Hugging Face's logo

RaikkonenWu
/
3dcvt-lrw1000

Video Classification
PyTorch
lip-reading
computer-vision
reproduction
3dcvt
Model card Files
xet
 Community
RaikkonenWu commited on
Commit
464090a
·
verified ·
1 Parent(s): b4f1045

Add files using upload-large-folder tool

Browse files
Files changed (6) hide show
  1. README.md +107 -0
  2. best_model.pth +3 -0
  3. logs/train.log +531 -0
  4. plots/learning_curve.png +0 -0
  5. results/per_class_acc_lrw1000_val.csv +1185 -0
  6. sha256.txt +1 -0
README.md ADDED
@@ -0,0 +1,107 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ library_name: pytorch
4
+ tags:
5
+ - pytorch
6
+ - lip-reading
7
+ - computer-vision
8
+ - video-classification
9
+ - reproduction
10
+ - 3dcvt
11
+ ---
12
+
13
+ # 3DCvT on LRW-1000
14
+
15
+ This repository provides the released checkpoint and evaluation artifacts for an unofficial PyTorch reproduction of:
16
+
17
+ **A Lip Reading Method Based on 3D Convolutional Vision Transformer**
18
+
19
+ Code repository:
20
+
21
+ - https://github.com/DPInnovationWorks/3DCvT_LipReading
22
+
23
+ ## Model Summary
24
+
25
+ - Task: Chinese word-level lip reading
26
+ - Dataset: LRW-1000
27
+ - Number of classes: 1184 in this processed split
28
+ - Framework: PyTorch
29
+ - Architecture: 3D CNN + CvT + BiGRU
30
+
31
+ ## Released Files
32
+
33
+ - `best_model.pth`: released checkpoint
34
+ - `sha256.txt`: checksum for the checkpoint
35
+ - `logs/train.log`: selected training log
36
+ - `results/per_class_acc_lrw1000_val.csv`: per-class validation summary
37
+ - `plots/learning_curve.png`: learning curve exported from training
38
+
39
+ ## Training Setup
40
+
41
+ Training settings from the released run:
42
+
43
+ - GPUs: 1 GPU
44
+ - Per-step batch size: 128
45
+ - Gradient accumulation: 2
46
+ - Effective batch size: 256
47
+ - Epochs: 120
48
+ - Optimizer: Adam
49
+ - Weight decay: 1e-4
50
+ - Learning rate: 6e-4
51
+ - Warmup epochs: 5
52
+ - Mixed precision: AMP enabled
53
+ - `torch.compile`: disabled
54
+
55
+ ## Evaluation Result
56
+
57
+ | Dataset | Split | Metric | Value |
58
+ | --- | --- | --- | --- |
59
+ | LRW-1000 | Validation | Top-1 Accuracy | 55.29% |
60
+
61
+ ## Intended Use
62
+
63
+ This checkpoint is intended for:
64
+
65
+ - research reproduction
66
+ - benchmark comparison
67
+ - qualitative inference demos
68
+
69
+ It is not intended as a production-ready commercial lip-reading system.
70
+
71
+ ## Limitations
72
+
73
+ - Performance depends on using the matching preprocessing pipeline
74
+ - This release does not include the raw LRW-1000 dataset
75
+ - Users must obtain the dataset according to its own terms
76
+ - This processed split uses 1184 classes in the generated vocabulary
77
+
78
+ ## Usage
79
+
80
+ Example inference command:
81
+
82
+ ```bash
83
+ python inference.py \
84
+ --dataset lrw1000 \
85
+ --pkl_path /path/to/sample.pkl \
86
+ --checkpoint /path/to/best_model.pth \
87
+ --gpu 0
88
+ ```
89
+
90
+ ## Notes
91
+
92
+ - The checkpoint is released for reproducibility
93
+ - Please use the matching code version when possible
94
+ - Local source artifact names were `best_model_for_lrw1000.pth` and `train_lrw1000.log`
95
+
96
+ ## Citation
97
+
98
+ If you use this release, please cite the original paper:
99
+
100
+ ```bibtex
101
+ @article{wu2022lip,
102
+ title={A Lip Reading Method Based on 3D Convolutional Vision Transformer},
103
+ author={Wu, Jiafeng and others},
104
+ journal={IEEE Access},
105
+ year={2022}
106
+ }
107
+ ```
best_model.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2ef301608cbdc841d2f1ba0255bef65f65c3b94cfcb9d4991f50167d80a15ef3
3
+ size 468858433
logs/train.log ADDED
@@ -0,0 +1,531 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2026-03-07 13:27:30,519 - DDP Initialized. World Size: 1
2
+ 2026-03-07 13:27:30,519 - Config: {
3
+ "dataset": "lrw1000",
4
+ "data_root": "/ssd2/3DCvT_data/data_LRW1000",
5
+ "exp_name": "3DCvT_LRW1000_new_version",
6
+ "batch_size": 32,
7
+ "epochs": 120,
8
+ "lr": 0.0006,
9
+ "num_workers": 8,
10
+ "num_classes": 1184,
11
+ "resume": null,
12
+ "warmup_epochs": 5,
13
+ "accum_steps": 4
14
+ }
15
+ 2026-03-07 13:27:30,520 - Effective batch size: 32 x 1 GPUs x 4 accum = 128
16
+ 2026-03-07 13:27:30,520 - Initializing Datasets (lrw1000)...
17
+ 2026-03-07 13:27:30,522 - Initialized LRW1000Dataset [train]. Found 1184 classes.
18
+ 2026-03-07 13:27:32,860 - Loaded 603193 samples for split 'train'.
19
+ 2026-03-07 13:27:32,860 - Initialized LRW1000Dataset [val]. Found 1184 classes.
20
+ 2026-03-07 13:27:33,039 - Loaded 63237 samples for split 'val'.
21
+ 2026-03-07 13:27:34,735 - Reverted SyncBN → BatchNorm in Stage 3 blocks (checkpoint compatibility).
22
+ 2026-03-07 13:27:34,808 - Start DDP Training...
23
+ 2026-03-07 13:29:10,789 - DDP Initialized. World Size: 1
24
+ 2026-03-07 13:29:10,790 - Config: {
25
+ "dataset": "lrw1000",
26
+ "data_root": "/ssd2/3DCvT_data/data_LRW1000",
27
+ "exp_name": "3DCvT_LRW1000_new_version",
28
+ "batch_size": 64,
29
+ "epochs": 120,
30
+ "lr": 0.0006,
31
+ "num_workers": 8,
32
+ "num_classes": 1184,
33
+ "resume": null,
34
+ "warmup_epochs": 5,
35
+ "accum_steps": 4
36
+ }
37
+ 2026-03-07 13:29:10,790 - Effective batch size: 64 x 1 GPUs x 4 accum = 256
38
+ 2026-03-07 13:29:10,790 - Initializing Datasets (lrw1000)...
39
+ 2026-03-07 13:29:10,791 - Initialized LRW1000Dataset [train]. Found 1184 classes.
40
+ 2026-03-07 13:29:13,078 - Loaded 603193 samples for split 'train'.
41
+ 2026-03-07 13:29:13,079 - Initialized LRW1000Dataset [val]. Found 1184 classes.
42
+ 2026-03-07 13:29:13,245 - Loaded 63237 samples for split 'val'.
43
+ 2026-03-07 13:29:14,873 - Reverted SyncBN → BatchNorm in Stage 3 blocks (checkpoint compatibility).
44
+ 2026-03-07 13:29:14,902 - Start DDP Training...
45
+ 2026-03-07 13:30:03,949 - Experiment Started: 3DCvT_LRW1000_new_version
46
+ 2026-03-07 13:30:03,949 - Config: {
47
+ "dataset": "lrw1000",
48
+ "data_root": "/ssd2/3DCvT_data/data_LRW1000",
49
+ "exp_name": "3DCvT_LRW1000_new_version",
50
+ "batch_size": 64,
51
+ "epochs": 120,
52
+ "lr": 0.0006,
53
+ "num_workers": 8,
54
+ "num_classes": 1184,
55
+ "gpu": "0",
56
+ "resume": "",
57
+ "warmup_epochs": 5,
58
+ "accum_steps": 4
59
+ }
60
+ 2026-03-07 13:30:03,949 - Effective batch size: 64 x 4 accum = 256
61
+ 2026-03-07 13:30:03,949 - Initializing Datasets (lrw1000)...
