840917794608, 'Affiliation-P': 0.8513575184285536, 'Affiliation-R': 0.9531531346865671, 'F1_T': 0.4384180790956478, 'Precision_T': 0.30124223602484423, 'Recall_T': 0.8049792531120299}
Results for 297_TODS_id_11_Synthetic_tr_500_1st_7.csv: {'filename': '297_TODS_id_11_Synthetic_tr_500_1st_7.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 98, 'train_index': '500', 'data_shape': '5000x1', 'output_length': 4500, 'label_length': 4500, 'aligned_length': 4500, 'AUC-PR': 0.2605940393618474, 'AUC-ROC': 0.9032782908344449, 'VUS-PR': 0.6730803365738892, 'VUS-ROC': 0.9634834899741348, 'Standard-F1': 0.4374999999996057, 'Standard-Precision': 0.29923664122137356, 'Standard-Recall': 0.8132780082987519, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8993840917794608, 'Affiliation-P': 0.8513575184285536, 'Affiliation-R': 0.9531531346865671, 'F1_T': 0.4384180790956478, 'Precision_T': 0.30124223602484423, 'Recall_T': 0.8049792531120299}
Logits metrics skipped for 297_TODS_id_11_Synthetic_tr_500_1st_7.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/297_TODS_id_11_Synthetic_tr_500_1st_7_results.pkl
Skipping file: 546_SMAP_id_16_Sensor_tr_2117_1st_4770.csv due to filter criteria.
Processing file: 114_WSD_id_86_WebService_tr_500_1st_403.csv, Data shape: (18243, 1)
Label shape: (18243,)
Train data shape: (500, 1), Test data shape: (17743, 1), Label test shape: (17743,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 114_WSD_id_86_WebService_tr_500_1st_403.csv...
Using CUDA device: 2
Data shape: (17743, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 17743, window_size 5000, adding 2257 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.87it/s]
Original shapes - Output: (20000,), Label: (17743,)
Aligned shapes - Output: (17743,), Label: (17743,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [00:01<00:00, 126.97it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 151.96it/s]
Table1-only metrics completed in 3.62s
Evaluation Result:  {'AUC-PR': 0.23415225064342848, 'AUC-ROC': 0.9131155197368, 'VUS-PR': 0.2988022710688163, 'VUS-ROC': 0.9892372344261995, 'Standard-F1': 0.2131661442002796, 'Standard-Precision': 0.13709677419354785, 'Standard-Recall': 0.47887323943661303, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9972791123010087, 'Affiliation-P': 0.9963707132337472, 'Affiliation-R': 0.9981891692690352, 'F1_T': 0.21428571428537968, 'Precision_T': 0.13584905660377306, 'Recall_T': 0.5070422535211196}
Results for 114_WSD_id_86_WebService_tr_500_1st_403.csv: {'filename': '114_WSD_id_86_WebService_tr_500_1st_403.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '18243x1', 'output_length': 20000, 'label_length': 17743, 'aligned_length': 17743, 'AUC-PR': 0.23415225064342848, 'AUC-ROC': 0.9131155197368, 'VUS-PR': 0.2988022710688163, 'VUS-ROC': 0.9892372344261995, 'Standard-F1': 0.2131661442002796, 'Standard-Precision': 0.13709677419354785, 'Standard-Recall': 0.47887323943661303, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9972791123010087, 'Affiliation-P': 0.9963707132337472, 'Affiliation-R': 0.9981891692690352, 'F1_T': 0.21428571428537968, 'Precision_T': 0.13584905660377306, 'Recall_T': 0.5070422535211196}
Logits metrics skipped for 114_WSD_id_86_WebService_tr_500_1st_403.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/114_WSD_id_86_WebService_tr_500_1st_403_results.pkl
Processing file: 031_WSD_id_3_WebService_tr_826_1st_926.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (826, 1), Test data shape: (17410, 1), Label test shape: (17410,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 031_WSD_id_3_WebService_tr_826_1st_926.csv...
Using CUDA device: 2
Data shape: (17410, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 17410, window_size 5000, adding 2590 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.49it/s]
Original shapes - Output: (20000,), Label: (17410,)
Aligned shapes - Output: (17410,), Label: (17410,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 239.84it/s]
Table1-only metrics completed in 1.46s
Evaluation Result:  {'AUC-PR': 0.563995437583089, 'AUC-ROC': 0.9986073163495305, 'VUS-PR': 0.41735356017229913, 'VUS-ROC': 0.998229361596785, 'Standard-F1': 0.3644859813081097, 'Standard-Precision': 0.2228571428571416, 'Standard-Recall': 0.9999999999999744, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8469231409201826, 'Affiliation-P': 0.7344913450994727, 'Affiliation-R': 0.9999970544485192, 'F1_T': 0.3644859813081097, 'Precision_T': 0.2228571428571416, 'Recall_T': 0.9999999999999744}
Results for 031_WSD_id_3_WebService_tr_826_1st_926.csv: {'filename': '031_WSD_id_3_WebService_tr_826_1st_926.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '826', 'data_shape': '18236x1', 'output_length': 20000, 'label_length': 17410, 'aligned_length': 17410, 'AUC-PR': 0.563995437583089, 'AUC-ROC': 0.9986073163495305, 'VUS-PR': 0.41735356017229913, 'VUS-ROC': 0.998229361596785, 'Standard-F1': 0.3644859813081097, 'Standard-Precision': 0.2228571428571416, 'Standard-Recall': 0.9999999999999744, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8469231409201826, 'Affiliation-P': 0.7344913450994727, 'Affiliation-R': 0.9999970544485192, 'F1_T': 0.3644859813081097, 'Precision_T': 0.2228571428571416, 'Recall_T': 0.9999999999999744}
Logits metrics skipped for 031_WSD_id_3_WebService_tr_826_1st_926.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/031_WSD_id_3_WebService_tr_826_1st_926_results.pkl
Processing file: 401_UCR_id_99_HumanActivity_tr_1671_1st_5920.csv, Data shape: (6684, 1)
Label shape: (6684,)
Train data shape: (1671, 1), Test data shape: (5013, 1), Label test shape: (5013,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 401_UCR_id_99_HumanActivity_tr_1671_1st_5920.csv...
Using CUDA device: 2
Data shape: (5013, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5013, window_size 5000, adding 4987 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.56it/s]
Original shapes - Output: (10000,), Label: (5013,)
Aligned shapes - Output: (5013,), Label: (5013,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 657.84it/s]
Table1-only metrics completed in 0.74s
Evaluation Result:  {'AUC-PR': 0.9157186232635082, 'AUC-ROC': 0.9982961893487885, 'VUS-PR': 0.8912667737263589, 'VUS-ROC': 0.9985254377159078, 'Standard-F1': 0.8363636363631237, 'Standard-Precision': 0.9019607843137077, 'Standard-Recall': 0.7796610169491393, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9997091460053227, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9994184611535432, 'F1_T': 0.8363636363631237, 'Precision_T': 0.9019607843137077, 'Recall_T': 0.7796610169491393}
Results for 401_UCR_id_99_HumanActivity_tr_1671_1st_5920.csv: {'filename': '401_UCR_id_99_HumanActivity_tr_1671_1st_5920.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 115, 'train_index': '1671', 'data_shape': '6684x1', 'output_length': 10000, 'label_length': 5013, 'aligned_length': 5013, 'AUC-PR': 0.9157186232635082, 'AUC-ROC': 0.9982961893487885, 'VUS-PR': 0.8912667737263589, 'VUS-ROC': 0.9985254377159078, 'Standard-F1': 0.8363636363631237, 'Standard-Precision': 0.9019607843137077, 'Standard-Recall': 0.7796610169491393, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9997091460053227, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9994184611535432, 'F1_T': 0.8363636363631237, 'Precision_T': 0.9019607843137077, 'Recall_T': 0.7796610169491393}
Logits metrics skipped for 401_UCR_id_99_HumanActivity_tr_1671_1st_5920.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/401_UCR_id_99_HumanActivity_tr_1671_1st_5920_results.pkl
Processing file: 474_UCR_id_172_Environment_tr_2046_1st_6520.csv, Data shape: (8184, 1)
Label shape: (8184,)
Train data shape: (2046, 1), Test data shape: (6138, 1), Label test shape: (6138,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 474_UCR_id_172_Environment_tr_2046_1st_6520.csv...
Using CUDA device: 2
Data shape: (6138, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6138, window_size 5000, adding 3862 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.81it/s]
Original shapes - Output: (10000,), Label: (6138,)
Aligned shapes - Output: (6138,), Label: (6138,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 503.83it/s]
Table1-only metrics completed in 0.52s
Evaluation Result:  {'AUC-PR': 0.311032860672421, 'AUC-ROC': 0.9928715516301384, 'VUS-PR': 0.3638326634899534, 'VUS-ROC': 0.9953278754322455, 'Standard-F1': 0.37209302325540294, 'Standard-Precision': 0.2580645161290281, 'Standard-Recall': 0.666666666666639, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8817306313373292, 'Affiliation-P': 0.7885580571584734, 'Affiliation-R': 0.9998710220484414, 'F1_T': 0.37209302325540294, 'Precision_T': 0.2580645161290281, 'Recall_T': 0.666666666666639}
Results for 474_UCR_id_172_Environment_tr_2046_1st_6520.csv: {'filename': '474_UCR_id_172_Environment_tr_2046_1st_6520.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '2046', 'data_shape': '8184x1', 'output_length': 10000, 'label_length': 6138, 'aligned_length': 6138, 'AUC-PR': 0.311032860672421, 'AUC-ROC': 0.9928715516301384, 'VUS-PR': 0.3638326634899534, 'VUS-ROC': 0.9953278754322455, 'Standard-F1': 0.37209302325540294, 'Standard-Precision': 0.2580645161290281, 'Standard-Recall': 0.666666666666639, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8817306313373292, 'Affiliation-P': 0.7885580571584734, 'Affiliation-R': 0.9998710220484414, 'F1_T': 0.37209302325540294, 'Precision_T': 0.2580645161290281, 'Recall_T': 0.666666666666639}
Logits metrics skipped for 474_UCR_id_172_Environment_tr_2046_1st_6520.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/474_UCR_id_172_Environment_tr_2046_1st_6520_results.pkl
Processing file: 268_IOPS_id_9_WebService_tr_6216_1st_6316.csv, Data shape: (29133, 1)
Label shape: (29133,)
Train data shape: (6216, 1), Test data shape: (22917, 1), Label test shape: (22917,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 268_IOPS_id_9_WebService_tr_6216_1st_6316.csv...
Using CUDA device: 2
Data shape: (22917, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22917, window_size 5000, adding 2083 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 153.01it/s]
Original shapes - Output: (25000,), Label: (22917,)
Aligned shapes - Output: (22917,), Label: (22917,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  46%|████▋     | 93/200 [00:01<00:01, 92.62it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 104.78it/s]
Table1-only metrics completed in 2.74s
Evaluation Result:  {'AUC-PR': 0.9535852306227045, 'AUC-ROC': 0.9948841183843233, 'VUS-PR': 0.9549156941651796, 'VUS-ROC': 0.9962217093997054, 'Standard-F1': 0.8925009235310847, 'Standard-Precision': 0.922137404580152, 'Standard-Recall': 0.8647100930565492, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9965465640135456, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9931168983779325, 'F1_T': 0.8885662431936919, 'Precision_T': 0.9013254786450656, 'Recall_T': 0.8761632068718678}
Results for 268_IOPS_id_9_WebService_tr_6216_1st_6316.csv: {'filename': '268_IOPS_id_9_WebService_tr_6216_1st_6316.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 20, 'train_index': '6216', 'data_shape': '29133x1', 'output_length': 25000, 'label_length': 22917, 'aligned_length': 22917, 'AUC-PR': 0.9535852306227045, 'AUC-ROC': 0.9948841183843233, 'VUS-PR': 0.9549156941651796, 'VUS-ROC': 0.9962217093997054, 'Standard-F1': 0.8925009235310847, 'Standard-Precision': 0.922137404580152, 'Standard-Recall': 0.8647100930565492, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9965465640135456, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9931168983779325, 'F1_T': 0.8885662431936919, 'Precision_T': 0.9013254786450656, 'Recall_T': 0.8761632068718678}
Logits metrics skipped for 268_IOPS_id_9_WebService_tr_6216_1st_6316.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/268_IOPS_id_9_WebService_tr_6216_1st_6316_results.pkl
Processing file: 723_YAHOO_id_173_WebService_tr_500_1st_1214.csv, Data shape: (1420, 1)
Label shape: (1420,)
Train data shape: (500, 1), Test data shape: (920, 1), Label test shape: (920,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 723_YAHOO_id_173_WebService_tr_500_1st_1214.csv...
Using CUDA device: 2
Data shape: (920, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 920, 1])
batched_mask shape: torch.Size([1, 920])

100%|██████████| 1/1 [00:00<00:00, 158.93it/s]
Original shapes - Output: (920,), Label: (920,)
Aligned shapes - Output: (920,), Label: (920,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2191.66it/s]
Table1-only metrics completed in 0.27s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.905329700552229, 'VUS-ROC': 0.9997936499664017, 'Standard-F1': 0.33333333333299997, 'Standard-Precision': 0.19999999999998, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9903904870380614, 'Affiliation-P': 0.9809639017093923, 'Affiliation-R': 1.0, 'F1_T': 0.33333333333299997, 'Precision_T': 0.19999999999998, 'Recall_T': 0.9999999999995}
Results for 723_YAHOO_id_173_WebService_tr_500_1st_1214.csv: {'filename': '723_YAHOO_id_173_WebService_tr_500_1st_1214.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 48, 'train_index': '500', 'data_shape': '1420x1', 'output_length': 920, 'label_length': 920, 'aligned_length': 920, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.905329700552229, 'VUS-ROC': 0.9997936499664017, 'Standard-F1': 0.33333333333299997, 'Standard-Precision': 0.19999999999998, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9903904870380614, 'Affiliation-P': 0.9809639017093923, 'Affiliation-R': 1.0, 'F1_T': 0.33333333333299997, 'Precision_T': 0.19999999999998, 'Recall_T': 0.9999999999995}
Logits metrics skipped for 723_YAHOO_id_173_WebService_tr_500_1st_1214.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/723_YAHOO_id_173_WebService_tr_500_1st_1214_results.pkl
Processing file: 595_YAHOO_id_45_WebService_tr_500_1st_248.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 595_YAHOO_id_45_WebService_tr_500_1st_248.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 165.09it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 952.87it/s]
Table1-only metrics completed in 0.47s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.9999444925344662, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9991383000605965, 'Affiliation-P': 0.9982780838961962, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Results for 595_YAHOO_id_45_WebService_tr_500_1st_248.csv: {'filename': '595_YAHOO_id_45_WebService_tr_500_1st_248.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.9999444925344662, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9991383000605965, 'Affiliation-P': 0.9982780838961962, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 595_YAHOO_id_45_WebService_tr_500_1st_248.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/595_YAHOO_id_45_WebService_tr_500_1st_248_results.pkl
Processing file: 012_NAB_id_12_Synthetic_tr_1007_1st_2787.csv, Data shape: (4031, 1)
Label shape: (4031,)
Train data shape: (1007, 1), Test data shape: (3024, 1), Label test shape: (3024,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 012_NAB_id_12_Synthetic_tr_1007_1st_2787.csv...
Using CUDA device: 2
Data shape: (3024, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 3024, 1])
batched_mask shape: torch.Size([1, 3024])

100%|██████████| 1/1 [00:00<00:00, 163.22it/s]
Original shapes - Output: (3024,), Label: (3024,)
Aligned shapes - Output: (3024,), Label: (3024,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1138.23it/s]
Table1-only metrics completed in 1.29s
Evaluation Result:  {'AUC-PR': 0.9762867985710588, 'AUC-ROC': 0.9942025070197952, 'VUS-PR': 0.9926892348321498, 'VUS-ROC': 0.9984667349085112, 'Standard-F1': 0.9120323559145644, 'Standard-Precision': 0.9474789915966366, 'Standard-Recall': 0.87914230019493, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9985143744160582, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9970331564507977, 'F1_T': 0.9111111111106097, 'Precision_T': 0.9454926624737925, 'Recall_T': 0.87914230019493}
Results for 012_NAB_id_12_Synthetic_tr_1007_1st_2787.csv: {'filename': '012_NAB_id_12_Synthetic_tr_1007_1st_2787.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 288, 'train_index': '1007', 'data_shape': '4031x1', 'output_length': 3024, 'label_length': 3024, 'aligned_length': 3024, 'AUC-PR': 0.9762867985710588, 'AUC-ROC': 0.9942025070197952, 'VUS-PR': 0.9926892348321498, 'VUS-ROC': 0.9984667349085112, 'Standard-F1': 0.9120323559145644, 'Standard-Precision': 0.9474789915966366, 'Standard-Recall': 0.87914230019493, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9985143744160582, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9970331564507977, 'F1_T': 0.9111111111106097, 'Precision_T': 0.9454926624737925, 'Recall_T': 0.87914230019493}
Logits metrics skipped for 012_NAB_id_12_Synthetic_tr_1007_1st_2787.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/012_NAB_id_12_Synthetic_tr_1007_1st_2787_results.pkl
Processing file: 651_YAHOO_id_101_WebService_tr_500_1st_103.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 651_YAHOO_id_101_WebService_tr_500_1st_103.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 165.48it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1638.98it/s]
Table1-only metrics completed in 0.32s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.7974693162566899, 'VUS-ROC': 0.999569669435519, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9532619479809435, 'Affiliation-P': 0.9106977109910112, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Results for 651_YAHOO_id_101_WebService_tr_500_1st_103.csv: {'filename': '651_YAHOO_id_101_WebService_tr_500_1st_103.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.7974693162566899, 'VUS-ROC': 0.999569669435519, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9532619479809435, 'Affiliation-P': 0.9106977109910112, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Logits metrics skipped for 651_YAHOO_id_101_WebService_tr_500_1st_103.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/651_YAHOO_id_101_WebService_tr_500_1st_103_results.pkl
Skipping file: 221_LTDB_id_6_Medical_tr_500_1st_26.csv due to filter criteria.
Skipping file: 258_TAO_id_2_Environment_tr_500_1st_4.csv due to filter criteria.
Skipping file: 170_MITDB_id_1_Medical_tr_17675_1st_17775.csv due to filter criteria.
Skipping file: 148_MSL_id_9_Sensor_tr_883_1st_1238.csv due to filter criteria.
Processing file: 329_UCR_id_27_Medical_tr_7500_1st_17210.csv, Data shape: (30001, 1)
Label shape: (30001,)
Train data shape: (7500, 1), Test data shape: (22501, 1), Label test shape: (22501,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 329_UCR_id_27_Medical_tr_7500_1st_17210.csv...
Using CUDA device: 2
Data shape: (22501, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22501, window_size 5000, adding 2499 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 144.47it/s]
Original shapes - Output: (25000,), Label: (22501,)
Aligned shapes - Output: (22501,), Label: (22501,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  65%|██████▌   | 130/200 [00:01<00:00, 129.03it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 137.33it/s]
Table1-only metrics completed in 1.71s
Evaluation Result:  {'AUC-PR': 0.55366634860157, 'AUC-ROC': 0.9622867578281591, 'VUS-PR': 0.2915424623497032, 'VUS-ROC': 0.976546309003091, 'Standard-F1': 0.2681159420286869, 'Standard-Precision': 0.16371681415929132, 'Standard-Recall': 0.7399999999999852, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666842169313524, 'Affiliation-P': 0.5000197443076604, 'Affiliation-R': 1.0, 'F1_T': 0.2681159420286869, 'Precision_T': 0.16371681415929132, 'Recall_T': 0.7399999999999852}
Results for 329_UCR_id_27_Medical_tr_7500_1st_17210.csv: {'filename': '329_UCR_id_27_Medical_tr_7500_1st_17210.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 34, 'train_index': '7500', 'data_shape': '30001x1', 'output_length': 25000, 'label_length': 22501, 'aligned_length': 22501, 'AUC-PR': 0.55366634860157, 'AUC-ROC': 0.9622867578281591, 'VUS-PR': 0.2915424623497032, 'VUS-ROC': 0.976546309003091, 'Standard-F1': 0.2681159420286869, 'Standard-Precision': 0.16371681415929132, 'Standard-Recall': 0.7399999999999852, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666842169313524, 'Affiliation-P': 0.5000197443076604, 'Affiliation-R': 1.0, 'F1_T': 0.2681159420286869, 'Precision_T': 0.16371681415929132, 'Recall_T': 0.7399999999999852}
Logits metrics skipped for 329_UCR_id_27_Medical_tr_7500_1st_17210.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/329_UCR_id_27_Medical_tr_7500_1st_17210_results.pkl
Processing file: 712_YAHOO_id_162_WebService_tr_500_1st_88.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 712_YAHOO_id_162_WebService_tr_500_1st_88.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 140.67it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1186.53it/s]
Table1-only metrics completed in 0.47s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5882352941171626, 'Standard-Precision': 0.41666666666663194, 'Standard-Recall': 0.9999999999997999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9993181043300369, 'Affiliation-P': 0.9986371379897786, 'Affiliation-R': 1.0, 'F1_T': 0.5882352941171626, 'Precision_T': 0.41666666666663194, 'Recall_T': 0.9999999999997999}
Results for 712_YAHOO_id_162_WebService_tr_500_1st_88.csv: {'filename': '712_YAHOO_id_162_WebService_tr_500_1st_88.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5882352941171626, 'Standard-Precision': 0.41666666666663194, 'Standard-Recall': 0.9999999999997999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9993181043300369, 'Affiliation-P': 0.9986371379897786, 'Affiliation-R': 1.0, 'F1_T': 0.5882352941171626, 'Precision_T': 0.41666666666663194, 'Recall_T': 0.9999999999997999}
Logits metrics skipped for 712_YAHOO_id_162_WebService_tr_500_1st_88.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/712_YAHOO_id_162_WebService_tr_500_1st_88_results.pkl
Processing file: 488_UCR_id_186_Medical_tr_7500_1st_16000.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 488_UCR_id_186_Medical_tr_7500_1st_16000.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 153.78it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  85%|████████▌ | 170/200 [00:01<00:00, 169.31it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 181.31it/s]
Table1-only metrics completed in 1.64s
Evaluation Result:  {'AUC-PR': 0.04469601759288082, 'AUC-ROC': 0.9351535714285714, 'VUS-PR': 0.052459651741537244, 'VUS-ROC': 0.9481418021690473, 'Standard-F1': 0.12598425196823818, 'Standard-Precision': 0.07476635514018677, 'Standard-Recall': 0.399999999999996, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6757215652668896, 'Affiliation-P': 0.5105974087161366, 'Affiliation-R': 0.9986928888888887, 'F1_T': 0.12158054711220388, 'Precision_T': 0.07168458781361994, 'Recall_T': 0.399999999999996}
Results for 488_UCR_id_186_Medical_tr_7500_1st_16000.csv: {'filename': '488_UCR_id_186_Medical_tr_7500_1st_16000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 91, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.04469601759288082, 'AUC-ROC': 0.9351535714285714, 'VUS-PR': 0.052459651741537244, 'VUS-ROC': 0.9481418021690473, 'Standard-F1': 0.12598425196823818, 'Standard-Precision': 0.07476635514018677, 'Standard-Recall': 0.399999999999996, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6757215652668896, 'Affiliation-P': 0.5105974087161366, 'Affiliation-R': 0.9986928888888887, 'F1_T': 0.12158054711220388, 'Precision_T': 0.07168458781361994, 'Recall_T': 0.399999999999996}
Logits metrics skipped for 488_UCR_id_186_Medical_tr_7500_1st_16000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/488_UCR_id_186_Medical_tr_7500_1st_16000_results.pkl
Processing file: 521_UCR_id_219_Facility_tr_3000_1st_6527.csv, Data shape: (12000, 1)
Label shape: (12000,)
Train data shape: (3000, 1), Test data shape: (9000, 1), Label test shape: (9000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 521_UCR_id_219_Facility_tr_3000_1st_6527.csv...
Using CUDA device: 2
Data shape: (9000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9000, window_size 5000, adding 1000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.52it/s]
Original shapes - Output: (10000,), Label: (9000,)
Aligned shapes - Output: (9000,), Label: (9000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 466.35it/s]
Table1-only metrics completed in 1.32s
Evaluation Result:  {'AUC-PR': 0.4354373889592209, 'AUC-ROC': 0.8302723275792976, 'VUS-PR': 0.38587876684743844, 'VUS-ROC': 0.8806709728622493, 'Standard-F1': 0.4615384615379662, 'Standard-Precision': 0.5333333333333274, 'Standard-Recall': 0.40677966101694574, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9827385860882116, 'Affiliation-P': 0.969072463768116, 'Affiliation-R': 0.9967956685499059, 'F1_T': 0.4615384615379662, 'Precision_T': 0.5333333333333274, 'Recall_T': 0.40677966101694574}
Results for 521_UCR_id_219_Facility_tr_3000_1st_6527.csv: {'filename': '521_UCR_id_219_Facility_tr_3000_1st_6527.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 218, 'train_index': '3000', 'data_shape': '12000x1', 'output_length': 10000, 'label_length': 9000, 'aligned_length': 9000, 'AUC-PR': 0.4354373889592209, 'AUC-ROC': 0.8302723275792976, 'VUS-PR': 0.38587876684743844, 'VUS-ROC': 0.8806709728622493, 'Standard-F1': 0.4615384615379662, 'Standard-Precision': 0.5333333333333274, 'Standard-Recall': 0.40677966101694574, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9827385860882116, 'Affiliation-P': 0.969072463768116, 'Affiliation-R': 0.9967956685499059, 'F1_T': 0.4615384615379662, 'Precision_T': 0.5333333333333274, 'Recall_T': 0.40677966101694574}
Logits metrics skipped for 521_UCR_id_219_Facility_tr_3000_1st_6527.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/521_UCR_id_219_Facility_tr_3000_1st_6527_results.pkl
Processing file: 475_UCR_id_173_Medical_tr_3000_1st_8612.csv, Data shape: (12000, 1)
Label shape: (12000,)
Train data shape: (3000, 1), Test data shape: (9000, 1), Label test shape: (9000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 475_UCR_id_173_Medical_tr_3000_1st_8612.csv...
Using CUDA device: 2
Data shape: (9000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9000, window_size 5000, adding 1000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.10it/s]
Original shapes - Output: (10000,), Label: (9000,)
Aligned shapes - Output: (9000,), Label: (9000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 524.98it/s]
Table1-only metrics completed in 1.27s
Evaluation Result:  {'AUC-PR': 0.7682573354365398, 'AUC-ROC': 0.9958764959186497, 'VUS-PR': 0.7533917208572739, 'VUS-ROC': 0.9965042813282474, 'Standard-F1': 0.6804123711335162, 'Standard-Precision': 0.7333333333333252, 'Standard-Recall': 0.6346153846153786, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9734959236739986, 'Affiliation-P': 0.9518550724637681, 'Affiliation-R': 0.9961436965811965, 'F1_T': 0.6804123711335162, 'Precision_T': 0.7333333333333252, 'Recall_T': 0.6346153846153786}
Results for 475_UCR_id_173_Medical_tr_3000_1st_8612.csv: {'filename': '475_UCR_id_173_Medical_tr_3000_1st_8612.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 218, 'train_index': '3000', 'data_shape': '12000x1', 'output_length': 10000, 'label_length': 9000, 'aligned_length': 9000, 'AUC-PR': 0.7682573354365398, 'AUC-ROC': 0.9958764959186497, 'VUS-PR': 0.7533917208572739, 'VUS-ROC': 0.9965042813282474, 'Standard-F1': 0.6804123711335162, 'Standard-Precision': 0.7333333333333252, 'Standard-Recall': 0.6346153846153786, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9734959236739986, 'Affiliation-P': 0.9518550724637681, 'Affiliation-R': 0.9961436965811965, 'F1_T': 0.6804123711335162, 'Precision_T': 0.7333333333333252, 'Recall_T': 0.6346153846153786}
Logits metrics skipped for 475_UCR_id_173_Medical_tr_3000_1st_8612.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/475_UCR_id_173_Medical_tr_3000_1st_8612_results.pkl
Processing file: 312_UCR_id_10_Facility_tr_19948_1st_52000.csv, Data shape: (79795, 1)
Label shape: (79795,)
Train data shape: (19948, 1), Test data shape: (59847, 1), Label test shape: (59847,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 312_UCR_id_10_Facility_tr_19948_1st_52000.csv...
Using CUDA device: 2
Data shape: (59847, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 59847, window_size 5000, adding 153 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([12, 5000, 1])
batched_mask shape: torch.Size([12, 5000])

100%|██████████| 1/1 [00:00<00:00, 92.79it/s]
Original shapes - Output: (60000,), Label: (59847,)
Aligned shapes - Output: (59847,), Label: (59847,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  32%|███▏      | 63/200 [00:01<00:02, 62.18it/s]

Calculating Affiliation Metrics:  64%|██████▍   | 128/200 [00:02<00:01, 63.76it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 73.52it/s]
Table1-only metrics completed in 5.07s
Evaluation Result:  {'AUC-PR': 0.05912486490454255, 'AUC-ROC': 0.8395408834353557, 'VUS-PR': 0.06436139601210109, 'VUS-ROC': 0.8603478702824161, 'Standard-F1': 0.1742107265116212, 'Standard-Precision': 0.11398705823792926, 'Standard-Recall': 0.36935483870967684, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7312014750758641, 'Affiliation-P': 0.5767985150674899, 'Affiliation-R': 0.9984857084490181, 'F1_T': 0.17378388414372267, 'Precision_T': 0.11287988422575972, 'Recall_T': 0.37741935483870903}
Results for 312_UCR_id_10_Facility_tr_19948_1st_52000.csv: {'filename': '312_UCR_id_10_Facility_tr_19948_1st_52000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 208, 'train_index': '19948', 'data_shape': '79795x1', 'output_length': 60000, 'label_length': 59847, 'aligned_length': 59847, 'AUC-PR': 0.05912486490454255, 'AUC-ROC': 0.8395408834353557, 'VUS-PR': 0.06436139601210109, 'VUS-ROC': 0.8603478702824161, 'Standard-F1': 0.1742107265116212, 'Standard-Precision': 0.11398705823792926, 'Standard-Recall': 0.36935483870967684, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7312014750758641, 'Affiliation-P': 0.5767985150674899, 'Affiliation-R': 0.9984857084490181, 'F1_T': 0.17378388414372267, 'Precision_T': 0.11287988422575972, 'Recall_T': 0.37741935483870903}
Logits metrics skipped for 312_UCR_id_10_Facility_tr_19948_1st_52000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/312_UCR_id_10_Facility_tr_19948_1st_52000_results.pkl
Processing file: 419_UCR_id_117_Facility_tr_3000_1st_6168.csv, Data shape: (12000, 1)
Label shape: (12000,)
Train data shape: (3000, 1), Test data shape: (9000, 1), Label test shape: (9000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 419_UCR_id_117_Facility_tr_3000_1st_6168.csv...
Using CUDA device: 2
Data shape: (9000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9000, window_size 5000, adding 1000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 154.37it/s]
Original shapes - Output: (10000,), Label: (9000,)
Aligned shapes - Output: (9000,), Label: (9000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 462.51it/s]
Table1-only metrics completed in 1.31s
Evaluation Result:  {'AUC-PR': 0.005700701131073453, 'AUC-ROC': 0.5713919566364855, 'VUS-PR': 0.010915640835421887, 'VUS-ROC': 0.6907392600931594, 'Standard-F1': 0.021133525456271108, 'Standard-Precision': 0.010679611650485435, 'Standard-Recall': 0.9999999999999772, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6747130590878914, 'Affiliation-P': 0.5091071512585275, 'Affiliation-R': 1.0, 'F1_T': 0.02111831053513628, 'Precision_T': 0.010671840892553963, 'Recall_T': 0.9999999999999772}
Results for 419_UCR_id_117_Facility_tr_3000_1st_6168.csv: {'filename': '419_UCR_id_117_Facility_tr_3000_1st_6168.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 218, 'train_index': '3000', 'data_shape': '12000x1', 'output_length': 10000, 'label_length': 9000, 'aligned_length': 9000, 'AUC-PR': 0.005700701131073453, 'AUC-ROC': 0.5713919566364855, 'VUS-PR': 0.010915640835421887, 'VUS-ROC': 0.6907392600931594, 'Standard-F1': 0.021133525456271108, 'Standard-Precision': 0.010679611650485435, 'Standard-Recall': 0.9999999999999772, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6747130590878914, 'Affiliation-P': 0.5091071512585275, 'Affiliation-R': 1.0, 'F1_T': 0.02111831053513628, 'Precision_T': 0.010671840892553963, 'Recall_T': 0.9999999999999772}
Logits metrics skipped for 419_UCR_id_117_Facility_tr_3000_1st_6168.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/419_UCR_id_117_Facility_tr_3000_1st_6168_results.pkl
Processing file: 682_YAHOO_id_132_WebService_tr_500_1st_354.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 682_YAHOO_id_132_WebService_tr_500_1st_354.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.59it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1128.63it/s]
Table1-only metrics completed in 0.39s
Evaluation Result:  {'AUC-PR': 0.9015471167369902, 'AUC-ROC': 0.9933580036056552, 'VUS-PR': 0.8545968668427263, 'VUS-ROC': 0.9924763586170068, 'Standard-F1': 0.7619047619041995, 'Standard-Precision': 0.6666666666666111, 'Standard-Recall': 0.8888888888887901, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9263607257203836, 'Affiliation-P': 0.9841269841269842, 'Affiliation-R': 0.875, 'F1_T': 0.7619047619041995, 'Precision_T': 0.6666666666666111, 'Recall_T': 0.8888888888887901}
Results for 682_YAHOO_id_132_WebService_tr_500_1st_354.csv: {'filename': '682_YAHOO_id_132_WebService_tr_500_1st_354.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 11, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9015471167369902, 'AUC-ROC': 0.9933580036056552, 'VUS-PR': 0.8545968668427263, 'VUS-ROC': 0.9924763586170068, 'Standard-F1': 0.7619047619041995, 'Standard-Precision': 0.6666666666666111, 'Standard-Recall': 0.8888888888887901, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9263607257203836, 'Affiliation-P': 0.9841269841269842, 'Affiliation-R': 0.875, 'F1_T': 0.7619047619041995, 'Precision_T': 0.6666666666666111, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 682_YAHOO_id_132_WebService_tr_500_1st_354.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/682_YAHOO_id_132_WebService_tr_500_1st_354_results.pkl
Processing file: 644_YAHOO_id_94_WebService_tr_500_1st_1078.csv, Data shape: (1435, 1)
Label shape: (1435,)
Train data shape: (500, 1), Test data shape: (935, 1), Label test shape: (935,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 644_YAHOO_id_94_WebService_tr_500_1st_1078.csv...
Using CUDA device: 2
Data shape: (935, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 935, 1])
batched_mask shape: torch.Size([1, 935])

100%|██████████| 1/1 [00:00<00:00, 156.41it/s]
Original shapes - Output: (935,), Label: (935,)
Aligned shapes - Output: (935,), Label: (935,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1375.72it/s]
Table1-only metrics completed in 0.26s
Evaluation Result:  {'AUC-PR': 0.8712988543877062, 'AUC-ROC': 0.9900930232558139, 'VUS-PR': 0.9286712656967522, 'VUS-ROC': 0.9948541349222643, 'Standard-F1': 0.8571428571423576, 'Standard-Precision': 0.7499999999999926, 'Standard-Recall': 0.9999999999999868, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.964921195560559, 'Affiliation-P': 0.9360749321640025, 'Affiliation-R': 0.9956018518518519, 'F1_T': 0.8488372093018239, 'Precision_T': 0.7525773195876212, 'Recall_T': 0.9733333333333204}
Results for 644_YAHOO_id_94_WebService_tr_500_1st_1078.csv: {'filename': '644_YAHOO_id_94_WebService_tr_500_1st_1078.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1435x1', 'output_length': 935, 'label_length': 935, 'aligned_length': 935, 'AUC-PR': 0.8712988543877062, 'AUC-ROC': 0.9900930232558139, 'VUS-PR': 0.9286712656967522, 'VUS-ROC': 0.9948541349222643, 'Standard-F1': 0.8571428571423576, 'Standard-Precision': 0.7499999999999926, 'Standard-Recall': 0.9999999999999868, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.964921195560559, 'Affiliation-P': 0.9360749321640025, 'Affiliation-R': 0.9956018518518519, 'F1_T': 0.8488372093018239, 'Precision_T': 0.7525773195876212, 'Recall_T': 0.9733333333333204}
Logits metrics skipped for 644_YAHOO_id_94_WebService_tr_500_1st_1078.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/644_YAHOO_id_94_WebService_tr_500_1st_1078_results.pkl
Processing file: 641_YAHOO_id_91_WebService_tr_500_1st_149.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 641_YAHOO_id_91_WebService_tr_500_1st_149.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.25it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1074.34it/s]
Table1-only metrics completed in 0.54s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9474402272851752, 'VUS-ROC': 0.9997759784375188, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Results for 641_YAHOO_id_91_WebService_tr_500_1st_149.csv: {'filename': '641_YAHOO_id_91_WebService_tr_500_1st_149.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9474402272851752, 'VUS-ROC': 0.9997759784375188, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Logits metrics skipped for 641_YAHOO_id_91_WebService_tr_500_1st_149.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/641_YAHOO_id_91_WebService_tr_500_1st_149_results.pkl
Processing file: 448_UCR_id_146_Facility_tr_7482_1st_23357.csv, Data shape: (29931, 1)
Label shape: (29931,)
Train data shape: (7482, 1), Test data shape: (22449, 1), Label test shape: (22449,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 448_UCR_id_146_Facility_tr_7482_1st_23357.csv...
Using CUDA device: 2
Data shape: (22449, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22449, window_size 5000, adding 2551 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 147.07it/s]
Original shapes - Output: (25000,), Label: (22449,)
Aligned shapes - Output: (22449,), Label: (22449,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [00:01<00:00, 117.67it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 142.07it/s]
Table1-only metrics completed in 2.36s
Evaluation Result:  {'AUC-PR': 0.05073249894323215, 'AUC-ROC': 0.8641408996936636, 'VUS-PR': 0.05695503410346619, 'VUS-ROC': 0.8727760260613403, 'Standard-F1': 0.13494949494936, 'Standard-Precision': 0.0727668845315904, 'Standard-Recall': 0.9277777777777751, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6695530090991535, 'Affiliation-P': 0.5032541797443576, 'Affiliation-R': 1.0, 'F1_T': 0.13439542483646502, 'Precision_T': 0.07253086419753085, 'Recall_T': 0.9138888888888863}
Results for 448_UCR_id_146_Facility_tr_7482_1st_23357.csv: {'filename': '448_UCR_id_146_Facility_tr_7482_1st_23357.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 168, 'train_index': '7482', 'data_shape': '29931x1', 'output_length': 25000, 'label_length': 22449, 'aligned_length': 22449, 'AUC-PR': 0.05073249894323215, 'AUC-ROC': 0.8641408996936636, 'VUS-PR': 0.05695503410346619, 'VUS-ROC': 0.8727760260613403, 'Standard-F1': 0.13494949494936, 'Standard-Precision': 0.0727668845315904, 'Standard-Recall': 0.9277777777777751, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6695530090991535, 'Affiliation-P': 0.5032541797443576, 'Affiliation-R': 1.0, 'F1_T': 0.13439542483646502, 'Precision_T': 0.07253086419753085, 'Recall_T': 0.9138888888888863}
Logits metrics skipped for 448_UCR_id_146_Facility_tr_7482_1st_23357.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/448_UCR_id_146_Facility_tr_7482_1st_23357_results.pkl
Processing file: 373_UCR_id_71_Sensor_tr_2827_1st_5988.csv, Data shape: (11308, 1)
Label shape: (11308,)
Train data shape: (2827, 1), Test data shape: (8481, 1), Label test shape: (8481,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 373_UCR_id_71_Sensor_tr_2827_1st_5988.csv...
Using CUDA device: 2
Data shape: (8481, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 8481, window_size 5000, adding 1519 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.04it/s]
Original shapes - Output: (10000,), Label: (8481,)
Aligned shapes - Output: (8481,), Label: (8481,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 459.07it/s]
Table1-only metrics completed in 0.85s
Evaluation Result:  {'AUC-PR': 0.11231671463387616, 'AUC-ROC': 0.9431600200676793, 'VUS-PR': 0.1486362831075714, 'VUS-ROC': 0.9567908124789212, 'Standard-F1': 0.23293172690715308, 'Standard-Precision': 0.1907894736842093, 'Standard-Recall': 0.2989690721649454, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.761206823637562, 'Affiliation-P': 0.6150789264237708, 'Affiliation-R': 0.99840334428565, 'F1_T': 0.23214285714236405, 'Precision_T': 0.20472440944881728, 'Recall_T': 0.26804123711339933}
Results for 373_UCR_id_71_Sensor_tr_2827_1st_5988.csv: {'filename': '373_UCR_id_71_Sensor_tr_2827_1st_5988.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 98, 'train_index': '2827', 'data_shape': '11308x1', 'output_length': 10000, 'label_length': 8481, 'aligned_length': 8481, 'AUC-PR': 0.11231671463387616, 'AUC-ROC': 0.9431600200676793, 'VUS-PR': 0.1486362831075714, 'VUS-ROC': 0.9567908124789212, 'Standard-F1': 0.23293172690715308, 'Standard-Precision': 0.1907894736842093, 'Standard-Recall': 0.2989690721649454, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.761206823637562, 'Affiliation-P': 0.6150789264237708, 'Affiliation-R': 0.99840334428565, 'F1_T': 0.23214285714236405, 'Precision_T': 0.20472440944881728, 'Recall_T': 0.26804123711339933}
Logits metrics skipped for 373_UCR_id_71_Sensor_tr_2827_1st_5988.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/373_UCR_id_71_Sensor_tr_2827_1st_5988_results.pkl
Processing file: 783_YAHOO_id_233_WebService_tr_500_1st_24.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 783_YAHOO_id_233_WebService_tr_500_1st_24.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.51it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1124.89it/s]
Table1-only metrics completed in 0.62s
Evaluation Result:  {'AUC-PR': 0.36516274732505144, 'AUC-ROC': 0.8845241484011765, 'VUS-PR': 0.31622522725811114, 'VUS-ROC': 0.9065271213681025, 'Standard-F1': 0.3809523809518549, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.44444444444439507, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8526952260079964, 'Affiliation-P': 0.7432159660951088, 'Affiliation-R': 1.0, 'F1_T': 0.3809523809518549, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.44444444444439507}
Results for 783_YAHOO_id_233_WebService_tr_500_1st_24.csv: {'filename': '783_YAHOO_id_233_WebService_tr_500_1st_24.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.36516274732505144, 'AUC-ROC': 0.8845241484011765, 'VUS-PR': 0.31622522725811114, 'VUS-ROC': 0.9065271213681025, 'Standard-F1': 0.3809523809518549, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.44444444444439507, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8526952260079964, 'Affiliation-P': 0.7432159660951088, 'Affiliation-R': 1.0, 'F1_T': 0.3809523809518549, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.44444444444439507}
Logits metrics skipped for 783_YAHOO_id_233_WebService_tr_500_1st_24.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/783_YAHOO_id_233_WebService_tr_500_1st_24_results.pkl
Processing file: 321_UCR_id_19_Medical_tr_1875_1st_4485.csv, Data shape: (7500, 1)
Label shape: (7500,)
Train data shape: (1875, 1), Test data shape: (5625, 1), Label test shape: (5625,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 321_UCR_id_19_Medical_tr_1875_1st_4485.csv...
Using CUDA device: 2
Data shape: (5625, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5625, window_size 5000, adding 4375 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.08it/s]
Original shapes - Output: (10000,), Label: (5625,)
Aligned shapes - Output: (5625,), Label: (5625,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 633.64it/s]
Table1-only metrics completed in 1.01s
Evaluation Result:  {'AUC-PR': 0.9934373092527506, 'AUC-ROC': 0.9998366900625903, 'VUS-PR': 0.9838113555166419, 'VUS-ROC': 0.9997718986200278, 'Standard-F1': 0.9753694581275691, 'Standard-Precision': 0.9801980198019705, 'Standard-Recall': 0.9705882352941082, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999006437240818, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9998013071895424, 'F1_T': 0.9345794392518288, 'Precision_T': 0.8928571428571349, 'Recall_T': 0.9803921568627355}
Results for 321_UCR_id_19_Medical_tr_1875_1st_4485.csv: {'filename': '321_UCR_id_19_Medical_tr_1875_1st_4485.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 183, 'train_index': '1875', 'data_shape': '7500x1', 'output_length': 10000, 'label_length': 5625, 'aligned_length': 5625, 'AUC-PR': 0.9934373092527506, 'AUC-ROC': 0.9998366900625903, 'VUS-PR': 0.9838113555166419, 'VUS-ROC': 0.9997718986200278, 'Standard-F1': 0.9753694581275691, 'Standard-Precision': 0.9801980198019705, 'Standard-Recall': 0.9705882352941082, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999006437240818, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9998013071895424, 'F1_T': 0.9345794392518288, 'Precision_T': 0.8928571428571349, 'Recall_T': 0.9803921568627355}
Logits metrics skipped for 321_UCR_id_19_Medical_tr_1875_1st_4485.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/321_UCR_id_19_Medical_tr_1875_1st_4485_results.pkl
Processing file: 399_UCR_id_97_Facility_tr_19948_1st_52000.csv, Data shape: (79795, 1)
Label shape: (79795,)
Train data shape: (19948, 1), Test data shape: (59847, 1), Label test shape: (59847,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 399_UCR_id_97_Facility_tr_19948_1st_52000.csv...
Using CUDA device: 2
Data shape: (59847, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 59847, window_size 5000, adding 153 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([12, 5000, 1])
batched_mask shape: torch.Size([12, 5000])

100%|██████████| 1/1 [00:00<00:00, 91.16it/s]
Original shapes - Output: (60000,), Label: (59847,)
Aligned shapes - Output: (59847,), Label: (59847,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  34%|███▎      | 67/200 [00:01<00:02, 66.10it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [00:02<00:00, 72.82it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 80.54it/s]
Table1-only metrics completed in 4.74s
Evaluation Result:  {'AUC-PR': 0.036903447015507705, 'AUC-ROC': 0.7461737426860134, 'VUS-PR': 0.0383416159983063, 'VUS-ROC': 0.7671382431511413, 'Standard-F1': 0.10300866904613866, 'Standard-Precision': 0.061175045427013915, 'Standard-Recall': 0.32580645161290267, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7271441445569227, 'Affiliation-P': 0.5724700368872184, 'Affiliation-R': 0.9963434446009367, 'F1_T': 0.10081613058063955, 'Precision_T': 0.059221658206429766, 'Recall_T': 0.33870967741935426}
Results for 399_UCR_id_97_Facility_tr_19948_1st_52000.csv: {'filename': '399_UCR_id_97_Facility_tr_19948_1st_52000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 208, 'train_index': '19948', 'data_shape': '79795x1', 'output_length': 60000, 'label_length': 59847, 'aligned_length': 59847, 'AUC-PR': 0.036903447015507705, 'AUC-ROC': 0.7461737426860134, 'VUS-PR': 0.0383416159983063, 'VUS-ROC': 0.7671382431511413, 'Standard-F1': 0.10300866904613866, 'Standard-Precision': 0.061175045427013915, 'Standard-Recall': 0.32580645161290267, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7271441445569227, 'Affiliation-P': 0.5724700368872184, 'Affiliation-R': 0.9963434446009367, 'F1_T': 0.10081613058063955, 'Precision_T': 0.059221658206429766, 'Recall_T': 0.33870967741935426}
Logits metrics skipped for 399_UCR_id_97_Facility_tr_19948_1st_52000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/399_UCR_id_97_Facility_tr_19948_1st_52000_results.pkl
Processing file: 482_UCR_id_180_Medical_tr_7500_1st_16000.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 482_UCR_id_180_Medical_tr_7500_1st_16000.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 147.86it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 175/200 [00:01<00:00, 174.77it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 186.07it/s]
Table1-only metrics completed in 1.61s
Evaluation Result:  {'AUC-PR': 0.06366276867761643, 'AUC-ROC': 0.9674200892857142, 'VUS-PR': 0.06681245976997373, 'VUS-ROC': 0.967824069923406, 'Standard-F1': 0.15123859191633088, 'Standard-Precision': 0.08695652173913031, 'Standard-Recall': 0.5799999999999942, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6776855905658291, 'Affiliation-P': 0.512858474368408, 'Affiliation-R': 0.9986364444444444, 'F1_T': 0.15084525357584616, 'Precision_T': 0.08669656203288477, 'Recall_T': 0.5799999999999942}
Results for 482_UCR_id_180_Medical_tr_7500_1st_16000.csv: {'filename': '482_UCR_id_180_Medical_tr_7500_1st_16000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 91, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.06366276867761643, 'AUC-ROC': 0.9674200892857142, 'VUS-PR': 0.06681245976997373, 'VUS-ROC': 0.967824069923406, 'Standard-F1': 0.15123859191633088, 'Standard-Precision': 0.08695652173913031, 'Standard-Recall': 0.5799999999999942, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6776855905658291, 'Affiliation-P': 0.512858474368408, 'Affiliation-R': 0.9986364444444444, 'F1_T': 0.15084525357584616, 'Precision_T': 0.08669656203288477, 'Recall_T': 0.5799999999999942}
Logits metrics skipped for 482_UCR_id_180_Medical_tr_7500_1st_16000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/482_UCR_id_180_Medical_tr_7500_1st_16000_results.pkl
Skipping file: 815_Exathlon_id_6_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 704_YAHOO_id_154_WebService_tr_500_1st_541.csv, Data shape: (1422, 1)
Label shape: (1422,)
Train data shape: (500, 1), Test data shape: (922, 1), Label test shape: (922,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 704_YAHOO_id_154_WebService_tr_500_1st_541.csv...
Using CUDA device: 2
Data shape: (922, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 922, 1])
batched_mask shape: torch.Size([1, 922])

100%|██████████| 1/1 [00:00<00:00, 158.33it/s]
Original shapes - Output: (922,), Label: (922,)
Aligned shapes - Output: (922,), Label: (922,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1924.93it/s]
Table1-only metrics completed in 0.30s
Evaluation Result:  {'AUC-PR': 0.27882370325267725, 'AUC-ROC': 0.8332140300644237, 'VUS-PR': 0.5173654990807991, 'VUS-ROC': 0.9431197916137376, 'Standard-F1': 0.39999999999953817, 'Standard-Precision': 0.30645161290322087, 'Standard-Recall': 0.5757575757575583, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9610501301031475, 'Affiliation-P': 0.940677730229969, 'Affiliation-R': 0.9823244799333742, 'F1_T': 0.3917525773191306, 'Precision_T': 0.2968749999999954, 'Recall_T': 0.5757575757575583}
Results for 704_YAHOO_id_154_WebService_tr_500_1st_541.csv: {'filename': '704_YAHOO_id_154_WebService_tr_500_1st_541.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1422x1', 'output_length': 922, 'label_length': 922, 'aligned_length': 922, 'AUC-PR': 0.27882370325267725, 'AUC-ROC': 0.8332140300644237, 'VUS-PR': 0.5173654990807991, 'VUS-ROC': 0.9431197916137376, 'Standard-F1': 0.39999999999953817, 'Standard-Precision': 0.30645161290322087, 'Standard-Recall': 0.5757575757575583, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9610501301031475, 'Affiliation-P': 0.940677730229969, 'Affiliation-R': 0.9823244799333742, 'F1_T': 0.3917525773191306, 'Precision_T': 0.2968749999999954, 'Recall_T': 0.5757575757575583}
Logits metrics skipped for 704_YAHOO_id_154_WebService_tr_500_1st_541.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/704_YAHOO_id_154_WebService_tr_500_1st_541_results.pkl
Skipping file: 820_Exathlon_id_11_Facility_tr_11665_1st_13484.csv due to filter criteria.
Processing file: 787_YAHOO_id_237_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 787_YAHOO_id_237_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 161.79it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1798.54it/s]
Table1-only metrics completed in 1.02s
Evaluation Result:  {'AUC-PR': 0.9172278338945006, 'AUC-ROC': 0.9990253411306043, 'VUS-PR': 0.9953243119048887, 'VUS-ROC': 0.9999570068340861, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Results for 787_YAHOO_id_237_Synthetic_tr_500_1st_623.csv: {'filename': '787_YAHOO_id_237_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 155, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9172278338945006, 'AUC-ROC': 0.9990253411306043, 'VUS-PR': 0.9953243119048887, 'VUS-ROC': 0.9999570068340861, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 787_YAHOO_id_237_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/787_YAHOO_id_237_Synthetic_tr_500_1st_623_results.pkl
Skipping file: 171_MITDB_id_2_Medical_tr_50000_1st_88864.csv due to filter criteria.
Processing file: 314_UCR_id_12_HumanActivity_tr_16250_1st_38500.csv, Data shape: (65000, 1)
Label shape: (65000,)
Train data shape: (16250, 1), Test data shape: (48750, 1), Label test shape: (48750,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 314_UCR_id_12_HumanActivity_tr_16250_1st_38500.csv...
Using CUDA device: 2
Data shape: (48750, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 48750, window_size 5000, adding 1250 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([10, 5000, 1])
batched_mask shape: torch.Size([10, 5000])

100%|██████████| 1/1 [00:00<00:00, 95.03it/s]
Original shapes - Output: (50000,), Label: (48750,)
Aligned shapes - Output: (48750,), Label: (48750,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  40%|████      | 81/200 [00:01<00:01, 80.05it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 176/200 [00:02<00:00, 88.36it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 91.53it/s]
Table1-only metrics completed in 3.24s
Evaluation Result:  {'AUC-PR': 0.023943648145606622, 'AUC-ROC': 0.7752694874441005, 'VUS-PR': 0.025275710695301802, 'VUS-ROC': 0.7990000964440755, 'Standard-F1': 0.07558733401387509, 'Standard-Precision': 0.05449189985272451, 'Standard-Recall': 0.1233333333333329, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8367121856272846, 'Affiliation-P': 0.7208643014128728, 'Affiliation-R': 0.996924923076923, 'F1_T': 0.07392996108908069, 'Precision_T': 0.05219780219780212, 'Recall_T': 0.12666666666666623}
Results for 314_UCR_id_12_HumanActivity_tr_16250_1st_38500.csv: {'filename': '314_UCR_id_12_HumanActivity_tr_16250_1st_38500.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '16250', 'data_shape': '65000x1', 'output_length': 50000, 'label_length': 48750, 'aligned_length': 48750, 'AUC-PR': 0.023943648145606622, 'AUC-ROC': 0.7752694874441005, 'VUS-PR': 0.025275710695301802, 'VUS-ROC': 0.7990000964440755, 'Standard-F1': 0.07558733401387509, 'Standard-Precision': 0.05449189985272451, 'Standard-Recall': 0.1233333333333329, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8367121856272846, 'Affiliation-P': 0.7208643014128728, 'Affiliation-R': 0.996924923076923, 'F1_T': 0.07392996108908069, 'Precision_T': 0.05219780219780212, 'Recall_T': 0.12666666666666623}
Logits metrics skipped for 314_UCR_id_12_HumanActivity_tr_16250_1st_38500.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/314_UCR_id_12_HumanActivity_tr_16250_1st_38500_results.pkl
Processing file: 588_YAHOO_id_38_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 588_YAHOO_id_38_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 156.90it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1682.72it/s]
Table1-only metrics completed in 1.16s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 588_YAHOO_id_38_Synthetic_tr_500_1st_658.csv: {'filename': '588_YAHOO_id_38_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 235, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 588_YAHOO_id_38_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/588_YAHOO_id_38_Synthetic_tr_500_1st_658_results.pkl
Skipping file: 849_OPPORTUNITY_id_8_HumanActivity_tr_1100_1st_1200.csv due to filter criteria.
Processing file: 602_YAHOO_id_52_WebService_tr_500_1st_595.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 602_YAHOO_id_52_WebService_tr_500_1st_595.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.57it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1217.05it/s]
Table1-only metrics completed in 0.42s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.9999444925344662, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9979011298173531, 'Affiliation-P': 0.9958110516934047, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Results for 602_YAHOO_id_52_WebService_tr_500_1st_595.csv: {'filename': '602_YAHOO_id_52_WebService_tr_500_1st_595.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.9999444925344662, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9979011298173531, 'Affiliation-P': 0.9958110516934047, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 602_YAHOO_id_52_WebService_tr_500_1st_595.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/602_YAHOO_id_52_WebService_tr_500_1st_595_results.pkl
Processing file: 295_TODS_id_9_Synthetic_tr_1250_1st_2046.csv, Data shape: (5000, 1)
Label shape: (5000,)
Train data shape: (1250, 1), Test data shape: (3750, 1), Label test shape: (3750,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 295_TODS_id_9_Synthetic_tr_1250_1st_2046.csv...
Using CUDA device: 2
Data shape: (3750, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 3750, 1])
batched_mask shape: torch.Size([1, 3750])

100%|██████████| 1/1 [00:00<00:00, 161.75it/s]
Original shapes - Output: (3750,), Label: (3750,)
Aligned shapes - Output: (3750,), Label: (3750,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 594.86it/s]
Table1-only metrics completed in 2.34s
Evaluation Result:  {'AUC-PR': 0.9187173608600698, 'AUC-ROC': 0.9815445665445667, 'VUS-PR': 0.9855837320972989, 'VUS-ROC': 0.997606775155055, 'Standard-F1': 0.854296388542462, 'Standard-Precision': 0.8305084745762691, 'Standard-Recall': 0.8794871794871771, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9790682217057568, 'Affiliation-P': 0.9674681037232975, 'Affiliation-R': 0.9909498905701681, 'F1_T': 0.8531994981174403, 'Precision_T': 0.8353808353808333, 'Recall_T': 0.8717948717948695}
Results for 295_TODS_id_9_Synthetic_tr_1250_1st_2046.csv: {'filename': '295_TODS_id_9_Synthetic_tr_1250_1st_2046.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 99, 'train_index': '1250', 'data_shape': '5000x1', 'output_length': 3750, 'label_length': 3750, 'aligned_length': 3750, 'AUC-PR': 0.9187173608600698, 'AUC-ROC': 0.9815445665445667, 'VUS-PR': 0.9855837320972989, 'VUS-ROC': 0.997606775155055, 'Standard-F1': 0.854296388542462, 'Standard-Precision': 0.8305084745762691, 'Standard-Recall': 0.8794871794871771, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9790682217057568, 'Affiliation-P': 0.9674681037232975, 'Affiliation-R': 0.9909498905701681, 'F1_T': 0.8531994981174403, 'Precision_T': 0.8353808353808333, 'Recall_T': 0.8717948717948695}
Logits metrics skipped for 295_TODS_id_9_Synthetic_tr_1250_1st_2046.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/295_TODS_id_9_Synthetic_tr_1250_1st_2046_results.pkl
Processing file: 742_YAHOO_id_192_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 742_YAHOO_id_192_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 158.33it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1816.49it/s]
Table1-only metrics completed in 1.16s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 742_YAHOO_id_192_Synthetic_tr_500_1st_658.csv: {'filename': '742_YAHOO_id_192_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 235, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 742_YAHOO_id_192_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/742_YAHOO_id_192_Synthetic_tr_500_1st_658_results.pkl
Processing file: 260_IOPS_id_1_WebService_tr_5000_1st_11804.csv, Data shape: (20000, 1)
Label shape: (20000,)
Train data shape: (5000, 1), Test data shape: (15000, 1), Label test shape: (15000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 260_IOPS_id_1_WebService_tr_5000_1st_11804.csv...
Using CUDA device: 2
Data shape: (15000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.73it/s]
Original shapes - Output: (15000,), Label: (15000,)
Aligned shapes - Output: (15000,), Label: (15000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 301.88it/s]
Table1-only metrics completed in 0.77s
Evaluation Result:  {'AUC-PR': 0.28964994206379924, 'AUC-ROC': 0.8050084841089779, 'VUS-PR': 0.15606944136355339, 'VUS-ROC': 0.8165717338984768, 'Standard-F1': 0.0487804878047213, 'Standard-Precision': 0.02666666666666649, 'Standard-Recall': 0.2857142857142653, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7028236484834697, 'Affiliation-P': 0.5418139029782807, 'Affiliation-R': 0.9999880952380952, 'F1_T': 0.0487804878047213, 'Precision_T': 0.02666666666666649, 'Recall_T': 0.2857142857142653}
Results for 260_IOPS_id_1_WebService_tr_5000_1st_11804.csv: {'filename': '260_IOPS_id_1_WebService_tr_5000_1st_11804.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 8, 'train_index': '5000', 'data_shape': '20000x1', 'output_length': 15000, 'label_length': 15000, 'aligned_length': 15000, 'AUC-PR': 0.28964994206379924, 'AUC-ROC': 0.8050084841089779, 'VUS-PR': 0.15606944136355339, 'VUS-ROC': 0.8165717338984768, 'Standard-F1': 0.0487804878047213, 'Standard-Precision': 0.02666666666666649, 'Standard-Recall': 0.2857142857142653, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7028236484834697, 'Affiliation-P': 0.5418139029782807, 'Affiliation-R': 0.9999880952380952, 'F1_T': 0.0487804878047213, 'Precision_T': 0.02666666666666649, 'Recall_T': 0.2857142857142653}
Logits metrics skipped for 260_IOPS_id_1_WebService_tr_5000_1st_11804.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/260_IOPS_id_1_WebService_tr_5000_1st_11804_results.pkl
Processing file: 014_NAB_id_14_WebService_tr_500_1st_1045.csv, Data shape: (1537, 1)
Label shape: (1537,)
Train data shape: (500, 1), Test data shape: (1037, 1), Label test shape: (1037,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 014_NAB_id_14_WebService_tr_500_1st_1045.csv...
Using CUDA device: 2
Data shape: (1037, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1037, 1])
batched_mask shape: torch.Size([1, 1037])

100%|██████████| 1/1 [00:00<00:00, 161.41it/s]
Original shapes - Output: (1037,), Label: (1037,)
Aligned shapes - Output: (1037,), Label: (1037,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2216.77it/s]
Table1-only metrics completed in 0.18s
Evaluation Result:  {'AUC-PR': 0.08927907404565447, 'AUC-ROC': 0.09976211715730002, 'VUS-PR': 0.09260697363885749, 'VUS-ROC': 0.1065854351938136, 'Standard-F1': 0.25214408233253444, 'Standard-Precision': 0.14497041420118328, 'Standard-Recall': 0.9671052631578884, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6757328439742505, 'Affiliation-P': 0.5102701027242974, 'Affiliation-R': 0.9999968278942293, 'F1_T': 0.2517123287668965, 'Precision_T': 0.1446850393700786, 'Recall_T': 0.9671052631578884}
Results for 014_NAB_id_14_WebService_tr_500_1st_1045.csv: {'filename': '014_NAB_id_14_WebService_tr_500_1st_1045.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 23, 'train_index': '500', 'data_shape': '1537x1', 'output_length': 1037, 'label_length': 1037, 'aligned_length': 1037, 'AUC-PR': 0.08927907404565447, 'AUC-ROC': 0.09976211715730002, 'VUS-PR': 0.09260697363885749, 'VUS-ROC': 0.1065854351938136, 'Standard-F1': 0.25214408233253444, 'Standard-Precision': 0.14497041420118328, 'Standard-Recall': 0.9671052631578884, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6757328439742505, 'Affiliation-P': 0.5102701027242974, 'Affiliation-R': 0.9999968278942293, 'F1_T': 0.2517123287668965, 'Precision_T': 0.1446850393700786, 'Recall_T': 0.9671052631578884}
Logits metrics skipped for 014_NAB_id_14_WebService_tr_500_1st_1045.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/014_NAB_id_14_WebService_tr_500_1st_1045_results.pkl
Processing file: 649_YAHOO_id_99_WebService_tr_500_1st_1386.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 649_YAHOO_id_99_WebService_tr_500_1st_1386.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 163.74it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1283.91it/s]
Table1-only metrics completed in 0.71s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9947465568649717, 'Affiliation-P': 0.98954802259887, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Results for 649_YAHOO_id_99_WebService_tr_500_1st_1386.csv: {'filename': '649_YAHOO_id_99_WebService_tr_500_1st_1386.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 168, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9947465568649717, 'Affiliation-P': 0.98954802259887, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 649_YAHOO_id_99_WebService_tr_500_1st_1386.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/649_YAHOO_id_99_WebService_tr_500_1st_1386_results.pkl
Processing file: 553_YAHOO_id_3_WebService_tr_500_1st_237.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 553_YAHOO_id_3_WebService_tr_500_1st_237.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 155.93it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1609.82it/s]
Table1-only metrics completed in 0.24s
Evaluation Result:  {'AUC-PR': 0.7916666666666666, 'AUC-ROC': 0.9985839705465873, 'VUS-PR': 0.41111111111111115, 'VUS-ROC': 0.9972151420749551, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9284530106350579, 'Affiliation-P': 0.8664603791060171, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Results for 553_YAHOO_id_3_WebService_tr_500_1st_237.csv: {'filename': '553_YAHOO_id_3_WebService_tr_500_1st_237.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 12, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.7916666666666666, 'AUC-ROC': 0.9985839705465873, 'VUS-PR': 0.41111111111111115, 'VUS-ROC': 0.9972151420749551, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9284530106350579, 'Affiliation-P': 0.8664603791060171, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Logits metrics skipped for 553_YAHOO_id_3_WebService_tr_500_1st_237.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/553_YAHOO_id_3_WebService_tr_500_1st_237_results.pkl
Skipping file: 177_MITDB_id_8_Medical_tr_17675_1st_17775.csv due to filter criteria.
Processing file: 289_TODS_id_3_Synthetic_tr_500_1st_26.csv, Data shape: (5000, 1)
Label shape: (5000,)
Train data shape: (500, 1), Test data shape: (4500, 1), Label test shape: (4500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 289_TODS_id_3_Synthetic_tr_500_1st_26.csv...
Using CUDA device: 2
Data shape: (4500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 4500, 1])
batched_mask shape: torch.Size([1, 4500])

100%|██████████| 1/1 [00:00<00:00, 155.84it/s]
Original shapes - Output: (4500,), Label: (4500,)
Aligned shapes - Output: (4500,), Label: (4500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  20%|█▉        | 39/200 [00:01<00:04, 37.83it/s]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [00:03<00:05, 23.11it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 104/200 [00:04<00:04, 19.45it/s]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [00:06<00:03, 18.56it/s]

Calculating Affiliation Metrics:  75%|███████▌  | 150/200 [00:07<00:02, 19.98it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:08<00:00, 24.72it/s]
Table1-only metrics completed in 14.67s
Evaluation Result:  {'AUC-PR': 0.6183852749745418, 'AUC-ROC': 0.8791662531681421, 'VUS-PR': 0.9367510109672866, 'VUS-ROC': 0.9806497139832416, 'Standard-F1': 0.5839793281648747, 'Standard-Precision': 0.6348314606741537, 'Standard-Recall': 0.5406698564593275, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7641797138474852, 'Affiliation-P': 0.8042679831146536, 'Affiliation-R': 0.7278980589763232, 'F1_T': 0.5839793281648747, 'Precision_T': 0.6348314606741537, 'Recall_T': 0.5406698564593275}
Results for 289_TODS_id_3_Synthetic_tr_500_1st_26.csv: {'filename': '289_TODS_id_3_Synthetic_tr_500_1st_26.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '500', 'data_shape': '5000x1', 'output_length': 4500, 'label_length': 4500, 'aligned_length': 4500, 'AUC-PR': 0.6183852749745418, 'AUC-ROC': 0.8791662531681421, 'VUS-PR': 0.9367510109672866, 'VUS-ROC': 0.9806497139832416, 'Standard-F1': 0.5839793281648747, 'Standard-Precision': 0.6348314606741537, 'Standard-Recall': 0.5406698564593275, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7641797138474852, 'Affiliation-P': 0.8042679831146536, 'Affiliation-R': 0.7278980589763232, 'F1_T': 0.5839793281648747, 'Precision_T': 0.6348314606741537, 'Recall_T': 0.5406698564593275}
Logits metrics skipped for 289_TODS_id_3_Synthetic_tr_500_1st_26.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/289_TODS_id_3_Synthetic_tr_500_1st_26_results.pkl
Processing file: 425_UCR_id_123_Medical_tr_16700_1st_16800.csv, Data shape: (200000, 1)
Label shape: (200000,)
Train data shape: (16700, 1), Test data shape: (183300, 1), Label test shape: (183300,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 425_UCR_id_123_Medical_tr_16700_1st_16800.csv...
Using CUDA device: 2
Data shape: (183300, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 183300, window_size 5000, adding 1700 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([37, 5000, 1])
batched_mask shape: torch.Size([37, 5000])

100%|██████████| 1/1 [00:00<00:00, 32.46it/s]
Original shapes - Output: (185000,), Label: (183300,)
Aligned shapes - Output: (183300,), Label: (183300,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  11%|█         | 22/200 [00:01<00:08, 21.88it/s]

Calculating Affiliation Metrics:  22%|██▏       | 44/200 [00:02<00:07, 21.02it/s]

Calculating Affiliation Metrics:  34%|███▎      | 67/200 [00:03<00:06, 21.60it/s]

Calculating Affiliation Metrics:  46%|████▌     | 91/200 [00:04<00:04, 22.35it/s]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [00:05<00:03, 23.86it/s]

Calculating Affiliation Metrics:  74%|███████▍  | 149/200 [00:06<00:01, 26.09it/s]

Calculating Affiliation Metrics:  92%|█████████▎| 185/200 [00:07<00:00, 29.17it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:07<00:00, 26.72it/s]
Table1-only metrics completed in 14.06s
Evaluation Result:  {'AUC-PR': 0.004970186681389349, 'AUC-ROC': 0.852990072859745, 'VUS-PR': 0.005099620331706235, 'VUS-ROC': 0.853505096434525, 'Standard-F1': 0.014895625721163451, 'Standard-Precision': 0.007506872488898287, 'Standard-Recall': 0.9466666666666634, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7201161092820837, 'Affiliation-P': 0.56264174782148, 'Affiliation-R': 1.0, 'F1_T': 0.014905467953228333, 'Precision_T': 0.007511662844943465, 'Recall_T': 0.9499999999999967}
Results for 425_UCR_id_123_Medical_tr_16700_1st_16800.csv: {'filename': '425_UCR_id_123_Medical_tr_16700_1st_16800.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '16700', 'data_shape': '200000x1', 'output_length': 185000, 'label_length': 183300, 'aligned_length': 183300, 'AUC-PR': 0.004970186681389349, 'AUC-ROC': 0.852990072859745, 'VUS-PR': 0.005099620331706235, 'VUS-ROC': 0.853505096434525, 'Standard-F1': 0.014895625721163451, 'Standard-Precision': 0.007506872488898287, 'Standard-Recall': 0.9466666666666634, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7201161092820837, 'Affiliation-P': 0.56264174782148, 'Affiliation-R': 1.0, 'F1_T': 0.014905467953228333, 'Precision_T': 0.007511662844943465, 'Recall_T': 0.9499999999999967}
Logits metrics skipped for 425_UCR_id_123_Medical_tr_16700_1st_16800.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/425_UCR_id_123_Medical_tr_16700_1st_16800_results.pkl
Processing file: 021_NAB_id_21_WebService_tr_500_1st_565.csv, Data shape: (1623, 1)
Label shape: (1623,)
Train data shape: (500, 1), Test data shape: (1123, 1), Label test shape: (1123,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 021_NAB_id_21_WebService_tr_500_1st_565.csv...
Using CUDA device: 2
Data shape: (1123, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1123, 1])
batched_mask shape: torch.Size([1, 1123])

100%|██████████| 1/1 [00:00<00:00, 147.52it/s]
Original shapes - Output: (1123,), Label: (1123,)
Aligned shapes - Output: (1123,), Label: (1123,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1571.52it/s]
Table1-only metrics completed in 0.28s
Evaluation Result:  {'AUC-PR': 0.3309606106733483, 'AUC-ROC': 0.7426596573208722, 'VUS-PR': 0.35111391413083626, 'VUS-ROC': 0.7758791430530149, 'Standard-F1': 0.4563758389257128, 'Standard-Precision': 0.35540069686411024, 'Standard-Recall': 0.6374999999999961, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8921812473414417, 'Affiliation-P': 0.8616211990762469, 'Affiliation-R': 0.9249888234788275, 'F1_T': 0.4553571428566816, 'Precision_T': 0.3541666666666654, 'Recall_T': 0.6374999999999961}
Results for 021_NAB_id_21_WebService_tr_500_1st_565.csv: {'filename': '021_NAB_id_21_WebService_tr_500_1st_565.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1623x1', 'output_length': 1123, 'label_length': 1123, 'aligned_length': 1123, 'AUC-PR': 0.3309606106733483, 'AUC-ROC': 0.7426596573208722, 'VUS-PR': 0.35111391413083626, 'VUS-ROC': 0.7758791430530149, 'Standard-F1': 0.4563758389257128, 'Standard-Precision': 0.35540069686411024, 'Standard-Recall': 0.6374999999999961, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8921812473414417, 'Affiliation-P': 0.8616211990762469, 'Affiliation-R': 0.9249888234788275, 'F1_T': 0.4553571428566816, 'Precision_T': 0.3541666666666654, 'Recall_T': 0.6374999999999961}
Logits metrics skipped for 021_NAB_id_21_WebService_tr_500_1st_565.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/021_NAB_id_21_WebService_tr_500_1st_565_results.pkl
Processing file: 646_YAHOO_id_96_WebService_tr_500_1st_121.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 646_YAHOO_id_96_WebService_tr_500_1st_121.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.39it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1215.11it/s]
Table1-only metrics completed in 0.42s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.999944492534466, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.998790886938668, 'Affiliation-P': 0.9975846942550606, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Results for 646_YAHOO_id_96_WebService_tr_500_1st_121.csv: {'filename': '646_YAHOO_id_96_WebService_tr_500_1st_121.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.999944492534466, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.998790886938668, 'Affiliation-P': 0.9975846942550606, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 646_YAHOO_id_96_WebService_tr_500_1st_121.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/646_YAHOO_id_96_WebService_tr_500_1st_121_results.pkl
Processing file: 636_YAHOO_id_86_WebService_tr_500_1st_20.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 636_YAHOO_id_86_WebService_tr_500_1st_20.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 160.63it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1351.73it/s]
Table1-only metrics completed in 0.49s
Evaluation Result:  {'AUC-PR': 0.737353206865402, 'AUC-ROC': 0.9931856899488927, 'VUS-PR': 0.576774816126784, 'VUS-ROC': 0.993787117760638, 'Standard-F1': 0.4444444444439506, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.6666666666665555, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9029147639085747, 'Affiliation-P': 0.8230124097972379, 'Affiliation-R': 1.0, 'F1_T': 0.4444444444439506, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.6666666666665555}
Results for 636_YAHOO_id_86_WebService_tr_500_1st_20.csv: {'filename': '636_YAHOO_id_86_WebService_tr_500_1st_20.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.737353206865402, 'AUC-ROC': 0.9931856899488927, 'VUS-PR': 0.576774816126784, 'VUS-ROC': 0.993787117760638, 'Standard-F1': 0.4444444444439506, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.6666666666665555, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9029147639085747, 'Affiliation-P': 0.8230124097972379, 'Affiliation-R': 1.0, 'F1_T': 0.4444444444439506, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.6666666666665555}
Logits metrics skipped for 636_YAHOO_id_86_WebService_tr_500_1st_20.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/636_YAHOO_id_86_WebService_tr_500_1st_20_results.pkl
Processing file: 756_YAHOO_id_206_WebService_tr_500_1st_55.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 756_YAHOO_id_206_WebService_tr_500_1st_55.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 165.30it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 793.14it/s]
Table1-only metrics completed in 1.89s
Evaluation Result:  {'AUC-PR': 0.9658119658119659, 'AUC-ROC': 0.9996204573488946, 'VUS-PR': 0.9844501629778668, 'VUS-ROC': 0.9999106134584574, 'Standard-F1': 0.7619047619041995, 'Standard-Precision': 0.6666666666666111, 'Standard-Recall': 0.8888888888887901, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.995690318608588, 'Affiliation-P': 0.9914176245210728, 'Affiliation-R': 1.0, 'F1_T': 0.7619047619041995, 'Precision_T': 0.6666666666666111, 'Recall_T': 0.8888888888887901}
Results for 756_YAHOO_id_206_WebService_tr_500_1st_55.csv: {'filename': '756_YAHOO_id_206_WebService_tr_500_1st_55.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9658119658119659, 'AUC-ROC': 0.9996204573488946, 'VUS-PR': 0.9844501629778668, 'VUS-ROC': 0.9999106134584574, 'Standard-F1': 0.7619047619041995, 'Standard-Precision': 0.6666666666666111, 'Standard-Recall': 0.8888888888887901, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.995690318608588, 'Affiliation-P': 0.9914176245210728, 'Affiliation-R': 1.0, 'F1_T': 0.7619047619041995, 'Precision_T': 0.6666666666666111, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 756_YAHOO_id_206_WebService_tr_500_1st_55.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/756_YAHOO_id_206_WebService_tr_500_1st_55_results.pkl
Processing file: 045_WSD_id_17_WebService_tr_2566_1st_2666.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (2566, 1), Test data shape: (15670, 1), Label test shape: (15670,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 045_WSD_id_17_WebService_tr_2566_1st_2666.csv...
Using CUDA device: 2
Data shape: (15670, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 15670, window_size 5000, adding 4330 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.63it/s]
Original shapes - Output: (20000,), Label: (15670,)
Aligned shapes - Output: (15670,), Label: (15670,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 215.19it/s]
Table1-only metrics completed in 1.05s
Evaluation Result:  {'AUC-PR': 0.3537352345507899, 'AUC-ROC': 0.8402174838270344, 'VUS-PR': 0.342885401386947, 'VUS-ROC': 0.9172085079572664, 'Standard-F1': 0.03680981595084889, 'Standard-Precision': 0.019108280254776948, 'Standard-Recall': 0.4999999999999167, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8524285285518828, 'Affiliation-P': 0.7428137942237062, 'Affiliation-R': 0.9999946457637283, 'F1_T': 0.03680981595084889, 'Precision_T': 0.019108280254776948, 'Recall_T': 0.4999999999999167}
Results for 045_WSD_id_17_WebService_tr_2566_1st_2666.csv: {'filename': '045_WSD_id_17_WebService_tr_2566_1st_2666.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 9, 'train_index': '2566', 'data_shape': '18236x1', 'output_length': 20000, 'label_length': 15670, 'aligned_length': 15670, 'AUC-PR': 0.3537352345507899, 'AUC-ROC': 0.8402174838270344, 'VUS-PR': 0.342885401386947, 'VUS-ROC': 0.9172085079572664, 'Standard-F1': 0.03680981595084889, 'Standard-Precision': 0.019108280254776948, 'Standard-Recall': 0.4999999999999167, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8524285285518828, 'Affiliation-P': 0.7428137942237062, 'Affiliation-R': 0.9999946457637283, 'F1_T': 0.03680981595084889, 'Precision_T': 0.019108280254776948, 'Recall_T': 0.4999999999999167}
Logits metrics skipped for 045_WSD_id_17_WebService_tr_2566_1st_2666.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/045_WSD_id_17_WebService_tr_2566_1st_2666_results.pkl
Processing file: 652_YAHOO_id_102_WebService_tr_500_1st_174.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 652_YAHOO_id_102_WebService_tr_500_1st_174.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.89it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 826.86it/s]
Table1-only metrics completed in 0.75s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9866975943855992, 'VUS-ROC': 0.9999145994652906, 'Standard-F1': 0.9565217391298525, 'Standard-Precision': 0.9166666666665902, 'Standard-Recall': 0.9999999999999091, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9995993589743584, 'Affiliation-P': 0.999199038846616, 'Affiliation-R': 1.0, 'F1_T': 0.9565217391298525, 'Precision_T': 0.9166666666665902, 'Recall_T': 0.9999999999999091}
Results for 652_YAHOO_id_102_WebService_tr_500_1st_174.csv: {'filename': '652_YAHOO_id_102_WebService_tr_500_1st_174.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9866975943855992, 'VUS-ROC': 0.9999145994652906, 'Standard-F1': 0.9565217391298525, 'Standard-Precision': 0.9166666666665902, 'Standard-Recall': 0.9999999999999091, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9995993589743584, 'Affiliation-P': 0.999199038846616, 'Affiliation-R': 1.0, 'F1_T': 0.9565217391298525, 'Precision_T': 0.9166666666665902, 'Recall_T': 0.9999999999999091}
Logits metrics skipped for 652_YAHOO_id_102_WebService_tr_500_1st_174.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/652_YAHOO_id_102_WebService_tr_500_1st_174_results.pkl
Processing file: 127_WSD_id_99_WebService_tr_1164_1st_1264.csv, Data shape: (18190, 1)
Label shape: (18190,)
Train data shape: (1164, 1), Test data shape: (17026, 1), Label test shape: (17026,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 127_WSD_id_99_WebService_tr_1164_1st_1264.csv...
Using CUDA device: 2
Data shape: (17026, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 17026, window_size 5000, adding 2974 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.90it/s]
Original shapes - Output: (20000,), Label: (17026,)
Aligned shapes - Output: (17026,), Label: (17026,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 136/200 [00:01<00:00, 135.19it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 157.56it/s]
Table1-only metrics completed in 4.49s
Evaluation Result:  {'AUC-PR': 0.27110785955079375, 'AUC-ROC': 0.8754731840071008, 'VUS-PR': 0.2302755303860028, 'VUS-ROC': 0.9705976266835312, 'Standard-F1': 0.2222222222217383, 'Standard-Precision': 0.18713450292397552, 'Standard-Recall': 0.2735042735042712, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9471261152852233, 'Affiliation-P': 0.9035365460746182, 'Affiliation-R': 0.9951346791433721, 'F1_T': 0.2222222222217383, 'Precision_T': 0.18713450292397552, 'Recall_T': 0.2735042735042712}
Results for 127_WSD_id_99_WebService_tr_1164_1st_1264.csv: {'filename': '127_WSD_id_99_WebService_tr_1164_1st_1264.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '1164', 'data_shape': '18190x1', 'output_length': 20000, 'label_length': 17026, 'aligned_length': 17026, 'AUC-PR': 0.27110785955079375, 'AUC-ROC': 0.8754731840071008, 'VUS-PR': 0.2302755303860028, 'VUS-ROC': 0.9705976266835312, 'Standard-F1': 0.2222222222217383, 'Standard-Precision': 0.18713450292397552, 'Standard-Recall': 0.2735042735042712, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9471261152852233, 'Affiliation-P': 0.9035365460746182, 'Affiliation-R': 0.9951346791433721, 'F1_T': 0.2222222222217383, 'Precision_T': 0.18713450292397552, 'Recall_T': 0.2735042735042712}
Logits metrics skipped for 127_WSD_id_99_WebService_tr_1164_1st_1264.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/127_WSD_id_99_WebService_tr_1164_1st_1264_results.pkl
Processing file: 317_UCR_id_15_HumanActivity_tr_16000_1st_38400.csv, Data shape: (64000, 1)
Label shape: (64000,)
Train data shape: (16000, 1), Test data shape: (48000, 1), Label test shape: (48000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 317_UCR_id_15_HumanActivity_tr_16000_1st_38400.csv...
Using CUDA device: 2
Data shape: (48000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 48000, window_size 5000, adding 2000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([10, 5000, 1])
batched_mask shape: torch.Size([10, 5000])

100%|██████████| 1/1 [00:00<00:00, 91.14it/s]
Original shapes - Output: (50000,), Label: (48000,)
Aligned shapes - Output: (48000,), Label: (48000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [00:01<00:01, 82.35it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 175/200 [00:02<00:00, 87.88it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 91.58it/s]
Table1-only metrics completed in 3.33s
Evaluation Result:  {'AUC-PR': 0.04716089906311863, 'AUC-ROC': 0.813205376059322, 'VUS-PR': 0.05135764062762604, 'VUS-ROC': 0.8247236177877392, 'Standard-F1': 0.12596324836967834, 'Standard-Precision': 0.0714525891055817, 'Standard-Recall': 0.5312499999999993, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8069519185217845, 'Affiliation-P': 0.6786644283746556, 'Affiliation-R': 0.99504453125, 'F1_T': 0.12535358046726072, 'Precision_T': 0.071150921074869, 'Recall_T': 0.5262499999999993}
Results for 317_UCR_id_15_HumanActivity_tr_16000_1st_38400.csv: {'filename': '317_UCR_id_15_HumanActivity_tr_16000_1st_38400.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '16000', 'data_shape': '64000x1', 'output_length': 50000, 'label_length': 48000, 'aligned_length': 48000, 'AUC-PR': 0.04716089906311863, 'AUC-ROC': 0.813205376059322, 'VUS-PR': 0.05135764062762604, 'VUS-ROC': 0.8247236177877392, 'Standard-F1': 0.12596324836967834, 'Standard-Precision': 0.0714525891055817, 'Standard-Recall': 0.5312499999999993, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8069519185217845, 'Affiliation-P': 0.6786644283746556, 'Affiliation-R': 0.99504453125, 'F1_T': 0.12535358046726072, 'Precision_T': 0.071150921074869, 'Recall_T': 0.5262499999999993}
Logits metrics skipped for 317_UCR_id_15_HumanActivity_tr_16000_1st_38400.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/317_UCR_id_15_HumanActivity_tr_16000_1st_38400_results.pkl
Processing file: 497_UCR_id_195_Facility_tr_32372_1st_114000.csv, Data shape: (129489, 1)
Label shape: (129489,)
Train data shape: (32372, 1), Test data shape: (97117, 1), Label test shape: (97117,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 497_UCR_id_195_Facility_tr_32372_1st_114000.csv...
Using CUDA device: 2
Data shape: (97117, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 97117, window_size 5000, adding 2883 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([20, 5000, 1])
batched_mask shape: torch.Size([20, 5000])

100%|██████████| 1/1 [00:00<00:00, 64.21it/s]
Original shapes - Output: (100000,), Label: (97117,)
Aligned shapes - Output: (97117,), Label: (97117,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  20%|█▉        | 39/200 [00:01<00:04, 38.75it/s]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [00:02<00:03, 36.55it/s]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [00:03<00:02, 36.55it/s]

Calculating Affiliation Metrics:  80%|████████  | 160/200 [00:04<00:01, 39.52it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:04<00:00, 42.05it/s]
Table1-only metrics completed in 10.92s
Evaluation Result:  {'AUC-PR': 0.6198533521199692, 'AUC-ROC': 0.974391579709574, 'VUS-PR': 0.4486106939742405, 'VUS-ROC': 0.976154435239943, 'Standard-F1': 0.39791356184758814, 'Standard-Precision': 0.2746913580246911, 'Standard-Recall': 0.7216216216216196, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8707383962257051, 'Affiliation-P': 0.7712254522842422, 'Affiliation-R': 0.9997366508883546, 'F1_T': 0.39791356184758814, 'Precision_T': 0.2746913580246911, 'Recall_T': 0.7216216216216196}
Results for 497_UCR_id_195_Facility_tr_32372_1st_114000.csv: {'filename': '497_UCR_id_195_Facility_tr_32372_1st_114000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 243, 'train_index': '32372', 'data_shape': '129489x1', 'output_length': 100000, 'label_length': 97117, 'aligned_length': 97117, 'AUC-PR': 0.6198533521199692, 'AUC-ROC': 0.974391579709574, 'VUS-PR': 0.4486106939742405, 'VUS-ROC': 0.976154435239943, 'Standard-F1': 0.39791356184758814, 'Standard-Precision': 0.2746913580246911, 'Standard-Recall': 0.7216216216216196, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8707383962257051, 'Affiliation-P': 0.7712254522842422, 'Affiliation-R': 0.9997366508883546, 'F1_T': 0.39791356184758814, 'Precision_T': 0.2746913580246911, 'Recall_T': 0.7216216216216196}
Logits metrics skipped for 497_UCR_id_195_Facility_tr_32372_1st_114000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/497_UCR_id_195_Facility_tr_32372_1st_114000_results.pkl
Processing file: 749_YAHOO_id_199_WebService_tr_500_1st_136.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 749_YAHOO_id_199_WebService_tr_500_1st_136.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 153.44it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1112.32it/s]
Table1-only metrics completed in 0.53s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9829823771486967, 'VUS-ROC': 0.9999274768937315, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Results for 749_YAHOO_id_199_WebService_tr_500_1st_136.csv: {'filename': '749_YAHOO_id_199_WebService_tr_500_1st_136.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9829823771486967, 'VUS-ROC': 0.9999274768937315, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Logits metrics skipped for 749_YAHOO_id_199_WebService_tr_500_1st_136.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/749_YAHOO_id_199_WebService_tr_500_1st_136_results.pkl
Processing file: 443_UCR_id_141_Medical_tr_7456_1st_17390.csv, Data shape: (29826, 1)
Label shape: (29826,)
Train data shape: (7456, 1), Test data shape: (22370, 1), Label test shape: (22370,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 443_UCR_id_141_Medical_tr_7456_1st_17390.csv...
Using CUDA device: 2
Data shape: (22370, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22370, window_size 5000, adding 2630 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 152.09it/s]
Original shapes - Output: (25000,), Label: (22370,)
Aligned shapes - Output: (22370,), Label: (22370,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  54%|█████▎    | 107/200 [00:01<00:00, 106.66it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 117.43it/s]
Table1-only metrics completed in 1.96s
Evaluation Result:  {'AUC-PR': 0.040818026837250984, 'AUC-ROC': 0.43166626314333145, 'VUS-PR': 0.020748917258955626, 'VUS-ROC': 0.45707850570585024, 'Standard-F1': 0.06030150753724825, 'Standard-Precision': 0.044776119402984906, 'Standard-Recall': 0.0923076923076916, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666795665851731, 'Affiliation-P': 0.5000145125487283, 'Affiliation-R': 1.0, 'F1_T': 0.05649717514077787, 'Precision_T': 0.04464285714285694, 'Recall_T': 0.07692307692307633}
Results for 443_UCR_id_141_Medical_tr_7456_1st_17390.csv: {'filename': '443_UCR_id_141_Medical_tr_7456_1st_17390.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 33, 'train_index': '7456', 'data_shape': '29826x1', 'output_length': 25000, 'label_length': 22370, 'aligned_length': 22370, 'AUC-PR': 0.040818026837250984, 'AUC-ROC': 0.43166626314333145, 'VUS-PR': 0.020748917258955626, 'VUS-ROC': 0.45707850570585024, 'Standard-F1': 0.06030150753724825, 'Standard-Precision': 0.044776119402984906, 'Standard-Recall': 0.0923076923076916, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666795665851731, 'Affiliation-P': 0.5000145125487283, 'Affiliation-R': 1.0, 'F1_T': 0.05649717514077787, 'Precision_T': 0.04464285714285694, 'Recall_T': 0.07692307692307633}
Logits metrics skipped for 443_UCR_id_141_Medical_tr_7456_1st_17390.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/443_UCR_id_141_Medical_tr_7456_1st_17390_results.pkl
Processing file: 066_WSD_id_38_WebService_tr_3320_1st_3420.csv, Data shape: (18302, 1)
Label shape: (18302,)
Train data shape: (3320, 1), Test data shape: (14982, 1), Label test shape: (14982,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 066_WSD_id_38_WebService_tr_3320_1st_3420.csv...
Using CUDA device: 2
Data shape: (14982, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 14982, window_size 5000, adding 18 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.19it/s]
Original shapes - Output: (15000,), Label: (14982,)
Aligned shapes - Output: (14982,), Label: (14982,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 211.62it/s]
Table1-only metrics completed in 3.04s
Evaluation Result:  {'AUC-PR': 0.6585640974963595, 'AUC-ROC': 0.9738138269766549, 'VUS-PR': 0.7702238186706527, 'VUS-ROC': 0.9960430906198202, 'Standard-F1': 0.5725190839689718, 'Standard-Precision': 0.49999999999999667, 'Standard-Recall': 0.6696428571428512, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9308743793919925, 'Affiliation-P': 0.9946671499953289, 'Affiliation-R': 0.874771115971788, 'F1_T': 0.5725190839689718, 'Precision_T': 0.49999999999999667, 'Recall_T': 0.6696428571428512}
Results for 066_WSD_id_38_WebService_tr_3320_1st_3420.csv: {'filename': '066_WSD_id_38_WebService_tr_3320_1st_3420.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '3320', 'data_shape': '18302x1', 'output_length': 15000, 'label_length': 14982, 'aligned_length': 14982, 'AUC-PR': 0.6585640974963595, 'AUC-ROC': 0.9738138269766549, 'VUS-PR': 0.7702238186706527, 'VUS-ROC': 0.9960430906198202, 'Standard-F1': 0.5725190839689718, 'Standard-Precision': 0.49999999999999667, 'Standard-Recall': 0.6696428571428512, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9308743793919925, 'Affiliation-P': 0.9946671499953289, 'Affiliation-R': 0.874771115971788, 'F1_T': 0.5725190839689718, 'Precision_T': 0.49999999999999667, 'Recall_T': 0.6696428571428512}
Logits metrics skipped for 066_WSD_id_38_WebService_tr_3320_1st_3420.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/066_WSD_id_38_WebService_tr_3320_1st_3420_results.pkl
Skipping file: 869_OPPORTUNITY_id_28_HumanActivity_tr_500_1st_566.csv due to filter criteria.
Skipping file: 169_Daphnet_id_1_HumanActivity_tr_9693_1st_20732.csv due to filter criteria.
Processing file: 717_YAHOO_id_167_WebService_tr_500_1st_84.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 717_YAHOO_id_167_WebService_tr_500_1st_84.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 160.54it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1536.25it/s]
Table1-only metrics completed in 0.38s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.999944492534466, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990982867448146, 'Affiliation-P': 0.9981981981981982, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Results for 717_YAHOO_id_167_WebService_tr_500_1st_84.csv: {'filename': '717_YAHOO_id_167_WebService_tr_500_1st_84.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.999944492534466, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990982867448146, 'Affiliation-P': 0.9981981981981982, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 717_YAHOO_id_167_WebService_tr_500_1st_84.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/717_YAHOO_id_167_WebService_tr_500_1st_84_results.pkl
Processing file: 566_YAHOO_id_16_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 566_YAHOO_id_16_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 160.90it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1725.07it/s]
Table1-only metrics completed in 1.02s
Evaluation Result:  {'AUC-PR': 0.9108175774842442, 'AUC-ROC': 0.9989035087719298, 'VUS-PR': 0.9934562892651673, 'VUS-ROC': 0.9999422005105032, 'Standard-F1': 0.8181818181812603, 'Standard-Precision': 0.692307692307639, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9960624297351877, 'Affiliation-P': 0.9928360285167797, 'Affiliation-R': 0.9993098688750863, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Results for 566_YAHOO_id_16_Synthetic_tr_500_1st_623.csv: {'filename': '566_YAHOO_id_16_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 155, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9108175774842442, 'AUC-ROC': 0.9989035087719298, 'VUS-PR': 0.9934562892651673, 'VUS-ROC': 0.9999422005105032, 'Standard-F1': 0.8181818181812603, 'Standard-Precision': 0.692307692307639, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9960624297351877, 'Affiliation-P': 0.9928360285167797, 'Affiliation-R': 0.9993098688750863, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 566_YAHOO_id_16_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/566_YAHOO_id_16_Synthetic_tr_500_1st_623_results.pkl
Skipping file: 853_OPPORTUNITY_id_12_HumanActivity_tr_808_1st_908.csv due to filter criteria.
Processing file: 392_UCR_id_90_HumanActivity_tr_48750_1st_130700.csv, Data shape: (195000, 1)
Label shape: (195000,)
Train data shape: (48750, 1), Test data shape: (146250, 1), Label test shape: (146250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 392_UCR_id_90_HumanActivity_tr_48750_1st_130700.csv...
Using CUDA device: 2
Data shape: (146250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 146250, window_size 5000, adding 3750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([30, 5000, 1])
batched_mask shape: torch.Size([30, 5000])

100%|██████████| 1/1 [00:00<00:00, 46.87it/s]
Original shapes - Output: (150000,), Label: (146250,)
Aligned shapes - Output: (146250,), Label: (146250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  14%|█▎        | 27/200 [00:01<00:06, 26.44it/s]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:02<00:05, 26.38it/s]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [00:03<00:04, 27.21it/s]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [00:04<00:02, 29.01it/s]

Calculating Affiliation Metrics:  77%|███████▋  | 154/200 [00:05<00:01, 32.48it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:05<00:00, 33.43it/s]
Table1-only metrics completed in 11.39s
Evaluation Result:  {'AUC-PR': 0.03824299001071792, 'AUC-ROC': 0.8866919389003322, 'VUS-PR': 0.0391614849903273, 'VUS-ROC': 0.8888464697473606, 'Standard-F1': 0.09770324462248393, 'Standard-Precision': 0.05473856209150326, 'Standard-Recall': 0.4542372881355929, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7056198878200877, 'Affiliation-P': 0.5451763394911058, 'Affiliation-R': 0.9998816746342171, 'F1_T': 0.0972886762358574, 'Precision_T': 0.054324163863051646, 'Recall_T': 0.46525423728813525}
Results for 392_UCR_id_90_HumanActivity_tr_48750_1st_130700.csv: {'filename': '392_UCR_id_90_HumanActivity_tr_48750_1st_130700.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '48750', 'data_shape': '195000x1', 'output_length': 150000, 'label_length': 146250, 'aligned_length': 146250, 'AUC-PR': 0.03824299001071792, 'AUC-ROC': 0.8866919389003322, 'VUS-PR': 0.0391614849903273, 'VUS-ROC': 0.8888464697473606, 'Standard-F1': 0.09770324462248393, 'Standard-Precision': 0.05473856209150326, 'Standard-Recall': 0.4542372881355929, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7056198878200877, 'Affiliation-P': 0.5451763394911058, 'Affiliation-R': 0.9998816746342171, 'F1_T': 0.0972886762358574, 'Precision_T': 0.054324163863051646, 'Recall_T': 0.46525423728813525}
Logits metrics skipped for 392_UCR_id_90_HumanActivity_tr_48750_1st_130700.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/392_UCR_id_90_HumanActivity_tr_48750_1st_130700_results.pkl
Skipping file: 243_SVDB_id_7_Medical_tr_11393_1st_11493.csv due to filter criteria.
Skipping file: 179_SMD_id_2_Facility_tr_5925_1st_17580.csv due to filter criteria.
Processing file: 726_YAHOO_id_176_WebService_tr_500_1st_221.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 726_YAHOO_id_176_WebService_tr_500_1st_221.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 149.22it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 805.45it/s]
Table1-only metrics completed in 0.92s
Evaluation Result:  {'AUC-PR': 0.9285233996772456, 'AUC-ROC': 0.9949904422912135, 'VUS-PR': 0.9460880091565487, 'VUS-ROC': 0.9964167914350059, 'Standard-F1': 0.8799999999994303, 'Standard-Precision': 0.9166666666665902, 'Standard-Recall': 0.8461538461537811, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9539596492570712, 'Affiliation-P': 0.9869803466987133, 'Affiliation-R': 0.9230769230769231, 'F1_T': 0.8799999999994303, 'Precision_T': 0.9166666666665902, 'Recall_T': 0.8461538461537811}
Results for 726_YAHOO_id_176_WebService_tr_500_1st_221.csv: {'filename': '726_YAHOO_id_176_WebService_tr_500_1st_221.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9285233996772456, 'AUC-ROC': 0.9949904422912135, 'VUS-PR': 0.9460880091565487, 'VUS-ROC': 0.9964167914350059, 'Standard-F1': 0.8799999999994303, 'Standard-Precision': 0.9166666666665902, 'Standard-Recall': 0.8461538461537811, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9539596492570712, 'Affiliation-P': 0.9869803466987133, 'Affiliation-R': 0.9230769230769231, 'F1_T': 0.8799999999994303, 'Precision_T': 0.9166666666665902, 'Recall_T': 0.8461538461537811}
Logits metrics skipped for 726_YAHOO_id_176_WebService_tr_500_1st_221.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/726_YAHOO_id_176_WebService_tr_500_1st_221_results.pkl
Processing file: 085_WSD_id_57_WebService_tr_780_1st_880.csv, Data shape: (15403, 1)
Label shape: (15403,)
Train data shape: (780, 1), Test data shape: (14623, 1), Label test shape: (14623,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 085_WSD_id_57_WebService_tr_780_1st_880.csv...
Using CUDA device: 2
Data shape: (14623, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 14623, window_size 5000, adding 377 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 155.90it/s]
Original shapes - Output: (15000,), Label: (14623,)
Aligned shapes - Output: (14623,), Label: (14623,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  70%|███████   | 140/200 [00:01<00:00, 139.65it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 156.42it/s]
Table1-only metrics completed in 1.93s
Evaluation Result:  {'AUC-PR': 0.11101125841134192, 'AUC-ROC': 0.7458387707275745, 'VUS-PR': 0.08621732651457131, 'VUS-ROC': 0.7451783556809991, 'Standard-F1': 0.21052631578900757, 'Standard-Precision': 0.16666666666666619, 'Standard-Recall': 0.2857142857142843, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.747525978274808, 'Affiliation-P': 0.6461817153558369, 'Affiliation-R': 0.8865719798707113, 'F1_T': 0.20494699646597037, 'Precision_T': 0.15977961432506843, 'Recall_T': 0.2857142857142843}
Results for 085_WSD_id_57_WebService_tr_780_1st_880.csv: {'filename': '085_WSD_id_57_WebService_tr_780_1st_880.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 11, 'train_index': '780', 'data_shape': '15403x1', 'output_length': 15000, 'label_length': 14623, 'aligned_length': 14623, 'AUC-PR': 0.11101125841134192, 'AUC-ROC': 0.7458387707275745, 'VUS-PR': 0.08621732651457131, 'VUS-ROC': 0.7451783556809991, 'Standard-F1': 0.21052631578900757, 'Standard-Precision': 0.16666666666666619, 'Standard-Recall': 0.2857142857142843, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.747525978274808, 'Affiliation-P': 0.6461817153558369, 'Affiliation-R': 0.8865719798707113, 'F1_T': 0.20494699646597037, 'Precision_T': 0.15977961432506843, 'Recall_T': 0.2857142857142843}
Logits metrics skipped for 085_WSD_id_57_WebService_tr_780_1st_880.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/085_WSD_id_57_WebService_tr_780_1st_880_results.pkl
Processing file: 697_YAHOO_id_147_WebService_tr_500_1st_225.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 697_YAHOO_id_147_WebService_tr_500_1st_225.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 158.84it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1261.98it/s]
Table1-only metrics completed in 0.50s
Evaluation Result:  {'AUC-PR': 0.9999999999999998, 'AUC-ROC': 1.0, 'VUS-PR': 0.9622747865054364, 'VUS-ROC': 0.9998391580102762, 'Standard-F1': 0.7368421052626148, 'Standard-Precision': 0.5833333333332847, 'Standard-Recall': 0.9999999999998571, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9756411829672348, 'Affiliation-P': 0.9524408505540581, 'Affiliation-R': 1.0, 'F1_T': 0.7368421052626148, 'Precision_T': 0.5833333333332847, 'Recall_T': 0.9999999999998571}
Results for 697_YAHOO_id_147_WebService_tr_500_1st_225.csv: {'filename': '697_YAHOO_id_147_WebService_tr_500_1st_225.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9999999999999998, 'AUC-ROC': 1.0, 'VUS-PR': 0.9622747865054364, 'VUS-ROC': 0.9998391580102762, 'Standard-F1': 0.7368421052626148, 'Standard-Precision': 0.5833333333332847, 'Standard-Recall': 0.9999999999998571, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9756411829672348, 'Affiliation-P': 0.9524408505540581, 'Affiliation-R': 1.0, 'F1_T': 0.7368421052626148, 'Precision_T': 0.5833333333332847, 'Recall_T': 0.9999999999998571}
Logits metrics skipped for 697_YAHOO_id_147_WebService_tr_500_1st_225.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/697_YAHOO_id_147_WebService_tr_500_1st_225_results.pkl
Processing file: 225_MGAB_id_1_Synthetic_tr_25000_1st_38478.csv, Data shape: (100000, 1)
Label shape: (100000,)
Train data shape: (25000, 1), Test data shape: (75000, 1), Label test shape: (75000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 225_MGAB_id_1_Synthetic_tr_25000_1st_38478.csv...
Using CUDA device: 2
Data shape: (75000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([15, 5000, 1])
batched_mask shape: torch.Size([15, 5000])

100%|██████████| 1/1 [00:00<00:00, 78.79it/s]
Original shapes - Output: (75000,), Label: (75000,)
Aligned shapes - Output: (75000,), Label: (75000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  18%|█▊        | 35/200 [00:01<00:04, 34.32it/s]

Calculating Affiliation Metrics:  35%|███▌      | 70/200 [00:02<00:04, 28.31it/s]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [00:03<00:03, 25.68it/s]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [00:04<00:03, 24.44it/s]

Calculating Affiliation Metrics:  76%|███████▌  | 151/200 [00:05<00:02, 24.31it/s]

Calculating Affiliation Metrics:  93%|█████████▎| 186/200 [00:06<00:00, 27.51it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:07<00:00, 28.01it/s]
Table1-only metrics completed in 9.05s
Evaluation Result:  {'AUC-PR': 0.017974526041022836, 'AUC-ROC': 0.6502511029411765, 'VUS-PR': 0.012972234447694799, 'VUS-ROC': 0.6593897809151613, 'Standard-F1': 0.04580152671736288, 'Standard-Precision': 0.02570379436964503, 'Standard-Recall': 0.20999999999999897, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.734264510134012, 'Affiliation-P': 0.6569308091127505, 'Affiliation-R': 0.8322349646178262, 'F1_T': 0.04273919378321456, 'Precision_T': 0.023668639053254427, 'Recall_T': 0.21999999999999892}
Results for 225_MGAB_id_1_Synthetic_tr_25000_1st_38478.csv: {'filename': '225_MGAB_id_1_Synthetic_tr_25000_1st_38478.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 49, 'train_index': '25000', 'data_shape': '100000x1', 'output_length': 75000, 'label_length': 75000, 'aligned_length': 75000, 'AUC-PR': 0.017974526041022836, 'AUC-ROC': 0.6502511029411765, 'VUS-PR': 0.012972234447694799, 'VUS-ROC': 0.6593897809151613, 'Standard-F1': 0.04580152671736288, 'Standard-Precision': 0.02570379436964503, 'Standard-Recall': 0.20999999999999897, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.734264510134012, 'Affiliation-P': 0.6569308091127505, 'Affiliation-R': 0.8322349646178262, 'F1_T': 0.04273919378321456, 'Precision_T': 0.023668639053254427, 'Recall_T': 0.21999999999999892}
Logits metrics skipped for 225_MGAB_id_1_Synthetic_tr_25000_1st_38478.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/225_MGAB_id_1_Synthetic_tr_25000_1st_38478_results.pkl
Processing file: 149_Stock_id_1_Finance_tr_500_1st_7.csv, Data shape: (20000, 1)
Label shape: (20000,)
Train data shape: (500, 1), Test data shape: (19500, 1), Label test shape: (19500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 149_Stock_id_1_Finance_tr_500_1st_7.csv...
Using CUDA device: 2
Data shape: (19500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 19500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 154.10it/s]
Original shapes - Output: (20000,), Label: (19500,)
Aligned shapes - Output: (19500,), Label: (19500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   4%|▍         | 9/200 [00:01<00:21,  8.71it/s]

Calculating Affiliation Metrics:   9%|▉         | 18/200 [00:03<00:37,  4.91it/s]

Calculating Affiliation Metrics:  12%|█▏        | 24/200 [00:05<00:45,  3.88it/s]

Calculating Affiliation Metrics:  14%|█▍        | 29/200 [00:07<00:53,  3.22it/s]

Calculating Affiliation Metrics:  16%|█▋        | 33/200 [00:09<00:58,  2.86it/s]

Calculating Affiliation Metrics:  18%|█▊        | 37/200 [00:11<01:02,  2.63it/s]

Calculating Affiliation Metrics:  20%|██        | 40/200 [00:12<01:04,  2.50it/s]

Calculating Affiliation Metrics:  22%|██▏       | 43/200 [00:14<01:05,  2.38it/s]

Calculating Affiliation Metrics:  23%|██▎       | 46/200 [00:15<01:08,  2.24it/s]

Calculating Affiliation Metrics:  24%|██▍       | 49/200 [00:17<01:10,  2.14it/s]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:19<01:12,  2.04it/s]

Calculating Affiliation Metrics:  28%|██▊       | 55/200 [00:20<01:13,  1.96it/s]

Calculating Affiliation Metrics:  28%|██▊       | 57/200 [00:21<01:14,  1.92it/s]

Calculating Affiliation Metrics:  30%|██▉       | 59/200 [00:23<01:14,  1.88it/s]

Calculating Affiliation Metrics:  30%|███       | 61/200 [00:24<01:15,  1.84it/s]

Calculating Affiliation Metrics:  32%|███▏      | 63/200 [00:25<01:15,  1.82it/s]

Calculating Affiliation Metrics:  32%|███▎      | 65/200 [00:26<01:15,  1.79it/s]

Calculating Affiliation Metrics:  34%|███▎      | 67/200 [00:27<01:15,  1.77it/s]

Calculating Affiliation Metrics:  34%|███▍      | 69/200 [00:28<01:14,  1.76it/s]

Calculating Affiliation Metrics:  36%|███▌      | 71/200 [00:29<01:13,  1.76it/s]

Calculating Affiliation Metrics:  36%|███▋      | 73/200 [00:31<01:12,  1.76it/s]

Calculating Affiliation Metrics:  38%|███▊      | 75/200 [00:32<01:11,  1.76it/s]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [00:33<01:09,  1.77it/s]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [00:34<01:07,  1.80it/s]

Calculating Affiliation Metrics:  40%|████      | 81/200 [00:35<01:05,  1.80it/s]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [00:36<01:05,  1.80it/s]

Calculating Affiliation Metrics:  42%|████▎     | 85/200 [00:37<01:04,  1.79it/s]

Calculating Affiliation Metrics:  44%|████▎     | 87/200 [00:38<01:03,  1.78it/s]

Calculating Affiliation Metrics:  44%|████▍     | 89/200 [00:39<01:02,  1.79it/s]

Calculating Affiliation Metrics:  46%|████▌     | 91/200 [00:41<01:00,  1.81it/s]

Calculating Affiliation Metrics:  46%|████▋     | 93/200 [00:42<00:59,  1.80it/s]

Calculating Affiliation Metrics:  48%|████▊     | 95/200 [00:43<00:58,  1.80it/s]

Calculating Affiliation Metrics:  48%|████▊     | 97/200 [00:44<00:56,  1.81it/s]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [00:45<00:54,  1.84it/s]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [00:46<00:53,  1.87it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 103/200 [00:47<00:51,  1.90it/s]

Calculating Affiliation Metrics:  53%|█████▎    | 106/200 [00:48<00:48,  1.95it/s]

Calculating Affiliation Metrics:  55%|█████▍    | 109/200 [00:50<00:45,  2.02it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [00:51<00:42,  2.08it/s]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [00:52<00:39,  2.14it/s]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [00:54<00:36,  2.22it/s]

Calculating Affiliation Metrics:  60%|██████    | 121/200 [00:55<00:34,  2.31it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 124/200 [00:56<00:32,  2.37it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [00:57<00:29,  2.46it/s]

Calculating Affiliation Metrics:  65%|██████▌   | 130/200 [00:58<00:27,  2.56it/s]

Calculating Affiliation Metrics:  66%|██████▋   | 133/200 [00:59<00:25,  2.62it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 136/200 [01:00<00:24,  2.66it/s]

Calculating Affiliation Metrics:  70%|██████▉   | 139/200 [01:02<00:22,  2.70it/s]

Calculating Affiliation Metrics:  71%|███████   | 142/200 [01:03<00:21,  2.76it/s]

Calculating Affiliation Metrics:  73%|███████▎  | 146/200 [01:04<00:18,  2.86it/s]

