CUDA Available:  True
cuDNN Version:  92000
Processing file: 097_SVDB_id_14_Medical_tr_1031_1st_1131.csv, Data shape: (230400, 2)
Label shape: (230400,)
Train data shape: (1031, 2), Test data shape: (229369, 2), Label test shape: (229369,)
Optimal Hyperparameters for Time_RCD: {'win_size': 5000, 'batch_size': 1}
Running Time_RCD on 097_SVDB_id_14_Medical_tr_1031_1st_1131.csv...
Using CUDA device: 2
Data shape: (229369, 2)
Config: TimeRCDConfig(ts_config=TimeSeriesConfig(d_model=512, d_proj=256, patch_size=32, num_query_tokens=1, num_layers=8, num_heads=8, d_ff_dropout=0.1, use_rope=True, activation='gelu', num_features=2), llm_config=LLMConfig(model_name='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', max_batch_size=20, num_fixed_tokens=30, num_prototype=1000, num_heads=8), deepspeed_config_path='experiments/configs/deepspeed_config.json', accelerate_config_path='experiments/configs/accelerate_config.yaml', batch_size=1, learning_rate=0.0001, num_epochs=1000, max_seq_len=512, dropout=0.1, accumulation_steps=1, weight_decay=1e-05, enable_ts_train=False, data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_dataset_20250614_133353', test_data_path='/home3/lantian/ChatTS/data/ChatTSQA/time_rcd_test_dataset_20250621_090751', pretrain_data_path='/home2/lijinbo/Projects/dataset/ChatTS/result', load_data=True, load_pretrain_path='experiments/checkpoints/pretrain/pretrain_checkpoint_latest.pth', load_path='experiments/checkpoints/checkpoint_step.pth', checkpoint_dir='experiments/checkpoints', log_freq=100, save_freq=10, save_step_freq=100, model_prefix='time_rcd_qa_by_pretrain', test_batch_limit=20, early_stopping_patience=7, seed=72, cuda_devices='2', dist_port='12355', device='cuda', use_multi_gpu=True, mixed_precision=True)
Checkpoint path: TSB_AD_Time_RCD/checkpoints/dataset_15_56.pth
Traceback (most recent call last):
  File "/timercd_full_eval_work/code/Time-RCD-main/main.py", line 153, in <module>
    output, logits = run_Unsupervise_AD(args.AD_Name, data_train, test_data, Multi=Multi, **Optimal_Det_HP)
                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/timercd_full_eval_work/code/Time-RCD-main/model_wrapper.py", line 23, in run_Unsupervise_AD
    results = function_to_call(testing_data, **kwargs)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/timercd_full_eval_work/code/Time-RCD-main/model_wrapper.py", line 465, in run_Time_RCD
    cls = TimeRCDPretrainTester(checkpoint_path, config)
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/timercd_full_eval_work/code/Time-RCD-main/models/TimeRCD.py", line 32, in __init__
    self.load_checkpoint(checkpoint_path)
  File "/timercd_full_eval_work/code/Time-RCD-main/models/TimeRCD.py", line 59, in load_checkpoint
    self.model.load_state_dict(new_state_dict)
  File "/root/.cache/uv/archive-v0/iJjGyJfr82rBS4IM-6eyX/lib/python3.12/site-packages/torch/nn/modules/module.py", line 2638, in load_state_dict
    raise RuntimeError(
RuntimeError: Error(s) in loading state_dict for TimeSeriesPretrainModel:
	size mismatch for ts_encoder.embedding_layer.weight: copying a param with shape torch.Size([512, 16]) from checkpoint, the shape in current model is torch.Size([512, 32]).
	size mismatch for ts_encoder.projection_layer.weight: copying a param with shape torch.Size([4096, 512]) from checkpoint, the shape in current model is torch.Size([8192, 512]).
	size mismatch for ts_encoder.projection_layer.bias: copying a param with shape torch.Size([4096]) from checkpoint, the shape in current model is torch.Size([8192]).