62
+ 2026-03-07 13:30:03,949 - Initialized LRW1000Dataset [train]. Found 1184 classes.
63
+ 2026-03-07 13:30:06,188 - Loaded 603193 samples for split 'train'.
64
+ 2026-03-07 13:30:06,188 - Initialized LRW1000Dataset [val]. Found 1184 classes.
65
+ 2026-03-07 13:30:06,346 - Loaded 63237 samples for split 'val'.
66
+ 2026-03-07 13:30:06,347 - Building Model...
67
+ 2026-03-07 13:30:10,713 - Start Training...
68
+ 2026-03-07 13:31:11,274 - Experiment Started: 3DCvT_LRW1000_new_version
69
+ 2026-03-07 13:31:11,274 - Config: {
70
+ "dataset": "lrw1000",
71
+ "data_root": "/ssd2/3DCvT_data/data_LRW1000",
72
+ "exp_name": "3DCvT_LRW1000_new_version",
73
+ "batch_size": 64,
74
+ "epochs": 120,
75
+ "lr": 0.0006,
76
+ "num_workers": 8,
77
+ "num_classes": 1184,
78
+ "gpu": "0",
79
+ "resume": "",
80
+ "warmup_epochs": 5,
81
+ "accum_steps": 4
82
+ }
83
+ 2026-03-07 13:31:11,274 - Effective batch size: 64 x 4 accum = 256
84
+ 2026-03-07 13:31:11,274 - Initializing Datasets (lrw1000)...
85
+ 2026-03-07 13:31:11,274 - Initialized LRW1000Dataset [train]. Found 1184 classes.
86
+ 2026-03-07 13:31:20,545 - Experiment Started: 3DCvT_LRW1000_new_version
87
+ 2026-03-07 13:31:20,545 - Config: {
88
+ "dataset": "lrw1000",
89
+ "data_root": "/ssd2/3DCvT_data/data_LRW1000",
90
+ "exp_name": "3DCvT_LRW1000_new_version",
91
+ "batch_size": 64,
92
+ "epochs": 120,
93
+ "lr": 0.0006,
94
+ "num_workers": 8,
95
+ "num_classes": 1184,
96
+ "gpu": "0",
97
+ "resume": "",
98
+ "warmup_epochs": 5,
99
+ "accum_steps": 4
100
+ }
101
+ 2026-03-07 13:31:20,545 - Effective batch size: 64 x 4 accum = 256
102
+ 2026-03-07 13:31:20,545 - Initializing Datasets (lrw1000)...
103
+ 2026-03-07 13:31:20,545 - Initialized LRW1000Dataset [train]. Found 1184 classes.
104
+ 2026-03-07 13:31:22,844 - Loaded 603193 samples for split 'train'.
105
+ 2026-03-07 13:31:22,844 - Initialized LRW1000Dataset [val]. Found 1184 classes.
106
+ 2026-03-07 13:31:23,008 - Loaded 63237 samples for split 'val'.
107
+ 2026-03-07 13:31:23,009 - Building Model...
108
+ 2026-03-07 13:31:25,648 - Start Training...
109
+ 2026-03-07 13:33:35,145 - Experiment Started: 3DCvT_LRW1000_new_version
110
+ 2026-03-07 13:33:35,145 - Config: {
111
+ "dataset": "lrw1000",
112
+ "data_root": "/ssd2/3DCvT_data/data_LRW1000",
113
+ "exp_name": "3DCvT_LRW1000_new_version",
114
+ "batch_size": 64,
115
+ "epochs": 120,
116
+ "lr": 0.0006,
117
+ "num_workers": 8,
118
+ "num_classes": 1184,
119
+ "gpu": "1",
120
+ "resume": "",
121
+ "warmup_epochs": 5,
122
+ "accum_steps": 4
123
+ }
124
+ 2026-03-07 13:33:35,145 - Effective batch size: 64 x 4 accum = 256
125
+ 2026-03-07 13:33:35,145 - Initializing Datasets (lrw1000)...
126
+ 2026-03-07 13:33:35,145 - Initialized LRW1000Dataset [train]. Found 1184 classes.
127
+ 2026-03-07 13:33:37,497 - Loaded 603193 samples for split 'train'.
128
+ 2026-03-07 13:33:37,498 - Initialized LRW1000Dataset [val]. Found 1184 classes.
129
+ 2026-03-07 13:33:37,662 - Loaded 63237 samples for split 'val'.
130
+ 2026-03-07 13:33:37,663 - Building Model...
131
+ 2026-03-07 13:33:40,535 - Start Training...
132
+ 2026-03-07 13:36:54,798 - Experiment Started: 3DCvT_LRW1000_new_version
133
+ 2026-03-07 13:36:54,798 - Config: {
134
+ "dataset": "lrw1000",
135
+ "data_root": "/ssd2/3DCvT_data/data_LRW1000",
136
+ "exp_name": "3DCvT_LRW1000_new_version",
137
+ "batch_size": 128,
138
+ "epochs": 120,
139
+ "lr": 0.0006,
140
+ "num_workers": 8,
141
+ "num_classes": 1184,
142
+ "gpu": "1",
143
+ "resume": "",
144
+ "warmup_epochs": 5,
145
+ "accum_steps": 2
146
+ }
147
+ 2026-03-07 13:36:54,798 - Effective batch size: 128 x 2 accum = 256
148
+ 2026-03-07 13:36:54,798 - Initializing Datasets (lrw1000)...
149
+ 2026-03-07 13:36:54,799 - Initialized LRW1000Dataset [train]. Found 1184 classes.
150
+ 2026-03-07 13:36:57,090 - Loaded 603193 samples for split 'train'.
151
+ 2026-03-07 13:36:57,090 - Initialized LRW1000Dataset [val]. Found 1184 classes.
152
+ 2026-03-07 13:36:57,250 - Loaded 63237 samples for split 'val'.
153
+ 2026-03-07 13:36:57,251 - Building Model...
154
+ 2026-03-07 13:37:00,242 - Start Training...
155
+ 2026-03-07 13:47:33,750 - Experiment Started: 3DCvT_LRW1000_new_version
156
+ 2026-03-07 13:47:33,751 - Config: {
157
+ "dataset": "lrw1000",
158
+ "data_root": "/ssd2/3DCvT_data/data_LRW1000",
159
+ "exp_name": "3DCvT_LRW1000_new_version",
160
+ "batch_size": 128,
161
+ "epochs": 120,
162
+ "lr": 0.0006,
163
+ "num_workers": 8,
164
+ "num_classes": 1184,
165
+ "gpu": "1",
166
+ "resume": "",
167
+ "warmup_epochs": 5,
168
+ "accum_steps": 2
169
+ }
170
+ 2026-03-07 13:47:33,751 - Effective batch size: 128 x 2 accum = 256
171
+ 2026-03-07 13:47:33,751 - Initializing Datasets (lrw1000)...
172
+ 2026-03-07 13:47:33,751 - Initialized LRW1000Dataset [train]. Found 1184 classes.
173
+ 2026-03-07 13:47:35,967 - Loaded 603193 samples for split 'train'.
174
+ 2026-03-07 13:47:35,968 - Initialized LRW1000Dataset [val]. Found 1184 classes.
175
+ 2026-03-07 13:47:36,125 - Loaded 63237 samples for split 'val'.
176
+ 2026-03-07 13:47:36,126 - Building Model...
177
+ 2026-03-07 13:47:38,905 - Start Training...
178
+ 2026-03-07 14:39:10,115 - Epoch [1/120] Completed in 3091s | ETA: 4 days, 6:10:53
179
+ 2026-03-07 14:39:10,116 - Train Loss: 6.1146 | Val Loss: 6.1266 | Val Acc: 8.96%
180
+ 2026-03-07 14:39:14,138 - New Best Accuracy: 8.96% - Saving Model...