Calculating Affiliation Metrics:  75%|███████▌  | 150/200 [01:05<00:17,  2.93it/s]

Calculating Affiliation Metrics:  77%|███████▋  | 154/200 [01:06<00:15,  3.02it/s]

Calculating Affiliation Metrics:  79%|███████▉  | 158/200 [01:08<00:13,  3.15it/s]

Calculating Affiliation Metrics:  81%|████████  | 162/200 [01:09<00:11,  3.30it/s]

Calculating Affiliation Metrics:  83%|████████▎ | 166/200 [01:10<00:09,  3.44it/s]

Calculating Affiliation Metrics:  86%|████████▌ | 171/200 [01:11<00:07,  3.68it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 176/200 [01:12<00:06,  3.98it/s]

Calculating Affiliation Metrics:  91%|█████████ | 182/200 [01:13<00:04,  4.48it/s]

Calculating Affiliation Metrics:  96%|█████████▌| 191/200 [01:14<00:01,  5.68it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [01:15<00:00,  2.67it/s]
Table1-only metrics completed in 108.30s
Evaluation Result:  {'AUC-PR': 0.14433650642473736, 'AUC-ROC': 0.6300900222802556, 'VUS-PR': 0.828321769385492, 'VUS-ROC': 0.9301619104122533, 'Standard-F1': 0.1729541497383331, 'Standard-Precision': 0.19350649350649324, 'Standard-Recall': 0.15634837355718767, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6783480309771007, 'Affiliation-P': 0.5239342503256972, 'Affiliation-R': 0.9618134751321611, 'F1_T': 0.1727536231879111, 'Precision_T': 0.19300518134715, 'Recall_T': 0.15634837355718767}
Results for 149_Stock_id_1_Finance_tr_500_1st_7.csv: {'filename': '149_Stock_id_1_Finance_tr_500_1st_7.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 112, 'train_index': '500', 'data_shape': '20000x1', 'output_length': 20000, 'label_length': 19500, 'aligned_length': 19500, 'AUC-PR': 0.14433650642473736, 'AUC-ROC': 0.6300900222802556, 'VUS-PR': 0.828321769385492, 'VUS-ROC': 0.9301619104122533, 'Standard-F1': 0.1729541497383331, 'Standard-Precision': 0.19350649350649324, 'Standard-Recall': 0.15634837355718767, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6783480309771007, 'Affiliation-P': 0.5239342503256972, 'Affiliation-R': 0.9618134751321611, 'F1_T': 0.1727536231879111, 'Precision_T': 0.19300518134715, 'Recall_T': 0.15634837355718767}
Logits metrics skipped for 149_Stock_id_1_Finance_tr_500_1st_7.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/149_Stock_id_1_Finance_tr_500_1st_7_results.pkl
Processing file: 019_NAB_id_19_Facility_tr_1007_1st_1171.csv, Data shape: (4031, 1)
Label shape: (4031,)
Train data shape: (1007, 1), Test data shape: (3024, 1), Label test shape: (3024,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 019_NAB_id_19_Facility_tr_1007_1st_1171.csv...
Using CUDA device: 2
Data shape: (3024, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 3024, 1])
batched_mask shape: torch.Size([1, 3024])

100%|██████████| 1/1 [00:00<00:00, 144.76it/s]
Original shapes - Output: (3024,), Label: (3024,)
Aligned shapes - Output: (3024,), Label: (3024,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 825.33it/s]
Table1-only metrics completed in 0.32s
Evaluation Result:  {'AUC-PR': 0.5419318474201057, 'AUC-ROC': 0.7363938643292683, 'VUS-PR': 0.5443888408027827, 'VUS-ROC': 0.7415495860042816, 'Standard-F1': 0.5398373983735273, 'Standard-Precision': 0.7720930232558104, 'Standard-Recall': 0.4149999999999989, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9826683390254606, 'Affiliation-P': 0.9979154899271094, 'Affiliation-R': 0.9678800988677532, 'F1_T': 0.53205128205082, 'Precision_T': 0.7410714285714253, 'Recall_T': 0.4149999999999989}
Results for 019_NAB_id_19_Facility_tr_1007_1st_1171.csv: {'filename': '019_NAB_id_19_Facility_tr_1007_1st_1171.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 8, 'train_index': '1007', 'data_shape': '4031x1', 'output_length': 3024, 'label_length': 3024, 'aligned_length': 3024, 'AUC-PR': 0.5419318474201057, 'AUC-ROC': 0.7363938643292683, 'VUS-PR': 0.5443888408027827, 'VUS-ROC': 0.7415495860042816, 'Standard-F1': 0.5398373983735273, 'Standard-Precision': 0.7720930232558104, 'Standard-Recall': 0.4149999999999989, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9826683390254606, 'Affiliation-P': 0.9979154899271094, 'Affiliation-R': 0.9678800988677532, 'F1_T': 0.53205128205082, 'Precision_T': 0.7410714285714253, 'Recall_T': 0.4149999999999989}
Logits metrics skipped for 019_NAB_id_19_Facility_tr_1007_1st_1171.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/019_NAB_id_19_Facility_tr_1007_1st_1171_results.pkl
Processing file: 707_YAHOO_id_157_WebService_tr_500_1st_137.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 707_YAHOO_id_157_WebService_tr_500_1st_137.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 166.13it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 898.21it/s]
Table1-only metrics completed in 0.63s
Evaluation Result:  {'AUC-PR': 0.9888888888888892, 'AUC-ROC': 0.9999051143372236, 'VUS-PR': 0.9815966236352034, 'VUS-ROC': 0.9999214168878408, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9820752335268865, 'Affiliation-P': 0.9647817460317459, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Results for 707_YAHOO_id_157_WebService_tr_500_1st_137.csv: {'filename': '707_YAHOO_id_157_WebService_tr_500_1st_137.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9888888888888892, 'AUC-ROC': 0.9999051143372236, 'VUS-PR': 0.9815966236352034, 'VUS-ROC': 0.9999214168878408, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9820752335268865, 'Affiliation-P': 0.9647817460317459, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 707_YAHOO_id_157_WebService_tr_500_1st_137.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/707_YAHOO_id_157_WebService_tr_500_1st_137_results.pkl
Processing file: 234_SED_id_1_Medical_tr_3740_1st_3840.csv, Data shape: (19999, 1)
Label shape: (19999,)
Train data shape: (3740, 1), Test data shape: (16259, 1), Label test shape: (16259,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 234_SED_id_1_Medical_tr_3740_1st_3840.csv...
Using CUDA device: 2
Data shape: (16259, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 16259, window_size 5000, adding 3741 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.74it/s]
Original shapes - Output: (20000,), Label: (16259,)
Aligned shapes - Output: (16259,), Label: (16259,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  60%|██████    | 120/200 [00:01<00:00, 119.26it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 138.88it/s]
Table1-only metrics completed in 5.70s
Evaluation Result:  {'AUC-PR': 0.48989815786440677, 'AUC-ROC': 0.904679499749439, 'VUS-PR': 0.6721420614137845, 'VUS-ROC': 0.9525012251770205, 'Standard-F1': 0.563190773666009, 'Standard-Precision': 0.5227475468331843, 'Standard-Recall': 0.610416666666666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8439840061928395, 'Affiliation-P': 0.8374177846554559, 'Affiliation-R': 0.8506540134536724, 'F1_T': 0.5633397312854911, 'Precision_T': 0.5222419928825619, 'Recall_T': 0.6114583333333327}
Results for 234_SED_id_1_Medical_tr_3740_1st_3840.csv: {'filename': '234_SED_id_1_Medical_tr_3740_1st_3840.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '3740', 'data_shape': '19999x1', 'output_length': 20000, 'label_length': 16259, 'aligned_length': 16259, 'AUC-PR': 0.48989815786440677, 'AUC-ROC': 0.904679499749439, 'VUS-PR': 0.6721420614137845, 'VUS-ROC': 0.9525012251770205, 'Standard-F1': 0.563190773666009, 'Standard-Precision': 0.5227475468331843, 'Standard-Recall': 0.610416666666666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8439840061928395, 'Affiliation-P': 0.8374177846554559, 'Affiliation-R': 0.8506540134536724, 'F1_T': 0.5633397312854911, 'Precision_T': 0.5222419928825619, 'Recall_T': 0.6114583333333327}
Logits metrics skipped for 234_SED_id_1_Medical_tr_3740_1st_3840.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/234_SED_id_1_Medical_tr_3740_1st_3840_results.pkl
Processing file: 374_UCR_id_72_Medical_tr_2500_1st_7000.csv, Data shape: (10001, 1)
Label shape: (10001,)
Train data shape: (2500, 1), Test data shape: (7501, 1), Label test shape: (7501,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 374_UCR_id_72_Medical_tr_2500_1st_7000.csv...
Using CUDA device: 2
Data shape: (7501, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 7501, window_size 5000, adding 2499 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 155.97it/s]
Original shapes - Output: (10000,), Label: (7501,)
Aligned shapes - Output: (7501,), Label: (7501,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 497.28it/s]
Table1-only metrics completed in 0.54s
Evaluation Result:  {'AUC-PR': 0.011076850409918928, 'AUC-ROC': 0.8328916253959755, 'VUS-PR': 0.012038809818711534, 'VUS-ROC': 0.8374510191785154, 'Standard-F1': 0.031908831908798264, 'Standard-Precision': 0.016231884057971005, 'Standard-Recall': 0.9333333333333023, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8207137159994652, 'Affiliation-P': 0.8171682372174939, 'Affiliation-R': 0.8242900946540461, 'F1_T': 0.03167420814476299, 'Precision_T': 0.01611047180667433, 'Recall_T': 0.9333333333333023}
Results for 374_UCR_id_72_Medical_tr_2500_1st_7000.csv: {'filename': '374_UCR_id_72_Medical_tr_2500_1st_7000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 26, 'train_index': '2500', 'data_shape': '10001x1', 'output_length': 10000, 'label_length': 7501, 'aligned_length': 7501, 'AUC-PR': 0.011076850409918928, 'AUC-ROC': 0.8328916253959755, 'VUS-PR': 0.012038809818711534, 'VUS-ROC': 0.8374510191785154, 'Standard-F1': 0.031908831908798264, 'Standard-Precision': 0.016231884057971005, 'Standard-Recall': 0.9333333333333023, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8207137159994652, 'Affiliation-P': 0.8171682372174939, 'Affiliation-R': 0.8242900946540461, 'F1_T': 0.03167420814476299, 'Precision_T': 0.01611047180667433, 'Recall_T': 0.9333333333333023}
Logits metrics skipped for 374_UCR_id_72_Medical_tr_2500_1st_7000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/374_UCR_id_72_Medical_tr_2500_1st_7000_results.pkl
Processing file: 686_YAHOO_id_136_WebService_tr_500_1st_755.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 686_YAHOO_id_136_WebService_tr_500_1st_755.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 163.44it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1185.35it/s]
Table1-only metrics completed in 0.43s
Evaluation Result:  {'AUC-PR': 0.7249999999999999, 'AUC-ROC': 0.9933617021276595, 'VUS-PR': 0.8222601236007508, 'VUS-ROC': 0.9955014238870115, 'Standard-F1': 0.4705882352936471, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.79999999999984, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9612924182737942, 'Affiliation-P': 0.9254697233231358, 'Affiliation-R': 1.0, 'F1_T': 0.4705882352936471, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.79999999999984}
Results for 686_YAHOO_id_136_WebService_tr_500_1st_755.csv: {'filename': '686_YAHOO_id_136_WebService_tr_500_1st_755.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.7249999999999999, 'AUC-ROC': 0.9933617021276595, 'VUS-PR': 0.8222601236007508, 'VUS-ROC': 0.9955014238870115, 'Standard-F1': 0.4705882352936471, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.79999999999984, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9612924182737942, 'Affiliation-P': 0.9254697233231358, 'Affiliation-R': 1.0, 'F1_T': 0.4705882352936471, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.79999999999984}
Logits metrics skipped for 686_YAHOO_id_136_WebService_tr_500_1st_755.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/686_YAHOO_id_136_WebService_tr_500_1st_755_results.pkl
Processing file: 275_IOPS_id_16_WebService_tr_1482_1st_1582.csv, Data shape: (8784, 1)
Label shape: (8784,)
Train data shape: (1482, 1), Test data shape: (7302, 1), Label test shape: (7302,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 275_IOPS_id_16_WebService_tr_1482_1st_1582.csv...
Using CUDA device: 2
Data shape: (7302, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 7302, window_size 5000, adding 2698 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.10it/s]
Original shapes - Output: (10000,), Label: (7302,)
Aligned shapes - Output: (7302,), Label: (7302,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 485.97it/s]
Table1-only metrics completed in 4.16s
Evaluation Result:  {'AUC-PR': 0.02740309967735058, 'AUC-ROC': 0.8010171386499052, 'VUS-PR': 0.0950564581693787, 'VUS-ROC': 0.7235685634178516, 'Standard-F1': 0.08154188287610477, 'Standard-Precision': 0.04303599374021906, 'Standard-Recall': 0.7746478873239329, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7358326676396119, 'Affiliation-P': 0.5908097584727796, 'Affiliation-R': 0.9752129775057069, 'F1_T': 0.08114199849726927, 'Precision_T': 0.04285714285714282, 'Recall_T': 0.7605633802816795}
Results for 275_IOPS_id_16_WebService_tr_1482_1st_1582.csv: {'filename': '275_IOPS_id_16_WebService_tr_1482_1st_1582.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 288, 'train_index': '1482', 'data_shape': '8784x1', 'output_length': 10000, 'label_length': 7302, 'aligned_length': 7302, 'AUC-PR': 0.02740309967735058, 'AUC-ROC': 0.8010171386499052, 'VUS-PR': 0.0950564581693787, 'VUS-ROC': 0.7235685634178516, 'Standard-F1': 0.08154188287610477, 'Standard-Precision': 0.04303599374021906, 'Standard-Recall': 0.7746478873239329, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7358326676396119, 'Affiliation-P': 0.5908097584727796, 'Affiliation-R': 0.9752129775057069, 'F1_T': 0.08114199849726927, 'Precision_T': 0.04285714285714282, 'Recall_T': 0.7605633802816795}
Logits metrics skipped for 275_IOPS_id_16_WebService_tr_1482_1st_1582.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/275_IOPS_id_16_WebService_tr_1482_1st_1582_results.pkl
Processing file: 571_YAHOO_id_21_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 571_YAHOO_id_21_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 158.17it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1887.64it/s]
Table1-only metrics completed in 0.35s
Evaluation Result:  {'AUC-PR': 0.9094356261022929, 'AUC-ROC': 0.9990253411306043, 'VUS-PR': 0.9820605815874914, 'VUS-ROC': 0.99983402391278, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 571_YAHOO_id_21_Synthetic_tr_500_1st_623.csv: {'filename': '571_YAHOO_id_21_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 35, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9094356261022929, 'AUC-ROC': 0.9990253411306043, 'VUS-PR': 0.9820605815874914, 'VUS-ROC': 0.99983402391278, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 571_YAHOO_id_21_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/571_YAHOO_id_21_Synthetic_tr_500_1st_623_results.pkl
Processing file: 731_YAHOO_id_181_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 731_YAHOO_id_181_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 161.57it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1989.92it/s]
Table1-only metrics completed in 0.80s
Evaluation Result:  {'AUC-PR': 0.9170113836780505, 'AUC-ROC': 0.9991471734892788, 'VUS-PR': 0.9939765667872729, 'VUS-ROC': 0.9999444710668464, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 731_YAHOO_id_181_Synthetic_tr_500_1st_623.csv: {'filename': '731_YAHOO_id_181_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 119, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9170113836780505, 'AUC-ROC': 0.9991471734892788, 'VUS-PR': 0.9939765667872729, 'VUS-ROC': 0.9999444710668464, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 731_YAHOO_id_181_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/731_YAHOO_id_181_Synthetic_tr_500_1st_623_results.pkl
Processing file: 491_UCR_id_189_Facility_tr_32372_1st_114000.csv, Data shape: (129489, 1)
Label shape: (129489,)
Train data shape: (32372, 1), Test data shape: (97117, 1), Label test shape: (97117,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 491_UCR_id_189_Facility_tr_32372_1st_114000.csv...
Using CUDA device: 2
Data shape: (97117, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 97117, window_size 5000, adding 2883 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([20, 5000, 1])
batched_mask shape: torch.Size([20, 5000])

100%|██████████| 1/1 [00:00<00:00, 63.48it/s]
Original shapes - Output: (100000,), Label: (97117,)
Aligned shapes - Output: (97117,), Label: (97117,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  20%|██        | 40/200 [00:01<00:04, 38.96it/s]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [00:02<00:03, 37.59it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 124/200 [00:03<00:01, 40.59it/s]

Calculating Affiliation Metrics:  94%|█████████▍| 189/200 [00:04<00:00, 49.98it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:04<00:00, 47.38it/s]
Table1-only metrics completed in 10.43s
Evaluation Result:  {'AUC-PR': 0.956949071160022, 'AUC-ROC': 0.9990758565319017, 'VUS-PR': 0.8423164672119675, 'VUS-ROC': 0.9991080397934239, 'Standard-F1': 0.53204172876264, 'Standard-Precision': 0.3672839506172836, 'Standard-Recall': 0.9648648648648622, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.937525335345112, 'Affiliation-P': 0.8824065813307681, 'Affiliation-R': 0.9999887569437699, 'F1_T': 0.53204172876264, 'Precision_T': 0.3672839506172836, 'Recall_T': 0.9648648648648622}
Results for 491_UCR_id_189_Facility_tr_32372_1st_114000.csv: {'filename': '491_UCR_id_189_Facility_tr_32372_1st_114000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 243, 'train_index': '32372', 'data_shape': '129489x1', 'output_length': 100000, 'label_length': 97117, 'aligned_length': 97117, 'AUC-PR': 0.956949071160022, 'AUC-ROC': 0.9990758565319017, 'VUS-PR': 0.8423164672119675, 'VUS-ROC': 0.9991080397934239, 'Standard-F1': 0.53204172876264, 'Standard-Precision': 0.3672839506172836, 'Standard-Recall': 0.9648648648648622, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.937525335345112, 'Affiliation-P': 0.8824065813307681, 'Affiliation-R': 0.9999887569437699, 'F1_T': 0.53204172876264, 'Precision_T': 0.3672839506172836, 'Recall_T': 0.9648648648648622}
Logits metrics skipped for 491_UCR_id_189_Facility_tr_32372_1st_114000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/491_UCR_id_189_Facility_tr_32372_1st_114000_results.pkl
Processing file: 469_UCR_id_167_Medical_tr_1875_1st_4187.csv, Data shape: (7500, 1)
Label shape: (7500,)
Train data shape: (1875, 1), Test data shape: (5625, 1), Label test shape: (5625,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 469_UCR_id_167_Medical_tr_1875_1st_4187.csv...
Using CUDA device: 2
Data shape: (5625, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5625, window_size 5000, adding 4375 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.82it/s]
Original shapes - Output: (10000,), Label: (5625,)
Aligned shapes - Output: (5625,), Label: (5625,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 704.98it/s]
Table1-only metrics completed in 0.96s
Evaluation Result:  {'AUC-PR': 0.6987851102751882, 'AUC-ROC': 0.9920391807658059, 'VUS-PR': 0.355957071948004, 'VUS-ROC': 0.9966917487090938, 'Standard-F1': 0.20895522388033685, 'Standard-Precision': 0.12280701754385749, 'Standard-Recall': 0.69999999999993, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8667057346919593, 'Affiliation-P': 0.7648187134502924, 'Affiliation-R': 0.9999111111111111, 'F1_T': 0.20895522388033685, 'Precision_T': 0.12280701754385749, 'Recall_T': 0.69999999999993}
Results for 469_UCR_id_167_Medical_tr_1875_1st_4187.csv: {'filename': '469_UCR_id_167_Medical_tr_1875_1st_4187.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 183, 'train_index': '1875', 'data_shape': '7500x1', 'output_length': 10000, 'label_length': 5625, 'aligned_length': 5625, 'AUC-PR': 0.6987851102751882, 'AUC-ROC': 0.9920391807658059, 'VUS-PR': 0.355957071948004, 'VUS-ROC': 0.9966917487090938, 'Standard-F1': 0.20895522388033685, 'Standard-Precision': 0.12280701754385749, 'Standard-Recall': 0.69999999999993, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8667057346919593, 'Affiliation-P': 0.7648187134502924, 'Affiliation-R': 0.9999111111111111, 'F1_T': 0.20895522388033685, 'Precision_T': 0.12280701754385749, 'Recall_T': 0.69999999999993}
Logits metrics skipped for 469_UCR_id_167_Medical_tr_1875_1st_4187.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/469_UCR_id_167_Medical_tr_1875_1st_4187_results.pkl
Skipping file: 256_SVDB_id_20_Medical_tr_6264_1st_6364.csv due to filter criteria.
Processing file: 754_YAHOO_id_204_WebService_tr_500_1st_28.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 754_YAHOO_id_204_WebService_tr_500_1st_28.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 159.43it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1217.86it/s]
Table1-only metrics completed in 0.52s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9769308846506938, 'VUS-ROC': 0.9999016779536335, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Results for 754_YAHOO_id_204_WebService_tr_500_1st_28.csv: {'filename': '754_YAHOO_id_204_WebService_tr_500_1st_28.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9769308846506938, 'VUS-ROC': 0.9999016779536335, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Logits metrics skipped for 754_YAHOO_id_204_WebService_tr_500_1st_28.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/754_YAHOO_id_204_WebService_tr_500_1st_28_results.pkl
Processing file: 808_YAHOO_id_258_WebService_tr_500_1st_142.csv, Data shape: (1439, 1)
Label shape: (1439,)
Train data shape: (500, 1), Test data shape: (939, 1), Label test shape: (939,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 808_YAHOO_id_258_WebService_tr_500_1st_142.csv...
Using CUDA device: 2
Data shape: (939, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 939, 1])
batched_mask shape: torch.Size([1, 939])

100%|██████████| 1/1 [00:00<00:00, 161.04it/s]
Original shapes - Output: (939,), Label: (939,)
Aligned shapes - Output: (939,), Label: (939,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1958.08it/s]
Table1-only metrics completed in 0.25s
Evaluation Result:  {'AUC-PR': 0.7838437839797028, 'AUC-ROC': 0.973952738990333, 'VUS-PR': 0.873535180827315, 'VUS-ROC': 0.9965684782179924, 'Standard-F1': 0.6666666666660988, 'Standard-Precision': 0.59999999999994, 'Standard-Recall': 0.7499999999999063, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9980855137204845, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9961783439490446, 'F1_T': 0.6666666666660988, 'Precision_T': 0.59999999999994, 'Recall_T': 0.7499999999999063}
Results for 808_YAHOO_id_258_WebService_tr_500_1st_142.csv: {'filename': '808_YAHOO_id_258_WebService_tr_500_1st_142.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1439x1', 'output_length': 939, 'label_length': 939, 'aligned_length': 939, 'AUC-PR': 0.7838437839797028, 'AUC-ROC': 0.973952738990333, 'VUS-PR': 0.873535180827315, 'VUS-ROC': 0.9965684782179924, 'Standard-F1': 0.6666666666660988, 'Standard-Precision': 0.59999999999994, 'Standard-Recall': 0.7499999999999063, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9980855137204845, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9961783439490446, 'F1_T': 0.6666666666660988, 'Precision_T': 0.59999999999994, 'Recall_T': 0.7499999999999063}
Logits metrics skipped for 808_YAHOO_id_258_WebService_tr_500_1st_142.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/808_YAHOO_id_258_WebService_tr_500_1st_142_results.pkl
Processing file: 764_YAHOO_id_214_WebService_tr_500_1st_599.csv, Data shape: (1427, 1)
Label shape: (1427,)
Train data shape: (500, 1), Test data shape: (927, 1), Label test shape: (927,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 764_YAHOO_id_214_WebService_tr_500_1st_599.csv...
Using CUDA device: 2
Data shape: (927, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 927, 1])
batched_mask shape: torch.Size([1, 927])

100%|██████████| 1/1 [00:00<00:00, 159.24it/s]
Original shapes - Output: (927,), Label: (927,)
Aligned shapes - Output: (927,), Label: (927,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1884.08it/s]
Table1-only metrics completed in 0.31s
Evaluation Result:  {'AUC-PR': 0.43606988306314576, 'AUC-ROC': 0.9103122730573712, 'VUS-PR': 0.46737727411677293, 'VUS-ROC': 0.964082837954463, 'Standard-F1': 0.42105263157840445, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.44444444444439507, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.967849261323568, 'Affiliation-P': 0.9412837261534753, 'Affiliation-R': 0.995957845852098, 'F1_T': 0.42105263157840445, 'Precision_T': 0.39999999999996, 'Recall_T': 0.44444444444439507}
Results for 764_YAHOO_id_214_WebService_tr_500_1st_599.csv: {'filename': '764_YAHOO_id_214_WebService_tr_500_1st_599.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1427x1', 'output_length': 927, 'label_length': 927, 'aligned_length': 927, 'AUC-PR': 0.43606988306314576, 'AUC-ROC': 0.9103122730573712, 'VUS-PR': 0.46737727411677293, 'VUS-ROC': 0.964082837954463, 'Standard-F1': 0.42105263157840445, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.44444444444439507, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.967849261323568, 'Affiliation-P': 0.9412837261534753, 'Affiliation-R': 0.995957845852098, 'F1_T': 0.42105263157840445, 'Precision_T': 0.39999999999996, 'Recall_T': 0.44444444444439507}
Logits metrics skipped for 764_YAHOO_id_214_WebService_tr_500_1st_599.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/764_YAHOO_id_214_WebService_tr_500_1st_599_results.pkl
Processing file: 090_WSD_id_62_WebService_tr_2281_1st_2381.csv, Data shape: (13236, 1)
Label shape: (13236,)
Train data shape: (2281, 1), Test data shape: (10955, 1), Label test shape: (10955,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 090_WSD_id_62_WebService_tr_2281_1st_2381.csv...
Using CUDA device: 2
Data shape: (10955, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 10955, window_size 5000, adding 4045 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.00it/s]
Original shapes - Output: (15000,), Label: (10955,)
Aligned shapes - Output: (10955,), Label: (10955,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 280.22it/s]
Table1-only metrics completed in 1.26s
Evaluation Result:  {'AUC-PR': 0.19572806204247128, 'AUC-ROC': 0.7359703840814437, 'VUS-PR': 0.10259019988403784, 'VUS-ROC': 0.8291604004502819, 'Standard-F1': 0.1999999999995103, 'Standard-Precision': 0.23636363636363422, 'Standard-Recall': 0.17333333333333217, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7619111680066338, 'Affiliation-P': 0.6158735830866129, 'Affiliation-R': 0.9987335216225208, 'F1_T': 0.1999999999995103, 'Precision_T': 0.23636363636363422, 'Recall_T': 0.17333333333333217}
Results for 090_WSD_id_62_WebService_tr_2281_1st_2381.csv: {'filename': '090_WSD_id_62_WebService_tr_2281_1st_2381.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 19, 'train_index': '2281', 'data_shape': '13236x1', 'output_length': 15000, 'label_length': 10955, 'aligned_length': 10955, 'AUC-PR': 0.19572806204247128, 'AUC-ROC': 0.7359703840814437, 'VUS-PR': 0.10259019988403784, 'VUS-ROC': 0.8291604004502819, 'Standard-F1': 0.1999999999995103, 'Standard-Precision': 0.23636363636363422, 'Standard-Recall': 0.17333333333333217, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7619111680066338, 'Affiliation-P': 0.6158735830866129, 'Affiliation-R': 0.9987335216225208, 'F1_T': 0.1999999999995103, 'Precision_T': 0.23636363636363422, 'Recall_T': 0.17333333333333217}
Logits metrics skipped for 090_WSD_id_62_WebService_tr_2281_1st_2381.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/090_WSD_id_62_WebService_tr_2281_1st_2381_results.pkl
Skipping file: 843_OPPORTUNITY_id_2_HumanActivity_tr_808_1st_908.csv due to filter criteria.
Processing file: 088_WSD_id_60_WebService_tr_4548_1st_5596.csv, Data shape: (18193, 1)
Label shape: (18193,)
Train data shape: (4548, 1), Test data shape: (13645, 1), Label test shape: (13645,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 088_WSD_id_60_WebService_tr_4548_1st_5596.csv...
Using CUDA device: 2
Data shape: (13645, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13645, window_size 5000, adding 1355 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 154.12it/s]
Original shapes - Output: (15000,), Label: (13645,)
Aligned shapes - Output: (13645,), Label: (13645,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 241.62it/s]
Table1-only metrics completed in 1.21s
Evaluation Result:  {'AUC-PR': 0.2240123571963481, 'AUC-ROC': 0.8795370109309933, 'VUS-PR': 0.1045584897893176, 'VUS-ROC': 0.9393805819960683, 'Standard-F1': 0.15306122448937357, 'Standard-Precision': 0.10948905109488971, 'Standard-Recall': 0.2542372881355889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8471965373864726, 'Affiliation-P': 0.7350449198129793, 'Affiliation-R': 0.9997338366782635, 'F1_T': 0.15306122448937357, 'Precision_T': 0.10948905109488971, 'Recall_T': 0.2542372881355889}
Results for 088_WSD_id_60_WebService_tr_4548_1st_5596.csv: {'filename': '088_WSD_id_60_WebService_tr_4548_1st_5596.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '4548', 'data_shape': '18193x1', 'output_length': 15000, 'label_length': 13645, 'aligned_length': 13645, 'AUC-PR': 0.2240123571963481, 'AUC-ROC': 0.8795370109309933, 'VUS-PR': 0.1045584897893176, 'VUS-ROC': 0.9393805819960683, 'Standard-F1': 0.15306122448937357, 'Standard-Precision': 0.10948905109488971, 'Standard-Recall': 0.2542372881355889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8471965373864726, 'Affiliation-P': 0.7350449198129793, 'Affiliation-R': 0.9997338366782635, 'F1_T': 0.15306122448937357, 'Precision_T': 0.10948905109488971, 'Recall_T': 0.2542372881355889}
Logits metrics skipped for 088_WSD_id_60_WebService_tr_4548_1st_5596.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/088_WSD_id_60_WebService_tr_4548_1st_5596_results.pkl
Processing file: 015_NAB_id_15_Synthetic_tr_1007_1st_2787.csv, Data shape: (4031, 1)
Label shape: (4031,)
Train data shape: (1007, 1), Test data shape: (3024, 1), Label test shape: (3024,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 015_NAB_id_15_Synthetic_tr_1007_1st_2787.csv...
Using CUDA device: 2
Data shape: (3024, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 3024, 1])
batched_mask shape: torch.Size([1, 3024])

100%|██████████| 1/1 [00:00<00:00, 160.11it/s]
Original shapes - Output: (3024,), Label: (3024,)
Aligned shapes - Output: (3024,), Label: (3024,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1100.90it/s]
Table1-only metrics completed in 1.28s
Evaluation Result:  {'AUC-PR': 0.8549791680249704, 'AUC-ROC': 0.9660705371996742, 'VUS-PR': 0.9296812759170177, 'VUS-ROC': 0.9855386628829449, 'Standard-F1': 0.8003629764060344, 'Standard-Precision': 0.7487266553480463, 'Standard-Recall': 0.8596491228070158, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9884530220726426, 'Affiliation-P': 0.9991326696370556, 'Affiliation-R': 0.9779992677166167, 'F1_T': 0.7981735159812356, 'Precision_T': 0.7508591065292083, 'Recall_T': 0.8518518518518502}
Results for 015_NAB_id_15_Synthetic_tr_1007_1st_2787.csv: {'filename': '015_NAB_id_15_Synthetic_tr_1007_1st_2787.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 288, 'train_index': '1007', 'data_shape': '4031x1', 'output_length': 3024, 'label_length': 3024, 'aligned_length': 3024, 'AUC-PR': 0.8549791680249704, 'AUC-ROC': 0.9660705371996742, 'VUS-PR': 0.9296812759170177, 'VUS-ROC': 0.9855386628829449, 'Standard-F1': 0.8003629764060344, 'Standard-Precision': 0.7487266553480463, 'Standard-Recall': 0.8596491228070158, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9884530220726426, 'Affiliation-P': 0.9991326696370556, 'Affiliation-R': 0.9779992677166167, 'F1_T': 0.7981735159812356, 'Precision_T': 0.7508591065292083, 'Recall_T': 0.8518518518518502}
Logits metrics skipped for 015_NAB_id_15_Synthetic_tr_1007_1st_2787.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/015_NAB_id_15_Synthetic_tr_1007_1st_2787_results.pkl
Processing file: 226_MGAB_id_2_Synthetic_tr_20000_1st_45134.csv, Data shape: (80000, 1)
Label shape: (80000,)
Train data shape: (20000, 1), Test data shape: (60000, 1), Label test shape: (60000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 226_MGAB_id_2_Synthetic_tr_20000_1st_45134.csv...
Using CUDA device: 2
Data shape: (60000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([12, 5000, 1])
batched_mask shape: torch.Size([12, 5000])

100%|██████████| 1/1 [00:00<00:00, 90.06it/s]
Original shapes - Output: (60000,), Label: (60000,)
Aligned shapes - Output: (60000,), Label: (60000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  24%|██▍       | 48/200 [00:01<00:03, 46.67it/s]

Calculating Affiliation Metrics:  48%|████▊     | 95/200 [00:02<00:02, 37.41it/s]

Calculating Affiliation Metrics:  67%|██████▋   | 134/200 [00:03<00:01, 34.74it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 177/200 [00:04<00:00, 37.42it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:04<00:00, 40.19it/s]
Table1-only metrics completed in 6.30s
Evaluation Result:  {'AUC-PR': 0.03039010462181728, 'AUC-ROC': 0.8075451649085963, 'VUS-PR': 0.04819944880019186, 'VUS-ROC': 0.8305686060410199, 'Standard-F1': 0.09631147540959015, 'Standard-Precision': 0.05622009569377984, 'Standard-Recall': 0.33571428571428336, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7864434929211253, 'Affiliation-P': 0.6683005023650253, 'Affiliation-R': 0.9553274655339667, 'F1_T': 0.09073359073335681, 'Precision_T': 0.05245535714285708, 'Recall_T': 0.33571428571428336}
Results for 226_MGAB_id_2_Synthetic_tr_20000_1st_45134.csv: {'filename': '226_MGAB_id_2_Synthetic_tr_20000_1st_45134.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 49, 'train_index': '20000', 'data_shape': '80000x1', 'output_length': 60000, 'label_length': 60000, 'aligned_length': 60000, 'AUC-PR': 0.03039010462181728, 'AUC-ROC': 0.8075451649085963, 'VUS-PR': 0.04819944880019186, 'VUS-ROC': 0.8305686060410199, 'Standard-F1': 0.09631147540959015, 'Standard-Precision': 0.05622009569377984, 'Standard-Recall': 0.33571428571428336, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7864434929211253, 'Affiliation-P': 0.6683005023650253, 'Affiliation-R': 0.9553274655339667, 'F1_T': 0.09073359073335681, 'Precision_T': 0.05245535714285708, 'Recall_T': 0.33571428571428336}
Logits metrics skipped for 226_MGAB_id_2_Synthetic_tr_20000_1st_45134.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/226_MGAB_id_2_Synthetic_tr_20000_1st_45134_results.pkl
Skipping file: 846_OPPORTUNITY_id_5_HumanActivity_tr_2085_1st_2185.csv due to filter criteria.
Processing file: 112_WSD_id_84_WebService_tr_2442_1st_2542.csv, Data shape: (18380, 1)
Label shape: (18380,)
Train data shape: (2442, 1), Test data shape: (15938, 1), Label test shape: (15938,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 112_WSD_id_84_WebService_tr_2442_1st_2542.csv...
Using CUDA device: 2
Data shape: (15938, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 15938, window_size 5000, adding 4062 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 154.99it/s]
Original shapes - Output: (20000,), Label: (15938,)
Aligned shapes - Output: (15938,), Label: (15938,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 209.37it/s]
Table1-only metrics completed in 2.93s
Evaluation Result:  {'AUC-PR': 0.23900215385935608, 'AUC-ROC': 0.8379932508436445, 'VUS-PR': 0.11726206114850511, 'VUS-ROC': 0.9614420868950837, 'Standard-F1': 0.13452914798165622, 'Standard-Precision': 0.09374999999999942, 'Standard-Recall': 0.2380952380952343, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8508087839997903, 'Affiliation-P': 0.7412809025386778, 'Affiliation-R': 0.9983146613457441, 'F1_T': 0.13452914798165622, 'Precision_T': 0.09374999999999942, 'Recall_T': 0.2380952380952343}
Results for 112_WSD_id_84_WebService_tr_2442_1st_2542.csv: {'filename': '112_WSD_id_84_WebService_tr_2442_1st_2542.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '2442', 'data_shape': '18380x1', 'output_length': 20000, 'label_length': 15938, 'aligned_length': 15938, 'AUC-PR': 0.23900215385935608, 'AUC-ROC': 0.8379932508436445, 'VUS-PR': 0.11726206114850511, 'VUS-ROC': 0.9614420868950837, 'Standard-F1': 0.13452914798165622, 'Standard-Precision': 0.09374999999999942, 'Standard-Recall': 0.2380952380952343, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8508087839997903, 'Affiliation-P': 0.7412809025386778, 'Affiliation-R': 0.9983146613457441, 'F1_T': 0.13452914798165622, 'Precision_T': 0.09374999999999942, 'Recall_T': 0.2380952380952343}
Logits metrics skipped for 112_WSD_id_84_WebService_tr_2442_1st_2542.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/112_WSD_id_84_WebService_tr_2442_1st_2542_results.pkl
Processing file: 078_WSD_id_50_WebService_tr_2449_1st_2549.csv, Data shape: (18381, 1)
Label shape: (18381,)
Train data shape: (2449, 1), Test data shape: (15932, 1), Label test shape: (15932,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 078_WSD_id_50_WebService_tr_2449_1st_2549.csv...
Using CUDA device: 2
Data shape: (15932, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 15932, window_size 5000, adding 4068 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 154.74it/s]
Original shapes - Output: (20000,), Label: (15932,)
Aligned shapes - Output: (15932,), Label: (15932,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 212.58it/s]
Table1-only metrics completed in 2.94s
Evaluation Result:  {'AUC-PR': 0.31313368202616454, 'AUC-ROC': 0.7721651505950987, 'VUS-PR': 0.26153999827124247, 'VUS-ROC': 0.9581652367727855, 'Standard-F1': 0.2060085836905551, 'Standard-Precision': 0.14999999999999908, 'Standard-Recall': 0.3287671232876668, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9819828322811488, 'Affiliation-P': 0.9664395667442071, 'Affiliation-R': 0.9980342354630592, 'F1_T': 0.2060085836905551, 'Precision_T': 0.14999999999999908, 'Recall_T': 0.3287671232876668}
Results for 078_WSD_id_50_WebService_tr_2449_1st_2549.csv: {'filename': '078_WSD_id_50_WebService_tr_2449_1st_2549.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '2449', 'data_shape': '18381x1', 'output_length': 20000, 'label_length': 15932, 'aligned_length': 15932, 'AUC-PR': 0.31313368202616454, 'AUC-ROC': 0.7721651505950987, 'VUS-PR': 0.26153999827124247, 'VUS-ROC': 0.9581652367727855, 'Standard-F1': 0.2060085836905551, 'Standard-Precision': 0.14999999999999908, 'Standard-Recall': 0.3287671232876668, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9819828322811488, 'Affiliation-P': 0.9664395667442071, 'Affiliation-R': 0.9980342354630592, 'F1_T': 0.2060085836905551, 'Precision_T': 0.14999999999999908, 'Recall_T': 0.3287671232876668}
Logits metrics skipped for 078_WSD_id_50_WebService_tr_2449_1st_2549.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/078_WSD_id_50_WebService_tr_2449_1st_2549_results.pkl
Processing file: 689_YAHOO_id_139_WebService_tr_500_1st_1073.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 689_YAHOO_id_139_WebService_tr_500_1st_1073.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 157.51it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1753.50it/s]
Table1-only metrics completed in 0.21s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9927443448570203, 'Affiliation-P': 0.9855932203389829, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Results for 689_YAHOO_id_139_WebService_tr_500_1st_1073.csv: {'filename': '689_YAHOO_id_139_WebService_tr_500_1st_1073.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9927443448570203, 'Affiliation-P': 0.9855932203389829, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 689_YAHOO_id_139_WebService_tr_500_1st_1073.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/689_YAHOO_id_139_WebService_tr_500_1st_1073_results.pkl
Processing file: 352_UCR_id_50_HumanActivity_tr_48750_1st_110800.csv, Data shape: (195000, 1)
Label shape: (195000,)
Train data shape: (48750, 1), Test data shape: (146250, 1), Label test shape: (146250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 352_UCR_id_50_HumanActivity_tr_48750_1st_110800.csv...
Using CUDA device: 2
Data shape: (146250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 146250, window_size 5000, adding 3750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([30, 5000, 1])
batched_mask shape: torch.Size([30, 5000])

100%|██████████| 1/1 [00:00<00:00, 46.80it/s]
Original shapes - Output: (150000,), Label: (146250,)
Aligned shapes - Output: (146250,), Label: (146250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  14%|█▎        | 27/200 [00:01<00:06, 26.50it/s]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:02<00:05, 26.45it/s]

Calculating Affiliation Metrics:  40%|████      | 81/200 [00:03<00:04, 25.99it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [00:04<00:03, 27.81it/s]

Calculating Affiliation Metrics:  75%|███████▌  | 150/200 [00:05<00:01, 31.32it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:06<00:00, 32.75it/s]
Table1-only metrics completed in 11.51s
Evaluation Result:  {'AUC-PR': 3.610238636773891e-05, 'AUC-ROC': 0.8106106708421938, 'VUS-PR': 0.0011184114252921575, 'VUS-ROC': 0.9531992381981854, 'Standard-F1': 7.171543316114458e-05, 'Standard-Precision': 3.585900240255316e-05, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7163121504966946, 'Affiliation-P': 0.5580132926116032, 'Affiliation-R': 0.9999931623931624, 'F1_T': 7.114399544671314e-05, 'Precision_T': 3.5573263135427415e-05, 'Recall_T': 0.9999999999989999}
Results for 352_UCR_id_50_HumanActivity_tr_48750_1st_110800.csv: {'filename': '352_UCR_id_50_HumanActivity_tr_48750_1st_110800.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '48750', 'data_shape': '195000x1', 'output_length': 150000, 'label_length': 146250, 'aligned_length': 146250, 'AUC-PR': 3.610238636773891e-05, 'AUC-ROC': 0.8106106708421938, 'VUS-PR': 0.0011184114252921575, 'VUS-ROC': 0.9531992381981854, 'Standard-F1': 7.171543316114458e-05, 'Standard-Precision': 3.585900240255316e-05, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7163121504966946, 'Affiliation-P': 0.5580132926116032, 'Affiliation-R': 0.9999931623931624, 'F1_T': 7.114399544671314e-05, 'Precision_T': 3.5573263135427415e-05, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 352_UCR_id_50_HumanActivity_tr_48750_1st_110800.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/352_UCR_id_50_HumanActivity_tr_48750_1st_110800_results.pkl
Skipping file: 143_MSL_id_4_Sensor_tr_980_1st_3550.csv due to filter criteria.
Processing file: 271_IOPS_id_12_WebService_tr_500_1st_341.csv, Data shape: (8784, 1)
Label shape: (8784,)
Train data shape: (500, 1), Test data shape: (8284, 1), Label test shape: (8284,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 271_IOPS_id_12_WebService_tr_500_1st_341.csv...
Using CUDA device: 2
Data shape: (8284, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 8284, window_size 5000, adding 1716 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 155.56it/s]
Original shapes - Output: (10000,), Label: (8284,)
Aligned shapes - Output: (8284,), Label: (8284,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 334.68it/s]
Table1-only metrics completed in 6.35s
Evaluation Result:  {'AUC-PR': 0.5742490323797403, 'AUC-ROC': 0.9479077202669055, 'VUS-PR': 0.528780999940877, 'VUS-ROC': 0.9556063470292907, 'Standard-F1': 0.6478873239431577, 'Standard-Precision': 0.6899999999999932, 'Standard-Recall': 0.6106194690265433, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8437539759054737, 'Affiliation-P': 0.7300648253265778, 'Affiliation-R': 0.99938255884319, 'F1_T': 0.6075949367083567, 'Precision_T': 0.580645161290318, 'Recall_T': 0.6371681415929148}
Results for 271_IOPS_id_12_WebService_tr_500_1st_341.csv: {'filename': '271_IOPS_id_12_WebService_tr_500_1st_341.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 288, 'train_index': '500', 'data_shape': '8784x1', 'output_length': 10000, 'label_length': 8284, 'aligned_length': 8284, 'AUC-PR': 0.5742490323797403, 'AUC-ROC': 0.9479077202669055, 'VUS-PR': 0.528780999940877, 'VUS-ROC': 0.9556063470292907, 'Standard-F1': 0.6478873239431577, 'Standard-Precision': 0.6899999999999932, 'Standard-Recall': 0.6106194690265433, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8437539759054737, 'Affiliation-P': 0.7300648253265778, 'Affiliation-R': 0.99938255884319, 'F1_T': 0.6075949367083567, 'Precision_T': 0.580645161290318, 'Recall_T': 0.6371681415929148}
Logits metrics skipped for 271_IOPS_id_12_WebService_tr_500_1st_341.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/271_IOPS_id_12_WebService_tr_500_1st_341_results.pkl
Processing file: 517_UCR_id_215_Facility_tr_7482_1st_24005.csv, Data shape: (29931, 1)
Label shape: (29931,)
Train data shape: (7482, 1), Test data shape: (22449, 1), Label test shape: (22449,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 517_UCR_id_215_Facility_tr_7482_1st_24005.csv...
Using CUDA device: 2
Data shape: (22449, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22449, window_size 5000, adding 2551 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 149.91it/s]
Original shapes - Output: (25000,), Label: (22449,)
Aligned shapes - Output: (22449,), Label: (22449,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  61%|██████    | 122/200 [00:01<00:00, 121.34it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 147.47it/s]
Table1-only metrics completed in 2.25s
Evaluation Result:  {'AUC-PR': 0.37691285025017235, 'AUC-ROC': 0.9906656388255798, 'VUS-PR': 0.36872296665551296, 'VUS-ROC': 0.9945335638296947, 'Standard-F1': 0.26936026935990026, 'Standard-Precision': 0.17777777777777698, 'Standard-Recall': 0.5555555555555479, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8023923552248682, 'Affiliation-P': 0.6700870054719902, 'Affiliation-R': 0.9997973802347049, 'F1_T': 0.26936026935990026, 'Precision_T': 0.17777777777777698, 'Recall_T': 0.5555555555555479}
Results for 517_UCR_id_215_Facility_tr_7482_1st_24005.csv: {'filename': '517_UCR_id_215_Facility_tr_7482_1st_24005.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 168, 'train_index': '7482', 'data_shape': '29931x1', 'output_length': 25000, 'label_length': 22449, 'aligned_length': 22449, 'AUC-PR': 0.37691285025017235, 'AUC-ROC': 0.9906656388255798, 'VUS-PR': 0.36872296665551296, 'VUS-ROC': 0.9945335638296947, 'Standard-F1': 0.26936026935990026, 'Standard-Precision': 0.17777777777777698, 'Standard-Recall': 0.5555555555555479, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8023923552248682, 'Affiliation-P': 0.6700870054719902, 'Affiliation-R': 0.9997973802347049, 'F1_T': 0.26936026935990026, 'Precision_T': 0.17777777777777698, 'Recall_T': 0.5555555555555479}
Logits metrics skipped for 517_UCR_id_215_Facility_tr_7482_1st_24005.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/517_UCR_id_215_Facility_tr_7482_1st_24005_results.pkl
Processing file: 492_UCR_id_190_HumanActivity_tr_12500_1st_46500.csv, Data shape: (50000, 1)
Label shape: (50000,)
Train data shape: (12500, 1), Test data shape: (37500, 1), Label test shape: (37500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 492_UCR_id_190_HumanActivity_tr_12500_1st_46500.csv...
Using CUDA device: 2
Data shape: (37500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 37500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([8, 5000, 1])
batched_mask shape: torch.Size([8, 5000])

100%|██████████| 1/1 [00:00<00:00, 128.38it/s]
Original shapes - Output: (40000,), Label: (37500,)
Aligned shapes - Output: (37500,), Label: (37500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [00:01<00:01, 98.69it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 126.20it/s]
Table1-only metrics completed in 2.51s
Evaluation Result:  {'AUC-PR': 0.056737493857395876, 'AUC-ROC': 0.899727105734767, 'VUS-PR': 0.05842733670463045, 'VUS-ROC': 0.9007196056665564, 'Standard-F1': 0.15617128463438357, 'Standard-Precision': 0.10437710437710426, 'Standard-Recall': 0.30999999999999894, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7065361147797488, 'Affiliation-P': 0.5468255172413793, 'Affiliation-R': 0.9980291555555556, 'F1_T': 0.15622457282306443, 'Precision_T': 0.10333692142088255, 'Recall_T': 0.3199999999999989}
Results for 492_UCR_id_190_HumanActivity_tr_12500_1st_46500.csv: {'filename': '492_UCR_id_190_HumanActivity_tr_12500_1st_46500.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '12500', 'data_shape': '50000x1', 'output_length': 40000, 'label_length': 37500, 'aligned_length': 37500, 'AUC-PR': 0.056737493857395876, 'AUC-ROC': 0.899727105734767, 'VUS-PR': 0.05842733670463045, 'VUS-ROC': 0.9007196056665564, 'Standard-F1': 0.15617128463438357, 'Standard-Precision': 0.10437710437710426, 'Standard-Recall': 0.30999999999999894, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7065361147797488, 'Affiliation-P': 0.5468255172413793, 'Affiliation-R': 0.9980291555555556, 'F1_T': 0.15622457282306443, 'Precision_T': 0.10333692142088255, 'Recall_T': 0.3199999999999989}
Logits metrics skipped for 492_UCR_id_190_HumanActivity_tr_12500_1st_46500.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/492_UCR_id_190_HumanActivity_tr_12500_1st_46500_results.pkl
Processing file: 713_YAHOO_id_163_WebService_tr_500_1st_66.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 713_YAHOO_id_163_WebService_tr_500_1st_66.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 162.70it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1066.54it/s]
Table1-only metrics completed in 0.62s
Evaluation Result:  {'AUC-PR': 1.0000000000000002, 'AUC-ROC': 1.0, 'VUS-PR': 0.9843944684836633, 'VUS-ROC': 0.999933359734119, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9955679448098824, 'Affiliation-P': 0.9911750024957572, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Results for 713_YAHOO_id_163_WebService_tr_500_1st_66.csv: {'filename': '713_YAHOO_id_163_WebService_tr_500_1st_66.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0000000000000002, 'AUC-ROC': 1.0, 'VUS-PR': 0.9843944684836633, 'VUS-ROC': 0.999933359734119, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9955679448098824, 'Affiliation-P': 0.9911750024957572, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 713_YAHOO_id_163_WebService_tr_500_1st_66.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/713_YAHOO_id_163_WebService_tr_500_1st_66_results.pkl
Processing file: 132_WSD_id_104_WebService_tr_4037_1st_4137.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4037, 1), Test data shape: (14199, 1), Label test shape: (14199,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 132_WSD_id_104_WebService_tr_4037_1st_4137.csv...
Using CUDA device: 2
Data shape: (14199, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 14199, window_size 5000, adding 801 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.92it/s]
Original shapes - Output: (15000,), Label: (14199,)
Aligned shapes - Output: (14199,), Label: (14199,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 282.18it/s]
Table1-only metrics completed in 1.58s
Evaluation Result:  {'AUC-PR': 0.7700989038222985, 'AUC-ROC': 0.9994063326564656, 'VUS-PR': 0.698234261717381, 'VUS-ROC': 0.9992019756876164, 'Standard-F1': 0.4893617021272847, 'Standard-Precision': 0.3239436619718287, 'Standard-Recall': 0.9999999999999782, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8919854424951331, 'Affiliation-P': 0.8050350708281868, 'Affiliation-R': 0.999992848636847, 'F1_T': 0.4893617021272847, 'Precision_T': 0.3239436619718287, 'Recall_T': 0.9999999999999782}
Results for 132_WSD_id_104_WebService_tr_4037_1st_4137.csv: {'filename': '132_WSD_id_104_WebService_tr_4037_1st_4137.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4037', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 14199, 'aligned_length': 14199, 'AUC-PR': 0.7700989038222985, 'AUC-ROC': 0.9994063326564656, 'VUS-PR': 0.698234261717381, 'VUS-ROC': 0.9992019756876164, 'Standard-F1': 0.4893617021272847, 'Standard-Precision': 0.3239436619718287, 'Standard-Recall': 0.9999999999999782, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8919854424951331, 'Affiliation-P': 0.8050350708281868, 'Affiliation-R': 0.999992848636847, 'F1_T': 0.4893617021272847, 'Precision_T': 0.3239436619718287, 'Recall_T': 0.9999999999999782}
Logits metrics skipped for 132_WSD_id_104_WebService_tr_4037_1st_4137.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/132_WSD_id_104_WebService_tr_4037_1st_4137_results.pkl
Processing file: 594_YAHOO_id_44_WebService_tr_500_1st_599.csv, Data shape: (1427, 1)
Label shape: (1427,)
Train data shape: (500, 1), Test data shape: (927, 1), Label test shape: (927,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 594_YAHOO_id_44_WebService_tr_500_1st_599.csv...
Using CUDA device: 2
Data shape: (927, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 927, 1])
batched_mask shape: torch.Size([1, 927])

100%|██████████| 1/1 [00:00<00:00, 152.85it/s]
Original shapes - Output: (927,), Label: (927,)
Aligned shapes - Output: (927,), Label: (927,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1447.90it/s]
Table1-only metrics completed in 0.34s
Evaluation Result:  {'AUC-PR': 0.95, 'AUC-ROC': 0.9997291440953413, 'VUS-PR': 0.902029697801132, 'VUS-ROC': 0.9995740149728931, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9995289684408849, 'Affiliation-P': 0.9990583804143126, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 594_YAHOO_id_44_WebService_tr_500_1st_599.csv: {'filename': '594_YAHOO_id_44_WebService_tr_500_1st_599.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1427x1', 'output_length': 927, 'label_length': 927, 'aligned_length': 927, 'AUC-PR': 0.95, 'AUC-ROC': 0.9997291440953413, 'VUS-PR': 0.902029697801132, 'VUS-ROC': 0.9995740149728931, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9995289684408849, 'Affiliation-P': 0.9990583804143126, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 594_YAHOO_id_44_WebService_tr_500_1st_599.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/594_YAHOO_id_44_WebService_tr_500_1st_599_results.pkl
Processing file: 709_YAHOO_id_159_WebService_tr_500_1st_481.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 709_YAHOO_id_159_WebService_tr_500_1st_481.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 163.22it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1199.23it/s]
Table1-only metrics completed in 0.31s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.6634571532069815, 'VUS-ROC': 0.9992855607606193, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9926542605288927, 'Affiliation-P': 0.9854156538648517, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Results for 709_YAHOO_id_159_WebService_tr_500_1st_481.csv: {'filename': '709_YAHOO_id_159_WebService_tr_500_1st_481.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.6634571532069815, 'VUS-ROC': 0.9992855607606193, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9926542605288927, 'Affiliation-P': 0.9854156538648517, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Logits metrics skipped for 709_YAHOO_id_159_WebService_tr_500_1st_481.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/709_YAHOO_id_159_WebService_tr_500_1st_481_results.pkl
Processing file: 414_UCR_id_112_HumanActivity_tr_48750_1st_143411.csv, Data shape: (195000, 1)
Label shape: (195000,)
Train data shape: (48750, 1), Test data shape: (146250, 1), Label test shape: (146250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 414_UCR_id_112_HumanActivity_tr_48750_1st_143411.csv...
Using CUDA device: 2
Data shape: (146250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 146250, window_size 5000, adding 3750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([30, 5000, 1])
batched_mask shape: torch.Size([30, 5000])

100%|██████████| 1/1 [00:00<00:00, 47.16it/s]
Original shapes - Output: (150000,), Label: (146250,)
Aligned shapes - Output: (146250,), Label: (146250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  14%|█▎        | 27/200 [00:01<00:06, 26.67it/s]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:02<00:05, 26.48it/s]

Calculating Affiliation Metrics:  41%|████      | 82/200 [00:03<00:04, 27.06it/s]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [00:04<00:03, 28.58it/s]

Calculating Affiliation Metrics:  76%|███████▌  | 152/200 [00:05<00:01, 32.07it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:06<00:00, 33.33it/s]
Table1-only metrics completed in 11.27s
Evaluation Result:  {'AUC-PR': 0.7260763502118734, 'AUC-ROC': 0.9906106055422511, 'VUS-PR': 0.1753181006688788, 'VUS-ROC': 0.9896610752559324, 'Standard-F1': 0.11004478566846615, 'Standard-Precision': 0.05878332194121664, 'Standard-Recall': 0.8599999999999914, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7484778102916593, 'Affiliation-P': 0.5980564125821269, 'Affiliation-R': 0.9999931623931624, 'F1_T': 0.11004478566846615, 'Precision_T': 0.05878332194121664, 'Recall_T': 0.8599999999999914}
Results for 414_UCR_id_112_HumanActivity_tr_48750_1st_143411.csv: {'filename': '414_UCR_id_112_HumanActivity_tr_48750_1st_143411.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '48750', 'data_shape': '195000x1', 'output_length': 150000, 'label_length': 146250, 'aligned_length': 146250, 'AUC-PR': 0.7260763502118734, 'AUC-ROC': 0.9906106055422511, 'VUS-PR': 0.1753181006688788, 'VUS-ROC': 0.9896610752559324, 'Standard-F1': 0.11004478566846615, 'Standard-Precision': 0.05878332194121664, 'Standard-Recall': 0.8599999999999914, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7484778102916593, 'Affiliation-P': 0.5980564125821269, 'Affiliation-R': 0.9999931623931624, 'F1_T': 0.11004478566846615, 'Precision_T': 0.05878332194121664, 'Recall_T': 0.8599999999999914}
Logits metrics skipped for 414_UCR_id_112_HumanActivity_tr_48750_1st_143411.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/414_UCR_id_112_HumanActivity_tr_48750_1st_143411_results.pkl
Processing file: 782_YAHOO_id_232_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 782_YAHOO_id_232_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 156.83it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1579.84it/s]
Table1-only metrics completed in 0.37s
Evaluation Result:  {'AUC-PR': 0.8936387686387687, 'AUC-ROC': 0.9989035087719299, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9969368810069911, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.9997412008281573, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Results for 782_YAHOO_id_232_Synthetic_tr_500_1st_623.csv: {'filename': '782_YAHOO_id_232_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 35, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.8936387686387687, 'AUC-ROC': 0.9989035087719299, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9969368810069911, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.9997412008281573, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 782_YAHOO_id_232_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/782_YAHOO_id_232_Synthetic_tr_500_1st_623_results.pkl
Processing file: 565_YAHOO_id_15_WebService_tr_500_1st_151.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 565_YAHOO_id_15_WebService_tr_500_1st_151.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 160.48it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1666.74it/s]
Table1-only metrics completed in 0.22s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9961718417694592, 'Affiliation-P': 0.9923728813559322, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Results for 565_YAHOO_id_15_WebService_tr_500_1st_151.csv: {'filename': '565_YAHOO_id_15_WebService_tr_500_1st_151.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9961718417694592, 'Affiliation-P': 0.9923728813559322, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 565_YAHOO_id_15_WebService_tr_500_1st_151.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/565_YAHOO_id_15_WebService_tr_500_1st_151_results.pkl
Skipping file: 536_SMAP_id_6_Sensor_tr_2160_1st_5600.csv due to filter criteria.
Processing file: 370_UCR_id_68_Medical_tr_9625_1st_12240.csv, Data shape: (38501, 1)
Label shape: (38501,)
Train data shape: (9625, 1), Test data shape: (28876, 1), Label test shape: (28876,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 370_UCR_id_68_Medical_tr_9625_1st_12240.csv...
Using CUDA device: 2
Data shape: (28876, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 28876, window_size 5000, adding 1124 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 142.41it/s]
Original shapes - Output: (30000,), Label: (28876,)
Aligned shapes - Output: (28876,), Label: (28876,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [00:01<00:00, 112.85it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 125.38it/s]
Table1-only metrics completed in 1.93s
Evaluation Result:  {'AUC-PR': 0.10370616495692694, 'AUC-ROC': 0.8859237425878432, 'VUS-PR': 0.10056719927550123, 'VUS-ROC': 0.9093391131018708, 'Standard-F1': 0.20168067226859804, 'Standard-Precision': 0.12456747404844247, 'Standard-Recall': 0.5294117647058746, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7728786765405292, 'Affiliation-P': 0.6299938376088974, 'Affiliation-R': 0.9995890134693577, 'F1_T': 0.20168067226859804, 'Precision_T': 0.12456747404844247, 'Recall_T': 0.5294117647058746}
Results for 370_UCR_id_68_Medical_tr_9625_1st_12240.csv: {'filename': '370_UCR_id_68_Medical_tr_9625_1st_12240.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 40, 'train_index': '9625', 'data_shape': '38501x1', 'output_length': 30000, 'label_length': 28876, 'aligned_length': 28876, 'AUC-PR': 0.10370616495692694, 'AUC-ROC': 0.8859237425878432, 'VUS-PR': 0.10056719927550123, 'VUS-ROC': 0.9093391131018708, 'Standard-F1': 0.20168067226859804, 'Standard-Precision': 0.12456747404844247, 'Standard-Recall': 0.5294117647058746, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7728786765405292, 'Affiliation-P': 0.6299938376088974, 'Affiliation-R': 0.9995890134693577, 'F1_T': 0.20168067226859804, 'Precision_T': 0.12456747404844247, 'Recall_T': 0.5294117647058746}
Logits metrics skipped for 370_UCR_id_68_Medical_tr_9625_1st_12240.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/370_UCR_id_68_Medical_tr_9625_1st_12240_results.pkl
Processing file: 118_WSD_id_90_WebService_tr_4559_1st_6317.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 118_WSD_id_90_WebService_tr_4559_1st_6317.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 150.84it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 293.38it/s]
Table1-only metrics completed in 1.26s
Evaluation Result:  {'AUC-PR': 0.16808704302847657, 'AUC-ROC': 0.9188240456340501, 'VUS-PR': 0.007460092067466729, 'VUS-ROC': 0.9185302914828192, 'Standard-F1': 0.014285714285672144, 'Standard-Precision': 0.007299270072992648, 'Standard-Recall': 0.33333333333322224, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666579153759318, 'Affiliation-P': 0.49999015486254195, 'Affiliation-R': 1.0, 'F1_T': 0.014285714285672144, 'Precision_T': 0.007299270072992648, 'Recall_T': 0.33333333333322224}
Results for 118_WSD_id_90_WebService_tr_4559_1st_6317.csv: {'filename': '118_WSD_id_90_WebService_tr_4559_1st_6317.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.16808704302847657, 'AUC-ROC': 0.9188240456340501, 'VUS-PR': 0.007460092067466729, 'VUS-ROC': 0.9185302914828192, 'Standard-F1': 0.014285714285672144, 'Standard-Precision': 0.007299270072992648, 'Standard-Recall': 0.33333333333322224, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666579153759318, 'Affiliation-P': 0.49999015486254195, 'Affiliation-R': 1.0, 'F1_T': 0.014285714285672144, 'Precision_T': 0.007299270072992648, 'Recall_T': 0.33333333333322224}
Logits metrics skipped for 118_WSD_id_90_WebService_tr_4559_1st_6317.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/118_WSD_id_90_WebService_tr_4559_1st_6317_results.pkl
Processing file: 678_YAHOO_id_128_WebService_tr_500_1st_665.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 678_YAHOO_id_128_WebService_tr_500_1st_665.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.07it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1417.22it/s]
Table1-only metrics completed in 0.34s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9152855044441917, 'VUS-ROC': 0.9998200011486014, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9979508196721306, 'Affiliation-P': 0.9959100204498977, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Results for 678_YAHOO_id_128_WebService_tr_500_1st_665.csv: {'filename': '678_YAHOO_id_128_WebService_tr_500_1st_665.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9152855044441917, 'VUS-ROC': 0.9998200011486014, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9979508196721306, 'Affiliation-P': 0.9959100204498977, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Logits metrics skipped for 678_YAHOO_id_128_WebService_tr_500_1st_665.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/678_YAHOO_id_128_WebService_tr_500_1st_665_results.pkl
Processing file: 755_YAHOO_id_205_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 755_YAHOO_id_205_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 160.71it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2268.96it/s]
Table1-only metrics completed in 0.50s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984775989560672, 'Affiliation-P': 0.9969598262757872, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 755_YAHOO_id_205_Synthetic_tr_500_1st_893.csv: {'filename': '755_YAHOO_id_205_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984775989560672, 'Affiliation-P': 0.9969598262757872, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 755_YAHOO_id_205_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/755_YAHOO_id_205_Synthetic_tr_500_1st_893_results.pkl
Processing file: 073_WSD_id_45_WebService_tr_3392_1st_3492.csv, Data shape: (15000, 1)
Label shape: (15000,)
Train data shape: (3392, 1), Test data shape: (11608, 1), Label test shape: (11608,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 073_WSD_id_45_WebService_tr_3392_1st_3492.csv...
Using CUDA device: 2
Data shape: (11608, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 11608, window_size 5000, adding 3392 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.10it/s]
Original shapes - Output: (15000,), Label: (11608,)
Aligned shapes - Output: (11608,), Label: (11608,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 340.03it/s]
Table1-only metrics completed in 0.71s
Evaluation Result:  {'AUC-PR': 0.7810598694721715, 'AUC-ROC': 0.97677123641054, 'VUS-PR': 0.7811265252281051, 'VUS-ROC': 0.9824033149355436, 'Standard-F1': 0.7830882352936147, 'Standard-Precision': 0.7717391304347797, 'Standard-Recall': 0.7947761194029821, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9983327269597163, 'Affiliation-P': 0.9984955155908732, 'Affiliation-R': 0.9981699914000429, 'F1_T': 0.7733812949635266, 'Precision_T': 0.7465277777777751, 'Recall_T': 0.8022388059701462}
Results for 073_WSD_id_45_WebService_tr_3392_1st_3492.csv: {'filename': '073_WSD_id_45_WebService_tr_3392_1st_3492.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 10, 'train_index': '3392', 'data_shape': '15000x1', 'output_length': 15000, 'label_length': 11608, 'aligned_length': 11608, 'AUC-PR': 0.7810598694721715, 'AUC-ROC': 0.97677123641054, 'VUS-PR': 0.7811265252281051, 'VUS-ROC': 0.9824033149355436, 'Standard-F1': 0.7830882352936147, 'Standard-Precision': 0.7717391304347797, 'Standard-Recall': 0.7947761194029821, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9983327269597163, 'Affiliation-P': 0.9984955155908732, 'Affiliation-R': 0.9981699914000429, 'F1_T': 0.7733812949635266, 'Precision_T': 0.7465277777777751, 'Recall_T': 0.8022388059701462}
Logits metrics skipped for 073_WSD_id_45_WebService_tr_3392_1st_3492.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/073_WSD_id_45_WebService_tr_3392_1st_3492_results.pkl
Processing file: 282_NEK_id_6_WebService_tr_500_1st_64.csv, Data shape: (1073, 1)
Label shape: (1073,)
Train data shape: (500, 1), Test data shape: (573, 1), Label test shape: (573,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 282_NEK_id_6_WebService_tr_500_1st_64.csv...
Using CUDA device: 2
Data shape: (573, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 573, 1])
batched_mask shape: torch.Size([1, 573])

100%|██████████| 1/1 [00:00<00:00, 155.22it/s]
Original shapes - Output: (573,), Label: (573,)
Aligned shapes - Output: (573,), Label: (573,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2333.82it/s]
Table1-only metrics completed in 0.33s
Evaluation Result:  {'AUC-PR': 0.5760610265792752, 'AUC-ROC': 0.9135695764909249, 'VUS-PR': 0.6779467072491925, 'VUS-ROC': 0.9277041629324719, 'Standard-F1': 0.5833333333328206, 'Standard-Precision': 0.636363636363617, 'Standard-Recall': 0.5384615384615247, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9235863691370518, 'Affiliation-P': 0.8614437404094168, 'Affiliation-R': 0.9953917050691244, 'F1_T': 0.5753424657529113, 'Precision_T': 0.6176470588235112, 'Recall_T': 0.5384615384615247}
Results for 282_NEK_id_6_WebService_tr_500_1st_64.csv: {'filename': '282_NEK_id_6_WebService_tr_500_1st_64.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1073x1', 'output_length': 573, 'label_length': 573, 'aligned_length': 573, 'AUC-PR': 0.5760610265792752, 'AUC-ROC': 0.9135695764909249, 'VUS-PR': 0.6779467072491925, 'VUS-ROC': 0.9277041629324719, 'Standard-F1': 0.5833333333328206, 'Standard-Precision': 0.636363636363617, 'Standard-Recall': 0.5384615384615247, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9235863691370518, 'Affiliation-P': 0.8614437404094168, 'Affiliation-R': 0.9953917050691244, 'F1_T': 0.5753424657529113, 'Precision_T': 0.6176470588235112, 'Recall_T': 0.5384615384615247}
Logits metrics skipped for 282_NEK_id_6_WebService_tr_500_1st_64.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/282_NEK_id_6_WebService_tr_500_1st_64_results.pkl
Processing file: 266_IOPS_id_7_WebService_tr_4010_1st_4110.csv, Data shape: (89155, 1)
Label shape: (89155,)
Train data shape: (4010, 1), Test data shape: (85145, 1), Label test shape: (85145,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 266_IOPS_id_7_WebService_tr_4010_1st_4110.csv...
Using CUDA device: 2
Data shape: (85145, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 85145, window_size 5000, adding 4855 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([18, 5000, 1])
batched_mask shape: torch.Size([18, 5000])

100%|██████████| 1/1 [00:00<00:00, 75.68it/s]
Original shapes - Output: (90000,), Label: (85145,)
Aligned shapes - Output: (85145,), Label: (85145,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  10%|█         | 20/200 [00:01<00:09, 18.85it/s]

Calculating Affiliation Metrics:  20%|█▉        | 39/200 [00:02<00:12, 13.07it/s]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:04<00:13, 11.00it/s]

Calculating Affiliation Metrics:  33%|███▎      | 66/200 [00:05<00:12, 10.43it/s]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [00:06<00:11, 10.78it/s]

Calculating Affiliation Metrics:  46%|████▌     | 91/200 [00:07<00:09, 11.19it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 104/200 [00:08<00:08, 11.60it/s]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [00:10<00:06, 12.06it/s]

Calculating Affiliation Metrics:  66%|██████▋   | 133/200 [00:11<00:05, 12.90it/s]

Calculating Affiliation Metrics:  76%|███████▌  | 152/200 [00:12<00:03, 14.63it/s]

Calculating Affiliation Metrics:  89%|████████▉ | 178/200 [00:13<00:01, 17.80it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:13<00:00, 14.71it/s]
Table1-only metrics completed in 14.80s
Evaluation Result:  {'AUC-PR': 0.5181449703621532, 'AUC-ROC': 0.976436310212628, 'VUS-PR': 0.47216977474027644, 'VUS-ROC': 0.978873312727036, 'Standard-F1': 0.4832693362584339, 'Standard-Precision': 0.4022831050228309, 'Standard-Recall': 0.6050824175824172, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9496156227615412, 'Affiliation-P': 0.9057406083938646, 'Affiliation-R': 0.9979577301402563, 'F1_T': 0.47616376892828827, 'Precision_T': 0.4023696682464453, 'Recall_T': 0.5831043956043952}
Results for 266_IOPS_id_7_WebService_tr_4010_1st_4110.csv: {'filename': '266_IOPS_id_7_WebService_tr_4010_1st_4110.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 6, 'train_index': '4010', 'data_shape': '89155x1', 'output_length': 90000, 'label_length': 85145, 'aligned_length': 85145, 'AUC-PR': 0.5181449703621532, 'AUC-ROC': 0.976436310212628, 'VUS-PR': 0.47216977474027644, 'VUS-ROC': 0.978873312727036, 'Standard-F1': 0.4832693362584339, 'Standard-Precision': 0.4022831050228309, 'Standard-Recall': 0.6050824175824172, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9496156227615412, 'Affiliation-P': 0.9057406083938646, 'Affiliation-R': 0.9979577301402563, 'F1_T': 0.47616376892828827, 'Precision_T': 0.4023696682464453, 'Recall_T': 0.5831043956043952}
Logits metrics skipped for 266_IOPS_id_7_WebService_tr_4010_1st_4110.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/266_IOPS_id_7_WebService_tr_4010_1st_4110_results.pkl
Processing file: 080_WSD_id_52_WebService_tr_1538_1st_1638.csv, Data shape: (18307, 1)
Label shape: (18307,)
Train data shape: (1538, 1), Test data shape: (16769, 1), Label test shape: (16769,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 080_WSD_id_52_WebService_tr_1538_1st_1638.csv...
Using CUDA device: 2
Data shape: (16769, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 16769, window_size 5000, adding 3231 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 152.36it/s]
Original shapes - Output: (20000,), Label: (16769,)
Aligned shapes - Output: (16769,), Label: (16769,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [00:01<00:00, 112.77it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 136.09it/s]
Table1-only metrics completed in 1.78s
Evaluation Result:  {'AUC-PR': 0.689805003257137, 'AUC-ROC': 0.9686362932085807, 'VUS-PR': 0.6254361105529018, 'VUS-ROC': 0.978773593900975, 'Standard-F1': 0.5075757575752908, 'Standard-Precision': 0.39880952380952145, 'Standard-Recall': 0.6979166666666594, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.981822292408061, 'Affiliation-P': 0.9653344978848192, 'Affiliation-R': 0.9988830928073374, 'F1_T': 0.5075757575752908, 'Precision_T': 0.39880952380952145, 'Recall_T': 0.6979166666666594}
Results for 080_WSD_id_52_WebService_tr_1538_1st_1638.csv: {'filename': '080_WSD_id_52_WebService_tr_1538_1st_1638.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 6, 'train_index': '1538', 'data_shape': '18307x1', 'output_length': 20000, 'label_length': 16769, 'aligned_length': 16769, 'AUC-PR': 0.689805003257137, 'AUC-ROC': 0.9686362932085807, 'VUS-PR': 0.6254361105529018, 'VUS-ROC': 0.978773593900975, 'Standard-F1': 0.5075757575752908, 'Standard-Precision': 0.39880952380952145, 'Standard-Recall': 0.6979166666666594, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.981822292408061, 'Affiliation-P': 0.9653344978848192, 'Affiliation-R': 0.9988830928073374, 'F1_T': 0.5075757575752908, 'Precision_T': 0.39880952380952145, 'Recall_T': 0.6979166666666594}
Logits metrics skipped for 080_WSD_id_52_WebService_tr_1538_1st_1638.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/080_WSD_id_52_WebService_tr_1538_1st_1638_results.pkl
Processing file: 099_WSD_id_71_WebService_tr_4559_1st_14411.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 099_WSD_id_71_WebService_tr_4559_1st_14411.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 156.98it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 271.42it/s]
Table1-only metrics completed in 0.99s
Evaluation Result:  {'AUC-PR': 0.12026495998276993, 'AUC-ROC': 0.6196921348796943, 'VUS-PR': 0.03434042276891543, 'VUS-ROC': 0.7249193677293282, 'Standard-F1': 0.07865168539290299, 'Standard-Precision': 0.05109489051094854, 'Standard-Recall': 0.170731707317069, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9216967511967048, 'Affiliation-P': 0.8562746777032492, 'Affiliation-R': 0.9979427068583433, 'F1_T': 0.07865168539290299, 'Precision_T': 0.05109489051094854, 'Recall_T': 0.170731707317069}
Results for 099_WSD_id_71_WebService_tr_4559_1st_14411.csv: {'filename': '099_WSD_id_71_WebService_tr_4559_1st_14411.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.12026495998276993, 'AUC-ROC': 0.6196921348796943, 'VUS-PR': 0.03434042276891543, 'VUS-ROC': 0.7249193677293282, 'Standard-F1': 0.07865168539290299, 'Standard-Precision': 0.05109489051094854, 'Standard-Recall': 0.170731707317069, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9216967511967048, 'Affiliation-P': 0.8562746777032492, 'Affiliation-R': 0.9979427068583433, 'F1_T': 0.07865168539290299, 'Precision_T': 0.05109489051094854, 'Recall_T': 0.170731707317069}
Logits metrics skipped for 099_WSD_id_71_WebService_tr_4559_1st_14411.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/099_WSD_id_71_WebService_tr_4559_1st_14411_results.pkl
Skipping file: 814_Exathlon_id_5_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 103_WSD_id_75_WebService_tr_1375_1st_1475.csv, Data shape: (18179, 1)
Label shape: (18179,)
Train data shape: (1375, 1), Test data shape: (16804, 1), Label test shape: (16804,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 103_WSD_id_75_WebService_tr_1375_1st_1475.csv...
Using CUDA device: 2
Data shape: (16804, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 16804, window_size 5000, adding 3196 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.60it/s]
Original shapes - Output: (20000,), Label: (16804,)
Aligned shapes - Output: (16804,), Label: (16804,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  80%|████████  | 161/200 [00:01<00:00, 160.67it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 179.84it/s]
Table1-only metrics completed in 1.52s
Evaluation Result:  {'AUC-PR': 0.7753314314198082, 'AUC-ROC': 0.9533482706334285, 'VUS-PR': 0.7766992243571111, 'VUS-ROC': 0.9890958496817255, 'Standard-F1': 0.615942028985028, 'Standard-Precision': 0.5029585798816538, 'Standard-Recall': 0.7943925233644786, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8957816768031471, 'Affiliation-P': 0.8115455187707826, 'Affiliation-R': 0.9995301103625306, 'F1_T': 0.615942028985028, 'Precision_T': 0.5029585798816538, 'Recall_T': 0.7943925233644786}
Results for 103_WSD_id_75_WebService_tr_1375_1st_1475.csv: {'filename': '103_WSD_id_75_WebService_tr_1375_1st_1475.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '1375', 'data_shape': '18179x1', 'output_length': 20000, 'label_length': 16804, 'aligned_length': 16804, 'AUC-PR': 0.7753314314198082, 'AUC-ROC': 0.9533482706334285, 'VUS-PR': 0.7766992243571111, 'VUS-ROC': 0.9890958496817255, 'Standard-F1': 0.615942028985028, 'Standard-Precision': 0.5029585798816538, 'Standard-Recall': 0.7943925233644786, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8957816768031471, 'Affiliation-P': 0.8115455187707826, 'Affiliation-R': 0.9995301103625306, 'F1_T': 0.615942028985028, 'Precision_T': 0.5029585798816538, 'Recall_T': 0.7943925233644786}
Logits metrics skipped for 103_WSD_id_75_WebService_tr_1375_1st_1475.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/103_WSD_id_75_WebService_tr_1375_1st_1475_results.pkl
Processing file: 789_YAHOO_id_239_WebService_tr_500_1st_421.csv, Data shape: (1461, 1)
Label shape: (1461,)
Train data shape: (500, 1), Test data shape: (961, 1), Label test shape: (961,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 789_YAHOO_id_239_WebService_tr_500_1st_421.csv...
Using CUDA device: 2
Data shape: (961, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 961, 1])
batched_mask shape: torch.Size([1, 961])

100%|██████████| 1/1 [00:00<00:00, 159.95it/s]
Original shapes - Output: (961,), Label: (961,)
Aligned shapes - Output: (961,), Label: (961,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Table1-only metrics completed in 0.01s
Evaluation Result:  {'AUC-PR': 0.0, 'AUC-ROC': nan, 'VUS-PR': 0.0, 'VUS-ROC': 0.0, 'Standard-F1': 0.0, 'Standard-Precision': 0.0, 'Standard-Recall': 0.0, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.0, 'Affiliation-P': 0.0, 'Affiliation-R': 0.0, 'F1_T': 0.0, 'Precision_T': 0.0, 'Recall_T': 0.0}
Results for 789_YAHOO_id_239_WebService_tr_500_1st_421.csv: {'filename': '789_YAHOO_id_239_WebService_tr_500_1st_421.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1461x1', 'output_length': 961, 'label_length': 961, 'aligned_length': 961, 'AUC-PR': 0.0, 'AUC-ROC': nan, 'VUS-PR': 0.0, 'VUS-ROC': 0.0, 'Standard-F1': 0.0, 'Standard-Precision': 0.0, 'Standard-Recall': 0.0, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.0, 'Affiliation-P': 0.0, 'Affiliation-R': 0.0, 'F1_T': 0.0, 'Precision_T': 0.0, 'Recall_T': 0.0}
Logits metrics skipped for 789_YAHOO_id_239_WebService_tr_500_1st_421.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/789_YAHOO_id_239_WebService_tr_500_1st_421_results.pkl
Skipping file: 845_OPPORTUNITY_id_4_HumanActivity_tr_7295_1st_27516.csv due to filter criteria.
Processing file: 718_YAHOO_id_168_WebService_tr_500_1st_263.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 718_YAHOO_id_168_WebService_tr_500_1st_263.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 166.24it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1010.36it/s]
Table1-only metrics completed in 0.69s
Evaluation Result:  {'AUC-PR': 0.8923976608187134, 'AUC-ROC': 0.9738115570737262, 'VUS-PR': 0.7108600837679383, 'VUS-ROC': 0.9628809350147535, 'Standard-F1': 0.7619047619041995, 'Standard-Precision': 0.6666666666666111, 'Standard-Recall': 0.8888888888887901, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9077422034714095, 'Affiliation-P': 0.8310695573484514, 'Affiliation-R': 1.0, 'F1_T': 0.7619047619041995, 'Precision_T': 0.6666666666666111, 'Recall_T': 0.8888888888887901}
Results for 718_YAHOO_id_168_WebService_tr_500_1st_263.csv: {'filename': '718_YAHOO_id_168_WebService_tr_500_1st_263.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.8923976608187134, 'AUC-ROC': 0.9738115570737262, 'VUS-PR': 0.7108600837679383, 'VUS-ROC': 0.9628809350147535, 'Standard-F1': 0.7619047619041995, 'Standard-Precision': 0.6666666666666111, 'Standard-Recall': 0.8888888888887901, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9077422034714095, 'Affiliation-P': 0.8310695573484514, 'Affiliation-R': 1.0, 'F1_T': 0.7619047619041995, 'Precision_T': 0.6666666666666111, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 718_YAHOO_id_168_WebService_tr_500_1st_263.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/718_YAHOO_id_168_WebService_tr_500_1st_263_results.pkl
Skipping file: 185_SMD_id_8_Facility_tr_5000_1st_15144.csv due to filter criteria.
Processing file: 017_NAB_id_17_Synthetic_tr_1007_1st_1805.csv, Data shape: (4031, 1)
Label shape: (4031,)
Train data shape: (1007, 1), Test data shape: (3024, 1), Label test shape: (3024,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 017_NAB_id_17_Synthetic_tr_1007_1st_1805.csv...
Using CUDA device: 2
Data shape: (3024, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 3024, 1])
batched_mask shape: torch.Size([1, 3024])

100%|██████████| 1/1 [00:00<00:00, 164.29it/s]
Original shapes - Output: (3024,), Label: (3024,)
Aligned shapes - Output: (3024,), Label: (3024,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 553.99it/s]
Table1-only metrics completed in 0.75s
Evaluation Result:  {'AUC-PR': 0.22825503826944454, 'AUC-ROC': 0.5019553263431128, 'VUS-PR': 0.23209443313343117, 'VUS-ROC': 0.5245308568563155, 'Standard-F1': 0.24285275130626957, 'Standard-Precision': 0.13854787793756573, 'Standard-Recall': 0.9825870646766144, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.968389287829748, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9387158124720314, 'F1_T': 0.2422569763873948, 'Precision_T': 0.13816019587268272, 'Recall_T': 0.9825870646766144}
Results for 017_NAB_id_17_Synthetic_tr_1007_1st_1805.csv: {'filename': '017_NAB_id_17_Synthetic_tr_1007_1st_1805.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '1007', 'data_shape': '4031x1', 'output_length': 3024, 'label_length': 3024, 'aligned_length': 3024, 'AUC-PR': 0.22825503826944454, 'AUC-ROC': 0.5019553263431128, 'VUS-PR': 0.23209443313343117, 'VUS-ROC': 0.5245308568563155, 'Standard-F1': 0.24285275130626957, 'Standard-Precision': 0.13854787793756573, 'Standard-Recall': 0.9825870646766144, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.968389287829748, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9387158124720314, 'F1_T': 0.2422569763873948, 'Precision_T': 0.13816019587268272, 'Recall_T': 0.9825870646766144}
Logits metrics skipped for 017_NAB_id_17_Synthetic_tr_1007_1st_1805.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/017_NAB_id_17_Synthetic_tr_1007_1st_1805_results.pkl
Processing file: 366_UCR_id_64_Medical_tr_2507_1st_8500.csv, Data shape: (10028, 1)
Label shape: (10028,)
Train data shape: (2507, 1), Test data shape: (7521, 1), Label test shape: (7521,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 366_UCR_id_64_Medical_tr_2507_1st_8500.csv...
Using CUDA device: 2
Data shape: (7521, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 7521, window_size 5000, adding 2479 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.27it/s]
Original shapes - Output: (10000,), Label: (7521,)
Aligned shapes - Output: (7521,), Label: (7521,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 511.72it/s]
Table1-only metrics completed in 0.53s
Evaluation Result:  {'AUC-PR': 0.0010582010582010583, 'AUC-ROC': 0.874468085106383, 'VUS-PR': 0.08165083452863382, 'VUS-ROC': 0.9392946053052426, 'Standard-F1': 0.002109704641348099, 'Standard-Precision': 0.0010559662090813082, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7707041798153775, 'Affiliation-P': 0.6269476940868715, 'Affiliation-R': 1.0, 'F1_T': 0.002072538860101552, 'Precision_T': 0.0010373443983402479, 'Recall_T': 0.9999999999989999}
Results for 366_UCR_id_64_Medical_tr_2507_1st_8500.csv: {'filename': '366_UCR_id_64_Medical_tr_2507_1st_8500.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 26, 'train_index': '2507', 'data_shape': '10028x1', 'output_length': 10000, 'label_length': 7521, 'aligned_length': 7521, 'AUC-PR': 0.0010582010582010583, 'AUC-ROC': 0.874468085106383, 'VUS-PR': 0.08165083452863382, 'VUS-ROC': 0.9392946053052426, 'Standard-F1': 0.002109704641348099, 'Standard-Precision': 0.0010559662090813082, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7707041798153775, 'Affiliation-P': 0.6269476940868715, 'Affiliation-R': 1.0, 'F1_T': 0.002072538860101552, 'Precision_T': 0.0010373443983402479, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 366_UCR_id_64_Medical_tr_2507_1st_8500.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/366_UCR_id_64_Medical_tr_2507_1st_8500_results.pkl
Skipping file: 844_OPPORTUNITY_id_3_HumanActivity_tr_500_1st_566.csv due to filter criteria.
Skipping file: 543_SMAP_id_13_Sensor_tr_2020_1st_4550.csv due to filter criteria.
Processing file: 802_YAHOO_id_252_WebService_tr_500_1st_63.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 802_YAHOO_id_252_WebService_tr_500_1st_63.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 101.77it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1359.36it/s]
Table1-only metrics completed in 0.35s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9504941880689357, 'VUS-ROC': 0.9998948142040058, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990591135682693, 'Affiliation-P': 0.9981199960067885, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Results for 802_YAHOO_id_252_WebService_tr_500_1st_63.csv: {'filename': '802_YAHOO_id_252_WebService_tr_500_1st_63.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9504941880689357, 'VUS-ROC': 0.9998948142040058, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990591135682693, 'Affiliation-P': 0.9981199960067885, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Logits metrics skipped for 802_YAHOO_id_252_WebService_tr_500_1st_63.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/802_YAHOO_id_252_WebService_tr_500_1st_63_results.pkl
Processing file: 514_UCR_id_212_Medical_tr_2499_1st_6508.csv, Data shape: (9998, 1)
Label shape: (9998,)
Train data shape: (2499, 1), Test data shape: (7499, 1), Label test shape: (7499,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 514_UCR_id_212_Medical_tr_2499_1st_6508.csv...
Using CUDA device: 2
Data shape: (7499, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 7499, window_size 5000, adding 2501 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 166.16it/s]
Original shapes - Output: (10000,), Label: (7499,)
Aligned shapes - Output: (7499,), Label: (7499,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 470.51it/s]
Table1-only metrics completed in 0.56s
Evaluation Result:  {'AUC-PR': 0.7574409372754307, 'AUC-ROC': 0.9961551886159216, 'VUS-PR': 0.6887289518207093, 'VUS-ROC': 0.9962367640309405, 'Standard-F1': 0.559999999999511, 'Standard-Precision': 0.46666666666666046, 'Standard-Recall': 0.699999999999986, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9949151407676505, 'Affiliation-P': 0.9904232845081098, 'Affiliation-R': 0.9994479263901853, 'F1_T': 0.559999999999511, 'Precision_T': 0.46666666666666046, 'Recall_T': 0.699999999999986}
Results for 514_UCR_id_212_Medical_tr_2499_1st_6508.csv: {'filename': '514_UCR_id_212_Medical_tr_2499_1st_6508.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 26, 'train_index': '2499', 'data_shape': '9998x1', 'output_length': 10000, 'label_length': 7499, 'aligned_length': 7499, 'AUC-PR': 0.7574409372754307, 'AUC-ROC': 0.9961551886159216, 'VUS-PR': 0.6887289518207093, 'VUS-ROC': 0.9962367640309405, 'Standard-F1': 0.559999999999511, 'Standard-Precision': 0.46666666666666046, 'Standard-Recall': 0.699999999999986, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9949151407676505, 'Affiliation-P': 0.9904232845081098, 'Affiliation-R': 0.9994479263901853, 'F1_T': 0.559999999999511, 'Precision_T': 0.46666666666666046, 'Recall_T': 0.699999999999986}
Logits metrics skipped for 514_UCR_id_212_Medical_tr_2499_1st_6508.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/514_UCR_id_212_Medical_tr_2499_1st_6508_results.pkl
Processing file: 413_UCR_id_111_Facility_tr_12500_1st_46600.csv, Data shape: (50000, 1)
Label shape: (50000,)
Train data shape: (12500, 1), Test data shape: (37500, 1), Label test shape: (37500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 413_UCR_id_111_Facility_tr_12500_1st_46600.csv...
Using CUDA device: 2
Data shape: (37500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 37500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([8, 5000, 1])
batched_mask shape: torch.Size([8, 5000])

100%|██████████| 1/1 [00:00<00:00, 129.01it/s]
Original shapes - Output: (40000,), Label: (37500,)
Aligned shapes - Output: (37500,), Label: (37500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  52%|█████▎    | 105/200 [00:01<00:00, 104.56it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 121.21it/s]
Table1-only metrics completed in 3.11s
Evaluation Result:  {'AUC-PR': 0.9713849640473929, 'AUC-ROC': 0.9984443548387096, 'VUS-PR': 0.9412279006232471, 'VUS-ROC': 0.9987297371005709, 'Standard-F1': 0.8681481481476516, 'Standard-Precision': 0.7813333333333312, 'Standard-Recall': 0.9766666666666634, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999710880530119, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9999421777777778, 'F1_T': 0.8681481481476516, 'Precision_T': 0.7813333333333312, 'Recall_T': 0.9766666666666634}
Results for 413_UCR_id_111_Facility_tr_12500_1st_46600.csv: {'filename': '413_UCR_id_111_Facility_tr_12500_1st_46600.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 208, 'train_index': '12500', 'data_shape': '50000x1', 'output_length': 40000, 'label_length': 37500, 'aligned_length': 37500, 'AUC-PR': 0.9713849640473929, 'AUC-ROC': 0.9984443548387096, 'VUS-PR': 0.9412279006232471, 'VUS-ROC': 0.9987297371005709, 'Standard-F1': 0.8681481481476516, 'Standard-Precision': 0.7813333333333312, 'Standard-Recall': 0.9766666666666634, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999710880530119, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9999421777777778, 'F1_T': 0.8681481481476516, 'Precision_T': 0.7813333333333312, 'Recall_T': 0.9766666666666634}
Logits metrics skipped for 413_UCR_id_111_Facility_tr_12500_1st_46600.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/413_UCR_id_111_Facility_tr_12500_1st_46600_results.pkl
Skipping file: 200_SMD_id_23_Facility_tr_7174_1st_21230.csv due to filter criteria.
Processing file: 292_TODS_id_6_Synthetic_tr_500_1st_11.csv, Data shape: (5000, 1)
Label shape: (5000,)
Train data shape: (500, 1), Test data shape: (4500, 1), Label test shape: (4500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 292_TODS_id_6_Synthetic_tr_500_1st_11.csv...
Using CUDA device: 2
Data shape: (4500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 4500, 1])
batched_mask shape: torch.Size([1, 4500])

100%|██████████| 1/1 [00:00<00:00, 158.11it/s]
Original shapes - Output: (4500,), Label: (4500,)
Aligned shapes - Output: (4500,), Label: (4500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  12%|█▏        | 24/200 [00:01<00:07, 22.53it/s]

Calculating Affiliation Metrics:  24%|██▎       | 47/200 [00:03<00:10, 14.04it/s]

Calculating Affiliation Metrics:  32%|███▏      | 63/200 [00:04<00:11, 11.77it/s]

Calculating Affiliation Metrics:  38%|███▊      | 76/200 [00:06<00:11, 10.81it/s]

Calculating Affiliation Metrics:  44%|████▍     | 88/200 [00:07<00:10, 10.41it/s]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [00:08<00:09, 10.16it/s]

Calculating Affiliation Metrics:  55%|█████▌    | 110/200 [00:09<00:09,  9.93it/s]

Calculating Affiliation Metrics:  60%|██████    | 120/200 [00:10<00:08,  9.87it/s]

Calculating Affiliation Metrics:  65%|██████▌   | 130/200 [00:11<00:07,  9.86it/s]

Calculating Affiliation Metrics:  70%|███████   | 141/200 [00:12<00:05, 10.09it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 155/200 [00:14<00:04, 10.99it/s]

Calculating Affiliation Metrics:  87%|████████▋ | 174/200 [00:15<00:01, 13.25it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:15<00:00, 12.70it/s]
Table1-only metrics completed in 23.64s
Evaluation Result:  {'AUC-PR': 0.5285981579569436, 'AUC-ROC': 0.7867431593125638, 'VUS-PR': 0.9515726275858002, 'VUS-ROC': 0.9774856590825503, 'Standard-F1': 0.552706552706107, 'Standard-Precision': 0.836206896551717, 'Standard-Recall': 0.4127659574468068, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7221022323978811, 'Affiliation-P': 0.671060733554793, 'Affiliation-R': 0.7815474531202131, 'F1_T': 0.5475504322762167, 'Precision_T': 0.8482142857142781, 'Recall_T': 0.4042553191489345}
Results for 292_TODS_id_6_Synthetic_tr_500_1st_11.csv: {'filename': '292_TODS_id_6_Synthetic_tr_500_1st_11.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '500', 'data_shape': '5000x1', 'output_length': 4500, 'label_length': 4500, 'aligned_length': 4500, 'AUC-PR': 0.5285981579569436, 'AUC-ROC': 0.7867431593125638, 'VUS-PR': 0.9515726275858002, 'VUS-ROC': 0.9774856590825503, 'Standard-F1': 0.552706552706107, 'Standard-Precision': 0.836206896551717, 'Standard-Recall': 0.4127659574468068, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7221022323978811, 'Affiliation-P': 0.671060733554793, 'Affiliation-R': 0.7815474531202131, 'F1_T': 0.5475504322762167, 'Precision_T': 0.8482142857142781, 'Recall_T': 0.4042553191489345}
Logits metrics skipped for 292_TODS_id_6_Synthetic_tr_500_1st_11.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/292_TODS_id_6_Synthetic_tr_500_1st_11_results.pkl
Processing file: 368_UCR_id_66_Facility_tr_50000_1st_125720.csv, Data shape: (299690, 1)
Label shape: (299690,)
Train data shape: (50000, 1), Test data shape: (249690, 1), Label test shape: (249690,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 368_UCR_id_66_Facility_tr_50000_1st_125720.csv...
Using CUDA device: 2
Data shape: (249690, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 249690, window_size 5000, adding 310 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([50, 5000, 1])
batched_mask shape: torch.Size([50, 5000])

100%|██████████| 1/1 [00:00<00:00, 28.02it/s]
Original shapes - Output: (250000,), Label: (249690,)
Aligned shapes - Output: (249690,), Label: (249690,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   9%|▉         | 18/200 [00:01<00:10, 16.87it/s]

Calculating Affiliation Metrics:  18%|█▊        | 35/200 [00:02<00:11, 14.40it/s]

Calculating Affiliation Metrics:  25%|██▌       | 50/200 [00:03<00:11, 13.09it/s]

Calculating Affiliation Metrics:  32%|███▏      | 64/200 [00:04<00:11, 12.23it/s]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [00:06<00:10, 11.41it/s]

Calculating Affiliation Metrics:  44%|████▍     | 89/200 [00:07<00:09, 11.20it/s]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [00:08<00:08, 11.20it/s]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [00:09<00:07, 11.20it/s]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [00:10<00:06, 11.62it/s]

Calculating Affiliation Metrics:  70%|███████   | 140/200 [00:11<00:04, 12.28it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 156/200 [00:12<00:03, 13.27it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 176/200 [00:13<00:01, 15.18it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:14<00:00, 13.93it/s]
Table1-only metrics completed in 22.92s
Evaluation Result:  {'AUC-PR': 0.36226345866545484, 'AUC-ROC': 0.9935508599174676, 'VUS-PR': 0.3713474259293979, 'VUS-ROC': 0.9949681757644833, 'Standard-F1': 0.3209405783282389, 'Standard-Precision': 0.2022426912294753, 'Standard-Recall': 0.7769230769230757, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8315095523945498, 'Affiliation-P': 0.7116319120171164, 'Affiliation-R': 0.9999568418685324, 'F1_T': 0.3209405783282389, 'Precision_T': 0.2022426912294753, 'Recall_T': 0.7769230769230757}
Results for 368_UCR_id_66_Facility_tr_50000_1st_125720.csv: {'filename': '368_UCR_id_66_Facility_tr_50000_1st_125720.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '50000', 'data_shape': '299690x1', 'output_length': 250000, 'label_length': 249690, 'aligned_length': 249690, 'AUC-PR': 0.36226345866545484, 'AUC-ROC': 0.9935508599174676, 'VUS-PR': 0.3713474259293979, 'VUS-ROC': 0.9949681757644833, 'Standard-F1': 0.3209405783282389, 'Standard-Precision': 0.2022426912294753, 'Standard-Recall': 0.7769230769230757, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8315095523945498, 'Affiliation-P': 0.7116319120171164, 'Affiliation-R': 0.9999568418685324, 'F1_T': 0.3209405783282389, 'Precision_T': 0.2022426912294753, 'Recall_T': 0.7769230769230757}
Logits metrics skipped for 368_UCR_id_66_Facility_tr_50000_1st_125720.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/368_UCR_id_66_Facility_tr_50000_1st_125720_results.pkl
Skipping file: 548_SMAP_id_18_Sensor_tr_2094_1st_5600.csv due to filter criteria.
Skipping file: 840_Exathlon_id_31_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 063_WSD_id_35_WebService_tr_2290_1st_2390.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (2290, 1), Test data shape: (15946, 1), Label test shape: (15946,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 063_WSD_id_35_WebService_tr_2290_1st_2390.csv...
Using CUDA device: 2
Data shape: (15946, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 15946, window_size 5000, adding 4054 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.16it/s]
Original shapes - Output: (20000,), Label: (15946,)
Aligned shapes - Output: (15946,), Label: (15946,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 256.70it/s]
Table1-only metrics completed in 1.95s
Evaluation Result:  {'AUC-PR': 0.04822467140570825, 'AUC-ROC': 0.45414811451903525, 'VUS-PR': 0.013532979216232088, 'VUS-ROC': 0.6277567379796151, 'Standard-F1': 0.026431718061257697, 'Standard-Precision': 0.018749999999999885, 'Standard-Recall': 0.04477611940298441, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7729505768549162, 'Affiliation-P': 0.6323318649353337, 'Affiliation-R': 0.9939967575971177, 'F1_T': 0.026431718061257697, 'Precision_T': 0.018749999999999885, 'Recall_T': 0.04477611940298441}
Results for 063_WSD_id_35_WebService_tr_2290_1st_2390.csv: {'filename': '063_WSD_id_35_WebService_tr_2290_1st_2390.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '2290', 'data_shape': '18236x1', 'output_length': 20000, 'label_length': 15946, 'aligned_length': 15946, 'AUC-PR': 0.04822467140570825, 'AUC-ROC': 0.45414811451903525, 'VUS-PR': 0.013532979216232088, 'VUS-ROC': 0.6277567379796151, 'Standard-F1': 0.026431718061257697, 'Standard-Precision': 0.018749999999999885, 'Standard-Recall': 0.04477611940298441, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7729505768549162, 'Affiliation-P': 0.6323318649353337, 'Affiliation-R': 0.9939967575971177, 'F1_T': 0.026431718061257697, 'Precision_T': 0.018749999999999885, 'Recall_T': 0.04477611940298441}
Logits metrics skipped for 063_WSD_id_35_WebService_tr_2290_1st_2390.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/063_WSD_id_35_WebService_tr_2290_1st_2390_results.pkl
Processing file: 672_YAHOO_id_122_WebService_tr_500_1st_857.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 672_YAHOO_id_122_WebService_tr_500_1st_857.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 158.85it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1479.47it/s]
Table1-only metrics completed in 0.23s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9978768577494687, 'Affiliation-P': 0.9957627118644067, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Results for 672_YAHOO_id_122_WebService_tr_500_1st_857.csv: {'filename': '672_YAHOO_id_122_WebService_tr_500_1st_857.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9978768577494687, 'Affiliation-P': 0.9957627118644067, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 672_YAHOO_id_122_WebService_tr_500_1st_857.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/672_YAHOO_id_122_WebService_tr_500_1st_857_results.pkl
Processing file: 503_UCR_id_201_Facility_tr_50000_1st_593450.csv, Data shape: (897667, 1)
Label shape: (897667,)
Train data shape: (50000, 1), Test data shape: (847667, 1), Label test shape: (847667,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 503_UCR_id_201_Facility_tr_50000_1st_593450.csv...
Using CUDA device: 2
Data shape: (847667, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 847667, window_size 5000, adding 2333 samples


  0%|          | 0/3 [00:00<?, ?it/s]batched_time_series shape: torch.Size([64, 5000, 1])
batched_mask shape: torch.Size([64, 5000])
batched_time_series shape: torch.Size([64, 5000, 1])
batched_mask shape: torch.Size([64, 5000])
batched_time_series shape: torch.Size([42, 5000, 1])
batched_mask shape: torch.Size([42, 5000])

100%|██████████| 3/3 [00:00<00:00, 26.09it/s]
Original shapes - Output: (850000,), Label: (847667,)
Aligned shapes - Output: (847667,), Label: (847667,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   4%|▎         | 7/200 [00:01<00:30,  6.26it/s]

Calculating Affiliation Metrics:   7%|▋         | 14/200 [00:02<00:31,  5.94it/s]

Calculating Affiliation Metrics:  10%|█         | 20/200 [00:03<00:31,  5.66it/s]

Calculating Affiliation Metrics:  13%|█▎        | 26/200 [00:04<00:32,  5.44it/s]

Calculating Affiliation Metrics:  16%|█▌        | 32/200 [00:05<00:32,  5.24it/s]

Calculating Affiliation Metrics:  19%|█▉        | 38/200 [00:07<00:31,  5.09it/s]

Calculating Affiliation Metrics:  22%|██▏       | 44/200 [00:08<00:31,  4.91it/s]

Calculating Affiliation Metrics:  24%|██▍       | 49/200 [00:09<00:31,  4.87it/s]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:10<00:30,  4.83it/s]

Calculating Affiliation Metrics:  30%|██▉       | 59/200 [00:11<00:29,  4.81it/s]

Calculating Affiliation Metrics:  32%|███▏      | 64/200 [00:12<00:28,  4.80it/s]

Calculating Affiliation Metrics:  34%|███▍      | 69/200 [00:13<00:27,  4.68it/s]

Calculating Affiliation Metrics:  37%|███▋      | 74/200 [00:14<00:26,  4.71it/s]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [00:15<00:25,  4.75it/s]

Calculating Affiliation Metrics:  42%|████▏     | 84/200 [00:16<00:24,  4.71it/s]

Calculating Affiliation Metrics:  44%|████▍     | 89/200 [00:17<00:23,  4.78it/s]

Calculating Affiliation Metrics:  48%|████▊     | 95/200 [00:19<00:21,  4.90it/s]

Calculating Affiliation Metrics:  50%|█████     | 100/200 [00:20<00:20,  4.92it/s]

Calculating Affiliation Metrics:  53%|█████▎    | 106/200 [00:21<00:18,  5.04it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [00:22<00:17,  5.13it/s]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [00:23<00:15,  5.15it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 124/200 [00:24<00:14,  5.29it/s]

Calculating Affiliation Metrics:  65%|██████▌   | 130/200 [00:25<00:12,  5.45it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 137/200 [00:26<00:11,  5.68it/s]

Calculating Affiliation Metrics:  72%|███████▏  | 144/200 [00:27<00:09,  5.92it/s]

Calculating Affiliation Metrics:  76%|███████▌  | 151/200 [00:28<00:07,  6.19it/s]

Calculating Affiliation Metrics:  80%|███████▉  | 159/200 [00:29<00:06,  6.52it/s]

Calculating Affiliation Metrics:  84%|████████▎ | 167/200 [00:30<00:04,  6.89it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 176/200 [00:32<00:03,  7.32it/s]

Calculating Affiliation Metrics:  93%|█████████▎| 186/200 [00:33<00:01,  8.08it/s]

Calculating Affiliation Metrics:  99%|█████████▉| 198/200 [00:34<00:00,  9.09it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:34<00:00,  5.84it/s]
Table1-only metrics completed in 61.34s
Evaluation Result:  {'AUC-PR': 0.0006414191791753829, 'AUC-ROC': 0.9378676359982209, 'VUS-PR': 0.0006797085184834382, 'VUS-ROC': 0.9387535222252814, 'Standard-F1': 0.001936468923393003, 'Standard-Precision': 0.00097054761282616, 'Standard-Recall': 0.40624999999999367, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6859152996972429, 'Affiliation-P': 0.5219754849071359, 'Affiliation-R': 0.999986875742479, 'F1_T': 0.0018225761717925563, 'Precision_T': 0.000913604766633565, 'Recall_T': 0.3593749999999944}
Results for 503_UCR_id_201_Facility_tr_50000_1st_593450.csv: {'filename': '503_UCR_id_201_Facility_tr_50000_1st_593450.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '50000', 'data_shape': '897667x1', 'output_length': 850000, 'label_length': 847667, 'aligned_length': 847667, 'AUC-PR': 0.0006414191791753829, 'AUC-ROC': 0.9378676359982209, 'VUS-PR': 0.0006797085184834382, 'VUS-ROC': 0.9387535222252814, 'Standard-F1': 0.001936468923393003, 'Standard-Precision': 0.00097054761282616, 'Standard-Recall': 0.40624999999999367, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6859152996972429, 'Affiliation-P': 0.5219754849071359, 'Affiliation-R': 0.999986875742479, 'F1_T': 0.0018225761717925563, 'Precision_T': 0.000913604766633565, 'Recall_T': 0.3593749999999944}
Logits metrics skipped for 503_UCR_id_201_Facility_tr_50000_1st_593450.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/503_UCR_id_201_Facility_tr_50000_1st_593450_results.pkl
Processing file: 152_Stock_id_4_Finance_tr_500_1st_2.csv, Data shape: (20000, 1)
Label shape: (20000,)
Train data shape: (500, 1), Test data shape: (19500, 1), Label test shape: (19500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 152_Stock_id_4_Finance_tr_500_1st_2.csv...
Using CUDA device: 2
Data shape: (19500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 19500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.03it/s]
Original shapes - Output: (20000,), Label: (19500,)
Aligned shapes - Output: (19500,), Label: (19500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   4%|▎         | 7/200 [00:01<00:32,  5.88it/s]

Calculating Affiliation Metrics:   6%|▋         | 13/200 [00:03<00:53,  3.51it/s]

Calculating Affiliation Metrics:   8%|▊         | 17/200 [00:05<01:11,  2.56it/s]

Calculating Affiliation Metrics:  10%|█         | 20/200 [00:08<01:27,  2.06it/s]

Calculating Affiliation Metrics:  12%|█▏        | 23/200 [00:10<01:44,  1.70it/s]

Calculating Affiliation Metrics:  12%|█▎        | 25/200 [00:12<01:55,  1.51it/s]

Calculating Affiliation Metrics:  14%|█▎        | 27/200 [00:14<02:06,  1.37it/s]

Calculating Affiliation Metrics:  14%|█▍        | 29/200 [00:16<02:15,  1.26it/s]

Calculating Affiliation Metrics:  16%|█▌        | 31/200 [00:18<02:23,  1.18it/s]

Calculating Affiliation Metrics:  16%|█▋        | 33/200 [00:20<02:29,  1.11it/s]

Calculating Affiliation Metrics:  18%|█▊        | 35/200 [00:22<02:35,  1.06it/s]

Calculating Affiliation Metrics:  18%|█▊        | 37/200 [00:24<02:41,  1.01it/s]

Calculating Affiliation Metrics:  20%|█▉        | 39/200 [00:27<02:47,  1.04s/it]

Calculating Affiliation Metrics:  20%|██        | 40/200 [00:28<02:51,  1.07s/it]

Calculating Affiliation Metrics:  20%|██        | 41/200 [00:29<02:54,  1.09s/it]

Calculating Affiliation Metrics:  21%|██        | 42/200 [00:30<02:56,  1.12s/it]

Calculating Affiliation Metrics:  22%|██▏       | 43/200 [00:31<02:59,  1.14s/it]

Calculating Affiliation Metrics:  22%|██▏       | 44/200 [00:33<03:01,  1.17s/it]

Calculating Affiliation Metrics:  22%|██▎       | 45/200 [00:34<03:04,  1.19s/it]

Calculating Affiliation Metrics:  23%|██▎       | 46/200 [00:35<03:07,  1.22s/it]

Calculating Affiliation Metrics:  24%|██▎       | 47/200 [00:37<03:10,  1.25s/it]

Calculating Affiliation Metrics:  24%|██▍       | 48/200 [00:38<03:12,  1.27s/it]

Calculating Affiliation Metrics:  24%|██▍       | 49/200 [00:39<03:13,  1.28s/it]

Calculating Affiliation Metrics:  25%|██▌       | 50/200 [00:41<03:14,  1.29s/it]

Calculating Affiliation Metrics:  26%|██▌       | 51/200 [00:42<03:12,  1.29s/it]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:43<03:10,  1.29s/it]

Calculating Affiliation Metrics:  26%|██▋       | 53/200 [00:44<03:07,  1.28s/it]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:46<03:03,  1.25s/it]

Calculating Affiliation Metrics:  28%|██▊       | 55/200 [00:47<03:00,  1.24s/it]

Calculating Affiliation Metrics:  28%|██▊       | 56/200 [00:48<02:58,  1.24s/it]

Calculating Affiliation Metrics:  28%|██▊       | 57/200 [00:49<02:58,  1.25s/it]

Calculating Affiliation Metrics:  29%|██▉       | 58/200 [00:50<02:56,  1.25s/it]

Calculating Affiliation Metrics:  30%|██▉       | 59/200 [00:52<02:56,  1.26s/it]

Calculating Affiliation Metrics:  30%|███       | 60/200 [00:53<02:57,  1.26s/it]

Calculating Affiliation Metrics:  30%|███       | 61/200 [00:54<02:57,  1.28s/it]