181
+ 2026-03-07 15:26:58,724 - Epoch [2/120] Completed in 2863s | ETA: 3 days, 21:52:18
182
+ 2026-03-07 15:26:58,724 - Train Loss: 5.5959 | Val Loss: 4.7110 | Val Acc: 20.13%
183
+ 2026-03-07 15:27:04,810 - New Best Accuracy: 20.13% - Saving Model...
184
+ 2026-03-07 16:14:43,682 - Epoch [3/120] Completed in 2854s | ETA: 3 days, 20:46:22
185
+ 2026-03-07 16:14:43,692 - Train Loss: 5.0246 | Val Loss: 4.1261 | Val Acc: 29.50%
186
+ 2026-03-07 16:14:47,546 - New Best Accuracy: 29.50% - Saving Model...
187
+ 2026-03-07 17:02:10,337 - Epoch [4/120] Completed in 2839s | ETA: 3 days, 19:30:06
188
+ 2026-03-07 17:02:10,445 - Train Loss: 4.7857 | Val Loss: 4.0181 | Val Acc: 31.03%
189
+ 2026-03-07 17:02:14,727 - New Best Accuracy: 31.03% - Saving Model...
190
+ 2026-03-07 17:49:00,047 - Epoch [5/120] Completed in 2802s | ETA: 3 days, 17:32:23
191
+ 2026-03-07 17:49:00,078 - Train Loss: 4.6874 | Val Loss: 3.9716 | Val Acc: 32.33%
192
+ 2026-03-07 17:49:03,689 - New Best Accuracy: 32.33% - Saving Model...
193
+ 2026-03-07 18:35:30,091 - Epoch [6/120] Completed in 2784s | ETA: 3 days, 16:10:05
194
+ 2026-03-07 18:35:30,102 - Train Loss: 4.6389 | Val Loss: 4.0567 | Val Acc: 30.94%
195
+ 2026-03-07 19:21:43,990 - Epoch [7/120] Completed in 2771s | ETA: 3 days, 14:58:51
196
+ 2026-03-07 19:21:44,042 - Train Loss: 4.5574 | Val Loss: 3.8917 | Val Acc: 33.41%
197
+ 2026-03-07 19:21:47,962 - New Best Accuracy: 33.41% - Saving Model...
198
+ 2026-03-07 20:07:59,380 - Epoch [8/120] Completed in 2768s | ETA: 3 days, 14:08:14
199
+ 2026-03-07 20:07:59,436 - Train Loss: 4.4765 | Val Loss: 3.8285 | Val Acc: 35.18%
200
+ 2026-03-07 20:08:02,930 - New Best Accuracy: 35.18% - Saving Model...
201
+ 2026-03-07 20:54:09,507 - Epoch [9/120] Completed in 2763s | ETA: 3 days, 13:13:10
202
+ 2026-03-07 20:54:09,591 - Train Loss: 4.4423 | Val Loss: 3.7680 | Val Acc: 36.24%
203
+ 2026-03-07 20:54:13,711 - New Best Accuracy: 36.24% - Saving Model...
204
+ 2026-03-07 21:40:16,305 - Epoch [10/120] Completed in 2760s | ETA: 3 days, 12:20:29
205
+ 2026-03-07 21:40:16,315 - Train Loss: 4.3743 | Val Loss: 3.7331 | Val Acc: 36.27%
206
+ 2026-03-07 21:40:19,393 - New Best Accuracy: 36.27% - Saving Model...
207
+ 2026-03-07 22:26:22,878 - Epoch [11/120] Completed in 2760s | ETA: 3 days, 11:34:01
208
+ 2026-03-07 22:26:22,890 - Train Loss: 4.3717 | Val Loss: 3.7375 | Val Acc: 37.03%
209
+ 2026-03-07 22:26:26,071 - New Best Accuracy: 37.03% - Saving Model...
210
+ 2026-03-07 23:12:19,349 - Epoch [12/120] Completed in 2750s | ETA: 3 days, 10:31:28
211
+ 2026-03-07 23:12:19,441 - Train Loss: 4.3413 | Val Loss: 3.6906 | Val Acc: 37.47%
212
+ 2026-03-07 23:12:25,000 - New Best Accuracy: 37.47% - Saving Model...
213
+ 2026-03-07 23:58:27,461 - Epoch [13/120] Completed in 2759s | ETA: 3 days, 10:01:34
214
+ 2026-03-07 23:58:27,515 - Train Loss: 4.3101 | Val Loss: 3.7782 | Val Acc: 36.47%
215
+ 2026-03-08 00:44:28,522 - Epoch [14/120] Completed in 2757s | ETA: 3 days, 9:10:58
216
+ 2026-03-08 00:44:28,725 - Train Loss: 4.2773 | Val Loss: 3.6685 | Val Acc: 38.19%
217
+ 2026-03-08 00:44:33,851 - New Best Accuracy: 38.19% - Saving Model...
218
+ 2026-03-08 01:30:36,273 - Epoch [15/120] Completed in 2759s | ETA: 3 days, 8:29:25
219
+ 2026-03-08 01:30:36,303 - Train Loss: 4.2530 | Val Loss: 3.6649 | Val Acc: 38.32%
220
+ 2026-03-08 01:30:39,781 - New Best Accuracy: 38.32% - Saving Model...
221
+ 2026-03-08 02:16:39,537 - Epoch [16/120] Completed in 2756s | ETA: 3 days, 7:38:33
222
+ 2026-03-08 02:16:39,724 - Train Loss: 4.2342 | Val Loss: 3.6691 | Val Acc: 38.20%
223
+ 2026-03-08 03:02:44,783 - Epoch [17/120] Completed in 2760s | ETA: 3 days, 6:58:13
224
+ 2026-03-08 03:02:44,820 - Train Loss: 4.2250 | Val Loss: 3.6091 | Val Acc: 38.71%
225
+ 2026-03-08 03:02:48,281 - New Best Accuracy: 38.71% - Saving Model...
226
+ 2026-03-08 03:48:49,500 - Epoch [18/120] Completed in 2758s | ETA: 3 days, 6:09:05
227
+ 2026-03-08 03:48:49,604 - Train Loss: 4.2004 | Val Loss: 3.6267 | Val Acc: 38.32%
228
+ 2026-03-08 04:34:53,259 - Epoch [19/120] Completed in 2758s | ETA: 3 days, 5:24:01
229
+ 2026-03-08 04:34:53,328 - Train Loss: 4.2055 | Val Loss: 3.6411 | Val Acc: 38.70%
230
+ 2026-03-08 05:20:55,701 - Epoch [20/120] Completed in 2758s | ETA: 3 days, 4:37:34
231
+ 2026-03-08 05:20:55,803 - Train Loss: 4.1628 | Val Loss: 3.6491 | Val Acc: 38.05%
232
+ 2026-03-08 06:06:58,019 - Epoch [21/120] Completed in 2756s | ETA: 3 days, 3:48:48
233
+ 2026-03-08 06:06:58,103 - Train Loss: 4.1641 | Val Loss: 3.6765 | Val Acc: 37.76%
234
+ 2026-03-08 06:53:03,234 - Epoch [22/120] Completed in 2759s | ETA: 3 days, 3:07:52
235
+ 2026-03-08 06:53:03,255 - Train Loss: 4.1442 | Val Loss: 3.5592 | Val Acc: 39.63%
236
+ 2026-03-08 06:53:06,593 - New Best Accuracy: 39.63% - Saving Model...
237
+ 2026-03-08 07:39:07,865 - Epoch [23/120] Completed in 2758s | ETA: 3 days, 2:19:29
238
+ 2026-03-08 07:39:08,096 - Train Loss: 4.1409 | Val Loss: 3.6482 | Val Acc: 38.85%
239
+ 2026-03-08 08:25:13,002 - Epoch [24/120] Completed in 2758s | ETA: 3 days, 1:33:37
240
+ 2026-03-08 08:25:13,019 - Train Loss: 4.1133 | Val Loss: 3.6647 | Val Acc: 38.52%
241
+ 2026-03-08 09:11:12,667 - Epoch [25/120] Completed in 2756s | ETA: 3 days, 0:43:47
242
+ 2026-03-08 09:11:12,878 - Train Loss: 4.0990 | Val Loss: 3.6565 | Val Acc: 38.09%
243
+ 2026-03-08 09:57:19,987 - Epoch [26/120] Completed in 2760s | ETA: 3 days, 0:05:19
244
+ 2026-03-08 09:57:19,988 - Train Loss: 4.1216 | Val Loss: 3.6347 | Val Acc: 37.98%
245
+ 2026-03-08 22:32:37,854 - Experiment Started: 3DCvT_LRW1000_new_version
246
+ 2026-03-08 22:32:37,869 - Config: {
247
+ "dataset": "lrw1000",
248
+ "data_root": "/ssd2/3DCvT_data/data_LRW1000",
249
+ "exp_name": "3DCvT_LRW1000_new_version",
250
+ "batch_size": 128,
251
+ "epochs": 120,
252
+ "lr": 0.0006,
253
+ "num_workers": 8,
254
+ "num_classes": 1184,
255
+ "gpu": "1",
256
+ "resume": "",
257
+ "warmup_epochs": 5,
258
+ "accum_steps": 2,
259
+ "use_compile": false
260
+ }
261
+ 2026-03-08 22:32:37,869 - Effective batch size: 128 x 2 accum = 256
262
+ 2026-03-08 22:32:37,869 - torch.compile: disabled (recommended for stability on RTX 20xx / checkpointing).
263
+ 2026-03-08 22:32:37,869 - Initializing Datasets (lrw1000)...
264
+ 2026-03-08 22:32:37,872 - Initialized LRW1000Dataset [train]. Found 1184 classes.
265
+ 2026-03-08 22:32:41,605 - Loaded 603193 samples for split 'train'.
266
+ 2026-03-08 22:32:41,606 - Initialized LRW1000Dataset [val]. Found 1184 classes.
267
+ 2026-03-08 22:32:41,981 - Loaded 63237 samples for split 'val'.
268
+ 2026-03-08 22:32:41,982 - Building Model...
269
+ 2026-03-08 22:32:45,660 - Start Training...
270
+ 2026-03-08 23:32:25,334 - Epoch [1/120] Completed in 3579s | ETA: 4 days, 22:19:41
271
+ 2026-03-08 23:32:25,395 - Train Loss: 6.1139 | Val Loss: 6.1373 | Val Acc: 8.96%
272
+ 2026-03-08 23:32:31,004 - New Best Accuracy: 8.96% - Saving Model...
273
+ 2026-03-09 00:31:25,158 - Epoch [2/120] Completed in 3531s | ETA: 4 days, 19:45:44
274
+ 2026-03-09 00:31:25,175 - Train Loss: 5.4274 | Val Loss: 4.0172 | Val Acc: 30.90%
275
+ 2026-03-09 00:31:27,994 - New Best Accuracy: 30.90% - Saving Model...
276
+ 2026-03-09 01:30:14,477 - Epoch [3/120] Completed in 3523s | ETA: 4 days, 18:31:22
277
+ 2026-03-09 01:30:14,555 - Train Loss: 4.5912 | Val Loss: 3.5136 | Val Acc: 39.33%
278
+ 2026-03-09 01:30:18,307 - New Best Accuracy: 39.33% - Saving Model...
279
+ 2026-03-09 02:28:43,353 - Epoch [4/120] Completed in 3502s | ETA: 4 days, 16:51:38
280
+ 2026-03-09 02:28:43,362 - Train Loss: 4.3607 | Val Loss: 3.4661 | Val Acc: 41.29%
281
+ 2026-03-09 02:28:46,182 - New Best Accuracy: 41.29% - Saving Model...
282
+ 2026-03-09 03:26:32,731 - Epoch [5/120] Completed in 3464s | ETA: 4 days, 14:39:44
283
+ 2026-03-09 03:26:32,805 - Train Loss: 4.2389 | Val Loss: 3.3259 | Val Acc: 43.45%
284
+ 2026-03-09 03:26:36,034 - New Best Accuracy: 43.45% - Saving Model...
285
+ 2026-03-09 04:24:01,406 - Epoch [6/120] Completed in 3443s | ETA: 4 days, 13:02:34
286
+ 2026-03-09 04:24:01,422 - Train Loss: 4.2117 | Val Loss: 3.3535 | Val Acc: 42.81%
287
+ 2026-03-09 05:21:20,464 - Epoch [7/120] Completed in 3436s | ETA: 4 days, 11:51:49
288
+ 2026-03-09 05:21:20,534 - Train Loss: 4.1278 | Val Loss: 3.2863 | Val Acc: 44.59%
289
+ 2026-03-09 05:21:23,967 - New Best Accuracy: 44.59% - Saving Model...
290
+ 2026-03-09 06:18:37,252 - Epoch [8/120] Completed in 3431s | ETA: 4 days, 10:44:55
291
+ 2026-03-09 06:18:37,262 - Train Loss: 4.0589 | Val Loss: 3.2972 | Val Acc: 44.68%
292
+ 2026-03-09 06:18:40,013 - New Best Accuracy: 44.68% - Saving Model...
293
+ 2026-03-09 07:15:49,174 - Epoch [9/120] Completed in 3427s | ETA: 4 days, 9:40:27
294
+ 2026-03-09 07:15:49,186 - Train Loss: 4.0015 | Val Loss: 3.2348 | Val Acc: 46.27%
295
+ 2026-03-09 07:15:51,838 - New Best Accuracy: 46.27% - Saving Model...
296
+ 2026-03-09 08:12:59,866 - Epoch [10/120] Completed in 3426s | ETA: 4 days, 8:41:11
297
+ 2026-03-09 08:12:59,938 - Train Loss: 3.9943 | Val Loss: 3.2154 | Val Acc: 46.89%
298
+ 2026-03-09 08:13:03,177 - New Best Accuracy: 46.89% - Saving Model...
299
+ 2026-03-09 09:10:11,694 - Epoch [11/120] Completed in 3424s | ETA: 4 days, 7:41:29
300
+ 2026-03-09 09:10:11,706 - Train Loss: 3.9392 | Val Loss: 3.2110 | Val Acc: 46.43%
301
+ 2026-03-09 10:07:18,028 - Epoch [12/120] Completed in 3423s | ETA: 4 days, 6:42:48
302
+ 2026-03-09 10:07:18,088 - Train Loss: 3.8925 | Val Loss: 3.1698 | Val Acc: 47.07%
303
+ 2026-03-09 10:07:21,390 - New Best Accuracy: 47.07% - Saving Model...
304
+ 2026-03-09 11:04:32,703 - Epoch [13/120] Completed in 3429s | ETA: 4 days, 5:55:54
305
+ 2026-03-09 11:04:32,718 - Train Loss: 3.8839 | Val Loss: 3.0736 | Val Acc: 48.91%
306
+ 2026-03-09 11:04:35,708 - New Best Accuracy: 48.91% - Saving Model...
307
+ 2026-03-09 12:01:38,239 - Epoch [14/120] Completed in 3420s | ETA: 4 days, 4:42:11
308
+ 2026-03-09 12:01:38,263 - Train Loss: 3.8637 | Val Loss: 3.1596 | Val Acc: 47.45%
309
+ 2026-03-09 12:58:42,245 - Epoch [15/120] Completed in 3420s | ETA: 4 days, 3:45:19
310
+ 2026-03-09 12:58:42,280 - Train Loss: 3.8370 | Val Loss: 3.1210 | Val Acc: 48.25%
311
+ 2026-03-09 13:55:48,145 - Epoch [16/120] Completed in 3422s | ETA: 4 days, 2:52:46
312
+ 2026-03-09 13:55:48,170 - Train Loss: 3.8118 | Val Loss: 3.1930 | Val Acc: 47.71%
313
+ 2026-03-09 14:52:52,972 - Epoch [17/120] Completed in 3421s | ETA: 4 days, 1:53:49
314
+ 2026-03-09 14:52:53,002 - Train Loss: 3.8038 | Val Loss: 3.0847 | Val Acc: 48.88%
315
+ 2026-03-09 15:49:57,051 - Epoch [18/120] Completed in 3421s | ETA: 4 days, 0:56:13
316
+ 2026-03-09 15:49:57,120 - Train Loss: 3.7720 | Val Loss: 3.2661 | Val Acc: 46.02%
317
+ 2026-03-09 16:47:02,991 - Epoch [19/120] Completed in 3422s | ETA: 4 days, 0:00:39
318
+ 2026-03-09 16:47:02,992 - Train Loss: 3.7863 | Val Loss: 3.2192 | Val Acc: 47.61%
319
+ 2026-03-09 17:44:08,478 - Epoch [20/120] Completed in 3422s | ETA: 3 days, 23:04:05
320
+ 2026-03-09 17:44:08,518 - Train Loss: 3.7505 | Val Loss: 3.1052 | Val Acc: 48.69%
321
+ 2026-03-09 18:41:14,096 - Epoch [21/120] Completed in 3420s | ETA: 3 days, 22:03:21
322
+ 2026-03-09 18:41:14,097 - Train Loss: 3.7487 | Val Loss: 3.0824 | Val Acc: 48.92%
323
+ 2026-03-09 18:41:16,873 - New Best Accuracy: 48.92% - Saving Model...