Calculating Affiliation Metrics:  31%|███       | 62/200 [00:56<02:58,  1.29s/it]

Calculating Affiliation Metrics:  32%|███▏      | 63/200 [00:57<02:58,  1.30s/it]

Calculating Affiliation Metrics:  32%|███▏      | 64/200 [00:58<02:58,  1.31s/it]

Calculating Affiliation Metrics:  32%|███▎      | 65/200 [01:00<02:58,  1.32s/it]

Calculating Affiliation Metrics:  33%|███▎      | 66/200 [01:01<02:58,  1.33s/it]

Calculating Affiliation Metrics:  34%|███▎      | 67/200 [01:02<02:58,  1.34s/it]

Calculating Affiliation Metrics:  34%|███▍      | 68/200 [01:04<02:59,  1.36s/it]

Calculating Affiliation Metrics:  34%|███▍      | 69/200 [01:05<02:59,  1.37s/it]

Calculating Affiliation Metrics:  35%|███▌      | 70/200 [01:07<02:59,  1.38s/it]

Calculating Affiliation Metrics:  36%|███▌      | 71/200 [01:08<02:58,  1.39s/it]

Calculating Affiliation Metrics:  36%|███▌      | 72/200 [01:09<02:56,  1.38s/it]

Calculating Affiliation Metrics:  36%|███▋      | 73/200 [01:11<02:54,  1.37s/it]

Calculating Affiliation Metrics:  37%|███▋      | 74/200 [01:12<02:51,  1.36s/it]

Calculating Affiliation Metrics:  38%|███▊      | 75/200 [01:13<02:48,  1.35s/it]

Calculating Affiliation Metrics:  38%|███▊      | 76/200 [01:15<02:47,  1.35s/it]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [01:16<02:45,  1.35s/it]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [01:18<02:47,  1.37s/it]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [01:19<02:45,  1.36s/it]

Calculating Affiliation Metrics:  40%|████      | 80/200 [01:20<02:42,  1.36s/it]

Calculating Affiliation Metrics:  40%|████      | 81/200 [01:22<02:41,  1.35s/it]

Calculating Affiliation Metrics:  41%|████      | 82/200 [01:23<02:39,  1.35s/it]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [01:24<02:38,  1.35s/it]

Calculating Affiliation Metrics:  42%|████▏     | 84/200 [01:26<02:37,  1.35s/it]

Calculating Affiliation Metrics:  42%|████▎     | 85/200 [01:27<02:34,  1.34s/it]

Calculating Affiliation Metrics:  43%|████▎     | 86/200 [01:28<02:32,  1.34s/it]

Calculating Affiliation Metrics:  44%|████▎     | 87/200 [01:30<02:28,  1.32s/it]

Calculating Affiliation Metrics:  44%|████▍     | 88/200 [01:31<02:25,  1.30s/it]

Calculating Affiliation Metrics:  44%|████▍     | 89/200 [01:32<02:23,  1.29s/it]

Calculating Affiliation Metrics:  45%|████▌     | 90/200 [01:33<02:21,  1.28s/it]

Calculating Affiliation Metrics:  46%|████▌     | 91/200 [01:35<02:19,  1.28s/it]

Calculating Affiliation Metrics:  46%|████▌     | 92/200 [01:36<02:17,  1.27s/it]

Calculating Affiliation Metrics:  46%|████▋     | 93/200 [01:37<02:16,  1.27s/it]

Calculating Affiliation Metrics:  47%|████▋     | 94/200 [01:38<02:14,  1.27s/it]

Calculating Affiliation Metrics:  48%|████▊     | 95/200 [01:40<02:11,  1.26s/it]

Calculating Affiliation Metrics:  48%|████▊     | 96/200 [01:41<02:09,  1.24s/it]

Calculating Affiliation Metrics:  48%|████▊     | 97/200 [01:42<02:06,  1.23s/it]

Calculating Affiliation Metrics:  49%|████▉     | 98/200 [01:43<02:05,  1.23s/it]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [01:44<02:03,  1.22s/it]

Calculating Affiliation Metrics:  50%|█████     | 100/200 [01:46<02:01,  1.22s/it]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [01:47<01:59,  1.21s/it]

Calculating Affiliation Metrics:  51%|█████     | 102/200 [01:48<01:58,  1.21s/it]

Calculating Affiliation Metrics:  52%|█████▏    | 103/200 [01:49<01:56,  1.20s/it]

Calculating Affiliation Metrics:  52%|█████▏    | 104/200 [01:50<01:54,  1.19s/it]

Calculating Affiliation Metrics:  52%|█████▎    | 105/200 [01:52<01:53,  1.19s/it]

Calculating Affiliation Metrics:  53%|█████▎    | 106/200 [01:53<01:52,  1.20s/it]

Calculating Affiliation Metrics:  54%|█████▎    | 107/200 [01:54<01:51,  1.19s/it]

Calculating Affiliation Metrics:  54%|█████▍    | 108/200 [01:55<01:50,  1.20s/it]

Calculating Affiliation Metrics:  55%|█████▍    | 109/200 [01:56<01:48,  1.19s/it]

Calculating Affiliation Metrics:  55%|█████▌    | 110/200 [01:58<01:46,  1.19s/it]

Calculating Affiliation Metrics:  56%|█████▌    | 111/200 [01:59<01:45,  1.19s/it]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [02:00<01:45,  1.19s/it]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [02:01<01:44,  1.20s/it]

Calculating Affiliation Metrics:  57%|█████▋    | 114/200 [02:02<01:44,  1.22s/it]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [02:04<01:43,  1.22s/it]

Calculating Affiliation Metrics:  58%|█████▊    | 116/200 [02:05<01:42,  1.22s/it]

Calculating Affiliation Metrics:  58%|█████▊    | 117/200 [02:06<01:41,  1.23s/it]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [02:07<01:41,  1.24s/it]

Calculating Affiliation Metrics:  60%|█████▉    | 119/200 [02:09<01:40,  1.24s/it]

Calculating Affiliation Metrics:  60%|██████    | 120/200 [02:10<01:39,  1.25s/it]

Calculating Affiliation Metrics:  60%|██████    | 121/200 [02:11<01:37,  1.24s/it]

Calculating Affiliation Metrics:  61%|██████    | 122/200 [02:12<01:36,  1.23s/it]

Calculating Affiliation Metrics:  62%|██████▏   | 123/200 [02:14<01:34,  1.23s/it]

Calculating Affiliation Metrics:  62%|██████▏   | 124/200 [02:15<01:32,  1.22s/it]

Calculating Affiliation Metrics:  62%|██████▎   | 125/200 [02:16<01:30,  1.20s/it]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [02:17<01:28,  1.20s/it]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [02:18<01:26,  1.19s/it]

Calculating Affiliation Metrics:  64%|██████▍   | 128/200 [02:19<01:24,  1.18s/it]

Calculating Affiliation Metrics:  64%|██████▍   | 129/200 [02:20<01:22,  1.16s/it]

Calculating Affiliation Metrics:  65%|██████▌   | 130/200 [02:22<01:20,  1.15s/it]

Calculating Affiliation Metrics:  66%|██████▌   | 131/200 [02:23<01:18,  1.13s/it]

Calculating Affiliation Metrics:  66%|██████▌   | 132/200 [02:24<01:16,  1.12s/it]

Calculating Affiliation Metrics:  66%|██████▋   | 133/200 [02:25<01:13,  1.10s/it]

Calculating Affiliation Metrics:  67%|██████▋   | 134/200 [02:26<01:10,  1.07s/it]

Calculating Affiliation Metrics:  68%|██████▊   | 136/200 [02:28<01:05,  1.02s/it]

Calculating Affiliation Metrics:  69%|██████▉   | 138/200 [02:30<01:01,  1.01it/s]

Calculating Affiliation Metrics:  70%|███████   | 140/200 [02:32<00:57,  1.03it/s]

Calculating Affiliation Metrics:  71%|███████   | 142/200 [02:33<00:54,  1.06it/s]

Calculating Affiliation Metrics:  72%|███████▏  | 144/200 [02:35<00:51,  1.09it/s]

Calculating Affiliation Metrics:  73%|███████▎  | 146/200 [02:37<00:47,  1.13it/s]

Calculating Affiliation Metrics:  74%|███████▍  | 148/200 [02:38<00:44,  1.17it/s]

Calculating Affiliation Metrics:  75%|███████▌  | 150/200 [02:40<00:40,  1.22it/s]

Calculating Affiliation Metrics:  76%|███████▌  | 152/200 [02:41<00:37,  1.28it/s]

Calculating Affiliation Metrics:  77%|███████▋  | 154/200 [02:42<00:34,  1.34it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 156/200 [02:44<00:31,  1.39it/s]

Calculating Affiliation Metrics:  79%|███████▉  | 158/200 [02:45<00:28,  1.46it/s]

Calculating Affiliation Metrics:  80%|████████  | 160/200 [02:46<00:26,  1.53it/s]

Calculating Affiliation Metrics:  81%|████████  | 162/200 [02:47<00:23,  1.62it/s]

Calculating Affiliation Metrics:  82%|████████▎ | 165/200 [02:49<00:19,  1.77it/s]

Calculating Affiliation Metrics:  84%|████████▍ | 168/200 [02:50<00:16,  1.90it/s]

Calculating Affiliation Metrics:  86%|████████▌ | 171/200 [02:51<00:14,  2.03it/s]

Calculating Affiliation Metrics:  87%|████████▋ | 174/200 [02:52<00:11,  2.19it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 177/200 [02:53<00:09,  2.37it/s]

Calculating Affiliation Metrics:  90%|█████████ | 180/200 [02:55<00:07,  2.53it/s]

Calculating Affiliation Metrics:  92%|█████████▏| 184/200 [02:56<00:05,  2.92it/s]

Calculating Affiliation Metrics:  94%|█████████▍| 189/200 [02:57<00:03,  3.40it/s]

Calculating Affiliation Metrics:  98%|█████████▊| 195/200 [02:58<00:01,  4.11it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [02:58<00:00,  1.12it/s]
Table1-only metrics completed in 221.26s
Evaluation Result:  {'AUC-PR': 0.17129475563345462, 'AUC-ROC': 0.5401420591068018, 'VUS-PR': 0.9046396243580893, 'VUS-ROC': 0.9412528278470259, 'Standard-F1': 0.2473086994469623, 'Standard-Precision': 0.14255366726296959, 'Standard-Recall': 0.9326993416239938, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6863766400742217, 'Affiliation-P': 0.5225064208001019, 'Affiliation-R': 1.0, 'F1_T': 0.2473327766952048, 'Precision_T': 0.1416277093072673, 'Recall_T': 0.9751280175566932}
Results for 152_Stock_id_4_Finance_tr_500_1st_2.csv: {'filename': '152_Stock_id_4_Finance_tr_500_1st_2.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 76, 'train_index': '500', 'data_shape': '20000x1', 'output_length': 20000, 'label_length': 19500, 'aligned_length': 19500, 'AUC-PR': 0.17129475563345462, 'AUC-ROC': 0.5401420591068018, 'VUS-PR': 0.9046396243580893, 'VUS-ROC': 0.9412528278470259, 'Standard-F1': 0.2473086994469623, 'Standard-Precision': 0.14255366726296959, 'Standard-Recall': 0.9326993416239938, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6863766400742217, 'Affiliation-P': 0.5225064208001019, 'Affiliation-R': 1.0, 'F1_T': 0.2473327766952048, 'Precision_T': 0.1416277093072673, 'Recall_T': 0.9751280175566932}
Logits metrics skipped for 152_Stock_id_4_Finance_tr_500_1st_2.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/152_Stock_id_4_Finance_tr_500_1st_2_results.pkl
Processing file: 162_Stock_id_14_Finance_tr_500_1st_4.csv, Data shape: (10000, 1)
Label shape: (10000,)
Train data shape: (500, 1), Test data shape: (9500, 1), Label test shape: (9500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 162_Stock_id_14_Finance_tr_500_1st_4.csv...
Using CUDA device: 2
Data shape: (9500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.74it/s]
Original shapes - Output: (10000,), Label: (9500,)
Aligned shapes - Output: (9500,), Label: (9500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   8%|▊         | 15/200 [00:01<00:12, 14.62it/s]

Calculating Affiliation Metrics:  15%|█▌        | 30/200 [00:03<00:22,  7.69it/s]

Calculating Affiliation Metrics:  20%|██        | 40/200 [00:06<00:30,  5.18it/s]

Calculating Affiliation Metrics:  24%|██▎       | 47/200 [00:09<00:36,  4.20it/s]

Calculating Affiliation Metrics:  26%|██▋       | 53/200 [00:11<00:41,  3.52it/s]

Calculating Affiliation Metrics:  29%|██▉       | 58/200 [00:14<00:45,  3.10it/s]

Calculating Affiliation Metrics:  31%|███       | 62/200 [00:16<00:49,  2.81it/s]

Calculating Affiliation Metrics:  32%|███▎      | 65/200 [00:17<00:51,  2.60it/s]

Calculating Affiliation Metrics:  34%|███▍      | 68/200 [00:19<00:54,  2.43it/s]

Calculating Affiliation Metrics:  36%|███▌      | 71/200 [00:20<00:55,  2.32it/s]

Calculating Affiliation Metrics:  37%|███▋      | 74/200 [00:22<00:56,  2.25it/s]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [00:23<00:56,  2.19it/s]

Calculating Affiliation Metrics:  40%|████      | 80/200 [00:24<00:55,  2.16it/s]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [00:26<00:54,  2.15it/s]

Calculating Affiliation Metrics:  43%|████▎     | 86/200 [00:27<00:52,  2.15it/s]

Calculating Affiliation Metrics:  44%|████▍     | 89/200 [00:29<00:51,  2.17it/s]

Calculating Affiliation Metrics:  46%|████▌     | 92/200 [00:30<00:48,  2.22it/s]

Calculating Affiliation Metrics:  48%|████▊     | 95/200 [00:31<00:46,  2.27it/s]

Calculating Affiliation Metrics:  49%|████▉     | 98/200 [00:32<00:43,  2.36it/s]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [00:33<00:40,  2.43it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 104/200 [00:35<00:38,  2.47it/s]

Calculating Affiliation Metrics:  54%|█████▎    | 107/200 [00:36<00:37,  2.50it/s]

Calculating Affiliation Metrics:  55%|█████▌    | 110/200 [00:37<00:35,  2.53it/s]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [00:38<00:33,  2.58it/s]

Calculating Affiliation Metrics:  58%|█████▊    | 116/200 [00:39<00:31,  2.67it/s]

Calculating Affiliation Metrics:  60%|█████▉    | 119/200 [00:40<00:29,  2.75it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 123/200 [00:41<00:26,  2.86it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [00:43<00:24,  3.03it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 131/200 [00:44<00:21,  3.23it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 135/200 [00:45<00:18,  3.42it/s]

Calculating Affiliation Metrics:  70%|███████   | 140/200 [00:46<00:16,  3.66it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [00:47<00:13,  3.93it/s]

Calculating Affiliation Metrics:  76%|███████▌  | 151/200 [00:48<00:11,  4.32it/s]

Calculating Affiliation Metrics:  79%|███████▉  | 158/200 [00:49<00:08,  4.88it/s]

Calculating Affiliation Metrics:  83%|████████▎ | 166/200 [00:50<00:05,  5.73it/s]

Calculating Affiliation Metrics:  89%|████████▉ | 178/200 [00:51<00:02,  7.50it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:52<00:00,  3.81it/s]
Table1-only metrics completed in 80.12s
Evaluation Result:  {'AUC-PR': 0.14699648975577267, 'AUC-ROC': 0.5188424964772882, 'VUS-PR': 0.9363982273168927, 'VUS-ROC': 0.9626418928630962, 'Standard-F1': 0.24447993986636254, 'Standard-Precision': 0.13944265809217574, 'Standard-Recall': 0.9908606245239902, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.68773072229537, 'Affiliation-P': 0.5240774389676504, 'Affiliation-R': 1.0, 'F1_T': 0.24402586449233402, 'Precision_T': 0.13913229322504805, 'Recall_T': 0.9916222391469909}
Results for 162_Stock_id_14_Finance_tr_500_1st_4.csv: {'filename': '162_Stock_id_14_Finance_tr_500_1st_4.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '10000x1', 'output_length': 10000, 'label_length': 9500, 'aligned_length': 9500, 'AUC-PR': 0.14699648975577267, 'AUC-ROC': 0.5188424964772882, 'VUS-PR': 0.9363982273168927, 'VUS-ROC': 0.9626418928630962, 'Standard-F1': 0.24447993986636254, 'Standard-Precision': 0.13944265809217574, 'Standard-Recall': 0.9908606245239902, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.68773072229537, 'Affiliation-P': 0.5240774389676504, 'Affiliation-R': 1.0, 'F1_T': 0.24402586449233402, 'Precision_T': 0.13913229322504805, 'Recall_T': 0.9916222391469909}
Logits metrics skipped for 162_Stock_id_14_Finance_tr_500_1st_4.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/162_Stock_id_14_Finance_tr_500_1st_4_results.pkl
Processing file: 584_YAHOO_id_34_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 584_YAHOO_id_34_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 158.03it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2122.84it/s]
Table1-only metrics completed in 0.71s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 584_YAHOO_id_34_Synthetic_tr_500_1st_658.csv: {'filename': '584_YAHOO_id_34_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 113, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 584_YAHOO_id_34_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/584_YAHOO_id_34_Synthetic_tr_500_1st_658_results.pkl
Processing file: 675_YAHOO_id_125_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 675_YAHOO_id_125_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 160.40it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1537.20it/s]
Table1-only metrics completed in 0.37s
Evaluation Result:  {'AUC-PR': 0.9535634118967453, 'AUC-ROC': 0.9995126705653022, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984359113984375, 'Affiliation-P': 0.9969624510347107, 'Affiliation-R': 0.9999137336093858, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 675_YAHOO_id_125_Synthetic_tr_500_1st_623.csv: {'filename': '675_YAHOO_id_125_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 35, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9535634118967453, 'AUC-ROC': 0.9995126705653022, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984359113984375, 'Affiliation-P': 0.9969624510347107, 'Affiliation-R': 0.9999137336093858, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 675_YAHOO_id_125_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/675_YAHOO_id_125_Synthetic_tr_500_1st_623_results.pkl
Processing file: 714_YAHOO_id_164_WebService_tr_500_1st_1.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 714_YAHOO_id_164_WebService_tr_500_1st_1.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 165.34it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1296.74it/s]
Table1-only metrics completed in 0.42s
Evaluation Result:  {'AUC-PR': 0.9666666666666666, 'AUC-ROC': 0.9998297872340425, 'VUS-PR': 0.742, 'VUS-ROC': 0.999115063829787, 'Standard-F1': 0.5882352941171626, 'Standard-Precision': 0.41666666666663194, 'Standard-Recall': 0.9999999999997999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.980254936265933, 'Affiliation-P': 0.9612745098039216, 'Affiliation-R': 1.0, 'F1_T': 0.5882352941171626, 'Precision_T': 0.41666666666663194, 'Recall_T': 0.9999999999997999}
Results for 714_YAHOO_id_164_WebService_tr_500_1st_1.csv: {'filename': '714_YAHOO_id_164_WebService_tr_500_1st_1.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9666666666666666, 'AUC-ROC': 0.9998297872340425, 'VUS-PR': 0.742, 'VUS-ROC': 0.999115063829787, 'Standard-F1': 0.5882352941171626, 'Standard-Precision': 0.41666666666663194, 'Standard-Recall': 0.9999999999997999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.980254936265933, 'Affiliation-P': 0.9612745098039216, 'Affiliation-R': 1.0, 'F1_T': 0.5882352941171626, 'Precision_T': 0.41666666666663194, 'Recall_T': 0.9999999999997999}
Logits metrics skipped for 714_YAHOO_id_164_WebService_tr_500_1st_1.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/714_YAHOO_id_164_WebService_tr_500_1st_1_results.pkl
Processing file: 349_UCR_id_47_Facility_tr_29954_1st_74900.csv, Data shape: (119819, 1)
Label shape: (119819,)
Train data shape: (29954, 1), Test data shape: (89865, 1), Label test shape: (89865,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 349_UCR_id_47_Facility_tr_29954_1st_74900.csv...
Using CUDA device: 2
Data shape: (89865, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 89865, window_size 5000, adding 135 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([18, 5000, 1])
batched_mask shape: torch.Size([18, 5000])

100%|██████████| 1/1 [00:00<00:00, 55.36it/s]
Original shapes - Output: (90000,), Label: (89865,)
Aligned shapes - Output: (89865,), Label: (89865,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  20%|██        | 40/200 [00:01<00:04, 39.72it/s]

Calculating Affiliation Metrics:  40%|████      | 80/200 [00:02<00:03, 35.94it/s]

Calculating Affiliation Metrics:  58%|█████▊    | 117/200 [00:03<00:02, 35.14it/s]

Calculating Affiliation Metrics:  80%|████████  | 161/200 [00:04<00:01, 38.42it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:04<00:00, 41.63it/s]
Table1-only metrics completed in 7.04s
Evaluation Result:  {'AUC-PR': 0.0006984667438833293, 'AUC-ROC': 0.2794483851143862, 'VUS-PR': 0.001069190999163487, 'VUS-ROC': 0.3876936302341462, 'Standard-F1': 0.0016541077007858947, 'Standard-Precision': 0.0008288819303739178, 'Standard-Recall': 0.3749999999999961, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7087119598867672, 'Affiliation-P': 0.5517948241023192, 'Affiliation-R': 0.9903410671562899, 'F1_T': 0.0016559337626450882, 'Precision_T': 0.000829799004241195, 'Recall_T': 0.3749999999999961}
Results for 349_UCR_id_47_Facility_tr_29954_1st_74900.csv: {'filename': '349_UCR_id_47_Facility_tr_29954_1st_74900.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 96, 'train_index': '29954', 'data_shape': '119819x1', 'output_length': 90000, 'label_length': 89865, 'aligned_length': 89865, 'AUC-PR': 0.0006984667438833293, 'AUC-ROC': 0.2794483851143862, 'VUS-PR': 0.001069190999163487, 'VUS-ROC': 0.3876936302341462, 'Standard-F1': 0.0016541077007858947, 'Standard-Precision': 0.0008288819303739178, 'Standard-Recall': 0.3749999999999961, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7087119598867672, 'Affiliation-P': 0.5517948241023192, 'Affiliation-R': 0.9903410671562899, 'F1_T': 0.0016559337626450882, 'Precision_T': 0.000829799004241195, 'Recall_T': 0.3749999999999961}
Logits metrics skipped for 349_UCR_id_47_Facility_tr_29954_1st_74900.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/349_UCR_id_47_Facility_tr_29954_1st_74900_results.pkl
Skipping file: 542_SMAP_id_12_Sensor_tr_1908_1st_4690.csv due to filter criteria.
Processing file: 453_UCR_id_151_Facility_tr_32000_1st_67995.csv, Data shape: (128001, 1)
Label shape: (128001,)
Train data shape: (32000, 1), Test data shape: (96001, 1), Label test shape: (96001,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 453_UCR_id_151_Facility_tr_32000_1st_67995.csv...
Using CUDA device: 2
Data shape: (96001, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 96001, window_size 5000, adding 3999 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([20, 5000, 1])
batched_mask shape: torch.Size([20, 5000])

100%|██████████| 1/1 [00:00<00:00, 64.44it/s]
Original shapes - Output: (100000,), Label: (96001,)
Aligned shapes - Output: (96001,), Label: (96001,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  20%|█▉        | 39/200 [00:01<00:04, 38.81it/s]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [00:02<00:03, 34.65it/s]

Calculating Affiliation Metrics:  57%|█████▋    | 114/200 [00:03<00:02, 35.19it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 155/200 [00:04<00:01, 37.28it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:04<00:00, 40.75it/s]
Table1-only metrics completed in 10.78s
Evaluation Result:  {'AUC-PR': 5.448403617740002e-05, 'AUC-ROC': 0.8088229166666667, 'VUS-PR': 0.6495749952752639, 'VUS-ROC': 0.9774954061895628, 'Standard-F1': 0.00010813733441459854, 'Standard-Precision': 5.407159078620093e-05, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7159654026913882, 'Affiliation-P': 0.5575936682006712, 'Affiliation-R': 0.9999895834418392, 'F1_T': 0.00010837758751479351, 'Precision_T': 5.419173034194982e-05, 'Recall_T': 0.9999999999989999}
Results for 453_UCR_id_151_Facility_tr_32000_1st_67995.csv: {'filename': '453_UCR_id_151_Facility_tr_32000_1st_67995.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 243, 'train_index': '32000', 'data_shape': '128001x1', 'output_length': 100000, 'label_length': 96001, 'aligned_length': 96001, 'AUC-PR': 5.448403617740002e-05, 'AUC-ROC': 0.8088229166666667, 'VUS-PR': 0.6495749952752639, 'VUS-ROC': 0.9774954061895628, 'Standard-F1': 0.00010813733441459854, 'Standard-Precision': 5.407159078620093e-05, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7159654026913882, 'Affiliation-P': 0.5575936682006712, 'Affiliation-R': 0.9999895834418392, 'F1_T': 0.00010837758751479351, 'Precision_T': 5.419173034194982e-05, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 453_UCR_id_151_Facility_tr_32000_1st_67995.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/453_UCR_id_151_Facility_tr_32000_1st_67995_results.pkl
Processing file: 323_UCR_id_21_HumanActivity_tr_48812_1st_128430.csv, Data shape: (195250, 1)
Label shape: (195250,)
Train data shape: (48812, 1), Test data shape: (146438, 1), Label test shape: (146438,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 323_UCR_id_21_HumanActivity_tr_48812_1st_128430.csv...
Using CUDA device: 2
Data shape: (146438, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 146438, window_size 5000, adding 3562 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([30, 5000, 1])
batched_mask shape: torch.Size([30, 5000])

100%|██████████| 1/1 [00:00<00:00, 46.90it/s]
Original shapes - Output: (150000,), Label: (146438,)
Aligned shapes - Output: (146438,), Label: (146438,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  14%|█▎        | 27/200 [00:01<00:06, 26.75it/s]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:02<00:05, 26.44it/s]

Calculating Affiliation Metrics:  41%|████      | 82/200 [00:03<00:04, 27.05it/s]

Calculating Affiliation Metrics:  57%|█████▋    | 114/200 [00:04<00:02, 28.69it/s]

Calculating Affiliation Metrics:  76%|███████▋  | 153/200 [00:05<00:01, 32.23it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:05<00:00, 33.43it/s]
Table1-only metrics completed in 11.19s
Evaluation Result:  {'AUC-PR': 8.540074298646397e-06, 'AUC-ROC': 0.20037968546200757, 'VUS-PR': 1.5685100578848533e-05, 'VUS-ROC': 0.24164030148810317, 'Standard-F1': 1.7040855450926597e-05, 'Standard-Precision': 8.520500323779012e-06, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.699919122942765, 'Affiliation-P': 0.5437380138285264, 'Affiliation-R': 0.9819787213701362, 'F1_T': 1.7010997609937823e-05, 'Precision_T': 8.505571149102662e-06, 'Recall_T': 0.9999999999989999}
Results for 323_UCR_id_21_HumanActivity_tr_48812_1st_128430.csv: {'filename': '323_UCR_id_21_HumanActivity_tr_48812_1st_128430.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '48812', 'data_shape': '195250x1', 'output_length': 150000, 'label_length': 146438, 'aligned_length': 146438, 'AUC-PR': 8.540074298646397e-06, 'AUC-ROC': 0.20037968546200757, 'VUS-PR': 1.5685100578848533e-05, 'VUS-ROC': 0.24164030148810317, 'Standard-F1': 1.7040855450926597e-05, 'Standard-Precision': 8.520500323779012e-06, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.699919122942765, 'Affiliation-P': 0.5437380138285264, 'Affiliation-R': 0.9819787213701362, 'F1_T': 1.7010997609937823e-05, 'Precision_T': 8.505571149102662e-06, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 323_UCR_id_21_HumanActivity_tr_48812_1st_128430.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/323_UCR_id_21_HumanActivity_tr_48812_1st_128430_results.pkl
Processing file: 356_UCR_id_54_Medical_tr_7500_1st_16000.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 356_UCR_id_54_Medical_tr_7500_1st_16000.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 145.72it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  90%|█████████ | 180/200 [00:01<00:00, 179.91it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 187.84it/s]
Table1-only metrics completed in 1.97s
Evaluation Result:  {'AUC-PR': 0.011792931372729812, 'AUC-ROC': 0.8059370535714286, 'VUS-PR': 0.016071329244701848, 'VUS-ROC': 0.8478434620971873, 'Standard-F1': 0.029069767441685383, 'Standard-Precision': 0.01609442060085835, 'Standard-Recall': 0.1499999999999985, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7076408346299604, 'Affiliation-P': 0.547676876131368, 'Affiliation-R': 0.9996017777777779, 'F1_T': 0.027247956403104547, 'Precision_T': 0.01498501498501497, 'Recall_T': 0.1499999999999985}
Results for 356_UCR_id_54_Medical_tr_7500_1st_16000.csv: {'filename': '356_UCR_id_54_Medical_tr_7500_1st_16000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 166, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.011792931372729812, 'AUC-ROC': 0.8059370535714286, 'VUS-PR': 0.016071329244701848, 'VUS-ROC': 0.8478434620971873, 'Standard-F1': 0.029069767441685383, 'Standard-Precision': 0.01609442060085835, 'Standard-Recall': 0.1499999999999985, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7076408346299604, 'Affiliation-P': 0.547676876131368, 'Affiliation-R': 0.9996017777777779, 'F1_T': 0.027247956403104547, 'Precision_T': 0.01498501498501497, 'Recall_T': 0.1499999999999985}
Logits metrics skipped for 356_UCR_id_54_Medical_tr_7500_1st_16000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/356_UCR_id_54_Medical_tr_7500_1st_16000_results.pkl
Processing file: 391_UCR_id_89_Medical_tr_1875_1st_5759.csv, Data shape: (7500, 1)
Label shape: (7500,)
Train data shape: (1875, 1), Test data shape: (5625, 1), Label test shape: (5625,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 391_UCR_id_89_Medical_tr_1875_1st_5759.csv...
Using CUDA device: 2
Data shape: (5625, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5625, window_size 5000, adding 4375 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.20it/s]
Original shapes - Output: (10000,), Label: (5625,)
Aligned shapes - Output: (5625,), Label: (5625,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 676.39it/s]
Table1-only metrics completed in 0.99s
Evaluation Result:  {'AUC-PR': 0.17217684853255757, 'AUC-ROC': 0.8857811070448307, 'VUS-PR': 0.19980582640146216, 'VUS-ROC': 0.9176478690695641, 'Standard-F1': 0.282608695651719, 'Standard-Precision': 0.21666666666666592, 'Standard-Recall': 0.4062499999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7645056479146868, 'Affiliation-P': 0.620718779956427, 'Affiliation-R': 0.9949911111111112, 'F1_T': 0.27896995708109296, 'Precision_T': 0.21241830065359407, 'Recall_T': 0.4062499999999975}
Results for 391_UCR_id_89_Medical_tr_1875_1st_5759.csv: {'filename': '391_UCR_id_89_Medical_tr_1875_1st_5759.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 183, 'train_index': '1875', 'data_shape': '7500x1', 'output_length': 10000, 'label_length': 5625, 'aligned_length': 5625, 'AUC-PR': 0.17217684853255757, 'AUC-ROC': 0.8857811070448307, 'VUS-PR': 0.19980582640146216, 'VUS-ROC': 0.9176478690695641, 'Standard-F1': 0.282608695651719, 'Standard-Precision': 0.21666666666666592, 'Standard-Recall': 0.4062499999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7645056479146868, 'Affiliation-P': 0.620718779956427, 'Affiliation-R': 0.9949911111111112, 'F1_T': 0.27896995708109296, 'Precision_T': 0.21241830065359407, 'Recall_T': 0.4062499999999975}
Logits metrics skipped for 391_UCR_id_89_Medical_tr_1875_1st_5759.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/391_UCR_id_89_Medical_tr_1875_1st_5759_results.pkl
Processing file: 708_YAHOO_id_158_WebService_tr_500_1st_165.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 708_YAHOO_id_158_WebService_tr_500_1st_165.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 160.47it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1631.44it/s]
Table1-only metrics completed in 0.27s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.5893449002318925, 'VUS-ROC': 0.9991284394172775, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9927203671466999, 'Affiliation-P': 0.9855459544383347, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Results for 708_YAHOO_id_158_WebService_tr_500_1st_165.csv: {'filename': '708_YAHOO_id_158_WebService_tr_500_1st_165.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.5893449002318925, 'VUS-ROC': 0.9991284394172775, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9927203671466999, 'Affiliation-P': 0.9855459544383347, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Logits metrics skipped for 708_YAHOO_id_158_WebService_tr_500_1st_165.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/708_YAHOO_id_158_WebService_tr_500_1st_165_results.pkl
Processing file: 645_YAHOO_id_95_WebService_tr_500_1st_1047.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 645_YAHOO_id_95_WebService_tr_500_1st_1047.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.37it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1491.95it/s]
Table1-only metrics completed in 0.24s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9956022130798688, 'Affiliation-P': 0.9912429378531072, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Results for 645_YAHOO_id_95_WebService_tr_500_1st_1047.csv: {'filename': '645_YAHOO_id_95_WebService_tr_500_1st_1047.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9956022130798688, 'Affiliation-P': 0.9912429378531072, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 645_YAHOO_id_95_WebService_tr_500_1st_1047.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/645_YAHOO_id_95_WebService_tr_500_1st_1047_results.pkl
Processing file: 702_YAHOO_id_152_WebService_tr_500_1st_8.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 702_YAHOO_id_152_WebService_tr_500_1st_8.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 167.73it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 750.88it/s]
Table1-only metrics completed in 0.82s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9915669033451497, 'VUS-ROC': 0.9999458235893512, 'Standard-F1': 0.9999999999994167, 'Standard-Precision': 0.9999999999999166, 'Standard-Recall': 0.9999999999999166, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.9999999999994167, 'Precision_T': 0.9999999999999166, 'Recall_T': 0.9999999999999166}
Results for 702_YAHOO_id_152_WebService_tr_500_1st_8.csv: {'filename': '702_YAHOO_id_152_WebService_tr_500_1st_8.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9915669033451497, 'VUS-ROC': 0.9999458235893512, 'Standard-F1': 0.9999999999994167, 'Standard-Precision': 0.9999999999999166, 'Standard-Recall': 0.9999999999999166, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.9999999999994167, 'Precision_T': 0.9999999999999166, 'Recall_T': 0.9999999999999166}
Logits metrics skipped for 702_YAHOO_id_152_WebService_tr_500_1st_8.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/702_YAHOO_id_152_WebService_tr_500_1st_8_results.pkl
Processing file: 507_UCR_id_205_HumanActivity_tr_16000_1st_31200.csv, Data shape: (64000, 1)
Label shape: (64000,)
Train data shape: (16000, 1), Test data shape: (48000, 1), Label test shape: (48000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 507_UCR_id_205_HumanActivity_tr_16000_1st_31200.csv...
Using CUDA device: 2
Data shape: (48000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 48000, window_size 5000, adding 2000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([10, 5000, 1])
batched_mask shape: torch.Size([10, 5000])

100%|██████████| 1/1 [00:00<00:00, 95.87it/s]
Original shapes - Output: (50000,), Label: (48000,)
Aligned shapes - Output: (48000,), Label: (48000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  44%|████▍     | 89/200 [00:01<00:01, 88.49it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 106.63it/s]
Table1-only metrics completed in 2.96s
Evaluation Result:  {'AUC-PR': 0.047144382166608975, 'AUC-ROC': 0.8570900333035498, 'VUS-PR': 0.05062464871934167, 'VUS-ROC': 0.8630939935829595, 'Standard-F1': 0.09674610086285858, 'Standard-Precision': 0.05099741256990234, 'Standard-Recall': 0.9399999999999985, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6840502822091912, 'Affiliation-P': 0.5198149085495171, 'Affiliation-R': 1.0, 'F1_T': 0.09649755550202513, 'Precision_T': 0.050900481948084886, 'Recall_T': 0.9261538461538447}
Results for 507_UCR_id_205_HumanActivity_tr_16000_1st_31200.csv: {'filename': '507_UCR_id_205_HumanActivity_tr_16000_1st_31200.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '16000', 'data_shape': '64000x1', 'output_length': 50000, 'label_length': 48000, 'aligned_length': 48000, 'AUC-PR': 0.047144382166608975, 'AUC-ROC': 0.8570900333035498, 'VUS-PR': 0.05062464871934167, 'VUS-ROC': 0.8630939935829595, 'Standard-F1': 0.09674610086285858, 'Standard-Precision': 0.05099741256990234, 'Standard-Recall': 0.9399999999999985, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6840502822091912, 'Affiliation-P': 0.5198149085495171, 'Affiliation-R': 1.0, 'F1_T': 0.09649755550202513, 'Precision_T': 0.050900481948084886, 'Recall_T': 0.9261538461538447}
Logits metrics skipped for 507_UCR_id_205_HumanActivity_tr_16000_1st_31200.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/507_UCR_id_205_HumanActivity_tr_16000_1st_31200_results.pkl
Processing file: 659_YAHOO_id_109_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 659_YAHOO_id_109_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 158.49it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1585.81it/s]
Table1-only metrics completed in 0.82s
Evaluation Result:  {'AUC-PR': 0.8517195767195769, 'AUC-ROC': 0.9987816764132553, 'VUS-PR': 0.9943053101472179, 'VUS-ROC': 0.9999613014421813, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 659_YAHOO_id_109_Synthetic_tr_500_1st_623.csv: {'filename': '659_YAHOO_id_109_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 118, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.8517195767195769, 'AUC-ROC': 0.9987816764132553, 'VUS-PR': 0.9943053101472179, 'VUS-ROC': 0.9999613014421813, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 659_YAHOO_id_109_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/659_YAHOO_id_109_Synthetic_tr_500_1st_623_results.pkl
Skipping file: 817_Exathlon_id_8_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 648_YAHOO_id_98_WebService_tr_500_1st_731.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 648_YAHOO_id_98_WebService_tr_500_1st_731.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 162.99it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1801.81it/s]
Table1-only metrics completed in 0.27s
Evaluation Result:  {'AUC-PR': 0.5126582278481012, 'AUC-ROC': 0.9673174872665535, 'VUS-PR': 0.35586770898566955, 'VUS-ROC': 0.9659868679429492, 'Standard-F1': 0.14285714285687756, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.49999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.941629974815398, 'Affiliation-P': 0.8896982646982647, 'Affiliation-R': 1.0, 'F1_T': 0.14285714285687756, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.49999999999975}
Results for 648_YAHOO_id_98_WebService_tr_500_1st_731.csv: {'filename': '648_YAHOO_id_98_WebService_tr_500_1st_731.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.5126582278481012, 'AUC-ROC': 0.9673174872665535, 'VUS-PR': 0.35586770898566955, 'VUS-ROC': 0.9659868679429492, 'Standard-F1': 0.14285714285687756, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.49999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.941629974815398, 'Affiliation-P': 0.8896982646982647, 'Affiliation-R': 1.0, 'F1_T': 0.14285714285687756, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.49999999999975}
Logits metrics skipped for 648_YAHOO_id_98_WebService_tr_500_1st_731.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/648_YAHOO_id_98_WebService_tr_500_1st_731_results.pkl
Processing file: 694_YAHOO_id_144_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 694_YAHOO_id_144_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 162.19it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2140.31it/s]
Table1-only metrics completed in 0.66s
Evaluation Result:  {'AUC-PR': 0.95, 'AUC-ROC': 0.9997273718647764, 'VUS-PR': 0.9893051815003151, 'VUS-ROC': 0.9999645556378418, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 694_YAHOO_id_144_Synthetic_tr_500_1st_658.csv: {'filename': '694_YAHOO_id_144_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 103, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.95, 'AUC-ROC': 0.9997273718647764, 'VUS-PR': 0.9893051815003151, 'VUS-ROC': 0.9999645556378418, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 694_YAHOO_id_144_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/694_YAHOO_id_144_Synthetic_tr_500_1st_658_results.pkl
Processing file: 093_WSD_id_65_WebService_tr_1125_1st_1225.csv, Data shape: (18179, 1)
Label shape: (18179,)
Train data shape: (1125, 1), Test data shape: (17054, 1), Label test shape: (17054,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 093_WSD_id_65_WebService_tr_1125_1st_1225.csv...
Using CUDA device: 2
Data shape: (17054, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 17054, window_size 5000, adding 2946 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.85it/s]
Original shapes - Output: (20000,), Label: (17054,)
Aligned shapes - Output: (17054,), Label: (17054,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  58%|█████▊    | 117/200 [00:01<00:00, 116.99it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 141.28it/s]
Table1-only metrics completed in 4.37s
Evaluation Result:  {'AUC-PR': 0.3510044034970304, 'AUC-ROC': 0.9482031675509309, 'VUS-PR': 0.47781047639810126, 'VUS-ROC': 0.9881191159766669, 'Standard-F1': 0.3900414937754672, 'Standard-Precision': 0.30819672131147435, 'Standard-Recall': 0.5310734463276806, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9558017408259187, 'Affiliation-P': 0.9217739669649171, 'Affiliation-R': 0.9924381192699209, 'F1_T': 0.36821705426311374, 'Precision_T': 0.28023598820058915, 'Recall_T': 0.5367231638418049}
Results for 093_WSD_id_65_WebService_tr_1125_1st_1225.csv: {'filename': '093_WSD_id_65_WebService_tr_1125_1st_1225.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '1125', 'data_shape': '18179x1', 'output_length': 20000, 'label_length': 17054, 'aligned_length': 17054, 'AUC-PR': 0.3510044034970304, 'AUC-ROC': 0.9482031675509309, 'VUS-PR': 0.47781047639810126, 'VUS-ROC': 0.9881191159766669, 'Standard-F1': 0.3900414937754672, 'Standard-Precision': 0.30819672131147435, 'Standard-Recall': 0.5310734463276806, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9558017408259187, 'Affiliation-P': 0.9217739669649171, 'Affiliation-R': 0.9924381192699209, 'F1_T': 0.36821705426311374, 'Precision_T': 0.28023598820058915, 'Recall_T': 0.5367231638418049}
Logits metrics skipped for 093_WSD_id_65_WebService_tr_1125_1st_1225.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/093_WSD_id_65_WebService_tr_1125_1st_1225_results.pkl
Processing file: 353_UCR_id_51_Medical_tr_1875_1st_3198.csv, Data shape: (7500, 1)
Label shape: (7500,)
Train data shape: (1875, 1), Test data shape: (5625, 1), Label test shape: (5625,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 353_UCR_id_51_Medical_tr_1875_1st_3198.csv...
Using CUDA device: 2
Data shape: (5625, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5625, window_size 5000, adding 4375 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.16it/s]
Original shapes - Output: (10000,), Label: (5625,)
Aligned shapes - Output: (5625,), Label: (5625,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 607.66it/s]
Table1-only metrics completed in 1.03s
Evaluation Result:  {'AUC-PR': 0.9956695194651647, 'AUC-ROC': 0.9998954340630075, 'VUS-PR': 0.987790819949307, 'VUS-ROC': 0.9998402333563412, 'Standard-F1': 0.9686098654703433, 'Standard-Precision': 0.9642857142857058, 'Standard-Recall': 0.9729729729729643, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9998770619651739, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9997541541541541, 'F1_T': 0.9686098654703433, 'Precision_T': 0.9642857142857058, 'Recall_T': 0.9729729729729643}
Results for 353_UCR_id_51_Medical_tr_1875_1st_3198.csv: {'filename': '353_UCR_id_51_Medical_tr_1875_1st_3198.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 183, 'train_index': '1875', 'data_shape': '7500x1', 'output_length': 10000, 'label_length': 5625, 'aligned_length': 5625, 'AUC-PR': 0.9956695194651647, 'AUC-ROC': 0.9998954340630075, 'VUS-PR': 0.987790819949307, 'VUS-ROC': 0.9998402333563412, 'Standard-F1': 0.9686098654703433, 'Standard-Precision': 0.9642857142857058, 'Standard-Recall': 0.9729729729729643, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9998770619651739, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9997541541541541, 'F1_T': 0.9686098654703433, 'Precision_T': 0.9642857142857058, 'Recall_T': 0.9729729729729643}
Logits metrics skipped for 353_UCR_id_51_Medical_tr_1875_1st_3198.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/353_UCR_id_51_Medical_tr_1875_1st_3198_results.pkl
Processing file: 460_UCR_id_158_Medical_tr_7487_1st_27862.csv, Data shape: (29950, 1)
Label shape: (29950,)
Train data shape: (7487, 1), Test data shape: (22463, 1), Label test shape: (22463,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 460_UCR_id_158_Medical_tr_7487_1st_27862.csv...
Using CUDA device: 2
Data shape: (22463, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22463, window_size 5000, adding 2537 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 148.70it/s]
Original shapes - Output: (25000,), Label: (22463,)
Aligned shapes - Output: (22463,), Label: (22463,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  54%|█████▎    | 107/200 [00:01<00:00, 106.16it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 121.83it/s]
Table1-only metrics completed in 1.89s
Evaluation Result:  {'AUC-PR': 0.7921625097989484, 'AUC-ROC': 0.99709220355851, 'VUS-PR': 0.49507859713354824, 'VUS-ROC': 0.9970876623521593, 'Standard-F1': 0.4067796610165844, 'Standard-Precision': 0.2666666666666655, 'Standard-Recall': 0.857142857142845, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8206984701213852, 'Affiliation-P': 0.6959582306242832, 'Affiliation-R': 0.999919232261306, 'F1_T': 0.4067796610165844, 'Precision_T': 0.2666666666666655, 'Recall_T': 0.857142857142845}
Results for 460_UCR_id_158_Medical_tr_7487_1st_27862.csv: {'filename': '460_UCR_id_158_Medical_tr_7487_1st_27862.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 33, 'train_index': '7487', 'data_shape': '29950x1', 'output_length': 25000, 'label_length': 22463, 'aligned_length': 22463, 'AUC-PR': 0.7921625097989484, 'AUC-ROC': 0.99709220355851, 'VUS-PR': 0.49507859713354824, 'VUS-ROC': 0.9970876623521593, 'Standard-F1': 0.4067796610165844, 'Standard-Precision': 0.2666666666666655, 'Standard-Recall': 0.857142857142845, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8206984701213852, 'Affiliation-P': 0.6959582306242832, 'Affiliation-R': 0.999919232261306, 'F1_T': 0.4067796610165844, 'Precision_T': 0.2666666666666655, 'Recall_T': 0.857142857142845}
Logits metrics skipped for 460_UCR_id_158_Medical_tr_7487_1st_27862.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/460_UCR_id_158_Medical_tr_7487_1st_27862_results.pkl
Skipping file: 822_Exathlon_id_13_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 422_UCR_id_120_Facility_tr_50000_1st_150720.csv, Data shape: (299690, 1)
Label shape: (299690,)
Train data shape: (50000, 1), Test data shape: (249690, 1), Label test shape: (249690,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 422_UCR_id_120_Facility_tr_50000_1st_150720.csv...
Using CUDA device: 2
Data shape: (249690, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 249690, window_size 5000, adding 310 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([50, 5000, 1])
batched_mask shape: torch.Size([50, 5000])

100%|██████████| 1/1 [00:00<00:00, 28.35it/s]
Original shapes - Output: (250000,), Label: (249690,)
Aligned shapes - Output: (249690,), Label: (249690,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   9%|▉         | 18/200 [00:01<00:10, 17.03it/s]

Calculating Affiliation Metrics:  18%|█▊        | 36/200 [00:02<00:11, 14.35it/s]

Calculating Affiliation Metrics:  26%|██▌       | 51/200 [00:03<00:11, 12.72it/s]

Calculating Affiliation Metrics:  32%|███▏      | 64/200 [00:05<00:11, 12.05it/s]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [00:06<00:10, 11.57it/s]

Calculating Affiliation Metrics:  44%|████▍     | 89/200 [00:07<00:09, 11.16it/s]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [00:08<00:08, 11.20it/s]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [00:09<00:07, 11.42it/s]

Calculating Affiliation Metrics:  62%|██████▎   | 125/200 [00:10<00:06, 11.55it/s]

Calculating Affiliation Metrics:  70%|██████▉   | 139/200 [00:11<00:05, 12.20it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 155/200 [00:12<00:03, 13.19it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 175/200 [00:13<00:01, 15.10it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:14<00:00, 13.91it/s]
Table1-only metrics completed in 23.29s
Evaluation Result:  {'AUC-PR': 0.010401504021521093, 'AUC-ROC': 0.6814119449455138, 'VUS-PR': 0.00844515841704922, 'VUS-ROC': 0.6981993222110218, 'Standard-F1': 0.030505243088328075, 'Standard-Precision': 0.019223067681217454, 'Standard-Recall': 0.07384615384615373, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7671818514079176, 'Affiliation-P': 0.6227519909548398, 'Affiliation-R': 0.9988332024017476, 'F1_T': 0.030505243088328075, 'Precision_T': 0.019223067681217454, 'Recall_T': 0.07384615384615373}
Results for 422_UCR_id_120_Facility_tr_50000_1st_150720.csv: {'filename': '422_UCR_id_120_Facility_tr_50000_1st_150720.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '50000', 'data_shape': '299690x1', 'output_length': 250000, 'label_length': 249690, 'aligned_length': 249690, 'AUC-PR': 0.010401504021521093, 'AUC-ROC': 0.6814119449455138, 'VUS-PR': 0.00844515841704922, 'VUS-ROC': 0.6981993222110218, 'Standard-F1': 0.030505243088328075, 'Standard-Precision': 0.019223067681217454, 'Standard-Recall': 0.07384615384615373, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7671818514079176, 'Affiliation-P': 0.6227519909548398, 'Affiliation-R': 0.9988332024017476, 'F1_T': 0.030505243088328075, 'Precision_T': 0.019223067681217454, 'Recall_T': 0.07384615384615373}
Logits metrics skipped for 422_UCR_id_120_Facility_tr_50000_1st_150720.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/422_UCR_id_120_Facility_tr_50000_1st_150720_results.pkl
Skipping file: 215_SMD_id_38_Facility_tr_5925_1st_17580.csv due to filter criteria.
Processing file: 151_Stock_id_3_Finance_tr_500_1st_62.csv, Data shape: (10000, 1)
Label shape: (10000,)
Train data shape: (500, 1), Test data shape: (9500, 1), Label test shape: (9500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 151_Stock_id_3_Finance_tr_500_1st_62.csv...
Using CUDA device: 2
Data shape: (9500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.49it/s]
Original shapes - Output: (10000,), Label: (9500,)
Aligned shapes - Output: (9500,), Label: (9500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  10%|█         | 21/200 [00:01<00:08, 20.63it/s]

Calculating Affiliation Metrics:  21%|██        | 42/200 [00:03<00:14, 11.00it/s]

Calculating Affiliation Metrics:  28%|██▊       | 56/200 [00:05<00:15,  9.20it/s]

Calculating Affiliation Metrics:  34%|███▎      | 67/200 [00:06<00:15,  8.74it/s]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [00:08<00:14,  8.24it/s]

Calculating Affiliation Metrics:  43%|████▎     | 86/200 [00:09<00:14,  7.66it/s]

Calculating Affiliation Metrics:  47%|████▋     | 94/200 [00:11<00:14,  7.28it/s]

Calculating Affiliation Metrics:  51%|█████     | 102/200 [00:12<00:13,  7.05it/s]

Calculating Affiliation Metrics:  55%|█████▌    | 110/200 [00:13<00:12,  6.97it/s]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [00:14<00:11,  7.02it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [00:15<00:09,  7.45it/s]

Calculating Affiliation Metrics:  70%|███████   | 140/200 [00:16<00:06,  8.76it/s]

Calculating Affiliation Metrics:  77%|███████▋  | 154/200 [00:17<00:04, 10.07it/s]

Calculating Affiliation Metrics:  86%|████████▋ | 173/200 [00:18<00:02, 12.47it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:19<00:00, 10.36it/s]
Table1-only metrics completed in 36.72s
Evaluation Result:  {'AUC-PR': 0.06936997638084577, 'AUC-ROC': 0.575175137024672, 'VUS-PR': 0.8165847783112596, 'VUS-ROC': 0.9286413515170522, 'Standard-F1': 0.11757789535527413, 'Standard-Precision': 0.08110300081102995, 'Standard-Recall': 0.21367521367521322, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6772535527974352, 'Affiliation-P': 0.5131864915225153, 'Affiliation-R': 0.9955257270693513, 'F1_T': 0.11743772241952763, 'Precision_T': 0.08128078817733984, 'Recall_T': 0.21153846153846106}
Results for 151_Stock_id_3_Finance_tr_500_1st_62.csv: {'filename': '151_Stock_id_3_Finance_tr_500_1st_62.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '10000x1', 'output_length': 10000, 'label_length': 9500, 'aligned_length': 9500, 'AUC-PR': 0.06936997638084577, 'AUC-ROC': 0.575175137024672, 'VUS-PR': 0.8165847783112596, 'VUS-ROC': 0.9286413515170522, 'Standard-F1': 0.11757789535527413, 'Standard-Precision': 0.08110300081102995, 'Standard-Recall': 0.21367521367521322, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6772535527974352, 'Affiliation-P': 0.5131864915225153, 'Affiliation-R': 0.9955257270693513, 'F1_T': 0.11743772241952763, 'Precision_T': 0.08128078817733984, 'Recall_T': 0.21153846153846106}
Logits metrics skipped for 151_Stock_id_3_Finance_tr_500_1st_62.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/151_Stock_id_3_Finance_tr_500_1st_62_results.pkl
Processing file: 026_NAB_id_26_Traffic_tr_624_1st_2261.csv, Data shape: (2499, 1)
Label shape: (2499,)
Train data shape: (624, 1), Test data shape: (1875, 1), Label test shape: (1875,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 026_NAB_id_26_Traffic_tr_624_1st_2261.csv...
Using CUDA device: 2
Data shape: (1875, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1875, 1])
batched_mask shape: torch.Size([1, 1875])

100%|██████████| 1/1 [00:00<00:00, 155.70it/s]
Original shapes - Output: (1875,), Label: (1875,)
Aligned shapes - Output: (1875,), Label: (1875,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1401.60it/s]
Table1-only metrics completed in 0.20s
Evaluation Result:  {'AUC-PR': 0.17211117659862812, 'AUC-ROC': 0.3744423853842087, 'VUS-PR': 0.1687006127801045, 'VUS-ROC': 0.37520542259694883, 'Standard-F1': 0.2194744976811417, 'Standard-Precision': 0.17359413202933943, 'Standard-Recall': 0.2983193277310912, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9621288154826436, 'Affiliation-P': 0.9777333333333333, 'Affiliation-R': 0.9470145658263305, 'F1_T': 0.21712538226253333, 'Precision_T': 0.1706730769230765, 'Recall_T': 0.2983193277310912}
Results for 026_NAB_id_26_Traffic_tr_624_1st_2261.csv: {'filename': '026_NAB_id_26_Traffic_tr_624_1st_2261.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 8, 'train_index': '624', 'data_shape': '2499x1', 'output_length': 1875, 'label_length': 1875, 'aligned_length': 1875, 'AUC-PR': 0.17211117659862812, 'AUC-ROC': 0.3744423853842087, 'VUS-PR': 0.1687006127801045, 'VUS-ROC': 0.37520542259694883, 'Standard-F1': 0.2194744976811417, 'Standard-Precision': 0.17359413202933943, 'Standard-Recall': 0.2983193277310912, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9621288154826436, 'Affiliation-P': 0.9777333333333333, 'Affiliation-R': 0.9470145658263305, 'F1_T': 0.21712538226253333, 'Precision_T': 0.1706730769230765, 'Recall_T': 0.2983193277310912}
Logits metrics skipped for 026_NAB_id_26_Traffic_tr_624_1st_2261.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/026_NAB_id_26_Traffic_tr_624_1st_2261_results.pkl
Processing file: 528_UCR_id_226_Medical_tr_7500_1st_17000.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 528_UCR_id_226_Medical_tr_7500_1st_17000.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 98.48it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 177/200 [00:01<00:00, 176.38it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 184.22it/s]
Table1-only metrics completed in 2.00s
Evaluation Result:  {'AUC-PR': 0.05534608416783964, 'AUC-ROC': 0.8565941964285715, 'VUS-PR': 0.04251763584757349, 'VUS-ROC': 0.9112200767569729, 'Standard-F1': 0.06083294337847882, 'Standard-Precision': 0.0319096710849288, 'Standard-Recall': 0.6499999999999936, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7151201997574622, 'Affiliation-P': 0.5573627019089574, 'Affiliation-R': 0.9974377777777778, 'F1_T': 0.06039255158520883, 'Precision_T': 0.03179650238473766, 'Recall_T': 0.599999999999994}
Results for 528_UCR_id_226_Medical_tr_7500_1st_17000.csv: {'filename': '528_UCR_id_226_Medical_tr_7500_1st_17000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 166, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.05534608416783964, 'AUC-ROC': 0.8565941964285715, 'VUS-PR': 0.04251763584757349, 'VUS-ROC': 0.9112200767569729, 'Standard-F1': 0.06083294337847882, 'Standard-Precision': 0.0319096710849288, 'Standard-Recall': 0.6499999999999936, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7151201997574622, 'Affiliation-P': 0.5573627019089574, 'Affiliation-R': 0.9974377777777778, 'F1_T': 0.06039255158520883, 'Precision_T': 0.03179650238473766, 'Recall_T': 0.599999999999994}
Logits metrics skipped for 528_UCR_id_226_Medical_tr_7500_1st_17000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/528_UCR_id_226_Medical_tr_7500_1st_17000_results.pkl
Skipping file: 197_SMD_id_20_Facility_tr_5000_1st_15144.csv due to filter criteria.
Processing file: 796_YAHOO_id_246_WebService_tr_500_1st_222.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 796_YAHOO_id_246_WebService_tr_500_1st_222.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 159.37it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1563.34it/s]
Table1-only metrics completed in 0.22s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9983022071307294, 'Affiliation-P': 0.9966101694915253, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Results for 796_YAHOO_id_246_WebService_tr_500_1st_222.csv: {'filename': '796_YAHOO_id_246_WebService_tr_500_1st_222.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9983022071307294, 'Affiliation-P': 0.9966101694915253, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 796_YAHOO_id_246_WebService_tr_500_1st_222.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/796_YAHOO_id_246_WebService_tr_500_1st_222_results.pkl
Processing file: 778_YAHOO_id_228_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 778_YAHOO_id_228_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 163.70it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1999.64it/s]
Table1-only metrics completed in 0.82s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 778_YAHOO_id_228_Synthetic_tr_500_1st_658.csv: {'filename': '778_YAHOO_id_228_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 128, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 778_YAHOO_id_228_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/778_YAHOO_id_228_Synthetic_tr_500_1st_658_results.pkl
Processing file: 427_UCR_id_125_Medical_tr_10403_1st_10995.csv, Data shape: (41612, 1)
Label shape: (41612,)
Train data shape: (10403, 1), Test data shape: (31209, 1), Label test shape: (31209,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 427_UCR_id_125_Medical_tr_10403_1st_10995.csv...
Using CUDA device: 2
Data shape: (31209, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 31209, window_size 5000, adding 3791 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([7, 5000, 1])
batched_mask shape: torch.Size([7, 5000])

100%|██████████| 1/1 [00:00<00:00, 135.08it/s]
Original shapes - Output: (35000,), Label: (31209,)
Aligned shapes - Output: (31209,), Label: (31209,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  48%|████▊     | 96/200 [00:01<00:01, 95.73it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 105.47it/s]
Table1-only metrics completed in 2.21s
Evaluation Result:  {'AUC-PR': 0.013646819893614295, 'AUC-ROC': 0.9426063949736005, 'VUS-PR': 0.016177624267090792, 'VUS-ROC': 0.9506261716766896, 'Standard-F1': 0.05202312138711037, 'Standard-Precision': 0.028753993610223547, 'Standard-Recall': 0.27272727272726444, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6764052527003159, 'Affiliation-P': 0.5110365193578145, 'Affiliation-R': 1.0, 'F1_T': 0.05202312138711037, 'Precision_T': 0.028753993610223547, 'Recall_T': 0.27272727272726444}
Results for 427_UCR_id_125_Medical_tr_10403_1st_10995.csv: {'filename': '427_UCR_id_125_Medical_tr_10403_1st_10995.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 33, 'train_index': '10403', 'data_shape': '41612x1', 'output_length': 35000, 'label_length': 31209, 'aligned_length': 31209, 'AUC-PR': 0.013646819893614295, 'AUC-ROC': 0.9426063949736005, 'VUS-PR': 0.016177624267090792, 'VUS-ROC': 0.9506261716766896, 'Standard-F1': 0.05202312138711037, 'Standard-Precision': 0.028753993610223547, 'Standard-Recall': 0.27272727272726444, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6764052527003159, 'Affiliation-P': 0.5110365193578145, 'Affiliation-R': 1.0, 'F1_T': 0.05202312138711037, 'Precision_T': 0.028753993610223547, 'Recall_T': 0.27272727272726444}
Logits metrics skipped for 427_UCR_id_125_Medical_tr_10403_1st_10995.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/427_UCR_id_125_Medical_tr_10403_1st_10995_results.pkl
Processing file: 420_UCR_id_118_Facility_tr_13750_1st_52600.csv, Data shape: (55000, 1)
Label shape: (55000,)
Train data shape: (13750, 1), Test data shape: (41250, 1), Label test shape: (41250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 420_UCR_id_118_Facility_tr_13750_1st_52600.csv...
Using CUDA device: 2
Data shape: (41250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 41250, window_size 5000, adding 3750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([9, 5000, 1])
batched_mask shape: torch.Size([9, 5000])

100%|██████████| 1/1 [00:00<00:00, 110.22it/s]
Original shapes - Output: (45000,), Label: (41250,)
Aligned shapes - Output: (41250,), Label: (41250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  48%|████▊     | 96/200 [00:01<00:01, 95.62it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 114.29it/s]
Table1-only metrics completed in 3.43s
Evaluation Result:  {'AUC-PR': 0.22410671857520723, 'AUC-ROC': 0.9302217417783192, 'VUS-PR': 0.1327615404059041, 'VUS-ROC': 0.9390038626970751, 'Standard-F1': 0.19902120717737373, 'Standard-Precision': 0.14769975786924902, 'Standard-Recall': 0.3049999999999985, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6849268404187255, 'Affiliation-P': 0.5210940711462451, 'Affiliation-R': 0.9990203636363636, 'F1_T': 0.19902120717737373, 'Precision_T': 0.14769975786924902, 'Recall_T': 0.3049999999999985}
Results for 420_UCR_id_118_Facility_tr_13750_1st_52600.csv: {'filename': '420_UCR_id_118_Facility_tr_13750_1st_52600.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 233, 'train_index': '13750', 'data_shape': '55000x1', 'output_length': 45000, 'label_length': 41250, 'aligned_length': 41250, 'AUC-PR': 0.22410671857520723, 'AUC-ROC': 0.9302217417783192, 'VUS-PR': 0.1327615404059041, 'VUS-ROC': 0.9390038626970751, 'Standard-F1': 0.19902120717737373, 'Standard-Precision': 0.14769975786924902, 'Standard-Recall': 0.3049999999999985, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6849268404187255, 'Affiliation-P': 0.5210940711462451, 'Affiliation-R': 0.9990203636363636, 'F1_T': 0.19902120717737373, 'Precision_T': 0.14769975786924902, 'Recall_T': 0.3049999999999985}
Logits metrics skipped for 420_UCR_id_118_Facility_tr_13750_1st_52600.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/420_UCR_id_118_Facility_tr_13750_1st_52600_results.pkl
Processing file: 803_YAHOO_id_253_WebService_tr_500_1st_58.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 803_YAHOO_id_253_WebService_tr_500_1st_58.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.25it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 853.01it/s]
Table1-only metrics completed in 0.81s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9935294223583143, 'VUS-ROC': 0.9999584598159071, 'Standard-F1': 0.9565217391298525, 'Standard-Precision': 0.9166666666665902, 'Standard-Recall': 0.9999999999999091, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996062992125979, 'Affiliation-P': 0.9992129083038175, 'Affiliation-R': 1.0, 'F1_T': 0.9565217391298525, 'Precision_T': 0.9166666666665902, 'Recall_T': 0.9999999999999091}
Results for 803_YAHOO_id_253_WebService_tr_500_1st_58.csv: {'filename': '803_YAHOO_id_253_WebService_tr_500_1st_58.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9935294223583143, 'VUS-ROC': 0.9999584598159071, 'Standard-F1': 0.9565217391298525, 'Standard-Precision': 0.9166666666665902, 'Standard-Recall': 0.9999999999999091, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996062992125979, 'Affiliation-P': 0.9992129083038175, 'Affiliation-R': 1.0, 'F1_T': 0.9565217391298525, 'Precision_T': 0.9166666666665902, 'Recall_T': 0.9999999999999091}
Logits metrics skipped for 803_YAHOO_id_253_WebService_tr_500_1st_58.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/803_YAHOO_id_253_WebService_tr_500_1st_58_results.pkl
Skipping file: 250_SVDB_id_14_Medical_tr_50000_1st_107354.csv due to filter criteria.
Skipping file: 850_OPPORTUNITY_id_9_HumanActivity_tr_895_1st_995.csv due to filter criteria.
Processing file: 092_WSD_id_64_WebService_tr_500_1st_118.csv, Data shape: (18242, 1)
Label shape: (18242,)
Train data shape: (500, 1), Test data shape: (17742, 1), Label test shape: (17742,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 092_WSD_id_64_WebService_tr_500_1st_118.csv...
Using CUDA device: 2
Data shape: (17742, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 17742, window_size 5000, adding 2258 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.34it/s]
Original shapes - Output: (20000,), Label: (17742,)
Aligned shapes - Output: (17742,), Label: (17742,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 206.76it/s]
Table1-only metrics completed in 2.25s
Evaluation Result:  {'AUC-PR': 0.549022813469066, 'AUC-ROC': 0.9717644898881482, 'VUS-PR': 0.27977342276625555, 'VUS-ROC': 0.9932323942037592, 'Standard-F1': 0.21100917431162505, 'Standard-Precision': 0.1292134831460667, 'Standard-Recall': 0.5749999999999856, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9949772622841812, 'Affiliation-P': 0.9906509823366346, 'Affiliation-R': 0.9993414946385565, 'F1_T': 0.21100917431162505, 'Precision_T': 0.1292134831460667, 'Recall_T': 0.5749999999999856}
Results for 092_WSD_id_64_WebService_tr_500_1st_118.csv: {'filename': '092_WSD_id_64_WebService_tr_500_1st_118.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '18242x1', 'output_length': 20000, 'label_length': 17742, 'aligned_length': 17742, 'AUC-PR': 0.549022813469066, 'AUC-ROC': 0.9717644898881482, 'VUS-PR': 0.27977342276625555, 'VUS-ROC': 0.9932323942037592, 'Standard-F1': 0.21100917431162505, 'Standard-Precision': 0.1292134831460667, 'Standard-Recall': 0.5749999999999856, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9949772622841812, 'Affiliation-P': 0.9906509823366346, 'Affiliation-R': 0.9993414946385565, 'F1_T': 0.21100917431162505, 'Precision_T': 0.1292134831460667, 'Recall_T': 0.5749999999999856}
Logits metrics skipped for 092_WSD_id_64_WebService_tr_500_1st_118.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/092_WSD_id_64_WebService_tr_500_1st_118_results.pkl
Skipping file: 862_OPPORTUNITY_id_21_HumanActivity_tr_500_1st_566.csv due to filter criteria.
Processing file: 128_WSD_id_100_WebService_tr_4068_1st_4168.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4068, 1), Test data shape: (14168, 1), Label test shape: (14168,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 128_WSD_id_100_WebService_tr_4068_1st_4168.csv...
Using CUDA device: 2
Data shape: (14168, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 14168, window_size 5000, adding 832 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.17it/s]
Original shapes - Output: (15000,), Label: (14168,)
Aligned shapes - Output: (14168,), Label: (14168,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 276.72it/s]
Table1-only metrics completed in 1.94s
Evaluation Result:  {'AUC-PR': 0.2674025734286249, 'AUC-ROC': 0.855699813860252, 'VUS-PR': 0.21686152011386262, 'VUS-ROC': 0.8871891276444255, 'Standard-F1': 0.2807017543854777, 'Standard-Precision': 0.3380281690140821, 'Standard-Recall': 0.2399999999999988, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8492925219765111, 'Affiliation-P': 0.7503117413779973, 'Affiliation-R': 0.9783569175659866, 'F1_T': 0.2807017543854777, 'Precision_T': 0.3380281690140821, 'Recall_T': 0.2399999999999988}
Results for 128_WSD_id_100_WebService_tr_4068_1st_4168.csv: {'filename': '128_WSD_id_100_WebService_tr_4068_1st_4168.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4068', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 14168, 'aligned_length': 14168, 'AUC-PR': 0.2674025734286249, 'AUC-ROC': 0.855699813860252, 'VUS-PR': 0.21686152011386262, 'VUS-ROC': 0.8871891276444255, 'Standard-F1': 0.2807017543854777, 'Standard-Precision': 0.3380281690140821, 'Standard-Recall': 0.2399999999999988, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8492925219765111, 'Affiliation-P': 0.7503117413779973, 'Affiliation-R': 0.9783569175659866, 'F1_T': 0.2807017543854777, 'Precision_T': 0.3380281690140821, 'Recall_T': 0.2399999999999988}
Logits metrics skipped for 128_WSD_id_100_WebService_tr_4068_1st_4168.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/128_WSD_id_100_WebService_tr_4068_1st_4168_results.pkl
Processing file: 596_YAHOO_id_46_WebService_tr_500_1st_666.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 596_YAHOO_id_46_WebService_tr_500_1st_666.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 162.37it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1278.02it/s]
Table1-only metrics completed in 0.30s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9547193755169392, 'VUS-ROC': 0.9999037276007114, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9956942949407961, 'Affiliation-P': 0.9914255091103966, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Results for 596_YAHOO_id_46_WebService_tr_500_1st_666.csv: {'filename': '596_YAHOO_id_46_WebService_tr_500_1st_666.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 12, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9547193755169392, 'VUS-ROC': 0.9999037276007114, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9956942949407961, 'Affiliation-P': 0.9914255091103966, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 596_YAHOO_id_46_WebService_tr_500_1st_666.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/596_YAHOO_id_46_WebService_tr_500_1st_666_results.pkl
Skipping file: 145_MSL_id_6_Sensor_tr_530_1st_630.csv due to filter criteria.
Processing file: 335_UCR_id_33_Medical_tr_1875_1st_4526.csv, Data shape: (7501, 1)
Label shape: (7501,)
Train data shape: (1875, 1), Test data shape: (5626, 1), Label test shape: (5626,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 335_UCR_id_33_Medical_tr_1875_1st_4526.csv...
Using CUDA device: 2
Data shape: (5626, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5626, window_size 5000, adding 4374 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.73it/s]
Original shapes - Output: (10000,), Label: (5626,)
Aligned shapes - Output: (5626,), Label: (5626,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 687.87it/s]
Table1-only metrics completed in 0.87s
Evaluation Result:  {'AUC-PR': 0.952345260967888, 'AUC-ROC': 0.9997200381224683, 'VUS-PR': 0.9332217867899557, 'VUS-ROC': 0.9997252477251869, 'Standard-F1': 0.6896551724133253, 'Standard-Precision': 0.526315789473675, 'Standard-Recall': 0.9999999999999667, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9997966332811237, 'Affiliation-P': 0.9996199909288157, 'Affiliation-R': 0.9999733380732313, 'F1_T': 0.6896551724133253, 'Precision_T': 0.526315789473675, 'Recall_T': 0.9999999999999667}
Results for 335_UCR_id_33_Medical_tr_1875_1st_4526.csv: {'filename': '335_UCR_id_33_Medical_tr_1875_1st_4526.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 154, 'train_index': '1875', 'data_shape': '7501x1', 'output_length': 10000, 'label_length': 5626, 'aligned_length': 5626, 'AUC-PR': 0.952345260967888, 'AUC-ROC': 0.9997200381224683, 'VUS-PR': 0.9332217867899557, 'VUS-ROC': 0.9997252477251869, 'Standard-F1': 0.6896551724133253, 'Standard-Precision': 0.526315789473675, 'Standard-Recall': 0.9999999999999667, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9997966332811237, 'Affiliation-P': 0.9996199909288157, 'Affiliation-R': 0.9999733380732313, 'F1_T': 0.6896551724133253, 'Precision_T': 0.526315789473675, 'Recall_T': 0.9999999999999667}
Logits metrics skipped for 335_UCR_id_33_Medical_tr_1875_1st_4526.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/335_UCR_id_33_Medical_tr_1875_1st_4526_results.pkl
Skipping file: 828_Exathlon_id_19_Facility_tr_11663_1st_13482.csv due to filter criteria.
Skipping file: 180_SMD_id_3_Facility_tr_5925_1st_17580.csv due to filter criteria.
Processing file: 345_UCR_id_43_Facility_tr_25000_1st_44630.csv, Data shape: (100000, 1)
Label shape: (100000,)
Train data shape: (25000, 1), Test data shape: (75000, 1), Label test shape: (75000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 345_UCR_id_43_Facility_tr_25000_1st_44630.csv...
Using CUDA device: 2
Data shape: (75000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([15, 5000, 1])
batched_mask shape: torch.Size([15, 5000])

100%|██████████| 1/1 [00:00<00:00, 78.77it/s]
Original shapes - Output: (75000,), Label: (75000,)
Aligned shapes - Output: (75000,), Label: (75000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:01<00:02, 51.84it/s]

Calculating Affiliation Metrics:  54%|█████▎    | 107/200 [00:02<00:01, 53.43it/s]

Calculating Affiliation Metrics:  92%|█████████▏| 183/200 [00:03<00:00, 63.44it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:03<00:00, 63.13it/s]
Table1-only metrics completed in 5.21s
Evaluation Result:  {'AUC-PR': 0.030138258081660804, 'AUC-ROC': 0.8883138469772133, 'VUS-PR': 0.031700823664572865, 'VUS-ROC': 0.8961283391068405, 'Standard-F1': 0.08764519535351173, 'Standard-Precision': 0.05079559363525089, 'Standard-Recall': 0.319230769230768, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8502300402942744, 'Affiliation-P': 0.7400077276746242, 'Affiliation-R': 0.9990336153846153, 'F1_T': 0.08494572911729291, 'Precision_T': 0.048413125336202234, 'Recall_T': 0.3461538461538448}
Results for 345_UCR_id_43_Facility_tr_25000_1st_44630.csv: {'filename': '345_UCR_id_43_Facility_tr_25000_1st_44630.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '25000', 'data_shape': '100000x1', 'output_length': 75000, 'label_length': 75000, 'aligned_length': 75000, 'AUC-PR': 0.030138258081660804, 'AUC-ROC': 0.8883138469772133, 'VUS-PR': 0.031700823664572865, 'VUS-ROC': 0.8961283391068405, 'Standard-F1': 0.08764519535351173, 'Standard-Precision': 0.05079559363525089, 'Standard-Recall': 0.319230769230768, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8502300402942744, 'Affiliation-P': 0.7400077276746242, 'Affiliation-R': 0.9990336153846153, 'F1_T': 0.08494572911729291, 'Precision_T': 0.048413125336202234, 'Recall_T': 0.3461538461538448}
Logits metrics skipped for 345_UCR_id_43_Facility_tr_25000_1st_44630.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/345_UCR_id_43_Facility_tr_25000_1st_44630_results.pkl
Processing file: 347_UCR_id_45_Sensor_tr_2851_1st_5365.csv, Data shape: (11406, 1)
Label shape: (11406,)
Train data shape: (2851, 1), Test data shape: (8555, 1), Label test shape: (8555,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 347_UCR_id_45_Sensor_tr_2851_1st_5365.csv...
Using CUDA device: 2
Data shape: (8555, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 8555, window_size 5000, adding 1445 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.47it/s]
Original shapes - Output: (10000,), Label: (8555,)
Aligned shapes - Output: (8555,), Label: (8555,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 338.38it/s]
Table1-only metrics completed in 1.00s
Evaluation Result:  {'AUC-PR': 0.06803591078655684, 'AUC-ROC': 0.9765261514441843, 'VUS-PR': 0.08490122179091074, 'VUS-ROC': 0.9885724521102753, 'Standard-F1': 0.15254237288113143, 'Standard-Precision': 0.08737864077669819, 'Standard-Recall': 0.59999999999996, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9175261666803952, 'Affiliation-P': 0.8476701144894639, 'Affiliation-R': 0.9999298655756867, 'F1_T': 0.13986013985995013, 'Precision_T': 0.07812499999999939, 'Recall_T': 0.6666666666666222}
Results for 347_UCR_id_45_Sensor_tr_2851_1st_5365.csv: {'filename': '347_UCR_id_45_Sensor_tr_2851_1st_5365.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 98, 'train_index': '2851', 'data_shape': '11406x1', 'output_length': 10000, 'label_length': 8555, 'aligned_length': 8555, 'AUC-PR': 0.06803591078655684, 'AUC-ROC': 0.9765261514441843, 'VUS-PR': 0.08490122179091074, 'VUS-ROC': 0.9885724521102753, 'Standard-F1': 0.15254237288113143, 'Standard-Precision': 0.08737864077669819, 'Standard-Recall': 0.59999999999996, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9175261666803952, 'Affiliation-P': 0.8476701144894639, 'Affiliation-R': 0.9999298655756867, 'F1_T': 0.13986013985995013, 'Precision_T': 0.07812499999999939, 'Recall_T': 0.6666666666666222}
Logits metrics skipped for 347_UCR_id_45_Sensor_tr_2851_1st_5365.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/347_UCR_id_45_Sensor_tr_2851_1st_5365_results.pkl
Processing file: 464_UCR_id_162_Medical_tr_7487_1st_16390.csv, Data shape: (29950, 1)
Label shape: (29950,)
Train data shape: (7487, 1), Test data shape: (22463, 1), Label test shape: (22463,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 464_UCR_id_162_Medical_tr_7487_1st_16390.csv...
Using CUDA device: 2
Data shape: (22463, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22463, window_size 5000, adding 2537 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 148.16it/s]
Original shapes - Output: (25000,), Label: (22463,)
Aligned shapes - Output: (22463,), Label: (22463,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  54%|█████▎    | 107/200 [00:01<00:00, 106.47it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 119.62it/s]
Table1-only metrics completed in 1.92s
Evaluation Result:  {'AUC-PR': 0.8322775828658051, 'AUC-ROC': 0.9955170210552906, 'VUS-PR': 0.34362720303305094, 'VUS-ROC': 0.9978237362402078, 'Standard-F1': 0.21960784313704554, 'Standard-Precision': 0.1244444444444439, 'Standard-Recall': 0.9333333333333023, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7980944832032892, 'Affiliation-P': 0.6640272558334867, 'Affiliation-R': 0.9999933223523126, 'F1_T': 0.21960784313704554, 'Precision_T': 0.1244444444444439, 'Recall_T': 0.9333333333333023}
Results for 464_UCR_id_162_Medical_tr_7487_1st_16390.csv: {'filename': '464_UCR_id_162_Medical_tr_7487_1st_16390.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 33, 'train_index': '7487', 'data_shape': '29950x1', 'output_length': 25000, 'label_length': 22463, 'aligned_length': 22463, 'AUC-PR': 0.8322775828658051, 'AUC-ROC': 0.9955170210552906, 'VUS-PR': 0.34362720303305094, 'VUS-ROC': 0.9978237362402078, 'Standard-F1': 0.21960784313704554, 'Standard-Precision': 0.1244444444444439, 'Standard-Recall': 0.9333333333333023, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7980944832032892, 'Affiliation-P': 0.6640272558334867, 'Affiliation-R': 0.9999933223523126, 'F1_T': 0.21960784313704554, 'Precision_T': 0.1244444444444439, 'Recall_T': 0.9333333333333023}
Logits metrics skipped for 464_UCR_id_162_Medical_tr_7487_1st_16390.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/464_UCR_id_162_Medical_tr_7487_1st_16390_results.pkl
Processing file: 761_YAHOO_id_211_WebService_tr_500_1st_7.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 761_YAHOO_id_211_WebService_tr_500_1st_7.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.39it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 736.87it/s]
Table1-only metrics completed in 0.74s
Evaluation Result:  {'AUC-PR': 0.8336196186463567, 'AUC-ROC': 0.9913679135236021, 'VUS-PR': 0.8517561151369486, 'VUS-ROC': 0.9949023391402655, 'Standard-F1': 0.7826086956516067, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.8181818181817438, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9199036959171887, 'Affiliation-P': 0.9309767949030396, 'Affiliation-R': 0.9090909090909091, 'F1_T': 0.7826086956516067, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.8181818181817438}
Results for 761_YAHOO_id_211_WebService_tr_500_1st_7.csv: {'filename': '761_YAHOO_id_211_WebService_tr_500_1st_7.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.8336196186463567, 'AUC-ROC': 0.9913679135236021, 'VUS-PR': 0.8517561151369486, 'VUS-ROC': 0.9949023391402655, 'Standard-F1': 0.7826086956516067, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.8181818181817438, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9199036959171887, 'Affiliation-P': 0.9309767949030396, 'Affiliation-R': 0.9090909090909091, 'F1_T': 0.7826086956516067, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.8181818181817438}
Logits metrics skipped for 761_YAHOO_id_211_WebService_tr_500_1st_7.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/761_YAHOO_id_211_WebService_tr_500_1st_7_results.pkl
Skipping file: 255_SVDB_id_19_Medical_tr_50000_1st_54982.csv due to filter criteria.
Processing file: 577_YAHOO_id_27_WebService_tr_500_1st_498.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 577_YAHOO_id_27_WebService_tr_500_1st_498.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 166.14it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1242.05it/s]
Table1-only metrics completed in 0.42s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.999944492534466, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9991830399719834, 'Affiliation-P': 0.998367413701716, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Results for 577_YAHOO_id_27_WebService_tr_500_1st_498.csv: {'filename': '577_YAHOO_id_27_WebService_tr_500_1st_498.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.999944492534466, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9991830399719834, 'Affiliation-P': 0.998367413701716, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 577_YAHOO_id_27_WebService_tr_500_1st_498.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/577_YAHOO_id_27_WebService_tr_500_1st_498_results.pkl
Processing file: 087_WSD_id_59_WebService_tr_4559_1st_12600.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 087_WSD_id_59_WebService_tr_4559_1st_12600.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.04it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 319.06it/s]
Table1-only metrics completed in 0.73s
Evaluation Result:  {'AUC-PR': 0.2858999229590884, 'AUC-ROC': 0.8553344065564173, 'VUS-PR': 0.11323861581928867, 'VUS-ROC': 0.8795802627554873, 'Standard-F1': 0.05405405405391573, 'Standard-Precision': 0.02919708029197059, 'Standard-Recall': 0.36363636363633056, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7171755460602885, 'Affiliation-P': 0.5591117214661661, 'Affiliation-R': 0.9998338285243308, 'F1_T': 0.05405405405391573, 'Precision_T': 0.02919708029197059, 'Recall_T': 0.36363636363633056}
Results for 087_WSD_id_59_WebService_tr_4559_1st_12600.csv: {'filename': '087_WSD_id_59_WebService_tr_4559_1st_12600.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 10, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.2858999229590884, 'AUC-ROC': 0.8553344065564173, 'VUS-PR': 0.11323861581928867, 'VUS-ROC': 0.8795802627554873, 'Standard-F1': 0.05405405405391573, 'Standard-Precision': 0.02919708029197059, 'Standard-Recall': 0.36363636363633056, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7171755460602885, 'Affiliation-P': 0.5591117214661661, 'Affiliation-R': 0.9998338285243308, 'F1_T': 0.05405405405391573, 'Precision_T': 0.02919708029197059, 'Recall_T': 0.36363636363633056}
Logits metrics skipped for 087_WSD_id_59_WebService_tr_4559_1st_12600.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/087_WSD_id_59_WebService_tr_4559_1st_12600_results.pkl
Processing file: 395_UCR_id_93_Medical_tr_7516_1st_17390.csv, Data shape: (30066, 1)
Label shape: (30066,)
Train data shape: (7516, 1), Test data shape: (22550, 1), Label test shape: (22550,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 395_UCR_id_93_Medical_tr_7516_1st_17390.csv...
Using CUDA device: 2
Data shape: (22550, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22550, window_size 5000, adding 2450 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 152.61it/s]
Original shapes - Output: (25000,), Label: (22550,)
Aligned shapes - Output: (22550,), Label: (22550,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  57%|█████▋    | 114/200 [00:01<00:00, 113.29it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 127.22it/s]
Table1-only metrics completed in 1.83s
Evaluation Result:  {'AUC-PR': 0.06172416364039362, 'AUC-ROC': 0.8331458862279559, 'VUS-PR': 0.0633639435115404, 'VUS-ROC': 0.8518666031436837, 'Standard-F1': 0.17415730337032193, 'Standard-Precision': 0.13716814159291976, 'Standard-Recall': 0.23846153846153664, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7385638637625555, 'Affiliation-P': 0.5856862869589848, 'Affiliation-R': 0.9994409005628518, 'F1_T': 0.17415730337032193, 'Precision_T': 0.13716814159291976, 'Recall_T': 0.23846153846153664}
Results for 395_UCR_id_93_Medical_tr_7516_1st_17390.csv: {'filename': '395_UCR_id_93_Medical_tr_7516_1st_17390.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 34, 'train_index': '7516', 'data_shape': '30066x1', 'output_length': 25000, 'label_length': 22550, 'aligned_length': 22550, 'AUC-PR': 0.06172416364039362, 'AUC-ROC': 0.8331458862279559, 'VUS-PR': 0.0633639435115404, 'VUS-ROC': 0.8518666031436837, 'Standard-F1': 0.17415730337032193, 'Standard-Precision': 0.13716814159291976, 'Standard-Recall': 0.23846153846153664, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7385638637625555, 'Affiliation-P': 0.5856862869589848, 'Affiliation-R': 0.9994409005628518, 'F1_T': 0.17415730337032193, 'Precision_T': 0.13716814159291976, 'Recall_T': 0.23846153846153664}
Logits metrics skipped for 395_UCR_id_93_Medical_tr_7516_1st_17390.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/395_UCR_id_93_Medical_tr_7516_1st_17390_results.pkl
Processing file: 586_YAHOO_id_36_WebService_tr_500_1st_398.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 586_YAHOO_id_36_WebService_tr_500_1st_398.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 162.44it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Table1-only metrics completed in 0.01s
Evaluation Result:  {'AUC-PR': 0.0, 'AUC-ROC': nan, 'VUS-PR': 0.0, 'VUS-ROC': 0.0, 'Standard-F1': 0.0, 'Standard-Precision': 0.0, 'Standard-Recall': 0.0, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.0, 'Affiliation-P': 0.0, 'Affiliation-R': 0.0, 'F1_T': 0.0, 'Precision_T': 0.0, 'Recall_T': 0.0}
Results for 586_YAHOO_id_36_WebService_tr_500_1st_398.csv: {'filename': '586_YAHOO_id_36_WebService_tr_500_1st_398.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.0, 'AUC-ROC': nan, 'VUS-PR': 0.0, 'VUS-ROC': 0.0, 'Standard-F1': 0.0, 'Standard-Precision': 0.0, 'Standard-Recall': 0.0, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.0, 'Affiliation-P': 0.0, 'Affiliation-R': 0.0, 'F1_T': 0.0, 'Precision_T': 0.0, 'Recall_T': 0.0}
Logits metrics skipped for 586_YAHOO_id_36_WebService_tr_500_1st_398.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/586_YAHOO_id_36_WebService_tr_500_1st_398_results.pkl
Processing file: 733_YAHOO_id_183_WebService_tr_500_1st_447.csv, Data shape: (1461, 1)
Label shape: (1461,)
Train data shape: (500, 1), Test data shape: (961, 1), Label test shape: (961,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 733_YAHOO_id_183_WebService_tr_500_1st_447.csv...
Using CUDA device: 2
Data shape: (961, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 961, 1])
batched_mask shape: torch.Size([1, 961])

100%|██████████| 1/1 [00:00<00:00, 159.86it/s]
Original shapes - Output: (961,), Label: (961,)
Aligned shapes - Output: (961,), Label: (961,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1894.00it/s]
Table1-only metrics completed in 1.08s
Evaluation Result:  {'AUC-PR': 0.29103302402704406, 'AUC-ROC': 0.9216267942583731, 'VUS-PR': 0.20681402277628197, 'VUS-ROC': 0.9461526655567268, 'Standard-F1': 0.2857142857137597, 'Standard-Precision': 0.29999999999997, 'Standard-Recall': 0.27272727272724795, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.736883451756132, 'Affiliation-P': 0.5833851617102429, 'Affiliation-R': 1.0, 'F1_T': 0.2857142857137597, 'Precision_T': 0.29999999999997, 'Recall_T': 0.27272727272724795}
Results for 733_YAHOO_id_183_WebService_tr_500_1st_447.csv: {'filename': '733_YAHOO_id_183_WebService_tr_500_1st_447.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '1461x1', 'output_length': 961, 'label_length': 961, 'aligned_length': 961, 'AUC-PR': 0.29103302402704406, 'AUC-ROC': 0.9216267942583731, 'VUS-PR': 0.20681402277628197, 'VUS-ROC': 0.9461526655567268, 'Standard-F1': 0.2857142857137597, 'Standard-Precision': 0.29999999999997, 'Standard-Recall': 0.27272727272724795, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.736883451756132, 'Affiliation-P': 0.5833851617102429, 'Affiliation-R': 1.0, 'F1_T': 0.2857142857137597, 'Precision_T': 0.29999999999997, 'Recall_T': 0.27272727272724795}
Logits metrics skipped for 733_YAHOO_id_183_WebService_tr_500_1st_447.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/733_YAHOO_id_183_WebService_tr_500_1st_447_results.pkl
Processing file: 679_YAHOO_id_129_WebService_tr_500_1st_428.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 679_YAHOO_id_129_WebService_tr_500_1st_428.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.56it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1892.43it/s]
Table1-only metrics completed in 0.26s
Evaluation Result:  {'AUC-PR': 0.5099009900990099, 'AUC-ROC': 0.9579796264855688, 'VUS-PR': 0.31074903786454816, 'VUS-ROC': 0.9514274747734345, 'Standard-F1': 0.14285714285687756, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.49999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8091402110526743, 'Affiliation-P': 0.6794588402115114, 'Affiliation-R': 1.0, 'F1_T': 0.14285714285687756, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.49999999999975}
Results for 679_YAHOO_id_129_WebService_tr_500_1st_428.csv: {'filename': '679_YAHOO_id_129_WebService_tr_500_1st_428.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.5099009900990099, 'AUC-ROC': 0.9579796264855688, 'VUS-PR': 0.31074903786454816, 'VUS-ROC': 0.9514274747734345, 'Standard-F1': 0.14285714285687756, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.49999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8091402110526743, 'Affiliation-P': 0.6794588402115114, 'Affiliation-R': 1.0, 'F1_T': 0.14285714285687756, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.49999999999975}
Logits metrics skipped for 679_YAHOO_id_129_WebService_tr_500_1st_428.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/679_YAHOO_id_129_WebService_tr_500_1st_428_results.pkl
Skipping file: 141_MSL_id_2_Sensor_tr_500_1st_550.csv due to filter criteria.
Processing file: 561_YAHOO_id_11_WebService_tr_500_1st_359.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 561_YAHOO_id_11_WebService_tr_500_1st_359.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 160.23it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1581.61it/s]
Table1-only metrics completed in 0.23s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984439100297067, 'Affiliation-P': 0.9968926553672316, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Results for 561_YAHOO_id_11_WebService_tr_500_1st_359.csv: {'filename': '561_YAHOO_id_11_WebService_tr_500_1st_359.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984439100297067, 'Affiliation-P': 0.9968926553672316, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 561_YAHOO_id_11_WebService_tr_500_1st_359.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/561_YAHOO_id_11_WebService_tr_500_1st_359_results.pkl
Processing file: 130_WSD_id_102_WebService_tr_2362_1st_2462.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (2362, 1), Test data shape: (15874, 1), Label test shape: (15874,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 130_WSD_id_102_WebService_tr_2362_1st_2462.csv...
Using CUDA device: 2
Data shape: (15874, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 15874, window_size 5000, adding 4126 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.17it/s]
Original shapes - Output: (20000,), Label: (15874,)
Aligned shapes - Output: (15874,), Label: (15874,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 202.69it/s]
Table1-only metrics completed in 1.65s
Evaluation Result:  {'AUC-PR': 0.15934859812490293, 'AUC-ROC': 0.5669140987209742, 'VUS-PR': 0.11571372991925075, 'VUS-ROC': 0.6825207086278716, 'Standard-F1': 0.19209039547973009, 'Standard-Precision': 0.21383647798742006, 'Standard-Recall': 0.17435897435897346, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8756791002281086, 'Affiliation-P': 0.7833070448926123, 'Affiliation-R': 0.9927498514926495, 'F1_T': 0.19209039547973009, 'Precision_T': 0.21383647798742006, 'Recall_T': 0.17435897435897346}
Results for 130_WSD_id_102_WebService_tr_2362_1st_2462.csv: {'filename': '130_WSD_id_102_WebService_tr_2362_1st_2462.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 29, 'train_index': '2362', 'data_shape': '18236x1', 'output_length': 20000, 'label_length': 15874, 'aligned_length': 15874, 'AUC-PR': 0.15934859812490293, 'AUC-ROC': 0.5669140987209742, 'VUS-PR': 0.11571372991925075, 'VUS-ROC': 0.6825207086278716, 'Standard-F1': 0.19209039547973009, 'Standard-Precision': 0.21383647798742006, 'Standard-Recall': 0.17435897435897346, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8756791002281086, 'Affiliation-P': 0.7833070448926123, 'Affiliation-R': 0.9927498514926495, 'F1_T': 0.19209039547973009, 'Precision_T': 0.21383647798742006, 'Recall_T': 0.17435897435897346}
Logits metrics skipped for 130_WSD_id_102_WebService_tr_2362_1st_2462.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/130_WSD_id_102_WebService_tr_2362_1st_2462_results.pkl
Processing file: 567_YAHOO_id_17_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 567_YAHOO_id_17_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 159.16it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1793.89it/s]
Table1-only metrics completed in 0.52s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984775989560672, 'Affiliation-P': 0.9969598262757872, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 567_YAHOO_id_17_Synthetic_tr_500_1st_893.csv: {'filename': '567_YAHOO_id_17_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984775989560672, 'Affiliation-P': 0.9969598262757872, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 567_YAHOO_id_17_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/567_YAHOO_id_17_Synthetic_tr_500_1st_893_results.pkl
Processing file: 230_MGAB_id_6_Synthetic_tr_25000_1st_42441.csv, Data shape: (100000, 1)
Label shape: (100000,)
Train data shape: (25000, 1), Test data shape: (75000, 1), Label test shape: (75000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 230_MGAB_id_6_Synthetic_tr_25000_1st_42441.csv...
Using CUDA device: 2
Data shape: (75000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([15, 5000, 1])
batched_mask shape: torch.Size([15, 5000])

100%|██████████| 1/1 [00:00<00:00, 79.91it/s]
Original shapes - Output: (75000,), Label: (75000,)
Aligned shapes - Output: (75000,), Label: (75000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  18%|█▊        | 35/200 [00:01<00:04, 33.74it/s]

Calculating Affiliation Metrics:  34%|███▍      | 69/200 [00:02<00:05, 25.80it/s]

Calculating Affiliation Metrics:  48%|████▊     | 96/200 [00:03<00:04, 23.31it/s]

Calculating Affiliation Metrics:  60%|██████    | 120/200 [00:05<00:03, 22.53it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [00:06<00:02, 23.28it/s]

Calculating Affiliation Metrics:  90%|████████▉ | 179/200 [00:07<00:00, 26.54it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:07<00:00, 27.20it/s]
Table1-only metrics completed in 9.27s
Evaluation Result:  {'AUC-PR': 0.017270402871159093, 'AUC-ROC': 0.6738053810160428, 'VUS-PR': 0.020846562684206708, 'VUS-ROC': 0.6618044745546936, 'Standard-F1': 0.05684210526282536, 'Standard-Precision': 0.03599999999999995, 'Standard-Recall': 0.13499999999999931, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7175031449839613, 'Affiliation-P': 0.6724558849355744, 'Affiliation-R': 0.7690190940409899, 'F1_T': 0.05684210526282536, 'Precision_T': 0.03599999999999995, 'Recall_T': 0.13499999999999931}
Results for 230_MGAB_id_6_Synthetic_tr_25000_1st_42441.csv: {'filename': '230_MGAB_id_6_Synthetic_tr_25000_1st_42441.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 49, 'train_index': '25000', 'data_shape': '100000x1', 'output_length': 75000, 'label_length': 75000, 'aligned_length': 75000, 'AUC-PR': 0.017270402871159093, 'AUC-ROC': 0.6738053810160428, 'VUS-PR': 0.020846562684206708, 'VUS-ROC': 0.6618044745546936, 'Standard-F1': 0.05684210526282536, 'Standard-Precision': 0.03599999999999995, 'Standard-Recall': 0.13499999999999931, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7175031449839613, 'Affiliation-P': 0.6724558849355744, 'Affiliation-R': 0.7690190940409899, 'F1_T': 0.05684210526282536, 'Precision_T': 0.03599999999999995, 'Recall_T': 0.13499999999999931}
Logits metrics skipped for 230_MGAB_id_6_Synthetic_tr_25000_1st_42441.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/230_MGAB_id_6_Synthetic_tr_25000_1st_42441_results.pkl
Processing file: 771_YAHOO_id_221_WebService_tr_500_1st_61.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 771_YAHOO_id_221_WebService_tr_500_1st_61.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.25it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 712.59it/s]
Table1-only metrics completed in 0.81s
Evaluation Result:  {'AUC-PR': 0.9999999999999999, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.9090909090903305, 'Standard-Precision': 0.8333333333332639, 'Standard-Recall': 0.9999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9976730735604659, 'Affiliation-P': 0.9953569511540525, 'Affiliation-R': 1.0, 'F1_T': 0.9090909090903305, 'Precision_T': 0.8333333333332639, 'Recall_T': 0.9999999999999}
Results for 771_YAHOO_id_221_WebService_tr_500_1st_61.csv: {'filename': '771_YAHOO_id_221_WebService_tr_500_1st_61.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9999999999999999, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.9090909090903305, 'Standard-Precision': 0.8333333333332639, 'Standard-Recall': 0.9999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9976730735604659, 'Affiliation-P': 0.9953569511540525, 'Affiliation-R': 1.0, 'F1_T': 0.9090909090903305, 'Precision_T': 0.8333333333332639, 'Recall_T': 0.9999999999999}
Logits metrics skipped for 771_YAHOO_id_221_WebService_tr_500_1st_61.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/771_YAHOO_id_221_WebService_tr_500_1st_61_results.pkl
Skipping file: 196_SMD_id_19_Facility_tr_7174_1st_21230.csv due to filter criteria.
Processing file: 467_UCR_id_165_Environment_tr_6166_1st_19280.csv, Data shape: (24667, 1)
Label shape: (24667,)
Train data shape: (6166, 1), Test data shape: (18501, 1), Label test shape: (18501,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 467_UCR_id_165_Environment_tr_6166_1st_19280.csv...
Using CUDA device: 2
Data shape: (18501, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 18501, window_size 5000, adding 1499 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.89it/s]
Original shapes - Output: (20000,), Label: (18501,)
Aligned shapes - Output: (18501,), Label: (18501,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 247.58it/s]
Table1-only metrics completed in 1.10s
Evaluation Result:  {'AUC-PR': 0.7456098541623999, 'AUC-ROC': 0.9831821465568943, 'VUS-PR': 0.7016995092747919, 'VUS-ROC': 0.9864286621192587, 'Standard-F1': 0.6589595375717533, 'Standard-Precision': 0.6129032258064483, 'Standard-Recall': 0.7124999999999956, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9997671036282412, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9995343157126643, 'F1_T': 0.6589595375717533, 'Precision_T': 0.6129032258064483, 'Recall_T': 0.7124999999999956}
Results for 467_UCR_id_165_Environment_tr_6166_1st_19280.csv: {'filename': '467_UCR_id_165_Environment_tr_6166_1st_19280.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 46, 'train_index': '6166', 'data_shape': '24667x1', 'output_length': 20000, 'label_length': 18501, 'aligned_length': 18501, 'AUC-PR': 0.7456098541623999, 'AUC-ROC': 0.9831821465568943, 'VUS-PR': 0.7016995092747919, 'VUS-ROC': 0.9864286621192587, 'Standard-F1': 0.6589595375717533, 'Standard-Precision': 0.6129032258064483, 'Standard-Recall': 0.7124999999999956, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9997671036282412, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9995343157126643, 'F1_T': 0.6589595375717533, 'Precision_T': 0.6129032258064483, 'Recall_T': 0.7124999999999956}
Logits metrics skipped for 467_UCR_id_165_Environment_tr_6166_1st_19280.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/467_UCR_id_165_Environment_tr_6166_1st_19280_results.pkl
Processing file: 156_Stock_id_8_Finance_tr_500_1st_24.csv, Data shape: (10000, 1)
Label shape: (10000,)
Train data shape: (500, 1), Test data shape: (9500, 1), Label test shape: (9500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 156_Stock_id_8_Finance_tr_500_1st_24.csv...
Using CUDA device: 2
Data shape: (9500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.58it/s]
Original shapes - Output: (10000,), Label: (9500,)
Aligned shapes - Output: (9500,), Label: (9500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  10%|█         | 20/200 [00:01<00:09, 19.63it/s]

Calculating Affiliation Metrics:  20%|██        | 40/200 [00:03<00:13, 11.96it/s]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:05<00:15,  9.65it/s]

Calculating Affiliation Metrics:  32%|███▎      | 65/200 [00:06<00:15,  8.55it/s]

Calculating Affiliation Metrics:  38%|███▊      | 75/200 [00:08<00:15,  7.84it/s]

Calculating Affiliation Metrics:  42%|████▏     | 84/200 [00:09<00:15,  7.63it/s]

Calculating Affiliation Metrics:  46%|████▌     | 92/200 [00:10<00:14,  7.52it/s]

Calculating Affiliation Metrics:  50%|█████     | 100/200 [00:11<00:13,  7.48it/s]

Calculating Affiliation Metrics:  54%|█████▍    | 108/200 [00:12<00:12,  7.56it/s]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [00:13<00:10,  8.10it/s]

Calculating Affiliation Metrics:  64%|██████▍   | 129/200 [00:14<00:08,  8.75it/s]

Calculating Affiliation Metrics:  70%|███████   | 140/200 [00:15<00:06,  9.20it/s]

Calculating Affiliation Metrics:  76%|███████▋  | 153/200 [00:16<00:04, 10.15it/s]

Calculating Affiliation Metrics:  86%|████████▋ | 173/200 [00:18<00:02, 12.88it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:18<00:00, 10.86it/s]
Table1-only metrics completed in 23.73s
Evaluation Result:  {'AUC-PR': 0.06543821407768097, 'AUC-ROC': 0.5702546292942493, 'VUS-PR': 0.14144531543466426, 'VUS-ROC': 0.5439036933814786, 'Standard-F1': 0.11753554502344908, 'Standard-Precision': 0.06871074990764682, 'Standard-Recall': 0.40611353711790305, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6767488578791832, 'Affiliation-P': 0.5138641378388557, 'Affiliation-R': 0.9908186940669539, 'F1_T': 0.117391304347582, 'Precision_T': 0.0684286748732802, 'Recall_T': 0.41266375545851436}
Results for 156_Stock_id_8_Finance_tr_500_1st_24.csv: {'filename': '156_Stock_id_8_Finance_tr_500_1st_24.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 6, 'train_index': '500', 'data_shape': '10000x1', 'output_length': 10000, 'label_length': 9500, 'aligned_length': 9500, 'AUC-PR': 0.06543821407768097, 'AUC-ROC': 0.5702546292942493, 'VUS-PR': 0.14144531543466426, 'VUS-ROC': 0.5439036933814786, 'Standard-F1': 0.11753554502344908, 'Standard-Precision': 0.06871074990764682, 'Standard-Recall': 0.40611353711790305, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6767488578791832, 'Affiliation-P': 0.5138641378388557, 'Affiliation-R': 0.9908186940669539, 'F1_T': 0.117391304347582, 'Precision_T': 0.0684286748732802, 'Recall_T': 0.41266375545851436}
Logits metrics skipped for 156_Stock_id_8_Finance_tr_500_1st_24.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/156_Stock_id_8_Finance_tr_500_1st_24_results.pkl
Processing file: 123_WSD_id_95_WebService_tr_4559_1st_4901.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 123_WSD_id_95_WebService_tr_4559_1st_4901.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.87it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 248.27it/s]
Table1-only metrics completed in 2.73s
Evaluation Result:  {'AUC-PR': 0.16367816391668874, 'AUC-ROC': 0.6608712297078252, 'VUS-PR': 0.04260195968265214, 'VUS-ROC': 0.919156744865971, 'Standard-F1': 0.0693641618493806, 'Standard-Precision': 0.04379562043795589, 'Standard-Recall': 0.16666666666666202, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7301813448685679, 'Affiliation-P': 0.5755314403188815, 'Affiliation-R': 0.998481248026618, 'F1_T': 0.0693641618493806, 'Precision_T': 0.04379562043795589, 'Recall_T': 0.16666666666666202}
Results for 123_WSD_id_95_WebService_tr_4559_1st_4901.csv: {'filename': '123_WSD_id_95_WebService_tr_4559_1st_4901.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.16367816391668874, 'AUC-ROC': 0.6608712297078252, 'VUS-PR': 0.04260195968265214, 'VUS-ROC': 0.919156744865971, 'Standard-F1': 0.0693641618493806, 'Standard-Precision': 0.04379562043795589, 'Standard-Recall': 0.16666666666666202, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7301813448685679, 'Affiliation-P': 0.5755314403188815, 'Affiliation-R': 0.998481248026618, 'F1_T': 0.0693641618493806, 'Precision_T': 0.04379562043795589, 'Recall_T': 0.16666666666666202}
Logits metrics skipped for 123_WSD_id_95_WebService_tr_4559_1st_4901.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/123_WSD_id_95_WebService_tr_4559_1st_4901_results.pkl
Processing file: 105_WSD_id_77_WebService_tr_2348_1st_2448.csv, Data shape: (18380, 1)
Label shape: (18380,)
Train data shape: (2348, 1), Test data shape: (16032, 1), Label test shape: (16032,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 105_WSD_id_77_WebService_tr_2348_1st_2448.csv...
Using CUDA device: 2
Data shape: (16032, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 16032, window_size 5000, adding 3968 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.60it/s]
Original shapes - Output: (20000,), Label: (16032,)
Aligned shapes - Output: (16032,), Label: (16032,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  76%|███████▌  | 152/200 [00:01<00:00, 151.85it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 170.19it/s]
Table1-only metrics completed in 3.98s
Evaluation Result:  {'AUC-PR': 0.403527760748173, 'AUC-ROC': 0.913034952309237, 'VUS-PR': 0.6995294086259158, 'VUS-ROC': 0.9912157631130754, 'Standard-F1': 0.3268482490267668, 'Standard-Precision': 0.2608695652173897, 'Standard-Recall': 0.43749999999999545, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9946029275230953, 'Affiliation-P': 0.9918307807104836, 'Affiliation-R': 0.9973906139562874, 'F1_T': 0.3268482490267668, 'Precision_T': 0.2608695652173897, 'Recall_T': 0.43749999999999545}
Results for 105_WSD_id_77_WebService_tr_2348_1st_2448.csv: {'filename': '105_WSD_id_77_WebService_tr_2348_1st_2448.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '2348', 'data_shape': '18380x1', 'output_length': 20000, 'label_length': 16032, 'aligned_length': 16032, 'AUC-PR': 0.403527760748173, 'AUC-ROC': 0.913034952309237, 'VUS-PR': 0.6995294086259158, 'VUS-ROC': 0.9912157631130754, 'Standard-F1': 0.3268482490267668, 'Standard-Precision': 0.2608695652173897, 'Standard-Recall': 0.43749999999999545, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9946029275230953, 'Affiliation-P': 0.9918307807104836, 'Affiliation-R': 0.9973906139562874, 'F1_T': 0.3268482490267668, 'Precision_T': 0.2608695652173897, 'Recall_T': 0.43749999999999545}
Logits metrics skipped for 105_WSD_id_77_WebService_tr_2348_1st_2448.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/105_WSD_id_77_WebService_tr_2348_1st_2448_results.pkl
Skipping file: 861_OPPORTUNITY_id_20_HumanActivity_tr_2085_1st_2185.csv due to filter criteria.
Processing file: 656_YAHOO_id_106_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 656_YAHOO_id_106_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 158.59it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1897.27it/s]
Table1-only metrics completed in 1.14s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 656_YAHOO_id_106_Synthetic_tr_500_1st_658.csv: {'filename': '656_YAHOO_id_106_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 235, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 656_YAHOO_id_106_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/656_YAHOO_id_106_Synthetic_tr_500_1st_658_results.pkl
Skipping file: 812_Exathlon_id_3_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 306_UCR_id_4_Medical_tr_2507_1st_8500.csv, Data shape: (10028, 1)
Label shape: (10028,)
Train data shape: (2507, 1), Test data shape: (7521, 1), Label test shape: (7521,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 306_UCR_id_4_Medical_tr_2507_1st_8500.csv...
Using CUDA device: 2
Data shape: (7521, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 7521, window_size 5000, adding 2479 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.68it/s]
Original shapes - Output: (10000,), Label: (7521,)
Aligned shapes - Output: (7521,), Label: (7521,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 498.66it/s]
Table1-only metrics completed in 0.54s
Evaluation Result:  {'AUC-PR': 0.0024630541871921183, 'AUC-ROC': 0.9461436170212766, 'VUS-PR': 0.10640343445313466, 'VUS-ROC': 0.9813382862182863, 'Standard-F1': 0.004807692307687487, 'Standard-Precision': 0.0024096385542168616, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.869041687135587, 'Affiliation-P': 0.7692776016627129, 'Affiliation-R': 0.9985374285334397, 'F1_T': 0.004878048780482914, 'Precision_T': 0.0024449877750611186, 'Recall_T': 0.9999999999989999}
Results for 306_UCR_id_4_Medical_tr_2507_1st_8500.csv: {'filename': '306_UCR_id_4_Medical_tr_2507_1st_8500.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 26, 'train_index': '2507', 'data_shape': '10028x1', 'output_length': 10000, 'label_length': 7521, 'aligned_length': 7521, 'AUC-PR': 0.0024630541871921183, 'AUC-ROC': 0.9461436170212766, 'VUS-PR': 0.10640343445313466, 'VUS-ROC': 0.9813382862182863, 'Standard-F1': 0.004807692307687487, 'Standard-Precision': 0.0024096385542168616, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.869041687135587, 'Affiliation-P': 0.7692776016627129, 'Affiliation-R': 0.9985374285334397, 'F1_T': 0.004878048780482914, 'Precision_T': 0.0024449877750611186, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 306_UCR_id_4_Medical_tr_2507_1st_8500.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/306_UCR_id_4_Medical_tr_2507_1st_8500_results.pkl
Skipping file: 203_SMD_id_26_Facility_tr_7174_1st_21230.csv due to filter criteria.
Processing file: 458_UCR_id_156_Facility_tr_20000_1st_56600.csv, Data shape: (80001, 1)
Label shape: (80001,)
Train data shape: (20000, 1), Test data shape: (60001, 1), Label test shape: (60001,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 458_UCR_id_156_Facility_tr_20000_1st_56600.csv...
Using CUDA device: 2
Data shape: (60001, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 60001, window_size 5000, adding 4999 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([13, 5000, 1])
batched_mask shape: torch.Size([13, 5000])

100%|██████████| 1/1 [00:00<00:00, 86.13it/s]
Original shapes - Output: (65000,), Label: (60001,)
Aligned shapes - Output: (60001,), Label: (60001,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  32%|███▎      | 65/200 [00:01<00:02, 64.70it/s]

Calculating Affiliation Metrics:  71%|███████   | 142/200 [00:02<00:00, 71.91it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 80.00it/s]
Table1-only metrics completed in 4.67s
Evaluation Result:  {'AUC-PR': 0.010008484718140113, 'AUC-ROC': 0.7711199700730863, 'VUS-PR': 0.0102717177331431, 'VUS-ROC': 0.7717307069296963, 'Standard-F1': 0.031651594478784545, 'Standard-Precision': 0.01611339956384783, 'Standard-Recall': 0.8866666666666636, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7385209500926024, 'Affiliation-P': 0.5940656231888409, 'Affiliation-R': 0.9758004033266112, 'F1_T': 0.03163867713643666, 'Precision_T': 0.01609611010672421, 'Recall_T': 0.9199999999999968}
Results for 458_UCR_id_156_Facility_tr_20000_1st_56600.csv: {'filename': '458_UCR_id_156_Facility_tr_20000_1st_56600.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 208, 'train_index': '20000', 'data_shape': '80001x1', 'output_length': 65000, 'label_length': 60001, 'aligned_length': 60001, 'AUC-PR': 0.010008484718140113, 'AUC-ROC': 0.7711199700730863, 'VUS-PR': 0.0102717177331431, 'VUS-ROC': 0.7717307069296963, 'Standard-F1': 0.031651594478784545, 'Standard-Precision': 0.01611339956384783, 'Standard-Recall': 0.8866666666666636, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7385209500926024, 'Affiliation-P': 0.5940656231888409, 'Affiliation-R': 0.9758004033266112, 'F1_T': 0.03163867713643666, 'Precision_T': 0.01609611010672421, 'Recall_T': 0.9199999999999968}
Logits metrics skipped for 458_UCR_id_156_Facility_tr_20000_1st_56600.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/458_UCR_id_156_Facility_tr_20000_1st_56600_results.pkl
Skipping file: 858_OPPORTUNITY_id_17_HumanActivity_tr_808_1st_908.csv due to filter criteria.
Processing file: 129_WSD_id_101_WebService_tr_4448_1st_9592.csv, Data shape: (17794, 1)
Label shape: (17794,)
Train data shape: (4448, 1), Test data shape: (13346, 1), Label test shape: (13346,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 129_WSD_id_101_WebService_tr_4448_1st_9592.csv...
Using CUDA device: 2
Data shape: (13346, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13346, window_size 5000, adding 1654 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 157.00it/s]
Original shapes - Output: (15000,), Label: (13346,)
Aligned shapes - Output: (13346,), Label: (13346,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 247.48it/s]
Table1-only metrics completed in 0.95s
Evaluation Result:  {'AUC-PR': 0.8605525215157961, 'AUC-ROC': 0.9960303166741709, 'VUS-PR': 0.4508465763708424, 'VUS-ROC': 0.9980996267858907, 'Standard-F1': 0.2337662337660047, 'Standard-Precision': 0.13432835820895422, 'Standard-Recall': 0.8999999999999551, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8887881624396531, 'Affiliation-P': 0.7998485178955432, 'Affiliation-R': 0.9999817432283389, 'F1_T': 0.2337662337660047, 'Precision_T': 0.13432835820895422, 'Recall_T': 0.8999999999999551}
Results for 129_WSD_id_101_WebService_tr_4448_1st_9592.csv: {'filename': '129_WSD_id_101_WebService_tr_4448_1st_9592.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 11, 'train_index': '4448', 'data_shape': '17794x1', 'output_length': 15000, 'label_length': 13346, 'aligned_length': 13346, 'AUC-PR': 0.8605525215157961, 'AUC-ROC': 0.9960303166741709, 'VUS-PR': 0.4508465763708424, 'VUS-ROC': 0.9980996267858907, 'Standard-F1': 0.2337662337660047, 'Standard-Precision': 0.13432835820895422, 'Standard-Recall': 0.8999999999999551, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8887881624396531, 'Affiliation-P': 0.7998485178955432, 'Affiliation-R': 0.9999817432283389, 'F1_T': 0.2337662337660047, 'Precision_T': 0.13432835820895422, 'Recall_T': 0.8999999999999551}
Logits metrics skipped for 129_WSD_id_101_WebService_tr_4448_1st_9592.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/129_WSD_id_101_WebService_tr_4448_1st_9592_results.pkl
Processing file: 759_YAHOO_id_209_WebService_tr_500_1st_189.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 759_YAHOO_id_209_WebService_tr_500_1st_189.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 160.50it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 812.24it/s]
Table1-only metrics completed in 0.81s
Evaluation Result:  {'AUC-PR': 0.9999999999999999, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.9090909090903305, 'Standard-Precision': 0.8333333333332639, 'Standard-Recall': 0.9999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9993013516962087, 'Affiliation-P': 0.9986036789297659, 'Affiliation-R': 1.0, 'F1_T': 0.9090909090903305, 'Precision_T': 0.8333333333332639, 'Recall_T': 0.9999999999999}
Results for 759_YAHOO_id_209_WebService_tr_500_1st_189.csv: {'filename': '759_YAHOO_id_209_WebService_tr_500_1st_189.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9999999999999999, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.9090909090903305, 'Standard-Precision': 0.8333333333332639, 'Standard-Recall': 0.9999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9993013516962087, 'Affiliation-P': 0.9986036789297659, 'Affiliation-R': 1.0, 'F1_T': 0.9090909090903305, 'Precision_T': 0.8333333333332639, 'Recall_T': 0.9999999999999}
Logits metrics skipped for 759_YAHOO_id_209_WebService_tr_500_1st_189.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/759_YAHOO_id_209_WebService_tr_500_1st_189_results.pkl
Processing file: 303_UCR_id_1_Medical_tr_2499_1st_8000.csv, Data shape: (9998, 1)
Label shape: (9998,)
Train data shape: (2499, 1), Test data shape: (7499, 1), Label test shape: (7499,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 303_UCR_id_1_Medical_tr_2499_1st_8000.csv...
Using CUDA device: 2
Data shape: (7499, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 7499, window_size 5000, adding 2501 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 167.69it/s]
Original shapes - Output: (10000,), Label: (7499,)
Aligned shapes - Output: (7499,), Label: (7499,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 470.69it/s]
Table1-only metrics completed in 0.56s
Evaluation Result:  {'AUC-PR': 0.06871939454659813, 'AUC-ROC': 0.9785656944072787, 'VUS-PR': 0.08358723095310445, 'VUS-ROC': 0.9790572023160219, 'Standard-F1': 0.19083969465631445, 'Standard-Precision': 0.1054852320675101, 'Standard-Recall': 0.99999999999996, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9043657052181394, 'Affiliation-P': 0.825426613173147, 'Affiliation-R': 1.0, 'F1_T': 0.1854838709675591, 'Precision_T': 0.10313901345291433, 'Recall_T': 0.9199999999999633}
Results for 303_UCR_id_1_Medical_tr_2499_1st_8000.csv: {'filename': '303_UCR_id_1_Medical_tr_2499_1st_8000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 26, 'train_index': '2499', 'data_shape': '9998x1', 'output_length': 10000, 'label_length': 7499, 'aligned_length': 7499, 'AUC-PR': 0.06871939454659813, 'AUC-ROC': 0.9785656944072787, 'VUS-PR': 0.08358723095310445, 'VUS-ROC': 0.9790572023160219, 'Standard-F1': 0.19083969465631445, 'Standard-Precision': 0.1054852320675101, 'Standard-Recall': 0.99999999999996, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9043657052181394, 'Affiliation-P': 0.825426613173147, 'Affiliation-R': 1.0, 'F1_T': 0.1854838709675591, 'Precision_T': 0.10313901345291433, 'Recall_T': 0.9199999999999633}
Logits metrics skipped for 303_UCR_id_1_Medical_tr_2499_1st_8000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/303_UCR_id_1_Medical_tr_2499_1st_8000_results.pkl
Processing file: 155_Stock_id_7_Finance_tr_500_1st_18.csv, Data shape: (20000, 1)
Label shape: (20000,)
Train data shape: (500, 1), Test data shape: (19500, 1), Label test shape: (19500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 155_Stock_id_7_Finance_tr_500_1st_18.csv...
Using CUDA device: 2
Data shape: (19500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 19500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.64it/s]
Original shapes - Output: (20000,), Label: (19500,)
Aligned shapes - Output: (19500,), Label: (19500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   4%|▍         | 9/200 [00:01<00:23,  8.20it/s]

Calculating Affiliation Metrics:   9%|▉         | 18/200 [00:03<00:39,  4.56it/s]

Calculating Affiliation Metrics:  12%|█▏        | 24/200 [00:05<00:46,  3.76it/s]

Calculating Affiliation Metrics:  14%|█▍        | 29/200 [00:07<00:51,  3.29it/s]

Calculating Affiliation Metrics:  16%|█▋        | 33/200 [00:09<00:54,  3.07it/s]

Calculating Affiliation Metrics:  18%|█▊        | 37/200 [00:10<00:56,  2.86it/s]

Calculating Affiliation Metrics:  20%|██        | 40/200 [00:12<00:59,  2.68it/s]

Calculating Affiliation Metrics:  22%|██▏       | 43/200 [00:13<01:02,  2.51it/s]

Calculating Affiliation Metrics:  23%|██▎       | 46/200 [00:15<01:06,  2.33it/s]

Calculating Affiliation Metrics:  24%|██▍       | 49/200 [00:16<01:09,  2.17it/s]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:18<01:12,  2.04it/s]

Calculating Affiliation Metrics:  28%|██▊       | 55/200 [00:20<01:14,  1.96it/s]

Calculating Affiliation Metrics:  28%|██▊       | 57/200 [00:21<01:14,  1.91it/s]

Calculating Affiliation Metrics:  30%|██▉       | 59/200 [00:22<01:15,  1.86it/s]

Calculating Affiliation Metrics:  30%|███       | 61/200 [00:23<01:16,  1.81it/s]

Calculating Affiliation Metrics:  32%|███▏      | 63/200 [00:25<01:16,  1.78it/s]

Calculating Affiliation Metrics:  32%|███▎      | 65/200 [00:26<01:16,  1.77it/s]

Calculating Affiliation Metrics:  34%|███▎      | 67/200 [00:27<01:15,  1.76it/s]

Calculating Affiliation Metrics:  34%|███▍      | 69/200 [00:28<01:14,  1.75it/s]

Calculating Affiliation Metrics:  36%|███▌      | 71/200 [00:29<01:14,  1.74it/s]

Calculating Affiliation Metrics:  36%|███▋      | 73/200 [00:30<01:13,  1.74it/s]

Calculating Affiliation Metrics:  38%|███▊      | 75/200 [00:31<01:12,  1.74it/s]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [00:33<01:11,  1.71it/s]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [00:34<01:11,  1.69it/s]

Calculating Affiliation Metrics:  40%|████      | 81/200 [00:35<01:12,  1.65it/s]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [00:36<01:11,  1.63it/s]

Calculating Affiliation Metrics:  42%|████▎     | 85/200 [00:38<01:11,  1.62it/s]

Calculating Affiliation Metrics:  44%|████▎     | 87/200 [00:39<01:10,  1.60it/s]

Calculating Affiliation Metrics:  44%|████▍     | 89/200 [00:40<01:09,  1.60it/s]

Calculating Affiliation Metrics:  46%|████▌     | 91/200 [00:41<01:08,  1.60it/s]

Calculating Affiliation Metrics:  46%|████▋     | 93/200 [00:43<01:06,  1.61it/s]

Calculating Affiliation Metrics:  48%|████▊     | 95/200 [00:44<01:04,  1.62it/s]

Calculating Affiliation Metrics:  48%|████▊     | 97/200 [00:45<01:03,  1.63it/s]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [00:46<01:00,  1.66it/s]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [00:47<00:59,  1.68it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 103/200 [00:49<00:57,  1.70it/s]

Calculating Affiliation Metrics:  52%|█████▎    | 105/200 [00:50<00:54,  1.73it/s]

Calculating Affiliation Metrics:  54%|█████▎    | 107/200 [00:51<00:52,  1.78it/s]

Calculating Affiliation Metrics:  55%|█████▍    | 109/200 [00:52<00:49,  1.84it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [00:53<00:46,  1.91it/s]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [00:55<00:43,  1.95it/s]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [00:56<00:41,  2.00it/s]

Calculating Affiliation Metrics:  60%|██████    | 121/200 [00:57<00:38,  2.06it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 124/200 [00:59<00:35,  2.14it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [01:00<00:32,  2.25it/s]

Calculating Affiliation Metrics:  65%|██████▌   | 130/200 [01:01<00:29,  2.34it/s]

Calculating Affiliation Metrics:  66%|██████▋   | 133/200 [01:02<00:27,  2.42it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 136/200 [01:03<00:25,  2.50it/s]

Calculating Affiliation Metrics:  70%|██████▉   | 139/200 [01:04<00:23,  2.56it/s]

Calculating Affiliation Metrics:  71%|███████   | 142/200 [01:06<00:22,  2.62it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [01:07<00:20,  2.67it/s]

Calculating Affiliation Metrics:  74%|███████▍  | 148/200 [01:08<00:19,  2.73it/s]

Calculating Affiliation Metrics:  76%|███████▌  | 152/200 [01:09<00:16,  2.85it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 156/200 [01:10<00:14,  2.96it/s]

Calculating Affiliation Metrics:  80%|████████  | 160/200 [01:11<00:12,  3.13it/s]

Calculating Affiliation Metrics:  82%|████████▏ | 164/200 [01:12<00:10,  3.28it/s]

Calculating Affiliation Metrics:  84%|████████▍ | 168/200 [01:13<00:09,  3.46it/s]

Calculating Affiliation Metrics:  86%|████████▋ | 173/200 [01:15<00:07,  3.74it/s]

Calculating Affiliation Metrics:  89%|████████▉ | 178/200 [01:16<00:05,  4.01it/s]

Calculating Affiliation Metrics:  92%|█████████▏| 184/200 [01:17<00:03,  4.45it/s]

Calculating Affiliation Metrics:  96%|█████████▋| 193/200 [01:18<00:01,  5.67it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [01:18<00:00,  2.54it/s]
Table1-only metrics completed in 127.54s
Evaluation Result:  {'AUC-PR': 0.145806438473018, 'AUC-ROC': 0.6324577883604849, 'VUS-PR': 0.9149833061868928, 'VUS-ROC': 0.9668608507581957, 'Standard-F1': 0.17647058823479458, 'Standard-Precision': 0.1830508474576269, 'Standard-Recall': 0.17034700315457396, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6812297517477471, 'Affiliation-P': 0.5274973800561625, 'Affiliation-R': 0.9614246952532624, 'F1_T': 0.17592083562347005, 'Precision_T': 0.18433179723502283, 'Recall_T': 0.16824395373291254}
Results for 155_Stock_id_7_Finance_tr_500_1st_18.csv: {'filename': '155_Stock_id_7_Finance_tr_500_1st_18.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 243, 'train_index': '500', 'data_shape': '20000x1', 'output_length': 20000, 'label_length': 19500, 'aligned_length': 19500, 'AUC-PR': 0.145806438473018, 'AUC-ROC': 0.6324577883604849, 'VUS-PR': 0.9149833061868928, 'VUS-ROC': 0.9668608507581957, 'Standard-F1': 0.17647058823479458, 'Standard-Precision': 0.1830508474576269, 'Standard-Recall': 0.17034700315457396, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6812297517477471, 'Affiliation-P': 0.5274973800561625, 'Affiliation-R': 0.9614246952532624, 'F1_T': 0.17592083562347005, 'Precision_T': 0.18433179723502283, 'Recall_T': 0.16824395373291254}
Logits metrics skipped for 155_Stock_id_7_Finance_tr_500_1st_18.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/155_Stock_id_7_Finance_tr_500_1st_18_results.pkl
Processing file: 333_UCR_id_31_Facility_tr_7500_1st_13400.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 333_UCR_id_31_Facility_tr_7500_1st_13400.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 149.32it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  86%|████████▌ | 172/200 [00:01<00:00, 171.80it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 183.57it/s]
Table1-only metrics completed in 2.15s
Evaluation Result:  {'AUC-PR': 0.07559409246821054, 'AUC-ROC': 0.8906702488687784, 'VUS-PR': 0.07848424310791188, 'VUS-ROC': 0.8964161736902222, 'Standard-F1': 0.16958211856153485, 'Standard-Precision': 0.09391819160387511, 'Standard-Recall': 0.8724999999999977, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7506939256809464, 'Affiliation-P': 0.6023019292604501, 'Affiliation-R': 0.9961103888888888, 'F1_T': 0.16830496283847432, 'Precision_T': 0.09307875894988064, 'Recall_T': 0.8774999999999977}
Results for 333_UCR_id_31_Facility_tr_7500_1st_13400.csv: {'filename': '333_UCR_id_31_Facility_tr_7500_1st_13400.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 189, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.07559409246821054, 'AUC-ROC': 0.8906702488687784, 'VUS-PR': 0.07848424310791188, 'VUS-ROC': 0.8964161736902222, 'Standard-F1': 0.16958211856153485, 'Standard-Precision': 0.09391819160387511, 'Standard-Recall': 0.8724999999999977, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7506939256809464, 'Affiliation-P': 0.6023019292604501, 'Affiliation-R': 0.9961103888888888, 'F1_T': 0.16830496283847432, 'Precision_T': 0.09307875894988064, 'Recall_T': 0.8774999999999977}
Logits metrics skipped for 333_UCR_id_31_Facility_tr_7500_1st_13400.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/333_UCR_id_31_Facility_tr_7500_1st_13400_results.pkl
Processing file: 752_YAHOO_id_202_WebService_tr_500_1st_95.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 752_YAHOO_id_202_WebService_tr_500_1st_95.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 159.34it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1121.01it/s]
Table1-only metrics completed in 0.61s
Evaluation Result:  {'AUC-PR': 0.9415204678362573, 'AUC-ROC': 0.9990511433722364, 'VUS-PR': 0.9632242898336746, 'VUS-ROC': 0.9996245382538202, 'Standard-F1': 0.7619047619041995, 'Standard-Precision': 0.6666666666666111, 'Standard-Recall': 0.8888888888887901, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9816699984468278, 'Affiliation-P': 0.9639998791644859, 'Affiliation-R': 1.0, 'F1_T': 0.7619047619041995, 'Precision_T': 0.6666666666666111, 'Recall_T': 0.8888888888887901}
Results for 752_YAHOO_id_202_WebService_tr_500_1st_95.csv: {'filename': '752_YAHOO_id_202_WebService_tr_500_1st_95.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9415204678362573, 'AUC-ROC': 0.9990511433722364, 'VUS-PR': 0.9632242898336746, 'VUS-ROC': 0.9996245382538202, 'Standard-F1': 0.7619047619041995, 'Standard-Precision': 0.6666666666666111, 'Standard-Recall': 0.8888888888887901, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9816699984468278, 'Affiliation-P': 0.9639998791644859, 'Affiliation-R': 1.0, 'F1_T': 0.7619047619041995, 'Precision_T': 0.6666666666666111, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 752_YAHOO_id_202_WebService_tr_500_1st_95.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/752_YAHOO_id_202_WebService_tr_500_1st_95_results.pkl
Processing file: 305_UCR_id_3_Medical_tr_3000_1st_5948.csv, Data shape: (12000, 1)
Label shape: (12000,)
Train data shape: (3000, 1), Test data shape: (9000, 1), Label test shape: (9000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 305_UCR_id_3_Medical_tr_3000_1st_5948.csv...
Using CUDA device: 2
Data shape: (9000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9000, window_size 5000, adding 1000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.65it/s]
Original shapes - Output: (10000,), Label: (9000,)
Aligned shapes - Output: (9000,), Label: (9000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 431.24it/s]
Table1-only metrics completed in 1.34s
Evaluation Result:  {'AUC-PR': 0.8901299940204848, 'AUC-ROC': 0.9969501830138345, 'VUS-PR': 0.8478705384582377, 'VUS-ROC': 0.9967665487933465, 'Standard-F1': 0.5925925925921394, 'Standard-Precision': 0.44444444444443953, 'Standard-Recall': 0.8888888888888691, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9399563417269052, 'Affiliation-P': 0.8867632850241545, 'Affiliation-R': 0.9999382716049384, 'F1_T': 0.5925925925921394, 'Precision_T': 0.44444444444443953, 'Recall_T': 0.8888888888888691}
Results for 305_UCR_id_3_Medical_tr_3000_1st_5948.csv: {'filename': '305_UCR_id_3_Medical_tr_3000_1st_5948.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 218, 'train_index': '3000', 'data_shape': '12000x1', 'output_length': 10000, 'label_length': 9000, 'aligned_length': 9000, 'AUC-PR': 0.8901299940204848, 'AUC-ROC': 0.9969501830138345, 'VUS-PR': 0.8478705384582377, 'VUS-ROC': 0.9967665487933465, 'Standard-F1': 0.5925925925921394, 'Standard-Precision': 0.44444444444443953, 'Standard-Recall': 0.8888888888888691, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9399563417269052, 'Affiliation-P': 0.8867632850241545, 'Affiliation-R': 0.9999382716049384, 'F1_T': 0.5925925925921394, 'Precision_T': 0.44444444444443953, 'Recall_T': 0.8888888888888691}
Logits metrics skipped for 305_UCR_id_3_Medical_tr_3000_1st_5948.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/305_UCR_id_3_Medical_tr_3000_1st_5948_results.pkl
Processing file: 738_YAHOO_id_188_WebService_tr_500_1st_26.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 738_YAHOO_id_188_WebService_tr_500_1st_26.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 166.84it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1258.26it/s]
Table1-only metrics completed in 0.32s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9408448782117509, 'VUS-ROC': 0.9998743193449048, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975867431084577, 'Affiliation-P': 0.9951851057935425, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Results for 738_YAHOO_id_188_WebService_tr_500_1st_26.csv: {'filename': '738_YAHOO_id_188_WebService_tr_500_1st_26.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9408448782117509, 'VUS-ROC': 0.9998743193449048, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975867431084577, 'Affiliation-P': 0.9951851057935425, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Logits metrics skipped for 738_YAHOO_id_188_WebService_tr_500_1st_26.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/738_YAHOO_id_188_WebService_tr_500_1st_26_results.pkl
Processing file: 115_WSD_id_87_WebService_tr_2252_1st_2352.csv, Data shape: (18214, 1)
Label shape: (18214,)
Train data shape: (2252, 1), Test data shape: (15962, 1), Label test shape: (15962,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 115_WSD_id_87_WebService_tr_2252_1st_2352.csv...
Using CUDA device: 2
Data shape: (15962, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 15962, window_size 5000, adding 4038 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.77it/s]
Original shapes - Output: (20000,), Label: (15962,)
Aligned shapes - Output: (15962,), Label: (15962,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  55%|█████▍    | 109/200 [00:01<00:00, 108.63it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 135.84it/s]
Table1-only metrics completed in 2.21s
Evaluation Result:  {'AUC-PR': 0.13206691575809235, 'AUC-ROC': 0.7458152173342234, 'VUS-PR': 0.07152593491317305, 'VUS-ROC': 0.8113497909150593, 'Standard-F1': 0.1407624633426096, 'Standard-Precision': 0.14999999999999908, 'Standard-Recall': 0.1325966850828722, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8785344947539301, 'Affiliation-P': 0.9123757329304464, 'Affiliation-R': 0.8471139051945705, 'F1_T': 0.1407624633426096, 'Precision_T': 0.14999999999999908, 'Recall_T': 0.1325966850828722}
Results for 115_WSD_id_87_WebService_tr_2252_1st_2352.csv: {'filename': '115_WSD_id_87_WebService_tr_2252_1st_2352.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 20, 'train_index': '2252', 'data_shape': '18214x1', 'output_length': 20000, 'label_length': 15962, 'aligned_length': 15962, 'AUC-PR': 0.13206691575809235, 'AUC-ROC': 0.7458152173342234, 'VUS-PR': 0.07152593491317305, 'VUS-ROC': 0.8113497909150593, 'Standard-F1': 0.1407624633426096, 'Standard-Precision': 0.14999999999999908, 'Standard-Recall': 0.1325966850828722, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8785344947539301, 'Affiliation-P': 0.9123757329304464, 'Affiliation-R': 0.8471139051945705, 'F1_T': 0.1407624633426096, 'Precision_T': 0.14999999999999908, 'Recall_T': 0.1325966850828722}
Logits metrics skipped for 115_WSD_id_87_WebService_tr_2252_1st_2352.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/115_WSD_id_87_WebService_tr_2252_1st_2352_results.pkl
Skipping file: 253_SVDB_id_17_Medical_tr_17803_1st_17903.csv due to filter criteria.
Processing file: 454_UCR_id_152_Medical_tr_9625_1st_20950.csv, Data shape: (38501, 1)
Label shape: (38501,)
Train data shape: (9625, 1), Test data shape: (28876, 1), Label test shape: (28876,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 454_UCR_id_152_Medical_tr_9625_1st_20950.csv...
Using CUDA device: 2
Data shape: (28876, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 28876, window_size 5000, adding 1124 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 140.68it/s]
Original shapes - Output: (30000,), Label: (28876,)
Aligned shapes - Output: (28876,), Label: (28876,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  55%|█████▌    | 110/200 [00:01<00:00, 109.14it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 127.00it/s]
Table1-only metrics completed in 1.91s
Evaluation Result:  {'AUC-PR': 0.44695438149807604, 'AUC-ROC': 0.9851500382928358, 'VUS-PR': 0.381674496124648, 'VUS-ROC': 0.9868145813016274, 'Standard-F1': 0.3599088838264278, 'Standard-Precision': 0.27335640138408207, 'Standard-Recall': 0.5266666666666632, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8916596875035184, 'Affiliation-P': 0.8047027251146613, 'Affiliation-R': 0.9996868218128089, 'F1_T': 0.3599088838264278, 'Precision_T': 0.27335640138408207, 'Recall_T': 0.5266666666666632}
Results for 454_UCR_id_152_Medical_tr_9625_1st_20950.csv: {'filename': '454_UCR_id_152_Medical_tr_9625_1st_20950.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 40, 'train_index': '9625', 'data_shape': '38501x1', 'output_length': 30000, 'label_length': 28876, 'aligned_length': 28876, 'AUC-PR': 0.44695438149807604, 'AUC-ROC': 0.9851500382928358, 'VUS-PR': 0.381674496124648, 'VUS-ROC': 0.9868145813016274, 'Standard-F1': 0.3599088838264278, 'Standard-Precision': 0.27335640138408207, 'Standard-Recall': 0.5266666666666632, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8916596875035184, 'Affiliation-P': 0.8047027251146613, 'Affiliation-R': 0.9996868218128089, 'F1_T': 0.3599088838264278, 'Precision_T': 0.27335640138408207, 'Recall_T': 0.5266666666666632}
Logits metrics skipped for 454_UCR_id_152_Medical_tr_9625_1st_20950.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/454_UCR_id_152_Medical_tr_9625_1st_20950_results.pkl
Processing file: 307_UCR_id_5_Facility_tr_29895_1st_39240.csv, Data shape: (119580, 1)
Label shape: (119580,)
Train data shape: (29895, 1), Test data shape: (89685, 1), Label test shape: (89685,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 307_UCR_id_5_Facility_tr_29895_1st_39240.csv...
Using CUDA device: 2
Data shape: (89685, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 89685, window_size 5000, adding 315 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([18, 5000, 1])
batched_mask shape: torch.Size([18, 5000])

100%|██████████| 1/1 [00:00<00:00, 74.35it/s]
Original shapes - Output: (90000,), Label: (89685,)
Aligned shapes - Output: (89685,), Label: (89685,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  20%|██        | 40/200 [00:01<00:04, 39.58it/s]

Calculating Affiliation Metrics:  40%|████      | 80/200 [00:02<00:03, 35.05it/s]

Calculating Affiliation Metrics:  58%|█████▊    | 116/200 [00:03<00:02, 34.13it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 157/200 [00:04<00:01, 36.54it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:04<00:00, 40.35it/s]
Table1-only metrics completed in 7.19s
Evaluation Result:  {'AUC-PR': 0.0007841494623003051, 'AUC-ROC': 0.36990433395841005, 'VUS-PR': 0.0009342377609142676, 'VUS-ROC': 0.4092729093200681, 'Standard-F1': 0.0025049439683560745, 'Standard-Precision': 0.0012540591915938433, 'Standard-Recall': 0.989583333333323, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7108712371734142, 'Affiliation-P': 0.5546962189194194, 'Affiliation-R': 0.9894519707866422, 'F1_T': 0.002505924331977886, 'Precision_T': 0.001254602481930997, 'Recall_T': 0.9583333333333234}
Results for 307_UCR_id_5_Facility_tr_29895_1st_39240.csv: {'filename': '307_UCR_id_5_Facility_tr_29895_1st_39240.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 96, 'train_index': '29895', 'data_shape': '119580x1', 'output_length': 90000, 'label_length': 89685, 'aligned_length': 89685, 'AUC-PR': 0.0007841494623003051, 'AUC-ROC': 0.36990433395841005, 'VUS-PR': 0.0009342377609142676, 'VUS-ROC': 0.4092729093200681, 'Standard-F1': 0.0025049439683560745, 'Standard-Precision': 0.0012540591915938433, 'Standard-Recall': 0.989583333333323, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7108712371734142, 'Affiliation-P': 0.5546962189194194, 'Affiliation-R': 0.9894519707866422, 'F1_T': 0.002505924331977886, 'Precision_T': 0.001254602481930997, 'Recall_T': 0.9583333333333234}
Logits metrics skipped for 307_UCR_id_5_Facility_tr_29895_1st_39240.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/307_UCR_id_5_Facility_tr_29895_1st_39240_results.pkl
Skipping file: 541_SMAP_id_11_Sensor_tr_1832_1st_5300.csv due to filter criteria.
Skipping file: 539_SMAP_id_9_Sensor_tr_1857_1st_4270.csv due to filter criteria.
Processing file: 501_UCR_id_199_Facility_tr_7482_1st_29480.csv, Data shape: (29931, 1)
Label shape: (29931,)
Train data shape: (7482, 1), Test data shape: (22449, 1), Label test shape: (22449,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 501_UCR_id_199_Facility_tr_7482_1st_29480.csv...
Using CUDA device: 2
Data shape: (22449, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22449, window_size 5000, adding 2551 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 148.03it/s]
Original shapes - Output: (25000,), Label: (22449,)
Aligned shapes - Output: (22449,), Label: (22449,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  60%|█████▉    | 119/200 [00:01<00:00, 118.37it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 145.00it/s]
Table1-only metrics completed in 2.28s
Evaluation Result:  {'AUC-PR': 0.005127676675221951, 'AUC-ROC': 0.9002136752136751, 'VUS-PR': 0.005660134499765322, 'VUS-ROC': 0.9051963824060687, 'Standard-F1': 0.013283520132815465, 'Standard-Precision': 0.006708595387840668, 'Standard-Recall': 0.666666666666639, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.69503887102679, 'Affiliation-P': 0.5326127005588834, 'Affiliation-R': 1.0, 'F1_T': 0.013008130081281481, 'Precision_T': 0.006568144499178979, 'Recall_T': 0.666666666666639}
Results for 501_UCR_id_199_Facility_tr_7482_1st_29480.csv: {'filename': '501_UCR_id_199_Facility_tr_7482_1st_29480.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 168, 'train_index': '7482', 'data_shape': '29931x1', 'output_length': 25000, 'label_length': 22449, 'aligned_length': 22449, 'AUC-PR': 0.005127676675221951, 'AUC-ROC': 0.9002136752136751, 'VUS-PR': 0.005660134499765322, 'VUS-ROC': 0.9051963824060687, 'Standard-F1': 0.013283520132815465, 'Standard-Precision': 0.006708595387840668, 'Standard-Recall': 0.666666666666639, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.69503887102679, 'Affiliation-P': 0.5326127005588834, 'Affiliation-R': 1.0, 'F1_T': 0.013008130081281481, 'Precision_T': 0.006568144499178979, 'Recall_T': 0.666666666666639}
Logits metrics skipped for 501_UCR_id_199_Facility_tr_7482_1st_29480.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/501_UCR_id_199_Facility_tr_7482_1st_29480_results.pkl
Processing file: 585_YAHOO_id_35_WebService_tr_500_1st_577.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 585_YAHOO_id_35_WebService_tr_500_1st_577.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 158.62it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1284.66it/s]
Table1-only metrics completed in 0.35s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.93193705339089, 'VUS-ROC': 0.9998553835992761, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9986965228805355, 'Affiliation-P': 0.9973964394426874, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Results for 585_YAHOO_id_35_WebService_tr_500_1st_577.csv: {'filename': '585_YAHOO_id_35_WebService_tr_500_1st_577.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.93193705339089, 'VUS-ROC': 0.9998553835992761, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9986965228805355, 'Affiliation-P': 0.9973964394426874, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Logits metrics skipped for 585_YAHOO_id_35_WebService_tr_500_1st_577.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/585_YAHOO_id_35_WebService_tr_500_1st_577_results.pkl
Processing file: 611_YAHOO_id_61_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 611_YAHOO_id_61_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 162.70it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1750.65it/s]
Table1-only metrics completed in 1.24s
Evaluation Result:  {'AUC-PR': 0.928030303030303, 'AUC-ROC': 0.9991471734892787, 'VUS-PR': 0.9980462196067569, 'VUS-ROC': 0.9999877578994462, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Results for 611_YAHOO_id_61_Synthetic_tr_500_1st_623.csv: {'filename': '611_YAHOO_id_61_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 235, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.928030303030303, 'AUC-ROC': 0.9991471734892787, 'VUS-PR': 0.9980462196067569, 'VUS-ROC': 0.9999877578994462, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 611_YAHOO_id_61_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/611_YAHOO_id_61_Synthetic_tr_500_1st_623_results.pkl
Processing file: 525_UCR_id_223_Facility_tr_50000_1st_250720.csv, Data shape: (299790, 1)
Label shape: (299790,)
Train data shape: (50000, 1), Test data shape: (249790, 1), Label test shape: (249790,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 525_UCR_id_223_Facility_tr_50000_1st_250720.csv...
Using CUDA device: 2
Data shape: (249790, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 249790, window_size 5000, adding 210 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([50, 5000, 1])
batched_mask shape: torch.Size([50, 5000])

100%|██████████| 1/1 [00:00<00:00, 28.44it/s]
Original shapes - Output: (250000,), Label: (249790,)
Aligned shapes - Output: (249790,), Label: (249790,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  10%|▉         | 19/200 [00:01<00:10, 17.96it/s]

Calculating Affiliation Metrics:  18%|█▊        | 37/200 [00:02<00:10, 15.34it/s]

Calculating Affiliation Metrics:  26%|██▋       | 53/200 [00:03<00:10, 14.19it/s]

Calculating Affiliation Metrics:  34%|███▍      | 68/200 [00:04<00:10, 13.17it/s]

Calculating Affiliation Metrics:  41%|████      | 82/200 [00:05<00:09, 13.07it/s]

Calculating Affiliation Metrics:  48%|████▊     | 96/200 [00:07<00:07, 13.01it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 111/200 [00:08<00:06, 13.38it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [00:09<00:05, 14.06it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [00:10<00:03, 15.14it/s]

Calculating Affiliation Metrics:  83%|████████▎ | 166/200 [00:11<00:02, 16.75it/s]

Calculating Affiliation Metrics:  97%|█████████▋| 194/200 [00:12<00:00, 19.98it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:12<00:00, 16.20it/s]
Table1-only metrics completed in 25.56s
Evaluation Result:  {'AUC-PR': 0.15843944367331197, 'AUC-ROC': 0.9784942818680877, 'VUS-PR': 0.10315815379508983, 'VUS-ROC': 0.9789460097778151, 'Standard-F1': 0.1905972045740051, 'Standard-Precision': 0.12009607686148915, 'Standard-Recall': 0.4615384615384608, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7128880685622035, 'Affiliation-P': 0.5539392181789391, 'Affiliation-R': 0.9997627576394941, 'F1_T': 0.1905972045740051, 'Precision_T': 0.12009607686148915, 'Recall_T': 0.4615384615384608}
Results for 525_UCR_id_223_Facility_tr_50000_1st_250720.csv: {'filename': '525_UCR_id_223_Facility_tr_50000_1st_250720.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 200, 'train_index': '50000', 'data_shape': '299790x1', 'output_length': 250000, 'label_length': 249790, 'aligned_length': 249790, 'AUC-PR': 0.15843944367331197, 'AUC-ROC': 0.9784942818680877, 'VUS-PR': 0.10315815379508983, 'VUS-ROC': 0.9789460097778151, 'Standard-F1': 0.1905972045740051, 'Standard-Precision': 0.12009607686148915, 'Standard-Recall': 0.4615384615384608, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7128880685622035, 'Affiliation-P': 0.5539392181789391, 'Affiliation-R': 0.9997627576394941, 'F1_T': 0.1905972045740051, 'Precision_T': 0.12009607686148915, 'Recall_T': 0.4615384615384608}
Logits metrics skipped for 525_UCR_id_223_Facility_tr_50000_1st_250720.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/525_UCR_id_223_Facility_tr_50000_1st_250720_results.pkl
Skipping file: 851_OPPORTUNITY_id_10_HumanActivity_tr_6213_1st_7311.csv due to filter criteria.
Processing file: 106_WSD_id_78_WebService_tr_3363_1st_3463.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (3363, 1), Test data shape: (14873, 1), Label test shape: (14873,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 106_WSD_id_78_WebService_tr_3363_1st_3463.csv...
Using CUDA device: 2
Data shape: (14873, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 14873, window_size 5000, adding 127 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.91it/s]
Original shapes - Output: (15000,), Label: (14873,)
Aligned shapes - Output: (14873,), Label: (14873,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 260.97it/s]
Table1-only metrics completed in 1.93s
Evaluation Result:  {'AUC-PR': 0.8663180632116819, 'AUC-ROC': 0.9518691871633048, 'VUS-PR': 0.9115643452329065, 'VUS-ROC': 0.9940833655979026, 'Standard-F1': 0.8771929824556417, 'Standard-Precision': 0.9831460674157249, 'Standard-Recall': 0.7918552036199059, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968715249616573, 'Affiliation-P': 0.9949718889767754, 'Affiliation-R': 0.998778428528258, 'F1_T': 0.8577878103832433, 'Precision_T': 0.855855855855852, 'Recall_T': 0.8597285067873265}
Results for 106_WSD_id_78_WebService_tr_3363_1st_3463.csv: {'filename': '106_WSD_id_78_WebService_tr_3363_1st_3463.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '3363', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 14873, 'aligned_length': 14873, 'AUC-PR': 0.8663180632116819, 'AUC-ROC': 0.9518691871633048, 'VUS-PR': 0.9115643452329065, 'VUS-ROC': 0.9940833655979026, 'Standard-F1': 0.8771929824556417, 'Standard-Precision': 0.9831460674157249, 'Standard-Recall': 0.7918552036199059, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968715249616573, 'Affiliation-P': 0.9949718889767754, 'Affiliation-R': 0.998778428528258, 'F1_T': 0.8577878103832433, 'Precision_T': 0.855855855855852, 'Recall_T': 0.8597285067873265}
Logits metrics skipped for 106_WSD_id_78_WebService_tr_3363_1st_3463.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/106_WSD_id_78_WebService_tr_3363_1st_3463_results.pkl
Skipping file: 140_MSL_id_1_Sensor_tr_530_1st_630.csv due to filter criteria.
Processing file: 272_IOPS_id_13_WebService_tr_3442_1st_3542.csv, Data shape: (149156, 1)
Label shape: (149156,)
Train data shape: (3442, 1), Test data shape: (145714, 1), Label test shape: (145714,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 272_IOPS_id_13_WebService_tr_3442_1st_3542.csv...
Using CUDA device: 2
Data shape: (145714, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 145714, window_size 5000, adding 4286 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([30, 5000, 1])
batched_mask shape: torch.Size([30, 5000])

100%|██████████| 1/1 [00:00<00:00, 47.03it/s]
Original shapes - Output: (150000,), Label: (145714,)
Aligned shapes - Output: (145714,), Label: (145714,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   7%|▋         | 14/200 [00:01<00:13, 13.41it/s]

Calculating Affiliation Metrics:  14%|█▍        | 28/200 [00:03<00:21,  8.04it/s]

Calculating Affiliation Metrics:  19%|█▉        | 38/200 [00:05<00:23,  6.81it/s]

Calculating Affiliation Metrics:  23%|██▎       | 46/200 [00:06<00:24,  6.25it/s]

Calculating Affiliation Metrics:  26%|██▋       | 53/200 [00:07<00:24,  5.97it/s]

Calculating Affiliation Metrics:  30%|███       | 60/200 [00:09<00:24,  5.78it/s]

Calculating Affiliation Metrics:  33%|███▎      | 66/200 [00:10<00:23,  5.78it/s]

Calculating Affiliation Metrics:  36%|███▌      | 72/200 [00:11<00:21,  5.83it/s]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [00:12<00:20,  5.94it/s]

Calculating Affiliation Metrics:  43%|████▎     | 86/200 [00:13<00:18,  6.19it/s]

Calculating Affiliation Metrics:  47%|████▋     | 94/200 [00:14<00:16,  6.58it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 103/200 [00:15<00:13,  7.12it/s]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [00:16<00:11,  7.80it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 124/200 [00:17<00:08,  8.57it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 135/200 [00:18<00:07,  9.21it/s]

Calculating Affiliation Metrics:  74%|███████▎  | 147/200 [00:19<00:05, 10.01it/s]

Calculating Affiliation Metrics:  80%|████████  | 161/200 [00:20<00:03, 11.02it/s]

Calculating Affiliation Metrics:  90%|█████████ | 180/200 [00:21<00:01, 13.17it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:22<00:00,  8.85it/s]
Table1-only metrics completed in 40.86s
Evaluation Result:  {'AUC-PR': 0.6047452431457273, 'AUC-ROC': 0.9613388328596479, 'VUS-PR': 0.4786069761092934, 'VUS-ROC': 0.979887357951032, 'Standard-F1': 0.3988522238159332, 'Standard-Precision': 0.2860082304526747, 'Standard-Recall': 0.6587677725118473, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9454052660280972, 'Affiliation-P': 0.9475439771528088, 'Affiliation-R': 0.9432761877740948, 'F1_T': 0.3988522238159332, 'Precision_T': 0.2860082304526747, 'Recall_T': 0.6587677725118473}
Results for 272_IOPS_id_13_WebService_tr_3442_1st_3542.csv: {'filename': '272_IOPS_id_13_WebService_tr_3442_1st_3542.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '3442', 'data_shape': '149156x1', 'output_length': 150000, 'label_length': 145714, 'aligned_length': 145714, 'AUC-PR': 0.6047452431457273, 'AUC-ROC': 0.9613388328596479, 'VUS-PR': 0.4786069761092934, 'VUS-ROC': 0.979887357951032, 'Standard-F1': 0.3988522238159332, 'Standard-Precision': 0.2860082304526747, 'Standard-Recall': 0.6587677725118473, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9454052660280972, 'Affiliation-P': 0.9475439771528088, 'Affiliation-R': 0.9432761877740948, 'F1_T': 0.3988522238159332, 'Precision_T': 0.2860082304526747, 'Recall_T': 0.6587677725118473}
Logits metrics skipped for 272_IOPS_id_13_WebService_tr_3442_1st_3542.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/272_IOPS_id_13_WebService_tr_3442_1st_3542_results.pkl
Processing file: 229_MGAB_id_5_Synthetic_tr_25000_1st_33098.csv, Data shape: (100000, 1)
Label shape: (100000,)
Train data shape: (25000, 1), Test data shape: (75000, 1), Label test shape: (75000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 229_MGAB_id_5_Synthetic_tr_25000_1st_33098.csv...
Using CUDA device: 2
Data shape: (75000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([15, 5000, 1])
batched_mask shape: torch.Size([15, 5000])

100%|██████████| 1/1 [00:00<00:00, 78.28it/s]
Original shapes - Output: (75000,), Label: (75000,)
Aligned shapes - Output: (75000,), Label: (75000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  18%|█▊        | 37/200 [00:01<00:04, 36.90it/s]

Calculating Affiliation Metrics:  37%|███▋      | 74/200 [00:02<00:04, 27.43it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 103/200 [00:03<00:03, 24.35it/s]

Calculating Affiliation Metrics:  64%|██████▍   | 129/200 [00:05<00:03, 23.42it/s]

Calculating Affiliation Metrics:  77%|███████▋  | 154/200 [00:06<00:01, 23.69it/s]

Calculating Affiliation Metrics:  97%|█████████▋| 194/200 [00:07<00:00, 28.54it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:07<00:00, 27.51it/s]
Table1-only metrics completed in 9.20s
Evaluation Result:  {'AUC-PR': 0.009489892661663571, 'AUC-ROC': 0.7447411096256684, 'VUS-PR': 0.014640315118823157, 'VUS-ROC': 0.7936510653144855, 'Standard-F1': 0.03271537622663225, 'Standard-Precision': 0.01835985312117502, 'Standard-Recall': 0.14999999999999925, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6671508606522938, 'Affiliation-P': 0.5005449161176613, 'Affiliation-R': 1.0, 'F1_T': 0.032054395337367064, 'Precision_T': 0.01775147928994082, 'Recall_T': 0.16499999999999918}
Results for 229_MGAB_id_5_Synthetic_tr_25000_1st_33098.csv: {'filename': '229_MGAB_id_5_Synthetic_tr_25000_1st_33098.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 49, 'train_index': '25000', 'data_shape': '100000x1', 'output_length': 75000, 'label_length': 75000, 'aligned_length': 75000, 'AUC-PR': 0.009489892661663571, 'AUC-ROC': 0.7447411096256684, 'VUS-PR': 0.014640315118823157, 'VUS-ROC': 0.7936510653144855, 'Standard-F1': 0.03271537622663225, 'Standard-Precision': 0.01835985312117502, 'Standard-Recall': 0.14999999999999925, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6671508606522938, 'Affiliation-P': 0.5005449161176613, 'Affiliation-R': 1.0, 'F1_T': 0.032054395337367064, 'Precision_T': 0.01775147928994082, 'Recall_T': 0.16499999999999918}
Logits metrics skipped for 229_MGAB_id_5_Synthetic_tr_25000_1st_33098.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/229_MGAB_id_5_Synthetic_tr_25000_1st_33098_results.pkl
Processing file: 527_UCR_id_225_Medical_tr_3000_1st_5948.csv, Data shape: (12000, 1)
Label shape: (12000,)
Train data shape: (3000, 1), Test data shape: (9000, 1), Label test shape: (9000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 527_UCR_id_225_Medical_tr_3000_1st_5948.csv...
Using CUDA device: 2
Data shape: (9000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9000, window_size 5000, adding 1000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.22it/s]
Original shapes - Output: (10000,), Label: (9000,)
Aligned shapes - Output: (9000,), Label: (9000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 529.92it/s]
Table1-only metrics completed in 1.26s
Evaluation Result:  {'AUC-PR': 0.9124781465674138, 'AUC-ROC': 0.9974688256095291, 'VUS-PR': 0.8856033248661049, 'VUS-ROC': 0.9980361082889423, 'Standard-F1': 0.5925925925921394, 'Standard-Precision': 0.44444444444443953, 'Standard-Recall': 0.8888888888888691, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9552088594021056, 'Affiliation-P': 0.9142850241545893, 'Affiliation-R': 0.9999679012345679, 'F1_T': 0.5925925925921394, 'Precision_T': 0.44444444444443953, 'Recall_T': 0.8888888888888691}
Results for 527_UCR_id_225_Medical_tr_3000_1st_5948.csv: {'filename': '527_UCR_id_225_Medical_tr_3000_1st_5948.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 218, 'train_index': '3000', 'data_shape': '12000x1', 'output_length': 10000, 'label_length': 9000, 'aligned_length': 9000, 'AUC-PR': 0.9124781465674138, 'AUC-ROC': 0.9974688256095291, 'VUS-PR': 0.8856033248661049, 'VUS-ROC': 0.9980361082889423, 'Standard-F1': 0.5925925925921394, 'Standard-Precision': 0.44444444444443953, 'Standard-Recall': 0.8888888888888691, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9552088594021056, 'Affiliation-P': 0.9142850241545893, 'Affiliation-R': 0.9999679012345679, 'F1_T': 0.5925925925921394, 'Precision_T': 0.44444444444443953, 'Recall_T': 0.8888888888888691}
Logits metrics skipped for 527_UCR_id_225_Medical_tr_3000_1st_5948.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/527_UCR_id_225_Medical_tr_3000_1st_5948_results.pkl
Processing file: 729_YAHOO_id_179_WebService_tr_500_1st_559.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 729_YAHOO_id_179_WebService_tr_500_1st_559.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 162.94it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1221.70it/s]
Table1-only metrics completed in 0.48s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.968749391105237, 'VUS-ROC': 0.9998668349842569, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Results for 729_YAHOO_id_179_WebService_tr_500_1st_559.csv: {'filename': '729_YAHOO_id_179_WebService_tr_500_1st_559.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.968749391105237, 'VUS-ROC': 0.9998668349842569, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999999999999996, 'Affiliation-P': 1.0, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Logits metrics skipped for 729_YAHOO_id_179_WebService_tr_500_1st_559.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/729_YAHOO_id_179_WebService_tr_500_1st_559_results.pkl
Processing file: 379_UCR_id_77_Facility_tr_32191_1st_124159.csv, Data shape: (128767, 1)
Label shape: (128767,)
Train data shape: (32191, 1), Test data shape: (96576, 1), Label test shape: (96576,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 379_UCR_id_77_Facility_tr_32191_1st_124159.csv...
Using CUDA device: 2
Data shape: (96576, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 96576, window_size 5000, adding 3424 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([20, 5000, 1])
batched_mask shape: torch.Size([20, 5000])

100%|██████████| 1/1 [00:00<00:00, 65.06it/s]
Original shapes - Output: (100000,), Label: (96576,)
Aligned shapes - Output: (96576,), Label: (96576,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  20%|█▉        | 39/200 [00:01<00:04, 38.74it/s]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [00:02<00:03, 38.02it/s]

Calculating Affiliation Metrics:  61%|██████    | 122/200 [00:03<00:01, 40.68it/s]

Calculating Affiliation Metrics:  93%|█████████▎| 186/200 [00:04<00:00, 49.66it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:04<00:00, 47.62it/s]
Table1-only metrics completed in 10.04s
Evaluation Result:  {'AUC-PR': 0.793677353348471, 'AUC-ROC': 0.9966611749979454, 'VUS-PR': 0.7281441027012344, 'VUS-ROC': 0.9967339902018263, 'Standard-F1': 0.6651785714280736, 'Standard-Precision': 0.6169772256728772, 'Standard-Recall': 0.7215496368038732, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9548118090934489, 'Affiliation-P': 0.9137213647981793, 'Affiliation-R': 0.9997719870245838, 'F1_T': 0.6651785714280736, 'Precision_T': 0.6169772256728772, 'Recall_T': 0.7215496368038732}
Results for 379_UCR_id_77_Facility_tr_32191_1st_124159.csv: {'filename': '379_UCR_id_77_Facility_tr_32191_1st_124159.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 243, 'train_index': '32191', 'data_shape': '128767x1', 'output_length': 100000, 'label_length': 96576, 'aligned_length': 96576, 'AUC-PR': 0.793677353348471, 'AUC-ROC': 0.9966611749979454, 'VUS-PR': 0.7281441027012344, 'VUS-ROC': 0.9967339902018263, 'Standard-F1': 0.6651785714280736, 'Standard-Precision': 0.6169772256728772, 'Standard-Recall': 0.7215496368038732, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9548118090934489, 'Affiliation-P': 0.9137213647981793, 'Affiliation-R': 0.9997719870245838, 'F1_T': 0.6651785714280736, 'Precision_T': 0.6169772256728772, 'Recall_T': 0.7215496368038732}
Logits metrics skipped for 379_UCR_id_77_Facility_tr_32191_1st_124159.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/379_UCR_id_77_Facility_tr_32191_1st_124159_results.pkl
Processing file: 520_UCR_id_218_Medical_tr_9567_1st_16000.csv, Data shape: (38269, 1)
Label shape: (38269,)
Train data shape: (9567, 1), Test data shape: (28702, 1), Label test shape: (28702,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 520_UCR_id_218_Medical_tr_9567_1st_16000.csv...
Using CUDA device: 2
Data shape: (28702, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 28702, window_size 5000, adding 1298 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 138.71it/s]
Original shapes - Output: (30000,), Label: (28702,)
Aligned shapes - Output: (28702,), Label: (28702,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 131/200 [00:01<00:00, 130.73it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 146.84it/s]
Table1-only metrics completed in 1.70s
Evaluation Result:  {'AUC-PR': 0.05421568514101813, 'AUC-ROC': 0.9714250751695684, 'VUS-PR': 0.06015973142561982, 'VUS-ROC': 0.9722736681976559, 'Standard-F1': 0.12914188615107625, 'Standard-Precision': 0.07056638811513458, 'Standard-Recall': 0.7599999999999925, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7480899692236227, 'Affiliation-P': 0.6405818169585683, 'Affiliation-R': 0.8989617448261445, 'F1_T': 0.12797074954279528, 'Precision_T': 0.07042253521126754, 'Recall_T': 0.6999999999999931}
Results for 520_UCR_id_218_Medical_tr_9567_1st_16000.csv: {'filename': '520_UCR_id_218_Medical_tr_9567_1st_16000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 40, 'train_index': '9567', 'data_shape': '38269x1', 'output_length': 30000, 'label_length': 28702, 'aligned_length': 28702, 'AUC-PR': 0.05421568514101813, 'AUC-ROC': 0.9714250751695684, 'VUS-PR': 0.06015973142561982, 'VUS-ROC': 0.9722736681976559, 'Standard-F1': 0.12914188615107625, 'Standard-Precision': 0.07056638811513458, 'Standard-Recall': 0.7599999999999925, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7480899692236227, 'Affiliation-P': 0.6405818169585683, 'Affiliation-R': 0.8989617448261445, 'F1_T': 0.12797074954279528, 'Precision_T': 0.07042253521126754, 'Recall_T': 0.6999999999999931}
Logits metrics skipped for 520_UCR_id_218_Medical_tr_9567_1st_16000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/520_UCR_id_218_Medical_tr_9567_1st_16000_results.pkl
Processing file: 309_UCR_id_7_Medical_tr_3000_1st_5948.csv, Data shape: (12000, 1)
Label shape: (12000,)
Train data shape: (3000, 1), Test data shape: (9000, 1), Label test shape: (9000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 309_UCR_id_7_Medical_tr_3000_1st_5948.csv...
Using CUDA device: 2
Data shape: (9000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9000, window_size 5000, adding 1000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.99it/s]
Original shapes - Output: (10000,), Label: (9000,)
Aligned shapes - Output: (9000,), Label: (9000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 525.03it/s]
Table1-only metrics completed in 1.26s
Evaluation Result:  {'AUC-PR': 0.9195429576222055, 'AUC-ROC': 0.9931608660586886, 'VUS-PR': 0.8876311618475222, 'VUS-ROC': 0.9965554739235103, 'Standard-F1': 0.6074074074069539, 'Standard-Precision': 0.4555555555555505, 'Standard-Recall': 0.9111111111110908, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9817282921703291, 'Affiliation-P': 0.964135265700483, 'Affiliation-R': 0.9999753086419754, 'F1_T': 0.6074074074069539, 'Precision_T': 0.4555555555555505, 'Recall_T': 0.9111111111110908}
Results for 309_UCR_id_7_Medical_tr_3000_1st_5948.csv: {'filename': '309_UCR_id_7_Medical_tr_3000_1st_5948.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 218, 'train_index': '3000', 'data_shape': '12000x1', 'output_length': 10000, 'label_length': 9000, 'aligned_length': 9000, 'AUC-PR': 0.9195429576222055, 'AUC-ROC': 0.9931608660586886, 'VUS-PR': 0.8876311618475222, 'VUS-ROC': 0.9965554739235103, 'Standard-F1': 0.6074074074069539, 'Standard-Precision': 0.4555555555555505, 'Standard-Recall': 0.9111111111110908, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9817282921703291, 'Affiliation-P': 0.964135265700483, 'Affiliation-R': 0.9999753086419754, 'F1_T': 0.6074074074069539, 'Precision_T': 0.4555555555555505, 'Recall_T': 0.9111111111110908}
Logits metrics skipped for 309_UCR_id_7_Medical_tr_3000_1st_5948.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/309_UCR_id_7_Medical_tr_3000_1st_5948_results.pkl
Processing file: 776_YAHOO_id_226_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 776_YAHOO_id_226_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 161.82it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1521.18it/s]
Table1-only metrics completed in 0.37s
Evaluation Result:  {'AUC-PR': 0.9419191919191919, 'AUC-ROC': 0.9992690058479532, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9981780632671686, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9963627533675843, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Results for 776_YAHOO_id_226_Synthetic_tr_500_1st_623.csv: {'filename': '776_YAHOO_id_226_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 35, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9419191919191919, 'AUC-ROC': 0.9992690058479532, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9981780632671686, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9963627533675843, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 776_YAHOO_id_226_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/776_YAHOO_id_226_Synthetic_tr_500_1st_623_results.pkl
Processing file: 010_NAB_id_10_WebService_tr_500_1st_271.csv, Data shape: (1000, 1)
Label shape: (1000,)
Train data shape: (500, 1), Test data shape: (500, 1), Label test shape: (500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 010_NAB_id_10_WebService_tr_500_1st_271.csv...
Using CUDA device: 2
Data shape: (500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 500, 1])
batched_mask shape: torch.Size([1, 500])

100%|██████████| 1/1 [00:00<00:00, 167.24it/s]
Original shapes - Output: (500,), Label: (500,)
Aligned shapes - Output: (500,), Label: (500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 3655.18it/s]
Table1-only metrics completed in 0.50s
Evaluation Result:  {'AUC-PR': 0.4275757288977256, 'AUC-ROC': 0.7351585545722714, 'VUS-PR': 0.5983596018269346, 'VUS-ROC': 0.8558988232027316, 'Standard-F1': 0.481012658227359, 'Standard-Precision': 0.6129032258064319, 'Standard-Recall': 0.3958333333333251, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.986989141789863, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9743124999999999, 'F1_T': 0.46913580246864134, 'Precision_T': 0.5757575757575583, 'Recall_T': 0.3958333333333251}
Results for 010_NAB_id_10_WebService_tr_500_1st_271.csv: {'filename': '010_NAB_id_10_WebService_tr_500_1st_271.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 142, 'train_index': '500', 'data_shape': '1000x1', 'output_length': 500, 'label_length': 500, 'aligned_length': 500, 'AUC-PR': 0.4275757288977256, 'AUC-ROC': 0.7351585545722714, 'VUS-PR': 0.5983596018269346, 'VUS-ROC': 0.8558988232027316, 'Standard-F1': 0.481012658227359, 'Standard-Precision': 0.6129032258064319, 'Standard-Recall': 0.3958333333333251, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.986989141789863, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9743124999999999, 'F1_T': 0.46913580246864134, 'Precision_T': 0.5757575757575583, 'Recall_T': 0.3958333333333251}
Logits metrics skipped for 010_NAB_id_10_WebService_tr_500_1st_271.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/010_NAB_id_10_WebService_tr_500_1st_271_results.pkl
Skipping file: 178_SMD_id_1_Facility_tr_6000_1st_10609.csv due to filter criteria.
Processing file: 388_UCR_id_86_Facility_tr_17500_1st_23355.csv, Data shape: (70001, 1)
Label shape: (70001,)
Train data shape: (17500, 1), Test data shape: (52501, 1), Label test shape: (52501,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 388_UCR_id_86_Facility_tr_17500_1st_23355.csv...
Using CUDA device: 2
Data shape: (52501, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 52501, window_size 5000, adding 2499 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([11, 5000, 1])
batched_mask shape: torch.Size([11, 5000])

100%|██████████| 1/1 [00:00<00:00, 94.03it/s]
Original shapes - Output: (55000,), Label: (52501,)
Aligned shapes - Output: (52501,), Label: (52501,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  38%|███▊      | 76/200 [00:01<00:01, 75.39it/s]

Calculating Affiliation Metrics:  84%|████████▎ | 167/200 [00:02<00:00, 84.48it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 89.22it/s]
Table1-only metrics completed in 4.98s
Evaluation Result:  {'AUC-PR': 0.8900634760022439, 'AUC-ROC': 0.9970968261230576, 'VUS-PR': 0.26295293050169277, 'VUS-ROC': 0.9984534208516955, 'Standard-F1': 0.1436077057791888, 'Standard-Precision': 0.07794676806083635, 'Standard-Recall': 0.9111111111110908, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7858497784660216, 'Affiliation-P': 0.6472454637227689, 'Affiliation-R': 0.9999932276422248, 'F1_T': 0.1436077057791888, 'Precision_T': 0.07794676806083635, 'Recall_T': 0.9111111111110908}
Results for 388_UCR_id_86_Facility_tr_17500_1st_23355.csv: {'filename': '388_UCR_id_86_Facility_tr_17500_1st_23355.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 294, 'train_index': '17500', 'data_shape': '70001x1', 'output_length': 55000, 'label_length': 52501, 'aligned_length': 52501, 'AUC-PR': 0.8900634760022439, 'AUC-ROC': 0.9970968261230576, 'VUS-PR': 0.26295293050169277, 'VUS-ROC': 0.9984534208516955, 'Standard-F1': 0.1436077057791888, 'Standard-Precision': 0.07794676806083635, 'Standard-Recall': 0.9111111111110908, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7858497784660216, 'Affiliation-P': 0.6472454637227689, 'Affiliation-R': 0.9999932276422248, 'F1_T': 0.1436077057791888, 'Precision_T': 0.07794676806083635, 'Recall_T': 0.9111111111110908}
Logits metrics skipped for 388_UCR_id_86_Facility_tr_17500_1st_23355.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/388_UCR_id_86_Facility_tr_17500_1st_23355_results.pkl
Processing file: 614_YAHOO_id_64_WebService_tr_500_1st_1204.csv, Data shape: (1434, 1)
Label shape: (1434,)
Train data shape: (500, 1), Test data shape: (934, 1), Label test shape: (934,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 614_YAHOO_id_64_WebService_tr_500_1st_1204.csv...
Using CUDA device: 2
Data shape: (934, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 934, 1])
batched_mask shape: torch.Size([1, 934])

100%|██████████| 1/1 [00:00<00:00, 159.87it/s]
Original shapes - Output: (934,), Label: (934,)
Aligned shapes - Output: (934,), Label: (934,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2346.89it/s]
Table1-only metrics completed in 0.99s
Evaluation Result:  {'AUC-PR': 0.12442888700597744, 'AUC-ROC': 0.8635620915032679, 'VUS-PR': 0.19050307153348253, 'VUS-ROC': 0.8565747748933353, 'Standard-F1': 0.3263157894733869, 'Standard-Precision': 0.19871794871794746, 'Standard-Recall': 0.911764705882326, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.707829739327667, 'Affiliation-P': 0.5480187269316772, 'Affiliation-R': 0.9992175273865416, 'F1_T': 0.32460732984263585, 'Precision_T': 0.19745222929936182, 'Recall_T': 0.911764705882326}
Results for 614_YAHOO_id_64_WebService_tr_500_1st_1204.csv: {'filename': '614_YAHOO_id_64_WebService_tr_500_1st_1204.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 168, 'train_index': '500', 'data_shape': '1434x1', 'output_length': 934, 'label_length': 934, 'aligned_length': 934, 'AUC-PR': 0.12442888700597744, 'AUC-ROC': 0.8635620915032679, 'VUS-PR': 0.19050307153348253, 'VUS-ROC': 0.8565747748933353, 'Standard-F1': 0.3263157894733869, 'Standard-Precision': 0.19871794871794746, 'Standard-Recall': 0.911764705882326, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.707829739327667, 'Affiliation-P': 0.5480187269316772, 'Affiliation-R': 0.9992175273865416, 'F1_T': 0.32460732984263585, 'Precision_T': 0.19745222929936182, 'Recall_T': 0.911764705882326}
Logits metrics skipped for 614_YAHOO_id_64_WebService_tr_500_1st_1204.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/614_YAHOO_id_64_WebService_tr_500_1st_1204_results.pkl
Skipping file: 182_SMD_id_5_Facility_tr_7174_1st_21230.csv due to filter criteria.
Processing file: 009_NAB_id_9_Traffic_tr_500_1st_438.csv, Data shape: (2161, 1)
Label shape: (2161,)
Train data shape: (500, 1), Test data shape: (1661, 1), Label test shape: (1661,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 009_NAB_id_9_Traffic_tr_500_1st_438.csv...
Using CUDA device: 2
Data shape: (1661, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1661, 1])
batched_mask shape: torch.Size([1, 1661])

100%|██████████| 1/1 [00:00<00:00, 166.94it/s]
Original shapes - Output: (1661,), Label: (1661,)
Aligned shapes - Output: (1661,), Label: (1661,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1287.37it/s]
Table1-only metrics completed in 0.59s
Evaluation Result:  {'AUC-PR': 0.3379511803561379, 'AUC-ROC': 0.7759632537336585, 'VUS-PR': 0.44744588975341065, 'VUS-ROC': 0.8468625684667785, 'Standard-F1': 0.3970588235289176, 'Standard-Precision': 0.4576271186440639, 'Standard-Recall': 0.3506493506493484, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9343079360363208, 'Affiliation-P': 0.8896748946417821, 'Affiliation-R': 0.9836557933336982, 'F1_T': 0.3945578231287502, 'Precision_T': 0.4142857142857113, 'Recall_T': 0.3766233766233742}
Results for 009_NAB_id_9_Traffic_tr_500_1st_438.csv: {'filename': '009_NAB_id_9_Traffic_tr_500_1st_438.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 128, 'train_index': '500', 'data_shape': '2161x1', 'output_length': 1661, 'label_length': 1661, 'aligned_length': 1661, 'AUC-PR': 0.3379511803561379, 'AUC-ROC': 0.7759632537336585, 'VUS-PR': 0.44744588975341065, 'VUS-ROC': 0.8468625684667785, 'Standard-F1': 0.3970588235289176, 'Standard-Precision': 0.4576271186440639, 'Standard-Recall': 0.3506493506493484, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9343079360363208, 'Affiliation-P': 0.8896748946417821, 'Affiliation-R': 0.9836557933336982, 'F1_T': 0.3945578231287502, 'Precision_T': 0.4142857142857113, 'Recall_T': 0.3766233766233742}
Logits metrics skipped for 009_NAB_id_9_Traffic_tr_500_1st_438.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/009_NAB_id_9_Traffic_tr_500_1st_438_results.pkl
Skipping file: 863_OPPORTUNITY_id_22_HumanActivity_tr_808_1st_908.csv due to filter criteria.
Processing file: 777_YAHOO_id_227_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 777_YAHOO_id_227_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 162.64it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1664.39it/s]
Table1-only metrics completed in 0.37s
Evaluation Result:  {'AUC-PR': 0.8870731120731121, 'AUC-ROC': 0.9990253411306044, 'VUS-PR': 0.9833386619355977, 'VUS-ROC': 0.9998788255120282, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984359113984375, 'Affiliation-P': 0.9969624510347107, 'Affiliation-R': 0.9999137336093858, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 777_YAHOO_id_227_Synthetic_tr_500_1st_623.csv: {'filename': '777_YAHOO_id_227_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 35, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.8870731120731121, 'AUC-ROC': 0.9990253411306044, 'VUS-PR': 0.9833386619355977, 'VUS-ROC': 0.9998788255120282, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984359113984375, 'Affiliation-P': 0.9969624510347107, 'Affiliation-R': 0.9999137336093858, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 777_YAHOO_id_227_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/777_YAHOO_id_227_Synthetic_tr_500_1st_623_results.pkl
Processing file: 781_YAHOO_id_231_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 781_YAHOO_id_231_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 163.74it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1619.13it/s]
Table1-only metrics completed in 0.37s
Evaluation Result:  {'AUC-PR': 0.9459876543209877, 'AUC-ROC': 0.9993908382066277, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 781_YAHOO_id_231_Synthetic_tr_500_1st_623.csv: {'filename': '781_YAHOO_id_231_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 35, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9459876543209877, 'AUC-ROC': 0.9993908382066277, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 781_YAHOO_id_231_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/781_YAHOO_id_231_Synthetic_tr_500_1st_623_results.pkl
Processing file: 340_UCR_id_38_Facility_tr_32500_1st_114283.csv, Data shape: (130001, 1)
Label shape: (130001,)
Train data shape: (32500, 1), Test data shape: (97501, 1), Label test shape: (97501,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 340_UCR_id_38_Facility_tr_32500_1st_114283.csv...
Using CUDA device: 2
Data shape: (97501, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 97501, window_size 5000, adding 2499 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([20, 5000, 1])
batched_mask shape: torch.Size([20, 5000])

100%|██████████| 1/1 [00:00<00:00, 65.24it/s]
Original shapes - Output: (100000,), Label: (97501,)
Aligned shapes - Output: (97501,), Label: (97501,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  20%|█▉        | 39/200 [00:01<00:04, 38.56it/s]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [00:02<00:03, 35.31it/s]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [00:03<00:02, 35.79it/s]

Calculating Affiliation Metrics:  80%|███████▉  | 159/200 [00:04<00:01, 38.89it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:04<00:00, 41.44it/s]
Table1-only metrics completed in 10.83s
Evaluation Result:  {'AUC-PR': 0.14392043677756106, 'AUC-ROC': 0.9757732674150034, 'VUS-PR': 0.06538076118843801, 'VUS-ROC': 0.9905162513020714, 'Standard-F1': 0.0690316395013178, 'Standard-Precision': 0.03688524590163931, 'Standard-Recall': 0.5373134328358129, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6724248768613167, 'Affiliation-P': 0.5065060840184736, 'Affiliation-R': 1.0, 'F1_T': 0.0690316395013178, 'Precision_T': 0.03688524590163931, 'Recall_T': 0.5373134328358129}
Results for 340_UCR_id_38_Facility_tr_32500_1st_114283.csv: {'filename': '340_UCR_id_38_Facility_tr_32500_1st_114283.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 243, 'train_index': '32500', 'data_shape': '130001x1', 'output_length': 100000, 'label_length': 97501, 'aligned_length': 97501, 'AUC-PR': 0.14392043677756106, 'AUC-ROC': 0.9757732674150034, 'VUS-PR': 0.06538076118843801, 'VUS-ROC': 0.9905162513020714, 'Standard-F1': 0.0690316395013178, 'Standard-Precision': 0.03688524590163931, 'Standard-Recall': 0.5373134328358129, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6724248768613167, 'Affiliation-P': 0.5065060840184736, 'Affiliation-R': 1.0, 'F1_T': 0.0690316395013178, 'Precision_T': 0.03688524590163931, 'Recall_T': 0.5373134328358129}
Logits metrics skipped for 340_UCR_id_38_Facility_tr_32500_1st_114283.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/340_UCR_id_38_Facility_tr_32500_1st_114283_results.pkl
Processing file: 280_NEK_id_4_WebService_tr_500_1st_231.csv, Data shape: (1073, 1)
Label shape: (1073,)
Train data shape: (500, 1), Test data shape: (573, 1), Label test shape: (573,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 280_NEK_id_4_WebService_tr_500_1st_231.csv...
Using CUDA device: 2
Data shape: (573, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 573, 1])
batched_mask shape: torch.Size([1, 573])

100%|██████████| 1/1 [00:00<00:00, 157.44it/s]
Original shapes - Output: (573,), Label: (573,)
Aligned shapes - Output: (573,), Label: (573,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2024.18it/s]
Table1-only metrics completed in 0.29s
Evaluation Result:  {'AUC-PR': 0.20961681248832992, 'AUC-ROC': 0.6523947750362844, 'VUS-PR': 0.25331982635672823, 'VUS-ROC': 0.7715096886914599, 'Standard-F1': 0.35199999999950593, 'Standard-Precision': 0.30555555555555136, 'Standard-Recall': 0.41509433962263365, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.941251292154005, 'Affiliation-P': 0.9401559454191033, 'Affiliation-R': 0.9423491941754811, 'F1_T': 0.3471074380160309, 'Precision_T': 0.30882352941176017, 'Recall_T': 0.39622641509433215}
Results for 280_NEK_id_4_WebService_tr_500_1st_231.csv: {'filename': '280_NEK_id_4_WebService_tr_500_1st_231.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1073x1', 'output_length': 573, 'label_length': 573, 'aligned_length': 573, 'AUC-PR': 0.20961681248832992, 'AUC-ROC': 0.6523947750362844, 'VUS-PR': 0.25331982635672823, 'VUS-ROC': 0.7715096886914599, 'Standard-F1': 0.35199999999950593, 'Standard-Precision': 0.30555555555555136, 'Standard-Recall': 0.41509433962263365, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.941251292154005, 'Affiliation-P': 0.9401559454191033, 'Affiliation-R': 0.9423491941754811, 'F1_T': 0.3471074380160309, 'Precision_T': 0.30882352941176017, 'Recall_T': 0.39622641509433215}
Logits metrics skipped for 280_NEK_id_4_WebService_tr_500_1st_231.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/280_NEK_id_4_WebService_tr_500_1st_231_results.pkl
Processing file: 139_WSD_id_111_WebService_tr_1701_1st_1801.csv, Data shape: (18381, 1)
Label shape: (18381,)
Train data shape: (1701, 1), Test data shape: (16680, 1), Label test shape: (16680,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 139_WSD_id_111_WebService_tr_1701_1st_1801.csv...
Using CUDA device: 2
Data shape: (16680, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 16680, window_size 5000, adding 3320 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.44it/s]
Original shapes - Output: (20000,), Label: (16680,)
Aligned shapes - Output: (16680,), Label: (16680,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 177/200 [00:01<00:00, 176.99it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 187.44it/s]
Table1-only metrics completed in 3.43s
Evaluation Result:  {'AUC-PR': 0.8287793632645909, 'AUC-ROC': 0.9776882787437716, 'VUS-PR': 0.8443729608574094, 'VUS-ROC': 0.9986894084345337, 'Standard-F1': 0.5905511811019072, 'Standard-Precision': 0.44910179640718295, 'Standard-Recall': 0.8620689655172316, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.952478219002508, 'Affiliation-P': 0.9102664964595097, 'Affiliation-R': 0.9987952804563407, 'F1_T': 0.5905511811019072, 'Precision_T': 0.44910179640718295, 'Recall_T': 0.8620689655172316}
Results for 139_WSD_id_111_WebService_tr_1701_1st_1801.csv: {'filename': '139_WSD_id_111_WebService_tr_1701_1st_1801.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '1701', 'data_shape': '18381x1', 'output_length': 20000, 'label_length': 16680, 'aligned_length': 16680, 'AUC-PR': 0.8287793632645909, 'AUC-ROC': 0.9776882787437716, 'VUS-PR': 0.8443729608574094, 'VUS-ROC': 0.9986894084345337, 'Standard-F1': 0.5905511811019072, 'Standard-Precision': 0.44910179640718295, 'Standard-Recall': 0.8620689655172316, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.952478219002508, 'Affiliation-P': 0.9102664964595097, 'Affiliation-R': 0.9987952804563407, 'F1_T': 0.5905511811019072, 'Precision_T': 0.44910179640718295, 'Recall_T': 0.8620689655172316}
Logits metrics skipped for 139_WSD_id_111_WebService_tr_1701_1st_1801.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/139_WSD_id_111_WebService_tr_1701_1st_1801_results.pkl
Processing file: 160_Stock_id_12_Finance_tr_500_1st_14.csv, Data shape: (20000, 1)
Label shape: (20000,)
Train data shape: (500, 1), Test data shape: (19500, 1), Label test shape: (19500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 160_Stock_id_12_Finance_tr_500_1st_14.csv...
Using CUDA device: 2
Data shape: (19500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 19500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.92it/s]
Original shapes - Output: (20000,), Label: (19500,)
Aligned shapes - Output: (19500,), Label: (19500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   4%|▍         | 8/200 [00:01<00:25,  7.63it/s]

Calculating Affiliation Metrics:   8%|▊         | 16/200 [00:03<00:42,  4.38it/s]

Calculating Affiliation Metrics:  11%|█         | 22/200 [00:05<00:49,  3.60it/s]

Calculating Affiliation Metrics:  14%|█▎        | 27/200 [00:07<00:53,  3.24it/s]

Calculating Affiliation Metrics:  16%|█▌        | 31/200 [00:08<00:55,  3.06it/s]

Calculating Affiliation Metrics:  18%|█▊        | 35/200 [00:10<00:56,  2.90it/s]

Calculating Affiliation Metrics:  19%|█▉        | 38/200 [00:11<00:57,  2.80it/s]

Calculating Affiliation Metrics:  20%|██        | 41/200 [00:12<00:58,  2.72it/s]

Calculating Affiliation Metrics:  22%|██▏       | 44/200 [00:14<00:59,  2.61it/s]

Calculating Affiliation Metrics:  24%|██▎       | 47/200 [00:15<01:00,  2.54it/s]

Calculating Affiliation Metrics:  25%|██▌       | 50/200 [00:16<01:00,  2.49it/s]

Calculating Affiliation Metrics:  26%|██▋       | 53/200 [00:17<00:59,  2.47it/s]

Calculating Affiliation Metrics:  28%|██▊       | 56/200 [00:19<00:59,  2.44it/s]

Calculating Affiliation Metrics:  30%|██▉       | 59/200 [00:20<00:58,  2.42it/s]

Calculating Affiliation Metrics:  31%|███       | 62/200 [00:21<00:57,  2.39it/s]

Calculating Affiliation Metrics:  32%|███▎      | 65/200 [00:23<00:57,  2.35it/s]

Calculating Affiliation Metrics:  34%|███▍      | 68/200 [00:24<00:56,  2.33it/s]

Calculating Affiliation Metrics:  36%|███▌      | 71/200 [00:25<00:56,  2.29it/s]

Calculating Affiliation Metrics:  37%|███▋      | 74/200 [00:27<00:56,  2.24it/s]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [00:28<00:56,  2.19it/s]

Calculating Affiliation Metrics:  40%|████      | 80/200 [00:30<00:56,  2.13it/s]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [00:31<00:57,  2.04it/s]

Calculating Affiliation Metrics:  43%|████▎     | 86/200 [00:33<00:57,  1.97it/s]

Calculating Affiliation Metrics:  44%|████▍     | 88/200 [00:34<00:57,  1.93it/s]

Calculating Affiliation Metrics:  45%|████▌     | 90/200 [00:35<00:57,  1.91it/s]

Calculating Affiliation Metrics:  46%|████▌     | 92/200 [00:36<00:56,  1.89it/s]

Calculating Affiliation Metrics:  47%|████▋     | 94/200 [00:37<00:56,  1.88it/s]

Calculating Affiliation Metrics:  48%|████▊     | 96/200 [00:38<00:55,  1.86it/s]

Calculating Affiliation Metrics:  49%|████▉     | 98/200 [00:40<00:55,  1.83it/s]

Calculating Affiliation Metrics:  50%|█████     | 100/200 [00:41<00:54,  1.82it/s]

Calculating Affiliation Metrics:  51%|█████     | 102/200 [00:42<00:54,  1.81it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 104/200 [00:43<00:53,  1.81it/s]

Calculating Affiliation Metrics:  53%|█████▎    | 106/200 [00:44<00:51,  1.81it/s]

Calculating Affiliation Metrics:  54%|█████▍    | 108/200 [00:45<00:50,  1.81it/s]

Calculating Affiliation Metrics:  55%|█████▌    | 110/200 [00:46<00:49,  1.81it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [00:47<00:48,  1.81it/s]

Calculating Affiliation Metrics:  57%|█████▋    | 114/200 [00:48<00:47,  1.81it/s]

Calculating Affiliation Metrics:  58%|█████▊    | 116/200 [00:50<00:46,  1.80it/s]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [00:51<00:46,  1.77it/s]

Calculating Affiliation Metrics:  60%|██████    | 120/200 [00:52<00:46,  1.74it/s]

Calculating Affiliation Metrics:  61%|██████    | 122/200 [00:53<00:45,  1.72it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 124/200 [00:54<00:44,  1.70it/s]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [00:55<00:43,  1.69it/s]

Calculating Affiliation Metrics:  64%|██████▍   | 128/200 [00:57<00:42,  1.69it/s]

Calculating Affiliation Metrics:  65%|██████▌   | 130/200 [00:58<00:41,  1.68it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 132/200 [00:59<00:40,  1.69it/s]

Calculating Affiliation Metrics:  67%|██████▋   | 134/200 [01:00<00:38,  1.71it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 136/200 [01:01<00:37,  1.73it/s]

Calculating Affiliation Metrics:  69%|██████▉   | 138/200 [01:02<00:35,  1.77it/s]

Calculating Affiliation Metrics:  70%|███████   | 140/200 [01:03<00:32,  1.82it/s]

Calculating Affiliation Metrics:  71%|███████   | 142/200 [01:04<00:31,  1.86it/s]

Calculating Affiliation Metrics:  72%|███████▏  | 144/200 [01:05<00:29,  1.89it/s]

Calculating Affiliation Metrics:  74%|███████▎  | 147/200 [01:07<00:26,  2.02it/s]

Calculating Affiliation Metrics:  75%|███████▌  | 150/200 [01:08<00:23,  2.13it/s]

Calculating Affiliation Metrics:  76%|███████▋  | 153/200 [01:09<00:20,  2.24it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 156/200 [01:10<00:18,  2.37it/s]

Calculating Affiliation Metrics:  80%|███████▉  | 159/200 [01:11<00:16,  2.50it/s]

Calculating Affiliation Metrics:  82%|████████▏ | 163/200 [01:13<00:13,  2.69it/s]

Calculating Affiliation Metrics:  84%|████████▎ | 167/200 [01:14<00:11,  2.88it/s]

Calculating Affiliation Metrics:  86%|████████▌ | 171/200 [01:15<00:09,  3.13it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 176/200 [01:16<00:06,  3.49it/s]

Calculating Affiliation Metrics:  91%|█████████ | 182/200 [01:17<00:04,  4.04it/s]

Calculating Affiliation Metrics:  95%|█████████▌| 190/200 [01:18<00:01,  5.06it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [01:19<00:00,  2.52it/s]
Table1-only metrics completed in 107.40s
Evaluation Result:  {'AUC-PR': 0.13992974961388083, 'AUC-ROC': 0.6448777471948844, 'VUS-PR': 0.6486359235420938, 'VUS-ROC': 0.8528736937461433, 'Standard-F1': 0.19418960244609165, 'Standard-Precision': 0.36389684813753476, 'Standard-Recall': 0.13242961418143886, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6784890169420454, 'Affiliation-P': 0.5233288150065958, 'Affiliation-R': 0.9644299938263776, 'F1_T': 0.1926541926537466, 'Precision_T': 0.2871900826446275, 'Recall_T': 0.14494264859228348}
Results for 160_Stock_id_12_Finance_tr_500_1st_14.csv: {'filename': '160_Stock_id_12_Finance_tr_500_1st_14.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 47, 'train_index': '500', 'data_shape': '20000x1', 'output_length': 20000, 'label_length': 19500, 'aligned_length': 19500, 'AUC-PR': 0.13992974961388083, 'AUC-ROC': 0.6448777471948844, 'VUS-PR': 0.6486359235420938, 'VUS-ROC': 0.8528736937461433, 'Standard-F1': 0.19418960244609165, 'Standard-Precision': 0.36389684813753476, 'Standard-Recall': 0.13242961418143886, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6784890169420454, 'Affiliation-P': 0.5233288150065958, 'Affiliation-R': 0.9644299938263776, 'F1_T': 0.1926541926537466, 'Precision_T': 0.2871900826446275, 'Recall_T': 0.14494264859228348}
Logits metrics skipped for 160_Stock_id_12_Finance_tr_500_1st_14.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/160_Stock_id_12_Finance_tr_500_1st_14_results.pkl
Processing file: 767_YAHOO_id_217_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 767_YAHOO_id_217_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 157.65it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1895.58it/s]
Table1-only metrics completed in 0.95s
Evaluation Result:  {'AUC-PR': 0.95, 'AUC-ROC': 0.9997273718647764, 'VUS-PR': 0.992632013705543, 'VUS-ROC': 0.9999757092987542, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 767_YAHOO_id_217_Synthetic_tr_500_1st_658.csv: {'filename': '767_YAHOO_id_217_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 157, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.95, 'AUC-ROC': 0.9997273718647764, 'VUS-PR': 0.992632013705543, 'VUS-ROC': 0.9999757092987542, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 767_YAHOO_id_217_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/767_YAHOO_id_217_Synthetic_tr_500_1st_658_results.pkl
Processing file: 064_WSD_id_36_WebService_tr_4559_1st_10195.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 064_WSD_id_36_WebService_tr_4559_1st_10195.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.90it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 303.01it/s]
Table1-only metrics completed in 1.24s
Evaluation Result:  {'AUC-PR': 0.33572863517139406, 'AUC-ROC': 0.8751245238792851, 'VUS-PR': 0.17288345725886187, 'VUS-ROC': 0.9693358168536159, 'Standard-F1': 0.13580246913553978, 'Standard-Precision': 0.08029197080291912, 'Standard-Recall': 0.4399999999999824, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7172754726229742, 'Affiliation-P': 0.5591954004970582, 'Affiliation-R': 0.9999546684214374, 'F1_T': 0.13580246913553978, 'Precision_T': 0.08029197080291912, 'Recall_T': 0.4399999999999824}
Results for 064_WSD_id_36_WebService_tr_4559_1st_10195.csv: {'filename': '064_WSD_id_36_WebService_tr_4559_1st_10195.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.33572863517139406, 'AUC-ROC': 0.8751245238792851, 'VUS-PR': 0.17288345725886187, 'VUS-ROC': 0.9693358168536159, 'Standard-F1': 0.13580246913553978, 'Standard-Precision': 0.08029197080291912, 'Standard-Recall': 0.4399999999999824, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7172754726229742, 'Affiliation-P': 0.5591954004970582, 'Affiliation-R': 0.9999546684214374, 'F1_T': 0.13580246913553978, 'Precision_T': 0.08029197080291912, 'Recall_T': 0.4399999999999824}
Logits metrics skipped for 064_WSD_id_36_WebService_tr_4559_1st_10195.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/064_WSD_id_36_WebService_tr_4559_1st_10195_results.pkl
Processing file: 573_YAHOO_id_23_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 573_YAHOO_id_23_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 159.97it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1816.27it/s]
Table1-only metrics completed in 0.35s
Evaluation Result:  {'AUC-PR': 0.9085858585858586, 'AUC-ROC': 0.9989035087719298, 'VUS-PR': 0.9820605815874914, 'VUS-ROC': 0.99983402391278, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9969368810069911, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.9997412008281573, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Results for 573_YAHOO_id_23_Synthetic_tr_500_1st_623.csv: {'filename': '573_YAHOO_id_23_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 35, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9085858585858586, 'AUC-ROC': 0.9989035087719298, 'VUS-PR': 0.9820605815874914, 'VUS-ROC': 0.99983402391278, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9969368810069911, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.9997412008281573, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 573_YAHOO_id_23_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/573_YAHOO_id_23_Synthetic_tr_500_1st_623_results.pkl
Processing file: 308_UCR_id_6_Facility_tr_20000_1st_56600.csv, Data shape: (80001, 1)
Label shape: (80001,)
Train data shape: (20000, 1), Test data shape: (60001, 1), Label test shape: (60001,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 308_UCR_id_6_Facility_tr_20000_1st_56600.csv...
Using CUDA device: 2
Data shape: (60001, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 60001, window_size 5000, adding 4999 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([13, 5000, 1])
batched_mask shape: torch.Size([13, 5000])

100%|██████████| 1/1 [00:00<00:00, 86.46it/s]
Original shapes - Output: (65000,), Label: (60001,)
Aligned shapes - Output: (60001,), Label: (60001,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  32%|███▏      | 64/200 [00:01<00:02, 63.14it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 132/200 [00:02<00:01, 65.76it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 74.06it/s]
Table1-only metrics completed in 5.08s
Evaluation Result:  {'AUC-PR': 0.006252003880863282, 'AUC-ROC': 0.6336641485625589, 'VUS-PR': 0.007170579283609837, 'VUS-ROC': 0.6530229973838714, 'Standard-F1': 0.019564312473963032, 'Standard-Precision': 0.009885025238362311, 'Standard-Recall': 0.9399999999999968, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.774766327319183, 'Affiliation-P': 0.6389951880175846, 'Affiliation-R': 0.9838002699955001, 'F1_T': 0.019476525661156503, 'Precision_T': 0.009839121728923565, 'Recall_T': 0.9499999999999967}
Results for 308_UCR_id_6_Facility_tr_20000_1st_56600.csv: {'filename': '308_UCR_id_6_Facility_tr_20000_1st_56600.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 208, 'train_index': '20000', 'data_shape': '80001x1', 'output_length': 65000, 'label_length': 60001, 'aligned_length': 60001, 'AUC-PR': 0.006252003880863282, 'AUC-ROC': 0.6336641485625589, 'VUS-PR': 0.007170579283609837, 'VUS-ROC': 0.6530229973838714, 'Standard-F1': 0.019564312473963032, 'Standard-Precision': 0.009885025238362311, 'Standard-Recall': 0.9399999999999968, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.774766327319183, 'Affiliation-P': 0.6389951880175846, 'Affiliation-R': 0.9838002699955001, 'F1_T': 0.019476525661156503, 'Precision_T': 0.009839121728923565, 'Recall_T': 0.9499999999999967}
Logits metrics skipped for 308_UCR_id_6_Facility_tr_20000_1st_56600.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/308_UCR_id_6_Facility_tr_20000_1st_56600_results.pkl
Processing file: 059_WSD_id_31_WebService_tr_4559_1st_9058.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 059_WSD_id_31_WebService_tr_4559_1st_9058.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.87it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 315.06it/s]
Table1-only metrics completed in 0.82s
Evaluation Result:  {'AUC-PR': 0.08492953909975234, 'AUC-ROC': 0.456580720773513, 'VUS-PR': 0.04705517819483831, 'VUS-ROC': 0.5425318415532344, 'Standard-F1': 0.037037037036832005, 'Standard-Precision': 0.02094240837696324, 'Standard-Recall': 0.15999999999999362, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666359000135664, 'Affiliation-P': 0.4999653883139271, 'Affiliation-R': 1.0, 'F1_T': 0.03703703703677557, 'Precision_T': 0.021897810218977944, 'Recall_T': 0.11999999999999521}
Results for 059_WSD_id_31_WebService_tr_4559_1st_9058.csv: {'filename': '059_WSD_id_31_WebService_tr_4559_1st_9058.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.08492953909975234, 'AUC-ROC': 0.456580720773513, 'VUS-PR': 0.04705517819483831, 'VUS-ROC': 0.5425318415532344, 'Standard-F1': 0.037037037036832005, 'Standard-Precision': 0.02094240837696324, 'Standard-Recall': 0.15999999999999362, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666359000135664, 'Affiliation-P': 0.4999653883139271, 'Affiliation-R': 1.0, 'F1_T': 0.03703703703677557, 'Precision_T': 0.021897810218977944, 'Recall_T': 0.11999999999999521}
Logits metrics skipped for 059_WSD_id_31_WebService_tr_4559_1st_9058.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/059_WSD_id_31_WebService_tr_4559_1st_9058_results.pkl
Processing file: 013_NAB_id_13_Traffic_tr_623_1st_2084.csv, Data shape: (2494, 1)
Label shape: (2494,)
Train data shape: (623, 1), Test data shape: (1871, 1), Label test shape: (1871,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 013_NAB_id_13_Traffic_tr_623_1st_2084.csv...
Using CUDA device: 2
Data shape: (1871, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1871, 1])
batched_mask shape: torch.Size([1, 1871])

100%|██████████| 1/1 [00:00<00:00, 158.57it/s]
Original shapes - Output: (1871,), Label: (1871,)
Aligned shapes - Output: (1871,), Label: (1871,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1323.67it/s]
Table1-only metrics completed in 1.18s
Evaluation Result:  {'AUC-PR': 0.5496589168653463, 'AUC-ROC': 0.834960646353825, 'VUS-PR': 0.656577327472884, 'VUS-ROC': 0.9118583821751965, 'Standard-F1': 0.5126353790608754, 'Standard-Precision': 0.4640522875816978, 'Standard-Recall': 0.572580645161288, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9520675443373032, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9085199520185021, 'F1_T': 0.5135623869796119, 'Precision_T': 0.4655737704918017, 'Recall_T': 0.572580645161288}
Results for 013_NAB_id_13_Traffic_tr_623_1st_2084.csv: {'filename': '013_NAB_id_13_Traffic_tr_623_1st_2084.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 247, 'train_index': '623', 'data_shape': '2494x1', 'output_length': 1871, 'label_length': 1871, 'aligned_length': 1871, 'AUC-PR': 0.5496589168653463, 'AUC-ROC': 0.834960646353825, 'VUS-PR': 0.656577327472884, 'VUS-ROC': 0.9118583821751965, 'Standard-F1': 0.5126353790608754, 'Standard-Precision': 0.4640522875816978, 'Standard-Recall': 0.572580645161288, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9520675443373032, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9085199520185021, 'F1_T': 0.5135623869796119, 'Precision_T': 0.4655737704918017, 'Recall_T': 0.572580645161288}
Logits metrics skipped for 013_NAB_id_13_Traffic_tr_623_1st_2084.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/013_NAB_id_13_Traffic_tr_623_1st_2084_results.pkl
Processing file: 498_UCR_id_196_Medical_tr_15000_1st_51370.csv, Data shape: (60000, 1)
Label shape: (60000,)
Train data shape: (15000, 1), Test data shape: (45000, 1), Label test shape: (45000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 498_UCR_id_196_Medical_tr_15000_1st_51370.csv...
Using CUDA device: 2
Data shape: (45000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([9, 5000, 1])
batched_mask shape: torch.Size([9, 5000])

100%|██████████| 1/1 [00:00<00:00, 112.18it/s]
Original shapes - Output: (45000,), Label: (45000,)
Aligned shapes - Output: (45000,), Label: (45000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  41%|████      | 82/200 [00:01<00:01, 81.52it/s]

Calculating Affiliation Metrics:  83%|████████▎ | 166/200 [00:02<00:00, 82.62it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 88.54it/s]
Table1-only metrics completed in 3.63s
Evaluation Result:  {'AUC-PR': 0.06719465772100816, 'AUC-ROC': 0.9396610569790045, 'VUS-PR': 0.08094092447164758, 'VUS-ROC': 0.9498545955192793, 'Standard-F1': 0.15907590759054457, 'Standard-Precision': 0.09060150375939846, 'Standard-Recall': 0.6513513513513496, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6783503761730757, 'Affiliation-P': 0.513260370936185, 'Affiliation-R': 1.0, 'F1_T': 0.1587092525516463, 'Precision_T': 0.09036370453693285, 'Recall_T': 0.6513513513513496}
Results for 498_UCR_id_196_Medical_tr_15000_1st_51370.csv: {'filename': '498_UCR_id_196_Medical_tr_15000_1st_51370.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 172, 'train_index': '15000', 'data_shape': '60000x1', 'output_length': 45000, 'label_length': 45000, 'aligned_length': 45000, 'AUC-PR': 0.06719465772100816, 'AUC-ROC': 0.9396610569790045, 'VUS-PR': 0.08094092447164758, 'VUS-ROC': 0.9498545955192793, 'Standard-F1': 0.15907590759054457, 'Standard-Precision': 0.09060150375939846, 'Standard-Recall': 0.6513513513513496, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6783503761730757, 'Affiliation-P': 0.513260370936185, 'Affiliation-R': 1.0, 'F1_T': 0.1587092525516463, 'Precision_T': 0.09036370453693285, 'Recall_T': 0.6513513513513496}
Logits metrics skipped for 498_UCR_id_196_Medical_tr_15000_1st_51370.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/498_UCR_id_196_Medical_tr_15000_1st_51370_results.pkl
Processing file: 519_UCR_id_217_Medical_tr_7516_1st_17390.csv, Data shape: (30066, 1)
Label shape: (30066,)
Train data shape: (7516, 1), Test data shape: (22550, 1), Label test shape: (22550,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 519_UCR_id_217_Medical_tr_7516_1st_17390.csv...
Using CUDA device: 2
Data shape: (22550, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22550, window_size 5000, adding 2450 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 147.46it/s]
Original shapes - Output: (25000,), Label: (22550,)
Aligned shapes - Output: (22550,), Label: (22550,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [00:01<00:00, 117.45it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 129.41it/s]
Table1-only metrics completed in 1.81s
Evaluation Result:  {'AUC-PR': 0.16613397448063577, 'AUC-ROC': 0.8239590338296852, 'VUS-PR': 0.13479373716934032, 'VUS-ROC': 0.8454499697651122, 'Standard-F1': 0.2528089887635799, 'Standard-Precision': 0.19911504424778673, 'Standard-Recall': 0.3461538461538435, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7564590284619376, 'Affiliation-P': 0.608565954798304, 'Affiliation-R': 0.999310421286031, 'F1_T': 0.2528089887635799, 'Precision_T': 0.19911504424778673, 'Recall_T': 0.3461538461538435}
Results for 519_UCR_id_217_Medical_tr_7516_1st_17390.csv: {'filename': '519_UCR_id_217_Medical_tr_7516_1st_17390.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 34, 'train_index': '7516', 'data_shape': '30066x1', 'output_length': 25000, 'label_length': 22550, 'aligned_length': 22550, 'AUC-PR': 0.16613397448063577, 'AUC-ROC': 0.8239590338296852, 'VUS-PR': 0.13479373716934032, 'VUS-ROC': 0.8454499697651122, 'Standard-F1': 0.2528089887635799, 'Standard-Precision': 0.19911504424778673, 'Standard-Recall': 0.3461538461538435, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7564590284619376, 'Affiliation-P': 0.608565954798304, 'Affiliation-R': 0.999310421286031, 'F1_T': 0.2528089887635799, 'Precision_T': 0.19911504424778673, 'Recall_T': 0.3461538461538435}
Logits metrics skipped for 519_UCR_id_217_Medical_tr_7516_1st_17390.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/519_UCR_id_217_Medical_tr_7516_1st_17390_results.pkl
Skipping file: 184_SMD_id_7_Facility_tr_7119_1st_15849.csv due to filter criteria.
Processing file: 261_IOPS_id_2_WebService_tr_2178_1st_2278.csv, Data shape: (8784, 1)
Label shape: (8784,)
Train data shape: (2178, 1), Test data shape: (6606, 1), Label test shape: (6606,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 261_IOPS_id_2_WebService_tr_2178_1st_2278.csv...
Using CUDA device: 2
Data shape: (6606, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6606, window_size 5000, adding 3394 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.71it/s]
Original shapes - Output: (10000,), Label: (6606,)
Aligned shapes - Output: (6606,), Label: (6606,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 573.33it/s]
Table1-only metrics completed in 2.06s
Evaluation Result:  {'AUC-PR': 0.18642666627210308, 'AUC-ROC': 0.9532575757575756, 'VUS-PR': 0.10850020632996844, 'VUS-ROC': 0.9536679292929247, 'Standard-F1': 0.05479452054779284, 'Standard-Precision': 0.029850746268656272, 'Standard-Recall': 0.33333333333327775, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7372919179525982, 'Affiliation-P': 0.5838973618962876, 'Affiliation-R': 1.0, 'F1_T': 0.05479452054779284, 'Precision_T': 0.029850746268656272, 'Recall_T': 0.33333333333327775}
Results for 261_IOPS_id_2_WebService_tr_2178_1st_2278.csv: {'filename': '261_IOPS_id_2_WebService_tr_2178_1st_2278.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 288, 'train_index': '2178', 'data_shape': '8784x1', 'output_length': 10000, 'label_length': 6606, 'aligned_length': 6606, 'AUC-PR': 0.18642666627210308, 'AUC-ROC': 0.9532575757575756, 'VUS-PR': 0.10850020632996844, 'VUS-ROC': 0.9536679292929247, 'Standard-F1': 0.05479452054779284, 'Standard-Precision': 0.029850746268656272, 'Standard-Recall': 0.33333333333327775, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7372919179525982, 'Affiliation-P': 0.5838973618962876, 'Affiliation-R': 1.0, 'F1_T': 0.05479452054779284, 'Precision_T': 0.029850746268656272, 'Recall_T': 0.33333333333327775}
Logits metrics skipped for 261_IOPS_id_2_WebService_tr_2178_1st_2278.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/261_IOPS_id_2_WebService_tr_2178_1st_2278_results.pkl
Processing file: 354_UCR_id_52_Facility_tr_7482_1st_18485.csv, Data shape: (29931, 1)
Label shape: (29931,)
Train data shape: (7482, 1), Test data shape: (22449, 1), Label test shape: (22449,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 354_UCR_id_52_Facility_tr_7482_1st_18485.csv...
Using CUDA device: 2
Data shape: (22449, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22449, window_size 5000, adding 2551 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 149.66it/s]
Original shapes - Output: (25000,), Label: (22449,)
Aligned shapes - Output: (22449,), Label: (22449,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  64%|██████▍   | 129/200 [00:01<00:00, 128.24it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 150.69it/s]
Table1-only metrics completed in 2.27s
Evaluation Result:  {'AUC-PR': 0.045917112112793465, 'AUC-ROC': 0.8575882157231363, 'VUS-PR': 0.047069163251470814, 'VUS-ROC': 0.858935541031706, 'Standard-F1': 0.11971679897004424, 'Standard-Precision': 0.0645086705202312, 'Standard-Recall': 0.8303571428571404, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6905367895369547, 'Affiliation-P': 0.5273444199545309, 'Affiliation-R': 0.9999963541699809, 'F1_T': 0.11911416899578076, 'Precision_T': 0.06419466975666278, 'Recall_T': 0.8244047619047594}
Results for 354_UCR_id_52_Facility_tr_7482_1st_18485.csv: {'filename': '354_UCR_id_52_Facility_tr_7482_1st_18485.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 168, 'train_index': '7482', 'data_shape': '29931x1', 'output_length': 25000, 'label_length': 22449, 'aligned_length': 22449, 'AUC-PR': 0.045917112112793465, 'AUC-ROC': 0.8575882157231363, 'VUS-PR': 0.047069163251470814, 'VUS-ROC': 0.858935541031706, 'Standard-F1': 0.11971679897004424, 'Standard-Precision': 0.0645086705202312, 'Standard-Recall': 0.8303571428571404, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6905367895369547, 'Affiliation-P': 0.5273444199545309, 'Affiliation-R': 0.9999963541699809, 'F1_T': 0.11911416899578076, 'Precision_T': 0.06419466975666278, 'Recall_T': 0.8244047619047594}
Logits metrics skipped for 354_UCR_id_52_Facility_tr_7482_1st_18485.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/354_UCR_id_52_Facility_tr_7482_1st_18485_results.pkl
Processing file: 104_WSD_id_76_WebService_tr_3047_1st_3147.csv, Data shape: (18084, 1)
Label shape: (18084,)
Train data shape: (3047, 1), Test data shape: (15037, 1), Label test shape: (15037,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 104_WSD_id_76_WebService_tr_3047_1st_3147.csv...
Using CUDA device: 2
Data shape: (15037, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 15037, window_size 5000, adding 4963 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.33it/s]
Original shapes - Output: (20000,), Label: (15037,)
Aligned shapes - Output: (15037,), Label: (15037,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 131/200 [00:01<00:00, 130.31it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 146.30it/s]
Table1-only metrics completed in 4.79s
Evaluation Result:  {'AUC-PR': 0.29379534823487186, 'AUC-ROC': 0.9043174424844054, 'VUS-PR': 0.3703099452805752, 'VUS-ROC': 0.9816244556506228, 'Standard-F1': 0.23655913978444795, 'Standard-Precision': 0.21854304635761446, 'Standard-Recall': 0.257812499999998, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9873462369247643, 'Affiliation-P': 0.9803635279315239, 'Affiliation-R': 0.9944291291637318, 'F1_T': 0.23655913978444795, 'Precision_T': 0.21854304635761446, 'Recall_T': 0.257812499999998}
Results for 104_WSD_id_76_WebService_tr_3047_1st_3147.csv: {'filename': '104_WSD_id_76_WebService_tr_3047_1st_3147.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '3047', 'data_shape': '18084x1', 'output_length': 20000, 'label_length': 15037, 'aligned_length': 15037, 'AUC-PR': 0.29379534823487186, 'AUC-ROC': 0.9043174424844054, 'VUS-PR': 0.3703099452805752, 'VUS-ROC': 0.9816244556506228, 'Standard-F1': 0.23655913978444795, 'Standard-Precision': 0.21854304635761446, 'Standard-Recall': 0.257812499999998, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9873462369247643, 'Affiliation-P': 0.9803635279315239, 'Affiliation-R': 0.9944291291637318, 'F1_T': 0.23655913978444795, 'Precision_T': 0.21854304635761446, 'Recall_T': 0.257812499999998}
Logits metrics skipped for 104_WSD_id_76_WebService_tr_3047_1st_3147.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/104_WSD_id_76_WebService_tr_3047_1st_3147_results.pkl
Processing file: 299_TODS_id_13_Synthetic_tr_500_1st_65.csv, Data shape: (5000, 1)
Label shape: (5000,)
Train data shape: (500, 1), Test data shape: (4500, 1), Label test shape: (4500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 299_TODS_id_13_Synthetic_tr_500_1st_65.csv...
Using CUDA device: 2
Data shape: (4500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 4500, 1])
batched_mask shape: torch.Size([1, 4500])

100%|██████████| 1/1 [00:00<00:00, 156.84it/s]
Original shapes - Output: (4500,), Label: (4500,)
Aligned shapes - Output: (4500,), Label: (4500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:01<00:02, 53.05it/s]

Calculating Affiliation Metrics:  54%|█████▍    | 108/200 [00:02<00:02, 34.86it/s]

Calculating Affiliation Metrics:  74%|███████▎  | 147/200 [00:04<00:01, 33.49it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:04<00:00, 40.74it/s]
Table1-only metrics completed in 11.45s
Evaluation Result:  {'AUC-PR': 0.6979781171940918, 'AUC-ROC': 0.9383320911661622, 'VUS-PR': 0.8890414098064117, 'VUS-ROC': 0.9853573784861578, 'Standard-F1': 0.6533333333328318, 'Standard-Precision': 0.6901408450704193, 'Standard-Recall': 0.6202531645569594, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8837650382429044, 'Affiliation-P': 0.8844729117274429, 'Affiliation-R': 0.8830582969222786, 'F1_T': 0.6484018264835186, 'Precision_T': 0.7064676616915387, 'Recall_T': 0.5991561181434574}
Results for 299_TODS_id_13_Synthetic_tr_500_1st_65.csv: {'filename': '299_TODS_id_13_Synthetic_tr_500_1st_65.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '500', 'data_shape': '5000x1', 'output_length': 4500, 'label_length': 4500, 'aligned_length': 4500, 'AUC-PR': 0.6979781171940918, 'AUC-ROC': 0.9383320911661622, 'VUS-PR': 0.8890414098064117, 'VUS-ROC': 0.9853573784861578, 'Standard-F1': 0.6533333333328318, 'Standard-Precision': 0.6901408450704193, 'Standard-Recall': 0.6202531645569594, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8837650382429044, 'Affiliation-P': 0.8844729117274429, 'Affiliation-R': 0.8830582969222786, 'F1_T': 0.6484018264835186, 'Precision_T': 0.7064676616915387, 'Recall_T': 0.5991561181434574}
Logits metrics skipped for 299_TODS_id_13_Synthetic_tr_500_1st_65.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/299_TODS_id_13_Synthetic_tr_500_1st_65_results.pkl
Processing file: 098_WSD_id_70_WebService_tr_3575_1st_3675.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (3575, 1), Test data shape: (14661, 1), Label test shape: (14661,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 098_WSD_id_70_WebService_tr_3575_1st_3675.csv...
Using CUDA device: 2
Data shape: (14661, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 14661, window_size 5000, adding 339 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.58it/s]
Original shapes - Output: (15000,), Label: (14661,)
Aligned shapes - Output: (14661,), Label: (14661,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 247.15it/s]
Table1-only metrics completed in 1.93s
Evaluation Result:  {'AUC-PR': 0.35060615072430257, 'AUC-ROC': 0.9589134589134589, 'VUS-PR': 0.07181378659818964, 'VUS-ROC': 0.9879372251960574, 'Standard-F1': 0.0648648648647716, 'Standard-Precision': 0.0340909090909089, 'Standard-Recall': 0.6666666666665926, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6895833087266705, 'Affiliation-P': 0.5262320858082207, 'Affiliation-R': 1.0, 'F1_T': 0.06410256410245456, 'Precision_T': 0.03401360544217664, 'Recall_T': 0.5555555555554939}
Results for 098_WSD_id_70_WebService_tr_3575_1st_3675.csv: {'filename': '098_WSD_id_70_WebService_tr_3575_1st_3675.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '3575', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 14661, 'aligned_length': 14661, 'AUC-PR': 0.35060615072430257, 'AUC-ROC': 0.9589134589134589, 'VUS-PR': 0.07181378659818964, 'VUS-ROC': 0.9879372251960574, 'Standard-F1': 0.0648648648647716, 'Standard-Precision': 0.0340909090909089, 'Standard-Recall': 0.6666666666665926, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6895833087266705, 'Affiliation-P': 0.5262320858082207, 'Affiliation-R': 1.0, 'F1_T': 0.06410256410245456, 'Precision_T': 0.03401360544217664, 'Recall_T': 0.5555555555554939}
Logits metrics skipped for 098_WSD_id_70_WebService_tr_3575_1st_3675.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/098_WSD_id_70_WebService_tr_3575_1st_3675_results.pkl
Processing file: 375_UCR_id_73_Medical_tr_47500_1st_123000.csv, Data shape: (190001, 1)
Label shape: (190001,)
Train data shape: (47500, 1), Test data shape: (142501, 1), Label test shape: (142501,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 375_UCR_id_73_Medical_tr_47500_1st_123000.csv...
Using CUDA device: 2
Data shape: (142501, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 142501, window_size 5000, adding 2499 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([29, 5000, 1])
batched_mask shape: torch.Size([29, 5000])

100%|██████████| 1/1 [00:00<00:00, 47.98it/s]
Original shapes - Output: (145000,), Label: (142501,)
Aligned shapes - Output: (142501,), Label: (142501,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  14%|█▎        | 27/200 [00:01<00:06, 26.05it/s]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:02<00:06, 22.66it/s]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [00:03<00:05, 22.65it/s]

Calculating Affiliation Metrics:  51%|█████     | 102/200 [00:04<00:04, 23.40it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 132/200 [00:05<00:02, 25.54it/s]