324
+ 2026-03-09 19:38:19,749 - Epoch [22/120] Completed in 3420s | ETA: 3 days, 21:06:56
325
+ 2026-03-09 19:38:19,769 - Train Loss: 3.7165 | Val Loss: 3.0937 | Val Acc: 49.24%
326
+ 2026-03-09 19:38:23,069 - New Best Accuracy: 49.24% - Saving Model...
327
+ 2026-03-09 20:35:25,989 - Epoch [23/120] Completed in 3421s | ETA: 3 days, 20:10:37
328
+ 2026-03-09 20:35:26,004 - Train Loss: 3.6894 | Val Loss: 3.0880 | Val Acc: 49.57%
329
+ 2026-03-09 20:35:28,812 - New Best Accuracy: 49.57% - Saving Model...
330
+ 2026-03-09 21:32:31,072 - Epoch [24/120] Completed in 3420s | ETA: 3 days, 19:12:26
331
+ 2026-03-09 21:32:31,072 - Train Loss: 3.7187 | Val Loss: 3.1269 | Val Acc: 49.02%
332
+ 2026-03-09 22:29:31,780 - Epoch [25/120] Completed in 3417s | ETA: 3 days, 18:11:27
333
+ 2026-03-09 22:29:31,781 - Train Loss: 3.7046 | Val Loss: 3.0835 | Val Acc: 49.52%
334
+ 2026-03-09 23:26:32,000 - Epoch [26/120] Completed in 3417s | ETA: 3 days, 17:13:33
335
+ 2026-03-09 23:26:32,000 - Train Loss: 3.6909 | Val Loss: 3.1447 | Val Acc: 48.62%
336
+ 2026-03-10 00:23:32,038 - Epoch [27/120] Completed in 3417s | ETA: 3 days, 16:16:58
337
+ 2026-03-10 00:23:32,057 - Train Loss: 3.6706 | Val Loss: 3.1064 | Val Acc: 48.84%
338
+ 2026-03-10 01:20:32,598 - Epoch [28/120] Completed in 3417s | ETA: 3 days, 15:20:27
339
+ 2026-03-10 01:20:32,666 - Train Loss: 3.6742 | Val Loss: 3.1448 | Val Acc: 47.85%
340
+ 2026-03-10 02:17:32,773 - Epoch [29/120] Completed in 3416s | ETA: 3 days, 14:21:46
341
+ 2026-03-10 02:17:32,805 - Train Loss: 3.6609 | Val Loss: 3.0539 | Val Acc: 49.69%
342
+ 2026-03-10 02:17:36,397 - New Best Accuracy: 49.69% - Saving Model...
343
+ 2026-03-10 03:14:35,293 - Epoch [30/120] Completed in 3416s | ETA: 3 days, 13:25:25
344
+ 2026-03-10 03:14:35,324 - Train Loss: 3.6600 | Val Loss: 3.1844 | Val Acc: 47.82%
345
+ 2026-03-10 04:11:33,759 - Epoch [31/120] Completed in 3414s | ETA: 3 days, 12:24:56
346
+ 2026-03-10 04:11:33,787 - Train Loss: 3.6417 | Val Loss: 3.1153 | Val Acc: 48.84%
347
+ 2026-03-10 05:08:33,230 - Epoch [32/120] Completed in 3416s | ETA: 3 days, 11:31:12
348
+ 2026-03-10 05:08:33,270 - Train Loss: 3.6493 | Val Loss: 3.0403 | Val Acc: 51.52%
349
+ 2026-03-10 05:08:36,814 - New Best Accuracy: 51.52% - Saving Model...
350
+ 2026-03-10 06:05:36,057 - Epoch [33/120] Completed in 3417s | ETA: 3 days, 10:34:53
351
+ 2026-03-10 06:05:36,082 - Train Loss: 3.6447 | Val Loss: 3.0588 | Val Acc: 50.09%
352
+ 2026-03-10 07:02:34,945 - Epoch [34/120] Completed in 3416s | ETA: 3 days, 9:36:42
353
+ 2026-03-10 07:02:34,971 - Train Loss: 3.6049 | Val Loss: 3.0544 | Val Acc: 50.60%
354
+ 2026-03-10 07:59:32,809 - Epoch [35/120] Completed in 3415s | ETA: 3 days, 8:38:16
355
+ 2026-03-10 07:59:32,828 - Train Loss: 3.6051 | Val Loss: 3.0683 | Val Acc: 50.15%
356
+ 2026-03-10 08:56:33,215 - Epoch [36/120] Completed in 3417s | ETA: 3 days, 7:44:21
357
+ 2026-03-10 08:56:33,273 - Train Loss: 3.6166 | Val Loss: 3.1145 | Val Acc: 49.42%
358
+ 2026-03-10 09:53:32,804 - Epoch [37/120] Completed in 3416s | ETA: 3 days, 6:45:42
359
+ 2026-03-10 09:53:32,836 - Train Loss: 3.5847 | Val Loss: 3.0647 | Val Acc: 50.04%
360
+ 2026-03-10 10:50:33,474 - Epoch [38/120] Completed in 3417s | ETA: 3 days, 5:51:05
361
+ 2026-03-10 10:50:33,485 - Train Loss: 3.5824 | Val Loss: 3.0365 | Val Acc: 51.10%
362
+ 2026-03-10 11:47:33,181 - Epoch [39/120] Completed in 3416s | ETA: 3 days, 4:52:25
363
+ 2026-03-10 11:47:33,199 - Train Loss: 3.5661 | Val Loss: 3.0442 | Val Acc: 50.47%
364
+ 2026-03-10 12:44:33,661 - Epoch [40/120] Completed in 3417s | ETA: 3 days, 3:56:35
365
+ 2026-03-10 12:44:33,661 - Train Loss: 3.5500 | Val Loss: 3.0243 | Val Acc: 50.68%
366
+ 2026-03-10 13:41:34,307 - Epoch [41/120] Completed in 3417s | ETA: 3 days, 2:59:20
367
+ 2026-03-10 13:41:34,320 - Train Loss: 3.5551 | Val Loss: 3.0164 | Val Acc: 51.17%
368
+ 2026-03-10 14:38:34,255 - Epoch [42/120] Completed in 3417s | ETA: 3 days, 2:02:23
369
+ 2026-03-10 14:38:34,282 - Train Loss: 3.5651 | Val Loss: 3.0644 | Val Acc: 50.71%
370
+ 2026-03-10 15:35:33,191 - Epoch [43/120] Completed in 3415s | ETA: 3 days, 1:03:17
371
+ 2026-03-10 15:35:33,212 - Train Loss: 3.5391 | Val Loss: 3.1033 | Val Acc: 50.31%
372
+ 2026-03-10 16:32:33,743 - Epoch [44/120] Completed in 3417s | ETA: 3 days, 0:09:00
373
+ 2026-03-10 16:32:33,744 - Train Loss: 3.5286 | Val Loss: 3.0567 | Val Acc: 50.88%
374
+ 2026-03-10 17:29:32,973 - Epoch [45/120] Completed in 3416s | ETA: 2 days, 23:10:33
375
+ 2026-03-10 17:29:32,996 - Train Loss: 3.4986 | Val Loss: 2.9846 | Val Acc: 51.51%
376
+ 2026-03-10 18:26:32,980 - Epoch [46/120] Completed in 3417s | ETA: 2 days, 22:14:49
377
+ 2026-03-10 18:26:33,015 - Train Loss: 3.5029 | Val Loss: 2.9726 | Val Acc: 52.56%
378
+ 2026-03-10 18:26:36,453 - New Best Accuracy: 52.56% - Saving Model...
379
+ 2026-03-10 19:23:36,323 - Epoch [47/120] Completed in 3417s | ETA: 2 days, 21:18:28
380
+ 2026-03-10 19:23:36,323 - Train Loss: 3.5171 | Val Loss: 3.0521 | Val Acc: 50.82%
381
+ 2026-03-10 20:20:36,070 - Epoch [48/120] Completed in 3417s | ETA: 2 days, 20:20:41
382
+ 2026-03-10 20:20:36,084 - Train Loss: 3.4932 | Val Loss: 2.9845 | Val Acc: 51.35%
383
+ 2026-03-10 21:17:36,907 - Epoch [49/120] Completed in 3418s | ETA: 2 days, 19:24:50
384
+ 2026-03-10 21:17:36,926 - Train Loss: 3.4877 | Val Loss: 2.9769 | Val Acc: 51.83%
385
+ 2026-03-10 22:14:40,225 - Epoch [50/120] Completed in 3419s | ETA: 2 days, 18:29:47
386
+ 2026-03-10 22:14:40,248 - Train Loss: 3.5020 | Val Loss: 2.9528 | Val Acc: 51.98%
387
+ 2026-03-10 23:11:41,697 - Epoch [51/120] Completed in 3417s | ETA: 2 days, 17:30:34
388
+ 2026-03-10 23:11:41,735 - Train Loss: 3.4634 | Val Loss: 3.0204 | Val Acc: 51.70%
389
+ 2026-03-11 00:08:41,867 - Epoch [52/120] Completed in 3417s | ETA: 2 days, 16:33:14
390
+ 2026-03-11 00:08:41,868 - Train Loss: 3.4536 | Val Loss: 3.0020 | Val Acc: 52.00%
391
+ 2026-03-11 01:05:41,837 - Epoch [53/120] Completed in 3417s | ETA: 2 days, 15:35:55
392
+ 2026-03-11 01:05:41,902 - Train Loss: 3.4476 | Val Loss: 2.9410 | Val Acc: 52.33%
393
+ 2026-03-11 02:02:43,245 - Epoch [54/120] Completed in 3417s | ETA: 2 days, 14:39:16
394
+ 2026-03-11 02:02:43,285 - Train Loss: 3.4324 | Val Loss: 3.0272 | Val Acc: 51.13%
395
+ 2026-03-11 02:59:42,827 - Epoch [55/120] Completed in 3416s | ETA: 2 days, 13:41:29
396
+ 2026-03-11 02:59:42,863 - Train Loss: 3.3916 | Val Loss: 2.9085 | Val Acc: 53.14%
397
+ 2026-03-11 02:59:45,589 - New Best Accuracy: 53.14% - Saving Model...
398
+ 2026-03-11 03:56:45,360 - Epoch [56/120] Completed in 3417s | ETA: 2 days, 12:45:34
399
+ 2026-03-11 03:56:45,433 - Train Loss: 3.3916 | Val Loss: 3.0134 | Val Acc: 51.60%
400
+ 2026-03-11 04:53:46,207 - Epoch [57/120] Completed in 3417s | ETA: 2 days, 11:48:02
401
+ 2026-03-11 04:53:46,218 - Train Loss: 3.4111 | Val Loss: 2.9538 | Val Acc: 53.22%
402
+ 2026-03-11 04:53:48,946 - New Best Accuracy: 53.22% - Saving Model...
403
+ 2026-03-11 05:50:49,239 - Epoch [58/120] Completed in 3418s | ETA: 2 days, 10:52:18
404
+ 2026-03-11 05:50:49,273 - Train Loss: 3.3877 | Val Loss: 2.9946 | Val Acc: 52.16%
405
+ 2026-03-11 06:47:49,578 - Epoch [59/120] Completed in 3417s | ETA: 2 days, 9:54:37
406
+ 2026-03-11 06:47:49,658 - Train Loss: 3.3948 | Val Loss: 2.9766 | Val Acc: 51.99%
407
+ 2026-03-11 07:44:51,653 - Epoch [60/120] Completed in 3418s | ETA: 2 days, 8:58:23
408
+ 2026-03-11 07:44:51,653 - Train Loss: 3.3584 | Val Loss: 2.9990 | Val Acc: 52.17%
409
+ 2026-03-11 08:41:51,977 - Epoch [61/120] Completed in 3416s | ETA: 2 days, 7:59:58
410
+ 2026-03-11 08:41:51,995 - Train Loss: 3.3452 | Val Loss: 3.0195 | Val Acc: 52.12%
411
+ 2026-03-11 09:38:51,789 - Epoch [62/120] Completed in 3417s | ETA: 2 days, 7:03:18
412
+ 2026-03-11 09:38:51,840 - Train Loss: 3.3557 | Val Loss: 2.9761 | Val Acc: 52.32%
413
+ 2026-03-11 10:35:54,447 - Epoch [63/120] Completed in 3419s | ETA: 2 days, 6:08:19
414
+ 2026-03-11 10:35:54,449 - Train Loss: 3.3435 | Val Loss: 2.9746 | Val Acc: 52.17%
415
+ 2026-03-11 11:32:55,904 - Epoch [64/120] Completed in 3418s | ETA: 2 days, 5:10:08
416
+ 2026-03-11 11:32:55,919 - Train Loss: 3.3323 | Val Loss: 2.8906 | Val Acc: 54.45%
417
+ 2026-03-11 11:32:58,516 - New Best Accuracy: 54.45% - Saving Model...