Calculating Affiliation Metrics:  86%|████████▌ | 171/200 [00:06<00:00, 29.89it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:06<00:00, 28.98it/s]
Table1-only metrics completed in 14.66s
Evaluation Result:  {'AUC-PR': 0.051331505269766514, 'AUC-ROC': 0.9782612522649863, 'VUS-PR': 0.05289180134948137, 'VUS-ROC': 0.9801326315262129, 'Standard-F1': 0.10690161527149859, 'Standard-Precision': 0.058615136876006425, 'Standard-Recall': 0.6066666666666646, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666664502706257, 'Affiliation-P': 0.49999975655449386, 'Affiliation-R': 1.0, 'F1_T': 0.10597755155297489, 'Precision_T': 0.057490795808552805, 'Recall_T': 0.6766666666666643}
Results for 375_UCR_id_73_Medical_tr_47500_1st_123000.csv: {'filename': '375_UCR_id_73_Medical_tr_47500_1st_123000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 200, 'train_index': '47500', 'data_shape': '190001x1', 'output_length': 145000, 'label_length': 142501, 'aligned_length': 142501, 'AUC-PR': 0.051331505269766514, 'AUC-ROC': 0.9782612522649863, 'VUS-PR': 0.05289180134948137, 'VUS-ROC': 0.9801326315262129, 'Standard-F1': 0.10690161527149859, 'Standard-Precision': 0.058615136876006425, 'Standard-Recall': 0.6066666666666646, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6666664502706257, 'Affiliation-P': 0.49999975655449386, 'Affiliation-R': 1.0, 'F1_T': 0.10597755155297489, 'Precision_T': 0.057490795808552805, 'Recall_T': 0.6766666666666643}
Logits metrics skipped for 375_UCR_id_73_Medical_tr_47500_1st_123000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/375_UCR_id_73_Medical_tr_47500_1st_123000_results.pkl
Processing file: 441_UCR_id_139_Facility_tr_7500_1st_13400.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 441_UCR_id_139_Facility_tr_7500_1st_13400.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 128.16it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  80%|████████  | 160/200 [00:01<00:00, 159.43it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 171.64it/s]
Table1-only metrics completed in 2.23s
Evaluation Result:  {'AUC-PR': 0.01621461431748305, 'AUC-ROC': 0.5078796945701357, 'VUS-PR': 0.020104976401610988, 'VUS-ROC': 0.5659117877619696, 'Standard-F1': 0.04761014519843085, 'Standard-Precision': 0.02441362561513389, 'Standard-Recall': 0.9549999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.674716408419773, 'Affiliation-P': 0.5092497604073076, 'Affiliation-R': 0.9994649444444446, 'F1_T': 0.04759248931023236, 'Precision_T': 0.024401092965622415, 'Recall_T': 0.9599999999999975}
Results for 441_UCR_id_139_Facility_tr_7500_1st_13400.csv: {'filename': '441_UCR_id_139_Facility_tr_7500_1st_13400.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 189, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.01621461431748305, 'AUC-ROC': 0.5078796945701357, 'VUS-PR': 0.020104976401610988, 'VUS-ROC': 0.5659117877619696, 'Standard-F1': 0.04761014519843085, 'Standard-Precision': 0.02441362561513389, 'Standard-Recall': 0.9549999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.674716408419773, 'Affiliation-P': 0.5092497604073076, 'Affiliation-R': 0.9994649444444446, 'F1_T': 0.04759248931023236, 'Precision_T': 0.024401092965622415, 'Recall_T': 0.9599999999999975}
Logits metrics skipped for 441_UCR_id_139_Facility_tr_7500_1st_13400.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/441_UCR_id_139_Facility_tr_7500_1st_13400_results.pkl
Processing file: 043_WSD_id_15_WebService_tr_4521_1st_6828.csv, Data shape: (18085, 1)
Label shape: (18085,)
Train data shape: (4521, 1), Test data shape: (13564, 1), Label test shape: (13564,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 043_WSD_id_15_WebService_tr_4521_1st_6828.csv...
Using CUDA device: 2
Data shape: (13564, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13564, window_size 5000, adding 1436 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.93it/s]
Original shapes - Output: (15000,), Label: (13564,)
Aligned shapes - Output: (13564,), Label: (13564,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 309.38it/s]
Table1-only metrics completed in 1.49s
Evaluation Result:  {'AUC-PR': 0.33645523285386003, 'AUC-ROC': 0.9395313959777745, 'VUS-PR': 0.09559729967092896, 'VUS-ROC': 0.9419932736673322, 'Standard-F1': 0.028169014084425712, 'Standard-Precision': 0.014705882352941069, 'Standard-Recall': 0.33333333333327775, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9474079667939949, 'Affiliation-P': 0.9001655355847004, 'Affiliation-R': 0.9998838060212368, 'F1_T': 0.028169014084425712, 'Precision_T': 0.014705882352941069, 'Recall_T': 0.33333333333327775}
Results for 043_WSD_id_15_WebService_tr_4521_1st_6828.csv: {'filename': '043_WSD_id_15_WebService_tr_4521_1st_6828.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4521', 'data_shape': '18085x1', 'output_length': 15000, 'label_length': 13564, 'aligned_length': 13564, 'AUC-PR': 0.33645523285386003, 'AUC-ROC': 0.9395313959777745, 'VUS-PR': 0.09559729967092896, 'VUS-ROC': 0.9419932736673322, 'Standard-F1': 0.028169014084425712, 'Standard-Precision': 0.014705882352941069, 'Standard-Recall': 0.33333333333327775, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9474079667939949, 'Affiliation-P': 0.9001655355847004, 'Affiliation-R': 0.9998838060212368, 'F1_T': 0.028169014084425712, 'Precision_T': 0.014705882352941069, 'Recall_T': 0.33333333333327775}
Logits metrics skipped for 043_WSD_id_15_WebService_tr_4521_1st_6828.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/043_WSD_id_15_WebService_tr_4521_1st_6828_results.pkl
Processing file: 690_YAHOO_id_140_WebService_tr_500_1st_696.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 690_YAHOO_id_140_WebService_tr_500_1st_696.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.70it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1132.40it/s]
Table1-only metrics completed in 0.52s
Evaluation Result:  {'AUC-PR': 0.8650793650793649, 'AUC-ROC': 0.9855072463768115, 'VUS-PR': 0.8716766678052059, 'VUS-ROC': 0.9903829735179462, 'Standard-F1': 0.6315789473678891, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.8571428571427346, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9507201490078214, 'Affiliation-P': 0.9204204008974569, 'Affiliation-R': 0.9830827067669173, 'F1_T': 0.6315789473678891, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.8571428571427346}
Results for 690_YAHOO_id_140_WebService_tr_500_1st_696.csv: {'filename': '690_YAHOO_id_140_WebService_tr_500_1st_696.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.8650793650793649, 'AUC-ROC': 0.9855072463768115, 'VUS-PR': 0.8716766678052059, 'VUS-ROC': 0.9903829735179462, 'Standard-F1': 0.6315789473678891, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.8571428571427346, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9507201490078214, 'Affiliation-P': 0.9204204008974569, 'Affiliation-R': 0.9830827067669173, 'F1_T': 0.6315789473678891, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.8571428571427346}
Logits metrics skipped for 690_YAHOO_id_140_WebService_tr_500_1st_696.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/690_YAHOO_id_140_WebService_tr_500_1st_696_results.pkl
Processing file: 364_UCR_id_62_Sensor_tr_2833_1st_5988.csv, Data shape: (11334, 1)
Label shape: (11334,)
Train data shape: (2833, 1), Test data shape: (8501, 1), Label test shape: (8501,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 364_UCR_id_62_Sensor_tr_2833_1st_5988.csv...
Using CUDA device: 2
Data shape: (8501, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 8501, window_size 5000, adding 1499 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.02it/s]
Original shapes - Output: (10000,), Label: (8501,)
Aligned shapes - Output: (8501,), Label: (8501,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 501.96it/s]
Table1-only metrics completed in 0.81s
Evaluation Result:  {'AUC-PR': 0.007375151685057283, 'AUC-ROC': 0.24697370422528298, 'VUS-PR': 0.008940963066999468, 'VUS-ROC': 0.29195298795110625, 'Standard-F1': 0.025942765445280618, 'Standard-Precision': 0.013141850697737432, 'Standard-Recall': 0.9999999999999898, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8511309102250438, 'Affiliation-P': 0.848264089313706, 'Affiliation-R': 0.8540171744500648, 'F1_T': 0.02583910495468946, 'Precision_T': 0.013088652003778166, 'Recall_T': 0.9999999999999898}
Results for 364_UCR_id_62_Sensor_tr_2833_1st_5988.csv: {'filename': '364_UCR_id_62_Sensor_tr_2833_1st_5988.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 98, 'train_index': '2833', 'data_shape': '11334x1', 'output_length': 10000, 'label_length': 8501, 'aligned_length': 8501, 'AUC-PR': 0.007375151685057283, 'AUC-ROC': 0.24697370422528298, 'VUS-PR': 0.008940963066999468, 'VUS-ROC': 0.29195298795110625, 'Standard-F1': 0.025942765445280618, 'Standard-Precision': 0.013141850697737432, 'Standard-Recall': 0.9999999999999898, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8511309102250438, 'Affiliation-P': 0.848264089313706, 'Affiliation-R': 0.8540171744500648, 'F1_T': 0.02583910495468946, 'Precision_T': 0.013088652003778166, 'Recall_T': 0.9999999999999898}
Logits metrics skipped for 364_UCR_id_62_Sensor_tr_2833_1st_5988.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/364_UCR_id_62_Sensor_tr_2833_1st_5988_results.pkl
Skipping file: 286_CATSv2_id_1_Sensor_tr_17792_1st_17892.csv due to filter criteria.
Processing file: 318_UCR_id_16_Medical_tr_2499_1st_7000.csv, Data shape: (9998, 1)
Label shape: (9998,)
Train data shape: (2499, 1), Test data shape: (7499, 1), Label test shape: (7499,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 318_UCR_id_16_Medical_tr_2499_1st_7000.csv...
Using CUDA device: 2
Data shape: (7499, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 7499, window_size 5000, adding 2501 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 164.77it/s]
Original shapes - Output: (10000,), Label: (7499,)
Aligned shapes - Output: (7499,), Label: (7499,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 482.27it/s]
Table1-only metrics completed in 0.55s
Evaluation Result:  {'AUC-PR': 0.007988075461520283, 'AUC-ROC': 0.6043442072761445, 'VUS-PR': 0.008678887287863752, 'VUS-ROC': 0.6105489253847667, 'Standard-F1': 0.02732240437155773, 'Standard-Precision': 0.01385041551246537, 'Standard-Recall': 0.99999999999998, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8203814884094996, 'Affiliation-P': 0.8168176809539517, 'Affiliation-R': 0.8239765302040272, 'F1_T': 0.027247956403242863, 'Precision_T': 0.013812154696132593, 'Recall_T': 0.99999999999998}
Results for 318_UCR_id_16_Medical_tr_2499_1st_7000.csv: {'filename': '318_UCR_id_16_Medical_tr_2499_1st_7000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 26, 'train_index': '2499', 'data_shape': '9998x1', 'output_length': 10000, 'label_length': 7499, 'aligned_length': 7499, 'AUC-PR': 0.007988075461520283, 'AUC-ROC': 0.6043442072761445, 'VUS-PR': 0.008678887287863752, 'VUS-ROC': 0.6105489253847667, 'Standard-F1': 0.02732240437155773, 'Standard-Precision': 0.01385041551246537, 'Standard-Recall': 0.99999999999998, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8203814884094996, 'Affiliation-P': 0.8168176809539517, 'Affiliation-R': 0.8239765302040272, 'F1_T': 0.027247956403242863, 'Precision_T': 0.013812154696132593, 'Recall_T': 0.99999999999998}
Logits metrics skipped for 318_UCR_id_16_Medical_tr_2499_1st_7000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/318_UCR_id_16_Medical_tr_2499_1st_7000_results.pkl
Processing file: 409_UCR_id_107_Facility_tr_20000_1st_67950.csv, Data shape: (80000, 1)
Label shape: (80000,)
Train data shape: (20000, 1), Test data shape: (60000, 1), Label test shape: (60000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 409_UCR_id_107_Facility_tr_20000_1st_67950.csv...
Using CUDA device: 2
Data shape: (60000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([12, 5000, 1])
batched_mask shape: torch.Size([12, 5000])

100%|██████████| 1/1 [00:00<00:00, 93.16it/s]
Original shapes - Output: (60000,), Label: (60000,)
Aligned shapes - Output: (60000,), Label: (60000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  32%|███▏      | 63/200 [00:01<00:02, 62.30it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [00:02<00:01, 63.14it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 70.04it/s]
Table1-only metrics completed in 5.83s
Evaluation Result:  {'AUC-PR': 0.43199961200409154, 'AUC-ROC': 0.7498457573221757, 'VUS-PR': 0.2354720560853367, 'VUS-ROC': 0.8654063649813648, 'Standard-F1': 0.30588235294076055, 'Standard-Precision': 0.21666666666666629, 'Standard-Recall': 0.5199999999999979, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6881017239581374, 'Affiliation-P': 0.5246505859202264, 'Affiliation-R': 0.9994837333333334, 'F1_T': 0.30588235294076055, 'Precision_T': 0.21666666666666629, 'Recall_T': 0.5199999999999979}
Results for 409_UCR_id_107_Facility_tr_20000_1st_67950.csv: {'filename': '409_UCR_id_107_Facility_tr_20000_1st_67950.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 276, 'train_index': '20000', 'data_shape': '80000x1', 'output_length': 60000, 'label_length': 60000, 'aligned_length': 60000, 'AUC-PR': 0.43199961200409154, 'AUC-ROC': 0.7498457573221757, 'VUS-PR': 0.2354720560853367, 'VUS-ROC': 0.8654063649813648, 'Standard-F1': 0.30588235294076055, 'Standard-Precision': 0.21666666666666629, 'Standard-Recall': 0.5199999999999979, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6881017239581374, 'Affiliation-P': 0.5246505859202264, 'Affiliation-R': 0.9994837333333334, 'F1_T': 0.30588235294076055, 'Precision_T': 0.21666666666666629, 'Recall_T': 0.5199999999999979}
Logits metrics skipped for 409_UCR_id_107_Facility_tr_20000_1st_67950.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/409_UCR_id_107_Facility_tr_20000_1st_67950_results.pkl
Processing file: 495_UCR_id_193_Environment_tr_2046_1st_6520.csv, Data shape: (8184, 1)
Label shape: (8184,)
Train data shape: (2046, 1), Test data shape: (6138, 1), Label test shape: (6138,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 495_UCR_id_193_Environment_tr_2046_1st_6520.csv...
Using CUDA device: 2
Data shape: (6138, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6138, window_size 5000, adding 3862 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.89it/s]
Original shapes - Output: (10000,), Label: (6138,)
Aligned shapes - Output: (6138,), Label: (6138,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 500.23it/s]
Table1-only metrics completed in 0.52s
Evaluation Result:  {'AUC-PR': 0.2663836778014016, 'AUC-ROC': 0.9912018863809835, 'VUS-PR': 0.31140508242130877, 'VUS-ROC': 0.9933530662836038, 'Standard-F1': 0.2909090909087445, 'Standard-Precision': 0.18604651162790484, 'Standard-Recall': 0.666666666666639, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7445307691211843, 'Affiliation-P': 0.5930298814253067, 'Affiliation-R': 1.0, 'F1_T': 0.2931034482755288, 'Precision_T': 0.18478260869565016, 'Recall_T': 0.7083333333333038}
Results for 495_UCR_id_193_Environment_tr_2046_1st_6520.csv: {'filename': '495_UCR_id_193_Environment_tr_2046_1st_6520.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '2046', 'data_shape': '8184x1', 'output_length': 10000, 'label_length': 6138, 'aligned_length': 6138, 'AUC-PR': 0.2663836778014016, 'AUC-ROC': 0.9912018863809835, 'VUS-PR': 0.31140508242130877, 'VUS-ROC': 0.9933530662836038, 'Standard-F1': 0.2909090909087445, 'Standard-Precision': 0.18604651162790484, 'Standard-Recall': 0.666666666666639, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7445307691211843, 'Affiliation-P': 0.5930298814253067, 'Affiliation-R': 1.0, 'F1_T': 0.2931034482755288, 'Precision_T': 0.18478260869565016, 'Recall_T': 0.7083333333333038}
Logits metrics skipped for 495_UCR_id_193_Environment_tr_2046_1st_6520.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/495_UCR_id_193_Environment_tr_2046_1st_6520_results.pkl
Processing file: 555_YAHOO_id_5_WebService_tr_500_1st_730.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 555_YAHOO_id_5_WebService_tr_500_1st_730.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.10it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1136.10it/s]
Table1-only metrics completed in 0.43s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.9999444925344662, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9987868417186013, 'Affiliation-P': 0.9975766233766233, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Results for 555_YAHOO_id_5_WebService_tr_500_1st_730.csv: {'filename': '555_YAHOO_id_5_WebService_tr_500_1st_730.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.9999444925344662, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9987868417186013, 'Affiliation-P': 0.9975766233766233, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 555_YAHOO_id_5_WebService_tr_500_1st_730.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/555_YAHOO_id_5_WebService_tr_500_1st_730_results.pkl
Processing file: 293_TODS_id_7_Synthetic_tr_500_1st_7.csv, Data shape: (5000, 1)
Label shape: (5000,)
Train data shape: (500, 1), Test data shape: (4500, 1), Label test shape: (4500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 293_TODS_id_7_Synthetic_tr_500_1st_7.csv...
Using CUDA device: 2
Data shape: (4500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 4500, 1])
batched_mask shape: torch.Size([1, 4500])

100%|██████████| 1/1 [00:00<00:00, 163.56it/s]
Original shapes - Output: (4500,), Label: (4500,)
Aligned shapes - Output: (4500,), Label: (4500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 304.38it/s]
Table1-only metrics completed in 4.25s
Evaluation Result:  {'AUC-PR': 0.36790740205381206, 'AUC-ROC': 0.9124051678700413, 'VUS-PR': 0.7541849176884685, 'VUS-ROC': 0.9706452664363763, 'Standard-F1': 0.4507042253517059, 'Standard-Precision': 0.3142389525368244, 'Standard-Recall': 0.7966804979253079, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8779521929441301, 'Affiliation-P': 0.8512615829782294, 'Affiliation-R': 0.9063707042882463, 'F1_T': 0.4472769409034203, 'Precision_T': 0.3102893890675236, 'Recall_T': 0.8008298755186689}
Results for 293_TODS_id_7_Synthetic_tr_500_1st_7.csv: {'filename': '293_TODS_id_7_Synthetic_tr_500_1st_7.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '500', 'data_shape': '5000x1', 'output_length': 4500, 'label_length': 4500, 'aligned_length': 4500, 'AUC-PR': 0.36790740205381206, 'AUC-ROC': 0.9124051678700413, 'VUS-PR': 0.7541849176884685, 'VUS-ROC': 0.9706452664363763, 'Standard-F1': 0.4507042253517059, 'Standard-Precision': 0.3142389525368244, 'Standard-Recall': 0.7966804979253079, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8779521929441301, 'Affiliation-P': 0.8512615829782294, 'Affiliation-R': 0.9063707042882463, 'F1_T': 0.4472769409034203, 'Precision_T': 0.3102893890675236, 'Recall_T': 0.8008298755186689}
Logits metrics skipped for 293_TODS_id_7_Synthetic_tr_500_1st_7.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/293_TODS_id_7_Synthetic_tr_500_1st_7_results.pkl
Processing file: 037_WSD_id_9_WebService_tr_4583_1st_6214.csv, Data shape: (18334, 1)
Label shape: (18334,)
Train data shape: (4583, 1), Test data shape: (13751, 1), Label test shape: (13751,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 037_WSD_id_9_WebService_tr_4583_1st_6214.csv...
Using CUDA device: 2
Data shape: (13751, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13751, window_size 5000, adding 1249 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.38it/s]
Original shapes - Output: (15000,), Label: (13751,)
Aligned shapes - Output: (13751,), Label: (13751,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 245.93it/s]
Table1-only metrics completed in 1.13s
Evaluation Result:  {'AUC-PR': 0.08940096781810476, 'AUC-ROC': 0.5692733531081959, 'VUS-PR': 0.03974267276805449, 'VUS-ROC': 0.7338470657672007, 'Standard-F1': 0.08219178082165181, 'Standard-Precision': 0.04878048780487785, 'Standard-Recall': 0.26086956521738563, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9485508427037848, 'Affiliation-P': 0.9032978245577056, 'Affiliation-R': 0.9985771212577768, 'F1_T': 0.08150470219411005, 'Precision_T': 0.04761904761904744, 'Recall_T': 0.2826086956521678}
Results for 037_WSD_id_9_WebService_tr_4583_1st_6214.csv: {'filename': '037_WSD_id_9_WebService_tr_4583_1st_6214.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '4583', 'data_shape': '18334x1', 'output_length': 15000, 'label_length': 13751, 'aligned_length': 13751, 'AUC-PR': 0.08940096781810476, 'AUC-ROC': 0.5692733531081959, 'VUS-PR': 0.03974267276805449, 'VUS-ROC': 0.7338470657672007, 'Standard-F1': 0.08219178082165181, 'Standard-Precision': 0.04878048780487785, 'Standard-Recall': 0.26086956521738563, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9485508427037848, 'Affiliation-P': 0.9032978245577056, 'Affiliation-R': 0.9985771212577768, 'F1_T': 0.08150470219411005, 'Precision_T': 0.04761904761904744, 'Recall_T': 0.2826086956521678}
Logits metrics skipped for 037_WSD_id_9_WebService_tr_4583_1st_6214.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/037_WSD_id_9_WebService_tr_4583_1st_6214_results.pkl
Processing file: 457_UCR_id_155_Medical_tr_49999_1st_143000.csv, Data shape: (199999, 1)
Label shape: (199999,)
Train data shape: (49999, 1), Test data shape: (150000, 1), Label test shape: (150000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 457_UCR_id_155_Medical_tr_49999_1st_143000.csv...
Using CUDA device: 2
Data shape: (150000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([30, 5000, 1])
batched_mask shape: torch.Size([30, 5000])

100%|██████████| 1/1 [00:00<00:00, 47.03it/s]
Original shapes - Output: (150000,), Label: (150000,)
Aligned shapes - Output: (150000,), Label: (150000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  12%|█▎        | 25/200 [00:01<00:07, 24.76it/s]

Calculating Affiliation Metrics:  25%|██▌       | 50/200 [00:02<00:06, 22.13it/s]

Calculating Affiliation Metrics:  36%|███▋      | 73/200 [00:03<00:05, 21.20it/s]

Calculating Affiliation Metrics:  48%|████▊     | 97/200 [00:04<00:04, 22.10it/s]

Calculating Affiliation Metrics:  62%|██████▎   | 125/200 [00:05<00:03, 23.93it/s]

Calculating Affiliation Metrics:  80%|███████▉  | 159/200 [00:06<00:01, 27.17it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:07<00:00, 27.58it/s]
Table1-only metrics completed in 11.79s
Evaluation Result:  {'AUC-PR': 0.033179719118246555, 'AUC-ROC': 0.971388810955244, 'VUS-PR': 0.034731267035135606, 'VUS-ROC': 0.971934403524595, 'Standard-F1': 0.08245877061458651, 'Standard-Precision': 0.04368546465448768, 'Standard-Recall': 0.7333333333333308, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6681179729041518, 'Affiliation-P': 0.5016344986357202, 'Affiliation-R': 1.0, 'F1_T': 0.08159831756034441, 'Precision_T': 0.043546576879910204, 'Recall_T': 0.6466666666666645}
Results for 457_UCR_id_155_Medical_tr_49999_1st_143000.csv: {'filename': '457_UCR_id_155_Medical_tr_49999_1st_143000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '49999', 'data_shape': '199999x1', 'output_length': 150000, 'label_length': 150000, 'aligned_length': 150000, 'AUC-PR': 0.033179719118246555, 'AUC-ROC': 0.971388810955244, 'VUS-PR': 0.034731267035135606, 'VUS-ROC': 0.971934403524595, 'Standard-F1': 0.08245877061458651, 'Standard-Precision': 0.04368546465448768, 'Standard-Recall': 0.7333333333333308, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6681179729041518, 'Affiliation-P': 0.5016344986357202, 'Affiliation-R': 1.0, 'F1_T': 0.08159831756034441, 'Precision_T': 0.043546576879910204, 'Recall_T': 0.6466666666666645}
Logits metrics skipped for 457_UCR_id_155_Medical_tr_49999_1st_143000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/457_UCR_id_155_Medical_tr_49999_1st_143000_results.pkl
Processing file: 166_Stock_id_18_Finance_tr_500_1st_18.csv, Data shape: (20000, 1)
Label shape: (20000,)
Train data shape: (500, 1), Test data shape: (19500, 1), Label test shape: (19500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 166_Stock_id_18_Finance_tr_500_1st_18.csv...
Using CUDA device: 2
Data shape: (19500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 19500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.33it/s]
Original shapes - Output: (20000,), Label: (19500,)
Aligned shapes - Output: (19500,), Label: (19500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   3%|▎         | 6/200 [00:01<00:33,  5.79it/s]

Calculating Affiliation Metrics:   6%|▌         | 12/200 [00:03<01:06,  2.81it/s]

Calculating Affiliation Metrics:   8%|▊         | 16/200 [00:06<01:22,  2.22it/s]

Calculating Affiliation Metrics:  10%|▉         | 19/200 [00:08<01:32,  1.96it/s]

Calculating Affiliation Metrics:  11%|█         | 22/200 [00:10<01:44,  1.70it/s]

Calculating Affiliation Metrics:  12%|█▏        | 24/200 [00:12<01:54,  1.54it/s]

Calculating Affiliation Metrics:  13%|█▎        | 26/200 [00:14<02:05,  1.39it/s]

Calculating Affiliation Metrics:  14%|█▍        | 28/200 [00:16<02:15,  1.27it/s]

Calculating Affiliation Metrics:  15%|█▌        | 30/200 [00:18<02:23,  1.19it/s]

Calculating Affiliation Metrics:  16%|█▌        | 32/200 [00:20<02:30,  1.12it/s]

Calculating Affiliation Metrics:  17%|█▋        | 34/200 [00:22<02:35,  1.07it/s]

Calculating Affiliation Metrics:  18%|█▊        | 36/200 [00:24<02:36,  1.05it/s]

Calculating Affiliation Metrics:  19%|█▉        | 38/200 [00:26<02:37,  1.03it/s]

Calculating Affiliation Metrics:  20%|██        | 40/200 [00:28<02:38,  1.01it/s]

Calculating Affiliation Metrics:  21%|██        | 42/200 [00:30<02:41,  1.02s/it]

Calculating Affiliation Metrics:  22%|██▏       | 43/200 [00:31<02:42,  1.04s/it]

Calculating Affiliation Metrics:  22%|██▏       | 44/200 [00:33<02:43,  1.05s/it]

Calculating Affiliation Metrics:  22%|██▎       | 45/200 [00:34<02:41,  1.04s/it]

Calculating Affiliation Metrics:  23%|██▎       | 46/200 [00:35<02:40,  1.04s/it]

Calculating Affiliation Metrics:  24%|██▎       | 47/200 [00:36<02:40,  1.05s/it]

Calculating Affiliation Metrics:  24%|██▍       | 48/200 [00:37<02:40,  1.06s/it]

Calculating Affiliation Metrics:  24%|██▍       | 49/200 [00:38<02:40,  1.06s/it]

Calculating Affiliation Metrics:  25%|██▌       | 50/200 [00:39<02:39,  1.07s/it]

Calculating Affiliation Metrics:  26%|██▌       | 51/200 [00:40<02:39,  1.07s/it]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:41<02:37,  1.07s/it]

Calculating Affiliation Metrics:  26%|██▋       | 53/200 [00:42<02:37,  1.07s/it]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:43<02:38,  1.09s/it]

Calculating Affiliation Metrics:  28%|██▊       | 55/200 [00:44<02:40,  1.10s/it]

Calculating Affiliation Metrics:  28%|██▊       | 56/200 [00:46<02:42,  1.13s/it]

Calculating Affiliation Metrics:  28%|██▊       | 57/200 [00:47<02:43,  1.15s/it]

Calculating Affiliation Metrics:  29%|██▉       | 58/200 [00:48<02:45,  1.16s/it]

Calculating Affiliation Metrics:  30%|██▉       | 59/200 [00:49<02:44,  1.17s/it]

Calculating Affiliation Metrics:  30%|███       | 60/200 [00:50<02:44,  1.17s/it]

Calculating Affiliation Metrics:  30%|███       | 61/200 [00:52<02:43,  1.18s/it]

Calculating Affiliation Metrics:  31%|███       | 62/200 [00:53<02:42,  1.18s/it]

Calculating Affiliation Metrics:  32%|███▏      | 63/200 [00:54<02:42,  1.19s/it]

Calculating Affiliation Metrics:  32%|███▏      | 64/200 [00:55<02:43,  1.20s/it]

Calculating Affiliation Metrics:  32%|███▎      | 65/200 [00:56<02:42,  1.20s/it]

Calculating Affiliation Metrics:  33%|███▎      | 66/200 [00:58<02:42,  1.21s/it]

Calculating Affiliation Metrics:  34%|███▎      | 67/200 [00:59<02:42,  1.22s/it]

Calculating Affiliation Metrics:  34%|███▍      | 68/200 [01:00<02:42,  1.23s/it]

Calculating Affiliation Metrics:  34%|███▍      | 69/200 [01:01<02:43,  1.25s/it]

Calculating Affiliation Metrics:  35%|███▌      | 70/200 [01:03<02:43,  1.26s/it]

Calculating Affiliation Metrics:  36%|███▌      | 71/200 [01:04<02:43,  1.27s/it]

Calculating Affiliation Metrics:  36%|███▌      | 72/200 [01:05<02:44,  1.29s/it]

Calculating Affiliation Metrics:  36%|███▋      | 73/200 [01:07<02:45,  1.30s/it]

Calculating Affiliation Metrics:  37%|███▋      | 74/200 [01:08<02:45,  1.31s/it]

Calculating Affiliation Metrics:  38%|███▊      | 75/200 [01:09<02:46,  1.33s/it]

Calculating Affiliation Metrics:  38%|███▊      | 76/200 [01:11<02:44,  1.33s/it]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [01:12<02:42,  1.32s/it]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [01:13<02:40,  1.32s/it]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [01:15<02:39,  1.32s/it]

Calculating Affiliation Metrics:  40%|████      | 80/200 [01:16<02:37,  1.31s/it]

Calculating Affiliation Metrics:  40%|████      | 81/200 [01:17<02:35,  1.30s/it]

Calculating Affiliation Metrics:  41%|████      | 82/200 [01:18<02:32,  1.29s/it]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [01:20<02:31,  1.29s/it]

Calculating Affiliation Metrics:  42%|████▏     | 84/200 [01:21<02:28,  1.28s/it]

Calculating Affiliation Metrics:  42%|████▎     | 85/200 [01:22<02:25,  1.27s/it]

Calculating Affiliation Metrics:  43%|████▎     | 86/200 [01:23<02:24,  1.27s/it]

Calculating Affiliation Metrics:  44%|████▎     | 87/200 [01:25<02:21,  1.26s/it]

Calculating Affiliation Metrics:  44%|████▍     | 88/200 [01:26<02:18,  1.24s/it]

Calculating Affiliation Metrics:  44%|████▍     | 89/200 [01:27<02:15,  1.22s/it]

Calculating Affiliation Metrics:  45%|████▌     | 90/200 [01:28<02:13,  1.21s/it]

Calculating Affiliation Metrics:  46%|████▌     | 91/200 [01:29<02:12,  1.21s/it]

Calculating Affiliation Metrics:  46%|████▌     | 92/200 [01:31<02:09,  1.20s/it]

Calculating Affiliation Metrics:  46%|████▋     | 93/200 [01:32<02:07,  1.19s/it]

Calculating Affiliation Metrics:  47%|████▋     | 94/200 [01:33<02:04,  1.17s/it]

Calculating Affiliation Metrics:  48%|████▊     | 95/200 [01:34<02:02,  1.17s/it]

Calculating Affiliation Metrics:  48%|████▊     | 96/200 [01:35<02:00,  1.16s/it]

Calculating Affiliation Metrics:  48%|████▊     | 97/200 [01:36<01:57,  1.14s/it]

Calculating Affiliation Metrics:  49%|████▉     | 98/200 [01:37<01:55,  1.13s/it]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [01:39<01:53,  1.12s/it]

Calculating Affiliation Metrics:  50%|█████     | 100/200 [01:40<01:52,  1.12s/it]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [01:41<01:51,  1.13s/it]

Calculating Affiliation Metrics:  51%|█████     | 102/200 [01:42<01:50,  1.12s/it]

Calculating Affiliation Metrics:  52%|█████▏    | 103/200 [01:43<01:49,  1.13s/it]

Calculating Affiliation Metrics:  52%|█████▏    | 104/200 [01:44<01:47,  1.12s/it]

Calculating Affiliation Metrics:  52%|█████▎    | 105/200 [01:45<01:46,  1.12s/it]

Calculating Affiliation Metrics:  53%|█████▎    | 106/200 [01:46<01:44,  1.12s/it]

Calculating Affiliation Metrics:  54%|█████▎    | 107/200 [01:47<01:42,  1.11s/it]

Calculating Affiliation Metrics:  54%|█████▍    | 108/200 [01:49<01:41,  1.10s/it]

Calculating Affiliation Metrics:  55%|█████▍    | 109/200 [01:50<01:40,  1.10s/it]

Calculating Affiliation Metrics:  55%|█████▌    | 110/200 [01:51<01:38,  1.10s/it]

Calculating Affiliation Metrics:  56%|█████▌    | 111/200 [01:52<01:37,  1.09s/it]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [01:53<01:36,  1.09s/it]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [01:54<01:34,  1.09s/it]

Calculating Affiliation Metrics:  57%|█████▋    | 114/200 [01:55<01:32,  1.07s/it]

Calculating Affiliation Metrics:  58%|█████▊    | 116/200 [01:57<01:26,  1.03s/it]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [01:59<01:22,  1.00s/it]

Calculating Affiliation Metrics:  60%|██████    | 120/200 [02:01<01:18,  1.02it/s]

Calculating Affiliation Metrics:  61%|██████    | 122/200 [02:03<01:14,  1.05it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 124/200 [02:04<01:10,  1.08it/s]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [02:06<01:06,  1.11it/s]

Calculating Affiliation Metrics:  64%|██████▍   | 128/200 [02:08<01:04,  1.12it/s]

Calculating Affiliation Metrics:  65%|██████▌   | 130/200 [02:10<01:02,  1.13it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 132/200 [02:11<00:59,  1.14it/s]

Calculating Affiliation Metrics:  67%|██████▋   | 134/200 [02:13<00:57,  1.15it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 136/200 [02:15<00:54,  1.17it/s]

Calculating Affiliation Metrics:  69%|██████▉   | 138/200 [02:16<00:52,  1.17it/s]

Calculating Affiliation Metrics:  70%|███████   | 140/200 [02:18<00:51,  1.17it/s]

Calculating Affiliation Metrics:  71%|███████   | 142/200 [02:20<00:48,  1.20it/s]

Calculating Affiliation Metrics:  72%|███████▏  | 144/200 [02:21<00:45,  1.23it/s]

Calculating Affiliation Metrics:  73%|███████▎  | 146/200 [02:23<00:42,  1.26it/s]

Calculating Affiliation Metrics:  74%|███████▍  | 148/200 [02:24<00:40,  1.30it/s]

Calculating Affiliation Metrics:  75%|███████▌  | 150/200 [02:25<00:37,  1.33it/s]

Calculating Affiliation Metrics:  76%|███████▌  | 152/200 [02:27<00:35,  1.34it/s]

Calculating Affiliation Metrics:  77%|███████▋  | 154/200 [02:28<00:33,  1.35it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 156/200 [02:30<00:31,  1.38it/s]

Calculating Affiliation Metrics:  79%|███████▉  | 158/200 [02:31<00:30,  1.40it/s]

Calculating Affiliation Metrics:  80%|████████  | 160/200 [02:32<00:27,  1.44it/s]

Calculating Affiliation Metrics:  81%|████████  | 162/200 [02:34<00:25,  1.49it/s]

Calculating Affiliation Metrics:  82%|████████▏ | 164/200 [02:35<00:23,  1.54it/s]

Calculating Affiliation Metrics:  83%|████████▎ | 166/200 [02:36<00:21,  1.60it/s]

Calculating Affiliation Metrics:  84%|████████▍ | 168/200 [02:37<00:19,  1.66it/s]

Calculating Affiliation Metrics:  86%|████████▌ | 171/200 [02:39<00:16,  1.79it/s]

Calculating Affiliation Metrics:  87%|████████▋ | 174/200 [02:40<00:13,  1.89it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 177/200 [02:41<00:11,  2.00it/s]

Calculating Affiliation Metrics:  90%|█████████ | 180/200 [02:43<00:09,  2.14it/s]

Calculating Affiliation Metrics:  92%|█████████▏| 183/200 [02:44<00:07,  2.32it/s]

Calculating Affiliation Metrics:  94%|█████████▎| 187/200 [02:45<00:04,  2.61it/s]

Calculating Affiliation Metrics:  96%|█████████▌| 192/200 [02:46<00:02,  3.17it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [02:47<00:00,  1.20it/s]
Table1-only metrics completed in 201.66s
Evaluation Result:  {'AUC-PR': 0.17610677844835737, 'AUC-ROC': 0.5466553539878185, 'VUS-PR': 0.8625164295623341, 'VUS-ROC': 0.9108754632017718, 'Standard-F1': 0.24844856938850857, 'Standard-Precision': 0.14617124394184167, 'Standard-Recall': 0.8273591806876369, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6865758822980783, 'Affiliation-P': 0.5227373801383902, 'Affiliation-R': 1.0, 'F1_T': 0.24837894274070518, 'Precision_T': 0.1461458872219348, 'Recall_T': 0.8266276517922455}
Results for 166_Stock_id_18_Finance_tr_500_1st_18.csv: {'filename': '166_Stock_id_18_Finance_tr_500_1st_18.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 50, 'train_index': '500', 'data_shape': '20000x1', 'output_length': 20000, 'label_length': 19500, 'aligned_length': 19500, 'AUC-PR': 0.17610677844835737, 'AUC-ROC': 0.5466553539878185, 'VUS-PR': 0.8625164295623341, 'VUS-ROC': 0.9108754632017718, 'Standard-F1': 0.24844856938850857, 'Standard-Precision': 0.14617124394184167, 'Standard-Recall': 0.8273591806876369, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6865758822980783, 'Affiliation-P': 0.5227373801383902, 'Affiliation-R': 1.0, 'F1_T': 0.24837894274070518, 'Precision_T': 0.1461458872219348, 'Recall_T': 0.8266276517922455}
Logits metrics skipped for 166_Stock_id_18_Finance_tr_500_1st_18.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/166_Stock_id_18_Finance_tr_500_1st_18_results.pkl
Skipping file: 813_Exathlon_id_4_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 516_UCR_id_214_HumanActivity_tr_2108_1st_4702.csv, Data shape: (8432, 1)
Label shape: (8432,)
Train data shape: (2108, 1), Test data shape: (6324, 1), Label test shape: (6324,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 516_UCR_id_214_HumanActivity_tr_2108_1st_4702.csv...
Using CUDA device: 2
Data shape: (6324, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6324, window_size 5000, adding 3676 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.28it/s]
Original shapes - Output: (10000,), Label: (6324,)
Aligned shapes - Output: (6324,), Label: (6324,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 463.83it/s]
Table1-only metrics completed in 0.55s
Evaluation Result:  {'AUC-PR': 0.01708480555752279, 'AUC-ROC': 0.9739832251938598, 'VUS-PR': 0.024913612464649695, 'VUS-ROC': 0.9782537624264713, 'Standard-F1': 0.04444444444438995, 'Standard-Precision': 0.02285714285714273, 'Standard-Recall': 0.79999999999984, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7794924616449264, 'Affiliation-P': 0.6386625540186995, 'Affiliation-R': 1.0, 'F1_T': 0.041450777202021644, 'Precision_T': 0.02127659574468074, 'Recall_T': 0.79999999999984}
Results for 516_UCR_id_214_HumanActivity_tr_2108_1st_4702.csv: {'filename': '516_UCR_id_214_HumanActivity_tr_2108_1st_4702.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '2108', 'data_shape': '8432x1', 'output_length': 10000, 'label_length': 6324, 'aligned_length': 6324, 'AUC-PR': 0.01708480555752279, 'AUC-ROC': 0.9739832251938598, 'VUS-PR': 0.024913612464649695, 'VUS-ROC': 0.9782537624264713, 'Standard-F1': 0.04444444444438995, 'Standard-Precision': 0.02285714285714273, 'Standard-Recall': 0.79999999999984, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7794924616449264, 'Affiliation-P': 0.6386625540186995, 'Affiliation-R': 1.0, 'F1_T': 0.041450777202021644, 'Precision_T': 0.02127659574468074, 'Recall_T': 0.79999999999984}
Logits metrics skipped for 516_UCR_id_214_HumanActivity_tr_2108_1st_4702.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/516_UCR_id_214_HumanActivity_tr_2108_1st_4702_results.pkl
Skipping file: 183_SMD_id_6_Facility_tr_5925_1st_17580.csv due to filter criteria.
Skipping file: 855_OPPORTUNITY_id_14_HumanActivity_tr_808_1st_908.csv due to filter criteria.
Processing file: 647_YAHOO_id_97_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 647_YAHOO_id_97_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 156.78it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2048.40it/s]
Table1-only metrics completed in 1.14s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 647_YAHOO_id_97_Synthetic_tr_500_1st_658.csv: {'filename': '647_YAHOO_id_97_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 235, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 647_YAHOO_id_97_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/647_YAHOO_id_97_Synthetic_tr_500_1st_658_results.pkl
Processing file: 794_YAHOO_id_244_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 794_YAHOO_id_244_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 160.30it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1819.41it/s]
Table1-only metrics completed in 0.18s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 794_YAHOO_id_244_Synthetic_tr_500_1st_658.csv: {'filename': '794_YAHOO_id_244_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 10, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 794_YAHOO_id_244_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/794_YAHOO_id_244_Synthetic_tr_500_1st_658_results.pkl
Skipping file: 147_MSL_id_8_Sensor_tr_590_1st_690.csv due to filter criteria.
Processing file: 355_UCR_id_53_HumanActivity_tr_48750_1st_143411.csv, Data shape: (195000, 1)
Label shape: (195000,)
Train data shape: (48750, 1), Test data shape: (146250, 1), Label test shape: (146250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 355_UCR_id_53_HumanActivity_tr_48750_1st_143411.csv...
Using CUDA device: 2
Data shape: (146250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 146250, window_size 5000, adding 3750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([30, 5000, 1])
batched_mask shape: torch.Size([30, 5000])

100%|██████████| 1/1 [00:00<00:00, 47.06it/s]
Original shapes - Output: (150000,), Label: (146250,)
Aligned shapes - Output: (146250,), Label: (146250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  13%|█▎        | 26/200 [00:01<00:06, 25.96it/s]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:02<00:06, 23.10it/s]

Calculating Affiliation Metrics:  38%|███▊      | 76/200 [00:03<00:05, 22.60it/s]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [00:04<00:04, 22.67it/s]

Calculating Affiliation Metrics:  62%|██████▎   | 125/200 [00:05<00:03, 23.73it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 156/200 [00:06<00:01, 26.07it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:07<00:00, 27.44it/s]
Table1-only metrics completed in 12.72s
Evaluation Result:  {'AUC-PR': 0.8773656031479573, 'AUC-ROC': 0.9979126240164214, 'VUS-PR': 0.15899957222353456, 'VUS-ROC': 0.9967060526888867, 'Standard-F1': 0.11900191938567664, 'Standard-Precision': 0.06356801093643195, 'Standard-Recall': 0.9299999999999907, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8179059826296531, 'Affiliation-P': 0.6919142740857026, 'Affiliation-R': 0.9999969230769231, 'F1_T': 0.11900191938567664, 'Precision_T': 0.06356801093643195, 'Recall_T': 0.9299999999999907}
Results for 355_UCR_id_53_HumanActivity_tr_48750_1st_143411.csv: {'filename': '355_UCR_id_53_HumanActivity_tr_48750_1st_143411.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '48750', 'data_shape': '195000x1', 'output_length': 150000, 'label_length': 146250, 'aligned_length': 146250, 'AUC-PR': 0.8773656031479573, 'AUC-ROC': 0.9979126240164214, 'VUS-PR': 0.15899957222353456, 'VUS-ROC': 0.9967060526888867, 'Standard-F1': 0.11900191938567664, 'Standard-Precision': 0.06356801093643195, 'Standard-Recall': 0.9299999999999907, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8179059826296531, 'Affiliation-P': 0.6919142740857026, 'Affiliation-R': 0.9999969230769231, 'F1_T': 0.11900191938567664, 'Precision_T': 0.06356801093643195, 'Recall_T': 0.9299999999999907}
Logits metrics skipped for 355_UCR_id_53_HumanActivity_tr_48750_1st_143411.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/355_UCR_id_53_HumanActivity_tr_48750_1st_143411_results.pkl
Skipping file: 540_SMAP_id_10_Sensor_tr_2051_1st_4575.csv due to filter criteria.
Skipping file: 531_SMAP_id_1_Sensor_tr_1811_1st_4510.csv due to filter criteria.
Processing file: 472_UCR_id_170_Facility_tr_32000_1st_67995.csv, Data shape: (128001, 1)
Label shape: (128001,)
Train data shape: (32000, 1), Test data shape: (96001, 1), Label test shape: (96001,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 472_UCR_id_170_Facility_tr_32000_1st_67995.csv...
Using CUDA device: 2
Data shape: (96001, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 96001, window_size 5000, adding 3999 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([20, 5000, 1])
batched_mask shape: torch.Size([20, 5000])

100%|██████████| 1/1 [00:00<00:00, 64.81it/s]
Original shapes - Output: (100000,), Label: (96001,)
Aligned shapes - Output: (96001,), Label: (96001,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  20%|█▉        | 39/200 [00:01<00:04, 38.71it/s]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [00:02<00:03, 37.98it/s]

Calculating Affiliation Metrics:  61%|██████    | 122/200 [00:03<00:01, 40.43it/s]

Calculating Affiliation Metrics:  92%|█████████▎| 185/200 [00:04<00:00, 49.11it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:04<00:00, 47.31it/s]
Table1-only metrics completed in 10.17s
Evaluation Result:  {'AUC-PR': 9.537434430138292e-05, 'AUC-ROC': 0.8907916666666666, 'VUS-PR': 0.6496317640115135, 'VUS-ROC': 0.9855309896785316, 'Standard-F1': 0.00018908953389410984, 'Standard-Precision': 9.4553706505295e-05, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7558357982853082, 'Affiliation-P': 0.6075087037029009, 'Affiliation-R': 0.9999895834418392, 'F1_T': 0.00018365472910909087, 'Precision_T': 9.183579759390209e-05, 'Recall_T': 0.9999999999989999}
Results for 472_UCR_id_170_Facility_tr_32000_1st_67995.csv: {'filename': '472_UCR_id_170_Facility_tr_32000_1st_67995.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 243, 'train_index': '32000', 'data_shape': '128001x1', 'output_length': 100000, 'label_length': 96001, 'aligned_length': 96001, 'AUC-PR': 9.537434430138292e-05, 'AUC-ROC': 0.8907916666666666, 'VUS-PR': 0.6496317640115135, 'VUS-ROC': 0.9855309896785316, 'Standard-F1': 0.00018908953389410984, 'Standard-Precision': 9.4553706505295e-05, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7558357982853082, 'Affiliation-P': 0.6075087037029009, 'Affiliation-R': 0.9999895834418392, 'F1_T': 0.00018365472910909087, 'Precision_T': 9.183579759390209e-05, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 472_UCR_id_170_Facility_tr_32000_1st_67995.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/472_UCR_id_170_Facility_tr_32000_1st_67995_results.pkl
Processing file: 800_YAHOO_id_250_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 800_YAHOO_id_250_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 155.37it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1817.31it/s]
Table1-only metrics completed in 0.90s
Evaluation Result:  {'AUC-PR': 0.9253086419753087, 'AUC-ROC': 0.9991471734892787, 'VUS-PR': 0.9945851708564544, 'VUS-ROC': 0.9999501361240065, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 800_YAHOO_id_250_Synthetic_tr_500_1st_623.csv: {'filename': '800_YAHOO_id_250_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 133, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9253086419753087, 'AUC-ROC': 0.9991471734892787, 'VUS-PR': 0.9945851708564544, 'VUS-ROC': 0.9999501361240065, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 800_YAHOO_id_250_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/800_YAHOO_id_250_Synthetic_tr_500_1st_623_results.pkl
Processing file: 296_TODS_id_10_Synthetic_tr_500_1st_26.csv, Data shape: (5000, 1)
Label shape: (5000,)
Train data shape: (500, 1), Test data shape: (4500, 1), Label test shape: (4500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 296_TODS_id_10_Synthetic_tr_500_1st_26.csv...
Using CUDA device: 2
Data shape: (4500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 4500, 1])
batched_mask shape: torch.Size([1, 4500])

100%|██████████| 1/1 [00:00<00:00, 159.29it/s]
Original shapes - Output: (4500,), Label: (4500,)
Aligned shapes - Output: (4500,), Label: (4500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  18%|█▊        | 36/200 [00:01<00:04, 35.54it/s]

Calculating Affiliation Metrics:  36%|███▌      | 72/200 [00:02<00:05, 22.87it/s]

Calculating Affiliation Metrics:  49%|████▉     | 98/200 [00:04<00:05, 19.46it/s]

Calculating Affiliation Metrics:  60%|█████▉    | 119/200 [00:06<00:04, 17.76it/s]

Calculating Affiliation Metrics:  69%|██████▉   | 138/200 [00:07<00:03, 16.78it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 156/200 [00:08<00:02, 16.92it/s]

Calculating Affiliation Metrics:  92%|█████████▏| 184/200 [00:09<00:00, 19.82it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:09<00:00, 20.79it/s]
Table1-only metrics completed in 16.46s
Evaluation Result:  {'AUC-PR': 0.5932668893294814, 'AUC-ROC': 0.8339997797499021, 'VUS-PR': 0.9444317289474355, 'VUS-ROC': 0.9791488625306022, 'Standard-F1': 0.6220735785948386, 'Standard-Precision': 0.8017241379310276, 'Standard-Recall': 0.5081967213114726, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7669326813584886, 'Affiliation-P': 0.7546414125055448, 'Affiliation-R': 0.7796309691668606, 'F1_T': 0.6305084745757958, 'Precision_T': 0.8303571428571355, 'Recall_T': 0.5081967213114726}
Results for 296_TODS_id_10_Synthetic_tr_500_1st_26.csv: {'filename': '296_TODS_id_10_Synthetic_tr_500_1st_26.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '500', 'data_shape': '5000x1', 'output_length': 4500, 'label_length': 4500, 'aligned_length': 4500, 'AUC-PR': 0.5932668893294814, 'AUC-ROC': 0.8339997797499021, 'VUS-PR': 0.9444317289474355, 'VUS-ROC': 0.9791488625306022, 'Standard-F1': 0.6220735785948386, 'Standard-Precision': 0.8017241379310276, 'Standard-Recall': 0.5081967213114726, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7669326813584886, 'Affiliation-P': 0.7546414125055448, 'Affiliation-R': 0.7796309691668606, 'F1_T': 0.6305084745757958, 'Precision_T': 0.8303571428571355, 'Recall_T': 0.5081967213114726}
Logits metrics skipped for 296_TODS_id_10_Synthetic_tr_500_1st_26.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/296_TODS_id_10_Synthetic_tr_500_1st_26_results.pkl
Processing file: 696_YAHOO_id_146_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 696_YAHOO_id_146_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 154.81it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2367.54it/s]
Table1-only metrics completed in 0.86s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9027877055039308, 'Affiliation-P': 0.822801302931596, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 696_YAHOO_id_146_Synthetic_tr_500_1st_893.csv: {'filename': '696_YAHOO_id_146_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 237, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9027877055039308, 'Affiliation-P': 0.822801302931596, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 696_YAHOO_id_146_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/696_YAHOO_id_146_Synthetic_tr_500_1st_893_results.pkl
Processing file: 035_WSD_id_7_WebService_tr_4555_1st_11911.csv, Data shape: (18220, 1)
Label shape: (18220,)
Train data shape: (4555, 1), Test data shape: (13665, 1), Label test shape: (13665,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 035_WSD_id_7_WebService_tr_4555_1st_11911.csv...
Using CUDA device: 2
Data shape: (13665, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13665, window_size 5000, adding 1335 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 157.89it/s]
Original shapes - Output: (15000,), Label: (13665,)
Aligned shapes - Output: (13665,), Label: (13665,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 275.37it/s]
Table1-only metrics completed in 1.05s
Evaluation Result:  {'AUC-PR': 0.2190436489712924, 'AUC-ROC': 0.7098286327294066, 'VUS-PR': 0.0421564347077929, 'VUS-ROC': 0.7896260536877985, 'Standard-F1': 0.08284023668608258, 'Standard-Precision': 0.05109489051094854, 'Standard-Recall': 0.21874999999999314, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8149091246183694, 'Affiliation-P': 0.688378907494826, 'Affiliation-R': 0.998429403727291, 'F1_T': 0.08284023668608258, 'Precision_T': 0.05109489051094854, 'Recall_T': 0.21874999999999314}
Results for 035_WSD_id_7_WebService_tr_4555_1st_11911.csv: {'filename': '035_WSD_id_7_WebService_tr_4555_1st_11911.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '4555', 'data_shape': '18220x1', 'output_length': 15000, 'label_length': 13665, 'aligned_length': 13665, 'AUC-PR': 0.2190436489712924, 'AUC-ROC': 0.7098286327294066, 'VUS-PR': 0.0421564347077929, 'VUS-ROC': 0.7896260536877985, 'Standard-F1': 0.08284023668608258, 'Standard-Precision': 0.05109489051094854, 'Standard-Recall': 0.21874999999999314, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8149091246183694, 'Affiliation-P': 0.688378907494826, 'Affiliation-R': 0.998429403727291, 'F1_T': 0.08284023668608258, 'Precision_T': 0.05109489051094854, 'Recall_T': 0.21874999999999314}
Logits metrics skipped for 035_WSD_id_7_WebService_tr_4555_1st_11911.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/035_WSD_id_7_WebService_tr_4555_1st_11911_results.pkl
Processing file: 631_YAHOO_id_81_WebService_tr_500_1st_96.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 631_YAHOO_id_81_WebService_tr_500_1st_96.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.21it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1054.46it/s]
Table1-only metrics completed in 0.49s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5882352941171626, 'Standard-Precision': 0.41666666666663194, 'Standard-Recall': 0.9999999999997999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9973262032085556, 'Affiliation-P': 0.9946666666666667, 'Affiliation-R': 1.0, 'F1_T': 0.5882352941171626, 'Precision_T': 0.41666666666663194, 'Recall_T': 0.9999999999997999}
Results for 631_YAHOO_id_81_WebService_tr_500_1st_96.csv: {'filename': '631_YAHOO_id_81_WebService_tr_500_1st_96.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5882352941171626, 'Standard-Precision': 0.41666666666663194, 'Standard-Recall': 0.9999999999997999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9973262032085556, 'Affiliation-P': 0.9946666666666667, 'Affiliation-R': 1.0, 'F1_T': 0.5882352941171626, 'Precision_T': 0.41666666666663194, 'Recall_T': 0.9999999999997999}
Logits metrics skipped for 631_YAHOO_id_81_WebService_tr_500_1st_96.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/631_YAHOO_id_81_WebService_tr_500_1st_96_results.pkl
Processing file: 626_YAHOO_id_76_WebService_tr_500_1st_883.csv, Data shape: (1420, 1)
Label shape: (1420,)
Train data shape: (500, 1), Test data shape: (920, 1), Label test shape: (920,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 626_YAHOO_id_76_WebService_tr_500_1st_883.csv...
Using CUDA device: 2
Data shape: (920, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 920, 1])
batched_mask shape: torch.Size([1, 920])

100%|██████████| 1/1 [00:00<00:00, 161.90it/s]
Original shapes - Output: (920,), Label: (920,)
Aligned shapes - Output: (920,), Label: (920,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2042.38it/s]
Table1-only metrics completed in 0.86s
Evaluation Result:  {'AUC-PR': 0.9622222222222222, 'AUC-ROC': 0.9995604395604396, 'VUS-PR': 0.9778095886913855, 'VUS-ROC': 0.9998516582385067, 'Standard-F1': 0.8999999999994099, 'Standard-Precision': 0.89999999999991, 'Standard-Recall': 0.89999999999991, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9993759341145335, 'Affiliation-P': 0.9988909683232827, 'Affiliation-R': 0.9998613710404104, 'F1_T': 0.8999999999994099, 'Precision_T': 0.89999999999991, 'Recall_T': 0.89999999999991}
Results for 626_YAHOO_id_76_WebService_tr_500_1st_883.csv: {'filename': '626_YAHOO_id_76_WebService_tr_500_1st_883.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '1420x1', 'output_length': 920, 'label_length': 920, 'aligned_length': 920, 'AUC-PR': 0.9622222222222222, 'AUC-ROC': 0.9995604395604396, 'VUS-PR': 0.9778095886913855, 'VUS-ROC': 0.9998516582385067, 'Standard-F1': 0.8999999999994099, 'Standard-Precision': 0.89999999999991, 'Standard-Recall': 0.89999999999991, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9993759341145335, 'Affiliation-P': 0.9988909683232827, 'Affiliation-R': 0.9998613710404104, 'F1_T': 0.8999999999994099, 'Precision_T': 0.89999999999991, 'Recall_T': 0.89999999999991}
Logits metrics skipped for 626_YAHOO_id_76_WebService_tr_500_1st_883.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/626_YAHOO_id_76_WebService_tr_500_1st_883_results.pkl
Processing file: 703_YAHOO_id_153_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 703_YAHOO_id_153_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 164.49it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1726.14it/s]
Table1-only metrics completed in 0.37s
Evaluation Result:  {'AUC-PR': 0.8927128427128426, 'AUC-ROC': 0.9987816764132554, 'VUS-PR': 0.9820605815874914, 'VUS-ROC': 0.9998340239127798, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Results for 703_YAHOO_id_153_Synthetic_tr_500_1st_623.csv: {'filename': '703_YAHOO_id_153_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 35, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.8927128427128426, 'AUC-ROC': 0.9987816764132554, 'VUS-PR': 0.9820605815874914, 'VUS-ROC': 0.9998340239127798, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 703_YAHOO_id_153_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/703_YAHOO_id_153_Synthetic_tr_500_1st_623_results.pkl
Processing file: 365_UCR_id_63_Facility_tr_50000_1st_210720.csv, Data shape: (299940, 1)
Label shape: (299940,)
Train data shape: (50000, 1), Test data shape: (249940, 1), Label test shape: (249940,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 365_UCR_id_63_Facility_tr_50000_1st_210720.csv...
Using CUDA device: 2
Data shape: (249940, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 249940, window_size 5000, adding 60 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([50, 5000, 1])
batched_mask shape: torch.Size([50, 5000])

100%|██████████| 1/1 [00:00<00:00, 28.29it/s]
Original shapes - Output: (250000,), Label: (249940,)
Aligned shapes - Output: (249940,), Label: (249940,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   9%|▉         | 18/200 [00:01<00:10, 17.91it/s]

Calculating Affiliation Metrics:  18%|█▊        | 36/200 [00:02<00:10, 15.71it/s]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:03<00:10, 14.54it/s]

Calculating Affiliation Metrics:  34%|███▎      | 67/200 [00:04<00:09, 13.80it/s]

Calculating Affiliation Metrics:  40%|████      | 81/200 [00:05<00:08, 13.49it/s]

Calculating Affiliation Metrics:  48%|████▊     | 95/200 [00:06<00:07, 13.35it/s]

Calculating Affiliation Metrics:  55%|█████▌    | 110/200 [00:07<00:06, 13.85it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [00:08<00:05, 14.55it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [00:09<00:03, 15.57it/s]

Calculating Affiliation Metrics:  83%|████████▎ | 166/200 [00:10<00:01, 17.04it/s]

Calculating Affiliation Metrics:  98%|█████████▊| 195/200 [00:11<00:00, 20.48it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:12<00:00, 16.63it/s]
Table1-only metrics completed in 20.42s
Evaluation Result:  {'AUC-PR': 0.7258267559824909, 'AUC-ROC': 0.9930522622710035, 'VUS-PR': 0.43084813079285317, 'VUS-ROC': 0.9930226325589053, 'Standard-F1': 0.3568253968250691, 'Standard-Precision': 0.22479999999999992, 'Standard-Recall': 0.8646153846153832, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.772527517065782, 'Affiliation-P': 0.629369951654639, 'Affiliation-R': 0.9999860551147659, 'F1_T': 0.3568253968250691, 'Precision_T': 0.22479999999999992, 'Recall_T': 0.8646153846153832}
Results for 365_UCR_id_63_Facility_tr_50000_1st_210720.csv: {'filename': '365_UCR_id_63_Facility_tr_50000_1st_210720.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 124, 'train_index': '50000', 'data_shape': '299940x1', 'output_length': 250000, 'label_length': 249940, 'aligned_length': 249940, 'AUC-PR': 0.7258267559824909, 'AUC-ROC': 0.9930522622710035, 'VUS-PR': 0.43084813079285317, 'VUS-ROC': 0.9930226325589053, 'Standard-F1': 0.3568253968250691, 'Standard-Precision': 0.22479999999999992, 'Standard-Recall': 0.8646153846153832, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.772527517065782, 'Affiliation-P': 0.629369951654639, 'Affiliation-R': 0.9999860551147659, 'F1_T': 0.3568253968250691, 'Precision_T': 0.22479999999999992, 'Recall_T': 0.8646153846153832}
Logits metrics skipped for 365_UCR_id_63_Facility_tr_50000_1st_210720.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/365_UCR_id_63_Facility_tr_50000_1st_210720_results.pkl
Processing file: 773_YAHOO_id_223_WebService_tr_500_1st_137.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 773_YAHOO_id_223_WebService_tr_500_1st_137.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 159.77it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 851.50it/s]
Table1-only metrics completed in 0.85s
Evaluation Result:  {'AUC-PR': 0.8690711462450594, 'AUC-ROC': 0.9930010109650829, 'VUS-PR': 0.887983274992354, 'VUS-ROC': 0.9946972702362876, 'Standard-F1': 0.7826086956516067, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.8181818181817438, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9674242185550916, 'Affiliation-P': 0.9369038437075801, 'Affiliation-R': 1.0, 'F1_T': 0.7826086956516067, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.8181818181817438}
Results for 773_YAHOO_id_223_WebService_tr_500_1st_137.csv: {'filename': '773_YAHOO_id_223_WebService_tr_500_1st_137.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.8690711462450594, 'AUC-ROC': 0.9930010109650829, 'VUS-PR': 0.887983274992354, 'VUS-ROC': 0.9946972702362876, 'Standard-F1': 0.7826086956516067, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.8181818181817438, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9674242185550916, 'Affiliation-P': 0.9369038437075801, 'Affiliation-R': 1.0, 'F1_T': 0.7826086956516067, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.8181818181817438}
Logits metrics skipped for 773_YAHOO_id_223_WebService_tr_500_1st_137.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/773_YAHOO_id_223_WebService_tr_500_1st_137_results.pkl
Processing file: 421_UCR_id_119_HumanActivity_tr_48750_1st_130700.csv, Data shape: (195000, 1)
Label shape: (195000,)
Train data shape: (48750, 1), Test data shape: (146250, 1), Label test shape: (146250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 421_UCR_id_119_HumanActivity_tr_48750_1st_130700.csv...
Using CUDA device: 2
Data shape: (146250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 146250, window_size 5000, adding 3750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([30, 5000, 1])
batched_mask shape: torch.Size([30, 5000])

100%|██████████| 1/1 [00:00<00:00, 46.88it/s]
Original shapes - Output: (150000,), Label: (146250,)
Aligned shapes - Output: (146250,), Label: (146250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  14%|█▎        | 27/200 [00:01<00:06, 26.17it/s]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:02<00:06, 23.56it/s]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [00:03<00:05, 23.15it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 103/200 [00:04<00:04, 23.65it/s]

Calculating Affiliation Metrics:  65%|██████▌   | 130/200 [00:05<00:02, 24.54it/s]

Calculating Affiliation Metrics:  81%|████████  | 162/200 [00:06<00:01, 26.80it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:07<00:00, 27.77it/s]
Table1-only metrics completed in 12.58s
Evaluation Result:  {'AUC-PR': 0.036878389014562425, 'AUC-ROC': 0.8848311744301114, 'VUS-PR': 0.03805772937765896, 'VUS-ROC': 0.8882921953757306, 'Standard-F1': 0.08703790360298554, 'Standard-Precision': 0.04892162019989478, 'Standard-Recall': 0.39406779661016916, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7224670283205034, 'Affiliation-P': 0.5669526786441073, 'Affiliation-R': 0.995543166739099, 'F1_T': 0.08745858968271684, 'Precision_T': 0.049227490676611606, 'Recall_T': 0.39152542372881327}
Results for 421_UCR_id_119_HumanActivity_tr_48750_1st_130700.csv: {'filename': '421_UCR_id_119_HumanActivity_tr_48750_1st_130700.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '48750', 'data_shape': '195000x1', 'output_length': 150000, 'label_length': 146250, 'aligned_length': 146250, 'AUC-PR': 0.036878389014562425, 'AUC-ROC': 0.8848311744301114, 'VUS-PR': 0.03805772937765896, 'VUS-ROC': 0.8882921953757306, 'Standard-F1': 0.08703790360298554, 'Standard-Precision': 0.04892162019989478, 'Standard-Recall': 0.39406779661016916, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7224670283205034, 'Affiliation-P': 0.5669526786441073, 'Affiliation-R': 0.995543166739099, 'F1_T': 0.08745858968271684, 'Precision_T': 0.049227490676611606, 'Recall_T': 0.39152542372881327}
Logits metrics skipped for 421_UCR_id_119_HumanActivity_tr_48750_1st_130700.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/421_UCR_id_119_HumanActivity_tr_48750_1st_130700_results.pkl
Processing file: 591_YAHOO_id_41_WebService_tr_500_1st_139.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 591_YAHOO_id_41_WebService_tr_500_1st_139.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 159.61it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 970.51it/s]
Table1-only metrics completed in 0.68s
Evaluation Result:  {'AUC-PR': 1.0000000000000002, 'AUC-ROC': 1.0, 'VUS-PR': 0.9904331281835925, 'VUS-ROC': 0.9999591444188849, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984970229419784, 'Affiliation-P': 0.9969985569839519, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Results for 591_YAHOO_id_41_WebService_tr_500_1st_139.csv: {'filename': '591_YAHOO_id_41_WebService_tr_500_1st_139.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0000000000000002, 'AUC-ROC': 1.0, 'VUS-PR': 0.9904331281835925, 'VUS-ROC': 0.9999591444188849, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984970229419784, 'Affiliation-P': 0.9969985569839519, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 591_YAHOO_id_41_WebService_tr_500_1st_139.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/591_YAHOO_id_41_WebService_tr_500_1st_139_results.pkl
Processing file: 411_UCR_id_109_Environment_tr_2046_1st_4852.csv, Data shape: (8184, 1)
Label shape: (8184,)
Train data shape: (2046, 1), Test data shape: (6138, 1), Label test shape: (6138,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 411_UCR_id_109_Environment_tr_2046_1st_4852.csv...
Using CUDA device: 2
Data shape: (6138, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6138, window_size 5000, adding 3862 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 164.61it/s]
Original shapes - Output: (10000,), Label: (6138,)
Aligned shapes - Output: (6138,), Label: (6138,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 485.78it/s]
Table1-only metrics completed in 0.53s
Evaluation Result:  {'AUC-PR': 0.95280249683099, 'AUC-ROC': 0.999661330049261, 'VUS-PR': 0.9127232996646849, 'VUS-ROC': 0.9995669630616212, 'Standard-F1': 0.8727272727267649, 'Standard-Precision': 0.7741935483870843, 'Standard-Recall': 0.9999999999999791, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9874245843330367, 'Affiliation-P': 0.9756882035758206, 'Affiliation-R': 0.9994467524709459, 'F1_T': 0.8727272727267649, 'Precision_T': 0.7741935483870843, 'Recall_T': 0.9999999999999791}
Results for 411_UCR_id_109_Environment_tr_2046_1st_4852.csv: {'filename': '411_UCR_id_109_Environment_tr_2046_1st_4852.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '2046', 'data_shape': '8184x1', 'output_length': 10000, 'label_length': 6138, 'aligned_length': 6138, 'AUC-PR': 0.95280249683099, 'AUC-ROC': 0.999661330049261, 'VUS-PR': 0.9127232996646849, 'VUS-ROC': 0.9995669630616212, 'Standard-F1': 0.8727272727267649, 'Standard-Precision': 0.7741935483870843, 'Standard-Recall': 0.9999999999999791, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9874245843330367, 'Affiliation-P': 0.9756882035758206, 'Affiliation-R': 0.9994467524709459, 'F1_T': 0.8727272727267649, 'Precision_T': 0.7741935483870843, 'Recall_T': 0.9999999999999791}
Logits metrics skipped for 411_UCR_id_109_Environment_tr_2046_1st_4852.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/411_UCR_id_109_Environment_tr_2046_1st_4852_results.pkl
Processing file: 506_UCR_id_204_Environment_tr_2046_1st_5549.csv, Data shape: (8184, 1)
Label shape: (8184,)
Train data shape: (2046, 1), Test data shape: (6138, 1), Label test shape: (6138,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 506_UCR_id_204_Environment_tr_2046_1st_5549.csv...
Using CUDA device: 2
Data shape: (6138, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6138, window_size 5000, adding 3862 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 166.39it/s]
Original shapes - Output: (10000,), Label: (6138,)
Aligned shapes - Output: (6138,), Label: (6138,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 466.19it/s]
Table1-only metrics completed in 0.55s
Evaluation Result:  {'AUC-PR': 0.9422889166018553, 'AUC-ROC': 0.9995142309797482, 'VUS-PR': 0.8958372581929637, 'VUS-ROC': 0.9993747854302057, 'Standard-F1': 0.8363636363631293, 'Standard-Precision': 0.7419354838709558, 'Standard-Recall': 0.9583333333333134, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9994850894598114, 'Affiliation-P': 0.9997424820525757, 'Affiliation-R': 0.9992278293689584, 'F1_T': 0.8363636363631293, 'Precision_T': 0.7419354838709558, 'Recall_T': 0.9583333333333134}
Results for 506_UCR_id_204_Environment_tr_2046_1st_5549.csv: {'filename': '506_UCR_id_204_Environment_tr_2046_1st_5549.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '2046', 'data_shape': '8184x1', 'output_length': 10000, 'label_length': 6138, 'aligned_length': 6138, 'AUC-PR': 0.9422889166018553, 'AUC-ROC': 0.9995142309797482, 'VUS-PR': 0.8958372581929637, 'VUS-ROC': 0.9993747854302057, 'Standard-F1': 0.8363636363631293, 'Standard-Precision': 0.7419354838709558, 'Standard-Recall': 0.9583333333333134, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9994850894598114, 'Affiliation-P': 0.9997424820525757, 'Affiliation-R': 0.9992278293689584, 'F1_T': 0.8363636363631293, 'Precision_T': 0.7419354838709558, 'Recall_T': 0.9583333333333134}
Logits metrics skipped for 506_UCR_id_204_Environment_tr_2046_1st_5549.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/506_UCR_id_204_Environment_tr_2046_1st_5549_results.pkl
Skipping file: 821_Exathlon_id_12_Facility_tr_6985_1st_7085.csv due to filter criteria.
Processing file: 775_YAHOO_id_225_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 775_YAHOO_id_225_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 163.30it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1908.30it/s]
Table1-only metrics completed in 1.14s
Evaluation Result:  {'AUC-PR': 0.8541666666666666, 'AUC-ROC': 0.9991821155943293, 'VUS-PR': 0.9946068281632673, 'VUS-ROC': 0.999982029434357, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9983183734580836, 'Affiliation-P': 0.9970279889508809, 'Affiliation-R': 0.9996121024049651, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 775_YAHOO_id_225_Synthetic_tr_500_1st_658.csv: {'filename': '775_YAHOO_id_225_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 235, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.8541666666666666, 'AUC-ROC': 0.9991821155943293, 'VUS-PR': 0.9946068281632673, 'VUS-ROC': 0.999982029434357, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9983183734580836, 'Affiliation-P': 0.9970279889508809, 'Affiliation-R': 0.9996121024049651, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 775_YAHOO_id_225_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/775_YAHOO_id_225_Synthetic_tr_500_1st_658_results.pkl
Processing file: 068_WSD_id_40_WebService_tr_4549_1st_13322.csv, Data shape: (18197, 1)
Label shape: (18197,)
Train data shape: (4549, 1), Test data shape: (13648, 1), Label test shape: (13648,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 068_WSD_id_40_WebService_tr_4549_1st_13322.csv...
Using CUDA device: 2
Data shape: (13648, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13648, window_size 5000, adding 1352 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 164.99it/s]
Original shapes - Output: (15000,), Label: (13648,)
Aligned shapes - Output: (13648,), Label: (13648,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 293.64it/s]
Table1-only metrics completed in 0.93s
Evaluation Result:  {'AUC-PR': 0.14638518347738003, 'AUC-ROC': 0.8152251798324349, 'VUS-PR': 0.033052483116540105, 'VUS-ROC': 0.9020849906542674, 'Standard-F1': 0.04819277108404841, 'Standard-Precision': 0.02919708029197059, 'Standard-Recall': 0.13793103448275387, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8164750203749969, 'Affiliation-P': 0.8157209510121679, 'Affiliation-R': 0.8172304851825173, 'F1_T': 0.04819277108404841, 'Precision_T': 0.02919708029197059, 'Recall_T': 0.13793103448275387}
Results for 068_WSD_id_40_WebService_tr_4549_1st_13322.csv: {'filename': '068_WSD_id_40_WebService_tr_4549_1st_13322.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '4549', 'data_shape': '18197x1', 'output_length': 15000, 'label_length': 13648, 'aligned_length': 13648, 'AUC-PR': 0.14638518347738003, 'AUC-ROC': 0.8152251798324349, 'VUS-PR': 0.033052483116540105, 'VUS-ROC': 0.9020849906542674, 'Standard-F1': 0.04819277108404841, 'Standard-Precision': 0.02919708029197059, 'Standard-Recall': 0.13793103448275387, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8164750203749969, 'Affiliation-P': 0.8157209510121679, 'Affiliation-R': 0.8172304851825173, 'F1_T': 0.04819277108404841, 'Precision_T': 0.02919708029197059, 'Recall_T': 0.13793103448275387}
Logits metrics skipped for 068_WSD_id_40_WebService_tr_4549_1st_13322.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/068_WSD_id_40_WebService_tr_4549_1st_13322_results.pkl
Processing file: 579_YAHOO_id_29_WebService_tr_500_1st_1342.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 579_YAHOO_id_29_WebService_tr_500_1st_1342.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 165.24it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1165.66it/s]
Table1-only metrics completed in 0.27s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.99816046412905, 'Affiliation-P': 0.9963276836158191, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Results for 579_YAHOO_id_29_WebService_tr_500_1st_1342.csv: {'filename': '579_YAHOO_id_29_WebService_tr_500_1st_1342.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.99816046412905, 'Affiliation-P': 0.9963276836158191, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 579_YAHOO_id_29_WebService_tr_500_1st_1342.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/579_YAHOO_id_29_WebService_tr_500_1st_1342_results.pkl
Processing file: 433_UCR_id_131_Facility_tr_50000_1st_106720.csv, Data shape: (299990, 1)
Label shape: (299990,)
Train data shape: (50000, 1), Test data shape: (249990, 1), Label test shape: (249990,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 433_UCR_id_131_Facility_tr_50000_1st_106720.csv...
Using CUDA device: 2
Data shape: (249990, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 249990, window_size 5000, adding 10 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([50, 5000, 1])
batched_mask shape: torch.Size([50, 5000])

100%|██████████| 1/1 [00:00<00:00, 28.40it/s]
Original shapes - Output: (250000,), Label: (249990,)
Aligned shapes - Output: (249990,), Label: (249990,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   9%|▉         | 18/200 [00:01<00:10, 17.26it/s]

Calculating Affiliation Metrics:  18%|█▊        | 36/200 [00:02<00:10, 15.83it/s]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:03<00:09, 14.80it/s]

Calculating Affiliation Metrics:  34%|███▎      | 67/200 [00:04<00:09, 14.22it/s]

Calculating Affiliation Metrics:  41%|████      | 82/200 [00:05<00:08, 14.00it/s]

Calculating Affiliation Metrics:  48%|████▊     | 97/200 [00:06<00:07, 13.85it/s]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [00:07<00:06, 14.39it/s]

Calculating Affiliation Metrics:  65%|██████▌   | 130/200 [00:08<00:04, 15.07it/s]

Calculating Affiliation Metrics:  75%|███████▌  | 150/200 [00:09<00:03, 16.32it/s]

Calculating Affiliation Metrics:  86%|████████▌ | 172/200 [00:10<00:01, 17.99it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:11<00:00, 17.05it/s]
Table1-only metrics completed in 16.58s
Evaluation Result:  {'AUC-PR': 0.13922931155819024, 'AUC-ROC': 0.9866724645371473, 'VUS-PR': 0.13528645064940625, 'VUS-ROC': 0.9871541459505582, 'Standard-F1': 0.22539682539649777, 'Standard-Precision': 0.14199999999999996, 'Standard-Recall': 0.5461538461538453, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7703656577864985, 'Affiliation-P': 0.6266722204289922, 'Affiliation-R': 0.9995610039786208, 'F1_T': 0.22539682539649777, 'Precision_T': 0.14199999999999996, 'Recall_T': 0.5461538461538453}
Results for 433_UCR_id_131_Facility_tr_50000_1st_106720.csv: {'filename': '433_UCR_id_131_Facility_tr_50000_1st_106720.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 64, 'train_index': '50000', 'data_shape': '299990x1', 'output_length': 250000, 'label_length': 249990, 'aligned_length': 249990, 'AUC-PR': 0.13922931155819024, 'AUC-ROC': 0.9866724645371473, 'VUS-PR': 0.13528645064940625, 'VUS-ROC': 0.9871541459505582, 'Standard-F1': 0.22539682539649777, 'Standard-Precision': 0.14199999999999996, 'Standard-Recall': 0.5461538461538453, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7703656577864985, 'Affiliation-P': 0.6266722204289922, 'Affiliation-R': 0.9995610039786208, 'F1_T': 0.22539682539649777, 'Precision_T': 0.14199999999999996, 'Recall_T': 0.5461538461538453}
Logits metrics skipped for 433_UCR_id_131_Facility_tr_50000_1st_106720.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/433_UCR_id_131_Facility_tr_50000_1st_106720_results.pkl
Skipping file: 534_SMAP_id_4_Sensor_tr_1968_1st_4370.csv due to filter criteria.
Skipping file: 247_SVDB_id_11_Medical_tr_20100_1st_23611.csv due to filter criteria.
Skipping file: 535_SMAP_id_5_Sensor_tr_1113_1st_1890.csv due to filter criteria.
Skipping file: 819_Exathlon_id_10_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 470_UCR_id_168_Medical_tr_10447_1st_10998.csv, Data shape: (41790, 1)
Label shape: (41790,)
Train data shape: (10447, 1), Test data shape: (31343, 1), Label test shape: (31343,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 470_UCR_id_168_Medical_tr_10447_1st_10998.csv...
Using CUDA device: 2
Data shape: (31343, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 31343, window_size 5000, adding 3657 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([7, 5000, 1])
batched_mask shape: torch.Size([7, 5000])

100%|██████████| 1/1 [00:00<00:00, 133.79it/s]
Original shapes - Output: (35000,), Label: (31343,)
Aligned shapes - Output: (31343,), Label: (31343,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  54%|█████▎    | 107/200 [00:01<00:00, 106.92it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 122.84it/s]
Table1-only metrics completed in 1.94s
Evaluation Result:  {'AUC-PR': 0.10884326225414168, 'AUC-ROC': 0.9942962986618976, 'VUS-PR': 0.11970441445594003, 'VUS-ROC': 0.9953980122935631, 'Standard-F1': 0.13372093023239814, 'Standard-Precision': 0.07324840764331186, 'Standard-Recall': 0.7666666666666412, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6735557661747975, 'Affiliation-P': 0.507790489037294, 'Affiliation-R': 1.0, 'F1_T': 0.13372093023239814, 'Precision_T': 0.07324840764331186, 'Recall_T': 0.7666666666666412}
Results for 470_UCR_id_168_Medical_tr_10447_1st_10998.csv: {'filename': '470_UCR_id_168_Medical_tr_10447_1st_10998.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 33, 'train_index': '10447', 'data_shape': '41790x1', 'output_length': 35000, 'label_length': 31343, 'aligned_length': 31343, 'AUC-PR': 0.10884326225414168, 'AUC-ROC': 0.9942962986618976, 'VUS-PR': 0.11970441445594003, 'VUS-ROC': 0.9953980122935631, 'Standard-F1': 0.13372093023239814, 'Standard-Precision': 0.07324840764331186, 'Standard-Recall': 0.7666666666666412, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6735557661747975, 'Affiliation-P': 0.507790489037294, 'Affiliation-R': 1.0, 'F1_T': 0.13372093023239814, 'Precision_T': 0.07324840764331186, 'Recall_T': 0.7666666666666412}
Logits metrics skipped for 470_UCR_id_168_Medical_tr_10447_1st_10998.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/470_UCR_id_168_Medical_tr_10447_1st_10998_results.pkl
Processing file: 660_YAHOO_id_110_WebService_tr_500_1st_307.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 660_YAHOO_id_110_WebService_tr_500_1st_307.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.85it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1337.37it/s]
Table1-only metrics completed in 0.35s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9504941880689357, 'VUS-ROC': 0.9998948142040056, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984164524323212, 'Affiliation-P': 0.996837912181118, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Results for 660_YAHOO_id_110_WebService_tr_500_1st_307.csv: {'filename': '660_YAHOO_id_110_WebService_tr_500_1st_307.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9504941880689357, 'VUS-ROC': 0.9998948142040056, 'Standard-F1': 0.3999999999996266, 'Standard-Precision': 0.24999999999997916, 'Standard-Recall': 0.9999999999996666, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984164524323212, 'Affiliation-P': 0.996837912181118, 'Affiliation-R': 1.0, 'F1_T': 0.3999999999996266, 'Precision_T': 0.24999999999997916, 'Recall_T': 0.9999999999996666}
Logits metrics skipped for 660_YAHOO_id_110_WebService_tr_500_1st_307.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/660_YAHOO_id_110_WebService_tr_500_1st_307_results.pkl
Skipping file: 866_OPPORTUNITY_id_25_HumanActivity_tr_6213_1st_7311.csv due to filter criteria.
Processing file: 397_UCR_id_95_Medical_tr_1830_1st_4675.csv, Data shape: (7321, 1)
Label shape: (7321,)
Train data shape: (1830, 1), Test data shape: (5491, 1), Label test shape: (5491,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 397_UCR_id_95_Medical_tr_1830_1st_4675.csv...
Using CUDA device: 2
Data shape: (5491, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5491, window_size 5000, adding 4509 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.70it/s]
Original shapes - Output: (10000,), Label: (5491,)
Aligned shapes - Output: (5491,), Label: (5491,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 735.21it/s]
Table1-only metrics completed in 0.97s
Evaluation Result:  {'AUC-PR': 0.7623992478878506, 'AUC-ROC': 0.9437460750734376, 'VUS-PR': 0.8011691372920728, 'VUS-ROC': 0.9614340493741719, 'Standard-F1': 0.712749615974925, 'Standard-Precision': 0.7918088737201338, 'Standard-Recall': 0.6480446927374283, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.998068030504781, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9961435116274574, 'F1_T': 0.7111111111106181, 'Precision_T': 0.8235294117647027, 'Recall_T': 0.6256983240223446}
Results for 397_UCR_id_95_Medical_tr_1830_1st_4675.csv: {'filename': '397_UCR_id_95_Medical_tr_1830_1st_4675.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 178, 'train_index': '1830', 'data_shape': '7321x1', 'output_length': 10000, 'label_length': 5491, 'aligned_length': 5491, 'AUC-PR': 0.7623992478878506, 'AUC-ROC': 0.9437460750734376, 'VUS-PR': 0.8011691372920728, 'VUS-ROC': 0.9614340493741719, 'Standard-F1': 0.712749615974925, 'Standard-Precision': 0.7918088737201338, 'Standard-Recall': 0.6480446927374283, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.998068030504781, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9961435116274574, 'F1_T': 0.7111111111106181, 'Precision_T': 0.8235294117647027, 'Recall_T': 0.6256983240223446}
Logits metrics skipped for 397_UCR_id_95_Medical_tr_1830_1st_4675.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/397_UCR_id_95_Medical_tr_1830_1st_4675_results.pkl
Processing file: 716_YAHOO_id_166_WebService_tr_500_1st_540.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 716_YAHOO_id_166_WebService_tr_500_1st_540.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 160.04it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1119.40it/s]
Table1-only metrics completed in 0.43s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.999944492534466, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9995629370629365, 'Affiliation-P': 0.9991262560069899, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Results for 716_YAHOO_id_166_WebService_tr_500_1st_540.csv: {'filename': '716_YAHOO_id_166_WebService_tr_500_1st_540.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.999944492534466, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9995629370629365, 'Affiliation-P': 0.9991262560069899, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 716_YAHOO_id_166_WebService_tr_500_1st_540.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/716_YAHOO_id_166_WebService_tr_500_1st_540_results.pkl
Skipping file: 827_Exathlon_id_18_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 438_UCR_id_136_Medical_tr_14750_1st_57960.csv, Data shape: (59000, 1)
Label shape: (59000,)
Train data shape: (14750, 1), Test data shape: (44250, 1), Label test shape: (44250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 438_UCR_id_136_Medical_tr_14750_1st_57960.csv...
Using CUDA device: 2
Data shape: (44250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 44250, window_size 5000, adding 750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([9, 5000, 1])
batched_mask shape: torch.Size([9, 5000])

100%|██████████| 1/1 [00:00<00:00, 112.18it/s]
Original shapes - Output: (45000,), Label: (44250,)
Aligned shapes - Output: (44250,), Label: (44250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  43%|████▎     | 86/200 [00:01<00:01, 85.46it/s]

Calculating Affiliation Metrics:  94%|█████████▍| 188/200 [00:02<00:00, 95.01it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 96.64it/s]
Table1-only metrics completed in 2.28s
Evaluation Result:  {'AUC-PR': 0.002831797380821541, 'AUC-ROC': 0.8511392405063292, 'VUS-PR': 0.0035477605411586423, 'VUS-ROC': 0.8781061914204032, 'Standard-F1': 0.00871080139371094, 'Standard-Precision': 0.004393673110720559, 'Standard-Recall': 0.49999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.666658576578355, 'Affiliation-P': 0.4999908987058726, 'Affiliation-R': 1.0, 'F1_T': 0.0077770576798315725, 'Precision_T': 0.003913894324853227, 'Recall_T': 0.59999999999994}
Results for 438_UCR_id_136_Medical_tr_14750_1st_57960.csv: {'filename': '438_UCR_id_136_Medical_tr_14750_1st_57960.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 6, 'train_index': '14750', 'data_shape': '59000x1', 'output_length': 45000, 'label_length': 44250, 'aligned_length': 44250, 'AUC-PR': 0.002831797380821541, 'AUC-ROC': 0.8511392405063292, 'VUS-PR': 0.0035477605411586423, 'VUS-ROC': 0.8781061914204032, 'Standard-F1': 0.00871080139371094, 'Standard-Precision': 0.004393673110720559, 'Standard-Recall': 0.49999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.666658576578355, 'Affiliation-P': 0.4999908987058726, 'Affiliation-R': 1.0, 'F1_T': 0.0077770576798315725, 'Precision_T': 0.003913894324853227, 'Recall_T': 0.59999999999994}
Logits metrics skipped for 438_UCR_id_136_Medical_tr_14750_1st_57960.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/438_UCR_id_136_Medical_tr_14750_1st_57960_results.pkl
Processing file: 757_YAHOO_id_207_WebService_tr_500_1st_33.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 757_YAHOO_id_207_WebService_tr_500_1st_33.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 157.03it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1599.95it/s]
Table1-only metrics completed in 0.34s
Evaluation Result:  {'AUC-PR': 0.5292290723325206, 'AUC-ROC': 0.93218537414966, 'VUS-PR': 0.21246899627109514, 'VUS-ROC': 0.914647108843537, 'Standard-F1': 0.2499999999995938, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.499999999999875, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8993165036314523, 'Affiliation-P': 0.8170527736615854, 'Affiliation-R': 1.0, 'F1_T': 0.2499999999995938, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.499999999999875}
Results for 757_YAHOO_id_207_WebService_tr_500_1st_33.csv: {'filename': '757_YAHOO_id_207_WebService_tr_500_1st_33.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.5292290723325206, 'AUC-ROC': 0.93218537414966, 'VUS-PR': 0.21246899627109514, 'VUS-ROC': 0.914647108843537, 'Standard-F1': 0.2499999999995938, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.499999999999875, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8993165036314523, 'Affiliation-P': 0.8170527736615854, 'Affiliation-R': 1.0, 'F1_T': 0.2499999999995938, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.499999999999875}
Logits metrics skipped for 757_YAHOO_id_207_WebService_tr_500_1st_33.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/757_YAHOO_id_207_WebService_tr_500_1st_33_results.pkl
Skipping file: 537_SMAP_id_7_Sensor_tr_2077_1st_5394.csv due to filter criteria.
Processing file: 269_IOPS_id_10_WebService_tr_1216_1st_1728.csv, Data shape: (4866, 1)
Label shape: (4866,)
Train data shape: (1216, 1), Test data shape: (3650, 1), Label test shape: (3650,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 269_IOPS_id_10_WebService_tr_1216_1st_1728.csv...
Using CUDA device: 2
Data shape: (3650, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 3650, 1])
batched_mask shape: torch.Size([1, 3650])

100%|██████████| 1/1 [00:00<00:00, 157.86it/s]
Original shapes - Output: (3650,), Label: (3650,)
Aligned shapes - Output: (3650,), Label: (3650,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 805.87it/s]
Table1-only metrics completed in 0.37s
Evaluation Result:  {'AUC-PR': 0.9480064088059136, 'AUC-ROC': 0.9960385153783813, 'VUS-PR': 0.9559504931381989, 'VUS-ROC': 0.9987112336893752, 'Standard-F1': 0.9247311827951804, 'Standard-Precision': 0.977272727272705, 'Standard-Recall': 0.8775510204081454, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9983995000175202, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9968041150488498, 'F1_T': 0.8846153846148691, 'Precision_T': 0.8363636363636211, 'Recall_T': 0.9387755102040625}
Results for 269_IOPS_id_10_WebService_tr_1216_1st_1728.csv: {'filename': '269_IOPS_id_10_WebService_tr_1216_1st_1728.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 15, 'train_index': '1216', 'data_shape': '4866x1', 'output_length': 3650, 'label_length': 3650, 'aligned_length': 3650, 'AUC-PR': 0.9480064088059136, 'AUC-ROC': 0.9960385153783813, 'VUS-PR': 0.9559504931381989, 'VUS-ROC': 0.9987112336893752, 'Standard-F1': 0.9247311827951804, 'Standard-Precision': 0.977272727272705, 'Standard-Recall': 0.8775510204081454, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9983995000175202, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9968041150488498, 'F1_T': 0.8846153846148691, 'Precision_T': 0.8363636363636211, 'Recall_T': 0.9387755102040625}
Logits metrics skipped for 269_IOPS_id_10_WebService_tr_1216_1st_1728.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/269_IOPS_id_10_WebService_tr_1216_1st_1728_results.pkl
Processing file: 164_Stock_id_16_Finance_tr_500_1st_8.csv, Data shape: (20000, 1)
Label shape: (20000,)
Train data shape: (500, 1), Test data shape: (19500, 1), Label test shape: (19500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 164_Stock_id_16_Finance_tr_500_1st_8.csv...
Using CUDA device: 2
Data shape: (19500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 19500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.69it/s]
Original shapes - Output: (20000,), Label: (19500,)
Aligned shapes - Output: (19500,), Label: (19500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   3%|▎         | 6/200 [00:01<00:38,  5.09it/s]

Calculating Affiliation Metrics:   6%|▌         | 12/200 [00:03<01:00,  3.10it/s]

Calculating Affiliation Metrics:   8%|▊         | 16/200 [00:05<01:15,  2.44it/s]

Calculating Affiliation Metrics:  10%|▉         | 19/200 [00:07<01:26,  2.10it/s]

Calculating Affiliation Metrics:  11%|█         | 22/200 [00:09<01:34,  1.88it/s]

Calculating Affiliation Metrics:  12%|█▏        | 24/200 [00:11<01:41,  1.74it/s]

Calculating Affiliation Metrics:  13%|█▎        | 26/200 [00:12<01:48,  1.60it/s]

Calculating Affiliation Metrics:  14%|█▍        | 28/200 [00:14<01:54,  1.51it/s]

Calculating Affiliation Metrics:  15%|█▌        | 30/200 [00:16<02:00,  1.41it/s]

Calculating Affiliation Metrics:  16%|█▌        | 32/200 [00:17<02:07,  1.32it/s]

Calculating Affiliation Metrics:  17%|█▋        | 34/200 [00:19<02:15,  1.23it/s]

Calculating Affiliation Metrics:  18%|█▊        | 36/200 [00:21<02:23,  1.14it/s]

Calculating Affiliation Metrics:  19%|█▉        | 38/200 [00:23<02:28,  1.09it/s]

Calculating Affiliation Metrics:  20%|██        | 40/200 [00:26<02:32,  1.05it/s]

Calculating Affiliation Metrics:  21%|██        | 42/200 [00:28<02:37,  1.00it/s]

Calculating Affiliation Metrics:  22%|██▏       | 44/200 [00:30<02:41,  1.04s/it]

Calculating Affiliation Metrics:  22%|██▎       | 45/200 [00:31<02:45,  1.07s/it]

Calculating Affiliation Metrics:  23%|██▎       | 46/200 [00:32<02:48,  1.10s/it]

Calculating Affiliation Metrics:  24%|██▎       | 47/200 [00:34<02:52,  1.13s/it]

Calculating Affiliation Metrics:  24%|██▍       | 48/200 [00:35<02:56,  1.16s/it]

Calculating Affiliation Metrics:  24%|██▍       | 49/200 [00:36<03:00,  1.19s/it]

Calculating Affiliation Metrics:  25%|██▌       | 50/200 [00:37<03:02,  1.22s/it]

Calculating Affiliation Metrics:  26%|██▌       | 51/200 [00:39<03:04,  1.24s/it]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:40<03:04,  1.25s/it]

Calculating Affiliation Metrics:  26%|██▋       | 53/200 [00:41<03:05,  1.26s/it]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:43<03:06,  1.28s/it]

Calculating Affiliation Metrics:  28%|██▊       | 55/200 [00:44<03:08,  1.30s/it]

Calculating Affiliation Metrics:  28%|██▊       | 56/200 [00:45<03:10,  1.32s/it]

Calculating Affiliation Metrics:  28%|██▊       | 57/200 [00:47<03:11,  1.34s/it]

Calculating Affiliation Metrics:  29%|██▉       | 58/200 [00:48<03:13,  1.36s/it]

Calculating Affiliation Metrics:  30%|██▉       | 59/200 [00:50<03:15,  1.39s/it]

Calculating Affiliation Metrics:  30%|███       | 60/200 [00:51<03:15,  1.40s/it]

Calculating Affiliation Metrics:  30%|███       | 61/200 [00:52<03:14,  1.40s/it]

Calculating Affiliation Metrics:  31%|███       | 62/200 [00:54<03:11,  1.39s/it]

Calculating Affiliation Metrics:  32%|███▏      | 63/200 [00:55<03:08,  1.38s/it]

Calculating Affiliation Metrics:  32%|███▏      | 64/200 [00:56<03:04,  1.36s/it]

Calculating Affiliation Metrics:  32%|███▎      | 65/200 [00:58<03:01,  1.35s/it]

Calculating Affiliation Metrics:  33%|███▎      | 66/200 [00:59<02:58,  1.33s/it]

Calculating Affiliation Metrics:  34%|███▎      | 67/200 [01:00<02:55,  1.32s/it]

Calculating Affiliation Metrics:  34%|███▍      | 68/200 [01:02<02:52,  1.31s/it]

Calculating Affiliation Metrics:  34%|███▍      | 69/200 [01:03<02:49,  1.30s/it]

Calculating Affiliation Metrics:  35%|███▌      | 70/200 [01:04<02:46,  1.28s/it]

Calculating Affiliation Metrics:  36%|███▌      | 71/200 [01:05<02:44,  1.28s/it]

Calculating Affiliation Metrics:  36%|███▌      | 72/200 [01:07<02:42,  1.27s/it]

Calculating Affiliation Metrics:  36%|███▋      | 73/200 [01:08<02:40,  1.27s/it]

Calculating Affiliation Metrics:  37%|███▋      | 74/200 [01:09<02:40,  1.28s/it]

Calculating Affiliation Metrics:  38%|███▊      | 75/200 [01:11<02:40,  1.29s/it]

Calculating Affiliation Metrics:  38%|███▊      | 76/200 [01:12<02:39,  1.29s/it]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [01:13<02:39,  1.30s/it]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [01:15<02:40,  1.31s/it]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [01:16<02:39,  1.32s/it]

Calculating Affiliation Metrics:  40%|████      | 80/200 [01:17<02:38,  1.32s/it]

Calculating Affiliation Metrics:  40%|████      | 81/200 [01:18<02:36,  1.32s/it]

Calculating Affiliation Metrics:  41%|████      | 82/200 [01:20<02:35,  1.32s/it]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [01:21<02:33,  1.31s/it]

Calculating Affiliation Metrics:  42%|████▏     | 84/200 [01:22<02:30,  1.30s/it]

Calculating Affiliation Metrics:  42%|████▎     | 85/200 [01:24<02:27,  1.28s/it]

Calculating Affiliation Metrics:  43%|████▎     | 86/200 [01:25<02:25,  1.28s/it]

Calculating Affiliation Metrics:  44%|████▎     | 87/200 [01:26<02:26,  1.30s/it]

Calculating Affiliation Metrics:  44%|████▍     | 88/200 [01:28<02:28,  1.32s/it]

Calculating Affiliation Metrics:  44%|████▍     | 89/200 [01:29<02:26,  1.32s/it]

Calculating Affiliation Metrics:  45%|████▌     | 90/200 [01:30<02:24,  1.32s/it]

Calculating Affiliation Metrics:  46%|████▌     | 91/200 [01:32<02:24,  1.33s/it]

Calculating Affiliation Metrics:  46%|████▌     | 92/200 [01:33<02:22,  1.32s/it]

Calculating Affiliation Metrics:  46%|████▋     | 93/200 [01:34<02:19,  1.31s/it]

Calculating Affiliation Metrics:  47%|████▋     | 94/200 [01:35<02:17,  1.30s/it]

Calculating Affiliation Metrics:  48%|████▊     | 95/200 [01:37<02:14,  1.29s/it]

Calculating Affiliation Metrics:  48%|████▊     | 96/200 [01:38<02:12,  1.28s/it]

Calculating Affiliation Metrics:  48%|████▊     | 97/200 [01:39<02:11,  1.28s/it]

Calculating Affiliation Metrics:  49%|████▉     | 98/200 [01:41<02:09,  1.27s/it]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [01:42<02:07,  1.27s/it]

Calculating Affiliation Metrics:  50%|█████     | 100/200 [01:43<02:04,  1.25s/it]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [01:44<02:02,  1.24s/it]

Calculating Affiliation Metrics:  51%|█████     | 102/200 [01:45<01:59,  1.22s/it]

Calculating Affiliation Metrics:  52%|█████▏    | 103/200 [01:47<01:58,  1.22s/it]

Calculating Affiliation Metrics:  52%|█████▏    | 104/200 [01:48<01:54,  1.20s/it]

Calculating Affiliation Metrics:  52%|█████▎    | 105/200 [01:49<01:51,  1.17s/it]

Calculating Affiliation Metrics:  53%|█████▎    | 106/200 [01:50<01:47,  1.15s/it]

Calculating Affiliation Metrics:  54%|█████▎    | 107/200 [01:51<01:44,  1.13s/it]

Calculating Affiliation Metrics:  54%|█████▍    | 108/200 [01:52<01:41,  1.11s/it]

Calculating Affiliation Metrics:  55%|█████▍    | 109/200 [01:53<01:39,  1.09s/it]

Calculating Affiliation Metrics:  55%|█████▌    | 110/200 [01:54<01:36,  1.07s/it]

Calculating Affiliation Metrics:  56%|█████▌    | 111/200 [01:55<01:34,  1.06s/it]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [01:56<01:32,  1.06s/it]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [01:57<01:32,  1.06s/it]

Calculating Affiliation Metrics:  57%|█████▋    | 114/200 [01:58<01:30,  1.06s/it]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [01:59<01:29,  1.05s/it]

Calculating Affiliation Metrics:  58%|█████▊    | 116/200 [02:00<01:27,  1.04s/it]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [02:02<01:22,  1.01s/it]

Calculating Affiliation Metrics:  60%|██████    | 120/200 [02:04<01:19,  1.01it/s]

Calculating Affiliation Metrics:  61%|██████    | 122/200 [02:06<01:16,  1.02it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 124/200 [02:08<01:12,  1.04it/s]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [02:10<01:08,  1.07it/s]

Calculating Affiliation Metrics:  64%|██████▍   | 128/200 [02:12<01:05,  1.10it/s]

Calculating Affiliation Metrics:  65%|██████▌   | 130/200 [02:13<01:02,  1.12it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 132/200 [02:15<00:59,  1.15it/s]

Calculating Affiliation Metrics:  67%|██████▋   | 134/200 [02:16<00:56,  1.18it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 136/200 [02:18<00:53,  1.19it/s]

Calculating Affiliation Metrics:  69%|██████▉   | 138/200 [02:20<00:50,  1.23it/s]

Calculating Affiliation Metrics:  70%|███████   | 140/200 [02:21<00:47,  1.28it/s]

Calculating Affiliation Metrics:  71%|███████   | 142/200 [02:22<00:43,  1.32it/s]

Calculating Affiliation Metrics:  72%|███████▏  | 144/200 [02:24<00:40,  1.37it/s]

Calculating Affiliation Metrics:  73%|███████▎  | 146/200 [02:25<00:38,  1.39it/s]

Calculating Affiliation Metrics:  74%|███████▍  | 148/200 [02:27<00:36,  1.41it/s]

Calculating Affiliation Metrics:  75%|███████▌  | 150/200 [02:28<00:34,  1.43it/s]

Calculating Affiliation Metrics:  76%|███████▌  | 152/200 [02:29<00:32,  1.47it/s]

Calculating Affiliation Metrics:  77%|███████▋  | 154/200 [02:30<00:30,  1.49it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 156/200 [02:32<00:29,  1.51it/s]

Calculating Affiliation Metrics:  79%|███████▉  | 158/200 [02:33<00:27,  1.54it/s]

Calculating Affiliation Metrics:  80%|████████  | 160/200 [02:34<00:25,  1.57it/s]

Calculating Affiliation Metrics:  81%|████████  | 162/200 [02:35<00:23,  1.61it/s]

Calculating Affiliation Metrics:  82%|████████▏ | 164/200 [02:36<00:21,  1.65it/s]

Calculating Affiliation Metrics:  83%|████████▎ | 166/200 [02:38<00:19,  1.71it/s]

Calculating Affiliation Metrics:  84%|████████▍ | 169/200 [02:39<00:16,  1.83it/s]

Calculating Affiliation Metrics:  86%|████████▌ | 172/200 [02:40<00:14,  1.99it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 175/200 [02:41<00:11,  2.19it/s]

Calculating Affiliation Metrics:  89%|████████▉ | 178/200 [02:42<00:09,  2.39it/s]

Calculating Affiliation Metrics:  90%|█████████ | 181/200 [02:43<00:07,  2.55it/s]

Calculating Affiliation Metrics:  92%|█████████▎| 185/200 [02:45<00:05,  2.74it/s]

Calculating Affiliation Metrics:  95%|█████████▌| 190/200 [02:46<00:03,  3.27it/s]

Calculating Affiliation Metrics: 100%|█████████▉| 199/200 [02:47<00:00,  4.78it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [02:47<00:00,  1.20it/s]
Table1-only metrics completed in 246.88s
Evaluation Result:  {'AUC-PR': 0.17231741870480738, 'AUC-ROC': 0.5481787169055296, 'VUS-PR': 0.9637982196542001, 'VUS-ROC': 0.9778263572201926, 'Standard-F1': 0.24655850540778515, 'Standard-Precision': 0.14793510324483775, 'Standard-Recall': 0.7396755162241886, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6866281774511278, 'Affiliation-P': 0.5233960076729521, 'Affiliation-R': 0.9978193505886354, 'F1_T': 0.24632420793574164, 'Precision_T': 0.1478254448792734, 'Recall_T': 0.7382005899705012}
Results for 164_Stock_id_16_Finance_tr_500_1st_8.csv: {'filename': '164_Stock_id_16_Finance_tr_500_1st_8.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 207, 'train_index': '500', 'data_shape': '20000x1', 'output_length': 20000, 'label_length': 19500, 'aligned_length': 19500, 'AUC-PR': 0.17231741870480738, 'AUC-ROC': 0.5481787169055296, 'VUS-PR': 0.9637982196542001, 'VUS-ROC': 0.9778263572201926, 'Standard-F1': 0.24655850540778515, 'Standard-Precision': 0.14793510324483775, 'Standard-Recall': 0.7396755162241886, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6866281774511278, 'Affiliation-P': 0.5233960076729521, 'Affiliation-R': 0.9978193505886354, 'F1_T': 0.24632420793574164, 'Precision_T': 0.1478254448792734, 'Recall_T': 0.7382005899705012}
Logits metrics skipped for 164_Stock_id_16_Finance_tr_500_1st_8.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/164_Stock_id_16_Finance_tr_500_1st_8_results.pkl
Processing file: 483_UCR_id_181_Medical_tr_2499_1st_6508.csv, Data shape: (9998, 1)
Label shape: (9998,)
Train data shape: (2499, 1), Test data shape: (7499, 1), Label test shape: (7499,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 483_UCR_id_181_Medical_tr_2499_1st_6508.csv...
Using CUDA device: 2
Data shape: (7499, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 7499, window_size 5000, adding 2501 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.89it/s]
Original shapes - Output: (10000,), Label: (7499,)
Aligned shapes - Output: (7499,), Label: (7499,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 513.06it/s]
Table1-only metrics completed in 0.53s
Evaluation Result:  {'AUC-PR': 0.923220972368547, 'AUC-ROC': 0.9989528795811519, 'VUS-PR': 0.8619506003510327, 'VUS-ROC': 0.9988253739646312, 'Standard-F1': 0.7359999999995083, 'Standard-Precision': 0.6133333333333252, 'Standard-Recall': 0.9199999999999816, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984120203945577, 'Affiliation-P': 0.9969960907068312, 'Affiliation-R': 0.999831977597013, 'F1_T': 0.7359999999995083, 'Precision_T': 0.6133333333333252, 'Recall_T': 0.9199999999999816}
Results for 483_UCR_id_181_Medical_tr_2499_1st_6508.csv: {'filename': '483_UCR_id_181_Medical_tr_2499_1st_6508.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 26, 'train_index': '2499', 'data_shape': '9998x1', 'output_length': 10000, 'label_length': 7499, 'aligned_length': 7499, 'AUC-PR': 0.923220972368547, 'AUC-ROC': 0.9989528795811519, 'VUS-PR': 0.8619506003510327, 'VUS-ROC': 0.9988253739646312, 'Standard-F1': 0.7359999999995083, 'Standard-Precision': 0.6133333333333252, 'Standard-Recall': 0.9199999999999816, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984120203945577, 'Affiliation-P': 0.9969960907068312, 'Affiliation-R': 0.999831977597013, 'F1_T': 0.7359999999995083, 'Precision_T': 0.6133333333333252, 'Recall_T': 0.9199999999999816}
Logits metrics skipped for 483_UCR_id_181_Medical_tr_2499_1st_6508.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/483_UCR_id_181_Medical_tr_2499_1st_6508_results.pkl
Processing file: 361_UCR_id_59_Medical_tr_3000_1st_7175.csv, Data shape: (12000, 1)
Label shape: (12000,)
Train data shape: (3000, 1), Test data shape: (9000, 1), Label test shape: (9000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 361_UCR_id_59_Medical_tr_3000_1st_7175.csv...
Using CUDA device: 2
Data shape: (9000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9000, window_size 5000, adding 1000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.35it/s]
Original shapes - Output: (10000,), Label: (9000,)
Aligned shapes - Output: (9000,), Label: (9000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 520.70it/s]
Table1-only metrics completed in 1.29s
Evaluation Result:  {'AUC-PR': 0.810063606605279, 'AUC-ROC': 0.9864476491359675, 'VUS-PR': 0.8047160214985806, 'VUS-ROC': 0.9902445264040542, 'Standard-F1': 0.7843137254897103, 'Standard-Precision': 0.9722222222222155, 'Standard-Recall': 0.6572769953051613, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9988016433858883, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9976061554512259, 'F1_T': 0.7758620689650378, 'Precision_T': 0.9999999999999927, 'Recall_T': 0.6338028169014055}
Results for 361_UCR_id_59_Medical_tr_3000_1st_7175.csv: {'filename': '361_UCR_id_59_Medical_tr_3000_1st_7175.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 218, 'train_index': '3000', 'data_shape': '12000x1', 'output_length': 10000, 'label_length': 9000, 'aligned_length': 9000, 'AUC-PR': 0.810063606605279, 'AUC-ROC': 0.9864476491359675, 'VUS-PR': 0.8047160214985806, 'VUS-ROC': 0.9902445264040542, 'Standard-F1': 0.7843137254897103, 'Standard-Precision': 0.9722222222222155, 'Standard-Recall': 0.6572769953051613, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9988016433858883, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9976061554512259, 'F1_T': 0.7758620689650378, 'Precision_T': 0.9999999999999927, 'Recall_T': 0.6338028169014055}
Logits metrics skipped for 361_UCR_id_59_Medical_tr_3000_1st_7175.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/361_UCR_id_59_Medical_tr_3000_1st_7175_results.pkl
Skipping file: 204_SMD_id_27_Facility_tr_5000_1st_15144.csv due to filter criteria.
Processing file: 638_YAHOO_id_88_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 638_YAHOO_id_88_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 158.63it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2258.99it/s]
Table1-only metrics completed in 0.47s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9628378378378373, 'Affiliation-P': 0.9283387622149837, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 638_YAHOO_id_88_Synthetic_tr_500_1st_893.csv: {'filename': '638_YAHOO_id_88_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 117, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9628378378378373, 'Affiliation-P': 0.9283387622149837, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 638_YAHOO_id_88_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/638_YAHOO_id_88_Synthetic_tr_500_1st_893_results.pkl
Skipping file: 829_Exathlon_id_20_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 154_Stock_id_6_Finance_tr_500_1st_38.csv, Data shape: (20000, 1)
Label shape: (20000,)
Train data shape: (500, 1), Test data shape: (19500, 1), Label test shape: (19500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 154_Stock_id_6_Finance_tr_500_1st_38.csv...
Using CUDA device: 2
Data shape: (19500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 19500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 164.73it/s]
Original shapes - Output: (20000,), Label: (19500,)
Aligned shapes - Output: (19500,), Label: (19500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   4%|▍         | 9/200 [00:01<00:21,  8.97it/s]

Calculating Affiliation Metrics:   9%|▉         | 18/200 [00:03<00:33,  5.45it/s]

Calculating Affiliation Metrics:  12%|█▎        | 25/200 [00:05<00:43,  4.04it/s]

Calculating Affiliation Metrics:  15%|█▌        | 30/200 [00:07<00:49,  3.42it/s]

Calculating Affiliation Metrics:  17%|█▋        | 34/200 [00:09<00:54,  3.05it/s]

Calculating Affiliation Metrics:  19%|█▉        | 38/200 [00:11<00:57,  2.81it/s]

Calculating Affiliation Metrics:  20%|██        | 41/200 [00:12<01:00,  2.65it/s]

Calculating Affiliation Metrics:  22%|██▏       | 44/200 [00:13<01:02,  2.50it/s]

Calculating Affiliation Metrics:  24%|██▎       | 47/200 [00:15<01:04,  2.37it/s]

Calculating Affiliation Metrics:  25%|██▌       | 50/200 [00:16<01:06,  2.26it/s]

Calculating Affiliation Metrics:  26%|██▋       | 53/200 [00:18<01:07,  2.18it/s]

Calculating Affiliation Metrics:  28%|██▊       | 56/200 [00:19<01:08,  2.11it/s]

Calculating Affiliation Metrics:  30%|██▉       | 59/200 [00:21<01:07,  2.08it/s]

Calculating Affiliation Metrics:  31%|███       | 62/200 [00:22<01:06,  2.08it/s]

Calculating Affiliation Metrics:  32%|███▎      | 65/200 [00:24<01:05,  2.07it/s]