418
+ 2026-03-11 12:29:57,530 - Epoch [65/120] Completed in 3417s | ETA: 2 days, 4:12:19
419
+ 2026-03-11 12:29:57,564 - Train Loss: 3.3051 | Val Loss: 2.9389 | Val Acc: 52.97%
420
+ 2026-03-11 13:26:57,857 - Epoch [66/120] Completed in 3417s | ETA: 2 days, 3:15:39
421
+ 2026-03-11 13:26:57,885 - Train Loss: 3.3266 | Val Loss: 2.9435 | Val Acc: 53.28%
422
+ 2026-03-11 14:23:59,820 - Epoch [67/120] Completed in 3418s | ETA: 2 days, 2:19:29
423
+ 2026-03-11 14:23:59,838 - Train Loss: 3.3100 | Val Loss: 2.9391 | Val Acc: 53.18%
424
+ 2026-03-11 15:21:00,492 - Epoch [68/120] Completed in 3417s | ETA: 2 days, 1:22:11
425
+ 2026-03-11 15:21:00,508 - Train Loss: 3.2992 | Val Loss: 2.9802 | Val Acc: 52.63%
426
+ 2026-03-11 16:18:00,018 - Epoch [69/120] Completed in 3416s | ETA: 2 days, 0:24:22
427
+ 2026-03-11 16:18:00,049 - Train Loss: 3.2812 | Val Loss: 2.9288 | Val Acc: 53.80%
428
+ 2026-03-11 17:15:02,113 - Epoch [70/120] Completed in 3418s | ETA: 1 day, 23:28:34
429
+ 2026-03-11 17:15:02,134 - Train Loss: 3.2531 | Val Loss: 2.9283 | Val Acc: 53.59%
430
+ 2026-03-11 18:12:01,397 - Epoch [71/120] Completed in 3415s | ETA: 1 day, 22:29:34
431
+ 2026-03-11 18:12:01,412 - Train Loss: 3.2429 | Val Loss: 2.9469 | Val Acc: 53.38%
432
+ 2026-03-11 19:09:01,503 - Epoch [72/120] Completed in 3417s | ETA: 1 day, 21:33:55
433
+ 2026-03-11 19:09:01,504 - Train Loss: 3.2276 | Val Loss: 2.9341 | Val Acc: 53.63%
434
+ 2026-03-11 20:06:00,430 - Epoch [73/120] Completed in 3415s | ETA: 1 day, 20:35:47
435
+ 2026-03-11 20:06:00,452 - Train Loss: 3.2019 | Val Loss: 2.9732 | Val Acc: 52.92%
436
+ 2026-03-11 21:03:02,527 - Epoch [74/120] Completed in 3418s | ETA: 1 day, 19:40:58
437
+ 2026-03-11 21:03:02,564 - Train Loss: 3.2033 | Val Loss: 2.9466 | Val Acc: 53.85%
438
+ 2026-03-11 22:00:04,961 - Epoch [75/120] Completed in 3419s | ETA: 1 day, 18:44:50
439
+ 2026-03-11 22:00:04,988 - Train Loss: 3.1955 | Val Loss: 2.9510 | Val Acc: 53.36%
440
+ 2026-03-11 22:57:09,110 - Epoch [76/120] Completed in 3420s | ETA: 1 day, 17:48:05
441
+ 2026-03-11 22:57:09,163 - Train Loss: 3.1709 | Val Loss: 2.9584 | Val Acc: 53.66%
442
+ 2026-03-11 23:54:09,474 - Epoch [77/120] Completed in 3416s | ETA: 1 day, 16:48:33
443
+ 2026-03-11 23:54:09,499 - Train Loss: 3.1468 | Val Loss: 2.9283 | Val Acc: 53.23%
444
+ 2026-03-12 00:51:09,678 - Epoch [78/120] Completed in 3417s | ETA: 1 day, 15:52:04
445
+ 2026-03-12 00:51:09,753 - Train Loss: 3.1605 | Val Loss: 2.9281 | Val Acc: 54.23%
446
+ 2026-03-12 01:48:08,810 - Epoch [79/120] Completed in 3416s | ETA: 1 day, 14:54:33
447
+ 2026-03-12 01:48:08,867 - Train Loss: 3.1382 | Val Loss: 3.0009 | Val Acc: 53.39%
448
+ 2026-03-12 02:45:10,047 - Epoch [80/120] Completed in 3417s | ETA: 1 day, 13:58:24
449
+ 2026-03-12 02:45:10,098 - Train Loss: 3.1162 | Val Loss: 2.8582 | Val Acc: 55.29%
450
+ 2026-03-12 02:45:12,996 - New Best Accuracy: 55.29% - Saving Model...
451
+ 2026-03-12 03:42:12,700 - Epoch [81/120] Completed in 3416s | ETA: 1 day, 13:00:57
452
+ 2026-03-12 03:42:12,729 - Train Loss: 3.1166 | Val Loss: 2.8631 | Val Acc: 54.96%
453
+ 2026-03-12 04:39:12,432 - Epoch [82/120] Completed in 3416s | ETA: 1 day, 12:04:03
454
+ 2026-03-12 04:39:12,484 - Train Loss: 3.1219 | Val Loss: 2.8845 | Val Acc: 54.78%
455
+ 2026-03-12 05:36:13,325 - Epoch [83/120] Completed in 3417s | ETA: 1 day, 11:07:27
456
+ 2026-03-12 05:36:13,413 - Train Loss: 3.0760 | Val Loss: 2.8952 | Val Acc: 54.93%
457
+ 2026-03-12 06:33:15,118 - Epoch [84/120] Completed in 3418s | ETA: 1 day, 10:11:01
458
+ 2026-03-12 06:33:15,146 - Train Loss: 3.0368 | Val Loss: 2.9457 | Val Acc: 54.25%
459
+ 2026-03-12 07:30:15,888 - Epoch [85/120] Completed in 3418s | ETA: 1 day, 9:14:09
460
+ 2026-03-12 07:30:15,889 - Train Loss: 3.0282 | Val Loss: 2.9537 | Val Acc: 54.24%
461
+ 2026-03-12 08:27:17,143 - Epoch [86/120] Completed in 3418s | ETA: 1 day, 8:17:07
462
+ 2026-03-12 08:27:17,163 - Train Loss: 3.0183 | Val Loss: 2.9113 | Val Acc: 54.73%
463
+ 2026-03-12 09:24:17,005 - Epoch [87/120] Completed in 3416s | ETA: 1 day, 7:19:17
464
+ 2026-03-12 09:24:17,018 - Train Loss: 2.9892 | Val Loss: 2.8974 | Val Acc: 54.99%
465
+ 2026-03-12 10:21:16,226 - Epoch [88/120] Completed in 3416s | ETA: 1 day, 6:22:12
466
+ 2026-03-12 10:21:16,226 - Train Loss: 2.9792 | Val Loss: 2.9543 | Val Acc: 54.20%
467
+ 2026-03-12 11:18:16,834 - Epoch [89/120] Completed in 3417s | ETA: 1 day, 5:25:51
468
+ 2026-03-12 11:18:16,848 - Train Loss: 2.9922 | Val Loss: 2.9240 | Val Acc: 54.68%
469
+ 2026-03-12 12:15:17,715 - Epoch [90/120] Completed in 3417s | ETA: 1 day, 4:28:52
470
+ 2026-03-12 12:15:17,756 - Train Loss: 2.9638 | Val Loss: 2.9329 | Val Acc: 54.40%
471
+ 2026-03-12 13:12:18,928 - Epoch [91/120] Completed in 3417s | ETA: 1 day, 3:31:56
472
+ 2026-03-12 13:12:18,942 - Train Loss: 2.9373 | Val Loss: 2.9200 | Val Acc: 54.56%
473
+ 2026-03-12 14:09:20,329 - Epoch [92/120] Completed in 3418s | ETA: 1 day, 2:35:14
474
+ 2026-03-12 14:09:20,384 - Train Loss: 2.9286 | Val Loss: 2.9743 | Val Acc: 54.36%
475
+ 2026-03-12 15:06:22,116 - Epoch [93/120] Completed in 3418s | ETA: 1 day, 1:38:24
476
+ 2026-03-12 15:06:22,156 - Train Loss: 2.9222 | Val Loss: 2.9585 | Val Acc: 53.95%
477
+ 2026-03-12 16:03:23,159 - Epoch [94/120] Completed in 3417s | ETA: 1 day, 0:40:59
478
+ 2026-03-12 16:03:23,172 - Train Loss: 2.9204 | Val Loss: 2.9674 | Val Acc: 54.20%
479
+ 2026-03-12 17:00:23,635 - Epoch [95/120] Completed in 3417s | ETA: 23:44:05
480
+ 2026-03-12 17:00:23,648 - Train Loss: 2.8748 | Val Loss: 2.9301 | Val Acc: 54.60%
481
+ 2026-03-12 17:57:23,681 - Epoch [96/120] Completed in 3417s | ETA: 22:46:51
482
+ 2026-03-12 17:57:23,736 - Train Loss: 2.8965 | Val Loss: 2.9597 | Val Acc: 54.27%