Calculating Affiliation Metrics:  34%|███▍      | 68/200 [00:25<01:04,  2.05it/s]

Calculating Affiliation Metrics:  36%|███▌      | 71/200 [00:27<01:03,  2.02it/s]

Calculating Affiliation Metrics:  37%|███▋      | 74/200 [00:28<01:02,  2.02it/s]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [00:30<01:01,  2.01it/s]

Calculating Affiliation Metrics:  40%|████      | 80/200 [00:31<00:59,  2.00it/s]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [00:33<00:58,  2.00it/s]

Calculating Affiliation Metrics:  43%|████▎     | 86/200 [00:34<00:56,  2.02it/s]

Calculating Affiliation Metrics:  44%|████▍     | 89/200 [00:36<00:55,  2.01it/s]

Calculating Affiliation Metrics:  46%|████▌     | 92/200 [00:37<00:54,  2.00it/s]

Calculating Affiliation Metrics:  47%|████▋     | 94/200 [00:38<00:53,  1.99it/s]

Calculating Affiliation Metrics:  48%|████▊     | 96/200 [00:39<00:52,  1.97it/s]

Calculating Affiliation Metrics:  49%|████▉     | 98/200 [00:40<00:52,  1.96it/s]

Calculating Affiliation Metrics:  50%|█████     | 100/200 [00:41<00:51,  1.95it/s]

Calculating Affiliation Metrics:  51%|█████     | 102/200 [00:42<00:50,  1.92it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 104/200 [00:44<00:50,  1.90it/s]

Calculating Affiliation Metrics:  53%|█████▎    | 106/200 [00:45<00:49,  1.89it/s]

Calculating Affiliation Metrics:  54%|█████▍    | 108/200 [00:46<00:49,  1.88it/s]

Calculating Affiliation Metrics:  55%|█████▌    | 110/200 [00:47<00:47,  1.88it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [00:48<00:46,  1.88it/s]

Calculating Affiliation Metrics:  57%|█████▋    | 114/200 [00:49<00:45,  1.90it/s]

Calculating Affiliation Metrics:  58%|█████▊    | 117/200 [00:50<00:42,  1.95it/s]

Calculating Affiliation Metrics:  60%|██████    | 120/200 [00:52<00:40,  1.99it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 123/200 [00:53<00:37,  2.04it/s]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [00:55<00:35,  2.09it/s]

Calculating Affiliation Metrics:  64%|██████▍   | 129/200 [00:56<00:32,  2.15it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 132/200 [00:57<00:30,  2.22it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 135/200 [00:58<00:28,  2.29it/s]

Calculating Affiliation Metrics:  69%|██████▉   | 138/200 [01:00<00:26,  2.35it/s]

Calculating Affiliation Metrics:  70%|███████   | 141/200 [01:01<00:23,  2.48it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [01:02<00:20,  2.66it/s]

Calculating Affiliation Metrics:  74%|███████▍  | 149/200 [01:03<00:17,  2.86it/s]

Calculating Affiliation Metrics:  76%|███████▋  | 153/200 [01:04<00:15,  3.08it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 157/200 [01:05<00:13,  3.28it/s]

Calculating Affiliation Metrics:  81%|████████  | 162/200 [01:06<00:10,  3.61it/s]

Calculating Affiliation Metrics:  84%|████████▎ | 167/200 [01:07<00:08,  3.86it/s]

Calculating Affiliation Metrics:  86%|████████▋ | 173/200 [01:09<00:06,  4.34it/s]

Calculating Affiliation Metrics:  90%|█████████ | 180/200 [01:10<00:04,  4.95it/s]

Calculating Affiliation Metrics:  94%|█████████▍| 189/200 [01:11<00:01,  5.93it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [01:11<00:00,  2.78it/s]
Table1-only metrics completed in 99.48s
Evaluation Result:  {'AUC-PR': 0.14011368739666885, 'AUC-ROC': 0.6421032185754476, 'VUS-PR': 0.618401993374832, 'VUS-ROC': 0.8282579850427653, 'Standard-F1': 0.17810550198475572, 'Standard-Precision': 0.19430693069306906, 'Standard-Recall': 0.16439790575916213, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6740550311822959, 'Affiliation-P': 0.5098899409556711, 'Affiliation-R': 0.9941254709508188, 'F1_T': 0.17772901810153058, 'Precision_T': 0.1866359447004606, 'Recall_T': 0.16963350785340295}
Results for 154_Stock_id_6_Finance_tr_500_1st_38.csv: {'filename': '154_Stock_id_6_Finance_tr_500_1st_38.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 42, 'train_index': '500', 'data_shape': '20000x1', 'output_length': 20000, 'label_length': 19500, 'aligned_length': 19500, 'AUC-PR': 0.14011368739666885, 'AUC-ROC': 0.6421032185754476, 'VUS-PR': 0.618401993374832, 'VUS-ROC': 0.8282579850427653, 'Standard-F1': 0.17810550198475572, 'Standard-Precision': 0.19430693069306906, 'Standard-Recall': 0.16439790575916213, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6740550311822959, 'Affiliation-P': 0.5098899409556711, 'Affiliation-R': 0.9941254709508188, 'F1_T': 0.17772901810153058, 'Precision_T': 0.1866359447004606, 'Recall_T': 0.16963350785340295}
Logits metrics skipped for 154_Stock_id_6_Finance_tr_500_1st_38.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/154_Stock_id_6_Finance_tr_500_1st_38_results.pkl
Skipping file: 174_MITDB_id_5_Medical_tr_50000_1st_539948.csv due to filter criteria.
Skipping file: 241_SVDB_id_5_Medical_tr_11587_1st_11687.csv due to filter criteria.
Processing file: 007_NAB_id_7_Traffic_tr_624_1st_2087.csv, Data shape: (2499, 1)
Label shape: (2499,)
Train data shape: (624, 1), Test data shape: (1875, 1), Label test shape: (1875,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 007_NAB_id_7_Traffic_tr_624_1st_2087.csv...
Using CUDA device: 2
Data shape: (1875, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1875, 1])
batched_mask shape: torch.Size([1, 1875])

100%|██████████| 1/1 [00:00<00:00, 160.98it/s]
Original shapes - Output: (1875,), Label: (1875,)
Aligned shapes - Output: (1875,), Label: (1875,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1246.90it/s]
Table1-only metrics completed in 0.89s
Evaluation Result:  {'AUC-PR': 0.2877633519194782, 'AUC-ROC': 0.5076655034994151, 'VUS-PR': 0.2962914430226261, 'VUS-ROC': 0.5557747037424682, 'Standard-F1': 0.3096085409247727, 'Standard-Precision': 0.2770700636942666, 'Standard-Recall': 0.3508064516129018, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9248091404997869, 'Affiliation-P': 1.0, 'Affiliation-R': 0.8601348610140456, 'F1_T': 0.30905861456433725, 'Precision_T': 0.2761904761904753, 'Recall_T': 0.3508064516129018}
Results for 007_NAB_id_7_Traffic_tr_624_1st_2087.csv: {'filename': '007_NAB_id_7_Traffic_tr_624_1st_2087.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '624', 'data_shape': '2499x1', 'output_length': 1875, 'label_length': 1875, 'aligned_length': 1875, 'AUC-PR': 0.2877633519194782, 'AUC-ROC': 0.5076655034994151, 'VUS-PR': 0.2962914430226261, 'VUS-ROC': 0.5557747037424682, 'Standard-F1': 0.3096085409247727, 'Standard-Precision': 0.2770700636942666, 'Standard-Recall': 0.3508064516129018, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9248091404997869, 'Affiliation-P': 1.0, 'Affiliation-R': 0.8601348610140456, 'F1_T': 0.30905861456433725, 'Precision_T': 0.2761904761904753, 'Recall_T': 0.3508064516129018}
Logits metrics skipped for 007_NAB_id_7_Traffic_tr_624_1st_2087.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/007_NAB_id_7_Traffic_tr_624_1st_2087_results.pkl
Processing file: 683_YAHOO_id_133_WebService_tr_500_1st_36.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 683_YAHOO_id_133_WebService_tr_500_1st_36.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.85it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 836.27it/s]
Table1-only metrics completed in 0.69s
Evaluation Result:  {'AUC-PR': 0.9513888888888891, 'AUC-ROC': 0.9993358003605656, 'VUS-PR': 0.8793752554671704, 'VUS-ROC': 0.9989321682631774, 'Standard-F1': 0.7619047619041995, 'Standard-Precision': 0.6666666666666111, 'Standard-Recall': 0.8888888888887901, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9081144229554076, 'Affiliation-P': 0.8316937617340844, 'Affiliation-R': 1.0, 'F1_T': 0.7619047619041995, 'Precision_T': 0.6666666666666111, 'Recall_T': 0.8888888888887901}
Results for 683_YAHOO_id_133_WebService_tr_500_1st_36.csv: {'filename': '683_YAHOO_id_133_WebService_tr_500_1st_36.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9513888888888891, 'AUC-ROC': 0.9993358003605656, 'VUS-PR': 0.8793752554671704, 'VUS-ROC': 0.9989321682631774, 'Standard-F1': 0.7619047619041995, 'Standard-Precision': 0.6666666666666111, 'Standard-Recall': 0.8888888888887901, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9081144229554076, 'Affiliation-P': 0.8316937617340844, 'Affiliation-R': 1.0, 'F1_T': 0.7619047619041995, 'Precision_T': 0.6666666666666111, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 683_YAHOO_id_133_WebService_tr_500_1st_36.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/683_YAHOO_id_133_WebService_tr_500_1st_36_results.pkl
Processing file: 452_UCR_id_150_Facility_tr_10000_1st_35774.csv, Data shape: (40000, 1)
Label shape: (40000,)
Train data shape: (10000, 1), Test data shape: (30000, 1), Label test shape: (30000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 452_UCR_id_150_Facility_tr_10000_1st_35774.csv...
Using CUDA device: 2
Data shape: (30000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 84.80it/s]
Original shapes - Output: (30000,), Label: (30000,)
Aligned shapes - Output: (30000,), Label: (30000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  64%|██████▍   | 129/200 [00:01<00:00, 128.24it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 144.43it/s]
Table1-only metrics completed in 2.10s
Evaluation Result:  {'AUC-PR': 0.1220922796362022, 'AUC-ROC': 0.8106844481605351, 'VUS-PR': 0.08603773936107596, 'VUS-ROC': 0.875261339467487, 'Standard-F1': 0.16499999999962417, 'Standard-Precision': 0.10999999999999963, 'Standard-Recall': 0.32999999999999674, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7098786592600494, 'Affiliation-P': 0.5504199659863945, 'Affiliation-R': 0.9994120000000001, 'F1_T': 0.16499999999962417, 'Precision_T': 0.10999999999999963, 'Recall_T': 0.32999999999999674}
Results for 452_UCR_id_150_Facility_tr_10000_1st_35774.csv: {'filename': '452_UCR_id_150_Facility_tr_10000_1st_35774.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 108, 'train_index': '10000', 'data_shape': '40000x1', 'output_length': 30000, 'label_length': 30000, 'aligned_length': 30000, 'AUC-PR': 0.1220922796362022, 'AUC-ROC': 0.8106844481605351, 'VUS-PR': 0.08603773936107596, 'VUS-ROC': 0.875261339467487, 'Standard-F1': 0.16499999999962417, 'Standard-Precision': 0.10999999999999963, 'Standard-Recall': 0.32999999999999674, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7098786592600494, 'Affiliation-P': 0.5504199659863945, 'Affiliation-R': 0.9994120000000001, 'F1_T': 0.16499999999962417, 'Precision_T': 0.10999999999999963, 'Recall_T': 0.32999999999999674}
Logits metrics skipped for 452_UCR_id_150_Facility_tr_10000_1st_35774.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/452_UCR_id_150_Facility_tr_10000_1st_35774_results.pkl
Processing file: 236_SED_id_3_Medical_tr_3740_1st_3840.csv, Data shape: (39999, 1)
Label shape: (39999,)
Train data shape: (3740, 1), Test data shape: (36259, 1), Label test shape: (36259,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 236_SED_id_3_Medical_tr_3740_1st_3840.csv...
Using CUDA device: 2
Data shape: (36259, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 36259, window_size 5000, adding 3741 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([8, 5000, 1])
batched_mask shape: torch.Size([8, 5000])

100%|██████████| 1/1 [00:00<00:00, 116.98it/s]
Original shapes - Output: (40000,), Label: (36259,)
Aligned shapes - Output: (36259,), Label: (36259,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  22%|██▏       | 44/200 [00:01<00:03, 42.96it/s]

Calculating Affiliation Metrics:  44%|████▎     | 87/200 [00:02<00:02, 38.29it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 135/200 [00:03<00:01, 42.37it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:03<00:00, 50.10it/s]
Table1-only metrics completed in 10.46s
Evaluation Result:  {'AUC-PR': 0.4564877044289441, 'AUC-ROC': 0.9280721277688692, 'VUS-PR': 0.576131307924147, 'VUS-ROC': 0.9600859219994065, 'Standard-F1': 0.520880361173329, 'Standard-Precision': 0.44546332046332027, 'Standard-Recall': 0.6270380434782605, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8548564727076424, 'Affiliation-P': 0.8142797797217416, 'Affiliation-R': 0.8996892376723143, 'F1_T': 0.5200464846014858, 'Precision_T': 0.45431472081218255, 'Recall_T': 0.6080163043478257}
Results for 236_SED_id_3_Medical_tr_3740_1st_3840.csv: {'filename': '236_SED_id_3_Medical_tr_3740_1st_3840.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '3740', 'data_shape': '39999x1', 'output_length': 40000, 'label_length': 36259, 'aligned_length': 36259, 'AUC-PR': 0.4564877044289441, 'AUC-ROC': 0.9280721277688692, 'VUS-PR': 0.576131307924147, 'VUS-ROC': 0.9600859219994065, 'Standard-F1': 0.520880361173329, 'Standard-Precision': 0.44546332046332027, 'Standard-Recall': 0.6270380434782605, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8548564727076424, 'Affiliation-P': 0.8142797797217416, 'Affiliation-R': 0.8996892376723143, 'F1_T': 0.5200464846014858, 'Precision_T': 0.45431472081218255, 'Recall_T': 0.6080163043478257}
Logits metrics skipped for 236_SED_id_3_Medical_tr_3740_1st_3840.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/236_SED_id_3_Medical_tr_3740_1st_3840_results.pkl
Processing file: 674_YAHOO_id_124_WebService_tr_500_1st_514.csv, Data shape: (1439, 1)
Label shape: (1439,)
Train data shape: (500, 1), Test data shape: (939, 1), Label test shape: (939,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 674_YAHOO_id_124_WebService_tr_500_1st_514.csv...
Using CUDA device: 2
Data shape: (939, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 939, 1])
batched_mask shape: torch.Size([1, 939])

100%|██████████| 1/1 [00:00<00:00, 157.98it/s]
Original shapes - Output: (939,), Label: (939,)
Aligned shapes - Output: (939,), Label: (939,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1847.11it/s]
Table1-only metrics completed in 0.20s
Evaluation Result:  {'AUC-PR': 0.7, 'AUC-ROC': 0.9983991462113128, 'VUS-PR': 0.625, 'VUS-ROC': 0.9975987193169688, 'Standard-F1': 0.33333333333299997, 'Standard-Precision': 0.19999999999998, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6662078255131576, 'Affiliation-P': 0.4994839812802713, 'Affiliation-R': 1.0, 'F1_T': 0.33333333333299997, 'Precision_T': 0.19999999999998, 'Recall_T': 0.9999999999995}
Results for 674_YAHOO_id_124_WebService_tr_500_1st_514.csv: {'filename': '674_YAHOO_id_124_WebService_tr_500_1st_514.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 23, 'train_index': '500', 'data_shape': '1439x1', 'output_length': 939, 'label_length': 939, 'aligned_length': 939, 'AUC-PR': 0.7, 'AUC-ROC': 0.9983991462113128, 'VUS-PR': 0.625, 'VUS-ROC': 0.9975987193169688, 'Standard-F1': 0.33333333333299997, 'Standard-Precision': 0.19999999999998, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6662078255131576, 'Affiliation-P': 0.4994839812802713, 'Affiliation-R': 1.0, 'F1_T': 0.33333333333299997, 'Precision_T': 0.19999999999998, 'Recall_T': 0.9999999999995}
Logits metrics skipped for 674_YAHOO_id_124_WebService_tr_500_1st_514.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/674_YAHOO_id_124_WebService_tr_500_1st_514_results.pkl
Processing file: 163_Stock_id_15_Finance_tr_500_1st_3.csv, Data shape: (10000, 1)
Label shape: (10000,)
Train data shape: (500, 1), Test data shape: (9500, 1), Label test shape: (9500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 163_Stock_id_15_Finance_tr_500_1st_3.csv...
Using CUDA device: 2
Data shape: (9500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 166.34it/s]
Original shapes - Output: (10000,), Label: (9500,)
Aligned shapes - Output: (9500,), Label: (9500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   7%|▋         | 14/200 [00:01<00:13, 13.85it/s]

Calculating Affiliation Metrics:  14%|█▍        | 28/200 [00:04<00:26,  6.39it/s]

Calculating Affiliation Metrics:  18%|█▊        | 37/200 [00:06<00:31,  5.21it/s]

Calculating Affiliation Metrics:  22%|██▏       | 44/200 [00:08<00:32,  4.81it/s]

Calculating Affiliation Metrics:  25%|██▌       | 50/200 [00:09<00:34,  4.40it/s]

Calculating Affiliation Metrics:  28%|██▊       | 55/200 [00:11<00:35,  4.03it/s]

Calculating Affiliation Metrics:  30%|███       | 60/200 [00:13<00:38,  3.68it/s]

Calculating Affiliation Metrics:  32%|███▏      | 64/200 [00:14<00:39,  3.44it/s]

Calculating Affiliation Metrics:  34%|███▍      | 68/200 [00:16<00:40,  3.22it/s]

Calculating Affiliation Metrics:  36%|███▌      | 72/200 [00:17<00:41,  3.07it/s]

Calculating Affiliation Metrics:  38%|███▊      | 76/200 [00:18<00:41,  2.97it/s]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [00:20<00:41,  2.91it/s]

Calculating Affiliation Metrics:  41%|████      | 82/200 [00:21<00:41,  2.84it/s]

Calculating Affiliation Metrics:  42%|████▎     | 85/200 [00:22<00:41,  2.79it/s]

Calculating Affiliation Metrics:  44%|████▍     | 88/200 [00:23<00:40,  2.77it/s]

Calculating Affiliation Metrics:  46%|████▌     | 91/200 [00:24<00:39,  2.75it/s]

Calculating Affiliation Metrics:  47%|████▋     | 94/200 [00:25<00:38,  2.74it/s]

Calculating Affiliation Metrics:  48%|████▊     | 97/200 [00:26<00:37,  2.75it/s]

Calculating Affiliation Metrics:  50%|█████     | 100/200 [00:27<00:36,  2.77it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 103/200 [00:28<00:34,  2.78it/s]

Calculating Affiliation Metrics:  53%|█████▎    | 106/200 [00:29<00:33,  2.80it/s]

Calculating Affiliation Metrics:  55%|█████▍    | 109/200 [00:30<00:32,  2.84it/s]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [00:32<00:29,  2.92it/s]

Calculating Affiliation Metrics:  58%|█████▊    | 117/200 [00:33<00:27,  3.02it/s]

Calculating Affiliation Metrics:  60%|██████    | 121/200 [00:34<00:24,  3.24it/s]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [00:35<00:20,  3.54it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 132/200 [00:36<00:16,  4.00it/s]

Calculating Affiliation Metrics:  70%|██████▉   | 139/200 [00:38<00:13,  4.64it/s]

Calculating Affiliation Metrics:  74%|███████▍  | 148/200 [00:39<00:09,  5.70it/s]

Calculating Affiliation Metrics:  80%|████████  | 161/200 [00:40<00:05,  7.65it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 176/200 [00:41<00:02,  9.62it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:42<00:00,  4.75it/s]
Table1-only metrics completed in 69.31s
Evaluation Result:  {'AUC-PR': 0.1574216638244147, 'AUC-ROC': 0.5363732841690734, 'VUS-PR': 0.9409669747523192, 'VUS-ROC': 0.9637284333508063, 'Standard-F1': 0.24827159219871042, 'Standard-Precision': 0.14389952153110044, 'Standard-Recall': 0.9038317054845975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6877686900058008, 'Affiliation-P': 0.5257200967431866, 'Affiliation-R': 0.994231934001671, 'F1_T': 0.2482401656312323, 'Precision_T': 0.14395485652539317, 'Recall_T': 0.9008264462809911}
Results for 163_Stock_id_15_Finance_tr_500_1st_3.csv: {'filename': '163_Stock_id_15_Finance_tr_500_1st_3.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '10000x1', 'output_length': 10000, 'label_length': 9500, 'aligned_length': 9500, 'AUC-PR': 0.1574216638244147, 'AUC-ROC': 0.5363732841690734, 'VUS-PR': 0.9409669747523192, 'VUS-ROC': 0.9637284333508063, 'Standard-F1': 0.24827159219871042, 'Standard-Precision': 0.14389952153110044, 'Standard-Recall': 0.9038317054845975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6877686900058008, 'Affiliation-P': 0.5257200967431866, 'Affiliation-R': 0.994231934001671, 'F1_T': 0.2482401656312323, 'Precision_T': 0.14395485652539317, 'Recall_T': 0.9008264462809911}
Logits metrics skipped for 163_Stock_id_15_Finance_tr_500_1st_3.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/163_Stock_id_15_Finance_tr_500_1st_3_results.pkl
Skipping file: 864_OPPORTUNITY_id_23_HumanActivity_tr_895_1st_995.csv due to filter criteria.
Processing file: 597_YAHOO_id_47_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 597_YAHOO_id_47_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 157.26it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1725.63it/s]
Table1-only metrics completed in 0.20s
Evaluation Result:  {'AUC-PR': 0.9013548180214848, 'AUC-ROC': 0.9989035087719298, 'VUS-PR': 0.9472392936188132, 'VUS-ROC': 0.9995067127965349, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Results for 597_YAHOO_id_47_Synthetic_tr_500_1st_623.csv: {'filename': '597_YAHOO_id_47_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 9, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9013548180214848, 'AUC-ROC': 0.9989035087719298, 'VUS-PR': 0.9472392936188132, 'VUS-ROC': 0.9995067127965349, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 597_YAHOO_id_47_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/597_YAHOO_id_47_Synthetic_tr_500_1st_623_results.pkl
Processing file: 736_YAHOO_id_186_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 736_YAHOO_id_186_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 163.46it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1917.02it/s]
Table1-only metrics completed in 0.35s
Evaluation Result:  {'AUC-PR': 0.9598765432098766, 'AUC-ROC': 0.9995126705653021, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984359113984375, 'Affiliation-P': 0.9969624510347107, 'Affiliation-R': 0.9999137336093858, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 736_YAHOO_id_186_Synthetic_tr_500_1st_623.csv: {'filename': '736_YAHOO_id_186_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 35, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9598765432098766, 'AUC-ROC': 0.9995126705653021, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984359113984375, 'Affiliation-P': 0.9969624510347107, 'Affiliation-R': 0.9999137336093858, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 736_YAHOO_id_186_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/736_YAHOO_id_186_Synthetic_tr_500_1st_623_results.pkl
Processing file: 276_IOPS_id_17_WebService_tr_19197_1st_19297.csv, Data shape: (91370, 1)
Label shape: (91370,)
Train data shape: (19197, 1), Test data shape: (72173, 1), Label test shape: (72173,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 276_IOPS_id_17_WebService_tr_19197_1st_19297.csv...
Using CUDA device: 2
Data shape: (72173, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 72173, window_size 5000, adding 2827 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([15, 5000, 1])
batched_mask shape: torch.Size([15, 5000])

100%|██████████| 1/1 [00:00<00:00, 80.40it/s]
Original shapes - Output: (75000,), Label: (72173,)
Aligned shapes - Output: (72173,), Label: (72173,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  18%|█▊        | 37/200 [00:01<00:04, 35.83it/s]

Calculating Affiliation Metrics:  36%|███▋      | 73/200 [00:02<00:04, 30.45it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 104/200 [00:03<00:03, 28.91it/s]

Calculating Affiliation Metrics:  70%|███████   | 140/200 [00:04<00:01, 31.34it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:05<00:00, 36.60it/s]
Table1-only metrics completed in 7.41s
Evaluation Result:  {'AUC-PR': 0.49733737714406195, 'AUC-ROC': 0.9683247590141184, 'VUS-PR': 0.4638202052812112, 'VUS-ROC': 0.9591842161196278, 'Standard-F1': 0.5461641391610533, 'Standard-Precision': 0.3831351689612014, 'Standard-Recall': 0.9506987577639748, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8645779789284823, 'Affiliation-P': 0.8344099758274747, 'Affiliation-R': 0.8970092521446207, 'F1_T': 0.5453533667219058, 'Precision_T': 0.3822749258854735, 'Recall_T': 0.9510869565217388}
Results for 276_IOPS_id_17_WebService_tr_19197_1st_19297.csv: {'filename': '276_IOPS_id_17_WebService_tr_19197_1st_19297.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '19197', 'data_shape': '91370x1', 'output_length': 75000, 'label_length': 72173, 'aligned_length': 72173, 'AUC-PR': 0.49733737714406195, 'AUC-ROC': 0.9683247590141184, 'VUS-PR': 0.4638202052812112, 'VUS-ROC': 0.9591842161196278, 'Standard-F1': 0.5461641391610533, 'Standard-Precision': 0.3831351689612014, 'Standard-Recall': 0.9506987577639748, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8645779789284823, 'Affiliation-P': 0.8344099758274747, 'Affiliation-R': 0.8970092521446207, 'F1_T': 0.5453533667219058, 'Precision_T': 0.3822749258854735, 'Recall_T': 0.9510869565217388}
Logits metrics skipped for 276_IOPS_id_17_WebService_tr_19197_1st_19297.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/276_IOPS_id_17_WebService_tr_19197_1st_19297_results.pkl
Processing file: 277_NEK_id_1_WebService_tr_500_1st_17.csv, Data shape: (1073, 1)
Label shape: (1073,)
Train data shape: (500, 1), Test data shape: (573, 1), Label test shape: (573,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 277_NEK_id_1_WebService_tr_500_1st_17.csv...
Using CUDA device: 2
Data shape: (573, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 573, 1])
batched_mask shape: torch.Size([1, 573])

100%|██████████| 1/1 [00:00<00:00, 157.70it/s]
Original shapes - Output: (573,), Label: (573,)
Aligned shapes - Output: (573,), Label: (573,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Table1-only metrics completed in 0.01s
Evaluation Result:  {'AUC-PR': 0.0, 'AUC-ROC': nan, 'VUS-PR': 0.0, 'VUS-ROC': 0.0, 'Standard-F1': 0.0, 'Standard-Precision': 0.0, 'Standard-Recall': 0.0, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.0, 'Affiliation-P': 0.0, 'Affiliation-R': 0.0, 'F1_T': 0.0, 'Precision_T': 0.0, 'Recall_T': 0.0}
Results for 277_NEK_id_1_WebService_tr_500_1st_17.csv: {'filename': '277_NEK_id_1_WebService_tr_500_1st_17.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '1073x1', 'output_length': 573, 'label_length': 573, 'aligned_length': 573, 'AUC-PR': 0.0, 'AUC-ROC': nan, 'VUS-PR': 0.0, 'VUS-ROC': 0.0, 'Standard-F1': 0.0, 'Standard-Precision': 0.0, 'Standard-Recall': 0.0, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.0, 'Affiliation-P': 0.0, 'Affiliation-R': 0.0, 'F1_T': 0.0, 'Precision_T': 0.0, 'Recall_T': 0.0}
Logits metrics skipped for 277_NEK_id_1_WebService_tr_500_1st_17.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/277_NEK_id_1_WebService_tr_500_1st_17_results.pkl
Skipping file: 210_SMD_id_33_Facility_tr_5925_1st_17580.csv due to filter criteria.
Skipping file: 198_SMD_id_21_Facility_tr_5925_1st_17580.csv due to filter criteria.
Processing file: 120_WSD_id_92_WebService_tr_2329_1st_2429.csv, Data shape: (17438, 1)
Label shape: (17438,)
Train data shape: (2329, 1), Test data shape: (15109, 1), Label test shape: (15109,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 120_WSD_id_92_WebService_tr_2329_1st_2429.csv...
Using CUDA device: 2
Data shape: (15109, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 15109, window_size 5000, adding 4891 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.40it/s]
Original shapes - Output: (20000,), Label: (15109,)
Aligned shapes - Output: (15109,), Label: (15109,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  90%|████████▉ | 179/200 [00:01<00:00, 178.72it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 189.97it/s]
Table1-only metrics completed in 3.83s
Evaluation Result:  {'AUC-PR': 0.19238047238650177, 'AUC-ROC': 0.8757720864035672, 'VUS-PR': 0.16859938569194982, 'VUS-ROC': 0.9646086374721758, 'Standard-F1': 0.15384615384567926, 'Standard-Precision': 0.12499999999999918, 'Standard-Recall': 0.1999999999999979, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9906962931162193, 'Affiliation-P': 0.9863231304690621, 'Affiliation-R': 0.9951084079547626, 'F1_T': 0.15384615384567926, 'Precision_T': 0.12499999999999918, 'Recall_T': 0.1999999999999979}
Results for 120_WSD_id_92_WebService_tr_2329_1st_2429.csv: {'filename': '120_WSD_id_92_WebService_tr_2329_1st_2429.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '2329', 'data_shape': '17438x1', 'output_length': 20000, 'label_length': 15109, 'aligned_length': 15109, 'AUC-PR': 0.19238047238650177, 'AUC-ROC': 0.8757720864035672, 'VUS-PR': 0.16859938569194982, 'VUS-ROC': 0.9646086374721758, 'Standard-F1': 0.15384615384567926, 'Standard-Precision': 0.12499999999999918, 'Standard-Recall': 0.1999999999999979, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9906962931162193, 'Affiliation-P': 0.9863231304690621, 'Affiliation-R': 0.9951084079547626, 'F1_T': 0.15384615384567926, 'Precision_T': 0.12499999999999918, 'Recall_T': 0.1999999999999979}
Logits metrics skipped for 120_WSD_id_92_WebService_tr_2329_1st_2429.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/120_WSD_id_92_WebService_tr_2329_1st_2429_results.pkl
Processing file: 428_UCR_id_126_Medical_tr_7487_1st_27862.csv, Data shape: (29950, 1)
Label shape: (29950,)
Train data shape: (7487, 1), Test data shape: (22463, 1), Label test shape: (22463,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 428_UCR_id_126_Medical_tr_7487_1st_27862.csv...
Using CUDA device: 2
Data shape: (22463, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22463, window_size 5000, adding 2537 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 149.36it/s]
Original shapes - Output: (25000,), Label: (22463,)
Aligned shapes - Output: (22463,), Label: (22463,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 103/200 [00:01<00:00, 102.71it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 116.79it/s]
Table1-only metrics completed in 1.96s
Evaluation Result:  {'AUC-PR': 0.8443897529682752, 'AUC-ROC': 0.9988950628704123, 'VUS-PR': 0.593147237518703, 'VUS-ROC': 0.9987211941200163, 'Standard-F1': 0.45423728813522823, 'Standard-Precision': 0.2977777777777765, 'Standard-Recall': 0.9571428571428435, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9114600749199813, 'Affiliation-P': 0.83736007964326, 'Affiliation-R': 0.9999478507513944, 'F1_T': 0.45423728813522823, 'Precision_T': 0.2977777777777765, 'Recall_T': 0.9571428571428435}
Results for 428_UCR_id_126_Medical_tr_7487_1st_27862.csv: {'filename': '428_UCR_id_126_Medical_tr_7487_1st_27862.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 33, 'train_index': '7487', 'data_shape': '29950x1', 'output_length': 25000, 'label_length': 22463, 'aligned_length': 22463, 'AUC-PR': 0.8443897529682752, 'AUC-ROC': 0.9988950628704123, 'VUS-PR': 0.593147237518703, 'VUS-ROC': 0.9987211941200163, 'Standard-F1': 0.45423728813522823, 'Standard-Precision': 0.2977777777777765, 'Standard-Recall': 0.9571428571428435, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9114600749199813, 'Affiliation-P': 0.83736007964326, 'Affiliation-R': 0.9999478507513944, 'F1_T': 0.45423728813522823, 'Precision_T': 0.2977777777777765, 'Recall_T': 0.9571428571428435}
Logits metrics skipped for 428_UCR_id_126_Medical_tr_7487_1st_27862.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/428_UCR_id_126_Medical_tr_7487_1st_27862_results.pkl
Processing file: 343_UCR_id_41_Facility_tr_14030_1st_52600.csv, Data shape: (56123, 1)
Label shape: (56123,)
Train data shape: (14030, 1), Test data shape: (42093, 1), Label test shape: (42093,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 343_UCR_id_41_Facility_tr_14030_1st_52600.csv...
Using CUDA device: 2
Data shape: (42093, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 42093, window_size 5000, adding 2907 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([9, 5000, 1])
batched_mask shape: torch.Size([9, 5000])

100%|██████████| 1/1 [00:00<00:00, 111.68it/s]
Original shapes - Output: (45000,), Label: (42093,)
Aligned shapes - Output: (42093,), Label: (42093,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  44%|████▍     | 88/200 [00:01<00:01, 87.51it/s]

Calculating Affiliation Metrics:  98%|█████████▊| 195/200 [00:02<00:00, 98.86it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 98.60it/s]
Table1-only metrics completed in 3.68s
Evaluation Result:  {'AUC-PR': 0.25624263981280665, 'AUC-ROC': 0.9518292435490416, 'VUS-PR': 0.14983497309797855, 'VUS-ROC': 0.9661785252294435, 'Standard-F1': 0.2415458937193693, 'Standard-Precision': 0.17814726840855064, 'Standard-Recall': 0.3749999999999981, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.821118669993552, 'Affiliation-P': 0.69683088587981, 'Affiliation-R': 0.9993672938493335, 'F1_T': 0.2415458937193693, 'Precision_T': 0.17814726840855064, 'Recall_T': 0.3749999999999981}
Results for 343_UCR_id_41_Facility_tr_14030_1st_52600.csv: {'filename': '343_UCR_id_41_Facility_tr_14030_1st_52600.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 233, 'train_index': '14030', 'data_shape': '56123x1', 'output_length': 45000, 'label_length': 42093, 'aligned_length': 42093, 'AUC-PR': 0.25624263981280665, 'AUC-ROC': 0.9518292435490416, 'VUS-PR': 0.14983497309797855, 'VUS-ROC': 0.9661785252294435, 'Standard-F1': 0.2415458937193693, 'Standard-Precision': 0.17814726840855064, 'Standard-Recall': 0.3749999999999981, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.821118669993552, 'Affiliation-P': 0.69683088587981, 'Affiliation-R': 0.9993672938493335, 'F1_T': 0.2415458937193693, 'Precision_T': 0.17814726840855064, 'Recall_T': 0.3749999999999981}
Logits metrics skipped for 343_UCR_id_41_Facility_tr_14030_1st_52600.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/343_UCR_id_41_Facility_tr_14030_1st_52600_results.pkl
Processing file: 408_UCR_id_106_Medical_tr_9594_1st_11111.csv, Data shape: (38379, 1)
Label shape: (38379,)
Train data shape: (9594, 1), Test data shape: (28785, 1), Label test shape: (28785,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 408_UCR_id_106_Medical_tr_9594_1st_11111.csv...
Using CUDA device: 2
Data shape: (28785, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 28785, window_size 5000, adding 1215 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 141.06it/s]
Original shapes - Output: (30000,), Label: (28785,)
Aligned shapes - Output: (28785,), Label: (28785,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  58%|█████▊    | 117/200 [00:01<00:00, 116.20it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 132.07it/s]
Table1-only metrics completed in 1.84s
Evaluation Result:  {'AUC-PR': 0.9949246361746362, 'AUC-ROC': 0.999968624716751, 'VUS-PR': 0.8506896551724137, 'VUS-ROC': 0.9996836325605712, 'Standard-F1': 0.5154639175253879, 'Standard-Precision': 0.347222222222221, 'Standard-Recall': 0.99999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9236091406856095, 'Affiliation-P': 0.8580613705653687, 'Affiliation-R': 0.9999996525968387, 'F1_T': 0.5154639175253879, 'Precision_T': 0.347222222222221, 'Recall_T': 0.99999999999999}
Results for 408_UCR_id_106_Medical_tr_9594_1st_11111.csv: {'filename': '408_UCR_id_106_Medical_tr_9594_1st_11111.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 40, 'train_index': '9594', 'data_shape': '38379x1', 'output_length': 30000, 'label_length': 28785, 'aligned_length': 28785, 'AUC-PR': 0.9949246361746362, 'AUC-ROC': 0.999968624716751, 'VUS-PR': 0.8506896551724137, 'VUS-ROC': 0.9996836325605712, 'Standard-F1': 0.5154639175253879, 'Standard-Precision': 0.347222222222221, 'Standard-Recall': 0.99999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9236091406856095, 'Affiliation-P': 0.8580613705653687, 'Affiliation-R': 0.9999996525968387, 'F1_T': 0.5154639175253879, 'Precision_T': 0.347222222222221, 'Recall_T': 0.99999999999999}
Logits metrics skipped for 408_UCR_id_106_Medical_tr_9594_1st_11111.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/408_UCR_id_106_Medical_tr_9594_1st_11111_results.pkl
Processing file: 693_YAHOO_id_143_WebService_tr_500_1st_1003.csv, Data shape: (1427, 1)
Label shape: (1427,)
Train data shape: (500, 1), Test data shape: (927, 1), Label test shape: (927,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 693_YAHOO_id_143_WebService_tr_500_1st_1003.csv...
Using CUDA device: 2
Data shape: (927, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 927, 1])
batched_mask shape: torch.Size([1, 927])

100%|██████████| 1/1 [00:00<00:00, 159.06it/s]
Original shapes - Output: (927,), Label: (927,)
Aligned shapes - Output: (927,), Label: (927,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2018.42it/s]
Table1-only metrics completed in 0.24s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.625, 'VUS-ROC': 0.9991891891891889, 'Standard-F1': 0.33333333333299997, 'Standard-Precision': 0.19999999999998, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9254037426300943, 'Affiliation-P': 0.8611641221374046, 'Affiliation-R': 1.0, 'F1_T': 0.33333333333299997, 'Precision_T': 0.19999999999998, 'Recall_T': 0.9999999999995}
Results for 693_YAHOO_id_143_WebService_tr_500_1st_1003.csv: {'filename': '693_YAHOO_id_143_WebService_tr_500_1st_1003.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1427x1', 'output_length': 927, 'label_length': 927, 'aligned_length': 927, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.625, 'VUS-ROC': 0.9991891891891889, 'Standard-F1': 0.33333333333299997, 'Standard-Precision': 0.19999999999998, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9254037426300943, 'Affiliation-P': 0.8611641221374046, 'Affiliation-R': 1.0, 'F1_T': 0.33333333333299997, 'Precision_T': 0.19999999999998, 'Recall_T': 0.9999999999995}
Logits metrics skipped for 693_YAHOO_id_143_WebService_tr_500_1st_1003.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/693_YAHOO_id_143_WebService_tr_500_1st_1003_results.pkl
Processing file: 061_WSD_id_33_WebService_tr_4559_1st_9722.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 061_WSD_id_33_WebService_tr_4559_1st_9722.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.36it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 305.66it/s]
Table1-only metrics completed in 1.54s
Evaluation Result:  {'AUC-PR': 0.009790800207445305, 'AUC-ROC': 0.8634107618279965, 'VUS-PR': 0.01149904181985504, 'VUS-ROC': 0.8946085093170049, 'Standard-F1': 0.023809523809230158, 'Standard-Precision': 0.014492753623188302, 'Standard-Recall': 0.06666666666666445, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7142268747084107, 'Affiliation-P': 0.5557487947526147, 'Affiliation-R': 0.9991441856018364, 'F1_T': 0.023809523809230158, 'Precision_T': 0.014492753623188302, 'Recall_T': 0.06666666666666445}
Results for 061_WSD_id_33_WebService_tr_4559_1st_9722.csv: {'filename': '061_WSD_id_33_WebService_tr_4559_1st_9722.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.009790800207445305, 'AUC-ROC': 0.8634107618279965, 'VUS-PR': 0.01149904181985504, 'VUS-ROC': 0.8946085093170049, 'Standard-F1': 0.023809523809230158, 'Standard-Precision': 0.014492753623188302, 'Standard-Recall': 0.06666666666666445, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7142268747084107, 'Affiliation-P': 0.5557487947526147, 'Affiliation-R': 0.9991441856018364, 'F1_T': 0.023809523809230158, 'Precision_T': 0.014492753623188302, 'Recall_T': 0.06666666666666445}
Logits metrics skipped for 061_WSD_id_33_WebService_tr_4559_1st_9722.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/061_WSD_id_33_WebService_tr_4559_1st_9722_results.pkl
Processing file: 033_WSD_id_5_WebService_tr_4559_1st_12588.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 033_WSD_id_5_WebService_tr_4559_1st_12588.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.01it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 322.33it/s]
Table1-only metrics completed in 1.20s
Evaluation Result:  {'AUC-PR': 0.2758258918897845, 'AUC-ROC': 0.9407628527514351, 'VUS-PR': 0.10568015019535658, 'VUS-ROC': 0.9571231322668646, 'Standard-F1': 0.0662251655627448, 'Standard-Precision': 0.03649635036496324, 'Standard-Recall': 0.3571428571428316, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7186049500757418, 'Affiliation-P': 0.5608728501143886, 'Affiliation-R': 0.9997649860558394, 'F1_T': 0.0662251655627448, 'Precision_T': 0.03649635036496324, 'Recall_T': 0.3571428571428316}
Results for 033_WSD_id_5_WebService_tr_4559_1st_12588.csv: {'filename': '033_WSD_id_5_WebService_tr_4559_1st_12588.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.2758258918897845, 'AUC-ROC': 0.9407628527514351, 'VUS-PR': 0.10568015019535658, 'VUS-ROC': 0.9571231322668646, 'Standard-F1': 0.0662251655627448, 'Standard-Precision': 0.03649635036496324, 'Standard-Recall': 0.3571428571428316, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7186049500757418, 'Affiliation-P': 0.5608728501143886, 'Affiliation-R': 0.9997649860558394, 'F1_T': 0.0662251655627448, 'Precision_T': 0.03649635036496324, 'Recall_T': 0.3571428571428316}
Logits metrics skipped for 033_WSD_id_5_WebService_tr_4559_1st_12588.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/033_WSD_id_5_WebService_tr_4559_1st_12588_results.pkl
Processing file: 590_YAHOO_id_40_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 590_YAHOO_id_40_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 158.37it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2010.34it/s]
Table1-only metrics completed in 1.14s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 590_YAHOO_id_40_Synthetic_tr_500_1st_658.csv: {'filename': '590_YAHOO_id_40_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 235, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 590_YAHOO_id_40_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/590_YAHOO_id_40_Synthetic_tr_500_1st_658_results.pkl
Processing file: 117_WSD_id_89_WebService_tr_3251_1st_3351.csv, Data shape: (16376, 1)
Label shape: (16376,)
Train data shape: (3251, 1), Test data shape: (13125, 1), Label test shape: (13125,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 117_WSD_id_89_WebService_tr_3251_1st_3351.csv...
Using CUDA device: 2
Data shape: (13125, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13125, window_size 5000, adding 1875 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.60it/s]
Original shapes - Output: (15000,), Label: (13125,)
Aligned shapes - Output: (13125,), Label: (13125,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 343.77it/s]
Table1-only metrics completed in 0.67s
Evaluation Result:  {'AUC-PR': 0.3363745576762166, 'AUC-ROC': 0.9626835340141239, 'VUS-PR': 0.3362607850633212, 'VUS-ROC': 0.9625819234872733, 'Standard-F1': 0.01481481481477114, 'Standard-Precision': 0.007575757575757519, 'Standard-Recall': 0.33333333333322224, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7073258157408719, 'Affiliation-P': 0.5471802750870771, 'Affiliation-R': 1.0, 'F1_T': 0.01481481481477114, 'Precision_T': 0.007575757575757519, 'Recall_T': 0.33333333333322224}
Results for 117_WSD_id_89_WebService_tr_3251_1st_3351.csv: {'filename': '117_WSD_id_89_WebService_tr_3251_1st_3351.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 7, 'train_index': '3251', 'data_shape': '16376x1', 'output_length': 15000, 'label_length': 13125, 'aligned_length': 13125, 'AUC-PR': 0.3363745576762166, 'AUC-ROC': 0.9626835340141239, 'VUS-PR': 0.3362607850633212, 'VUS-ROC': 0.9625819234872733, 'Standard-F1': 0.01481481481477114, 'Standard-Precision': 0.007575757575757519, 'Standard-Recall': 0.33333333333322224, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7073258157408719, 'Affiliation-P': 0.5471802750870771, 'Affiliation-R': 1.0, 'F1_T': 0.01481481481477114, 'Precision_T': 0.007575757575757519, 'Recall_T': 0.33333333333322224}
Logits metrics skipped for 117_WSD_id_89_WebService_tr_3251_1st_3351.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/117_WSD_id_89_WebService_tr_3251_1st_3351_results.pkl
Processing file: 344_UCR_id_42_Sensor_tr_2851_1st_5365.csv, Data shape: (11406, 1)
Label shape: (11406,)
Train data shape: (2851, 1), Test data shape: (8555, 1), Label test shape: (8555,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 344_UCR_id_42_Sensor_tr_2851_1st_5365.csv...
Using CUDA device: 2
Data shape: (8555, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 8555, window_size 5000, adding 1445 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.06it/s]
Original shapes - Output: (10000,), Label: (8555,)
Aligned shapes - Output: (8555,), Label: (8555,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 502.02it/s]
Table1-only metrics completed in 0.80s
Evaluation Result:  {'AUC-PR': 0.004805753986181762, 'AUC-ROC': 0.8196330991412959, 'VUS-PR': 0.009554441902362834, 'VUS-ROC': 0.8775167490678083, 'Standard-F1': 0.012531328320783368, 'Standard-Precision': 0.006325110689437061, 'Standard-Recall': 0.6666666666666222, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7478724724074326, 'Affiliation-P': 0.643298492653555, 'Affiliation-R': 0.8930450029222676, 'F1_T': 0.012422360248426698, 'Precision_T': 0.006276150627615059, 'Recall_T': 0.59999999999996}
Results for 344_UCR_id_42_Sensor_tr_2851_1st_5365.csv: {'filename': '344_UCR_id_42_Sensor_tr_2851_1st_5365.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 98, 'train_index': '2851', 'data_shape': '11406x1', 'output_length': 10000, 'label_length': 8555, 'aligned_length': 8555, 'AUC-PR': 0.004805753986181762, 'AUC-ROC': 0.8196330991412959, 'VUS-PR': 0.009554441902362834, 'VUS-ROC': 0.8775167490678083, 'Standard-F1': 0.012531328320783368, 'Standard-Precision': 0.006325110689437061, 'Standard-Recall': 0.6666666666666222, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7478724724074326, 'Affiliation-P': 0.643298492653555, 'Affiliation-R': 0.8930450029222676, 'F1_T': 0.012422360248426698, 'Precision_T': 0.006276150627615059, 'Recall_T': 0.59999999999996}
Logits metrics skipped for 344_UCR_id_42_Sensor_tr_2851_1st_5365.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/344_UCR_id_42_Sensor_tr_2851_1st_5365_results.pkl
Processing file: 788_YAHOO_id_238_WebService_tr_500_1st_973.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 788_YAHOO_id_238_WebService_tr_500_1st_973.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 163.91it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1607.45it/s]
Table1-only metrics completed in 0.22s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.987776109800557, 'Affiliation-P': 0.9758474576271187, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Results for 788_YAHOO_id_238_WebService_tr_500_1st_973.csv: {'filename': '788_YAHOO_id_238_WebService_tr_500_1st_973.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.987776109800557, 'Affiliation-P': 0.9758474576271187, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 788_YAHOO_id_238_WebService_tr_500_1st_973.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/788_YAHOO_id_238_WebService_tr_500_1st_973_results.pkl
Processing file: 005_NAB_id_5_Traffic_tr_594_1st_1645.csv, Data shape: (2379, 1)
Label shape: (2379,)
Train data shape: (594, 1), Test data shape: (1785, 1), Label test shape: (1785,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 005_NAB_id_5_Traffic_tr_594_1st_1645.csv...
Using CUDA device: 2
Data shape: (1785, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1785, 1])
batched_mask shape: torch.Size([1, 1785])

100%|██████████| 1/1 [00:00<00:00, 165.48it/s]
Original shapes - Output: (1785,), Label: (1785,)
Aligned shapes - Output: (1785,), Label: (1785,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1536.10it/s]
Table1-only metrics completed in 0.24s
Evaluation Result:  {'AUC-PR': 0.08594843019837205, 'AUC-ROC': 0.2693557060833383, 'VUS-PR': 0.08787242325186616, 'VUS-ROC': 0.27363987092539543, 'Standard-F1': 0.24211597151555497, 'Standard-Precision': 0.1377314814814814, 'Standard-Recall': 0.9999999999999958, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6781068580932894, 'Affiliation-P': 0.5129815993410652, 'Affiliation-R': 1.0, 'F1_T': 0.24162436548202082, 'Precision_T': 0.1374133949191685, 'Recall_T': 0.9999999999999958}
Results for 005_NAB_id_5_Traffic_tr_594_1st_1645.csv: {'filename': '005_NAB_id_5_Traffic_tr_594_1st_1645.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 22, 'train_index': '594', 'data_shape': '2379x1', 'output_length': 1785, 'label_length': 1785, 'aligned_length': 1785, 'AUC-PR': 0.08594843019837205, 'AUC-ROC': 0.2693557060833383, 'VUS-PR': 0.08787242325186616, 'VUS-ROC': 0.27363987092539543, 'Standard-F1': 0.24211597151555497, 'Standard-Precision': 0.1377314814814814, 'Standard-Recall': 0.9999999999999958, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6781068580932894, 'Affiliation-P': 0.5129815993410652, 'Affiliation-R': 1.0, 'F1_T': 0.24162436548202082, 'Precision_T': 0.1374133949191685, 'Recall_T': 0.9999999999999958}
Logits metrics skipped for 005_NAB_id_5_Traffic_tr_594_1st_1645.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/005_NAB_id_5_Traffic_tr_594_1st_1645_results.pkl
Processing file: 022_NAB_id_22_Facility_tr_1007_1st_2980.csv, Data shape: (4031, 1)
Label shape: (4031,)
Train data shape: (1007, 1), Test data shape: (3024, 1), Label test shape: (3024,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 022_NAB_id_22_Facility_tr_1007_1st_2980.csv...
Using CUDA device: 2
Data shape: (3024, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 3024, 1])
batched_mask shape: torch.Size([1, 3024])

100%|██████████| 1/1 [00:00<00:00, 165.84it/s]
Original shapes - Output: (3024,), Label: (3024,)
Aligned shapes - Output: (3024,), Label: (3024,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 856.58it/s]
Table1-only metrics completed in 1.00s
Evaluation Result:  {'AUC-PR': 0.22677808994726392, 'AUC-ROC': 0.48893769968051115, 'VUS-PR': 0.3180411194506214, 'VUS-ROC': 0.6029101494366738, 'Standard-F1': 0.3720930232553226, 'Standard-Precision': 0.3274853801169586, 'Standard-Recall': 0.4307692307692299, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8808487307322049, 'Affiliation-P': 0.8276509369241876, 'Affiliation-R': 0.9413549209001459, 'F1_T': 0.3718263718258822, 'Precision_T': 0.3238231098430808, 'Recall_T': 0.43653846153846065}
Results for 022_NAB_id_22_Facility_tr_1007_1st_2980.csv: {'filename': '022_NAB_id_22_Facility_tr_1007_1st_2980.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '1007', 'data_shape': '4031x1', 'output_length': 3024, 'label_length': 3024, 'aligned_length': 3024, 'AUC-PR': 0.22677808994726392, 'AUC-ROC': 0.48893769968051115, 'VUS-PR': 0.3180411194506214, 'VUS-ROC': 0.6029101494366738, 'Standard-F1': 0.3720930232553226, 'Standard-Precision': 0.3274853801169586, 'Standard-Recall': 0.4307692307692299, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8808487307322049, 'Affiliation-P': 0.8276509369241876, 'Affiliation-R': 0.9413549209001459, 'F1_T': 0.3718263718258822, 'Precision_T': 0.3238231098430808, 'Recall_T': 0.43653846153846065}
Logits metrics skipped for 022_NAB_id_22_Facility_tr_1007_1st_2980.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/022_NAB_id_22_Facility_tr_1007_1st_2980_results.pkl
Processing file: 281_NEK_id_5_WebService_tr_500_1st_974.csv, Data shape: (1073, 1)
Label shape: (1073,)
Train data shape: (500, 1), Test data shape: (573, 1), Label test shape: (573,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 281_NEK_id_5_WebService_tr_500_1st_974.csv...
Using CUDA device: 2
Data shape: (573, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 573, 1])
batched_mask shape: torch.Size([1, 573])

100%|██████████| 1/1 [00:00<00:00, 164.45it/s]
Original shapes - Output: (573,), Label: (573,)
Aligned shapes - Output: (573,), Label: (573,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 3582.85it/s]
Table1-only metrics completed in 0.15s
Evaluation Result:  {'AUC-PR': 0.9644154340658254, 'AUC-ROC': 0.9831302717900656, 'VUS-PR': 0.9659888217698578, 'VUS-ROC': 0.9859780632764127, 'Standard-F1': 0.9585798816562943, 'Standard-Precision': 0.9999999999999877, 'Standard-Recall': 0.9204545454545351, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9997520407049973, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9995042043471364, 'F1_T': 0.9529411764700777, 'Precision_T': 0.9878048780487685, 'Recall_T': 0.9204545454545351}
Results for 281_NEK_id_5_WebService_tr_500_1st_974.csv: {'filename': '281_NEK_id_5_WebService_tr_500_1st_974.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1073x1', 'output_length': 573, 'label_length': 573, 'aligned_length': 573, 'AUC-PR': 0.9644154340658254, 'AUC-ROC': 0.9831302717900656, 'VUS-PR': 0.9659888217698578, 'VUS-ROC': 0.9859780632764127, 'Standard-F1': 0.9585798816562943, 'Standard-Precision': 0.9999999999999877, 'Standard-Recall': 0.9204545454545351, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9997520407049973, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9995042043471364, 'F1_T': 0.9529411764700777, 'Precision_T': 0.9878048780487685, 'Recall_T': 0.9204545454545351}
Logits metrics skipped for 281_NEK_id_5_WebService_tr_500_1st_974.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/281_NEK_id_5_WebService_tr_500_1st_974_results.pkl
Processing file: 049_WSD_id_21_WebService_tr_4297_1st_7097.csv, Data shape: (17189, 1)
Label shape: (17189,)
Train data shape: (4297, 1), Test data shape: (12892, 1), Label test shape: (12892,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 049_WSD_id_21_WebService_tr_4297_1st_7097.csv...
Using CUDA device: 2
Data shape: (12892, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 12892, window_size 5000, adding 2108 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.75it/s]
Original shapes - Output: (15000,), Label: (12892,)
Aligned shapes - Output: (12892,), Label: (12892,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 296.34it/s]
Table1-only metrics completed in 0.79s
Evaluation Result:  {'AUC-PR': 0.8369617672808398, 'AUC-ROC': 0.9776493575224925, 'VUS-PR': 0.4413415861307373, 'VUS-ROC': 0.9829540145996959, 'Standard-F1': 0.3249999999996835, 'Standard-Precision': 0.20155038759689767, 'Standard-Recall': 0.8387096774193278, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8567297986742229, 'Affiliation-P': 0.7494248819049957, 'Affiliation-R': 0.9998982480791835, 'F1_T': 0.3249999999996835, 'Precision_T': 0.20155038759689767, 'Recall_T': 0.8387096774193278}
Results for 049_WSD_id_21_WebService_tr_4297_1st_7097.csv: {'filename': '049_WSD_id_21_WebService_tr_4297_1st_7097.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 6, 'train_index': '4297', 'data_shape': '17189x1', 'output_length': 15000, 'label_length': 12892, 'aligned_length': 12892, 'AUC-PR': 0.8369617672808398, 'AUC-ROC': 0.9776493575224925, 'VUS-PR': 0.4413415861307373, 'VUS-ROC': 0.9829540145996959, 'Standard-F1': 0.3249999999996835, 'Standard-Precision': 0.20155038759689767, 'Standard-Recall': 0.8387096774193278, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8567297986742229, 'Affiliation-P': 0.7494248819049957, 'Affiliation-R': 0.9998982480791835, 'F1_T': 0.3249999999996835, 'Precision_T': 0.20155038759689767, 'Recall_T': 0.8387096774193278}
Logits metrics skipped for 049_WSD_id_21_WebService_tr_4297_1st_7097.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/049_WSD_id_21_WebService_tr_4297_1st_7097_results.pkl
Processing file: 684_YAHOO_id_134_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 684_YAHOO_id_134_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 162.57it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1900.24it/s]
Table1-only metrics completed in 0.72s
Evaluation Result:  {'AUC-PR': 0.9253086419753087, 'AUC-ROC': 0.9991471734892788, 'VUS-PR': 0.9930845166427625, 'VUS-ROC': 0.9999361563702144, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 684_YAHOO_id_134_Synthetic_tr_500_1st_623.csv: {'filename': '684_YAHOO_id_134_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 103, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9253086419753087, 'AUC-ROC': 0.9991471734892788, 'VUS-PR': 0.9930845166427625, 'VUS-ROC': 0.9999361563702144, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 684_YAHOO_id_134_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/684_YAHOO_id_134_Synthetic_tr_500_1st_623_results.pkl
Skipping file: 868_OPPORTUNITY_id_27_HumanActivity_tr_2085_1st_2185.csv due to filter criteria.
Processing file: 560_YAHOO_id_10_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 560_YAHOO_id_10_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 162.79it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1952.18it/s]
Table1-only metrics completed in 0.63s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984775989560672, 'Affiliation-P': 0.9969598262757872, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 560_YAHOO_id_10_Synthetic_tr_500_1st_893.csv: {'filename': '560_YAHOO_id_10_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 158, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984775989560672, 'Affiliation-P': 0.9969598262757872, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 560_YAHOO_id_10_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/560_YAHOO_id_10_Synthetic_tr_500_1st_893_results.pkl
Processing file: 053_WSD_id_25_WebService_tr_4559_1st_9198.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 053_WSD_id_25_WebService_tr_4559_1st_9198.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.20it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 315.02it/s]
Table1-only metrics completed in 0.74s
Evaluation Result:  {'AUC-PR': 0.10454709466230863, 'AUC-ROC': 0.9442949468564222, 'VUS-PR': 0.11850707829763486, 'VUS-ROC': 0.9619246538854924, 'Standard-F1': 0.09039548022585157, 'Standard-Precision': 0.04878048780487775, 'Standard-Recall': 0.615384615384568, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9275317986325697, 'Affiliation-P': 0.8651602764823628, 'Affiliation-R': 0.9995950528962153, 'F1_T': 0.08256880733933664, 'Precision_T': 0.04390243902439003, 'Recall_T': 0.692307692307639}
Results for 053_WSD_id_25_WebService_tr_4559_1st_9198.csv: {'filename': '053_WSD_id_25_WebService_tr_4559_1st_9198.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 10, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.10454709466230863, 'AUC-ROC': 0.9442949468564222, 'VUS-PR': 0.11850707829763486, 'VUS-ROC': 0.9619246538854924, 'Standard-F1': 0.09039548022585157, 'Standard-Precision': 0.04878048780487775, 'Standard-Recall': 0.615384615384568, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9275317986325697, 'Affiliation-P': 0.8651602764823628, 'Affiliation-R': 0.9995950528962153, 'F1_T': 0.08256880733933664, 'Precision_T': 0.04390243902439003, 'Recall_T': 0.692307692307639}
Logits metrics skipped for 053_WSD_id_25_WebService_tr_4559_1st_9198.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/053_WSD_id_25_WebService_tr_4559_1st_9198_results.pkl
Processing file: 762_YAHOO_id_212_WebService_tr_500_1st_1055.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 762_YAHOO_id_212_WebService_tr_500_1st_1055.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 166.58it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1555.07it/s]
Table1-only metrics completed in 0.28s
Evaluation Result:  {'AUC-PR': 0.5169491525423728, 'AUC-ROC': 0.9758064516129032, 'VUS-PR': 0.30773338988595417, 'VUS-ROC': 0.9704311098639458, 'Standard-F1': 0.14285714285687756, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.49999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8325383658233093, 'Affiliation-P': 0.7749999999999999, 'Affiliation-R': 0.8993055555555556, 'F1_T': 0.14285714285687756, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.49999999999975}
Results for 762_YAHOO_id_212_WebService_tr_500_1st_1055.csv: {'filename': '762_YAHOO_id_212_WebService_tr_500_1st_1055.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.5169491525423728, 'AUC-ROC': 0.9758064516129032, 'VUS-PR': 0.30773338988595417, 'VUS-ROC': 0.9704311098639458, 'Standard-F1': 0.14285714285687756, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.49999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8325383658233093, 'Affiliation-P': 0.7749999999999999, 'Affiliation-R': 0.8993055555555556, 'F1_T': 0.14285714285687756, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.49999999999975}
Logits metrics skipped for 762_YAHOO_id_212_WebService_tr_500_1st_1055.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/762_YAHOO_id_212_WebService_tr_500_1st_1055_results.pkl
Skipping file: 816_Exathlon_id_7_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 552_YAHOO_id_2_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 552_YAHOO_id_2_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 165.50it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2119.33it/s]
Table1-only metrics completed in 0.43s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7958884656132357, 'Affiliation-P': 0.6609757010745441, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 552_YAHOO_id_2_Synthetic_tr_500_1st_893.csv: {'filename': '552_YAHOO_id_2_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7958884656132357, 'Affiliation-P': 0.6609757010745441, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 552_YAHOO_id_2_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/552_YAHOO_id_2_Synthetic_tr_500_1st_893_results.pkl
Processing file: 468_UCR_id_166_Facility_tr_10000_1st_35774.csv, Data shape: (40000, 1)
Label shape: (40000,)
Train data shape: (10000, 1), Test data shape: (30000, 1), Label test shape: (30000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 468_UCR_id_166_Facility_tr_10000_1st_35774.csv...
Using CUDA device: 2
Data shape: (30000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 142.91it/s]
Original shapes - Output: (30000,), Label: (30000,)
Aligned shapes - Output: (30000,), Label: (30000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 123/200 [00:01<00:00, 122.62it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 135.41it/s]
Table1-only metrics completed in 2.19s
Evaluation Result:  {'AUC-PR': 0.18111635099269585, 'AUC-ROC': 0.7508501672240803, 'VUS-PR': 0.15412211312954632, 'VUS-ROC': 0.8445826146924946, 'Standard-F1': 0.2199999999996239, 'Standard-Precision': 0.14666666666666617, 'Standard-Recall': 0.4399999999999956, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7256903780680495, 'Affiliation-P': 0.5696681476569407, 'Affiliation-R': 0.9994120000000001, 'F1_T': 0.2199999999996239, 'Precision_T': 0.14666666666666617, 'Recall_T': 0.4399999999999956}
Results for 468_UCR_id_166_Facility_tr_10000_1st_35774.csv: {'filename': '468_UCR_id_166_Facility_tr_10000_1st_35774.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 108, 'train_index': '10000', 'data_shape': '40000x1', 'output_length': 30000, 'label_length': 30000, 'aligned_length': 30000, 'AUC-PR': 0.18111635099269585, 'AUC-ROC': 0.7508501672240803, 'VUS-PR': 0.15412211312954632, 'VUS-ROC': 0.8445826146924946, 'Standard-F1': 0.2199999999996239, 'Standard-Precision': 0.14666666666666617, 'Standard-Recall': 0.4399999999999956, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7256903780680495, 'Affiliation-P': 0.5696681476569407, 'Affiliation-R': 0.9994120000000001, 'F1_T': 0.2199999999996239, 'Precision_T': 0.14666666666666617, 'Recall_T': 0.4399999999999956}
Logits metrics skipped for 468_UCR_id_166_Facility_tr_10000_1st_35774.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/468_UCR_id_166_Facility_tr_10000_1st_35774_results.pkl
Skipping file: 836_Exathlon_id_27_Facility_tr_10766_1st_12590.csv due to filter criteria.
Skipping file: 856_OPPORTUNITY_id_15_HumanActivity_tr_808_1st_908.csv due to filter criteria.
Processing file: 086_WSD_id_58_WebService_tr_4436_1st_9529.csv, Data shape: (17746, 1)
Label shape: (17746,)
Train data shape: (4436, 1), Test data shape: (13310, 1), Label test shape: (13310,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 086_WSD_id_58_WebService_tr_4436_1st_9529.csv...
Using CUDA device: 2
Data shape: (13310, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13310, window_size 5000, adding 1690 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.32it/s]
Original shapes - Output: (15000,), Label: (13310,)
Aligned shapes - Output: (13310,), Label: (13310,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 248.84it/s]
Table1-only metrics completed in 0.97s
Evaluation Result:  {'AUC-PR': 0.6906970883601364, 'AUC-ROC': 0.9965805085716343, 'VUS-PR': 0.446511563206249, 'VUS-ROC': 0.9983161216420805, 'Standard-F1': 0.2675159235666255, 'Standard-Precision': 0.1567164179104466, 'Standard-Recall': 0.91304347826083, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8900277447229898, 'Affiliation-P': 0.8019796468069431, 'Affiliation-R': 0.9997935200793397, 'F1_T': 0.2675159235666255, 'Precision_T': 0.1567164179104466, 'Recall_T': 0.91304347826083}
Results for 086_WSD_id_58_WebService_tr_4436_1st_9529.csv: {'filename': '086_WSD_id_58_WebService_tr_4436_1st_9529.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 11, 'train_index': '4436', 'data_shape': '17746x1', 'output_length': 15000, 'label_length': 13310, 'aligned_length': 13310, 'AUC-PR': 0.6906970883601364, 'AUC-ROC': 0.9965805085716343, 'VUS-PR': 0.446511563206249, 'VUS-ROC': 0.9983161216420805, 'Standard-F1': 0.2675159235666255, 'Standard-Precision': 0.1567164179104466, 'Standard-Recall': 0.91304347826083, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8900277447229898, 'Affiliation-P': 0.8019796468069431, 'Affiliation-R': 0.9997935200793397, 'F1_T': 0.2675159235666255, 'Precision_T': 0.1567164179104466, 'Recall_T': 0.91304347826083}
Logits metrics skipped for 086_WSD_id_58_WebService_tr_4436_1st_9529.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/086_WSD_id_58_WebService_tr_4436_1st_9529_results.pkl
Processing file: 426_UCR_id_124_Sensor_tr_2827_1st_4711.csv, Data shape: (11308, 1)
Label shape: (11308,)
Train data shape: (2827, 1), Test data shape: (8481, 1), Label test shape: (8481,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 426_UCR_id_124_Sensor_tr_2827_1st_4711.csv...
Using CUDA device: 2
Data shape: (8481, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 8481, window_size 5000, adding 1519 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 164.22it/s]
Original shapes - Output: (10000,), Label: (8481,)
Aligned shapes - Output: (8481,), Label: (8481,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 488.73it/s]
Table1-only metrics completed in 0.82s
Evaluation Result:  {'AUC-PR': 0.9711240214469661, 'AUC-ROC': 0.999476588917805, 'VUS-PR': 0.9399543236749442, 'VUS-ROC': 0.9993151755447267, 'Standard-F1': 0.9289617486333722, 'Standard-Precision': 0.9999999999999883, 'Standard-Recall': 0.8673469387755014, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999518708984976, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9999037464295941, 'F1_T': 0.9289617486333722, 'Precision_T': 0.9999999999999883, 'Recall_T': 0.8673469387755014}
Results for 426_UCR_id_124_Sensor_tr_2827_1st_4711.csv: {'filename': '426_UCR_id_124_Sensor_tr_2827_1st_4711.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 98, 'train_index': '2827', 'data_shape': '11308x1', 'output_length': 10000, 'label_length': 8481, 'aligned_length': 8481, 'AUC-PR': 0.9711240214469661, 'AUC-ROC': 0.999476588917805, 'VUS-PR': 0.9399543236749442, 'VUS-ROC': 0.9993151755447267, 'Standard-F1': 0.9289617486333722, 'Standard-Precision': 0.9999999999999883, 'Standard-Recall': 0.8673469387755014, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9999518708984976, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9999037464295941, 'F1_T': 0.9289617486333722, 'Precision_T': 0.9999999999999883, 'Recall_T': 0.8673469387755014}
Logits metrics skipped for 426_UCR_id_124_Sensor_tr_2827_1st_4711.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/426_UCR_id_124_Sensor_tr_2827_1st_4711_results.pkl
Processing file: 627_YAHOO_id_77_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 627_YAHOO_id_77_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 164.57it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1801.10it/s]
Table1-only metrics completed in 0.78s
Evaluation Result:  {'AUC-PR': 0.9172278338945006, 'AUC-ROC': 0.9990253411306043, 'VUS-PR': 0.9937762152964235, 'VUS-ROC': 0.9999426047313061, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Results for 627_YAHOO_id_77_Synthetic_tr_500_1st_623.csv: {'filename': '627_YAHOO_id_77_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 115, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9172278338945006, 'AUC-ROC': 0.9990253411306043, 'VUS-PR': 0.9937762152964235, 'VUS-ROC': 0.9999426047313061, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 627_YAHOO_id_77_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/627_YAHOO_id_77_Synthetic_tr_500_1st_623_results.pkl
Processing file: 633_YAHOO_id_83_WebService_tr_500_1st_223.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 633_YAHOO_id_83_WebService_tr_500_1st_223.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.65it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1411.23it/s]
Table1-only metrics completed in 0.30s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.8493019703517117, 'VUS-ROC': 0.999680059109235, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9956427015250539, 'Affiliation-P': 0.9913232104121474, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Results for 633_YAHOO_id_83_WebService_tr_500_1st_223.csv: {'filename': '633_YAHOO_id_83_WebService_tr_500_1st_223.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.8493019703517117, 'VUS-ROC': 0.999680059109235, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9956427015250539, 'Affiliation-P': 0.9913232104121474, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Logits metrics skipped for 633_YAHOO_id_83_WebService_tr_500_1st_223.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/633_YAHOO_id_83_WebService_tr_500_1st_223_results.pkl
Processing file: 496_UCR_id_194_Environment_tr_2046_1st_5391.csv, Data shape: (8184, 1)
Label shape: (8184,)
Train data shape: (2046, 1), Test data shape: (6138, 1), Label test shape: (6138,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 496_UCR_id_194_Environment_tr_2046_1st_5391.csv...
Using CUDA device: 2
Data shape: (6138, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6138, window_size 5000, adding 3862 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.57it/s]
Original shapes - Output: (10000,), Label: (6138,)
Aligned shapes - Output: (6138,), Label: (6138,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 486.40it/s]
Table1-only metrics completed in 0.53s
Evaluation Result:  {'AUC-PR': 0.006756756756756757, 'AUC-ROC': 0.9760469284666775, 'VUS-PR': 0.02812441513696605, 'VUS-ROC': 0.9911263807008551, 'Standard-F1': 0.012578616352188597, 'Standard-Precision': 0.006329113924050593, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7535695149934774, 'Affiliation-P': 0.6046416296156151, 'Affiliation-R': 0.9998370804822417, 'F1_T': 0.012987012986999915, 'Precision_T': 0.006535947712418258, 'Recall_T': 0.9999999999989999}
Results for 496_UCR_id_194_Environment_tr_2046_1st_5391.csv: {'filename': '496_UCR_id_194_Environment_tr_2046_1st_5391.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '2046', 'data_shape': '8184x1', 'output_length': 10000, 'label_length': 6138, 'aligned_length': 6138, 'AUC-PR': 0.006756756756756757, 'AUC-ROC': 0.9760469284666775, 'VUS-PR': 0.02812441513696605, 'VUS-ROC': 0.9911263807008551, 'Standard-F1': 0.012578616352188597, 'Standard-Precision': 0.006329113924050593, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7535695149934774, 'Affiliation-P': 0.6046416296156151, 'Affiliation-R': 0.9998370804822417, 'F1_T': 0.012987012986999915, 'Precision_T': 0.006535947712418258, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 496_UCR_id_194_Environment_tr_2046_1st_5391.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/496_UCR_id_194_Environment_tr_2046_1st_5391_results.pkl
Processing file: 383_UCR_id_81_Medical_tr_1875_1st_3198.csv, Data shape: (7500, 1)
Label shape: (7500,)
Train data shape: (1875, 1), Test data shape: (5625, 1), Label test shape: (5625,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 383_UCR_id_81_Medical_tr_1875_1st_3198.csv...
Using CUDA device: 2
Data shape: (5625, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5625, window_size 5000, adding 4375 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.79it/s]
Original shapes - Output: (10000,), Label: (5625,)
Aligned shapes - Output: (5625,), Label: (5625,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 652.21it/s]
Table1-only metrics completed in 1.00s
Evaluation Result:  {'AUC-PR': 0.9999195624195624, 'AUC-ROC': 0.9999983661572345, 'VUS-PR': 0.9998906878745715, 'VUS-ROC': 0.9999988798092911, 'Standard-F1': 0.9955156950667557, 'Standard-Precision': 0.9910714285714197, 'Standard-Recall': 0.999999999999991, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9998730569605826, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9997461461461461, 'F1_T': 0.9955156950667557, 'Precision_T': 0.9910714285714197, 'Recall_T': 0.999999999999991}
Results for 383_UCR_id_81_Medical_tr_1875_1st_3198.csv: {'filename': '383_UCR_id_81_Medical_tr_1875_1st_3198.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 183, 'train_index': '1875', 'data_shape': '7500x1', 'output_length': 10000, 'label_length': 5625, 'aligned_length': 5625, 'AUC-PR': 0.9999195624195624, 'AUC-ROC': 0.9999983661572345, 'VUS-PR': 0.9998906878745715, 'VUS-ROC': 0.9999988798092911, 'Standard-F1': 0.9955156950667557, 'Standard-Precision': 0.9910714285714197, 'Standard-Recall': 0.999999999999991, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9998730569605826, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9997461461461461, 'F1_T': 0.9955156950667557, 'Precision_T': 0.9910714285714197, 'Recall_T': 0.999999999999991}
Logits metrics skipped for 383_UCR_id_81_Medical_tr_1875_1st_3198.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/383_UCR_id_81_Medical_tr_1875_1st_3198_results.pkl
Processing file: 639_YAHOO_id_89_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 639_YAHOO_id_89_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 161.79it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2294.46it/s]
Table1-only metrics completed in 0.42s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9628378378378373, 'Affiliation-P': 0.9283387622149837, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 639_YAHOO_id_89_Synthetic_tr_500_1st_893.csv: {'filename': '639_YAHOO_id_89_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 103, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9628378378378373, 'Affiliation-P': 0.9283387622149837, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 639_YAHOO_id_89_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/639_YAHOO_id_89_Synthetic_tr_500_1st_893_results.pkl
Skipping file: 195_SMD_id_18_Facility_tr_5925_1st_17580.csv due to filter criteria.
Processing file: 698_YAHOO_id_148_WebService_tr_500_1st_186.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 698_YAHOO_id_148_WebService_tr_500_1st_186.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 167.20it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 993.01it/s]
Table1-only metrics completed in 0.47s
Evaluation Result:  {'AUC-PR': 0.8957382039573822, 'AUC-ROC': 0.9870006641996395, 'VUS-PR': 0.8236840198276193, 'VUS-ROC': 0.9861123988376376, 'Standard-F1': 0.7619047619041995, 'Standard-Precision': 0.6666666666666111, 'Standard-Recall': 0.8888888888887901, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9243745536631587, 'Affiliation-P': 0.9628113026819923, 'Affiliation-R': 0.8888888888888888, 'F1_T': 0.7619047619041995, 'Precision_T': 0.6666666666666111, 'Recall_T': 0.8888888888887901}
Results for 698_YAHOO_id_148_WebService_tr_500_1st_186.csv: {'filename': '698_YAHOO_id_148_WebService_tr_500_1st_186.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 12, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.8957382039573822, 'AUC-ROC': 0.9870006641996395, 'VUS-PR': 0.8236840198276193, 'VUS-ROC': 0.9861123988376376, 'Standard-F1': 0.7619047619041995, 'Standard-Precision': 0.6666666666666111, 'Standard-Recall': 0.8888888888887901, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9243745536631587, 'Affiliation-P': 0.9628113026819923, 'Affiliation-R': 0.8888888888888888, 'F1_T': 0.7619047619041995, 'Precision_T': 0.6666666666666111, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 698_YAHOO_id_148_WebService_tr_500_1st_186.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/698_YAHOO_id_148_WebService_tr_500_1st_186_results.pkl
Processing file: 610_YAHOO_id_60_WebService_tr_500_1st_878.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 610_YAHOO_id_60_WebService_tr_500_1st_878.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.37it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1601.22it/s]
Table1-only metrics completed in 0.23s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9539007092198575, 'Affiliation-P': 0.911864406779661, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Results for 610_YAHOO_id_60_WebService_tr_500_1st_878.csv: {'filename': '610_YAHOO_id_60_WebService_tr_500_1st_878.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9539007092198575, 'Affiliation-P': 0.911864406779661, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 610_YAHOO_id_60_WebService_tr_500_1st_878.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/610_YAHOO_id_60_WebService_tr_500_1st_878_results.pkl
Processing file: 231_MGAB_id_7_Synthetic_tr_25000_1st_35544.csv, Data shape: (100000, 1)
Label shape: (100000,)
Train data shape: (25000, 1), Test data shape: (75000, 1), Label test shape: (75000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 231_MGAB_id_7_Synthetic_tr_25000_1st_35544.csv...
Using CUDA device: 2
Data shape: (75000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([15, 5000, 1])
batched_mask shape: torch.Size([15, 5000])

100%|██████████| 1/1 [00:00<00:00, 80.66it/s]
Original shapes - Output: (75000,), Label: (75000,)
Aligned shapes - Output: (75000,), Label: (75000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  19%|█▉        | 38/200 [00:01<00:04, 36.85it/s]

Calculating Affiliation Metrics:  38%|███▊      | 75/200 [00:02<00:04, 27.85it/s]

Calculating Affiliation Metrics:  52%|█████▎    | 105/200 [00:04<00:03, 24.75it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 131/200 [00:05<00:02, 23.82it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 156/200 [00:06<00:01, 23.90it/s]

Calculating Affiliation Metrics:  96%|█████████▌| 192/200 [00:07<00:00, 27.48it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:07<00:00, 27.39it/s]
Table1-only metrics completed in 9.23s
Evaluation Result:  {'AUC-PR': 0.011094005642740722, 'AUC-ROC': 0.6888332553475935, 'VUS-PR': 0.013838187195074127, 'VUS-ROC': 0.7182506357937917, 'Standard-F1': 0.06315789473650955, 'Standard-Precision': 0.039999999999999945, 'Standard-Recall': 0.14999999999999925, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7536041537756796, 'Affiliation-P': 0.6315807719602089, 'Affiliation-R': 0.934069257751063, 'F1_T': 0.06315789473650955, 'Precision_T': 0.039999999999999945, 'Recall_T': 0.14999999999999925}
Results for 231_MGAB_id_7_Synthetic_tr_25000_1st_35544.csv: {'filename': '231_MGAB_id_7_Synthetic_tr_25000_1st_35544.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 49, 'train_index': '25000', 'data_shape': '100000x1', 'output_length': 75000, 'label_length': 75000, 'aligned_length': 75000, 'AUC-PR': 0.011094005642740722, 'AUC-ROC': 0.6888332553475935, 'VUS-PR': 0.013838187195074127, 'VUS-ROC': 0.7182506357937917, 'Standard-F1': 0.06315789473650955, 'Standard-Precision': 0.039999999999999945, 'Standard-Recall': 0.14999999999999925, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7536041537756796, 'Affiliation-P': 0.6315807719602089, 'Affiliation-R': 0.934069257751063, 'F1_T': 0.06315789473650955, 'Precision_T': 0.039999999999999945, 'Recall_T': 0.14999999999999925}
Logits metrics skipped for 231_MGAB_id_7_Synthetic_tr_25000_1st_35544.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/231_MGAB_id_7_Synthetic_tr_25000_1st_35544_results.pkl
Processing file: 004_NAB_id_4_Facility_tr_1007_1st_1437.csv, Data shape: (4031, 1)
Label shape: (4031,)
Train data shape: (1007, 1), Test data shape: (3024, 1), Label test shape: (3024,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 004_NAB_id_4_Facility_tr_1007_1st_1437.csv...
Using CUDA device: 2
Data shape: (3024, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 3024, 1])
batched_mask shape: torch.Size([1, 3024])

100%|██████████| 1/1 [00:00<00:00, 159.14it/s]
Original shapes - Output: (3024,), Label: (3024,)
Aligned shapes - Output: (3024,), Label: (3024,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 716.41it/s]
Table1-only metrics completed in 0.49s
Evaluation Result:  {'AUC-PR': 0.3471525589227749, 'AUC-ROC': 0.8184065371085077, 'VUS-PR': 0.35423100776053945, 'VUS-ROC': 0.8261721694707574, 'Standard-F1': 0.4674104521429154, 'Standard-Precision': 0.30591852421214427, 'Standard-Recall': 0.9900497512437786, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9475419729632703, 'Affiliation-P': 0.9210722819593787, 'Affiliation-R': 0.9755780456184684, 'F1_T': 0.46584938703992046, 'Precision_T': 0.3043478260869563, 'Recall_T': 0.9925373134328332}
Results for 004_NAB_id_4_Facility_tr_1007_1st_1437.csv: {'filename': '004_NAB_id_4_Facility_tr_1007_1st_1437.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 50, 'train_index': '1007', 'data_shape': '4031x1', 'output_length': 3024, 'label_length': 3024, 'aligned_length': 3024, 'AUC-PR': 0.3471525589227749, 'AUC-ROC': 0.8184065371085077, 'VUS-PR': 0.35423100776053945, 'VUS-ROC': 0.8261721694707574, 'Standard-F1': 0.4674104521429154, 'Standard-Precision': 0.30591852421214427, 'Standard-Recall': 0.9900497512437786, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9475419729632703, 'Affiliation-P': 0.9210722819593787, 'Affiliation-R': 0.9755780456184684, 'F1_T': 0.46584938703992046, 'Precision_T': 0.3043478260869563, 'Recall_T': 0.9925373134328332}
Logits metrics skipped for 004_NAB_id_4_Facility_tr_1007_1st_1437.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/004_NAB_id_4_Facility_tr_1007_1st_1437_results.pkl
Processing file: 367_UCR_id_65_Medical_tr_9567_1st_16000.csv, Data shape: (38269, 1)
Label shape: (38269,)
Train data shape: (9567, 1), Test data shape: (28702, 1), Label test shape: (28702,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 367_UCR_id_65_Medical_tr_9567_1st_16000.csv...
Using CUDA device: 2
Data shape: (28702, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 28702, window_size 5000, adding 1298 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 142.48it/s]
Original shapes - Output: (30000,), Label: (28702,)
Aligned shapes - Output: (28702,), Label: (28702,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  64%|██████▍   | 129/200 [00:01<00:00, 128.36it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 146.54it/s]
Table1-only metrics completed in 1.69s
Evaluation Result:  {'AUC-PR': 0.11453408969076657, 'AUC-ROC': 0.9854219984616459, 'VUS-PR': 0.13481517035697635, 'VUS-ROC': 0.9866981850929848, 'Standard-F1': 0.26799999999967894, 'Standard-Precision': 0.16749999999999957, 'Standard-Recall': 0.6699999999999934, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.728048186652205, 'Affiliation-P': 0.572504019950625, 'Affiliation-R': 0.9996418368057975, 'F1_T': 0.2646502835535677, 'Precision_T': 0.16317016317016278, 'Recall_T': 0.6999999999999931}
Results for 367_UCR_id_65_Medical_tr_9567_1st_16000.csv: {'filename': '367_UCR_id_65_Medical_tr_9567_1st_16000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 40, 'train_index': '9567', 'data_shape': '38269x1', 'output_length': 30000, 'label_length': 28702, 'aligned_length': 28702, 'AUC-PR': 0.11453408969076657, 'AUC-ROC': 0.9854219984616459, 'VUS-PR': 0.13481517035697635, 'VUS-ROC': 0.9866981850929848, 'Standard-F1': 0.26799999999967894, 'Standard-Precision': 0.16749999999999957, 'Standard-Recall': 0.6699999999999934, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.728048186652205, 'Affiliation-P': 0.572504019950625, 'Affiliation-R': 0.9996418368057975, 'F1_T': 0.2646502835535677, 'Precision_T': 0.16317016317016278, 'Recall_T': 0.6999999999999931}
Logits metrics skipped for 367_UCR_id_65_Medical_tr_9567_1st_16000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/367_UCR_id_65_Medical_tr_9567_1st_16000_results.pkl
Processing file: 600_YAHOO_id_50_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 600_YAHOO_id_50_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 161.66it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1824.34it/s]
Table1-only metrics completed in 0.94s
Evaluation Result:  {'AUC-PR': 0.8875, 'AUC-ROC': 0.9994547437295529, 'VUS-PR': 0.9920041351172983, 'VUS-ROC': 0.9999732326753041, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 600_YAHOO_id_50_Synthetic_tr_500_1st_658.csv: {'filename': '600_YAHOO_id_50_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 151, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.8875, 'AUC-ROC': 0.9994547437295529, 'VUS-PR': 0.9920041351172983, 'VUS-ROC': 0.9999732326753041, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 600_YAHOO_id_50_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/600_YAHOO_id_50_Synthetic_tr_500_1st_658_results.pkl
Processing file: 083_WSD_id_55_WebService_tr_4599_1st_4863.csv, Data shape: (18397, 1)
Label shape: (18397,)
Train data shape: (4599, 1), Test data shape: (13798, 1), Label test shape: (13798,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 083_WSD_id_55_WebService_tr_4599_1st_4863.csv...
Using CUDA device: 2
Data shape: (13798, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13798, window_size 5000, adding 1202 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.84it/s]
Original shapes - Output: (15000,), Label: (13798,)
Aligned shapes - Output: (13798,), Label: (13798,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 272.05it/s]
Table1-only metrics completed in 1.64s
Evaluation Result:  {'AUC-PR': 0.9968656075398666, 'AUC-ROC': 0.9997959387752288, 'VUS-PR': 0.999128018871209, 'VUS-ROC': 0.9999728288039794, 'Standard-F1': 0.9838969404181779, 'Standard-Precision': 0.9760383386581454, 'Standard-Recall': 0.9918831168831153, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996106511012626, 'Affiliation-P': 0.9994738580327757, 'Affiliation-R': 0.9997474816192642, 'F1_T': 0.98861788617836, 'Precision_T': 0.9902280130293143, 'Recall_T': 0.9870129870129853}
Results for 083_WSD_id_55_WebService_tr_4599_1st_4863.csv: {'filename': '083_WSD_id_55_WebService_tr_4599_1st_4863.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4599', 'data_shape': '18397x1', 'output_length': 15000, 'label_length': 13798, 'aligned_length': 13798, 'AUC-PR': 0.9968656075398666, 'AUC-ROC': 0.9997959387752288, 'VUS-PR': 0.999128018871209, 'VUS-ROC': 0.9999728288039794, 'Standard-F1': 0.9838969404181779, 'Standard-Precision': 0.9760383386581454, 'Standard-Recall': 0.9918831168831153, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996106511012626, 'Affiliation-P': 0.9994738580327757, 'Affiliation-R': 0.9997474816192642, 'F1_T': 0.98861788617836, 'Precision_T': 0.9902280130293143, 'Recall_T': 0.9870129870129853}
Logits metrics skipped for 083_WSD_id_55_WebService_tr_4599_1st_4863.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/083_WSD_id_55_WebService_tr_4599_1st_4863_results.pkl
Skipping file: 252_SVDB_id_16_Medical_tr_12167_1st_12267.csv due to filter criteria.
Processing file: 695_YAHOO_id_145_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 695_YAHOO_id_145_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 160.71it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2204.13it/s]
Table1-only metrics completed in 1.38s
Evaluation Result:  {'AUC-PR': 0.95, 'AUC-ROC': 0.9997273718647764, 'VUS-PR': 0.9959177149998506, 'VUS-ROC': 0.9999866260478654, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 695_YAHOO_id_145_Synthetic_tr_500_1st_658.csv: {'filename': '695_YAHOO_id_145_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 303, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.95, 'AUC-ROC': 0.9997273718647764, 'VUS-PR': 0.9959177149998506, 'VUS-ROC': 0.9999866260478654, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9996119518820329, 'Affiliation-P': 0.9992242048099302, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 695_YAHOO_id_145_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/695_YAHOO_id_145_Synthetic_tr_500_1st_658_results.pkl
Processing file: 511_UCR_id_209_Medical_tr_7500_1st_11800.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 511_UCR_id_209_Medical_tr_7500_1st_11800.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 153.12it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  91%|█████████ | 182/200 [00:01<00:00, 181.89it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 190.59it/s]
Table1-only metrics completed in 1.60s
Evaluation Result:  {'AUC-PR': 0.2880283903596697, 'AUC-ROC': 0.9772075075075075, 'VUS-PR': 0.30376324129332, 'VUS-ROC': 0.9777296844748256, 'Standard-F1': 0.4745395449616403, 'Standard-Precision': 0.3515248796147667, 'Standard-Recall': 0.7299999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7609618319862308, 'Affiliation-P': 0.6143914409108755, 'Affiliation-R': 0.9993745185185184, 'F1_T': 0.4731934731930172, 'Precision_T': 0.3637992831541212, 'Recall_T': 0.6766666666666643}
Results for 511_UCR_id_209_Medical_tr_7500_1st_11800.csv: {'filename': '511_UCR_id_209_Medical_tr_7500_1st_11800.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 91, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.2880283903596697, 'AUC-ROC': 0.9772075075075075, 'VUS-PR': 0.30376324129332, 'VUS-ROC': 0.9777296844748256, 'Standard-F1': 0.4745395449616403, 'Standard-Precision': 0.3515248796147667, 'Standard-Recall': 0.7299999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7609618319862308, 'Affiliation-P': 0.6143914409108755, 'Affiliation-R': 0.9993745185185184, 'F1_T': 0.4731934731930172, 'Precision_T': 0.3637992831541212, 'Recall_T': 0.6766666666666643}
Logits metrics skipped for 511_UCR_id_209_Medical_tr_7500_1st_11800.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/511_UCR_id_209_Medical_tr_7500_1st_11800_results.pkl
Skipping file: 199_SMD_id_22_Facility_tr_5925_1st_17580.csv due to filter criteria.
Processing file: 334_UCR_id_32_HumanActivity_tr_1671_1st_2764.csv, Data shape: (6684, 1)
Label shape: (6684,)
Train data shape: (1671, 1), Test data shape: (5013, 1), Label test shape: (5013,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 334_UCR_id_32_HumanActivity_tr_1671_1st_2764.csv...
Using CUDA device: 2
Data shape: (5013, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5013, window_size 5000, adding 4987 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 164.37it/s]
Original shapes - Output: (10000,), Label: (5013,)
Aligned shapes - Output: (5013,), Label: (5013,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 566.58it/s]
Table1-only metrics completed in 0.80s
Evaluation Result:  {'AUC-PR': 0.03774575435081746, 'AUC-ROC': 0.4394029921006082, 'VUS-PR': 0.042232533738274854, 'VUS-ROC': 0.4728150905827318, 'Standard-F1': 0.11328527291440917, 'Standard-Precision': 0.06022447303586092, 'Standard-Recall': 0.9523809523809483, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.746014827631484, 'Affiliation-P': 0.6005011788973488, 'Affiliation-R': 0.9846045303854999, 'F1_T': 0.11334180432009272, 'Precision_T': 0.06020518358531316, 'Recall_T': 0.9653679653679612}
Results for 334_UCR_id_32_HumanActivity_tr_1671_1st_2764.csv: {'filename': '334_UCR_id_32_HumanActivity_tr_1671_1st_2764.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 115, 'train_index': '1671', 'data_shape': '6684x1', 'output_length': 10000, 'label_length': 5013, 'aligned_length': 5013, 'AUC-PR': 0.03774575435081746, 'AUC-ROC': 0.4394029921006082, 'VUS-PR': 0.042232533738274854, 'VUS-ROC': 0.4728150905827318, 'Standard-F1': 0.11328527291440917, 'Standard-Precision': 0.06022447303586092, 'Standard-Recall': 0.9523809523809483, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.746014827631484, 'Affiliation-P': 0.6005011788973488, 'Affiliation-R': 0.9846045303854999, 'F1_T': 0.11334180432009272, 'Precision_T': 0.06020518358531316, 'Recall_T': 0.9653679653679612}
Logits metrics skipped for 334_UCR_id_32_HumanActivity_tr_1671_1st_2764.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/334_UCR_id_32_HumanActivity_tr_1671_1st_2764_results.pkl
Skipping file: 237_SVDB_id_1_Medical_tr_11587_1st_11687.csv due to filter criteria.
Processing file: 435_UCR_id_133_Medical_tr_14750_1st_47830.csv, Data shape: (59000, 1)
Label shape: (59000,)
Train data shape: (14750, 1), Test data shape: (44250, 1), Label test shape: (44250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 435_UCR_id_133_Medical_tr_14750_1st_47830.csv...
Using CUDA device: 2
Data shape: (44250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 44250, window_size 5000, adding 750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([9, 5000, 1])
batched_mask shape: torch.Size([9, 5000])

100%|██████████| 1/1 [00:00<00:00, 111.47it/s]
Original shapes - Output: (45000,), Label: (44250,)
Aligned shapes - Output: (44250,), Label: (44250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  42%|████▏     | 84/200 [00:01<00:01, 83.29it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 177/200 [00:02<00:00, 88.75it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 93.14it/s]
Table1-only metrics completed in 2.36s
Evaluation Result:  {'AUC-PR': 0.0007218659885033592, 'AUC-ROC': 0.6991103323536061, 'VUS-PR': 0.0007654568068796993, 'VUS-ROC': 0.7068089945685668, 'Standard-F1': 0.0021082591050419035, 'Standard-Precision': 0.0010552419142088323, 'Standard-Recall': 0.99999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7032255218675394, 'Affiliation-P': 0.5425144439293584, 'Affiliation-R': 0.999231638418079, 'F1_T': 0.0021069265209354773, 'Precision_T': 0.0010545742156604272, 'Recall_T': 0.99999999999995}
Results for 435_UCR_id_133_Medical_tr_14750_1st_47830.csv: {'filename': '435_UCR_id_133_Medical_tr_14750_1st_47830.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 6, 'train_index': '14750', 'data_shape': '59000x1', 'output_length': 45000, 'label_length': 44250, 'aligned_length': 44250, 'AUC-PR': 0.0007218659885033592, 'AUC-ROC': 0.6991103323536061, 'VUS-PR': 0.0007654568068796993, 'VUS-ROC': 0.7068089945685668, 'Standard-F1': 0.0021082591050419035, 'Standard-Precision': 0.0010552419142088323, 'Standard-Recall': 0.99999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7032255218675394, 'Affiliation-P': 0.5425144439293584, 'Affiliation-R': 0.999231638418079, 'F1_T': 0.0021069265209354773, 'Precision_T': 0.0010545742156604272, 'Recall_T': 0.99999999999995}
Logits metrics skipped for 435_UCR_id_133_Medical_tr_14750_1st_47830.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/435_UCR_id_133_Medical_tr_14750_1st_47830_results.pkl
Processing file: 725_YAHOO_id_175_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 725_YAHOO_id_175_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 160.58it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1623.48it/s]
Table1-only metrics completed in 0.35s
Evaluation Result:  {'AUC-PR': 0.928030303030303, 'AUC-ROC': 0.9991471734892787, 'VUS-PR': 0.9865829970585211, 'VUS-ROC': 0.999915369549478, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Results for 725_YAHOO_id_175_Synthetic_tr_500_1st_623.csv: {'filename': '725_YAHOO_id_175_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 29, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.928030303030303, 'AUC-ROC': 0.9991471734892787, 'VUS-PR': 0.9865829970585211, 'VUS-ROC': 0.999915369549478, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 725_YAHOO_id_175_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/725_YAHOO_id_175_Synthetic_tr_500_1st_623_results.pkl
Skipping file: 837_Exathlon_id_28_Facility_tr_8898_1st_8998.csv due to filter criteria.
Skipping file: 181_SMD_id_4_Facility_tr_5000_1st_15144.csv due to filter criteria.
Processing file: 406_UCR_id_104_Facility_tr_20000_1st_67950.csv, Data shape: (80000, 1)
Label shape: (80000,)
Train data shape: (20000, 1), Test data shape: (60000, 1), Label test shape: (60000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 406_UCR_id_104_Facility_tr_20000_1st_67950.csv...
Using CUDA device: 2
Data shape: (60000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([12, 5000, 1])
batched_mask shape: torch.Size([12, 5000])

100%|██████████| 1/1 [00:00<00:00, 93.24it/s]
Original shapes - Output: (60000,), Label: (60000,)
Aligned shapes - Output: (60000,), Label: (60000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  34%|███▍      | 69/200 [00:01<00:01, 68.63it/s]

Calculating Affiliation Metrics:  73%|███████▎  | 146/200 [00:02<00:00, 73.25it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 80.15it/s]
Table1-only metrics completed in 5.29s
Evaluation Result:  {'AUC-PR': 0.46166353975003416, 'AUC-ROC': 0.9264119497907949, 'VUS-PR': 0.2704524640582537, 'VUS-ROC': 0.9538397145407788, 'Standard-F1': 0.3341176470584076, 'Standard-Precision': 0.23666666666666628, 'Standard-Recall': 0.5679999999999977, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.799643975545801, 'Affiliation-P': 0.6664769867549669, 'Affiliation-R': 0.9993143, 'F1_T': 0.3341176470584076, 'Precision_T': 0.23666666666666628, 'Recall_T': 0.5679999999999977}
Results for 406_UCR_id_104_Facility_tr_20000_1st_67950.csv: {'filename': '406_UCR_id_104_Facility_tr_20000_1st_67950.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 276, 'train_index': '20000', 'data_shape': '80000x1', 'output_length': 60000, 'label_length': 60000, 'aligned_length': 60000, 'AUC-PR': 0.46166353975003416, 'AUC-ROC': 0.9264119497907949, 'VUS-PR': 0.2704524640582537, 'VUS-ROC': 0.9538397145407788, 'Standard-F1': 0.3341176470584076, 'Standard-Precision': 0.23666666666666628, 'Standard-Recall': 0.5679999999999977, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.799643975545801, 'Affiliation-P': 0.6664769867549669, 'Affiliation-R': 0.9993143, 'F1_T': 0.3341176470584076, 'Precision_T': 0.23666666666666628, 'Recall_T': 0.5679999999999977}
Logits metrics skipped for 406_UCR_id_104_Facility_tr_20000_1st_67950.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/406_UCR_id_104_Facility_tr_20000_1st_67950_results.pkl
Processing file: 445_UCR_id_143_Environment_tr_6166_1st_19280.csv, Data shape: (24667, 1)
Label shape: (24667,)
Train data shape: (6166, 1), Test data shape: (18501, 1), Label test shape: (18501,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 445_UCR_id_143_Environment_tr_6166_1st_19280.csv...
Using CUDA device: 2
Data shape: (18501, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 18501, window_size 5000, adding 1499 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.50it/s]
Original shapes - Output: (20000,), Label: (18501,)
Aligned shapes - Output: (18501,), Label: (18501,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 213.59it/s]
Table1-only metrics completed in 1.22s
Evaluation Result:  {'AUC-PR': 0.8724228838249158, 'AUC-ROC': 0.9823523799138542, 'VUS-PR': 0.8820998846604524, 'VUS-ROC': 0.9912332295960242, 'Standard-F1': 0.7919075144503652, 'Standard-Precision': 0.7365591397849423, 'Standard-Recall': 0.8562499999999946, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9997541758428342, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9995084725149992, 'F1_T': 0.7919075144503652, 'Precision_T': 0.7365591397849423, 'Recall_T': 0.8562499999999946}
Results for 445_UCR_id_143_Environment_tr_6166_1st_19280.csv: {'filename': '445_UCR_id_143_Environment_tr_6166_1st_19280.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 45, 'train_index': '6166', 'data_shape': '24667x1', 'output_length': 20000, 'label_length': 18501, 'aligned_length': 18501, 'AUC-PR': 0.8724228838249158, 'AUC-ROC': 0.9823523799138542, 'VUS-PR': 0.8820998846604524, 'VUS-ROC': 0.9912332295960242, 'Standard-F1': 0.7919075144503652, 'Standard-Precision': 0.7365591397849423, 'Standard-Recall': 0.8562499999999946, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9997541758428342, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9995084725149992, 'F1_T': 0.7919075144503652, 'Precision_T': 0.7365591397849423, 'Recall_T': 0.8562499999999946}
Logits metrics skipped for 445_UCR_id_143_Environment_tr_6166_1st_19280.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/445_UCR_id_143_Environment_tr_6166_1st_19280_results.pkl
Processing file: 466_UCR_id_164_Medical_tr_1900_1st_5684.csv, Data shape: (7601, 1)
Label shape: (7601,)
Train data shape: (1900, 1), Test data shape: (5701, 1), Label test shape: (5701,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 466_UCR_id_164_Medical_tr_1900_1st_5684.csv...
Using CUDA device: 2
Data shape: (5701, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5701, window_size 5000, adding 4299 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.73it/s]
Original shapes - Output: (10000,), Label: (5701,)
Aligned shapes - Output: (5701,), Label: (5701,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 629.00it/s]
Table1-only metrics completed in 0.99s
Evaluation Result:  {'AUC-PR': 0.9208866903446082, 'AUC-ROC': 0.9942867474236124, 'VUS-PR': 0.940486839934681, 'VUS-ROC': 0.9975017616788768, 'Standard-F1': 0.8706624605673204, 'Standard-Precision': 0.9387755102040752, 'Standard-Recall': 0.8117647058823482, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9995119734261653, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9990244229598522, 'F1_T': 0.8589743589738575, 'Precision_T': 0.9436619718309793, 'Recall_T': 0.7882352941176425}
Results for 466_UCR_id_164_Medical_tr_1900_1st_5684.csv: {'filename': '466_UCR_id_164_Medical_tr_1900_1st_5684.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 173, 'train_index': '1900', 'data_shape': '7601x1', 'output_length': 10000, 'label_length': 5701, 'aligned_length': 5701, 'AUC-PR': 0.9208866903446082, 'AUC-ROC': 0.9942867474236124, 'VUS-PR': 0.940486839934681, 'VUS-ROC': 0.9975017616788768, 'Standard-F1': 0.8706624605673204, 'Standard-Precision': 0.9387755102040752, 'Standard-Recall': 0.8117647058823482, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9995119734261653, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9990244229598522, 'F1_T': 0.8589743589738575, 'Precision_T': 0.9436619718309793, 'Recall_T': 0.7882352941176425}
Logits metrics skipped for 466_UCR_id_164_Medical_tr_1900_1st_5684.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/466_UCR_id_164_Medical_tr_1900_1st_5684_results.pkl
Processing file: 168_Stock_id_20_Finance_tr_500_1st_1.csv, Data shape: (10000, 1)
Label shape: (10000,)
Train data shape: (500, 1), Test data shape: (9500, 1), Label test shape: (9500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 168_Stock_id_20_Finance_tr_500_1st_1.csv...
Using CUDA device: 2
Data shape: (9500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.64it/s]
Original shapes - Output: (10000,), Label: (9500,)
Aligned shapes - Output: (9500,), Label: (9500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   7%|▋         | 14/200 [00:01<00:13, 13.39it/s]

Calculating Affiliation Metrics:  14%|█▍        | 28/200 [00:03<00:22,  7.72it/s]

Calculating Affiliation Metrics:  19%|█▉        | 38/200 [00:05<00:28,  5.73it/s]

Calculating Affiliation Metrics:  22%|██▎       | 45/200 [00:08<00:34,  4.51it/s]

Calculating Affiliation Metrics:  26%|██▌       | 51/200 [00:10<00:39,  3.78it/s]

Calculating Affiliation Metrics:  28%|██▊       | 56/200 [00:12<00:43,  3.34it/s]

Calculating Affiliation Metrics:  30%|███       | 60/200 [00:14<00:45,  3.11it/s]

Calculating Affiliation Metrics:  32%|███▏      | 64/200 [00:16<00:46,  2.94it/s]

Calculating Affiliation Metrics:  34%|███▎      | 67/200 [00:17<00:46,  2.85it/s]

Calculating Affiliation Metrics:  35%|███▌      | 70/200 [00:18<00:46,  2.79it/s]

Calculating Affiliation Metrics:  36%|███▋      | 73/200 [00:19<00:46,  2.76it/s]

Calculating Affiliation Metrics:  38%|███▊      | 76/200 [00:20<00:44,  2.77it/s]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [00:21<00:43,  2.79it/s]

Calculating Affiliation Metrics:  41%|████      | 82/200 [00:22<00:42,  2.81it/s]

Calculating Affiliation Metrics:  42%|████▎     | 85/200 [00:23<00:41,  2.80it/s]

Calculating Affiliation Metrics:  44%|████▍     | 88/200 [00:24<00:40,  2.80it/s]

Calculating Affiliation Metrics:  46%|████▌     | 91/200 [00:25<00:39,  2.79it/s]

Calculating Affiliation Metrics:  47%|████▋     | 94/200 [00:26<00:37,  2.79it/s]

Calculating Affiliation Metrics:  48%|████▊     | 97/200 [00:27<00:36,  2.84it/s]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [00:29<00:32,  3.01it/s]

Calculating Affiliation Metrics:  52%|█████▎    | 105/200 [00:30<00:30,  3.11it/s]

Calculating Affiliation Metrics:  55%|█████▍    | 109/200 [00:31<00:28,  3.18it/s]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [00:32<00:26,  3.31it/s]

Calculating Affiliation Metrics:  58%|█████▊    | 117/200 [00:33<00:24,  3.45it/s]

Calculating Affiliation Metrics:  61%|██████    | 122/200 [00:34<00:21,  3.64it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [00:36<00:19,  3.79it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 132/200 [00:37<00:17,  3.92it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 137/200 [00:38<00:15,  4.00it/s]

Calculating Affiliation Metrics:  71%|███████   | 142/200 [00:39<00:14,  4.08it/s]

Calculating Affiliation Metrics:  74%|███████▎  | 147/200 [00:40<00:12,  4.21it/s]

Calculating Affiliation Metrics:  76%|███████▌  | 152/200 [00:41<00:10,  4.41it/s]

Calculating Affiliation Metrics:  79%|███████▉  | 158/200 [00:42<00:08,  4.71it/s]

Calculating Affiliation Metrics:  82%|████████▎ | 165/200 [00:44<00:06,  5.21it/s]

Calculating Affiliation Metrics:  87%|████████▋ | 174/200 [00:45<00:04,  6.13it/s]

Calculating Affiliation Metrics:  96%|█████████▋| 193/200 [00:46<00:00,  9.72it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:46<00:00,  4.33it/s]
Table1-only metrics completed in 73.71s
Evaluation Result:  {'AUC-PR': 0.15772348202477798, 'AUC-ROC': 0.5369891275931994, 'VUS-PR': 0.9407654980913913, 'VUS-ROC': 0.9640393568071192, 'Standard-F1': 0.24555532315391468, 'Standard-Precision': 0.1423375363724539, 'Standard-Recall': 0.8934550989345503, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6864554112558211, 'Affiliation-P': 0.5239778233496456, 'Affiliation-R': 0.9949852778800148, 'F1_T': 0.2454707299191259, 'Precision_T': 0.14231936854887672, 'Recall_T': 0.8919330289193297}
Results for 168_Stock_id_20_Finance_tr_500_1st_1.csv: {'filename': '168_Stock_id_20_Finance_tr_500_1st_1.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '10000x1', 'output_length': 10000, 'label_length': 9500, 'aligned_length': 9500, 'AUC-PR': 0.15772348202477798, 'AUC-ROC': 0.5369891275931994, 'VUS-PR': 0.9407654980913913, 'VUS-ROC': 0.9640393568071192, 'Standard-F1': 0.24555532315391468, 'Standard-Precision': 0.1423375363724539, 'Standard-Recall': 0.8934550989345503, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6864554112558211, 'Affiliation-P': 0.5239778233496456, 'Affiliation-R': 0.9949852778800148, 'F1_T': 0.2454707299191259, 'Precision_T': 0.14231936854887672, 'Recall_T': 0.8919330289193297}
Logits metrics skipped for 168_Stock_id_20_Finance_tr_500_1st_1.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/168_Stock_id_20_Finance_tr_500_1st_1_results.pkl
Processing file: 432_UCR_id_130_Facility_tr_23000_1st_26970.csv, Data shape: (92000, 1)
Label shape: (92000,)
Train data shape: (23000, 1), Test data shape: (69000, 1), Label test shape: (69000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 432_UCR_id_130_Facility_tr_23000_1st_26970.csv...
Using CUDA device: 2
Data shape: (69000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 69000, window_size 5000, adding 1000 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([14, 5000, 1])
batched_mask shape: torch.Size([14, 5000])

100%|██████████| 1/1 [00:00<00:00, 82.40it/s]
Original shapes - Output: (70000,), Label: (69000,)
Aligned shapes - Output: (69000,), Label: (69000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:01<00:02, 51.97it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 104/200 [00:02<00:02, 47.46it/s]

Calculating Affiliation Metrics:  76%|███████▌  | 152/200 [00:03<00:01, 47.27it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:03<00:00, 53.72it/s]
Table1-only metrics completed in 7.11s
Evaluation Result:  {'AUC-PR': 0.04584233350415426, 'AUC-ROC': 0.9593174187287725, 'VUS-PR': 0.048837110174382685, 'VUS-ROC': 0.9595745540731186, 'Standard-F1': 0.12320428536631207, 'Standard-Precision': 0.06645652744943524, 'Standard-Recall': 0.8433333333333305, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9467749069218072, 'Affiliation-P': 0.9012461749432681, 'Affiliation-R': 0.9971483816425123, 'F1_T': 0.12215132178656353, 'Precision_T': 0.06555772994129157, 'Recall_T': 0.8933333333333303}
Results for 432_UCR_id_130_Facility_tr_23000_1st_26970.csv: {'filename': '432_UCR_id_130_Facility_tr_23000_1st_26970.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 262, 'train_index': '23000', 'data_shape': '92000x1', 'output_length': 70000, 'label_length': 69000, 'aligned_length': 69000, 'AUC-PR': 0.04584233350415426, 'AUC-ROC': 0.9593174187287725, 'VUS-PR': 0.048837110174382685, 'VUS-ROC': 0.9595745540731186, 'Standard-F1': 0.12320428536631207, 'Standard-Precision': 0.06645652744943524, 'Standard-Recall': 0.8433333333333305, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9467749069218072, 'Affiliation-P': 0.9012461749432681, 'Affiliation-R': 0.9971483816425123, 'F1_T': 0.12215132178656353, 'Precision_T': 0.06555772994129157, 'Recall_T': 0.8933333333333303}
Logits metrics skipped for 432_UCR_id_130_Facility_tr_23000_1st_26970.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/432_UCR_id_130_Facility_tr_23000_1st_26970_results.pkl
Skipping file: 238_SVDB_id_2_Medical_tr_25000_1st_71963.csv due to filter criteria.
Processing file: 382_UCR_id_80_Medical_tr_7500_1st_11800.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 382_UCR_id_80_Medical_tr_7500_1st_11800.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 149.38it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  80%|███████▉  | 159/200 [00:01<00:00, 158.46it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 174.68it/s]
Table1-only metrics completed in 1.70s
Evaluation Result:  {'AUC-PR': 0.30827637831405175, 'AUC-ROC': 0.9794298048048047, 'VUS-PR': 0.31825374980978566, 'VUS-ROC': 0.9799190449713787, 'Standard-F1': 0.4766467065863647, 'Standard-Precision': 0.3719626168224292, 'Standard-Recall': 0.6633333333333311, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7958913318736345, 'Affiliation-P': 0.6613017647058823, 'Affiliation-R': 0.9992636296296297, 'F1_T': 0.4778554778550218, 'Precision_T': 0.36738351254480217, 'Recall_T': 0.683333333333331}
Results for 382_UCR_id_80_Medical_tr_7500_1st_11800.csv: {'filename': '382_UCR_id_80_Medical_tr_7500_1st_11800.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 91, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.30827637831405175, 'AUC-ROC': 0.9794298048048047, 'VUS-PR': 0.31825374980978566, 'VUS-ROC': 0.9799190449713787, 'Standard-F1': 0.4766467065863647, 'Standard-Precision': 0.3719626168224292, 'Standard-Recall': 0.6633333333333311, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7958913318736345, 'Affiliation-P': 0.6613017647058823, 'Affiliation-R': 0.9992636296296297, 'F1_T': 0.4778554778550218, 'Precision_T': 0.36738351254480217, 'Recall_T': 0.683333333333331}
Logits metrics skipped for 382_UCR_id_80_Medical_tr_7500_1st_11800.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/382_UCR_id_80_Medical_tr_7500_1st_11800_results.pkl
Processing file: 673_YAHOO_id_123_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 673_YAHOO_id_123_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 148.23it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2129.50it/s]
Table1-only metrics completed in 0.39s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9702593917710192, 'Affiliation-P': 0.9422366992399566, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 673_YAHOO_id_123_Synthetic_tr_500_1st_893.csv: {'filename': '673_YAHOO_id_123_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 91, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9702593917710192, 'Affiliation-P': 0.9422366992399566, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 673_YAHOO_id_123_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/673_YAHOO_id_123_Synthetic_tr_500_1st_893_results.pkl
Processing file: 710_YAHOO_id_160_WebService_tr_500_1st_253.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 710_YAHOO_id_160_WebService_tr_500_1st_253.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 163.76it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1275.83it/s]
Table1-only metrics completed in 0.43s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.8125, 'VUS-ROC': 0.999601403061224, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9927920792079202, 'Affiliation-P': 0.9856873230575652, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Results for 710_YAHOO_id_160_WebService_tr_500_1st_253.csv: {'filename': '710_YAHOO_id_160_WebService_tr_500_1st_253.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.8125, 'VUS-ROC': 0.999601403061224, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9927920792079202, 'Affiliation-P': 0.9856873230575652, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 710_YAHOO_id_160_WebService_tr_500_1st_253.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/710_YAHOO_id_160_WebService_tr_500_1st_253_results.pkl
Processing file: 023_NAB_id_23_Facility_tr_4512_1st_16551.csv, Data shape: (18049, 1)
Label shape: (18049,)
Train data shape: (4512, 1), Test data shape: (13537, 1), Label test shape: (13537,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 023_NAB_id_23_Facility_tr_4512_1st_16551.csv...
Using CUDA device: 2
Data shape: (13537, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13537, window_size 5000, adding 1463 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.47it/s]
Original shapes - Output: (15000,), Label: (13537,)
Aligned shapes - Output: (13537,), Label: (13537,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 263.79it/s]
Table1-only metrics completed in 0.88s
Evaluation Result:  {'AUC-PR': 0.6968368673631021, 'AUC-ROC': 0.7902135150214407, 'VUS-PR': 0.6976759201021514, 'VUS-ROC': 0.7904637455181489, 'Standard-F1': 0.7351652727991994, 'Standard-Precision': 0.9111549851924966, 'Standard-Recall': 0.6161548731642186, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9917503556538153, 'Affiliation-P': 0.9933352837174279, 'Affiliation-R': 0.9901704772352647, 'F1_T': 0.734399999999519, 'Precision_T': 0.9161676646706578, 'Recall_T': 0.612817089452603}
Results for 023_NAB_id_23_Facility_tr_4512_1st_16551.csv: {'filename': '023_NAB_id_23_Facility_tr_4512_1st_16551.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 12, 'train_index': '4512', 'data_shape': '18049x1', 'output_length': 15000, 'label_length': 13537, 'aligned_length': 13537, 'AUC-PR': 0.6968368673631021, 'AUC-ROC': 0.7902135150214407, 'VUS-PR': 0.6976759201021514, 'VUS-ROC': 0.7904637455181489, 'Standard-F1': 0.7351652727991994, 'Standard-Precision': 0.9111549851924966, 'Standard-Recall': 0.6161548731642186, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9917503556538153, 'Affiliation-P': 0.9933352837174279, 'Affiliation-R': 0.9901704772352647, 'F1_T': 0.734399999999519, 'Precision_T': 0.9161676646706578, 'Recall_T': 0.612817089452603}
Logits metrics skipped for 023_NAB_id_23_Facility_tr_4512_1st_16551.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/023_NAB_id_23_Facility_tr_4512_1st_16551_results.pkl
Skipping file: 854_OPPORTUNITY_id_13_HumanActivity_tr_1477_1st_1577.csv due to filter criteria.
Processing file: 768_YAHOO_id_218_WebService_tr_500_1st_431.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 768_YAHOO_id_218_WebService_tr_500_1st_431.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 165.23it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 382.81it/s]
Table1-only metrics completed in 0.93s
Evaluation Result:  {'AUC-PR': 0.9999999999999999, 'AUC-ROC': 1.0, 'VUS-PR': 0.9921143946478479, 'VUS-ROC': 0.9999492711591995, 'Standard-F1': 0.9655172413787443, 'Standard-Precision': 0.9333333333332711, 'Standard-Recall': 0.9999999999999286, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9978698413291466, 'Affiliation-P': 0.9957487385197982, 'Affiliation-R': 1.0, 'F1_T': 0.923076923076355, 'Precision_T': 0.9999999999999166, 'Recall_T': 0.8571428571427959}
Results for 768_YAHOO_id_218_WebService_tr_500_1st_431.csv: {'filename': '768_YAHOO_id_218_WebService_tr_500_1st_431.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 12, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9999999999999999, 'AUC-ROC': 1.0, 'VUS-PR': 0.9921143946478479, 'VUS-ROC': 0.9999492711591995, 'Standard-F1': 0.9655172413787443, 'Standard-Precision': 0.9333333333332711, 'Standard-Recall': 0.9999999999999286, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9978698413291466, 'Affiliation-P': 0.9957487385197982, 'Affiliation-R': 1.0, 'F1_T': 0.923076923076355, 'Precision_T': 0.9999999999999166, 'Recall_T': 0.8571428571427959}
Logits metrics skipped for 768_YAHOO_id_218_WebService_tr_500_1st_431.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/768_YAHOO_id_218_WebService_tr_500_1st_431_results.pkl
Processing file: 568_YAHOO_id_18_WebService_tr_500_1st_333.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 568_YAHOO_id_18_WebService_tr_500_1st_333.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.56it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1459.24it/s]
Table1-only metrics completed in 0.24s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9967408247130503, 'Affiliation-P': 0.9935028248587571, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Results for 568_YAHOO_id_18_WebService_tr_500_1st_333.csv: {'filename': '568_YAHOO_id_18_WebService_tr_500_1st_333.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.15384615384598815, 'Standard-Precision': 0.08333333333332639, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9967408247130503, 'Affiliation-P': 0.9935028248587571, 'Affiliation-R': 1.0, 'F1_T': 0.15384615384598815, 'Precision_T': 0.08333333333332639, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 568_YAHOO_id_18_WebService_tr_500_1st_333.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/568_YAHOO_id_18_WebService_tr_500_1st_333_results.pkl
Skipping file: 867_OPPORTUNITY_id_26_HumanActivity_tr_808_1st_908.csv due to filter criteria.
Processing file: 779_YAHOO_id_229_WebService_tr_500_1st_25.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 779_YAHOO_id_229_WebService_tr_500_1st_25.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 157.32it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1234.91it/s]
Table1-only metrics completed in 0.59s
Evaluation Result:  {'AUC-PR': 0.5779482175383814, 'AUC-ROC': 0.9031569965870307, 'VUS-PR': 0.5985566084057246, 'VUS-ROC': 0.9284606130971169, 'Standard-F1': 0.49999999999947, 'Standard-Precision': 0.41666666666663194, 'Standard-Recall': 0.6249999999999218, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8813882543867252, 'Affiliation-P': 0.7879304484716529, 'Affiliation-R': 1.0, 'F1_T': 0.49999999999947, 'Precision_T': 0.41666666666663194, 'Recall_T': 0.6249999999999218}
Results for 779_YAHOO_id_229_WebService_tr_500_1st_25.csv: {'filename': '779_YAHOO_id_229_WebService_tr_500_1st_25.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.5779482175383814, 'AUC-ROC': 0.9031569965870307, 'VUS-PR': 0.5985566084057246, 'VUS-ROC': 0.9284606130971169, 'Standard-F1': 0.49999999999947, 'Standard-Precision': 0.41666666666663194, 'Standard-Recall': 0.6249999999999218, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8813882543867252, 'Affiliation-P': 0.7879304484716529, 'Affiliation-R': 1.0, 'F1_T': 0.49999999999947, 'Precision_T': 0.41666666666663194, 'Recall_T': 0.6249999999999218}
Logits metrics skipped for 779_YAHOO_id_229_WebService_tr_500_1st_25.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/779_YAHOO_id_229_WebService_tr_500_1st_25_results.pkl
Skipping file: 172_MITDB_id_3_Medical_tr_27274_1st_27374.csv due to filter criteria.
Skipping file: 533_SMAP_id_3_Sensor_tr_2052_1st_5300.csv due to filter criteria.
Processing file: 671_YAHOO_id_121_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 671_YAHOO_id_121_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 156.76it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1780.04it/s]
Table1-only metrics completed in 0.18s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 671_YAHOO_id_121_Synthetic_tr_500_1st_658.csv: {'filename': '671_YAHOO_id_121_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 10, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 671_YAHOO_id_121_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/671_YAHOO_id_121_Synthetic_tr_500_1st_658_results.pkl
Processing file: 508_UCR_id_206_Medical_tr_9567_1st_16000.csv, Data shape: (38269, 1)
Label shape: (38269,)
Train data shape: (9567, 1), Test data shape: (28702, 1), Label test shape: (28702,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 508_UCR_id_206_Medical_tr_9567_1st_16000.csv...
Using CUDA device: 2
Data shape: (28702, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 28702, window_size 5000, adding 1298 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 144.50it/s]
Original shapes - Output: (30000,), Label: (28702,)
Aligned shapes - Output: (28702,), Label: (28702,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [00:01<00:00, 114.39it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 128.38it/s]
Table1-only metrics completed in 1.89s
Evaluation Result:  {'AUC-PR': 0.02338367957656044, 'AUC-ROC': 0.9329001468428781, 'VUS-PR': 0.025343627347554916, 'VUS-ROC': 0.9346503199713612, 'Standard-F1': 0.06207211319023689, 'Standard-Precision': 0.032520325203252015, 'Standard-Recall': 0.6799999999999933, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6803003846322706, 'Affiliation-P': 0.5471747733547344, 'Affiliation-R': 0.8990314263814367, 'F1_T': 0.06112359550553208, 'Precision_T': 0.03199999999999999, 'Recall_T': 0.6799999999999933}
Results for 508_UCR_id_206_Medical_tr_9567_1st_16000.csv: {'filename': '508_UCR_id_206_Medical_tr_9567_1st_16000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 40, 'train_index': '9567', 'data_shape': '38269x1', 'output_length': 30000, 'label_length': 28702, 'aligned_length': 28702, 'AUC-PR': 0.02338367957656044, 'AUC-ROC': 0.9329001468428781, 'VUS-PR': 0.025343627347554916, 'VUS-ROC': 0.9346503199713612, 'Standard-F1': 0.06207211319023689, 'Standard-Precision': 0.032520325203252015, 'Standard-Recall': 0.6799999999999933, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6803003846322706, 'Affiliation-P': 0.5471747733547344, 'Affiliation-R': 0.8990314263814367, 'F1_T': 0.06112359550553208, 'Precision_T': 0.03199999999999999, 'Recall_T': 0.6799999999999933}
Logits metrics skipped for 508_UCR_id_206_Medical_tr_9567_1st_16000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/508_UCR_id_206_Medical_tr_9567_1st_16000_results.pkl
Skipping file: 842_OPPORTUNITY_id_1_HumanActivity_tr_7295_1st_27516.csv due to filter criteria.
Processing file: 563_YAHOO_id_13_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 563_YAHOO_id_13_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 155.84it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1278.68it/s]
Table1-only metrics completed in 0.40s
Evaluation Result:  {'AUC-PR': 0.9361111111111111, 'AUC-ROC': 0.9992690058479532, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 563_YAHOO_id_13_Synthetic_tr_500_1st_623.csv: {'filename': '563_YAHOO_id_13_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 35, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9361111111111111, 'AUC-ROC': 0.9992690058479532, 'VUS-PR': 0.9885322491644004, 'VUS-ROC': 0.9999277632142323, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 563_YAHOO_id_13_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/563_YAHOO_id_13_Synthetic_tr_500_1st_623_results.pkl
Processing file: 338_UCR_id_36_Facility_tr_9000_1st_12400.csv, Data shape: (36001, 1)
Label shape: (36001,)
Train data shape: (9000, 1), Test data shape: (27001, 1), Label test shape: (27001,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 338_UCR_id_36_Facility_tr_9000_1st_12400.csv...
Using CUDA device: 2
Data shape: (27001, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 27001, window_size 5000, adding 2999 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 92.00it/s]
Original shapes - Output: (30000,), Label: (27001,)
Aligned shapes - Output: (27001,), Label: (27001,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  70%|██████▉   | 139/200 [00:01<00:00, 138.55it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 152.42it/s]
Table1-only metrics completed in 2.42s
Evaluation Result:  {'AUC-PR': 0.0729165091558159, 'AUC-ROC': 0.8049164505093793, 'VUS-PR': 0.07778438277166448, 'VUS-ROC': 0.8353167161101392, 'Standard-F1': 0.17355371900798838, 'Standard-Precision': 0.10396039603960391, 'Standard-Recall': 0.5249999999999987, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8594746896669713, 'Affiliation-P': 0.7567211562970434, 'Affiliation-R': 0.9945181659938521, 'F1_T': 0.17423926636070575, 'Precision_T': 0.10455227613806899, 'Recall_T': 0.5224999999999986}
Results for 338_UCR_id_36_Facility_tr_9000_1st_12400.csv: {'filename': '338_UCR_id_36_Facility_tr_9000_1st_12400.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 188, 'train_index': '9000', 'data_shape': '36001x1', 'output_length': 30000, 'label_length': 27001, 'aligned_length': 27001, 'AUC-PR': 0.0729165091558159, 'AUC-ROC': 0.8049164505093793, 'VUS-PR': 0.07778438277166448, 'VUS-ROC': 0.8353167161101392, 'Standard-F1': 0.17355371900798838, 'Standard-Precision': 0.10396039603960391, 'Standard-Recall': 0.5249999999999987, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8594746896669713, 'Affiliation-P': 0.7567211562970434, 'Affiliation-R': 0.9945181659938521, 'F1_T': 0.17423926636070575, 'Precision_T': 0.10455227613806899, 'Recall_T': 0.5224999999999986}
Logits metrics skipped for 338_UCR_id_36_Facility_tr_9000_1st_12400.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/338_UCR_id_36_Facility_tr_9000_1st_12400_results.pkl
Skipping file: 865_OPPORTUNITY_id_24_HumanActivity_tr_2085_1st_2185.csv due to filter criteria.
Processing file: 264_IOPS_id_5_WebService_tr_11665_1st_11765.csv, Data shape: (60000, 1)
Label shape: (60000,)
Train data shape: (11665, 1), Test data shape: (48335, 1), Label test shape: (48335,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 264_IOPS_id_5_WebService_tr_11665_1st_11765.csv...
Using CUDA device: 2
Data shape: (48335, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 48335, window_size 5000, adding 1665 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([10, 5000, 1])
batched_mask shape: torch.Size([10, 5000])

100%|██████████| 1/1 [00:00<00:00, 96.00it/s]
Original shapes - Output: (50000,), Label: (48335,)
Aligned shapes - Output: (48335,), Label: (48335,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  30%|███       | 60/200 [00:01<00:02, 59.54it/s]

Calculating Affiliation Metrics:  60%|██████    | 120/200 [00:02<00:01, 54.63it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:03<00:00, 64.22it/s]
Table1-only metrics completed in 3.40s
Evaluation Result:  {'AUC-PR': 0.02501850427599022, 'AUC-ROC': 0.47915501915694114, 'VUS-PR': 0.02411418571182062, 'VUS-ROC': 0.4834715668252911, 'Standard-F1': 0.04639922667921466, 'Standard-Precision': 0.029657089898053744, 'Standard-Recall': 0.10654827968923407, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7861251310693557, 'Affiliation-P': 0.6476162924280825, 'Affiliation-R': 0.9999999783559798, 'F1_T': 0.04596552585525943, 'Precision_T': 0.029658284977433905, 'Recall_T': 0.10210876803551598}
Results for 264_IOPS_id_5_WebService_tr_11665_1st_11765.csv: {'filename': '264_IOPS_id_5_WebService_tr_11665_1st_11765.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 9, 'train_index': '11665', 'data_shape': '60000x1', 'output_length': 50000, 'label_length': 48335, 'aligned_length': 48335, 'AUC-PR': 0.02501850427599022, 'AUC-ROC': 0.47915501915694114, 'VUS-PR': 0.02411418571182062, 'VUS-ROC': 0.4834715668252911, 'Standard-F1': 0.04639922667921466, 'Standard-Precision': 0.029657089898053744, 'Standard-Recall': 0.10654827968923407, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7861251310693557, 'Affiliation-P': 0.6476162924280825, 'Affiliation-R': 0.9999999783559798, 'F1_T': 0.04596552585525943, 'Precision_T': 0.029658284977433905, 'Recall_T': 0.10210876803551598}
Logits metrics skipped for 264_IOPS_id_5_WebService_tr_11665_1st_11765.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/264_IOPS_id_5_WebService_tr_11665_1st_11765_results.pkl
Processing file: 500_UCR_id_198_Facility_tr_7482_1st_23357.csv, Data shape: (29931, 1)
Label shape: (29931,)
Train data shape: (7482, 1), Test data shape: (22449, 1), Label test shape: (22449,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 500_UCR_id_198_Facility_tr_7482_1st_23357.csv...
Using CUDA device: 2
Data shape: (22449, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22449, window_size 5000, adding 2551 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 146.58it/s]
Original shapes - Output: (25000,), Label: (22449,)
Aligned shapes - Output: (22449,), Label: (22449,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [00:01<00:00, 125.77it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 150.21it/s]
Table1-only metrics completed in 2.28s
Evaluation Result:  {'AUC-PR': 0.05495567783638525, 'AUC-ROC': 0.8735855830705076, 'VUS-PR': 0.0604516857985436, 'VUS-ROC': 0.8801798986433391, 'Standard-F1': 0.13041385948011122, 'Standard-Precision': 0.07139093782929398, 'Standard-Recall': 0.7527777777777757, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.684131752408399, 'Affiliation-P': 0.5199090058298376, 'Affiliation-R': 1.0, 'F1_T': 0.13000242541822524, 'Precision_T': 0.07121977145894232, 'Recall_T': 0.7444444444444424}
Results for 500_UCR_id_198_Facility_tr_7482_1st_23357.csv: {'filename': '500_UCR_id_198_Facility_tr_7482_1st_23357.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 168, 'train_index': '7482', 'data_shape': '29931x1', 'output_length': 25000, 'label_length': 22449, 'aligned_length': 22449, 'AUC-PR': 0.05495567783638525, 'AUC-ROC': 0.8735855830705076, 'VUS-PR': 0.0604516857985436, 'VUS-ROC': 0.8801798986433391, 'Standard-F1': 0.13041385948011122, 'Standard-Precision': 0.07139093782929398, 'Standard-Recall': 0.7527777777777757, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.684131752408399, 'Affiliation-P': 0.5199090058298376, 'Affiliation-R': 1.0, 'F1_T': 0.13000242541822524, 'Precision_T': 0.07121977145894232, 'Recall_T': 0.7444444444444424}
Logits metrics skipped for 500_UCR_id_198_Facility_tr_7482_1st_23357.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/500_UCR_id_198_Facility_tr_7482_1st_23357_results.pkl
Processing file: 161_Stock_id_13_Finance_tr_500_1st_3.csv, Data shape: (10000, 1)
Label shape: (10000,)
Train data shape: (500, 1), Test data shape: (9500, 1), Label test shape: (9500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 161_Stock_id_13_Finance_tr_500_1st_3.csv...
Using CUDA device: 2
Data shape: (9500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.46it/s]
Original shapes - Output: (10000,), Label: (9500,)
Aligned shapes - Output: (9500,), Label: (9500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  12%|█▏        | 23/200 [00:01<00:08, 21.35it/s]

Calculating Affiliation Metrics:  22%|██▎       | 45/200 [00:03<00:12, 12.03it/s]

Calculating Affiliation Metrics:  30%|███       | 60/200 [00:05<00:13, 10.12it/s]

Calculating Affiliation Metrics:  36%|███▌      | 72/200 [00:07<00:15,  8.34it/s]

Calculating Affiliation Metrics:  41%|████      | 82/200 [00:09<00:15,  7.52it/s]

Calculating Affiliation Metrics:  45%|████▌     | 90/200 [00:10<00:15,  7.01it/s]

Calculating Affiliation Metrics:  49%|████▉     | 98/200 [00:11<00:15,  6.69it/s]

Calculating Affiliation Metrics:  52%|█████▎    | 105/200 [00:13<00:14,  6.61it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [00:14<00:13,  6.60it/s]

Calculating Affiliation Metrics:  60%|█████▉    | 119/200 [00:15<00:12,  6.66it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [00:16<00:10,  6.80it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 135/200 [00:17<00:09,  7.07it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [00:18<00:07,  7.82it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 157/200 [00:19<00:04,  8.86it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 175/200 [00:20<00:02, 11.35it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:20<00:00,  9.61it/s]
Table1-only metrics completed in 26.84s
Evaluation Result:  {'AUC-PR': 0.0757123520792001, 'AUC-ROC': 0.6234082545576511, 'VUS-PR': 0.18644945777322627, 'VUS-ROC': 0.6237362681356547, 'Standard-F1': 0.13968668407268356, 'Standard-Precision': 0.1003752345215759, 'Standard-Recall': 0.2296137339055789, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6774450643077143, 'Affiliation-P': 0.515771946752783, 'Affiliation-R': 0.9867503951582532, 'F1_T': 0.13917808219140038, 'Precision_T': 0.09345106696100067, 'Recall_T': 0.2725321888412011}
Results for 161_Stock_id_13_Finance_tr_500_1st_3.csv: {'filename': '161_Stock_id_13_Finance_tr_500_1st_3.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 7, 'train_index': '500', 'data_shape': '10000x1', 'output_length': 10000, 'label_length': 9500, 'aligned_length': 9500, 'AUC-PR': 0.0757123520792001, 'AUC-ROC': 0.6234082545576511, 'VUS-PR': 0.18644945777322627, 'VUS-ROC': 0.6237362681356547, 'Standard-F1': 0.13968668407268356, 'Standard-Precision': 0.1003752345215759, 'Standard-Recall': 0.2296137339055789, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6774450643077143, 'Affiliation-P': 0.515771946752783, 'Affiliation-R': 0.9867503951582532, 'F1_T': 0.13917808219140038, 'Precision_T': 0.09345106696100067, 'Recall_T': 0.2725321888412011}
Logits metrics skipped for 161_Stock_id_13_Finance_tr_500_1st_3.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/161_Stock_id_13_Finance_tr_500_1st_3_results.pkl
Processing file: 581_YAHOO_id_31_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 581_YAHOO_id_31_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 152.89it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2331.71it/s]
Table1-only metrics completed in 0.50s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7469387755102036, 'Affiliation-P': 0.5960912052117264, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 581_YAHOO_id_31_Synthetic_tr_500_1st_893.csv: {'filename': '581_YAHOO_id_31_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7469387755102036, 'Affiliation-P': 0.5960912052117264, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 581_YAHOO_id_31_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/581_YAHOO_id_31_Synthetic_tr_500_1st_893_results.pkl
Processing file: 513_UCR_id_211_Medical_tr_9967_1st_26929.csv, Data shape: (39870, 1)
Label shape: (39870,)
Train data shape: (9967, 1), Test data shape: (29903, 1), Label test shape: (29903,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 513_UCR_id_211_Medical_tr_9967_1st_26929.csv...
Using CUDA device: 2
Data shape: (29903, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 29903, window_size 5000, adding 97 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 89.17it/s]
Original shapes - Output: (30000,), Label: (29903,)
Aligned shapes - Output: (29903,), Label: (29903,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  55%|█████▌    | 110/200 [00:01<00:00, 109.61it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 126.08it/s]
Table1-only metrics completed in 1.90s
Evaluation Result:  {'AUC-PR': 0.1308818110519085, 'AUC-ROC': 0.9822504439902155, 'VUS-PR': 0.10712870359921399, 'VUS-ROC': 0.9848334930674811, 'Standard-F1': 0.15311004784664334, 'Standard-Precision': 0.08938547486033493, 'Standard-Recall': 0.5333333333333244, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7569731437199254, 'Affiliation-P': 0.608975695010541, 'Affiliation-R': 1.0, 'F1_T': 0.1388888888886103, 'Precision_T': 0.08333333333333305, 'Recall_T': 0.4166666666666597}
Results for 513_UCR_id_211_Medical_tr_9967_1st_26929.csv: {'filename': '513_UCR_id_211_Medical_tr_9967_1st_26929.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 37, 'train_index': '9967', 'data_shape': '39870x1', 'output_length': 30000, 'label_length': 29903, 'aligned_length': 29903, 'AUC-PR': 0.1308818110519085, 'AUC-ROC': 0.9822504439902155, 'VUS-PR': 0.10712870359921399, 'VUS-ROC': 0.9848334930674811, 'Standard-F1': 0.15311004784664334, 'Standard-Precision': 0.08938547486033493, 'Standard-Recall': 0.5333333333333244, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7569731437199254, 'Affiliation-P': 0.608975695010541, 'Affiliation-R': 1.0, 'F1_T': 0.1388888888886103, 'Precision_T': 0.08333333333333305, 'Recall_T': 0.4166666666666597}
Logits metrics skipped for 513_UCR_id_211_Medical_tr_9967_1st_26929.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/513_UCR_id_211_Medical_tr_9967_1st_26929_results.pkl
Processing file: 556_YAHOO_id_6_WebService_tr_500_1st_359.csv, Data shape: (1441, 1)
Label shape: (1441,)
Train data shape: (500, 1), Test data shape: (941, 1), Label test shape: (941,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 556_YAHOO_id_6_WebService_tr_500_1st_359.csv...
Using CUDA device: 2
Data shape: (941, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 941, 1])
batched_mask shape: torch.Size([1, 941])

100%|██████████| 1/1 [00:00<00:00, 147.87it/s]
Original shapes - Output: (941,), Label: (941,)
Aligned shapes - Output: (941,), Label: (941,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1847.59it/s]
Table1-only metrics completed in 0.31s
Evaluation Result:  {'AUC-PR': 0.4390242717580024, 'AUC-ROC': 0.9438845822566753, 'VUS-PR': 0.5234497341697817, 'VUS-ROC': 0.9530348651685772, 'Standard-F1': 0.743119266054572, 'Standard-Precision': 0.5912408759124045, 'Standard-Recall': 0.9999999999999877, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9720070069346896, 'Affiliation-P': 0.9455385527836345, 'Affiliation-R': 1.0, 'F1_T': 0.7330316742076738, 'Precision_T': 0.5785714285714244, 'Recall_T': 0.9999999999999877}
Results for 556_YAHOO_id_6_WebService_tr_500_1st_359.csv: {'filename': '556_YAHOO_id_6_WebService_tr_500_1st_359.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1441x1', 'output_length': 941, 'label_length': 941, 'aligned_length': 941, 'AUC-PR': 0.4390242717580024, 'AUC-ROC': 0.9438845822566753, 'VUS-PR': 0.5234497341697817, 'VUS-ROC': 0.9530348651685772, 'Standard-F1': 0.743119266054572, 'Standard-Precision': 0.5912408759124045, 'Standard-Recall': 0.9999999999999877, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9720070069346896, 'Affiliation-P': 0.9455385527836345, 'Affiliation-R': 1.0, 'F1_T': 0.7330316742076738, 'Precision_T': 0.5785714285714244, 'Recall_T': 0.9999999999999877}
Logits metrics skipped for 556_YAHOO_id_6_WebService_tr_500_1st_359.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/556_YAHOO_id_6_WebService_tr_500_1st_359_results.pkl
Processing file: 227_MGAB_id_3_Synthetic_tr_25000_1st_44340.csv, Data shape: (100000, 1)
Label shape: (100000,)
Train data shape: (25000, 1), Test data shape: (75000, 1), Label test shape: (75000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 227_MGAB_id_3_Synthetic_tr_25000_1st_44340.csv...
Using CUDA device: 2
Data shape: (75000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([15, 5000, 1])
batched_mask shape: torch.Size([15, 5000])

100%|██████████| 1/1 [00:00<00:00, 79.81it/s]
Original shapes - Output: (75000,), Label: (75000,)
Aligned shapes - Output: (75000,), Label: (75000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  18%|█▊        | 36/200 [00:01<00:04, 35.61it/s]

Calculating Affiliation Metrics:  36%|███▌      | 72/200 [00:02<00:04, 26.48it/s]

Calculating Affiliation Metrics:  50%|█████     | 100/200 [00:04<00:04, 23.51it/s]

Calculating Affiliation Metrics:  62%|██████▎   | 125/200 [00:05<00:03, 22.59it/s]

Calculating Affiliation Metrics:  75%|███████▌  | 150/200 [00:06<00:02, 23.09it/s]

Calculating Affiliation Metrics:  92%|█████████▎| 185/200 [00:07<00:00, 26.58it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:07<00:00, 26.89it/s]
Table1-only metrics completed in 9.32s
Evaluation Result:  {'AUC-PR': 0.022744397084709735, 'AUC-ROC': 0.7368782754010695, 'VUS-PR': 0.021522587061035162, 'VUS-ROC': 0.8157131426284919, 'Standard-F1': 0.0637522768667329, 'Standard-Precision': 0.0389755011135857, 'Standard-Recall': 0.17499999999999913, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7077572883560243, 'Affiliation-P': 0.584189994212588, 'Affiliation-R': 0.8976212551542757, 'F1_T': 0.06363636363610643, 'Precision_T': 0.03749999999999997, 'Recall_T': 0.20999999999999897}
Results for 227_MGAB_id_3_Synthetic_tr_25000_1st_44340.csv: {'filename': '227_MGAB_id_3_Synthetic_tr_25000_1st_44340.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 48, 'train_index': '25000', 'data_shape': '100000x1', 'output_length': 75000, 'label_length': 75000, 'aligned_length': 75000, 'AUC-PR': 0.022744397084709735, 'AUC-ROC': 0.7368782754010695, 'VUS-PR': 0.021522587061035162, 'VUS-ROC': 0.8157131426284919, 'Standard-F1': 0.0637522768667329, 'Standard-Precision': 0.0389755011135857, 'Standard-Recall': 0.17499999999999913, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7077572883560243, 'Affiliation-P': 0.584189994212588, 'Affiliation-R': 0.8976212551542757, 'F1_T': 0.06363636363610643, 'Precision_T': 0.03749999999999997, 'Recall_T': 0.20999999999999897}
Logits metrics skipped for 227_MGAB_id_3_Synthetic_tr_25000_1st_44340.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/227_MGAB_id_3_Synthetic_tr_25000_1st_44340_results.pkl
Processing file: 402_UCR_id_100_Medical_tr_7500_1st_19510.csv, Data shape: (30000, 1)
Label shape: (30000,)
Train data shape: (7500, 1), Test data shape: (22500, 1), Label test shape: (22500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 402_UCR_id_100_Medical_tr_7500_1st_19510.csv...
Using CUDA device: 2
Data shape: (22500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 143.44it/s]
Original shapes - Output: (25000,), Label: (22500,)
Aligned shapes - Output: (22500,), Label: (22500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  86%|████████▌ | 172/200 [00:01<00:00, 171.75it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 184.73it/s]
Table1-only metrics completed in 1.21s
Evaluation Result:  {'AUC-PR': 0.008535620247842308, 'AUC-ROC': 0.7624705357142858, 'VUS-PR': 0.008672983843473637, 'VUS-ROC': 0.7647572687394374, 'Standard-F1': 0.022316684378294715, 'Standard-Precision': 0.011308562197092083, 'Standard-Recall': 0.8399999999999916, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7323130580523742, 'Affiliation-P': 0.6562276803118909, 'Affiliation-R': 0.8283555555555556, 'F1_T': 0.02231964926262822, 'Precision_T': 0.011310084825636191, 'Recall_T': 0.8399999999999916}
Results for 402_UCR_id_100_Medical_tr_7500_1st_19510.csv: {'filename': '402_UCR_id_100_Medical_tr_7500_1st_19510.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 6, 'train_index': '7500', 'data_shape': '30000x1', 'output_length': 25000, 'label_length': 22500, 'aligned_length': 22500, 'AUC-PR': 0.008535620247842308, 'AUC-ROC': 0.7624705357142858, 'VUS-PR': 0.008672983843473637, 'VUS-ROC': 0.7647572687394374, 'Standard-F1': 0.022316684378294715, 'Standard-Precision': 0.011308562197092083, 'Standard-Recall': 0.8399999999999916, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7323130580523742, 'Affiliation-P': 0.6562276803118909, 'Affiliation-R': 0.8283555555555556, 'F1_T': 0.02231964926262822, 'Precision_T': 0.011310084825636191, 'Recall_T': 0.8399999999999916}
Logits metrics skipped for 402_UCR_id_100_Medical_tr_7500_1st_19510.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/402_UCR_id_100_Medical_tr_7500_1st_19510_results.pkl
Processing file: 283_NEK_id_7_WebService_tr_500_1st_777.csv, Data shape: (1073, 1)
Label shape: (1073,)
Train data shape: (500, 1), Test data shape: (573, 1), Label test shape: (573,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 283_NEK_id_7_WebService_tr_500_1st_777.csv...
Using CUDA device: 2
Data shape: (573, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 573, 1])
batched_mask shape: torch.Size([1, 573])

100%|██████████| 1/1 [00:00<00:00, 158.49it/s]
Original shapes - Output: (573,), Label: (573,)
Aligned shapes - Output: (573,), Label: (573,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 3020.78it/s]
Table1-only metrics completed in 0.21s
Evaluation Result:  {'AUC-PR': 0.9174913171755024, 'AUC-ROC': 0.9642935982339955, 'VUS-PR': 0.9366073856572413, 'VUS-ROC': 0.9754309469400024, 'Standard-F1': 0.8544600938962137, 'Standard-Precision': 0.9784946236559036, 'Standard-Recall': 0.7583333333333271, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9925162343001357, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9851436500425095, 'F1_T': 0.8504672897191257, 'Precision_T': 0.9680851063829685, 'Recall_T': 0.7583333333333271}
Results for 283_NEK_id_7_WebService_tr_500_1st_777.csv: {'filename': '283_NEK_id_7_WebService_tr_500_1st_777.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1073x1', 'output_length': 573, 'label_length': 573, 'aligned_length': 573, 'AUC-PR': 0.9174913171755024, 'AUC-ROC': 0.9642935982339955, 'VUS-PR': 0.9366073856572413, 'VUS-ROC': 0.9754309469400024, 'Standard-F1': 0.8544600938962137, 'Standard-Precision': 0.9784946236559036, 'Standard-Recall': 0.7583333333333271, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9925162343001357, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9851436500425095, 'F1_T': 0.8504672897191257, 'Precision_T': 0.9680851063829685, 'Recall_T': 0.7583333333333271}
Logits metrics skipped for 283_NEK_id_7_WebService_tr_500_1st_777.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/283_NEK_id_7_WebService_tr_500_1st_777_results.pkl
Processing file: 494_UCR_id_192_Facility_tr_22500_1st_72150.csv, Data shape: (90000, 1)
Label shape: (90000,)
Train data shape: (22500, 1), Test data shape: (67500, 1), Label test shape: (67500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 494_UCR_id_192_Facility_tr_22500_1st_72150.csv...
Using CUDA device: 2
Data shape: (67500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 67500, window_size 5000, adding 2500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([14, 5000, 1])
batched_mask shape: torch.Size([14, 5000])

100%|██████████| 1/1 [00:00<00:00, 90.07it/s]
Original shapes - Output: (70000,), Label: (67500,)
Aligned shapes - Output: (67500,), Label: (67500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:01<00:02, 51.27it/s]

Calculating Affiliation Metrics:  58%|█████▊    | 116/200 [00:02<00:01, 58.41it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 70.41it/s]
Table1-only metrics completed in 4.63s
Evaluation Result:  {'AUC-PR': 0.04193665232362856, 'AUC-ROC': 0.8111720559941904, 'VUS-PR': 0.046665208242554115, 'VUS-ROC': 0.8140088705872501, 'Standard-F1': 0.16763378465471512, 'Standard-Precision': 0.10779436152570472, 'Standard-Recall': 0.37681159420289745, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7398367404907708, 'Affiliation-P': 0.5880407843137254, 'Affiliation-R': 0.9972706816961888, 'F1_T': 0.1602373887237098, 'Precision_T': 0.10074626865671635, 'Recall_T': 0.3913043478260858}
Results for 494_UCR_id_192_Facility_tr_22500_1st_72150.csv: {'filename': '494_UCR_id_192_Facility_tr_22500_1st_72150.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '22500', 'data_shape': '90000x1', 'output_length': 70000, 'label_length': 67500, 'aligned_length': 67500, 'AUC-PR': 0.04193665232362856, 'AUC-ROC': 0.8111720559941904, 'VUS-PR': 0.046665208242554115, 'VUS-ROC': 0.8140088705872501, 'Standard-F1': 0.16763378465471512, 'Standard-Precision': 0.10779436152570472, 'Standard-Recall': 0.37681159420289745, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7398367404907708, 'Affiliation-P': 0.5880407843137254, 'Affiliation-R': 0.9972706816961888, 'F1_T': 0.1602373887237098, 'Precision_T': 0.10074626865671635, 'Recall_T': 0.3913043478260858}
Logits metrics skipped for 494_UCR_id_192_Facility_tr_22500_1st_72150.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/494_UCR_id_192_Facility_tr_22500_1st_72150_results.pkl
Skipping file: 538_SMAP_id_8_Sensor_tr_2007_1st_3650.csv due to filter criteria.
Skipping file: 189_SMD_id_12_Facility_tr_5925_1st_17580.csv due to filter criteria.
Processing file: 657_YAHOO_id_107_WebService_tr_500_1st_1260.csv, Data shape: (1440, 1)
Label shape: (1440,)
Train data shape: (500, 1), Test data shape: (940, 1), Label test shape: (940,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 657_YAHOO_id_107_WebService_tr_500_1st_1260.csv...
Using CUDA device: 2
Data shape: (940, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 940, 1])
batched_mask shape: torch.Size([1, 940])

100%|██████████| 1/1 [00:00<00:00, 156.11it/s]
Original shapes - Output: (940,), Label: (940,)
Aligned shapes - Output: (940,), Label: (940,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2158.27it/s]
Table1-only metrics completed in 0.50s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7928598946962881, 'Affiliation-P': 0.6568085106382979, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 657_YAHOO_id_107_WebService_tr_500_1st_1260.csv: {'filename': '657_YAHOO_id_107_WebService_tr_500_1st_1260.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '1440x1', 'output_length': 940, 'label_length': 940, 'aligned_length': 940, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7928598946962881, 'Affiliation-P': 0.6568085106382979, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 657_YAHOO_id_107_WebService_tr_500_1st_1260.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/657_YAHOO_id_107_WebService_tr_500_1st_1260_results.pkl
Processing file: 304_UCR_id_2_Medical_tr_2499_1st_7000.csv, Data shape: (9998, 1)
Label shape: (9998,)
Train data shape: (2499, 1), Test data shape: (7499, 1), Label test shape: (7499,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 304_UCR_id_2_Medical_tr_2499_1st_7000.csv...
Using CUDA device: 2
Data shape: (7499, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 7499, window_size 5000, adding 2501 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.78it/s]
Original shapes - Output: (10000,), Label: (7499,)
Aligned shapes - Output: (7499,), Label: (7499,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 467.28it/s]
Table1-only metrics completed in 0.56s
Evaluation Result:  {'AUC-PR': 0.2306072056980967, 'AUC-ROC': 0.988862934622097, 'VUS-PR': 0.23878486570650825, 'VUS-ROC': 0.98913517472988, 'Standard-F1': 0.38427947598218803, 'Standard-Precision': 0.24581005586592042, 'Standard-Recall': 0.8799999999999824, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9256256465301029, 'Affiliation-P': 0.8624781549820678, 'Affiliation-R': 0.9987505000666757, 'F1_T': 0.38983050847423906, 'Precision_T': 0.24731182795698792, 'Recall_T': 0.9199999999999816}
Results for 304_UCR_id_2_Medical_tr_2499_1st_7000.csv: {'filename': '304_UCR_id_2_Medical_tr_2499_1st_7000.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 26, 'train_index': '2499', 'data_shape': '9998x1', 'output_length': 10000, 'label_length': 7499, 'aligned_length': 7499, 'AUC-PR': 0.2306072056980967, 'AUC-ROC': 0.988862934622097, 'VUS-PR': 0.23878486570650825, 'VUS-ROC': 0.98913517472988, 'Standard-F1': 0.38427947598218803, 'Standard-Precision': 0.24581005586592042, 'Standard-Recall': 0.8799999999999824, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9256256465301029, 'Affiliation-P': 0.8624781549820678, 'Affiliation-R': 0.9987505000666757, 'F1_T': 0.38983050847423906, 'Precision_T': 0.24731182795698792, 'Recall_T': 0.9199999999999816}
Logits metrics skipped for 304_UCR_id_2_Medical_tr_2499_1st_7000.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/304_UCR_id_2_Medical_tr_2499_1st_7000_results.pkl
Processing file: 664_YAHOO_id_114_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 664_YAHOO_id_114_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 154.54it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2008.75it/s]
Table1-only metrics completed in 0.95s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 664_YAHOO_id_114_Synthetic_tr_500_1st_658.csv: {'filename': '664_YAHOO_id_114_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 157, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 664_YAHOO_id_114_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/664_YAHOO_id_114_Synthetic_tr_500_1st_658_results.pkl
Processing file: 167_Stock_id_19_Finance_tr_500_1st_7.csv, Data shape: (10000, 1)
Label shape: (10000,)
Train data shape: (500, 1), Test data shape: (9500, 1), Label test shape: (9500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 167_Stock_id_19_Finance_tr_500_1st_7.csv...
Using CUDA device: 2
Data shape: (9500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.41it/s]
Original shapes - Output: (10000,), Label: (9500,)
Aligned shapes - Output: (9500,), Label: (9500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  10%|█         | 20/200 [00:01<00:09, 19.90it/s]

Calculating Affiliation Metrics:  20%|██        | 40/200 [00:03<00:15, 10.39it/s]

Calculating Affiliation Metrics:  26%|██▋       | 53/200 [00:05<00:15,  9.20it/s]

Calculating Affiliation Metrics:  32%|███▏      | 64/200 [00:06<00:15,  8.68it/s]

Calculating Affiliation Metrics:  37%|███▋      | 74/200 [00:07<00:14,  8.43it/s]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [00:09<00:13,  8.39it/s]

Calculating Affiliation Metrics:  46%|████▌     | 92/200 [00:10<00:12,  8.47it/s]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [00:11<00:11,  8.54it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 111/200 [00:12<00:10,  8.75it/s]

Calculating Affiliation Metrics:  60%|██████    | 121/200 [00:13<00:08,  8.96it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 131/200 [00:14<00:07,  8.93it/s]

Calculating Affiliation Metrics:  70%|███████   | 141/200 [00:15<00:06,  9.12it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 157/200 [00:16<00:03, 10.94it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:17<00:00, 11.50it/s]
Table1-only metrics completed in 24.92s
Evaluation Result:  {'AUC-PR': 0.0737448399822435, 'AUC-ROC': 0.5837934676082548, 'VUS-PR': 0.22883524881401465, 'VUS-ROC': 0.6231558346712925, 'Standard-F1': 0.12175261729322695, 'Standard-Precision': 0.07440758293838859, 'Standard-Recall': 0.3347547974413639, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.673247998416826, 'Affiliation-P': 0.5097799420526378, 'Affiliation-R': 0.9910382513661202, 'F1_T': 0.12189440993759029, 'Precision_T': 0.07451352634076884, 'Recall_T': 0.3347547974413639}
Results for 167_Stock_id_19_Finance_tr_500_1st_7.csv: {'filename': '167_Stock_id_19_Finance_tr_500_1st_7.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 10, 'train_index': '500', 'data_shape': '10000x1', 'output_length': 10000, 'label_length': 9500, 'aligned_length': 9500, 'AUC-PR': 0.0737448399822435, 'AUC-ROC': 0.5837934676082548, 'VUS-PR': 0.22883524881401465, 'VUS-ROC': 0.6231558346712925, 'Standard-F1': 0.12175261729322695, 'Standard-Precision': 0.07440758293838859, 'Standard-Recall': 0.3347547974413639, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.673247998416826, 'Affiliation-P': 0.5097799420526378, 'Affiliation-R': 0.9910382513661202, 'F1_T': 0.12189440993759029, 'Precision_T': 0.07451352634076884, 'Recall_T': 0.3347547974413639}
Logits metrics skipped for 167_Stock_id_19_Finance_tr_500_1st_7.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/167_Stock_id_19_Finance_tr_500_1st_7_results.pkl
Processing file: 376_UCR_id_74_Medical_tr_7500_1st_17210.csv, Data shape: (30001, 1)
Label shape: (30001,)
Train data shape: (7500, 1), Test data shape: (22501, 1), Label test shape: (22501,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 376_UCR_id_74_Medical_tr_7500_1st_17210.csv...
Using CUDA device: 2
Data shape: (22501, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22501, window_size 5000, adding 2499 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 144.50it/s]
Original shapes - Output: (25000,), Label: (22501,)
Aligned shapes - Output: (22501,), Label: (22501,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [00:01<00:00, 126.53it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 135.03it/s]
Table1-only metrics completed in 1.73s
Evaluation Result:  {'AUC-PR': 0.4881809686863274, 'AUC-ROC': 0.9437869137232194, 'VUS-PR': 0.24608529067854631, 'VUS-ROC': 0.9666186883576114, 'Standard-F1': 0.2608695652170927, 'Standard-Precision': 0.1592920353982294, 'Standard-Recall': 0.7199999999999855, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7521157446843275, 'Affiliation-P': 0.6027502057177292, 'Affiliation-R': 0.9998968934714013, 'F1_T': 0.2608695652170927, 'Precision_T': 0.1592920353982294, 'Recall_T': 0.7199999999999855}
Results for 376_UCR_id_74_Medical_tr_7500_1st_17210.csv: {'filename': '376_UCR_id_74_Medical_tr_7500_1st_17210.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 33, 'train_index': '7500', 'data_shape': '30001x1', 'output_length': 25000, 'label_length': 22501, 'aligned_length': 22501, 'AUC-PR': 0.4881809686863274, 'AUC-ROC': 0.9437869137232194, 'VUS-PR': 0.24608529067854631, 'VUS-ROC': 0.9666186883576114, 'Standard-F1': 0.2608695652170927, 'Standard-Precision': 0.1592920353982294, 'Standard-Recall': 0.7199999999999855, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7521157446843275, 'Affiliation-P': 0.6027502057177292, 'Affiliation-R': 0.9998968934714013, 'F1_T': 0.2608695652170927, 'Precision_T': 0.1592920353982294, 'Recall_T': 0.7199999999999855}
Logits metrics skipped for 376_UCR_id_74_Medical_tr_7500_1st_17210.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/376_UCR_id_74_Medical_tr_7500_1st_17210_results.pkl
Processing file: 512_UCR_id_210_HumanActivity_tr_1671_1st_2764.csv, Data shape: (6684, 1)
Label shape: (6684,)
Train data shape: (1671, 1), Test data shape: (5013, 1), Label test shape: (5013,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 512_UCR_id_210_HumanActivity_tr_1671_1st_2764.csv...
Using CUDA device: 2
Data shape: (5013, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5013, window_size 5000, adding 4987 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 153.25it/s]
Original shapes - Output: (10000,), Label: (5013,)
Aligned shapes - Output: (5013,), Label: (5013,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 524.64it/s]
Table1-only metrics completed in 0.83s
Evaluation Result:  {'AUC-PR': 0.037664100810142966, 'AUC-ROC': 0.44779666172388877, 'VUS-PR': 0.044132476220550956, 'VUS-ROC': 0.5005327637940794, 'Standard-F1': 0.10316875460564089, 'Standard-Precision': 0.054687499999999986, 'Standard-Recall': 0.9090909090909052, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7082489423397534, 'Affiliation-P': 0.5490182971083694, 'Affiliation-R': 0.9975729769266574, 'F1_T': 0.10300640709698476, 'Precision_T': 0.05461196759864122, 'Recall_T': 0.9047619047619009}
Results for 512_UCR_id_210_HumanActivity_tr_1671_1st_2764.csv: {'filename': '512_UCR_id_210_HumanActivity_tr_1671_1st_2764.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 115, 'train_index': '1671', 'data_shape': '6684x1', 'output_length': 10000, 'label_length': 5013, 'aligned_length': 5013, 'AUC-PR': 0.037664100810142966, 'AUC-ROC': 0.44779666172388877, 'VUS-PR': 0.044132476220550956, 'VUS-ROC': 0.5005327637940794, 'Standard-F1': 0.10316875460564089, 'Standard-Precision': 0.054687499999999986, 'Standard-Recall': 0.9090909090909052, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7082489423397534, 'Affiliation-P': 0.5490182971083694, 'Affiliation-R': 0.9975729769266574, 'F1_T': 0.10300640709698476, 'Precision_T': 0.05461196759864122, 'Recall_T': 0.9047619047619009}
Logits metrics skipped for 512_UCR_id_210_HumanActivity_tr_1671_1st_2764.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/512_UCR_id_210_HumanActivity_tr_1671_1st_2764_results.pkl
Processing file: 029_WSD_id_1_WebService_tr_4559_1st_10201.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 029_WSD_id_1_WebService_tr_4559_1st_10201.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.94it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 304.05it/s]
Table1-only metrics completed in 1.23s
Evaluation Result:  {'AUC-PR': 0.15906258118251573, 'AUC-ROC': 0.7338008486276013, 'VUS-PR': 0.06216082368193045, 'VUS-ROC': 0.9116687449877406, 'Standard-F1': 0.04907975460095825, 'Standard-Precision': 0.02919708029197059, 'Standard-Recall': 0.15384615384614794, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.837157978214942, 'Affiliation-P': 0.7202889013926903, 'Affiliation-R': 0.9992969668337075, 'F1_T': 0.04907975460095825, 'Precision_T': 0.02919708029197059, 'Recall_T': 0.15384615384614794}
Results for 029_WSD_id_1_WebService_tr_4559_1st_10201.csv: {'filename': '029_WSD_id_1_WebService_tr_4559_1st_10201.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.15906258118251573, 'AUC-ROC': 0.7338008486276013, 'VUS-PR': 0.06216082368193045, 'VUS-ROC': 0.9116687449877406, 'Standard-F1': 0.04907975460095825, 'Standard-Precision': 0.02919708029197059, 'Standard-Recall': 0.15384615384614794, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.837157978214942, 'Affiliation-P': 0.7202889013926903, 'Affiliation-R': 0.9992969668337075, 'F1_T': 0.04907975460095825, 'Precision_T': 0.02919708029197059, 'Recall_T': 0.15384615384614794}
Logits metrics skipped for 029_WSD_id_1_WebService_tr_4559_1st_10201.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/029_WSD_id_1_WebService_tr_4559_1st_10201_results.pkl
Processing file: 798_YAHOO_id_248_WebService_tr_500_1st_105.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 798_YAHOO_id_248_WebService_tr_500_1st_105.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 162.14it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 815.30it/s]
Table1-only metrics completed in 0.73s
Evaluation Result:  {'AUC-PR': 0.9999999999999999, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.9090909090903305, 'Standard-Precision': 0.8333333333332639, 'Standard-Recall': 0.9999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9993339409661355, 'Affiliation-P': 0.998668768610965, 'Affiliation-R': 1.0, 'F1_T': 0.9090909090903305, 'Precision_T': 0.8333333333332639, 'Recall_T': 0.9999999999999}
Results for 798_YAHOO_id_248_WebService_tr_500_1st_105.csv: {'filename': '798_YAHOO_id_248_WebService_tr_500_1st_105.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9999999999999999, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.9090909090903305, 'Standard-Precision': 0.8333333333332639, 'Standard-Recall': 0.9999999999999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9993339409661355, 'Affiliation-P': 0.998668768610965, 'Affiliation-R': 1.0, 'F1_T': 0.9090909090903305, 'Precision_T': 0.8333333333332639, 'Recall_T': 0.9999999999999}
Logits metrics skipped for 798_YAHOO_id_248_WebService_tr_500_1st_105.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/798_YAHOO_id_248_WebService_tr_500_1st_105_results.pkl
Processing file: 294_TODS_id_8_Synthetic_tr_500_1st_200.csv, Data shape: (5000, 1)
Label shape: (5000,)
Train data shape: (500, 1), Test data shape: (4500, 1), Label test shape: (4500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 294_TODS_id_8_Synthetic_tr_500_1st_200.csv...
Using CUDA device: 2
Data shape: (4500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 4500, 1])
batched_mask shape: torch.Size([1, 4500])

100%|██████████| 1/1 [00:00<00:00, 162.75it/s]
Original shapes - Output: (4500,), Label: (4500,)
Aligned shapes - Output: (4500,), Label: (4500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 373.44it/s]
Table1-only metrics completed in 3.14s
Evaluation Result:  {'AUC-PR': 0.9356369436390223, 'AUC-ROC': 0.9920114496988491, 'VUS-PR': 0.9777079606855196, 'VUS-ROC': 0.9978978038618639, 'Standard-F1': 0.8542199488486037, 'Standard-Precision': 0.8978494623655889, 'Standard-Recall': 0.8146341463414614, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9680329234325911, 'Affiliation-P': 0.9433538289925832, 'Affiliation-R': 0.9940379663927265, 'F1_T': 0.8502538071060974, 'Precision_T': 0.8862433862433838, 'Recall_T': 0.8170731707317053}
Results for 294_TODS_id_8_Synthetic_tr_500_1st_200.csv: {'filename': '294_TODS_id_8_Synthetic_tr_500_1st_200.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '500', 'data_shape': '5000x1', 'output_length': 4500, 'label_length': 4500, 'aligned_length': 4500, 'AUC-PR': 0.9356369436390223, 'AUC-ROC': 0.9920114496988491, 'VUS-PR': 0.9777079606855196, 'VUS-ROC': 0.9978978038618639, 'Standard-F1': 0.8542199488486037, 'Standard-Precision': 0.8978494623655889, 'Standard-Recall': 0.8146341463414614, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9680329234325911, 'Affiliation-P': 0.9433538289925832, 'Affiliation-R': 0.9940379663927265, 'F1_T': 0.8502538071060974, 'Precision_T': 0.8862433862433838, 'Recall_T': 0.8170731707317053}
Logits metrics skipped for 294_TODS_id_8_Synthetic_tr_500_1st_200.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/294_TODS_id_8_Synthetic_tr_500_1st_200_results.pkl
Processing file: 348_UCR_id_46_Facility_tr_9000_1st_12400.csv, Data shape: (36001, 1)
Label shape: (36001,)
Train data shape: (9000, 1), Test data shape: (27001, 1), Label test shape: (27001,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 348_UCR_id_46_Facility_tr_9000_1st_12400.csv...
Using CUDA device: 2
Data shape: (27001, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 27001, window_size 5000, adding 2999 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 139.46it/s]
Original shapes - Output: (30000,), Label: (27001,)
Aligned shapes - Output: (27001,), Label: (27001,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 136/200 [00:01<00:00, 135.59it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 149.45it/s]
Table1-only metrics completed in 2.49s
Evaluation Result:  {'AUC-PR': 0.027162383955169813, 'AUC-ROC': 0.5872360061651818, 'VUS-PR': 0.029032762193997443, 'VUS-ROC': 0.6245592673364748, 'Standard-F1': 0.09933774834415898, 'Standard-Precision': 0.05646817248459957, 'Standard-Recall': 0.4124999999999989, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7180447793572569, 'Affiliation-P': 0.5610407691954086, 'Affiliation-R': 0.9970686733824672, 'F1_T': 0.09909909909888767, 'Precision_T': 0.05631399317406142, 'Recall_T': 0.4124999999999989}
Results for 348_UCR_id_46_Facility_tr_9000_1st_12400.csv: {'filename': '348_UCR_id_46_Facility_tr_9000_1st_12400.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 188, 'train_index': '9000', 'data_shape': '36001x1', 'output_length': 30000, 'label_length': 27001, 'aligned_length': 27001, 'AUC-PR': 0.027162383955169813, 'AUC-ROC': 0.5872360061651818, 'VUS-PR': 0.029032762193997443, 'VUS-ROC': 0.6245592673364748, 'Standard-F1': 0.09933774834415898, 'Standard-Precision': 0.05646817248459957, 'Standard-Recall': 0.4124999999999989, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7180447793572569, 'Affiliation-P': 0.5610407691954086, 'Affiliation-R': 0.9970686733824672, 'F1_T': 0.09909909909888767, 'Precision_T': 0.05631399317406142, 'Recall_T': 0.4124999999999989}
Logits metrics skipped for 348_UCR_id_46_Facility_tr_9000_1st_12400.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/348_UCR_id_46_Facility_tr_9000_1st_12400_results.pkl
Processing file: 471_UCR_id_169_Facility_tr_10000_1st_35774.csv, Data shape: (40000, 1)
Label shape: (40000,)
Train data shape: (10000, 1), Test data shape: (30000, 1), Label test shape: (30000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 471_UCR_id_169_Facility_tr_10000_1st_35774.csv...
Using CUDA device: 2
Data shape: (30000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 138.42it/s]
Original shapes - Output: (30000,), Label: (30000,)
Aligned shapes - Output: (30000,), Label: (30000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  62%|██████▎   | 125/200 [00:01<00:00, 124.57it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 141.74it/s]
Table1-only metrics completed in 2.12s
Evaluation Result:  {'AUC-PR': 0.02738209722451071, 'AUC-ROC': 0.7802852842809364, 'VUS-PR': 0.0423041079472156, 'VUS-ROC': 0.8569544029311842, 'Standard-F1': 0.10051993067562298, 'Standard-Precision': 0.060796645702305946, 'Standard-Recall': 0.2899999999999971, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6922189408752507, 'Affiliation-P': 0.5294575168324791, 'Affiliation-R': 0.9994666666666666, 'F1_T': 0.09489051094860641, 'Precision_T': 0.058035714285714156, 'Recall_T': 0.2599999999999974}
Results for 471_UCR_id_169_Facility_tr_10000_1st_35774.csv: {'filename': '471_UCR_id_169_Facility_tr_10000_1st_35774.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 108, 'train_index': '10000', 'data_shape': '40000x1', 'output_length': 30000, 'label_length': 30000, 'aligned_length': 30000, 'AUC-PR': 0.02738209722451071, 'AUC-ROC': 0.7802852842809364, 'VUS-PR': 0.0423041079472156, 'VUS-ROC': 0.8569544029311842, 'Standard-F1': 0.10051993067562298, 'Standard-Precision': 0.060796645702305946, 'Standard-Recall': 0.2899999999999971, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6922189408752507, 'Affiliation-P': 0.5294575168324791, 'Affiliation-R': 0.9994666666666666, 'F1_T': 0.09489051094860641, 'Precision_T': 0.058035714285714156, 'Recall_T': 0.2599999999999974}
Logits metrics skipped for 471_UCR_id_169_Facility_tr_10000_1st_35774.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/471_UCR_id_169_Facility_tr_10000_1st_35774_results.pkl
Processing file: 232_MGAB_id_8_Synthetic_tr_25000_1st_32708.csv, Data shape: (100000, 1)
Label shape: (100000,)
Train data shape: (25000, 1), Test data shape: (75000, 1), Label test shape: (75000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 232_MGAB_id_8_Synthetic_tr_25000_1st_32708.csv...
Using CUDA device: 2
Data shape: (75000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([15, 5000, 1])
batched_mask shape: torch.Size([15, 5000])

100%|██████████| 1/1 [00:00<00:00, 78.87it/s]
Original shapes - Output: (75000,), Label: (75000,)
Aligned shapes - Output: (75000,), Label: (75000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  17%|█▋        | 34/200 [00:01<00:04, 33.31it/s]

Calculating Affiliation Metrics:  34%|███▍      | 68/200 [00:02<00:04, 26.63it/s]

Calculating Affiliation Metrics:  48%|████▊     | 96/200 [00:03<00:04, 24.81it/s]

Calculating Affiliation Metrics:  61%|██████    | 122/200 [00:04<00:03, 24.19it/s]

Calculating Affiliation Metrics:  74%|███████▍  | 149/200 [00:05<00:02, 24.91it/s]

Calculating Affiliation Metrics:  94%|█████████▎| 187/200 [00:06<00:00, 28.98it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:07<00:00, 28.50it/s]
Table1-only metrics completed in 8.95s
Evaluation Result:  {'AUC-PR': 0.02283519612571222, 'AUC-ROC': 0.8357771390374331, 'VUS-PR': 0.033338136104599646, 'VUS-ROC': 0.8540282904788785, 'Standard-F1': 0.0732738374822158, 'Standard-Precision': 0.04043545878693622, 'Standard-Recall': 0.38999999999999807, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.853720846629357, 'Affiliation-P': 0.7722363163664836, 'Affiliation-R': 0.9544300211298224, 'F1_T': 0.0728508984942393, 'Precision_T': 0.04034427111350186, 'Recall_T': 0.3749999999999981}
Results for 232_MGAB_id_8_Synthetic_tr_25000_1st_32708.csv: {'filename': '232_MGAB_id_8_Synthetic_tr_25000_1st_32708.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 49, 'train_index': '25000', 'data_shape': '100000x1', 'output_length': 75000, 'label_length': 75000, 'aligned_length': 75000, 'AUC-PR': 0.02283519612571222, 'AUC-ROC': 0.8357771390374331, 'VUS-PR': 0.033338136104599646, 'VUS-ROC': 0.8540282904788785, 'Standard-F1': 0.0732738374822158, 'Standard-Precision': 0.04043545878693622, 'Standard-Recall': 0.38999999999999807, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.853720846629357, 'Affiliation-P': 0.7722363163664836, 'Affiliation-R': 0.9544300211298224, 'F1_T': 0.0728508984942393, 'Precision_T': 0.04034427111350186, 'Recall_T': 0.3749999999999981}
Logits metrics skipped for 232_MGAB_id_8_Synthetic_tr_25000_1st_32708.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/232_MGAB_id_8_Synthetic_tr_25000_1st_32708_results.pkl
Processing file: 551_YAHOO_id_1_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 551_YAHOO_id_1_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 153.76it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2312.53it/s]
Table1-only metrics completed in 0.43s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984775989560672, 'Affiliation-P': 0.9969598262757872, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 551_YAHOO_id_1_Synthetic_tr_500_1st_893.csv: {'filename': '551_YAHOO_id_1_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 105, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984775989560672, 'Affiliation-P': 0.9969598262757872, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 551_YAHOO_id_1_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/551_YAHOO_id_1_Synthetic_tr_500_1st_893_results.pkl
Processing file: 795_YAHOO_id_245_WebService_tr_500_1st_84.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 795_YAHOO_id_245_WebService_tr_500_1st_84.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.13it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1014.82it/s]
Table1-only metrics completed in 0.48s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5882352941171626, 'Standard-Precision': 0.41666666666663194, 'Standard-Recall': 0.9999999999997999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9991181657848319, 'Affiliation-P': 0.9982378854625551, 'Affiliation-R': 1.0, 'F1_T': 0.5882352941171626, 'Precision_T': 0.41666666666663194, 'Recall_T': 0.9999999999997999}
Results for 795_YAHOO_id_245_WebService_tr_500_1st_84.csv: {'filename': '795_YAHOO_id_245_WebService_tr_500_1st_84.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5882352941171626, 'Standard-Precision': 0.41666666666663194, 'Standard-Recall': 0.9999999999997999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9991181657848319, 'Affiliation-P': 0.9982378854625551, 'Affiliation-R': 1.0, 'F1_T': 0.5882352941171626, 'Precision_T': 0.41666666666663194, 'Recall_T': 0.9999999999997999}
Logits metrics skipped for 795_YAHOO_id_245_WebService_tr_500_1st_84.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/795_YAHOO_id_245_WebService_tr_500_1st_84_results.pkl
Processing file: 502_UCR_id_200_Medical_tr_7487_1st_16390.csv, Data shape: (29950, 1)
Label shape: (29950,)
Train data shape: (7487, 1), Test data shape: (22463, 1), Label test shape: (22463,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 502_UCR_id_200_Medical_tr_7487_1st_16390.csv...
Using CUDA device: 2
Data shape: (22463, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22463, window_size 5000, adding 2537 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 150.22it/s]
Original shapes - Output: (25000,), Label: (22463,)
Aligned shapes - Output: (22463,), Label: (22463,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [00:01<00:00, 114.19it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 125.49it/s]
Table1-only metrics completed in 1.84s
Evaluation Result:  {'AUC-PR': 0.8350095860718039, 'AUC-ROC': 0.9949716934872732, 'VUS-PR': 0.3653989147606959, 'VUS-ROC': 0.9979197432241821, 'Standard-F1': 0.21960784313704554, 'Standard-Precision': 0.1244444444444439, 'Standard-Recall': 0.9333333333333023, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7481061874410742, 'Affiliation-P': 0.5975817066334058, 'Affiliation-R': 0.9999940643131668, 'F1_T': 0.21960784313704554, 'Precision_T': 0.1244444444444439, 'Recall_T': 0.9333333333333023}
Results for 502_UCR_id_200_Medical_tr_7487_1st_16390.csv: {'filename': '502_UCR_id_200_Medical_tr_7487_1st_16390.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 33, 'train_index': '7487', 'data_shape': '29950x1', 'output_length': 25000, 'label_length': 22463, 'aligned_length': 22463, 'AUC-PR': 0.8350095860718039, 'AUC-ROC': 0.9949716934872732, 'VUS-PR': 0.3653989147606959, 'VUS-ROC': 0.9979197432241821, 'Standard-F1': 0.21960784313704554, 'Standard-Precision': 0.1244444444444439, 'Standard-Recall': 0.9333333333333023, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7481061874410742, 'Affiliation-P': 0.5975817066334058, 'Affiliation-R': 0.9999940643131668, 'F1_T': 0.21960784313704554, 'Precision_T': 0.1244444444444439, 'Recall_T': 0.9333333333333023}
Logits metrics skipped for 502_UCR_id_200_Medical_tr_7487_1st_16390.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/502_UCR_id_200_Medical_tr_7487_1st_16390_results.pkl
Processing file: 346_UCR_id_44_Medical_tr_7456_1st_17390.csv, Data shape: (29826, 1)
Label shape: (29826,)
Train data shape: (7456, 1), Test data shape: (22370, 1), Label test shape: (22370,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 346_UCR_id_44_Medical_tr_7456_1st_17390.csv...
Using CUDA device: 2
Data shape: (22370, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 22370, window_size 5000, adding 2630 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([5, 5000, 1])
batched_mask shape: torch.Size([5, 5000])

100%|██████████| 1/1 [00:00<00:00, 90.50it/s]
Original shapes - Output: (25000,), Label: (22370,)
Aligned shapes - Output: (22370,), Label: (22370,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  54%|█████▍    | 108/200 [00:01<00:00, 107.22it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 117.40it/s]
Table1-only metrics completed in 1.95s
Evaluation Result:  {'AUC-PR': 0.026174407902863458, 'AUC-ROC': 0.35318345323741007, 'VUS-PR': 0.019390752156254548, 'VUS-ROC': 0.38916774710330737, 'Standard-F1': 0.05649717514077787, 'Standard-Precision': 0.04464285714285694, 'Standard-Recall': 0.07692307692307633, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6668414247426527, 'Affiliation-P': 0.5001993817538761, 'Affiliation-R': 0.9999889962518483, 'F1_T': 0.05649717514077787, 'Precision_T': 0.04464285714285694, 'Recall_T': 0.07692307692307633}
Results for 346_UCR_id_44_Medical_tr_7456_1st_17390.csv: {'filename': '346_UCR_id_44_Medical_tr_7456_1st_17390.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 33, 'train_index': '7456', 'data_shape': '29826x1', 'output_length': 25000, 'label_length': 22370, 'aligned_length': 22370, 'AUC-PR': 0.026174407902863458, 'AUC-ROC': 0.35318345323741007, 'VUS-PR': 0.019390752156254548, 'VUS-ROC': 0.38916774710330737, 'Standard-F1': 0.05649717514077787, 'Standard-Precision': 0.04464285714285694, 'Standard-Recall': 0.07692307692307633, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6668414247426527, 'Affiliation-P': 0.5001993817538761, 'Affiliation-R': 0.9999889962518483, 'F1_T': 0.05649717514077787, 'Precision_T': 0.04464285714285694, 'Recall_T': 0.07692307692307633}
Logits metrics skipped for 346_UCR_id_44_Medical_tr_7456_1st_17390.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/346_UCR_id_44_Medical_tr_7456_1st_17390_results.pkl
Skipping file: 209_SMD_id_32_Facility_tr_5925_1st_17580.csv due to filter criteria.
Processing file: 126_WSD_id_98_WebService_tr_657_1st_757.csv, Data shape: (18883, 1)
Label shape: (18883,)
Train data shape: (657, 1), Test data shape: (18226, 1), Label test shape: (18226,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 126_WSD_id_98_WebService_tr_657_1st_757.csv...
Using CUDA device: 2
Data shape: (18226, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 18226, window_size 5000, adding 1774 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.47it/s]
Original shapes - Output: (20000,), Label: (18226,)
Aligned shapes - Output: (18226,), Label: (18226,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  55%|█████▍    | 109/200 [00:01<00:00, 108.49it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 122.45it/s]
Table1-only metrics completed in 2.46s
Evaluation Result:  {'AUC-PR': 0.38952652745976213, 'AUC-ROC': 0.9282020678899505, 'VUS-PR': 0.368110720004322, 'VUS-ROC': 0.9422641730344903, 'Standard-F1': 0.4426877470350714, 'Standard-Precision': 0.44094488188976205, 'Standard-Recall': 0.4444444444444427, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8208652774358097, 'Affiliation-P': 0.8227080939499897, 'Affiliation-R': 0.8190306980666314, 'F1_T': 0.44137931034433814, 'Precision_T': 0.5245901639344234, 'Recall_T': 0.38095238095237943}
Results for 126_WSD_id_98_WebService_tr_657_1st_757.csv: {'filename': '126_WSD_id_98_WebService_tr_657_1st_757.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 12, 'train_index': '657', 'data_shape': '18883x1', 'output_length': 20000, 'label_length': 18226, 'aligned_length': 18226, 'AUC-PR': 0.38952652745976213, 'AUC-ROC': 0.9282020678899505, 'VUS-PR': 0.368110720004322, 'VUS-ROC': 0.9422641730344903, 'Standard-F1': 0.4426877470350714, 'Standard-Precision': 0.44094488188976205, 'Standard-Recall': 0.4444444444444427, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8208652774358097, 'Affiliation-P': 0.8227080939499897, 'Affiliation-R': 0.8190306980666314, 'F1_T': 0.44137931034433814, 'Precision_T': 0.5245901639344234, 'Recall_T': 0.38095238095237943}
Logits metrics skipped for 126_WSD_id_98_WebService_tr_657_1st_757.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/126_WSD_id_98_WebService_tr_657_1st_757_results.pkl
Processing file: 608_YAHOO_id_58_WebService_tr_500_1st_72.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 608_YAHOO_id_58_WebService_tr_500_1st_72.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 161.82it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 994.02it/s]
Table1-only metrics completed in 0.66s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9925262083569195, 'VUS-ROC': 0.9999680851922617, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.99638658455901, 'Affiliation-P': 0.9927991886409736, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Results for 608_YAHOO_id_58_WebService_tr_500_1st_72.csv: {'filename': '608_YAHOO_id_58_WebService_tr_500_1st_72.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.9925262083569195, 'VUS-ROC': 0.9999680851922617, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.99638658455901, 'Affiliation-P': 0.9927991886409736, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 608_YAHOO_id_58_WebService_tr_500_1st_72.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/608_YAHOO_id_58_WebService_tr_500_1st_72_results.pkl
Processing file: 576_YAHOO_id_26_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 576_YAHOO_id_26_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 159.76it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1875.40it/s]
Table1-only metrics completed in 0.76s
Evaluation Result:  {'AUC-PR': 0.8875, 'AUC-ROC': 0.9994547437295529, 'VUS-PR': 0.9901300613047874, 'VUS-ROC': 0.9999668696881604, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 576_YAHOO_id_26_Synthetic_tr_500_1st_658.csv: {'filename': '576_YAHOO_id_26_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 119, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.8875, 'AUC-ROC': 0.9994547437295529, 'VUS-PR': 0.9901300613047874, 'VUS-ROC': 0.9999668696881604, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 576_YAHOO_id_26_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/576_YAHOO_id_26_Synthetic_tr_500_1st_658_results.pkl
Processing file: 067_WSD_id_39_WebService_tr_4559_1st_13973.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 067_WSD_id_39_WebService_tr_4559_1st_13973.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.80it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 318.04it/s]
Table1-only metrics completed in 0.82s
Evaluation Result:  {'AUC-PR': 0.5770690391257107, 'AUC-ROC': 0.8197639225181599, 'VUS-PR': 0.30467188584403354, 'VUS-ROC': 0.8789683682225459, 'Standard-F1': 0.2918918918915045, 'Standard-Precision': 0.19708029197080149, 'Standard-Recall': 0.5624999999999882, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7220815895592111, 'Affiliation-P': 0.5651818932238932, 'Affiliation-R': 0.9995719699739221, 'F1_T': 0.2918918918915045, 'Precision_T': 0.19708029197080149, 'Recall_T': 0.5624999999999882}
Results for 067_WSD_id_39_WebService_tr_4559_1st_13973.csv: {'filename': '067_WSD_id_39_WebService_tr_4559_1st_13973.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.5770690391257107, 'AUC-ROC': 0.8197639225181599, 'VUS-PR': 0.30467188584403354, 'VUS-ROC': 0.8789683682225459, 'Standard-F1': 0.2918918918915045, 'Standard-Precision': 0.19708029197080149, 'Standard-Recall': 0.5624999999999882, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7220815895592111, 'Affiliation-P': 0.5651818932238932, 'Affiliation-R': 0.9995719699739221, 'F1_T': 0.2918918918915045, 'Precision_T': 0.19708029197080149, 'Recall_T': 0.5624999999999882}
Logits metrics skipped for 067_WSD_id_39_WebService_tr_4559_1st_13973.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/067_WSD_id_39_WebService_tr_4559_1st_13973_results.pkl
Processing file: 668_YAHOO_id_118_WebService_tr_500_1st_112.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 668_YAHOO_id_118_WebService_tr_500_1st_112.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.84it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1022.28it/s]
Table1-only metrics completed in 0.57s
Evaluation Result:  {'AUC-PR': 0.9999999999999998, 'AUC-ROC': 1.0, 'VUS-PR': 0.941044018530149, 'VUS-ROC': 0.9997485634402468, 'Standard-F1': 0.7368421052626148, 'Standard-Precision': 0.5833333333332847, 'Standard-Recall': 0.9999999999998571, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9827837758972857, 'Affiliation-P': 0.9661503155479059, 'Affiliation-R': 1.0, 'F1_T': 0.7368421052626148, 'Precision_T': 0.5833333333332847, 'Recall_T': 0.9999999999998571}
Results for 668_YAHOO_id_118_WebService_tr_500_1st_112.csv: {'filename': '668_YAHOO_id_118_WebService_tr_500_1st_112.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9999999999999998, 'AUC-ROC': 1.0, 'VUS-PR': 0.941044018530149, 'VUS-ROC': 0.9997485634402468, 'Standard-F1': 0.7368421052626148, 'Standard-Precision': 0.5833333333332847, 'Standard-Recall': 0.9999999999998571, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9827837758972857, 'Affiliation-P': 0.9661503155479059, 'Affiliation-R': 1.0, 'F1_T': 0.7368421052626148, 'Precision_T': 0.5833333333332847, 'Recall_T': 0.9999999999998571}
Logits metrics skipped for 668_YAHOO_id_118_WebService_tr_500_1st_112.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/668_YAHOO_id_118_WebService_tr_500_1st_112_results.pkl
Processing file: 705_YAHOO_id_155_WebService_tr_500_1st_1279.csv, Data shape: (1420, 1)
Label shape: (1420,)
Train data shape: (500, 1), Test data shape: (920, 1), Label test shape: (920,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 705_YAHOO_id_155_WebService_tr_500_1st_1279.csv...
Using CUDA device: 2
Data shape: (920, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 920, 1])
batched_mask shape: torch.Size([1, 920])

100%|██████████| 1/1 [00:00<00:00, 161.86it/s]
Original shapes - Output: (920,), Label: (920,)
Aligned shapes - Output: (920,), Label: (920,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1839.26it/s]
Table1-only metrics completed in 0.66s
Evaluation Result:  {'AUC-PR': 0.8277777777777778, 'AUC-ROC': 0.9974471188913202, 'VUS-PR': 0.9065549907511754, 'VUS-ROC': 0.999475133879746, 'Standard-F1': 0.624999999999453, 'Standard-Precision': 0.49999999999995, 'Standard-Recall': 0.8333333333331945, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9924084383040578, 'Affiliation-P': 0.9898929845422116, 'Affiliation-R': 0.9949367088607595, 'F1_T': 0.624999999999453, 'Precision_T': 0.49999999999995, 'Recall_T': 0.8333333333331945}
Results for 705_YAHOO_id_155_WebService_tr_500_1st_1279.csv: {'filename': '705_YAHOO_id_155_WebService_tr_500_1st_1279.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 124, 'train_index': '500', 'data_shape': '1420x1', 'output_length': 920, 'label_length': 920, 'aligned_length': 920, 'AUC-PR': 0.8277777777777778, 'AUC-ROC': 0.9974471188913202, 'VUS-PR': 0.9065549907511754, 'VUS-ROC': 0.999475133879746, 'Standard-F1': 0.624999999999453, 'Standard-Precision': 0.49999999999995, 'Standard-Recall': 0.8333333333331945, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9924084383040578, 'Affiliation-P': 0.9898929845422116, 'Affiliation-R': 0.9949367088607595, 'F1_T': 0.624999999999453, 'Precision_T': 0.49999999999995, 'Recall_T': 0.8333333333331945}
Logits metrics skipped for 705_YAHOO_id_155_WebService_tr_500_1st_1279.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/705_YAHOO_id_155_WebService_tr_500_1st_1279_results.pkl
Processing file: 006_NAB_id_6_Traffic_tr_2579_1st_5839.csv, Data shape: (10319, 1)
Label shape: (10319,)
Train data shape: (2579, 1), Test data shape: (7740, 1), Label test shape: (7740,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 006_NAB_id_6_Traffic_tr_2579_1st_5839.csv...
Using CUDA device: 2
Data shape: (7740, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 7740, window_size 5000, adding 2260 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.30it/s]
Original shapes - Output: (10000,), Label: (7740,)
Aligned shapes - Output: (7740,), Label: (7740,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 369.88it/s]
Table1-only metrics completed in 2.08s
Evaluation Result:  {'AUC-PR': 0.4512830324152621, 'AUC-ROC': 0.802859736952527, 'VUS-PR': 0.5406924575701928, 'VUS-ROC': 0.875598998015136, 'Standard-F1': 0.5117845117840376, 'Standard-Precision': 0.4163116250760801, 'Standard-Recall': 0.6640776699029121, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8806459573839333, 'Affiliation-P': 0.7954889662936305, 'Affiliation-R': 0.9862207626363497, 'F1_T': 0.5123595505613234, 'Precision_T': 0.4170731707317071, 'Recall_T': 0.6640776699029121}
Results for 006_NAB_id_6_Traffic_tr_2579_1st_5839.csv: {'filename': '006_NAB_id_6_Traffic_tr_2579_1st_5839.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '2579', 'data_shape': '10319x1', 'output_length': 10000, 'label_length': 7740, 'aligned_length': 7740, 'AUC-PR': 0.4512830324152621, 'AUC-ROC': 0.802859736952527, 'VUS-PR': 0.5406924575701928, 'VUS-ROC': 0.875598998015136, 'Standard-F1': 0.5117845117840376, 'Standard-Precision': 0.4163116250760801, 'Standard-Recall': 0.6640776699029121, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8806459573839333, 'Affiliation-P': 0.7954889662936305, 'Affiliation-R': 0.9862207626363497, 'F1_T': 0.5123595505613234, 'Precision_T': 0.4170731707317071, 'Recall_T': 0.6640776699029121}
Logits metrics skipped for 006_NAB_id_6_Traffic_tr_2579_1st_5839.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/006_NAB_id_6_Traffic_tr_2579_1st_5839_results.pkl
Processing file: 455_UCR_id_153_Facility_tr_50000_1st_163632.csv, Data shape: (298340, 1)
Label shape: (298340,)
Train data shape: (50000, 1), Test data shape: (248340, 1), Label test shape: (248340,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 455_UCR_id_153_Facility_tr_50000_1st_163632.csv...
Using CUDA device: 2
Data shape: (248340, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 248340, window_size 5000, adding 1660 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([50, 5000, 1])
batched_mask shape: torch.Size([50, 5000])

100%|██████████| 1/1 [00:00<00:00, 28.26it/s]
Original shapes - Output: (250000,), Label: (248340,)
Aligned shapes - Output: (248340,), Label: (248340,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  10%|▉         | 19/200 [00:01<00:10, 18.05it/s]

Calculating Affiliation Metrics:  19%|█▉        | 38/200 [00:02<00:10, 15.59it/s]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:03<00:10, 14.54it/s]

Calculating Affiliation Metrics:  34%|███▍      | 69/200 [00:04<00:09, 14.15it/s]

Calculating Affiliation Metrics:  42%|████▏     | 84/200 [00:05<00:08, 14.13it/s]

Calculating Affiliation Metrics:  50%|█████     | 100/200 [00:06<00:06, 14.43it/s]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [00:07<00:05, 14.45it/s]

Calculating Affiliation Metrics:  66%|██████▋   | 133/200 [00:08<00:04, 15.34it/s]

Calculating Affiliation Metrics:  76%|███████▋  | 153/200 [00:09<00:02, 16.58it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 177/200 [00:10<00:01, 18.58it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:11<00:00, 17.18it/s]
Table1-only metrics completed in 16.69s
Evaluation Result:  {'AUC-PR': 0.03059752799152799, 'AUC-ROC': 0.8457496308291186, 'VUS-PR': 0.03208870449906984, 'VUS-ROC': 0.8467632478323083, 'Standard-F1': 0.10322370563302581, 'Standard-Precision': 0.06440471353108491, 'Standard-Recall': 0.2598360655737703, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6929087368482086, 'Affiliation-P': 0.5306785670635068, 'Affiliation-R': 0.998000855186636, 'F1_T': 0.10308943089399088, 'Precision_T': 0.06430020283975658, 'Recall_T': 0.2598360655737703}
Results for 455_UCR_id_153_Facility_tr_50000_1st_163632.csv: {'filename': '455_UCR_id_153_Facility_tr_50000_1st_163632.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 64, 'train_index': '50000', 'data_shape': '298340x1', 'output_length': 250000, 'label_length': 248340, 'aligned_length': 248340, 'AUC-PR': 0.03059752799152799, 'AUC-ROC': 0.8457496308291186, 'VUS-PR': 0.03208870449906984, 'VUS-ROC': 0.8467632478323083, 'Standard-F1': 0.10322370563302581, 'Standard-Precision': 0.06440471353108491, 'Standard-Recall': 0.2598360655737703, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6929087368482086, 'Affiliation-P': 0.5306785670635068, 'Affiliation-R': 0.998000855186636, 'F1_T': 0.10308943089399088, 'Precision_T': 0.06430020283975658, 'Recall_T': 0.2598360655737703}
Logits metrics skipped for 455_UCR_id_153_Facility_tr_50000_1st_163632.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/455_UCR_id_153_Facility_tr_50000_1st_163632_results.pkl
Processing file: 574_YAHOO_id_24_WebService_tr_500_1st_202.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 574_YAHOO_id_24_WebService_tr_500_1st_202.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 151.53it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Table1-only metrics completed in 0.01s
Evaluation Result:  {'AUC-PR': 0.0, 'AUC-ROC': nan, 'VUS-PR': 0.0, 'VUS-ROC': 0.0, 'Standard-F1': 0.0, 'Standard-Precision': 0.0, 'Standard-Recall': 0.0, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.0, 'Affiliation-P': 0.0, 'Affiliation-R': 0.0, 'F1_T': 0.0, 'Precision_T': 0.0, 'Recall_T': 0.0}
Results for 574_YAHOO_id_24_WebService_tr_500_1st_202.csv: {'filename': '574_YAHOO_id_24_WebService_tr_500_1st_202.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.0, 'AUC-ROC': nan, 'VUS-PR': 0.0, 'VUS-ROC': 0.0, 'Standard-F1': 0.0, 'Standard-Precision': 0.0, 'Standard-Recall': 0.0, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.0, 'Affiliation-P': 0.0, 'Affiliation-R': 0.0, 'F1_T': 0.0, 'Precision_T': 0.0, 'Recall_T': 0.0}
Logits metrics skipped for 574_YAHOO_id_24_WebService_tr_500_1st_202.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/574_YAHOO_id_24_WebService_tr_500_1st_202_results.pkl
Processing file: 394_UCR_id_92_Medical_tr_9967_1st_28995.csv, Data shape: (39870, 1)
Label shape: (39870,)
Train data shape: (9967, 1), Test data shape: (29903, 1), Label test shape: (29903,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 394_UCR_id_92_Medical_tr_9967_1st_28995.csv...
Using CUDA device: 2
Data shape: (29903, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 29903, window_size 5000, adding 97 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 143.02it/s]
Original shapes - Output: (30000,), Label: (29903,)
Aligned shapes - Output: (29903,), Label: (29903,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  51%|█████     | 102/200 [00:01<00:00, 101.63it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 125.96it/s]
Table1-only metrics completed in 1.89s
Evaluation Result:  {'AUC-PR': 0.8658015109476159, 'AUC-ROC': 0.9980321783561981, 'VUS-PR': 0.6233803525951289, 'VUS-ROC': 0.9976348937799926, 'Standard-F1': 0.43076923076887347, 'Standard-Precision': 0.279999999999999, 'Standard-Recall': 0.933333333333323, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8804906380387894, 'Affiliation-P': 0.7865052854117939, 'Affiliation-R': 0.9999864376298179, 'F1_T': 0.43076923076887347, 'Precision_T': 0.279999999999999, 'Recall_T': 0.933333333333323}
Results for 394_UCR_id_92_Medical_tr_9967_1st_28995.csv: {'filename': '394_UCR_id_92_Medical_tr_9967_1st_28995.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 33, 'train_index': '9967', 'data_shape': '39870x1', 'output_length': 30000, 'label_length': 29903, 'aligned_length': 29903, 'AUC-PR': 0.8658015109476159, 'AUC-ROC': 0.9980321783561981, 'VUS-PR': 0.6233803525951289, 'VUS-ROC': 0.9976348937799926, 'Standard-F1': 0.43076923076887347, 'Standard-Precision': 0.279999999999999, 'Standard-Recall': 0.933333333333323, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8804906380387894, 'Affiliation-P': 0.7865052854117939, 'Affiliation-R': 0.9999864376298179, 'F1_T': 0.43076923076887347, 'Precision_T': 0.279999999999999, 'Recall_T': 0.933333333333323}
Logits metrics skipped for 394_UCR_id_92_Medical_tr_9967_1st_28995.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/394_UCR_id_92_Medical_tr_9967_1st_28995_results.pkl
Processing file: 385_UCR_id_83_Environment_tr_2046_1st_6006.csv, Data shape: (8184, 1)
Label shape: (8184,)
Train data shape: (2046, 1), Test data shape: (6138, 1), Label test shape: (6138,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 385_UCR_id_83_Environment_tr_2046_1st_6006.csv...
Using CUDA device: 2
Data shape: (6138, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6138, window_size 5000, adding 3862 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.41it/s]
Original shapes - Output: (10000,), Label: (6138,)
Aligned shapes - Output: (6138,), Label: (6138,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 498.92it/s]
Table1-only metrics completed in 0.52s
Evaluation Result:  {'AUC-PR': 0.21070440066958193, 'AUC-ROC': 0.9612616310892172, 'VUS-PR': 0.16313858275403276, 'VUS-ROC': 0.9669950551349159, 'Standard-F1': 0.21587301587275617, 'Standard-Precision': 0.12734082397003696, 'Standard-Recall': 0.7083333333333186, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7380223102051031, 'Affiliation-P': 0.5848140709405498, 'Affiliation-R': 1.0, 'F1_T': 0.2118380062302739, 'Precision_T': 0.12454212454212407, 'Recall_T': 0.7083333333333186}
Results for 385_UCR_id_83_Environment_tr_2046_1st_6006.csv: {'filename': '385_UCR_id_83_Environment_tr_2046_1st_6006.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '2046', 'data_shape': '8184x1', 'output_length': 10000, 'label_length': 6138, 'aligned_length': 6138, 'AUC-PR': 0.21070440066958193, 'AUC-ROC': 0.9612616310892172, 'VUS-PR': 0.16313858275403276, 'VUS-ROC': 0.9669950551349159, 'Standard-F1': 0.21587301587275617, 'Standard-Precision': 0.12734082397003696, 'Standard-Recall': 0.7083333333333186, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7380223102051031, 'Affiliation-P': 0.5848140709405498, 'Affiliation-R': 1.0, 'F1_T': 0.2118380062302739, 'Precision_T': 0.12454212454212407, 'Recall_T': 0.7083333333333186}
Logits metrics skipped for 385_UCR_id_83_Environment_tr_2046_1st_6006.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/385_UCR_id_83_Environment_tr_2046_1st_6006_results.pkl
Skipping file: 249_SVDB_id_13_Medical_tr_50000_1st_54982.csv due to filter criteria.
Processing file: 055_WSD_id_27_WebService_tr_3768_1st_3868.csv, Data shape: (18191, 1)
Label shape: (18191,)
Train data shape: (3768, 1), Test data shape: (14423, 1), Label test shape: (14423,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 055_WSD_id_27_WebService_tr_3768_1st_3868.csv...
Using CUDA device: 2
Data shape: (14423, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 14423, window_size 5000, adding 577 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.55it/s]
Original shapes - Output: (15000,), Label: (14423,)
Aligned shapes - Output: (14423,), Label: (14423,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 216.41it/s]
Table1-only metrics completed in 3.43s
Evaluation Result:  {'AUC-PR': 0.25449077107717294, 'AUC-ROC': 0.8656275424418289, 'VUS-PR': 0.37142371322009454, 'VUS-ROC': 0.9756609935697319, 'Standard-F1': 0.22580645161241578, 'Standard-Precision': 0.19310344827586073, 'Standard-Recall': 0.27184466019417214, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9241917770677711, 'Affiliation-P': 0.9825521809241431, 'Affiliation-R': 0.8723755091707072, 'F1_T': 0.22580645161241578, 'Precision_T': 0.19310344827586073, 'Recall_T': 0.27184466019417214}
Results for 055_WSD_id_27_WebService_tr_3768_1st_3868.csv: {'filename': '055_WSD_id_27_WebService_tr_3768_1st_3868.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '3768', 'data_shape': '18191x1', 'output_length': 15000, 'label_length': 14423, 'aligned_length': 14423, 'AUC-PR': 0.25449077107717294, 'AUC-ROC': 0.8656275424418289, 'VUS-PR': 0.37142371322009454, 'VUS-ROC': 0.9756609935697319, 'Standard-F1': 0.22580645161241578, 'Standard-Precision': 0.19310344827586073, 'Standard-Recall': 0.27184466019417214, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9241917770677711, 'Affiliation-P': 0.9825521809241431, 'Affiliation-R': 0.8723755091707072, 'F1_T': 0.22580645161241578, 'Precision_T': 0.19310344827586073, 'Recall_T': 0.27184466019417214}
Logits metrics skipped for 055_WSD_id_27_WebService_tr_3768_1st_3868.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/055_WSD_id_27_WebService_tr_3768_1st_3868_results.pkl
Processing file: 607_YAHOO_id_57_WebService_tr_500_1st_426.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 607_YAHOO_id_57_WebService_tr_500_1st_426.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 159.88it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 957.37it/s]
Table1-only metrics completed in 0.78s
Evaluation Result:  {'AUC-PR': 0.7357485750789882, 'AUC-ROC': 0.9154677657671669, 'VUS-PR': 0.7550755695996468, 'VUS-ROC': 0.9248104910936736, 'Standard-F1': 0.6428571428566199, 'Standard-Precision': 0.5294117647058513, 'Standard-Recall': 0.8181818181817438, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.895106379504596, 'Affiliation-P': 0.9879967558799676, 'Affiliation-R': 0.8181818181818182, 'F1_T': 0.6206896551719001, 'Precision_T': 0.49999999999997224, 'Recall_T': 0.8181818181817438}
Results for 607_YAHOO_id_57_WebService_tr_500_1st_426.csv: {'filename': '607_YAHOO_id_57_WebService_tr_500_1st_426.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.7357485750789882, 'AUC-ROC': 0.9154677657671669, 'VUS-PR': 0.7550755695996468, 'VUS-ROC': 0.9248104910936736, 'Standard-F1': 0.6428571428566199, 'Standard-Precision': 0.5294117647058513, 'Standard-Recall': 0.8181818181817438, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.895106379504596, 'Affiliation-P': 0.9879967558799676, 'Affiliation-R': 0.8181818181818182, 'F1_T': 0.6206896551719001, 'Precision_T': 0.49999999999997224, 'Recall_T': 0.8181818181817438}
Logits metrics skipped for 607_YAHOO_id_57_WebService_tr_500_1st_426.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/607_YAHOO_id_57_WebService_tr_500_1st_426_results.pkl
Skipping file: 848_OPPORTUNITY_id_7_HumanActivity_tr_1477_1st_1577.csv due to filter criteria.
Skipping file: 835_Exathlon_id_26_Facility_tr_10766_1st_12590.csv due to filter criteria.
Skipping file: 857_OPPORTUNITY_id_16_HumanActivity_tr_500_1st_566.csv due to filter criteria.
Processing file: 578_YAHOO_id_28_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 578_YAHOO_id_28_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 160.78it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1402.07it/s]
Table1-only metrics completed in 0.39s
Evaluation Result:  {'AUC-PR': 0.8643658810325479, 'AUC-ROC': 0.9987816764132554, 'VUS-PR': 0.9754756358460717, 'VUS-ROC': 0.9998240929858491, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984359113984375, 'Affiliation-P': 0.9969624510347107, 'Affiliation-R': 0.9999137336093858, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 578_YAHOO_id_28_Synthetic_tr_500_1st_623.csv: {'filename': '578_YAHOO_id_28_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 35, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.8643658810325479, 'AUC-ROC': 0.9987816764132554, 'VUS-PR': 0.9754756358460717, 'VUS-ROC': 0.9998240929858491, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984359113984375, 'Affiliation-P': 0.9969624510347107, 'Affiliation-R': 0.9999137336093858, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 578_YAHOO_id_28_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/578_YAHOO_id_28_Synthetic_tr_500_1st_623_results.pkl
Processing file: 416_UCR_id_114_Medical_tr_50000_1st_91200.csv, Data shape: (211502, 1)
Label shape: (211502,)
Train data shape: (50000, 1), Test data shape: (161502, 1), Label test shape: (161502,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 416_UCR_id_114_Medical_tr_50000_1st_91200.csv...
Using CUDA device: 2
Data shape: (161502, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 161502, window_size 5000, adding 3498 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([33, 5000, 1])
batched_mask shape: torch.Size([33, 5000])

100%|██████████| 1/1 [00:00<00:00, 40.93it/s]
Original shapes - Output: (165000,), Label: (161502,)
Aligned shapes - Output: (161502,), Label: (161502,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  14%|█▎        | 27/200 [00:01<00:06, 26.18it/s]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:02<00:05, 24.79it/s]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [00:03<00:04, 24.52it/s]

Calculating Affiliation Metrics:  52%|█████▎    | 105/200 [00:04<00:03, 25.00it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 135/200 [00:05<00:02, 26.56it/s]

Calculating Affiliation Metrics:  84%|████████▍ | 169/200 [00:06<00:01, 28.90it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:06<00:00, 29.20it/s]
Table1-only metrics completed in 11.58s
Evaluation Result:  {'AUC-PR': 0.01965838204927775, 'AUC-ROC': 0.9281980348070211, 'VUS-PR': 0.02065013723154514, 'VUS-ROC': 0.9290308706937495, 'Standard-F1': 0.050408997954960406, 'Standard-Precision': 0.025865687303252886, 'Standard-Recall': 0.985999999999998, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.719142671229455, 'Affiliation-P': 0.5614541565841203, 'Affiliation-R': 1.0, 'F1_T': 0.0502727782592681, 'Precision_T': 0.02579396222466384, 'Recall_T': 0.985999999999998}
Results for 416_UCR_id_114_Medical_tr_50000_1st_91200.csv: {'filename': '416_UCR_id_114_Medical_tr_50000_1st_91200.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '50000', 'data_shape': '211502x1', 'output_length': 165000, 'label_length': 161502, 'aligned_length': 161502, 'AUC-PR': 0.01965838204927775, 'AUC-ROC': 0.9281980348070211, 'VUS-PR': 0.02065013723154514, 'VUS-ROC': 0.9290308706937495, 'Standard-F1': 0.050408997954960406, 'Standard-Precision': 0.025865687303252886, 'Standard-Recall': 0.985999999999998, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.719142671229455, 'Affiliation-P': 0.5614541565841203, 'Affiliation-R': 1.0, 'F1_T': 0.0502727782592681, 'Precision_T': 0.02579396222466384, 'Recall_T': 0.985999999999998}
Logits metrics skipped for 416_UCR_id_114_Medical_tr_50000_1st_91200.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/416_UCR_id_114_Medical_tr_50000_1st_91200_results.pkl
Skipping file: 259_TAO_id_3_Environment_tr_500_1st_7.csv due to filter criteria.
Processing file: 622_YAHOO_id_72_WebService_tr_500_1st_421.csv, Data shape: (1461, 1)
Label shape: (1461,)
Train data shape: (500, 1), Test data shape: (961, 1), Label test shape: (961,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 622_YAHOO_id_72_WebService_tr_500_1st_421.csv...
Using CUDA device: 2
Data shape: (961, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 961, 1])
batched_mask shape: torch.Size([1, 961])

100%|██████████| 1/1 [00:00<00:00, 152.97it/s]
Original shapes - Output: (961,), Label: (961,)
Aligned shapes - Output: (961,), Label: (961,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2425.25it/s]
Table1-only metrics completed in 0.17s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8468922486201098, 'Affiliation-P': 0.7344432882414151, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 622_YAHOO_id_72_WebService_tr_500_1st_421.csv: {'filename': '622_YAHOO_id_72_WebService_tr_500_1st_421.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1461x1', 'output_length': 961, 'label_length': 961, 'aligned_length': 961, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8468922486201098, 'Affiliation-P': 0.7344432882414151, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 622_YAHOO_id_72_WebService_tr_500_1st_421.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/622_YAHOO_id_72_WebService_tr_500_1st_421_results.pkl
Processing file: 784_YAHOO_id_234_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 784_YAHOO_id_234_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 163.64it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1936.27it/s]
Table1-only metrics completed in 0.88s
Evaluation Result:  {'AUC-PR': 0.9253086419753087, 'AUC-ROC': 0.9991471734892787, 'VUS-PR': 0.994505992222397, 'VUS-ROC': 0.9999493994853684, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 784_YAHOO_id_234_Synthetic_tr_500_1st_623.csv: {'filename': '784_YAHOO_id_234_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 131, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9253086419753087, 'AUC-ROC': 0.9991471734892787, 'VUS-PR': 0.994505992222397, 'VUS-ROC': 0.9999493994853684, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 784_YAHOO_id_234_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/784_YAHOO_id_234_Synthetic_tr_500_1st_623_results.pkl
Processing file: 666_YAHOO_id_116_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 666_YAHOO_id_116_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 159.78it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2247.55it/s]
Table1-only metrics completed in 0.71s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9684139784946231, 'Affiliation-P': 0.9387622149837134, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 666_YAHOO_id_116_Synthetic_tr_500_1st_893.csv: {'filename': '666_YAHOO_id_116_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 187, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9684139784946231, 'Affiliation-P': 0.9387622149837134, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 666_YAHOO_id_116_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/666_YAHOO_id_116_Synthetic_tr_500_1st_893_results.pkl
Processing file: 328_UCR_id_26_Medical_tr_1913_1st_3474.csv, Data shape: (7654, 1)
Label shape: (7654,)
Train data shape: (1913, 1), Test data shape: (5741, 1), Label test shape: (5741,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 328_UCR_id_26_Medical_tr_1913_1st_3474.csv...
Using CUDA device: 2
Data shape: (5741, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5741, window_size 5000, adding 4259 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.53it/s]
Original shapes - Output: (10000,), Label: (5741,)
Aligned shapes - Output: (5741,), Label: (5741,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 767.99it/s]
Table1-only metrics completed in 0.86s
Evaluation Result:  {'AUC-PR': 0.5686667503309416, 'AUC-ROC': 0.9696666782162779, 'VUS-PR': 0.7261604321025472, 'VUS-ROC': 0.9858924029758874, 'Standard-F1': 0.5234567901229817, 'Standard-Precision': 0.4239999999999983, 'Standard-Recall': 0.6838709677419311, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9919584256695997, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9840451534238726, 'F1_T': 0.5206812652063403, 'Precision_T': 0.41796874999999833, 'Recall_T': 0.6903225806451568}
Results for 328_UCR_id_26_Medical_tr_1913_1st_3474.csv: {'filename': '328_UCR_id_26_Medical_tr_1913_1st_3474.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 154, 'train_index': '1913', 'data_shape': '7654x1', 'output_length': 10000, 'label_length': 5741, 'aligned_length': 5741, 'AUC-PR': 0.5686667503309416, 'AUC-ROC': 0.9696666782162779, 'VUS-PR': 0.7261604321025472, 'VUS-ROC': 0.9858924029758874, 'Standard-F1': 0.5234567901229817, 'Standard-Precision': 0.4239999999999983, 'Standard-Recall': 0.6838709677419311, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9919584256695997, 'Affiliation-P': 1.0, 'Affiliation-R': 0.9840451534238726, 'F1_T': 0.5206812652063403, 'Precision_T': 0.41796874999999833, 'Recall_T': 0.6903225806451568}
Logits metrics skipped for 328_UCR_id_26_Medical_tr_1913_1st_3474.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/328_UCR_id_26_Medical_tr_1913_1st_3474_results.pkl
Processing file: 461_UCR_id_159_Medical_tr_45096_1st_131200.csv, Data shape: (180386, 1)
Label shape: (180386,)
Train data shape: (45096, 1), Test data shape: (135290, 1), Label test shape: (135290,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 461_UCR_id_159_Medical_tr_45096_1st_131200.csv...
Using CUDA device: 2
Data shape: (135290, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 135290, window_size 5000, adding 4710 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([28, 5000, 1])
batched_mask shape: torch.Size([28, 5000])

100%|██████████| 1/1 [00:00<00:00, 51.57it/s]
Original shapes - Output: (140000,), Label: (135290,)
Aligned shapes - Output: (135290,), Label: (135290,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  14%|█▍        | 28/200 [00:01<00:06, 27.91it/s]

Calculating Affiliation Metrics:  28%|██▊       | 56/200 [00:02<00:05, 25.01it/s]

Calculating Affiliation Metrics:  41%|████      | 82/200 [00:03<00:04, 24.81it/s]

Calculating Affiliation Metrics:  55%|█████▍    | 109/200 [00:04<00:03, 25.61it/s]

Calculating Affiliation Metrics:  70%|███████   | 141/200 [00:05<00:02, 27.70it/s]

Calculating Affiliation Metrics:  93%|█████████▎| 186/200 [00:06<00:00, 33.29it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:06<00:00, 30.77it/s]
Table1-only metrics completed in 10.53s
Evaluation Result:  {'AUC-PR': 0.006746892076196716, 'AUC-ROC': 0.7481160249276653, 'VUS-PR': 0.007903332350770006, 'VUS-ROC': 0.7688953134989025, 'Standard-F1': 0.01697741397602477, 'Standard-Precision': 0.008597308223734711, 'Standard-Recall': 0.6719999999999986, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6737767767301368, 'Affiliation-P': 0.5082665027928845, 'Affiliation-R': 0.9991303939685122, 'F1_T': 0.016922343016370297, 'Precision_T': 0.008568740623967046, 'Recall_T': 0.6739999999999986}
Results for 461_UCR_id_159_Medical_tr_45096_1st_131200.csv: {'filename': '461_UCR_id_159_Medical_tr_45096_1st_131200.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 100, 'train_index': '45096', 'data_shape': '180386x1', 'output_length': 140000, 'label_length': 135290, 'aligned_length': 135290, 'AUC-PR': 0.006746892076196716, 'AUC-ROC': 0.7481160249276653, 'VUS-PR': 0.007903332350770006, 'VUS-ROC': 0.7688953134989025, 'Standard-F1': 0.01697741397602477, 'Standard-Precision': 0.008597308223734711, 'Standard-Recall': 0.6719999999999986, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6737767767301368, 'Affiliation-P': 0.5082665027928845, 'Affiliation-R': 0.9991303939685122, 'F1_T': 0.016922343016370297, 'Precision_T': 0.008568740623967046, 'Recall_T': 0.6739999999999986}
Logits metrics skipped for 461_UCR_id_159_Medical_tr_45096_1st_131200.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/461_UCR_id_159_Medical_tr_45096_1st_131200_results.pkl
Processing file: 465_UCR_id_163_HumanActivity_tr_1671_1st_5920.csv, Data shape: (6684, 1)
Label shape: (6684,)
Train data shape: (1671, 1), Test data shape: (5013, 1), Label test shape: (5013,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 465_UCR_id_163_HumanActivity_tr_1671_1st_5920.csv...
Using CUDA device: 2
Data shape: (5013, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5013, window_size 5000, adding 4987 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 158.86it/s]
Original shapes - Output: (10000,), Label: (5013,)
Aligned shapes - Output: (5013,), Label: (5013,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 539.11it/s]
Table1-only metrics completed in 0.80s
Evaluation Result:  {'AUC-PR': 0.9665694070928227, 'AUC-ROC': 0.9996955037189602, 'VUS-PR': 0.9785638152364249, 'VUS-ROC': 0.9998480053238209, 'Standard-F1': 0.9579831932767947, 'Standard-Precision': 0.9499999999999842, 'Standard-Recall': 0.9661016949152378, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.999858981975477, 'Affiliation-P': 0.9997653160605954, 'Affiliation-R': 0.9999526654427303, 'F1_T': 0.9090909090903951, 'Precision_T': 0.9803921568627258, 'Recall_T': 0.8474576271186297}
Results for 465_UCR_id_163_HumanActivity_tr_1671_1st_5920.csv: {'filename': '465_UCR_id_163_HumanActivity_tr_1671_1st_5920.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 115, 'train_index': '1671', 'data_shape': '6684x1', 'output_length': 10000, 'label_length': 5013, 'aligned_length': 5013, 'AUC-PR': 0.9665694070928227, 'AUC-ROC': 0.9996955037189602, 'VUS-PR': 0.9785638152364249, 'VUS-ROC': 0.9998480053238209, 'Standard-F1': 0.9579831932767947, 'Standard-Precision': 0.9499999999999842, 'Standard-Recall': 0.9661016949152378, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.999858981975477, 'Affiliation-P': 0.9997653160605954, 'Affiliation-R': 0.9999526654427303, 'F1_T': 0.9090909090903951, 'Precision_T': 0.9803921568627258, 'Recall_T': 0.8474576271186297}
Logits metrics skipped for 465_UCR_id_163_HumanActivity_tr_1671_1st_5920.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/465_UCR_id_163_HumanActivity_tr_1671_1st_5920_results.pkl
Processing file: 522_UCR_id_220_Medical_tr_10217_1st_27929.csv, Data shape: (40870, 1)
Label shape: (40870,)
Train data shape: (10217, 1), Test data shape: (30653, 1), Label test shape: (30653,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 522_UCR_id_220_Medical_tr_10217_1st_27929.csv...
Using CUDA device: 2
Data shape: (30653, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 30653, window_size 5000, adding 4347 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([7, 5000, 1])
batched_mask shape: torch.Size([7, 5000])

100%|██████████| 1/1 [00:00<00:00, 136.38it/s]
Original shapes - Output: (35000,), Label: (30653,)
Aligned shapes - Output: (30653,), Label: (30653,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  48%|████▊     | 97/200 [00:01<00:01, 96.11it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 111.71it/s]
Table1-only metrics completed in 2.11s
Evaluation Result:  {'AUC-PR': 0.03800992505688653, 'AUC-ROC': 0.9391462099173014, 'VUS-PR': 0.0739997228078983, 'VUS-ROC': 0.9586070563575944, 'Standard-F1': 0.09264305177084314, 'Standard-Precision': 0.05537459283387604, 'Standard-Recall': 0.2833333333333286, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8349188693256825, 'Affiliation-P': 0.7171603474058741, 'Affiliation-R': 0.9989470851140183, 'F1_T': 0.09264305177084314, 'Precision_T': 0.05537459283387604, 'Recall_T': 0.2833333333333286}
Results for 522_UCR_id_220_Medical_tr_10217_1st_27929.csv: {'filename': '522_UCR_id_220_Medical_tr_10217_1st_27929.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 37, 'train_index': '10217', 'data_shape': '40870x1', 'output_length': 35000, 'label_length': 30653, 'aligned_length': 30653, 'AUC-PR': 0.03800992505688653, 'AUC-ROC': 0.9391462099173014, 'VUS-PR': 0.0739997228078983, 'VUS-ROC': 0.9586070563575944, 'Standard-F1': 0.09264305177084314, 'Standard-Precision': 0.05537459283387604, 'Standard-Recall': 0.2833333333333286, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8349188693256825, 'Affiliation-P': 0.7171603474058741, 'Affiliation-R': 0.9989470851140183, 'F1_T': 0.09264305177084314, 'Precision_T': 0.05537459283387604, 'Recall_T': 0.2833333333333286}
Logits metrics skipped for 522_UCR_id_220_Medical_tr_10217_1st_27929.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/522_UCR_id_220_Medical_tr_10217_1st_27929_results.pkl
Processing file: 060_WSD_id_32_WebService_tr_4576_1st_4766.csv, Data shape: (18305, 1)
Label shape: (18305,)
Train data shape: (4576, 1), Test data shape: (13729, 1), Label test shape: (13729,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 060_WSD_id_32_WebService_tr_4576_1st_4766.csv...
Using CUDA device: 2
Data shape: (13729, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13729, window_size 5000, adding 1271 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.05it/s]
Original shapes - Output: (15000,), Label: (13729,)
Aligned shapes - Output: (13729,), Label: (13729,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 229.96it/s]
Table1-only metrics completed in 1.33s
Evaluation Result:  {'AUC-PR': 0.16463348277731488, 'AUC-ROC': 0.7395179984373583, 'VUS-PR': 0.06423043586652259, 'VUS-ROC': 0.829939863126514, 'Standard-F1': 0.12499999999951023, 'Standard-Precision': 0.10869565217391226, 'Standard-Recall': 0.14705882352941033, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8779456672464714, 'Affiliation-P': 0.8832784475954603, 'Affiliation-R': 0.8726768936202353, 'F1_T': 0.12499999999951023, 'Precision_T': 0.10869565217391226, 'Recall_T': 0.14705882352941033}
Results for 060_WSD_id_32_WebService_tr_4576_1st_4766.csv: {'filename': '060_WSD_id_32_WebService_tr_4576_1st_4766.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 20, 'train_index': '4576', 'data_shape': '18305x1', 'output_length': 15000, 'label_length': 13729, 'aligned_length': 13729, 'AUC-PR': 0.16463348277731488, 'AUC-ROC': 0.7395179984373583, 'VUS-PR': 0.06423043586652259, 'VUS-ROC': 0.829939863126514, 'Standard-F1': 0.12499999999951023, 'Standard-Precision': 0.10869565217391226, 'Standard-Recall': 0.14705882352941033, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8779456672464714, 'Affiliation-P': 0.8832784475954603, 'Affiliation-R': 0.8726768936202353, 'F1_T': 0.12499999999951023, 'Precision_T': 0.10869565217391226, 'Recall_T': 0.14705882352941033}
Logits metrics skipped for 060_WSD_id_32_WebService_tr_4576_1st_4766.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/060_WSD_id_32_WebService_tr_4576_1st_4766_results.pkl
Processing file: 157_Stock_id_9_Finance_tr_500_1st_19.csv, Data shape: (10000, 1)
Label shape: (10000,)
Train data shape: (500, 1), Test data shape: (9500, 1), Label test shape: (9500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 157_Stock_id_9_Finance_tr_500_1st_19.csv...
Using CUDA device: 2
Data shape: (9500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 164.83it/s]
Original shapes - Output: (10000,), Label: (9500,)
Aligned shapes - Output: (9500,), Label: (9500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   8%|▊         | 16/200 [00:01<00:11, 15.56it/s]

Calculating Affiliation Metrics:  16%|█▌        | 32/200 [00:03<00:21,  7.76it/s]

Calculating Affiliation Metrics:  21%|██        | 42/200 [00:06<00:28,  5.49it/s]

Calculating Affiliation Metrics:  24%|██▍       | 49/200 [00:09<00:33,  4.46it/s]

Calculating Affiliation Metrics:  28%|██▊       | 55/200 [00:11<00:38,  3.81it/s]

Calculating Affiliation Metrics:  30%|███       | 60/200 [00:13<00:40,  3.42it/s]

Calculating Affiliation Metrics:  32%|███▏      | 64/200 [00:15<00:43,  3.16it/s]

Calculating Affiliation Metrics:  34%|███▍      | 68/200 [00:16<00:44,  2.96it/s]

Calculating Affiliation Metrics:  36%|███▌      | 72/200 [00:18<00:44,  2.84it/s]

Calculating Affiliation Metrics:  38%|███▊      | 75/200 [00:19<00:44,  2.80it/s]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [00:20<00:43,  2.78it/s]

Calculating Affiliation Metrics:  40%|████      | 81/200 [00:21<00:42,  2.77it/s]

Calculating Affiliation Metrics:  42%|████▏     | 84/200 [00:22<00:42,  2.75it/s]

Calculating Affiliation Metrics:  44%|████▎     | 87/200 [00:23<00:41,  2.71it/s]

Calculating Affiliation Metrics:  45%|████▌     | 90/200 [00:25<00:41,  2.63it/s]

Calculating Affiliation Metrics:  46%|████▋     | 93/200 [00:26<00:41,  2.59it/s]

Calculating Affiliation Metrics:  48%|████▊     | 96/200 [00:27<00:40,  2.55it/s]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [00:28<00:39,  2.53it/s]

Calculating Affiliation Metrics:  51%|█████     | 102/200 [00:29<00:39,  2.51it/s]

Calculating Affiliation Metrics:  52%|█████▎    | 105/200 [00:31<00:37,  2.51it/s]

Calculating Affiliation Metrics:  54%|█████▍    | 108/200 [00:32<00:35,  2.56it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 111/200 [00:33<00:34,  2.62it/s]

Calculating Affiliation Metrics:  57%|█████▋    | 114/200 [00:34<00:32,  2.68it/s]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [00:35<00:29,  2.80it/s]

Calculating Affiliation Metrics:  61%|██████    | 122/200 [00:36<00:27,  2.89it/s]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [00:38<00:24,  3.00it/s]

Calculating Affiliation Metrics:  65%|██████▌   | 130/200 [00:39<00:21,  3.18it/s]

Calculating Affiliation Metrics:  67%|██████▋   | 134/200 [00:40<00:19,  3.36it/s]

Calculating Affiliation Metrics:  70%|██████▉   | 139/200 [00:41<00:16,  3.60it/s]

Calculating Affiliation Metrics:  72%|███████▏  | 144/200 [00:42<00:14,  3.84it/s]

Calculating Affiliation Metrics:  74%|███████▍  | 149/200 [00:43<00:12,  4.04it/s]

Calculating Affiliation Metrics:  77%|███████▋  | 154/200 [00:44<00:10,  4.28it/s]

Calculating Affiliation Metrics:  80%|████████  | 160/200 [00:45<00:08,  4.59it/s]

Calculating Affiliation Metrics:  84%|████████▎ | 167/200 [00:46<00:06,  5.23it/s]

Calculating Affiliation Metrics:  89%|████████▉ | 178/200 [00:47<00:03,  6.83it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:48<00:00,  4.11it/s]
Table1-only metrics completed in 61.67s
Evaluation Result:  {'AUC-PR': 0.15369066300680626, 'AUC-ROC': 0.5264637385813444, 'VUS-PR': 0.5271140798665929, 'VUS-ROC': 0.6971525236908624, 'Standard-F1': 0.24578403055973835, 'Standard-Precision': 0.14024455077086653, 'Standard-Recall': 0.9932228915662644, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6856038914794116, 'Affiliation-P': 0.5216113217583169, 'Affiliation-R': 1.0, 'F1_T': 0.24578403055973835, 'Precision_T': 0.14024455077086653, 'Recall_T': 0.9932228915662644}
Results for 157_Stock_id_9_Finance_tr_500_1st_19.csv: {'filename': '157_Stock_id_9_Finance_tr_500_1st_19.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 12, 'train_index': '500', 'data_shape': '10000x1', 'output_length': 10000, 'label_length': 9500, 'aligned_length': 9500, 'AUC-PR': 0.15369066300680626, 'AUC-ROC': 0.5264637385813444, 'VUS-PR': 0.5271140798665929, 'VUS-ROC': 0.6971525236908624, 'Standard-F1': 0.24578403055973835, 'Standard-Precision': 0.14024455077086653, 'Standard-Recall': 0.9932228915662644, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6856038914794116, 'Affiliation-P': 0.5216113217583169, 'Affiliation-R': 1.0, 'F1_T': 0.24578403055973835, 'Precision_T': 0.14024455077086653, 'Recall_T': 0.9932228915662644}
Logits metrics skipped for 157_Stock_id_9_Finance_tr_500_1st_19.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/157_Stock_id_9_Finance_tr_500_1st_19_results.pkl
Processing file: 490_UCR_id_188_Medical_tr_14750_1st_47530.csv, Data shape: (59000, 1)
Label shape: (59000,)
Train data shape: (14750, 1), Test data shape: (44250, 1), Label test shape: (44250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 490_UCR_id_188_Medical_tr_14750_1st_47530.csv...
Using CUDA device: 2
Data shape: (44250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 44250, window_size 5000, adding 750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([9, 5000, 1])
batched_mask shape: torch.Size([9, 5000])

100%|██████████| 1/1 [00:00<00:00, 83.35it/s]
Original shapes - Output: (45000,), Label: (44250,)
Aligned shapes - Output: (44250,), Label: (44250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  42%|████▎     | 85/200 [00:01<00:01, 84.58it/s]

Calculating Affiliation Metrics:  93%|█████████▎| 186/200 [00:02<00:00, 93.93it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:02<00:00, 95.92it/s]
Table1-only metrics completed in 2.30s
Evaluation Result:  {'AUC-PR': 0.060177153956412474, 'AUC-ROC': 0.9461133649255951, 'VUS-PR': 0.06557293512628942, 'VUS-ROC': 0.9463217017924468, 'Standard-F1': 0.18333333333309898, 'Standard-Precision': 0.10602409638554211, 'Standard-Recall': 0.6769230769230743, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8114384185442854, 'Affiliation-P': 0.6829249213199824, 'Affiliation-R': 0.9995312038244243, 'F1_T': 0.17711442786047113, 'Precision_T': 0.10171428571428566, 'Recall_T': 0.684615384615382}
Results for 490_UCR_id_188_Medical_tr_14750_1st_47530.csv: {'filename': '490_UCR_id_188_Medical_tr_14750_1st_47530.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 6, 'train_index': '14750', 'data_shape': '59000x1', 'output_length': 45000, 'label_length': 44250, 'aligned_length': 44250, 'AUC-PR': 0.060177153956412474, 'AUC-ROC': 0.9461133649255951, 'VUS-PR': 0.06557293512628942, 'VUS-ROC': 0.9463217017924468, 'Standard-F1': 0.18333333333309898, 'Standard-Precision': 0.10602409638554211, 'Standard-Recall': 0.6769230769230743, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8114384185442854, 'Affiliation-P': 0.6829249213199824, 'Affiliation-R': 0.9995312038244243, 'F1_T': 0.17711442786047113, 'Precision_T': 0.10171428571428566, 'Recall_T': 0.684615384615382}
Logits metrics skipped for 490_UCR_id_188_Medical_tr_14750_1st_47530.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/490_UCR_id_188_Medical_tr_14750_1st_47530_results.pkl
Processing file: 801_YAHOO_id_251_WebService_tr_500_1st_190.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 801_YAHOO_id_251_WebService_tr_500_1st_190.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 159.63it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 914.96it/s]
Table1-only metrics completed in 0.53s
Evaluation Result:  {'AUC-PR': 0.8708333333333333, 'AUC-ROC': 0.9990062464508802, 'VUS-PR': 0.9213710636617735, 'VUS-ROC': 0.9996460971722928, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9939835408437694, 'Affiliation-P': 0.9880390442890442, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Results for 801_YAHOO_id_251_WebService_tr_500_1st_190.csv: {'filename': '801_YAHOO_id_251_WebService_tr_500_1st_190.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.8708333333333333, 'AUC-ROC': 0.9990062464508802, 'VUS-PR': 0.9213710636617735, 'VUS-ROC': 0.9996460971722928, 'Standard-F1': 0.666666666666148, 'Standard-Precision': 0.4999999999999583, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9939835408437694, 'Affiliation-P': 0.9880390442890442, 'Affiliation-R': 1.0, 'F1_T': 0.666666666666148, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.9999999999998334}
Logits metrics skipped for 801_YAHOO_id_251_WebService_tr_500_1st_190.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/801_YAHOO_id_251_WebService_tr_500_1st_190_results.pkl
Processing file: 056_WSD_id_28_WebService_tr_4559_1st_9689.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 056_WSD_id_28_WebService_tr_4559_1st_9689.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.97it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 304.27it/s]
Table1-only metrics completed in 0.74s
Evaluation Result:  {'AUC-PR': 0.24549772020670854, 'AUC-ROC': 0.9806611072736464, 'VUS-PR': 0.16500737910672342, 'VUS-ROC': 0.9813157518400499, 'Standard-F1': 0.18079096045163817, 'Standard-Precision': 0.11510791366906392, 'Standard-Recall': 0.42105263157893624, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7509045001710081, 'Affiliation-P': 0.6012316339942629, 'Affiliation-R': 0.9997979704690547, 'F1_T': 0.18079096045163817, 'Precision_T': 0.11510791366906392, 'Recall_T': 0.42105263157893624}
Results for 056_WSD_id_28_WebService_tr_4559_1st_9689.csv: {'filename': '056_WSD_id_28_WebService_tr_4559_1st_9689.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 6, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.24549772020670854, 'AUC-ROC': 0.9806611072736464, 'VUS-PR': 0.16500737910672342, 'VUS-ROC': 0.9813157518400499, 'Standard-F1': 0.18079096045163817, 'Standard-Precision': 0.11510791366906392, 'Standard-Recall': 0.42105263157893624, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7509045001710081, 'Affiliation-P': 0.6012316339942629, 'Affiliation-R': 0.9997979704690547, 'F1_T': 0.18079096045163817, 'Precision_T': 0.11510791366906392, 'Recall_T': 0.42105263157893624}
Logits metrics skipped for 056_WSD_id_28_WebService_tr_4559_1st_9689.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/056_WSD_id_28_WebService_tr_4559_1st_9689_results.pkl
Processing file: 165_Stock_id_17_Finance_tr_500_1st_7.csv, Data shape: (20000, 1)
Label shape: (20000,)
Train data shape: (500, 1), Test data shape: (19500, 1), Label test shape: (19500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 165_Stock_id_17_Finance_tr_500_1st_7.csv...
Using CUDA device: 2
Data shape: (19500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 19500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.31it/s]
Original shapes - Output: (20000,), Label: (19500,)
Aligned shapes - Output: (19500,), Label: (19500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   4%|▎         | 7/200 [00:01<00:29,  6.56it/s]

Calculating Affiliation Metrics:   7%|▋         | 14/200 [00:04<00:58,  3.16it/s]

Calculating Affiliation Metrics:  10%|▉         | 19/200 [00:07<01:22,  2.20it/s]

Calculating Affiliation Metrics:  11%|█         | 22/200 [00:10<01:37,  1.82it/s]

Calculating Affiliation Metrics:  12%|█▎        | 25/200 [00:12<01:52,  1.56it/s]

Calculating Affiliation Metrics:  14%|█▎        | 27/200 [00:14<02:01,  1.43it/s]

Calculating Affiliation Metrics:  14%|█▍        | 29/200 [00:16<02:07,  1.34it/s]

Calculating Affiliation Metrics:  16%|█▌        | 31/200 [00:18<02:13,  1.26it/s]

Calculating Affiliation Metrics:  16%|█▋        | 33/200 [00:20<02:20,  1.19it/s]

Calculating Affiliation Metrics:  18%|█▊        | 35/200 [00:22<02:27,  1.12it/s]

Calculating Affiliation Metrics:  18%|█▊        | 37/200 [00:24<02:33,  1.06it/s]

Calculating Affiliation Metrics:  20%|█▉        | 39/200 [00:26<02:38,  1.01it/s]

Calculating Affiliation Metrics:  20%|██        | 41/200 [00:29<02:43,  1.03s/it]

Calculating Affiliation Metrics:  21%|██        | 42/200 [00:30<02:47,  1.06s/it]

Calculating Affiliation Metrics:  22%|██▏       | 43/200 [00:31<02:50,  1.09s/it]

Calculating Affiliation Metrics:  22%|██▏       | 44/200 [00:32<02:53,  1.11s/it]

Calculating Affiliation Metrics:  22%|██▎       | 45/200 [00:33<02:54,  1.13s/it]

Calculating Affiliation Metrics:  23%|██▎       | 46/200 [00:34<02:56,  1.15s/it]

Calculating Affiliation Metrics:  24%|██▎       | 47/200 [00:36<02:59,  1.17s/it]

Calculating Affiliation Metrics:  24%|██▍       | 48/200 [00:37<03:01,  1.20s/it]

Calculating Affiliation Metrics:  24%|██▍       | 49/200 [00:38<03:03,  1.22s/it]

Calculating Affiliation Metrics:  25%|██▌       | 50/200 [00:40<03:06,  1.24s/it]

Calculating Affiliation Metrics:  26%|██▌       | 51/200 [00:41<03:10,  1.28s/it]

Calculating Affiliation Metrics:  26%|██▌       | 52/200 [00:42<03:12,  1.30s/it]

Calculating Affiliation Metrics:  26%|██▋       | 53/200 [00:44<03:14,  1.33s/it]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:45<03:16,  1.35s/it]

Calculating Affiliation Metrics:  28%|██▊       | 55/200 [00:46<03:18,  1.37s/it]

Calculating Affiliation Metrics:  28%|██▊       | 56/200 [00:48<03:19,  1.38s/it]

Calculating Affiliation Metrics:  28%|██▊       | 57/200 [00:49<03:19,  1.40s/it]

Calculating Affiliation Metrics:  29%|██▉       | 58/200 [00:51<03:19,  1.41s/it]

Calculating Affiliation Metrics:  30%|██▉       | 59/200 [00:52<03:19,  1.41s/it]

Calculating Affiliation Metrics:  30%|███       | 60/200 [00:54<03:18,  1.42s/it]

Calculating Affiliation Metrics:  30%|███       | 61/200 [00:55<03:16,  1.41s/it]

Calculating Affiliation Metrics:  31%|███       | 62/200 [00:56<03:13,  1.41s/it]

Calculating Affiliation Metrics:  32%|███▏      | 63/200 [00:58<03:10,  1.39s/it]

Calculating Affiliation Metrics:  32%|███▏      | 64/200 [00:59<03:06,  1.37s/it]

Calculating Affiliation Metrics:  32%|███▎      | 65/200 [01:00<03:03,  1.36s/it]

Calculating Affiliation Metrics:  33%|███▎      | 66/200 [01:02<03:02,  1.36s/it]

Calculating Affiliation Metrics:  34%|███▎      | 67/200 [01:03<03:00,  1.36s/it]

Calculating Affiliation Metrics:  34%|███▍      | 68/200 [01:04<02:58,  1.35s/it]

Calculating Affiliation Metrics:  34%|███▍      | 69/200 [01:06<02:56,  1.35s/it]

Calculating Affiliation Metrics:  35%|███▌      | 70/200 [01:07<02:55,  1.35s/it]

Calculating Affiliation Metrics:  36%|███▌      | 71/200 [01:08<02:53,  1.35s/it]

Calculating Affiliation Metrics:  36%|███▌      | 72/200 [01:10<02:52,  1.35s/it]

Calculating Affiliation Metrics:  36%|███▋      | 73/200 [01:11<02:52,  1.36s/it]

Calculating Affiliation Metrics:  37%|███▋      | 74/200 [01:13<02:53,  1.38s/it]

Calculating Affiliation Metrics:  38%|███▊      | 75/200 [01:14<02:56,  1.41s/it]

Calculating Affiliation Metrics:  38%|███▊      | 76/200 [01:16<02:55,  1.41s/it]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [01:17<02:53,  1.41s/it]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [01:18<02:51,  1.40s/it]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [01:20<02:49,  1.40s/it]

Calculating Affiliation Metrics:  40%|████      | 80/200 [01:21<02:46,  1.39s/it]

Calculating Affiliation Metrics:  40%|████      | 81/200 [01:22<02:44,  1.38s/it]

Calculating Affiliation Metrics:  41%|████      | 82/200 [01:24<02:42,  1.38s/it]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [01:25<02:40,  1.38s/it]

Calculating Affiliation Metrics:  42%|████▏     | 84/200 [01:27<02:38,  1.37s/it]

Calculating Affiliation Metrics:  42%|████▎     | 85/200 [01:28<02:36,  1.36s/it]

Calculating Affiliation Metrics:  43%|████▎     | 86/200 [01:29<02:33,  1.35s/it]

Calculating Affiliation Metrics:  44%|████▎     | 87/200 [01:31<02:29,  1.33s/it]

Calculating Affiliation Metrics:  44%|████▍     | 88/200 [01:32<02:26,  1.31s/it]

Calculating Affiliation Metrics:  44%|████▍     | 89/200 [01:33<02:22,  1.29s/it]

Calculating Affiliation Metrics:  45%|████▌     | 90/200 [01:34<02:19,  1.26s/it]

Calculating Affiliation Metrics:  46%|████▌     | 91/200 [01:35<02:15,  1.24s/it]

Calculating Affiliation Metrics:  46%|████▌     | 92/200 [01:37<02:12,  1.23s/it]

Calculating Affiliation Metrics:  46%|████▋     | 93/200 [01:38<02:08,  1.20s/it]

Calculating Affiliation Metrics:  47%|████▋     | 94/200 [01:39<02:05,  1.18s/it]

Calculating Affiliation Metrics:  48%|████▊     | 95/200 [01:40<02:03,  1.18s/it]

Calculating Affiliation Metrics:  48%|████▊     | 96/200 [01:41<02:01,  1.17s/it]

Calculating Affiliation Metrics:  48%|████▊     | 97/200 [01:42<01:59,  1.16s/it]

Calculating Affiliation Metrics:  49%|████▉     | 98/200 [01:43<01:57,  1.15s/it]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [01:45<01:55,  1.15s/it]

Calculating Affiliation Metrics:  50%|█████     | 100/200 [01:46<01:54,  1.15s/it]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [01:47<01:54,  1.16s/it]

Calculating Affiliation Metrics:  51%|█████     | 102/200 [01:48<01:54,  1.16s/it]

Calculating Affiliation Metrics:  52%|█████▏    | 103/200 [01:49<01:53,  1.17s/it]

Calculating Affiliation Metrics:  52%|█████▏    | 104/200 [01:50<01:52,  1.17s/it]

Calculating Affiliation Metrics:  52%|█████▎    | 105/200 [01:52<01:53,  1.19s/it]

Calculating Affiliation Metrics:  53%|█████▎    | 106/200 [01:53<01:53,  1.21s/it]

Calculating Affiliation Metrics:  54%|█████▎    | 107/200 [01:54<01:53,  1.22s/it]

Calculating Affiliation Metrics:  54%|█████▍    | 108/200 [01:55<01:53,  1.23s/it]

Calculating Affiliation Metrics:  55%|█████▍    | 109/200 [01:57<01:51,  1.23s/it]

Calculating Affiliation Metrics:  55%|█████▌    | 110/200 [01:58<01:49,  1.22s/it]

Calculating Affiliation Metrics:  56%|█████▌    | 111/200 [01:59<01:48,  1.22s/it]

Calculating Affiliation Metrics:  56%|█████▌    | 112/200 [02:00<01:47,  1.22s/it]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [02:02<01:45,  1.22s/it]

Calculating Affiliation Metrics:  57%|█████▋    | 114/200 [02:03<01:43,  1.21s/it]

Calculating Affiliation Metrics:  57%|█████▊    | 115/200 [02:04<01:40,  1.19s/it]

Calculating Affiliation Metrics:  58%|█████▊    | 116/200 [02:05<01:38,  1.17s/it]

Calculating Affiliation Metrics:  58%|█████▊    | 117/200 [02:06<01:36,  1.16s/it]

Calculating Affiliation Metrics:  59%|█████▉    | 118/200 [02:07<01:34,  1.15s/it]

Calculating Affiliation Metrics:  60%|█████▉    | 119/200 [02:08<01:31,  1.13s/it]

Calculating Affiliation Metrics:  60%|██████    | 120/200 [02:09<01:30,  1.13s/it]

Calculating Affiliation Metrics:  60%|██████    | 121/200 [02:11<01:27,  1.11s/it]

Calculating Affiliation Metrics:  61%|██████    | 122/200 [02:12<01:25,  1.09s/it]

Calculating Affiliation Metrics:  62%|██████▏   | 123/200 [02:13<01:23,  1.08s/it]

Calculating Affiliation Metrics:  62%|██████▏   | 124/200 [02:14<01:21,  1.07s/it]

Calculating Affiliation Metrics:  62%|██████▎   | 125/200 [02:15<01:18,  1.05s/it]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [02:17<01:14,  1.01s/it]

Calculating Affiliation Metrics:  64%|██████▍   | 129/200 [02:19<01:09,  1.02it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 131/200 [02:20<01:06,  1.04it/s]

Calculating Affiliation Metrics:  66%|██████▋   | 133/200 [02:22<01:02,  1.08it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 135/200 [02:24<00:58,  1.11it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 137/200 [02:25<00:55,  1.13it/s]

Calculating Affiliation Metrics:  70%|██████▉   | 139/200 [02:27<00:52,  1.16it/s]

Calculating Affiliation Metrics:  70%|███████   | 141/200 [02:29<00:48,  1.22it/s]

Calculating Affiliation Metrics:  72%|███████▏  | 143/200 [02:30<00:45,  1.25it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [02:31<00:42,  1.30it/s]

Calculating Affiliation Metrics:  74%|███████▎  | 147/200 [02:33<00:39,  1.34it/s]

Calculating Affiliation Metrics:  74%|███████▍  | 149/200 [02:34<00:36,  1.39it/s]

Calculating Affiliation Metrics:  76%|███████▌  | 151/200 [02:35<00:33,  1.45it/s]

Calculating Affiliation Metrics:  76%|███████▋  | 153/200 [02:37<00:30,  1.52it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 155/200 [02:38<00:28,  1.57it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 157/200 [02:39<00:26,  1.63it/s]

Calculating Affiliation Metrics:  80%|███████▉  | 159/200 [02:40<00:24,  1.69it/s]

Calculating Affiliation Metrics:  80%|████████  | 161/200 [02:41<00:22,  1.77it/s]

Calculating Affiliation Metrics:  82%|████████▏ | 164/200 [02:42<00:19,  1.86it/s]

Calculating Affiliation Metrics:  84%|████████▎ | 167/200 [02:44<00:16,  1.95it/s]

Calculating Affiliation Metrics:  85%|████████▌ | 170/200 [02:45<00:14,  2.07it/s]

Calculating Affiliation Metrics:  86%|████████▋ | 173/200 [02:46<00:12,  2.20it/s]

Calculating Affiliation Metrics:  88%|████████▊ | 177/200 [02:48<00:09,  2.49it/s]

Calculating Affiliation Metrics:  90%|█████████ | 181/200 [02:49<00:06,  2.76it/s]

Calculating Affiliation Metrics:  93%|█████████▎| 186/200 [02:50<00:04,  3.22it/s]

Calculating Affiliation Metrics:  96%|█████████▋| 193/200 [02:51<00:01,  4.14it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [02:51<00:00,  1.16it/s]
Table1-only metrics completed in 263.22s
Evaluation Result:  {'AUC-PR': 0.17479633931537666, 'AUC-ROC': 0.5440779340550128, 'VUS-PR': 0.972750927914976, 'VUS-ROC': 0.9824071522621759, 'Standard-F1': 0.24652177866669184, 'Standard-Precision': 0.140714211564414, 'Standard-Recall': 0.9937683284457475, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6867004190413174, 'Affiliation-P': 0.5232403951205336, 'Affiliation-R': 0.9986909441180365, 'F1_T': 0.2463577361228711, 'Precision_T': 0.14058534058534058, 'Recall_T': 0.9948680351906155}
Results for 165_Stock_id_17_Finance_tr_500_1st_7.csv: {'filename': '165_Stock_id_17_Finance_tr_500_1st_7.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 249, 'train_index': '500', 'data_shape': '20000x1', 'output_length': 20000, 'label_length': 19500, 'aligned_length': 19500, 'AUC-PR': 0.17479633931537666, 'AUC-ROC': 0.5440779340550128, 'VUS-PR': 0.972750927914976, 'VUS-ROC': 0.9824071522621759, 'Standard-F1': 0.24652177866669184, 'Standard-Precision': 0.140714211564414, 'Standard-Recall': 0.9937683284457475, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6867004190413174, 'Affiliation-P': 0.5232403951205336, 'Affiliation-R': 0.9986909441180365, 'F1_T': 0.2463577361228711, 'Precision_T': 0.14058534058534058, 'Recall_T': 0.9948680351906155}
Logits metrics skipped for 165_Stock_id_17_Finance_tr_500_1st_7.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/165_Stock_id_17_Finance_tr_500_1st_7_results.pkl
Skipping file: 550_SWaT_id_1_Sensor_tr_43700_1st_43800.csv due to filter criteria.
Skipping file: 240_SVDB_id_4_Medical_tr_18462_1st_18562.csv due to filter criteria.
Skipping file: 870_OPPORTUNITY_id_29_HumanActivity_tr_1477_1st_1577.csv due to filter criteria.
Processing file: 570_YAHOO_id_20_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 570_YAHOO_id_20_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 144.29it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1789.18it/s]
Table1-only metrics completed in 0.84s
Evaluation Result:  {'AUC-PR': 0.95, 'AUC-ROC': 0.9997273718647764, 'VUS-PR': 0.9911705794968647, 'VUS-ROC': 0.9999704051661991, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 570_YAHOO_id_20_Synthetic_tr_500_1st_658.csv: {'filename': '570_YAHOO_id_20_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 135, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.95, 'AUC-ROC': 0.9997273718647764, 'VUS-PR': 0.9911705794968647, 'VUS-ROC': 0.9999704051661991, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 570_YAHOO_id_20_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/570_YAHOO_id_20_Synthetic_tr_500_1st_658_results.pkl
Skipping file: 220_LTDB_id_5_Medical_tr_500_1st_26.csv due to filter criteria.
Processing file: 273_IOPS_id_14_WebService_tr_5569_1st_5669.csv, Data shape: (149122, 1)
Label shape: (149122,)
Train data shape: (5569, 1), Test data shape: (143553, 1), Label test shape: (143553,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 273_IOPS_id_14_WebService_tr_5569_1st_5669.csv...
Using CUDA device: 2
Data shape: (143553, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 143553, window_size 5000, adding 1447 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([29, 5000, 1])
batched_mask shape: torch.Size([29, 5000])

100%|██████████| 1/1 [00:00<00:00, 47.89it/s]
Original shapes - Output: (145000,), Label: (143553,)
Aligned shapes - Output: (143553,), Label: (143553,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   8%|▊         | 17/200 [00:01<00:11, 16.42it/s]

Calculating Affiliation Metrics:  17%|█▋        | 34/200 [00:02<00:13, 12.49it/s]

Calculating Affiliation Metrics:  24%|██▍       | 48/200 [00:04<00:13, 11.08it/s]

Calculating Affiliation Metrics:  30%|███       | 60/200 [00:05<00:13, 10.23it/s]

Calculating Affiliation Metrics:  36%|███▌      | 71/200 [00:06<00:13,  9.46it/s]

Calculating Affiliation Metrics:  40%|████      | 81/200 [00:08<00:13,  9.10it/s]

Calculating Affiliation Metrics:  46%|████▌     | 91/200 [00:09<00:12,  8.84it/s]

Calculating Affiliation Metrics:  50%|█████     | 100/200 [00:10<00:11,  8.85it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 111/200 [00:11<00:09,  9.28it/s]

Calculating Affiliation Metrics:  61%|██████    | 122/200 [00:12<00:08,  9.75it/s]

Calculating Affiliation Metrics:  66%|██████▋   | 133/200 [00:13<00:06, 10.10it/s]

Calculating Affiliation Metrics:  72%|███████▎  | 145/200 [00:14<00:05, 10.44it/s]

Calculating Affiliation Metrics:  79%|███████▉  | 158/200 [00:15<00:03, 11.10it/s]

Calculating Affiliation Metrics:  87%|████████▋ | 174/200 [00:16<00:02, 12.36it/s]

Calculating Affiliation Metrics:  98%|█████████▊| 197/200 [00:17<00:00, 15.39it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:17<00:00, 11.42it/s]
Table1-only metrics completed in 35.75s
Evaluation Result:  {'AUC-PR': 0.1528126902406788, 'AUC-ROC': 0.9470853518665746, 'VUS-PR': 0.1564035407214747, 'VUS-ROC': 0.9725777205387005, 'Standard-F1': 0.18326693227051688, 'Standard-Precision': 0.12649999999999995, 'Standard-Recall': 0.33245729303547916, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9069040799677742, 'Affiliation-P': 0.8480851201472647, 'Affiliation-R': 0.9744898416816652, 'F1_T': 0.1797520661153156, 'Precision_T': 0.12179188054129719, 'Recall_T': 0.3429697766097236}
Results for 273_IOPS_id_14_WebService_tr_5569_1st_5669.csv: {'filename': '273_IOPS_id_14_WebService_tr_5569_1st_5669.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '5569', 'data_shape': '149122x1', 'output_length': 145000, 'label_length': 143553, 'aligned_length': 143553, 'AUC-PR': 0.1528126902406788, 'AUC-ROC': 0.9470853518665746, 'VUS-PR': 0.1564035407214747, 'VUS-ROC': 0.9725777205387005, 'Standard-F1': 0.18326693227051688, 'Standard-Precision': 0.12649999999999995, 'Standard-Recall': 0.33245729303547916, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9069040799677742, 'Affiliation-P': 0.8480851201472647, 'Affiliation-R': 0.9744898416816652, 'F1_T': 0.1797520661153156, 'Precision_T': 0.12179188054129719, 'Recall_T': 0.3429697766097236}
Logits metrics skipped for 273_IOPS_id_14_WebService_tr_5569_1st_5669.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/273_IOPS_id_14_WebService_tr_5569_1st_5669_results.pkl
Processing file: 758_YAHOO_id_208_WebService_tr_500_1st_24.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 758_YAHOO_id_208_WebService_tr_500_1st_24.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 148.57it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1072.16it/s]
Table1-only metrics completed in 0.56s
Evaluation Result:  {'AUC-PR': 0.9285714285714284, 'AUC-ROC': 0.9991474850809889, 'VUS-PR': 0.9459475794989811, 'VUS-ROC': 0.9996277620830446, 'Standard-F1': 0.6363636363631445, 'Standard-Precision': 0.46666666666663553, 'Standard-Recall': 0.9999999999998571, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.992156885923176, 'Affiliation-P': 0.9844358433028384, 'Affiliation-R': 1.0, 'F1_T': 0.6315789473678891, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.8571428571427346}
Results for 758_YAHOO_id_208_WebService_tr_500_1st_24.csv: {'filename': '758_YAHOO_id_208_WebService_tr_500_1st_24.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9285714285714284, 'AUC-ROC': 0.9991474850809889, 'VUS-PR': 0.9459475794989811, 'VUS-ROC': 0.9996277620830446, 'Standard-F1': 0.6363636363631445, 'Standard-Precision': 0.46666666666663553, 'Standard-Recall': 0.9999999999998571, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.992156885923176, 'Affiliation-P': 0.9844358433028384, 'Affiliation-R': 1.0, 'F1_T': 0.6315789473678891, 'Precision_T': 0.4999999999999583, 'Recall_T': 0.8571428571427346}
Logits metrics skipped for 758_YAHOO_id_208_WebService_tr_500_1st_24.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/758_YAHOO_id_208_WebService_tr_500_1st_24_results.pkl
Processing file: 119_WSD_id_91_WebService_tr_4559_1st_12353.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 119_WSD_id_91_WebService_tr_4559_1st_12353.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.81it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 340.64it/s]
Table1-only metrics completed in 1.17s
Evaluation Result:  {'AUC-PR': 0.034002091428609056, 'AUC-ROC': 0.6288432766641802, 'VUS-PR': 0.03576900399417265, 'VUS-ROC': 0.7659028973002364, 'Standard-F1': 0.023255813953170903, 'Standard-Precision': 0.014492753623188302, 'Standard-Recall': 0.05882352941176297, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7733756586058212, 'Affiliation-P': 0.6309312259129629, 'Affiliation-R': 0.9988946664430195, 'F1_T': 0.023255813953170903, 'Precision_T': 0.014492753623188302, 'Recall_T': 0.05882352941176297}
Results for 119_WSD_id_91_WebService_tr_4559_1st_12353.csv: {'filename': '119_WSD_id_91_WebService_tr_4559_1st_12353.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.034002091428609056, 'AUC-ROC': 0.6288432766641802, 'VUS-PR': 0.03576900399417265, 'VUS-ROC': 0.7659028973002364, 'Standard-F1': 0.023255813953170903, 'Standard-Precision': 0.014492753623188302, 'Standard-Recall': 0.05882352941176297, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7733756586058212, 'Affiliation-P': 0.6309312259129629, 'Affiliation-R': 0.9988946664430195, 'F1_T': 0.023255813953170903, 'Precision_T': 0.014492753623188302, 'Recall_T': 0.05882352941176297}
Logits metrics skipped for 119_WSD_id_91_WebService_tr_4559_1st_12353.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/119_WSD_id_91_WebService_tr_4559_1st_12353_results.pkl
Processing file: 417_UCR_id_115_Sensor_tr_2750_1st_5400.csv, Data shape: (11000, 1)
Label shape: (11000,)
Train data shape: (2750, 1), Test data shape: (8250, 1), Label test shape: (8250,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 417_UCR_id_115_Sensor_tr_2750_1st_5400.csv...
Using CUDA device: 2
Data shape: (8250, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 8250, window_size 5000, adding 1750 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 160.54it/s]
Original shapes - Output: (10000,), Label: (8250,)
Aligned shapes - Output: (8250,), Label: (8250,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 469.28it/s]
Table1-only metrics completed in 0.78s
Evaluation Result:  {'AUC-PR': 0.06778585092630762, 'AUC-ROC': 0.7963773291925466, 'VUS-PR': 0.10113351998031422, 'VUS-ROC': 0.826772209369218, 'Standard-F1': 0.16553287981824305, 'Standard-Precision': 0.10703812316715526, 'Standard-Recall': 0.36499999999999816, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9457274276467513, 'Affiliation-P': 0.917064935064935, 'Affiliation-R': 0.976239393939394, 'F1_T': 0.16591928251086252, 'Precision_T': 0.10693641618497093, 'Recall_T': 0.36999999999999816}
Results for 417_UCR_id_115_Sensor_tr_2750_1st_5400.csv: {'filename': '417_UCR_id_115_Sensor_tr_2750_1st_5400.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 81, 'train_index': '2750', 'data_shape': '11000x1', 'output_length': 10000, 'label_length': 8250, 'aligned_length': 8250, 'AUC-PR': 0.06778585092630762, 'AUC-ROC': 0.7963773291925466, 'VUS-PR': 0.10113351998031422, 'VUS-ROC': 0.826772209369218, 'Standard-F1': 0.16553287981824305, 'Standard-Precision': 0.10703812316715526, 'Standard-Recall': 0.36499999999999816, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9457274276467513, 'Affiliation-P': 0.917064935064935, 'Affiliation-R': 0.976239393939394, 'F1_T': 0.16591928251086252, 'Precision_T': 0.10693641618497093, 'Recall_T': 0.36999999999999816}
Logits metrics skipped for 417_UCR_id_115_Sensor_tr_2750_1st_5400.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/417_UCR_id_115_Sensor_tr_2750_1st_5400_results.pkl
Processing file: 564_YAHOO_id_14_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 564_YAHOO_id_14_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 157.95it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2044.87it/s]
Table1-only metrics completed in 0.92s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 564_YAHOO_id_14_Synthetic_tr_500_1st_658.csv: {'filename': '564_YAHOO_id_14_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 151, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 564_YAHOO_id_14_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/564_YAHOO_id_14_Synthetic_tr_500_1st_658_results.pkl
Processing file: 040_WSD_id_12_WebService_tr_4559_1st_9714.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 040_WSD_id_12_WebService_tr_4559_1st_9714.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.44it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 295.79it/s]
Table1-only metrics completed in 1.58s
Evaluation Result:  {'AUC-PR': 0.013750992365051621, 'AUC-ROC': 0.8639675106783336, 'VUS-PR': 0.013833540466314511, 'VUS-ROC': 0.8867126708144968, 'Standard-F1': 0.0272263187747702, 'Standard-Precision': 0.01393728222996515, 'Standard-Recall': 0.5853658536585222, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7603884425426297, 'Affiliation-P': 0.6365826701616129, 'Affiliation-R': 0.9439779896654606, 'F1_T': 0.02674094707516425, 'Precision_T': 0.01368301026225769, 'Recall_T': 0.5853658536585222}
Results for 040_WSD_id_12_WebService_tr_4559_1st_9714.csv: {'filename': '040_WSD_id_12_WebService_tr_4559_1st_9714.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.013750992365051621, 'AUC-ROC': 0.8639675106783336, 'VUS-PR': 0.013833540466314511, 'VUS-ROC': 0.8867126708144968, 'Standard-F1': 0.0272263187747702, 'Standard-Precision': 0.01393728222996515, 'Standard-Recall': 0.5853658536585222, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7603884425426297, 'Affiliation-P': 0.6365826701616129, 'Affiliation-R': 0.9439779896654606, 'F1_T': 0.02674094707516425, 'Precision_T': 0.01368301026225769, 'Recall_T': 0.5853658536585222}
Logits metrics skipped for 040_WSD_id_12_WebService_tr_4559_1st_9714.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/040_WSD_id_12_WebService_tr_4559_1st_9714_results.pkl
Processing file: 589_YAHOO_id_39_WebService_tr_500_1st_10.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 589_YAHOO_id_39_WebService_tr_500_1st_10.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 162.54it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 792.96it/s]
Table1-only metrics completed in 0.72s
Evaluation Result:  {'AUC-PR': 1.0000000000000002, 'AUC-ROC': 1.0, 'VUS-PR': 0.9904331281835925, 'VUS-ROC': 0.9999591444188849, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9943670953342538, 'Affiliation-P': 0.9887972944409216, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Results for 589_YAHOO_id_39_WebService_tr_500_1st_10.csv: {'filename': '589_YAHOO_id_39_WebService_tr_500_1st_10.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0000000000000002, 'AUC-ROC': 1.0, 'VUS-PR': 0.9904331281835925, 'VUS-ROC': 0.9999591444188849, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9943670953342538, 'Affiliation-P': 0.9887972944409216, 'Affiliation-R': 1.0, 'F1_T': 0.8571428571422857, 'Precision_T': 0.7499999999999375, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 589_YAHOO_id_39_WebService_tr_500_1st_10.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/589_YAHOO_id_39_WebService_tr_500_1st_10_results.pkl
Skipping file: 826_Exathlon_id_17_Facility_tr_11665_1st_13484.csv due to filter criteria.
Processing file: 360_UCR_id_58_Facility_tr_14030_1st_52600.csv, Data shape: (56123, 1)
Label shape: (56123,)
Train data shape: (14030, 1), Test data shape: (42093, 1), Label test shape: (42093,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 360_UCR_id_58_Facility_tr_14030_1st_52600.csv...
Using CUDA device: 2
Data shape: (42093, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 42093, window_size 5000, adding 2907 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([9, 5000, 1])
batched_mask shape: torch.Size([9, 5000])

100%|██████████| 1/1 [00:00<00:00, 110.14it/s]
Original shapes - Output: (45000,), Label: (42093,)
Aligned shapes - Output: (42093,), Label: (42093,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [00:01<00:01, 98.87it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 114.06it/s]
Table1-only metrics completed in 3.46s
Evaluation Result:  {'AUC-PR': 0.23719314699928765, 'AUC-ROC': 0.9203621726780131, 'VUS-PR': 0.11412683445218495, 'VUS-ROC': 0.9315180221029392, 'Standard-F1': 0.23011363636322893, 'Standard-Precision': 0.1607142857142854, 'Standard-Recall': 0.404999999999998, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7620149594898862, 'Affiliation-P': 0.6158090980198361, 'Affiliation-R': 0.9992600313591333, 'F1_T': 0.228663446054313, 'Precision_T': 0.1686460807600946, 'Recall_T': 0.35499999999999826}
Results for 360_UCR_id_58_Facility_tr_14030_1st_52600.csv: {'filename': '360_UCR_id_58_Facility_tr_14030_1st_52600.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 233, 'train_index': '14030', 'data_shape': '56123x1', 'output_length': 45000, 'label_length': 42093, 'aligned_length': 42093, 'AUC-PR': 0.23719314699928765, 'AUC-ROC': 0.9203621726780131, 'VUS-PR': 0.11412683445218495, 'VUS-ROC': 0.9315180221029392, 'Standard-F1': 0.23011363636322893, 'Standard-Precision': 0.1607142857142854, 'Standard-Recall': 0.404999999999998, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7620149594898862, 'Affiliation-P': 0.6158090980198361, 'Affiliation-R': 0.9992600313591333, 'F1_T': 0.228663446054313, 'Precision_T': 0.1686460807600946, 'Recall_T': 0.35499999999999826}
Logits metrics skipped for 360_UCR_id_58_Facility_tr_14030_1st_52600.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/360_UCR_id_58_Facility_tr_14030_1st_52600_results.pkl
Processing file: 158_Stock_id_10_Finance_tr_500_1st_2.csv, Data shape: (20000, 1)
Label shape: (20000,)
Train data shape: (500, 1), Test data shape: (19500, 1), Label test shape: (19500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 158_Stock_id_10_Finance_tr_500_1st_2.csv...
Using CUDA device: 2
Data shape: (19500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 19500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 153.89it/s]
Original shapes - Output: (20000,), Label: (19500,)
Aligned shapes - Output: (19500,), Label: (19500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   4%|▍         | 9/200 [00:01<00:25,  7.49it/s]

Calculating Affiliation Metrics:   8%|▊         | 17/200 [00:03<00:39,  4.64it/s]

Calculating Affiliation Metrics:  12%|█▏        | 23/200 [00:05<00:47,  3.73it/s]

Calculating Affiliation Metrics:  14%|█▍        | 28/200 [00:07<00:56,  3.07it/s]

Calculating Affiliation Metrics:  16%|█▌        | 32/200 [00:09<01:00,  2.77it/s]

Calculating Affiliation Metrics:  18%|█▊        | 35/200 [00:11<01:03,  2.59it/s]

Calculating Affiliation Metrics:  19%|█▉        | 38/200 [00:12<01:06,  2.44it/s]

Calculating Affiliation Metrics:  20%|██        | 41/200 [00:14<01:08,  2.31it/s]

Calculating Affiliation Metrics:  22%|██▏       | 44/200 [00:15<01:09,  2.23it/s]

Calculating Affiliation Metrics:  24%|██▎       | 47/200 [00:17<01:11,  2.15it/s]

Calculating Affiliation Metrics:  25%|██▌       | 50/200 [00:18<01:11,  2.09it/s]

Calculating Affiliation Metrics:  26%|██▋       | 53/200 [00:20<01:13,  2.01it/s]

Calculating Affiliation Metrics:  28%|██▊       | 56/200 [00:21<01:14,  1.95it/s]

Calculating Affiliation Metrics:  29%|██▉       | 58/200 [00:23<01:15,  1.89it/s]

Calculating Affiliation Metrics:  30%|███       | 60/200 [00:24<01:16,  1.83it/s]

Calculating Affiliation Metrics:  31%|███       | 62/200 [00:25<01:17,  1.78it/s]

Calculating Affiliation Metrics:  32%|███▏      | 64/200 [00:26<01:17,  1.76it/s]

Calculating Affiliation Metrics:  33%|███▎      | 66/200 [00:27<01:16,  1.75it/s]

Calculating Affiliation Metrics:  34%|███▍      | 68/200 [00:28<01:15,  1.75it/s]

Calculating Affiliation Metrics:  35%|███▌      | 70/200 [00:30<01:13,  1.76it/s]

Calculating Affiliation Metrics:  36%|███▌      | 72/200 [00:31<01:12,  1.77it/s]

Calculating Affiliation Metrics:  37%|███▋      | 74/200 [00:32<01:11,  1.77it/s]

Calculating Affiliation Metrics:  38%|███▊      | 76/200 [00:33<01:09,  1.78it/s]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [00:34<01:07,  1.80it/s]

Calculating Affiliation Metrics:  40%|████      | 80/200 [00:35<01:05,  1.82it/s]

Calculating Affiliation Metrics:  41%|████      | 82/200 [00:36<01:04,  1.83it/s]

Calculating Affiliation Metrics:  42%|████▏     | 84/200 [00:37<01:02,  1.84it/s]

Calculating Affiliation Metrics:  43%|████▎     | 86/200 [00:38<01:01,  1.86it/s]

Calculating Affiliation Metrics:  44%|████▍     | 88/200 [00:39<00:59,  1.88it/s]

Calculating Affiliation Metrics:  45%|████▌     | 90/200 [00:40<00:58,  1.89it/s]

Calculating Affiliation Metrics:  46%|████▋     | 93/200 [00:42<00:55,  1.94it/s]

Calculating Affiliation Metrics:  48%|████▊     | 96/200 [00:43<00:53,  1.96it/s]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [00:45<00:50,  1.99it/s]

Calculating Affiliation Metrics:  51%|█████     | 102/200 [00:46<00:48,  2.02it/s]

Calculating Affiliation Metrics:  52%|█████▎    | 105/200 [00:48<00:46,  2.07it/s]

Calculating Affiliation Metrics:  54%|█████▍    | 108/200 [00:49<00:43,  2.10it/s]

Calculating Affiliation Metrics:  56%|█████▌    | 111/200 [00:50<00:41,  2.12it/s]

Calculating Affiliation Metrics:  57%|█████▋    | 114/200 [00:52<00:40,  2.14it/s]

Calculating Affiliation Metrics:  58%|█████▊    | 117/200 [00:53<00:38,  2.16it/s]

Calculating Affiliation Metrics:  60%|██████    | 120/200 [00:54<00:36,  2.20it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 123/200 [00:56<00:34,  2.23it/s]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [00:57<00:33,  2.24it/s]

Calculating Affiliation Metrics:  64%|██████▍   | 129/200 [00:58<00:31,  2.27it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 132/200 [01:00<00:29,  2.30it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 135/200 [01:01<00:28,  2.32it/s]

Calculating Affiliation Metrics:  69%|██████▉   | 138/200 [01:02<00:26,  2.32it/s]

Calculating Affiliation Metrics:  70%|███████   | 141/200 [01:03<00:25,  2.34it/s]

Calculating Affiliation Metrics:  72%|███████▏  | 144/200 [01:05<00:23,  2.41it/s]

Calculating Affiliation Metrics:  74%|███████▎  | 147/200 [01:06<00:21,  2.48it/s]

Calculating Affiliation Metrics:  75%|███████▌  | 150/200 [01:07<00:19,  2.59it/s]

Calculating Affiliation Metrics:  77%|███████▋  | 154/200 [01:08<00:16,  2.79it/s]

Calculating Affiliation Metrics:  79%|███████▉  | 158/200 [01:09<00:14,  2.99it/s]

Calculating Affiliation Metrics:  81%|████████  | 162/200 [01:10<00:11,  3.22it/s]

Calculating Affiliation Metrics:  84%|████████▎ | 167/200 [01:11<00:09,  3.61it/s]

Calculating Affiliation Metrics:  86%|████████▋ | 173/200 [01:12<00:06,  4.20it/s]

Calculating Affiliation Metrics:  90%|█████████ | 181/200 [01:13<00:03,  5.15it/s]

Calculating Affiliation Metrics:  96%|█████████▌| 191/200 [01:14<00:01,  6.44it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [01:15<00:00,  2.65it/s]
Table1-only metrics completed in 102.86s
Evaluation Result:  {'AUC-PR': 0.1308607150795984, 'AUC-ROC': 0.6328900823800425, 'VUS-PR': 0.5973826432286204, 'VUS-ROC': 0.8286353328487733, 'Standard-F1': 0.15559440559392076, 'Standard-Precision': 0.1324404761904761, 'Standard-Recall': 0.1885593220338981, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6819643739468558, 'Affiliation-P': 0.5402924318541252, 'Affiliation-R': 0.9243383105022418, 'F1_T': 0.15532286212866028, 'Precision_T': 0.13204747774480705, 'Recall_T': 0.1885593220338981}
Results for 158_Stock_id_10_Finance_tr_500_1st_2.csv: {'filename': '158_Stock_id_10_Finance_tr_500_1st_2.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 42, 'train_index': '500', 'data_shape': '20000x1', 'output_length': 20000, 'label_length': 19500, 'aligned_length': 19500, 'AUC-PR': 0.1308607150795984, 'AUC-ROC': 0.6328900823800425, 'VUS-PR': 0.5973826432286204, 'VUS-ROC': 0.8286353328487733, 'Standard-F1': 0.15559440559392076, 'Standard-Precision': 0.1324404761904761, 'Standard-Recall': 0.1885593220338981, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.6819643739468558, 'Affiliation-P': 0.5402924318541252, 'Affiliation-R': 0.9243383105022418, 'F1_T': 0.15532286212866028, 'Precision_T': 0.13204747774480705, 'Recall_T': 0.1885593220338981}
Logits metrics skipped for 158_Stock_id_10_Finance_tr_500_1st_2.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/158_Stock_id_10_Finance_tr_500_1st_2_results.pkl
Processing file: 121_WSD_id_93_WebService_tr_4559_1st_11749.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 121_WSD_id_93_WebService_tr_4559_1st_11749.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 145.82it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 274.28it/s]
Table1-only metrics completed in 1.05s
Evaluation Result:  {'AUC-PR': 0.12922989759982842, 'AUC-ROC': 0.7653625853287438, 'VUS-PR': 0.0360269238636087, 'VUS-ROC': 0.8581293494001421, 'Standard-F1': 0.07729468599000826, 'Standard-Precision': 0.04878048780487775, 'Standard-Recall': 0.18604651162790264, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8065432579843516, 'Affiliation-P': 0.6768615359236891, 'Affiliation-R': 0.9976941666288205, 'F1_T': 0.06666666666630229, 'Precision_T': 0.04379562043795589, 'Recall_T': 0.13953488372092698}
Results for 121_WSD_id_93_WebService_tr_4559_1st_11749.csv: {'filename': '121_WSD_id_93_WebService_tr_4559_1st_11749.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.12922989759982842, 'AUC-ROC': 0.7653625853287438, 'VUS-PR': 0.0360269238636087, 'VUS-ROC': 0.8581293494001421, 'Standard-F1': 0.07729468599000826, 'Standard-Precision': 0.04878048780487775, 'Standard-Recall': 0.18604651162790264, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8065432579843516, 'Affiliation-P': 0.6768615359236891, 'Affiliation-R': 0.9976941666288205, 'F1_T': 0.06666666666630229, 'Precision_T': 0.04379562043795589, 'Recall_T': 0.13953488372092698}
Logits metrics skipped for 121_WSD_id_93_WebService_tr_4559_1st_11749.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/121_WSD_id_93_WebService_tr_4559_1st_11749_results.pkl
Processing file: 076_WSD_id_48_WebService_tr_3821_1st_7083.csv, Data shape: (15285, 1)
Label shape: (15285,)
Train data shape: (3821, 1), Test data shape: (11464, 1), Label test shape: (11464,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 076_WSD_id_48_WebService_tr_3821_1st_7083.csv...
Using CUDA device: 2
Data shape: (11464, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 11464, window_size 5000, adding 3536 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.10it/s]
Original shapes - Output: (15000,), Label: (11464,)
Aligned shapes - Output: (11464,), Label: (11464,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  80%|███████▉  | 159/200 [00:01<00:00, 158.81it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 177.67it/s]
Table1-only metrics completed in 1.23s
Evaluation Result:  {'AUC-PR': 0.21005334155559538, 'AUC-ROC': 0.9928609534895, 'VUS-PR': 0.21304489911418711, 'VUS-ROC': 0.994866212103044, 'Standard-F1': 0.0967741935482509, 'Standard-Precision': 0.05217391304347781, 'Standard-Recall': 0.6666666666665926, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.788202449414801, 'Affiliation-P': 0.6504622638184996, 'Affiliation-R': 0.9999490393925496, 'F1_T': 0.0967741935482509, 'Precision_T': 0.05217391304347781, 'Recall_T': 0.6666666666665926}
Results for 076_WSD_id_48_WebService_tr_3821_1st_7083.csv: {'filename': '076_WSD_id_48_WebService_tr_3821_1st_7083.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 11, 'train_index': '3821', 'data_shape': '15285x1', 'output_length': 15000, 'label_length': 11464, 'aligned_length': 11464, 'AUC-PR': 0.21005334155559538, 'AUC-ROC': 0.9928609534895, 'VUS-PR': 0.21304489911418711, 'VUS-ROC': 0.994866212103044, 'Standard-F1': 0.0967741935482509, 'Standard-Precision': 0.05217391304347781, 'Standard-Recall': 0.6666666666665926, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.788202449414801, 'Affiliation-P': 0.6504622638184996, 'Affiliation-R': 0.9999490393925496, 'F1_T': 0.0967741935482509, 'Precision_T': 0.05217391304347781, 'Recall_T': 0.6666666666665926}
Logits metrics skipped for 076_WSD_id_48_WebService_tr_3821_1st_7083.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/076_WSD_id_48_WebService_tr_3821_1st_7083_results.pkl
Processing file: 337_UCR_id_35_HumanActivity_tr_50000_1st_110260.csv, Data shape: (200000, 1)
Label shape: (200000,)
Train data shape: (50000, 1), Test data shape: (150000, 1), Label test shape: (150000,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 337_UCR_id_35_HumanActivity_tr_50000_1st_110260.csv...
Using CUDA device: 2
Data shape: (150000, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([30, 5000, 1])
batched_mask shape: torch.Size([30, 5000])

100%|██████████| 1/1 [00:00<00:00, 46.67it/s]
Original shapes - Output: (150000,), Label: (150000,)
Aligned shapes - Output: (150000,), Label: (150000,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  14%|█▎        | 27/200 [00:01<00:06, 26.02it/s]

Calculating Affiliation Metrics:  27%|██▋       | 54/200 [00:02<00:06, 23.26it/s]

Calculating Affiliation Metrics:  39%|███▉      | 78/200 [00:03<00:05, 22.36it/s]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [00:04<00:04, 22.37it/s]

Calculating Affiliation Metrics:  63%|██████▎   | 126/200 [00:05<00:03, 23.10it/s]

Calculating Affiliation Metrics:  76%|███████▋  | 153/200 [00:06<00:01, 24.15it/s]

Calculating Affiliation Metrics:  97%|█████████▋| 194/200 [00:07<00:00, 29.39it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:07<00:00, 26.36it/s]
Table1-only metrics completed in 12.83s
Evaluation Result:  {'AUC-PR': 0.03185107497254552, 'AUC-ROC': 0.9850249349165049, 'VUS-PR': 0.03528602353735027, 'VUS-ROC': 0.9850685467948751, 'Standard-F1': 0.07939339875098862, 'Standard-Precision': 0.04258373205741625, 'Standard-Recall': 0.5855263157894699, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7326675703213004, 'Affiliation-P': 0.578122230327144, 'Affiliation-R': 0.9999870614035087, 'F1_T': 0.07862818904212261, 'Precision_T': 0.04198302813756139, 'Recall_T': 0.6184210526315749}
Results for 337_UCR_id_35_HumanActivity_tr_50000_1st_110260.csv: {'filename': '337_UCR_id_35_HumanActivity_tr_50000_1st_110260.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '50000', 'data_shape': '200000x1', 'output_length': 150000, 'label_length': 150000, 'aligned_length': 150000, 'AUC-PR': 0.03185107497254552, 'AUC-ROC': 0.9850249349165049, 'VUS-PR': 0.03528602353735027, 'VUS-ROC': 0.9850685467948751, 'Standard-F1': 0.07939339875098862, 'Standard-Precision': 0.04258373205741625, 'Standard-Recall': 0.5855263157894699, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7326675703213004, 'Affiliation-P': 0.578122230327144, 'Affiliation-R': 0.9999870614035087, 'F1_T': 0.07862818904212261, 'Precision_T': 0.04198302813756139, 'Recall_T': 0.6184210526315749}
Logits metrics skipped for 337_UCR_id_35_HumanActivity_tr_50000_1st_110260.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/337_UCR_id_35_HumanActivity_tr_50000_1st_110260_results.pkl
Processing file: 150_Stock_id_2_Finance_tr_500_1st_7.csv, Data shape: (10000, 1)
Label shape: (10000,)
Train data shape: (500, 1), Test data shape: (9500, 1), Label test shape: (9500,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 150_Stock_id_2_Finance_tr_500_1st_7.csv...
Using CUDA device: 2
Data shape: (9500, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 9500, window_size 5000, adding 500 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 157.00it/s]
Original shapes - Output: (10000,), Label: (9500,)
Aligned shapes - Output: (9500,), Label: (9500,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:   7%|▋         | 14/200 [00:01<00:13, 13.82it/s]

Calculating Affiliation Metrics:  14%|█▍        | 28/200 [00:03<00:25,  6.62it/s]

Calculating Affiliation Metrics:  18%|█▊        | 37/200 [00:06<00:30,  5.32it/s]

Calculating Affiliation Metrics:  22%|██▏       | 44/200 [00:08<00:36,  4.26it/s]

Calculating Affiliation Metrics:  25%|██▌       | 50/200 [00:11<00:42,  3.50it/s]

Calculating Affiliation Metrics:  28%|██▊       | 55/200 [00:13<00:47,  3.05it/s]

Calculating Affiliation Metrics:  30%|██▉       | 59/200 [00:15<00:50,  2.77it/s]

Calculating Affiliation Metrics:  31%|███       | 62/200 [00:17<00:53,  2.57it/s]

Calculating Affiliation Metrics:  32%|███▎      | 65/200 [00:18<00:56,  2.38it/s]

Calculating Affiliation Metrics:  34%|███▍      | 68/200 [00:20<00:59,  2.23it/s]

Calculating Affiliation Metrics:  36%|███▌      | 71/200 [00:21<01:00,  2.13it/s]

Calculating Affiliation Metrics:  37%|███▋      | 74/200 [00:23<01:01,  2.05it/s]

Calculating Affiliation Metrics:  38%|███▊      | 77/200 [00:25<01:01,  1.98it/s]

Calculating Affiliation Metrics:  40%|███▉      | 79/200 [00:26<01:02,  1.94it/s]

Calculating Affiliation Metrics:  40%|████      | 81/200 [00:27<01:02,  1.91it/s]

Calculating Affiliation Metrics:  42%|████▏     | 83/200 [00:28<01:02,  1.88it/s]

Calculating Affiliation Metrics:  42%|████▎     | 85/200 [00:29<01:02,  1.84it/s]

Calculating Affiliation Metrics:  44%|████▎     | 87/200 [00:30<01:01,  1.84it/s]

Calculating Affiliation Metrics:  44%|████▍     | 89/200 [00:31<01:00,  1.83it/s]

Calculating Affiliation Metrics:  46%|████▌     | 91/200 [00:33<00:59,  1.83it/s]

Calculating Affiliation Metrics:  46%|████▋     | 93/200 [00:34<00:58,  1.83it/s]

Calculating Affiliation Metrics:  48%|████▊     | 95/200 [00:35<00:56,  1.84it/s]

Calculating Affiliation Metrics:  48%|████▊     | 97/200 [00:36<00:55,  1.85it/s]

Calculating Affiliation Metrics:  50%|████▉     | 99/200 [00:37<00:53,  1.87it/s]

Calculating Affiliation Metrics:  50%|█████     | 101/200 [00:38<00:52,  1.90it/s]

Calculating Affiliation Metrics:  52%|█████▏    | 104/200 [00:39<00:49,  1.95it/s]

Calculating Affiliation Metrics:  54%|█████▎    | 107/200 [00:41<00:46,  2.00it/s]

Calculating Affiliation Metrics:  55%|█████▌    | 110/200 [00:42<00:43,  2.07it/s]

Calculating Affiliation Metrics:  56%|█████▋    | 113/200 [00:43<00:40,  2.17it/s]

Calculating Affiliation Metrics:  58%|█████▊    | 116/200 [00:44<00:36,  2.31it/s]

Calculating Affiliation Metrics:  60%|█████▉    | 119/200 [00:45<00:32,  2.48it/s]

Calculating Affiliation Metrics:  62%|██████▏   | 123/200 [00:47<00:28,  2.72it/s]

Calculating Affiliation Metrics:  64%|██████▎   | 127/200 [00:48<00:25,  2.88it/s]

Calculating Affiliation Metrics:  66%|██████▌   | 131/200 [00:49<00:22,  3.08it/s]

Calculating Affiliation Metrics:  68%|██████▊   | 135/200 [00:50<00:19,  3.30it/s]

Calculating Affiliation Metrics:  70%|███████   | 140/200 [00:51<00:16,  3.68it/s]

Calculating Affiliation Metrics:  73%|███████▎  | 146/200 [00:52<00:12,  4.29it/s]

Calculating Affiliation Metrics:  77%|███████▋  | 154/200 [00:53<00:08,  5.27it/s]

Calculating Affiliation Metrics:  82%|████████▏ | 164/200 [00:54<00:05,  6.53it/s]

Calculating Affiliation Metrics:  90%|█████████ | 180/200 [00:55<00:02,  9.17it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:56<00:00,  3.55it/s]
Table1-only metrics completed in 69.95s
Evaluation Result:  {'AUC-PR': 0.14882836290416537, 'AUC-ROC': 0.5250054556532551, 'VUS-PR': 0.7145690150040707, 'VUS-ROC': 0.8299720035711085, 'Standard-F1': 0.2498837965973445, 'Standard-Precision': 0.14290271132376392, 'Standard-Recall': 0.9940828402366857, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.68757328913071, 'Affiliation-P': 0.52401228840748, 'Affiliation-R': 0.9995715509854327, 'F1_T': 0.2498837965973445, 'Precision_T': 0.14290271132376392, 'Recall_T': 0.9940828402366857}
Results for 150_Stock_id_2_Finance_tr_500_1st_7.csv: {'filename': '150_Stock_id_2_Finance_tr_500_1st_7.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '10000x1', 'output_length': 10000, 'label_length': 9500, 'aligned_length': 9500, 'AUC-PR': 0.14882836290416537, 'AUC-ROC': 0.5250054556532551, 'VUS-PR': 0.7145690150040707, 'VUS-ROC': 0.8299720035711085, 'Standard-F1': 0.2498837965973445, 'Standard-Precision': 0.14290271132376392, 'Standard-Recall': 0.9940828402366857, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.68757328913071, 'Affiliation-P': 0.52401228840748, 'Affiliation-R': 0.9995715509854327, 'F1_T': 0.2498837965973445, 'Precision_T': 0.14290271132376392, 'Recall_T': 0.9940828402366857}
Logits metrics skipped for 150_Stock_id_2_Finance_tr_500_1st_7.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/150_Stock_id_2_Finance_tr_500_1st_7_results.pkl
Skipping file: 219_LTDB_id_4_Medical_tr_3618_1st_3718.csv due to filter criteria.
Processing file: 745_YAHOO_id_195_WebService_tr_500_1st_326.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 745_YAHOO_id_195_WebService_tr_500_1st_326.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 146.77it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 963.19it/s]
Table1-only metrics completed in 0.44s
Evaluation Result:  {'AUC-PR': 0.9659090909090909, 'AUC-ROC': 0.9996800341296929, 'VUS-PR': 0.9048682374366387, 'VUS-ROC': 0.9994028510666338, 'Standard-F1': 0.7999999999994398, 'Standard-Precision': 0.6666666666666111, 'Standard-Recall': 0.999999999999875, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9841269841269835, 'Affiliation-P': 0.96875, 'Affiliation-R': 1.0, 'F1_T': 0.7999999999994398, 'Precision_T': 0.6666666666666111, 'Recall_T': 0.999999999999875}
Results for 745_YAHOO_id_195_WebService_tr_500_1st_326.csv: {'filename': '745_YAHOO_id_195_WebService_tr_500_1st_326.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 11, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.9659090909090909, 'AUC-ROC': 0.9996800341296929, 'VUS-PR': 0.9048682374366387, 'VUS-ROC': 0.9994028510666338, 'Standard-F1': 0.7999999999994398, 'Standard-Precision': 0.6666666666666111, 'Standard-Recall': 0.999999999999875, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9841269841269835, 'Affiliation-P': 0.96875, 'Affiliation-R': 1.0, 'F1_T': 0.7999999999994398, 'Precision_T': 0.6666666666666111, 'Recall_T': 0.999999999999875}
Logits metrics skipped for 745_YAHOO_id_195_WebService_tr_500_1st_326.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/745_YAHOO_id_195_WebService_tr_500_1st_326_results.pkl
Skipping file: 254_SVDB_id_18_Medical_tr_30824_1st_30924.csv due to filter criteria.
Processing file: 036_WSD_id_8_WebService_tr_4504_1st_6898.csv, Data shape: (18016, 1)
Label shape: (18016,)
Train data shape: (4504, 1), Test data shape: (13512, 1), Label test shape: (13512,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 036_WSD_id_8_WebService_tr_4504_1st_6898.csv...
Using CUDA device: 2
Data shape: (13512, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13512, window_size 5000, adding 1488 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.23it/s]
Original shapes - Output: (15000,), Label: (13512,)
Aligned shapes - Output: (13512,), Label: (13512,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 290.90it/s]
Table1-only metrics completed in 1.53s
Evaluation Result:  {'AUC-PR': 0.3371640029131417, 'AUC-ROC': 0.9597339454069795, 'VUS-PR': 0.09803485333957579, 'VUS-ROC': 0.9698151711663535, 'Standard-F1': 0.028169014084425712, 'Standard-Precision': 0.014705882352941069, 'Standard-Recall': 0.33333333333327775, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9396821582957225, 'Affiliation-P': 0.8863172449857757, 'Affiliation-R': 0.9998849514321155, 'F1_T': 0.028169014084425712, 'Precision_T': 0.014705882352941069, 'Recall_T': 0.33333333333327775}
Results for 036_WSD_id_8_WebService_tr_4504_1st_6898.csv: {'filename': '036_WSD_id_8_WebService_tr_4504_1st_6898.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '4504', 'data_shape': '18016x1', 'output_length': 15000, 'label_length': 13512, 'aligned_length': 13512, 'AUC-PR': 0.3371640029131417, 'AUC-ROC': 0.9597339454069795, 'VUS-PR': 0.09803485333957579, 'VUS-ROC': 0.9698151711663535, 'Standard-F1': 0.028169014084425712, 'Standard-Precision': 0.014705882352941069, 'Standard-Recall': 0.33333333333327775, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9396821582957225, 'Affiliation-P': 0.8863172449857757, 'Affiliation-R': 0.9998849514321155, 'F1_T': 0.028169014084425712, 'Precision_T': 0.014705882352941069, 'Recall_T': 0.33333333333327775}
Logits metrics skipped for 036_WSD_id_8_WebService_tr_4504_1st_6898.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/036_WSD_id_8_WebService_tr_4504_1st_6898_results.pkl
Processing file: 765_YAHOO_id_215_WebService_tr_500_1st_92.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 765_YAHOO_id_215_WebService_tr_500_1st_92.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.66it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1156.16it/s]
Table1-only metrics completed in 0.58s
Evaluation Result:  {'AUC-PR': 0.5967276951225383, 'AUC-ROC': 0.9529898916088174, 'VUS-PR': 0.4094892004342908, 'VUS-ROC': 0.950979413341698, 'Standard-F1': 0.42105263157843764, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.5714285714284898, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8258723174656304, 'Affiliation-P': 0.7033922543100043, 'Affiliation-R': 1.0, 'F1_T': 0.42105263157843764, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.5714285714284898}
Results for 765_YAHOO_id_215_WebService_tr_500_1st_92.csv: {'filename': '765_YAHOO_id_215_WebService_tr_500_1st_92.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.5967276951225383, 'AUC-ROC': 0.9529898916088174, 'VUS-PR': 0.4094892004342908, 'VUS-ROC': 0.950979413341698, 'Standard-F1': 0.42105263157843764, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.5714285714284898, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8258723174656304, 'Affiliation-P': 0.7033922543100043, 'Affiliation-R': 1.0, 'F1_T': 0.42105263157843764, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.5714285714284898}
Logits metrics skipped for 765_YAHOO_id_215_WebService_tr_500_1st_92.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/765_YAHOO_id_215_WebService_tr_500_1st_92_results.pkl
Processing file: 279_NEK_id_3_WebService_tr_500_1st_974.csv, Data shape: (1073, 1)
Label shape: (1073,)
Train data shape: (500, 1), Test data shape: (573, 1), Label test shape: (573,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 279_NEK_id_3_WebService_tr_500_1st_974.csv...
Using CUDA device: 2
Data shape: (573, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 573, 1])
batched_mask shape: torch.Size([1, 573])

100%|██████████| 1/1 [00:00<00:00, 160.49it/s]
Original shapes - Output: (573,), Label: (573,)
Aligned shapes - Output: (573,), Label: (573,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2710.69it/s]
Table1-only metrics completed in 0.17s
Evaluation Result:  {'AUC-PR': 0.996255055752604, 'AUC-ROC': 0.9992476403264785, 'VUS-PR': 0.9963199470233262, 'VUS-ROC': 0.9993312176265161, 'Standard-F1': 0.9720670391056345, 'Standard-Precision': 0.9886363636363524, 'Standard-Recall': 0.9560439560439455, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.999501122474432, 'Affiliation-P': 1.0, 'Affiliation-R': 0.99900274245824, 'F1_T': 0.9720670391056345, 'Precision_T': 0.9886363636363524, 'Recall_T': 0.9560439560439455}
Results for 279_NEK_id_3_WebService_tr_500_1st_974.csv: {'filename': '279_NEK_id_3_WebService_tr_500_1st_974.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1073x1', 'output_length': 573, 'label_length': 573, 'aligned_length': 573, 'AUC-PR': 0.996255055752604, 'AUC-ROC': 0.9992476403264785, 'VUS-PR': 0.9963199470233262, 'VUS-ROC': 0.9993312176265161, 'Standard-F1': 0.9720670391056345, 'Standard-Precision': 0.9886363636363524, 'Standard-Recall': 0.9560439560439455, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.999501122474432, 'Affiliation-P': 1.0, 'Affiliation-R': 0.99900274245824, 'F1_T': 0.9720670391056345, 'Precision_T': 0.9886363636363524, 'Recall_T': 0.9560439560439455}
Logits metrics skipped for 279_NEK_id_3_WebService_tr_500_1st_974.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/279_NEK_id_3_WebService_tr_500_1st_974_results.pkl
Skipping file: 173_MITDB_id_4_Medical_tr_13620_1st_13720.csv due to filter criteria.
Processing file: 732_YAHOO_id_182_Synthetic_tr_500_1st_658.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 732_YAHOO_id_182_Synthetic_tr_500_1st_658.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 162.98it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2078.65it/s]
Table1-only metrics completed in 1.11s
Evaluation Result:  {'AUC-PR': 0.95, 'AUC-ROC': 0.9997273718647764, 'VUS-PR': 0.9936017243543619, 'VUS-ROC': 0.9999789456448585, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Results for 732_YAHOO_id_182_Synthetic_tr_500_1st_658.csv: {'filename': '732_YAHOO_id_182_Synthetic_tr_500_1st_658.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 184, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.95, 'AUC-ROC': 0.9997273718647764, 'VUS-PR': 0.9936017243543619, 'VUS-ROC': 0.9999789456448585, 'Standard-F1': 0.5714285714280816, 'Standard-Precision': 0.39999999999996, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9990950226244338, 'Affiliation-P': 0.9981916817359855, 'Affiliation-R': 1.0, 'F1_T': 0.5714285714280816, 'Precision_T': 0.39999999999996, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 732_YAHOO_id_182_Synthetic_tr_500_1st_658.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/732_YAHOO_id_182_Synthetic_tr_500_1st_658_results.pkl
Processing file: 624_YAHOO_id_74_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 624_YAHOO_id_74_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 161.90it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1752.81it/s]
Table1-only metrics completed in 0.53s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984775989560672, 'Affiliation-P': 0.9969598262757872, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 624_YAHOO_id_74_Synthetic_tr_500_1st_893.csv: {'filename': '624_YAHOO_id_74_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984775989560672, 'Affiliation-P': 0.9969598262757872, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 624_YAHOO_id_74_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/624_YAHOO_id_74_Synthetic_tr_500_1st_893_results.pkl
Processing file: 642_YAHOO_id_92_WebService_tr_500_1st_433.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 642_YAHOO_id_92_WebService_tr_500_1st_433.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 164.72it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1593.22it/s]
Table1-only metrics completed in 0.25s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.951611191584658, 'VUS-ROC': 0.9998971195953598, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9992706053975196, 'Affiliation-P': 0.9985422740524782, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Results for 642_YAHOO_id_92_WebService_tr_500_1st_433.csv: {'filename': '642_YAHOO_id_92_WebService_tr_500_1st_433.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 11, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.951611191584658, 'VUS-ROC': 0.9998971195953598, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9992706053975196, 'Affiliation-P': 0.9985422740524782, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 642_YAHOO_id_92_WebService_tr_500_1st_433.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/642_YAHOO_id_92_WebService_tr_500_1st_433_results.pkl
Processing file: 048_WSD_id_20_WebService_tr_4559_1st_11813.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 048_WSD_id_20_WebService_tr_4559_1st_11813.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 164.13it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 254.76it/s]
Table1-only metrics completed in 1.04s
Evaluation Result:  {'AUC-PR': 0.2350238952860224, 'AUC-ROC': 0.6861434870760811, 'VUS-PR': 0.2322618927831072, 'VUS-ROC': 0.8128967076115892, 'Standard-F1': 0.263157894736437, 'Standard-Precision': 0.18248175182481619, 'Standard-Recall': 0.47169811320753824, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.901433846451856, 'Affiliation-P': 0.821287253815526, 'Affiliation-R': 0.9989144666154146, 'F1_T': 0.263157894736437, 'Precision_T': 0.18248175182481619, 'Recall_T': 0.47169811320753824}
Results for 048_WSD_id_20_WebService_tr_4559_1st_11813.csv: {'filename': '048_WSD_id_20_WebService_tr_4559_1st_11813.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 30, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.2350238952860224, 'AUC-ROC': 0.6861434870760811, 'VUS-PR': 0.2322618927831072, 'VUS-ROC': 0.8128967076115892, 'Standard-F1': 0.263157894736437, 'Standard-Precision': 0.18248175182481619, 'Standard-Recall': 0.47169811320753824, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.901433846451856, 'Affiliation-P': 0.821287253815526, 'Affiliation-R': 0.9989144666154146, 'F1_T': 0.263157894736437, 'Precision_T': 0.18248175182481619, 'Recall_T': 0.47169811320753824}
Logits metrics skipped for 048_WSD_id_20_WebService_tr_4559_1st_11813.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/048_WSD_id_20_WebService_tr_4559_1st_11813_results.pkl
Processing file: 772_YAHOO_id_222_WebService_tr_500_1st_85.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 772_YAHOO_id_222_WebService_tr_500_1st_85.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 166.52it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 900.51it/s]
Table1-only metrics completed in 0.70s
Evaluation Result:  {'AUC-PR': 0.8913951545530495, 'AUC-ROC': 0.9629945915172218, 'VUS-PR': 0.878478995863544, 'VUS-ROC': 0.9627751691801639, 'Standard-F1': 0.7619047619041995, 'Standard-Precision': 0.6666666666666111, 'Standard-Recall': 0.8888888888887901, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9399222155477527, 'Affiliation-P': 0.9971723665550496, 'Affiliation-R': 0.8888888888888888, 'F1_T': 0.7619047619041995, 'Precision_T': 0.6666666666666111, 'Recall_T': 0.8888888888887901}
Results for 772_YAHOO_id_222_WebService_tr_500_1st_85.csv: {'filename': '772_YAHOO_id_222_WebService_tr_500_1st_85.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 0.8913951545530495, 'AUC-ROC': 0.9629945915172218, 'VUS-PR': 0.878478995863544, 'VUS-ROC': 0.9627751691801639, 'Standard-F1': 0.7619047619041995, 'Standard-Precision': 0.6666666666666111, 'Standard-Recall': 0.8888888888887901, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9399222155477527, 'Affiliation-P': 0.9971723665550496, 'Affiliation-R': 0.8888888888888888, 'F1_T': 0.7619047619041995, 'Precision_T': 0.6666666666666111, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 772_YAHOO_id_222_WebService_tr_500_1st_85.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/772_YAHOO_id_222_WebService_tr_500_1st_85_results.pkl
Skipping file: 824_Exathlon_id_15_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 371_UCR_id_69_Medical_tr_1853_1st_6200.csv, Data shape: (7415, 1)
Label shape: (7415,)
Train data shape: (1853, 1), Test data shape: (5562, 1), Label test shape: (5562,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 371_UCR_id_69_Medical_tr_1853_1st_6200.csv...
Using CUDA device: 2
Data shape: (5562, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 5562, window_size 5000, adding 4438 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.43it/s]
Original shapes - Output: (10000,), Label: (5562,)
Aligned shapes - Output: (5562,), Label: (5562,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 717.30it/s]
Table1-only metrics completed in 0.94s
Evaluation Result:  {'AUC-PR': 0.465922988527882, 'AUC-ROC': 0.862718188165474, 'VUS-PR': 0.4595418213438537, 'VUS-ROC': 0.886416644657899, 'Standard-F1': 0.47284345047873394, 'Standard-Precision': 0.5174825174825138, 'Standard-Recall': 0.4352941176470563, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9933087943878934, 'Affiliation-P': 1.0, 'Affiliation-R': 0.986706538062906, 'F1_T': 0.4805194805189828, 'Precision_T': 0.5362318840579672, 'Recall_T': 0.4352941176470563}
Results for 371_UCR_id_69_Medical_tr_1853_1st_6200.csv: {'filename': '371_UCR_id_69_Medical_tr_1853_1st_6200.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 173, 'train_index': '1853', 'data_shape': '7415x1', 'output_length': 10000, 'label_length': 5562, 'aligned_length': 5562, 'AUC-PR': 0.465922988527882, 'AUC-ROC': 0.862718188165474, 'VUS-PR': 0.4595418213438537, 'VUS-ROC': 0.886416644657899, 'Standard-F1': 0.47284345047873394, 'Standard-Precision': 0.5174825174825138, 'Standard-Recall': 0.4352941176470563, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9933087943878934, 'Affiliation-P': 1.0, 'Affiliation-R': 0.986706538062906, 'F1_T': 0.4805194805189828, 'Precision_T': 0.5362318840579672, 'Recall_T': 0.4352941176470563}
Logits metrics skipped for 371_UCR_id_69_Medical_tr_1853_1st_6200.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/371_UCR_id_69_Medical_tr_1853_1st_6200_results.pkl
Processing file: 025_NAB_id_25_WebService_tr_3958_1st_4614.csv, Data shape: (15832, 1)
Label shape: (15832,)
Train data shape: (3958, 1), Test data shape: (11874, 1), Label test shape: (11874,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 025_NAB_id_25_WebService_tr_3958_1st_4614.csv...
Using CUDA device: 2
Data shape: (11874, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 11874, window_size 5000, adding 3126 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.50it/s]
Original shapes - Output: (15000,), Label: (11874,)
Aligned shapes - Output: (11874,), Label: (11874,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  69%|██████▉   | 138/200 [00:01<00:00, 137.26it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 155.64it/s]
Table1-only metrics completed in 1.46s
Evaluation Result:  {'AUC-PR': 0.09771394430605605, 'AUC-ROC': 0.3564089810212659, 'VUS-PR': 0.09800637210134674, 'VUS-ROC': 0.35896932852124086, 'Standard-F1': 0.23191404695560974, 'Standard-Precision': 0.13263541192535272, 'Standard-Recall': 0.9221518987341767, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7228114676964422, 'Affiliation-P': 0.5848980008380272, 'Affiliation-R': 0.94582910660111, 'F1_T': 0.23193224095600687, 'Precision_T': 0.132543200940921, 'Recall_T': 0.9272151898734172}
Results for 025_NAB_id_25_WebService_tr_3958_1st_4614.csv: {'filename': '025_NAB_id_25_WebService_tr_3958_1st_4614.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 16, 'train_index': '3958', 'data_shape': '15832x1', 'output_length': 15000, 'label_length': 11874, 'aligned_length': 11874, 'AUC-PR': 0.09771394430605605, 'AUC-ROC': 0.3564089810212659, 'VUS-PR': 0.09800637210134674, 'VUS-ROC': 0.35896932852124086, 'Standard-F1': 0.23191404695560974, 'Standard-Precision': 0.13263541192535272, 'Standard-Recall': 0.9221518987341767, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7228114676964422, 'Affiliation-P': 0.5848980008380272, 'Affiliation-R': 0.94582910660111, 'F1_T': 0.23193224095600687, 'Precision_T': 0.132543200940921, 'Recall_T': 0.9272151898734172}
Logits metrics skipped for 025_NAB_id_25_WebService_tr_3958_1st_4614.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/025_NAB_id_25_WebService_tr_3958_1st_4614_results.pkl
Processing file: 612_YAHOO_id_62_WebService_tr_500_1st_97.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 612_YAHOO_id_62_WebService_tr_500_1st_97.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 163.08it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 663.02it/s]
Table1-only metrics completed in 1.01s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.9999999999994333, 'Standard-Precision': 0.9999999999999333, 'Standard-Recall': 0.9999999999999333, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984583364008545, 'Affiliation-P': 0.9969214189380702, 'Affiliation-R': 1.0, 'F1_T': 0.9090909090903582, 'Precision_T': 0.8333333333332871, 'Recall_T': 0.9999999999999333}
Results for 612_YAHOO_id_62_WebService_tr_500_1st_97.csv: {'filename': '612_YAHOO_id_62_WebService_tr_500_1st_97.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.9999999999994333, 'Standard-Precision': 0.9999999999999333, 'Standard-Recall': 0.9999999999999333, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9984583364008545, 'Affiliation-P': 0.9969214189380702, 'Affiliation-R': 1.0, 'F1_T': 0.9090909090903582, 'Precision_T': 0.8333333333332871, 'Recall_T': 0.9999999999999333}
Logits metrics skipped for 612_YAHOO_id_62_WebService_tr_500_1st_97.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/612_YAHOO_id_62_WebService_tr_500_1st_97_results.pkl
Processing file: 650_YAHOO_id_100_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 650_YAHOO_id_100_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 160.07it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1745.31it/s]
Table1-only metrics completed in 0.79s
Evaluation Result:  {'AUC-PR': 0.9013548180214848, 'AUC-ROC': 0.9989035087719298, 'VUS-PR': 0.9937762152964235, 'VUS-ROC': 0.9999426047313061, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Results for 650_YAHOO_id_100_Synthetic_tr_500_1st_623.csv: {'filename': '650_YAHOO_id_100_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 115, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9013548180214848, 'AUC-ROC': 0.9989035087719298, 'VUS-PR': 0.9937762152964235, 'VUS-ROC': 0.9999426047313061, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9968939875976737, 'Affiliation-P': 0.9941482496697848, 'Affiliation-R': 0.999654934437543, 'F1_T': 0.7826086956516295, 'Precision_T': 0.6428571428570969, 'Recall_T': 0.9999999999998889}
Logits metrics skipped for 650_YAHOO_id_100_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/650_YAHOO_id_100_Synthetic_tr_500_1st_623_results.pkl
Processing file: 790_YAHOO_id_240_WebService_tr_500_1st_119.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 790_YAHOO_id_240_WebService_tr_500_1st_119.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 165.99it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1276.31it/s]
Table1-only metrics completed in 0.41s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.999944492534466, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9987876914864471, 'Affiliation-P': 0.9975783187976339, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Results for 790_YAHOO_id_240_WebService_tr_500_1st_119.csv: {'filename': '790_YAHOO_id_240_WebService_tr_500_1st_119.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.973893419489478, 'VUS-ROC': 0.999944492534466, 'Standard-F1': 0.4999999999995625, 'Standard-Precision': 0.33333333333330556, 'Standard-Recall': 0.99999999999975, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9987876914864471, 'Affiliation-P': 0.9975783187976339, 'Affiliation-R': 1.0, 'F1_T': 0.4999999999995625, 'Precision_T': 0.33333333333330556, 'Recall_T': 0.99999999999975}
Logits metrics skipped for 790_YAHOO_id_240_WebService_tr_500_1st_119.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/790_YAHOO_id_240_WebService_tr_500_1st_119_results.pkl
Processing file: 688_YAHOO_id_138_WebService_tr_500_1st_482.csv, Data shape: (1680, 1)
Label shape: (1680,)
Train data shape: (500, 1), Test data shape: (1180, 1), Label test shape: (1180,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 688_YAHOO_id_138_WebService_tr_500_1st_482.csv...
Using CUDA device: 2
Data shape: (1180, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 1180, 1])
batched_mask shape: torch.Size([1, 1180])

100%|██████████| 1/1 [00:00<00:00, 166.83it/s]
Original shapes - Output: (1180,), Label: (1180,)
Aligned shapes - Output: (1180,), Label: (1180,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1545.72it/s]
Table1-only metrics completed in 0.28s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.8493019703517117, 'VUS-ROC': 0.999680059109235, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9959349593495929, 'Affiliation-P': 0.9919028340080972, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Results for 688_YAHOO_id_138_WebService_tr_500_1st_482.csv: {'filename': '688_YAHOO_id_138_WebService_tr_500_1st_482.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1680x1', 'output_length': 1180, 'label_length': 1180, 'aligned_length': 1180, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 0.8493019703517117, 'VUS-ROC': 0.999680059109235, 'Standard-F1': 0.285714285714, 'Standard-Precision': 0.16666666666665278, 'Standard-Recall': 0.9999999999995, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9959349593495929, 'Affiliation-P': 0.9919028340080972, 'Affiliation-R': 1.0, 'F1_T': 0.285714285714, 'Precision_T': 0.16666666666665278, 'Recall_T': 0.9999999999995}
Logits metrics skipped for 688_YAHOO_id_138_WebService_tr_500_1st_482.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/688_YAHOO_id_138_WebService_tr_500_1st_482_results.pkl
Processing file: 739_YAHOO_id_189_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 739_YAHOO_id_189_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 163.49it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2423.25it/s]
Table1-only metrics completed in 0.49s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7950782675815867, 'Affiliation-P': 0.6598588490770901, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 739_YAHOO_id_189_Synthetic_tr_500_1st_893.csv: {'filename': '739_YAHOO_id_189_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7950782675815867, 'Affiliation-P': 0.6598588490770901, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 739_YAHOO_id_189_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/739_YAHOO_id_189_Synthetic_tr_500_1st_893_results.pkl
Processing file: 113_WSD_id_85_WebService_tr_500_1st_335.csv, Data shape: (18242, 1)
Label shape: (18242,)
Train data shape: (500, 1), Test data shape: (17742, 1), Label test shape: (17742,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 113_WSD_id_85_WebService_tr_500_1st_335.csv...
Using CUDA device: 2
Data shape: (17742, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 17742, window_size 5000, adding 2258 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 165.65it/s]
Original shapes - Output: (20000,), Label: (17742,)
Aligned shapes - Output: (17742,), Label: (17742,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  78%|███████▊  | 155/200 [00:01<00:00, 154.75it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 175.29it/s]
Table1-only metrics completed in 3.42s
Evaluation Result:  {'AUC-PR': 0.30219545409317305, 'AUC-ROC': 0.8070630761751664, 'VUS-PR': 0.3570386382359561, 'VUS-ROC': 0.9664646454415015, 'Standard-F1': 0.21245421245375712, 'Standard-Precision': 0.16292134831460583, 'Standard-Recall': 0.3052631578947336, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9180900689287864, 'Affiliation-P': 0.8500107032759895, 'Affiliation-R': 0.9980241839813859, 'F1_T': 0.21245421245375712, 'Precision_T': 0.16292134831460583, 'Recall_T': 0.3052631578947336}
Results for 113_WSD_id_85_WebService_tr_500_1st_335.csv: {'filename': '113_WSD_id_85_WebService_tr_500_1st_335.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '18242x1', 'output_length': 20000, 'label_length': 17742, 'aligned_length': 17742, 'AUC-PR': 0.30219545409317305, 'AUC-ROC': 0.8070630761751664, 'VUS-PR': 0.3570386382359561, 'VUS-ROC': 0.9664646454415015, 'Standard-F1': 0.21245421245375712, 'Standard-Precision': 0.16292134831460583, 'Standard-Recall': 0.3052631578947336, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9180900689287864, 'Affiliation-P': 0.8500107032759895, 'Affiliation-R': 0.9980241839813859, 'F1_T': 0.21245421245375712, 'Precision_T': 0.16292134831460583, 'Recall_T': 0.3052631578947336}
Logits metrics skipped for 113_WSD_id_85_WebService_tr_500_1st_335.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/113_WSD_id_85_WebService_tr_500_1st_335_results.pkl
Processing file: 473_UCR_id_171_HumanActivity_tr_2617_1st_7290.csv, Data shape: (10468, 1)
Label shape: (10468,)
Train data shape: (2617, 1), Test data shape: (7851, 1), Label test shape: (7851,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 473_UCR_id_171_HumanActivity_tr_2617_1st_7290.csv...
Using CUDA device: 2
Data shape: (7851, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 7851, window_size 5000, adding 2149 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 161.25it/s]
Original shapes - Output: (10000,), Label: (7851,)
Aligned shapes - Output: (7851,), Label: (7851,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 351.61it/s]
Table1-only metrics completed in 0.70s
Evaluation Result:  {'AUC-PR': 0.0009314276017119566, 'AUC-ROC': 0.5500531123858083, 'VUS-PR': 0.001547393651727866, 'VUS-ROC': 0.6515797926666964, 'Standard-F1': 0.0026160889470215848, 'Standard-Precision': 0.001309757694826457, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9066144046612342, 'Affiliation-P': 0.8960196153356261, 'Affiliation-R': 0.9174627435995415, 'F1_T': 0.0026041666666640644, 'Precision_T': 0.0013037809647979137, 'Recall_T': 0.9999999999998334}
Results for 473_UCR_id_171_HumanActivity_tr_2617_1st_7290.csv: {'filename': '473_UCR_id_171_HumanActivity_tr_2617_1st_7290.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '2617', 'data_shape': '10468x1', 'output_length': 10000, 'label_length': 7851, 'aligned_length': 7851, 'AUC-PR': 0.0009314276017119566, 'AUC-ROC': 0.5500531123858083, 'VUS-PR': 0.001547393651727866, 'VUS-ROC': 0.6515797926666964, 'Standard-F1': 0.0026160889470215848, 'Standard-Precision': 0.001309757694826457, 'Standard-Recall': 0.9999999999998334, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9066144046612342, 'Affiliation-P': 0.8960196153356261, 'Affiliation-R': 0.9174627435995415, 'F1_T': 0.0026041666666640644, 'Precision_T': 0.0013037809647979137, 'Recall_T': 0.9999999999998334}
Logits metrics skipped for 473_UCR_id_171_HumanActivity_tr_2617_1st_7290.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/473_UCR_id_171_HumanActivity_tr_2617_1st_7290_results.pkl
Processing file: 302_Power_id_1_Facility_tr_7685_1st_7785.csv, Data shape: (35040, 1)
Label shape: (35040,)
Train data shape: (7685, 1), Test data shape: (27355, 1), Label test shape: (27355,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 302_Power_id_1_Facility_tr_7685_1st_7785.csv...
Using CUDA device: 2
Data shape: (27355, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 27355, window_size 5000, adding 2645 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([6, 5000, 1])
batched_mask shape: torch.Size([6, 5000])

100%|██████████| 1/1 [00:00<00:00, 143.13it/s]
Original shapes - Output: (30000,), Label: (27355,)
Aligned shapes - Output: (27355,), Label: (27355,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  51%|█████     | 102/200 [00:01<00:00, 101.83it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 120.33it/s]
Table1-only metrics completed in 3.10s
Evaluation Result:  {'AUC-PR': 0.1367011080129613, 'AUC-ROC': 0.5831801615000343, 'VUS-PR': 0.14632884726955503, 'VUS-ROC': 0.6049295275618976, 'Standard-F1': 0.23108816396741047, 'Standard-Precision': 0.1509761831748952, 'Standard-Recall': 0.4923333333333332, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8500440132679958, 'Affiliation-P': 0.774658763211324, 'Affiliation-R': 0.9416830919602465, 'F1_T': 0.23103287116337445, 'Precision_T': 0.1507420207359219, 'Recall_T': 0.4943333333333332}
Results for 302_Power_id_1_Facility_tr_7685_1st_7785.csv: {'filename': '302_Power_id_1_Facility_tr_7685_1st_7785.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 96, 'train_index': '7685', 'data_shape': '35040x1', 'output_length': 30000, 'label_length': 27355, 'aligned_length': 27355, 'AUC-PR': 0.1367011080129613, 'AUC-ROC': 0.5831801615000343, 'VUS-PR': 0.14632884726955503, 'VUS-ROC': 0.6049295275618976, 'Standard-F1': 0.23108816396741047, 'Standard-Precision': 0.1509761831748952, 'Standard-Recall': 0.4923333333333332, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.8500440132679958, 'Affiliation-P': 0.774658763211324, 'Affiliation-R': 0.9416830919602465, 'F1_T': 0.23103287116337445, 'Precision_T': 0.1507420207359219, 'Recall_T': 0.4943333333333332}
Logits metrics skipped for 302_Power_id_1_Facility_tr_7685_1st_7785.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/302_Power_id_1_Facility_tr_7685_1st_7785_results.pkl
Processing file: 504_UCR_id_202_Environment_tr_2046_1st_5549.csv, Data shape: (8184, 1)
Label shape: (8184,)
Train data shape: (2046, 1), Test data shape: (6138, 1), Label test shape: (6138,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 504_UCR_id_202_Environment_tr_2046_1st_5549.csv...
Using CUDA device: 2
Data shape: (6138, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 6138, window_size 5000, adding 3862 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([2, 5000, 1])
batched_mask shape: torch.Size([2, 5000])

100%|██████████| 1/1 [00:00<00:00, 163.50it/s]
Original shapes - Output: (10000,), Label: (6138,)
Aligned shapes - Output: (6138,), Label: (6138,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 503.92it/s]
Table1-only metrics completed in 0.52s
Evaluation Result:  {'AUC-PR': 0.8567482973395507, 'AUC-ROC': 0.99790469348659, 'VUS-PR': 0.8237276179189983, 'VUS-ROC': 0.998217240809187, 'Standard-F1': 0.6909090909085864, 'Standard-Precision': 0.6129032258064417, 'Standard-Recall': 0.7916666666666502, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9994850894598114, 'Affiliation-P': 0.9997424820525757, 'Affiliation-R': 0.9992278293689584, 'F1_T': 0.6909090909085864, 'Precision_T': 0.6129032258064417, 'Recall_T': 0.7916666666666502}
Results for 504_UCR_id_202_Environment_tr_2046_1st_5549.csv: {'filename': '504_UCR_id_202_Environment_tr_2046_1st_5549.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '2046', 'data_shape': '8184x1', 'output_length': 10000, 'label_length': 6138, 'aligned_length': 6138, 'AUC-PR': 0.8567482973395507, 'AUC-ROC': 0.99790469348659, 'VUS-PR': 0.8237276179189983, 'VUS-ROC': 0.998217240809187, 'Standard-F1': 0.6909090909085864, 'Standard-Precision': 0.6129032258064417, 'Standard-Recall': 0.7916666666666502, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9994850894598114, 'Affiliation-P': 0.9997424820525757, 'Affiliation-R': 0.9992278293689584, 'F1_T': 0.6909090909085864, 'Precision_T': 0.6129032258064417, 'Recall_T': 0.7916666666666502}
Logits metrics skipped for 504_UCR_id_202_Environment_tr_2046_1st_5549.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/504_UCR_id_202_Environment_tr_2046_1st_5549_results.pkl
Processing file: 284_NEK_id_8_WebService_tr_500_1st_976.csv, Data shape: (1073, 1)
Label shape: (1073,)
Train data shape: (500, 1), Test data shape: (573, 1), Label test shape: (573,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 284_NEK_id_8_WebService_tr_500_1st_976.csv...
Using CUDA device: 2
Data shape: (573, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 573, 1])
batched_mask shape: torch.Size([1, 573])

100%|██████████| 1/1 [00:00<00:00, 160.19it/s]
Original shapes - Output: (573,), Label: (573,)
Aligned shapes - Output: (573,), Label: (573,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 3188.70it/s]
Table1-only metrics completed in 0.15s
Evaluation Result:  {'AUC-PR': 0.9635816044349628, 'AUC-ROC': 0.9923159067694308, 'VUS-PR': 0.9666682571799946, 'VUS-ROC': 0.9950839400155271, 'Standard-F1': 0.9485714285709179, 'Standard-Precision': 0.9651162790697563, 'Standard-Recall': 0.932584269662911, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9967420033438145, 'Affiliation-P': 0.9948734729493892, 'Affiliation-R': 0.998617565739161, 'F1_T': 0.9425287356316734, 'Precision_T': 0.9647058823529299, 'Recall_T': 0.9213483146067313}
Results for 284_NEK_id_8_WebService_tr_500_1st_976.csv: {'filename': '284_NEK_id_8_WebService_tr_500_1st_976.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 24, 'train_index': '500', 'data_shape': '1073x1', 'output_length': 573, 'label_length': 573, 'aligned_length': 573, 'AUC-PR': 0.9635816044349628, 'AUC-ROC': 0.9923159067694308, 'VUS-PR': 0.9666682571799946, 'VUS-ROC': 0.9950839400155271, 'Standard-F1': 0.9485714285709179, 'Standard-Precision': 0.9651162790697563, 'Standard-Recall': 0.932584269662911, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9967420033438145, 'Affiliation-P': 0.9948734729493892, 'Affiliation-R': 0.998617565739161, 'F1_T': 0.9425287356316734, 'Precision_T': 0.9647058823529299, 'Recall_T': 0.9213483146067313}
Logits metrics skipped for 284_NEK_id_8_WebService_tr_500_1st_976.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/284_NEK_id_8_WebService_tr_500_1st_976_results.pkl
Processing file: 575_YAHOO_id_25_Synthetic_tr_500_1st_623.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 575_YAHOO_id_25_Synthetic_tr_500_1st_623.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 158.55it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 1624.41it/s]
Table1-only metrics completed in 1.04s
Evaluation Result:  {'AUC-PR': 0.9253086419753087, 'AUC-ROC': 0.9991471734892787, 'VUS-PR': 0.9953814724877547, 'VUS-ROC': 0.9999575377048268, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Results for 575_YAHOO_id_25_Synthetic_tr_500_1st_623.csv: {'filename': '575_YAHOO_id_25_Synthetic_tr_500_1st_623.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 157, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 0.9253086419753087, 'AUC-ROC': 0.9991471734892787, 'VUS-PR': 0.9953814724877547, 'VUS-ROC': 0.9999575377048268, 'Standard-F1': 0.8571428571422857, 'Standard-Precision': 0.7499999999999375, 'Standard-Recall': 0.9999999999998889, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9975000515719773, 'Affiliation-P': 0.9951834463571553, 'Affiliation-R': 0.9998274672187716, 'F1_T': 0.8421052631573075, 'Precision_T': 0.79999999999992, 'Recall_T': 0.8888888888887901}
Logits metrics skipped for 575_YAHOO_id_25_Synthetic_tr_500_1st_623.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/575_YAHOO_id_25_Synthetic_tr_500_1st_623_results.pkl
Processing file: 084_WSD_id_56_WebService_tr_4559_1st_9616.csv, Data shape: (18236, 1)
Label shape: (18236,)
Train data shape: (4559, 1), Test data shape: (13677, 1), Label test shape: (13677,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 084_WSD_id_56_WebService_tr_4559_1st_9616.csv...
Using CUDA device: 2
Data shape: (13677, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 13677, window_size 5000, adding 1323 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 164.40it/s]
Original shapes - Output: (15000,), Label: (13677,)
Aligned shapes - Output: (13677,), Label: (13677,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 275.39it/s]
Table1-only metrics completed in 0.87s
Evaluation Result:  {'AUC-PR': 0.08357130697415546, 'AUC-ROC': 0.9678563850005013, 'VUS-PR': 0.08980935448095562, 'VUS-ROC': 0.974698501402346, 'Standard-F1': 0.16825396825378772, 'Standard-Precision': 0.09347442680775997, 'Standard-Recall': 0.8412698412698278, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7495411153153823, 'Affiliation-P': 0.5998247444836609, 'Affiliation-R': 0.9988556851904078, 'F1_T': 0.16556291390709738, 'Precision_T': 0.0924214417744915, 'Recall_T': 0.7936507936507811}
Results for 084_WSD_id_56_WebService_tr_4559_1st_9616.csv: {'filename': '084_WSD_id_56_WebService_tr_4559_1st_9616.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 19, 'train_index': '4559', 'data_shape': '18236x1', 'output_length': 15000, 'label_length': 13677, 'aligned_length': 13677, 'AUC-PR': 0.08357130697415546, 'AUC-ROC': 0.9678563850005013, 'VUS-PR': 0.08980935448095562, 'VUS-ROC': 0.974698501402346, 'Standard-F1': 0.16825396825378772, 'Standard-Precision': 0.09347442680775997, 'Standard-Recall': 0.8412698412698278, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.7495411153153823, 'Affiliation-P': 0.5998247444836609, 'Affiliation-R': 0.9988556851904078, 'F1_T': 0.16556291390709738, 'Precision_T': 0.0924214417744915, 'Recall_T': 0.7936507936507811}
Logits metrics skipped for 084_WSD_id_56_WebService_tr_4559_1st_9616.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/084_WSD_id_56_WebService_tr_4559_1st_9616_results.pkl
Processing file: 042_WSD_id_14_WebService_tr_3015_1st_3115.csv, Data shape: (18308, 1)
Label shape: (18308,)
Train data shape: (3015, 1), Test data shape: (15293, 1), Label test shape: (15293,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 042_WSD_id_14_WebService_tr_3015_1st_3115.csv...
Using CUDA device: 2
Data shape: (15293, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 15293, window_size 5000, adding 4707 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([4, 5000, 1])
batched_mask shape: torch.Size([4, 5000])

100%|██████████| 1/1 [00:00<00:00, 159.86it/s]
Original shapes - Output: (20000,), Label: (15293,)
Aligned shapes - Output: (15293,), Label: (15293,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]

Calculating Affiliation Metrics:  84%|████████▍ | 168/200 [00:01<00:00, 167.94it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:01<00:00, 180.75it/s]
Table1-only metrics completed in 1.45s
Evaluation Result:  {'AUC-PR': 0.3644289690415825, 'AUC-ROC': 0.871012391247034, 'VUS-PR': 0.2776211145303648, 'VUS-ROC': 0.9239109353230536, 'Standard-F1': 0.37956204379512454, 'Standard-Precision': 0.3398692810457494, 'Standard-Recall': 0.42975206611569894, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.974753295274677, 'Affiliation-P': 0.9531148156592941, 'Affiliation-R': 0.9973971118630947, 'F1_T': 0.37956204379512454, 'Precision_T': 0.3398692810457494, 'Recall_T': 0.42975206611569894}
Results for 042_WSD_id_14_WebService_tr_3015_1st_3115.csv: {'filename': '042_WSD_id_14_WebService_tr_3015_1st_3115.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 10, 'train_index': '3015', 'data_shape': '18308x1', 'output_length': 20000, 'label_length': 15293, 'aligned_length': 15293, 'AUC-PR': 0.3644289690415825, 'AUC-ROC': 0.871012391247034, 'VUS-PR': 0.2776211145303648, 'VUS-ROC': 0.9239109353230536, 'Standard-F1': 0.37956204379512454, 'Standard-Precision': 0.3398692810457494, 'Standard-Recall': 0.42975206611569894, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.974753295274677, 'Affiliation-P': 0.9531148156592941, 'Affiliation-R': 0.9973971118630947, 'F1_T': 0.37956204379512454, 'Precision_T': 0.3398692810457494, 'Recall_T': 0.42975206611569894}
Logits metrics skipped for 042_WSD_id_14_WebService_tr_3015_1st_3115.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/042_WSD_id_14_WebService_tr_3015_1st_3115_results.pkl
Processing file: 135_WSD_id_107_WebService_tr_3779_1st_3879.csv, Data shape: (15506, 1)
Label shape: (15506,)
Train data shape: (3779, 1), Test data shape: (11727, 1), Label test shape: (11727,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 135_WSD_id_107_WebService_tr_3779_1st_3879.csv...
Using CUDA device: 2
Data shape: (11727, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda
Padding AnomalyClipDataset: original length 11727, window_size 5000, adding 3273 samples


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([3, 5000, 1])
batched_mask shape: torch.Size([3, 5000])

100%|██████████| 1/1 [00:00<00:00, 162.54it/s]
Original shapes - Output: (15000,), Label: (11727,)
Aligned shapes - Output: (11727,), Label: (11727,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 306.45it/s]
Table1-only metrics completed in 0.91s
Evaluation Result:  {'AUC-PR': 0.7972056500663106, 'AUC-ROC': 0.991333390425623, 'VUS-PR': 0.647616119370028, 'VUS-ROC': 0.9974003994290012, 'Standard-F1': 0.5238095238090995, 'Standard-Precision': 0.37288135593220023, 'Standard-Recall': 0.8799999999999824, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9875059826626559, 'Affiliation-P': 0.975981441049316, 'Affiliation-R': 0.9993059436534214, 'F1_T': 0.5238095238090995, 'Precision_T': 0.37288135593220023, 'Recall_T': 0.8799999999999824}
Results for 135_WSD_id_107_WebService_tr_3779_1st_3879.csv: {'filename': '135_WSD_id_107_WebService_tr_3779_1st_3879.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 19, 'train_index': '3779', 'data_shape': '15506x1', 'output_length': 15000, 'label_length': 11727, 'aligned_length': 11727, 'AUC-PR': 0.7972056500663106, 'AUC-ROC': 0.991333390425623, 'VUS-PR': 0.647616119370028, 'VUS-ROC': 0.9974003994290012, 'Standard-F1': 0.5238095238090995, 'Standard-Precision': 0.37288135593220023, 'Standard-Recall': 0.8799999999999824, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.9875059826626559, 'Affiliation-P': 0.975981441049316, 'Affiliation-R': 0.9993059436534214, 'F1_T': 0.5238095238090995, 'Precision_T': 0.37288135593220023, 'Recall_T': 0.8799999999999824}
Logits metrics skipped for 135_WSD_id_107_WebService_tr_3779_1st_3879.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/135_WSD_id_107_WebService_tr_3779_1st_3879_results.pkl
Skipping file: 839_Exathlon_id_30_Facility_tr_10766_1st_12590.csv due to filter criteria.
Processing file: 601_YAHOO_id_51_Synthetic_tr_500_1st_893.csv, Data shape: (1421, 1)
Label shape: (1421,)
Train data shape: (500, 1), Test data shape: (921, 1), Label test shape: (921,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 64}
Running Time_RCD on 601_YAHOO_id_51_Synthetic_tr_500_1st_893.csv...
Using CUDA device: 2
Data shape: (921, 1)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=16, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=1), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=64, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: best_model/pretrain_checkpoint_best_uni.pth
Successfully loaded checkpoint from best_model/pretrain_checkpoint_best_uni.pth
Model loaded on device: cuda


  0%|          | 0/1 [00:00<?, ?it/s]batched_time_series shape: torch.Size([1, 921, 1])
batched_mask shape: torch.Size([1, 921])

100%|██████████| 1/1 [00:00<00:00, 163.77it/s]
Original shapes - Output: (921,), Label: (921,)
Aligned shapes - Output: (921,), Label: (921,)


Calculating Affiliation Metrics:   0%|          | 0/200 [00:00<?, ?it/s]
Calculating Affiliation Metrics: 100%|██████████| 200/200 [00:00<00:00, 2295.76it/s]
Table1-only metrics completed in 0.49s
Evaluation Result:  {'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.799767540152155, 'Affiliation-P': 0.6663438682982656, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Results for 601_YAHOO_id_51_Synthetic_tr_500_1st_893.csv: {'filename': '601_YAHOO_id_51_Synthetic_tr_500_1st_893.csv', 'AD_Name': 'Time_RCD', 'sliding_window': 125, 'train_index': '500', 'data_shape': '1421x1', 'output_length': 921, 'label_length': 921, 'aligned_length': 921, 'AUC-PR': 1.0, 'AUC-ROC': 1.0, 'VUS-PR': 1.0, 'VUS-ROC': 1.0, 'Standard-F1': 0.18181818181798345, 'Standard-Precision': 0.09999999999999, 'Standard-Recall': 0.9999999999989999, 'PA-F1': 0.0, 'PA-Precision': 0.0, 'PA-Recall': 0.0, 'Affiliation-F': 0.799767540152155, 'Affiliation-P': 0.6663438682982656, 'Affiliation-R': 1.0, 'F1_T': 0.18181818181798345, 'Precision_T': 0.09999999999999, 'Recall_T': 0.9999999999989999}
Logits metrics skipped for 601_YAHOO_id_51_Synthetic_tr_500_1st_893.csv
Results saved to /timercd_full_eval_work/code/Time-RCD-main/Uni_Time_RCD/601_YAHOO_id_51_Synthetic_tr_500_1st_893_results.pkl
Skipping file: 218_LTDB_id_3_Medical_tr_500_1st_266.csv due to filter criteria.

All results saved to Uni_Time_RCD.csv
Total file processed: 700
Results shape: (700, 24)