483
+ 2026-03-12 18:54:23,853 - Epoch [97/120] Completed in 3416s | ETA: 21:49:49
484
+ 2026-03-12 18:54:23,872 - Train Loss: 2.8752 | Val Loss: 2.9376 | Val Acc: 54.51%
485
+ 2026-03-12 19:51:23,769 - Epoch [98/120] Completed in 3417s | ETA: 20:53:00
486
+ 2026-03-12 19:51:23,769 - Train Loss: 2.8899 | Val Loss: 2.9197 | Val Acc: 55.02%
487
+ 2026-03-12 20:48:22,338 - Epoch [99/120] Completed in 3415s | ETA: 19:55:35
488
+ 2026-03-12 20:48:22,338 - Train Loss: 2.8395 | Val Loss: 2.9611 | Val Acc: 54.53%
489
+ 2026-03-12 21:45:24,378 - Epoch [100/120] Completed in 3419s | ETA: 18:59:43
490
+ 2026-03-12 21:45:24,379 - Train Loss: 2.8442 | Val Loss: 2.9590 | Val Acc: 54.56%
491
+ 2026-03-12 22:42:27,399 - Epoch [101/120] Completed in 3418s | ETA: 18:02:40
492
+ 2026-03-12 22:42:27,413 - Train Loss: 2.7988 | Val Loss: 2.9433 | Val Acc: 54.59%
493
+ 2026-03-12 23:39:27,139 - Epoch [102/120] Completed in 3417s | ETA: 17:05:12
494
+ 2026-03-12 23:39:27,150 - Train Loss: 2.8174 | Val Loss: 2.9209 | Val Acc: 55.02%
495
+ 2026-03-13 00:36:29,351 - Epoch [103/120] Completed in 3419s | ETA: 16:08:51
496
+ 2026-03-13 00:36:29,460 - Train Loss: 2.7868 | Val Loss: 2.9363 | Val Acc: 54.94%
497
+ 2026-03-13 01:33:31,803 - Epoch [104/120] Completed in 3418s | ETA: 15:11:37
498
+ 2026-03-13 01:33:31,834 - Train Loss: 2.8287 | Val Loss: 2.9620 | Val Acc: 55.08%
499
+ 2026-03-13 02:30:34,133 - Epoch [105/120] Completed in 3419s | ETA: 14:14:51
500
+ 2026-03-13 02:30:34,157 - Train Loss: 2.7896 | Val Loss: 2.9275 | Val Acc: 55.08%
501
+ 2026-03-13 03:27:37,318 - Epoch [106/120] Completed in 3420s | ETA: 13:18:06
502
+ 2026-03-13 03:27:37,360 - Train Loss: 2.7728 | Val Loss: 2.9330 | Val Acc: 55.12%
503
+ 2026-03-13 04:24:39,530 - Epoch [107/120] Completed in 3418s | ETA: 12:20:45
504
+ 2026-03-13 04:24:39,540 - Train Loss: 2.7787 | Val Loss: 2.9365 | Val Acc: 55.02%
505
+ 2026-03-13 05:21:40,244 - Epoch [108/120] Completed in 3418s | ETA: 11:23:38
506
+ 2026-03-13 05:21:40,253 - Train Loss: 2.7443 | Val Loss: 2.9508 | Val Acc: 55.18%
507
+ 2026-03-13 06:18:40,598 - Epoch [109/120] Completed in 3417s | ETA: 10:26:35
508
+ 2026-03-13 06:18:40,613 - Train Loss: 2.7623 | Val Loss: 2.9520 | Val Acc: 54.85%
509
+ 2026-03-13 07:15:41,488 - Epoch [110/120] Completed in 3417s | ETA: 9:29:36
510
+ 2026-03-13 07:15:41,574 - Train Loss: 2.7811 | Val Loss: 2.9439 | Val Acc: 55.12%
511
+ 2026-03-13 08:12:42,116 - Epoch [111/120] Completed in 3416s | ETA: 8:32:28
512
+ 2026-03-13 08:12:42,129 - Train Loss: 2.7385 | Val Loss: 2.9456 | Val Acc: 54.93%
513
+ 2026-03-13 09:09:41,211 - Epoch [112/120] Completed in 3416s | ETA: 7:35:32
514
+ 2026-03-13 09:09:41,224 - Train Loss: 2.7537 | Val Loss: 2.9456 | Val Acc: 54.91%
515
+ 2026-03-13 10:06:39,463 - Epoch [113/120] Completed in 3415s | ETA: 6:38:26
516
+ 2026-03-13 10:06:39,480 - Train Loss: 2.7413 | Val Loss: 2.9600 | Val Acc: 54.95%
517
+ 2026-03-13 11:03:39,346 - Epoch [114/120] Completed in 3416s | ETA: 5:41:41
518
+ 2026-03-13 11:03:39,359 - Train Loss: 2.7361 | Val Loss: 2.9484 | Val Acc: 55.13%
519
+ 2026-03-13 12:00:37,701 - Epoch [115/120] Completed in 3415s | ETA: 4:44:38
520
+ 2026-03-13 12:00:37,713 - Train Loss: 2.7186 | Val Loss: 2.9513 | Val Acc: 55.07%
521
+ 2026-03-13 12:57:34,386 - Epoch [116/120] Completed in 3414s | ETA: 3:47:36
522
+ 2026-03-13 12:57:34,400 - Train Loss: 2.7412 | Val Loss: 2.9570 | Val Acc: 55.06%
523
+ 2026-03-13 13:54:33,470 - Epoch [117/120] Completed in 3415s | ETA: 2:50:47
524
+ 2026-03-13 13:54:33,470 - Train Loss: 2.7303 | Val Loss: 2.9650 | Val Acc: 54.93%
525
+ 2026-03-13 14:51:33,476 - Epoch [118/120] Completed in 3416s | ETA: 1:53:52
526
+ 2026-03-13 14:51:33,499 - Train Loss: 2.7389 | Val Loss: 2.9476 | Val Acc: 55.02%
527
+ 2026-03-13 15:48:33,104 - Epoch [119/120] Completed in 3416s | ETA: 0:56:56
528
+ 2026-03-13 15:48:33,108 - Train Loss: 2.7346 | Val Loss: 2.9424 | Val Acc: 55.02%
529
+ 2026-03-13 16:45:32,606 - Epoch [120/120] Completed in 3416s | ETA: 0:00:00
530
+ 2026-03-13 16:45:32,625 - Train Loss: 2.7154 | Val Loss: 2.9443 | Val Acc: 55.12%
531
+ 2026-03-13 16:45:36,121 - Training Complete. Total Time: 4 days, 18:12:50. Best Accuracy: 55.29%
plots/learning_curve.png ADDED
results/per_class_acc_lrw1000_val.csv ADDED
@@ -0,0 +1,1185 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Rank,Class,Accuracy(%),Correct,Total
2
+ 1,中共中央政治局,100.00,19,19
3
+ 2,中华民族,100.00,6,6
4
+ 3,中央气象台,100.00,33,33
5
+ 4,丰硕,100.00,8,8
6
+ 5,二维码,100.00,7,7
7
+ 6,互联网,100.00,10,10
8
+ 7,保障,100.00,9,9
9
+ 8,办法,100.00,11,11
10
+ 9,华北,100.00,12,12
11
+ 10,卡塔尔,100.00,14,14
12
+ 11,印发,100.00,1,1
13
+ 12,广东,100.00,6,6
14
+ 13,开幕,100.00,19,19
15
+ 14,打造,100.00,6,6
16
+ 15,新闻联播,100.00,6,6
17
+ 16,方案,100.00,13,13
18
+ 17,研发,100.00,7,7
19
+ 18,福建,100.00,6,6
20
+ 19,发布,97.56,80,82
21
+ 20,习近平,97.30,36,37
22
+ 21,建设,96.88,31,32
23
+ 22,北京时间,96.15,25,26
24
+ 23,进一步,95.65,22,23
25
+ 24,变化,95.24,20,21
26
+ 25,委员会,95.24,20,21
27
+ 26,维护,95.24,20,21
28
+ 27,采取,95.24,20,21
29
+ 28,方面,95.00,95,100
30
+ 29,二十三,94.29,33,35
31
+ 30,报告,94.12,16,17
32
+ 31,超过,94.12,32,34
33
+ 32,峰会,93.33,28,30
34
+ 33,南部,93.10,27,29
35
+ 34,目前,92.97,119,128
36
+ 35,人民币,92.86,13,14
37
+ 36,国民党,92.86,13,14
38
+ 37,标准,92.86,13,14
39
+ 38,部署,92.86,13,14
40
+ 39,台湾,92.75,64,69
41
+ 40,关注,92.64,277,299
42
+ 41,依法,92.31,12,13
43
+ 42,部分,92.31,36,39
44
+ 43,双方,92.06,58,63
45
+ 44,外交部,92.06,58,63
46
+ 45,对外,91.67,11,12
47
+ 46,明确,91.67,22,24
48
+ 47,发表,91.53,54,59
49
+ 48,法国,91.43,32,35
50
+ 49,联邦,91.30,21,23
51
+ 50,表示,91.10,174,191
52
+ 51,调查,91.03,71,78
53
+ 52,二十五,90.91,20,22
54
+ 53,分别,90.91,10,11
55
+ 54,扶贫,90.91,10,11
56
+ 55,明显,90.91,10,11
57
+ 56,目标,90.91,20,22
58
+ 57,二零一七,90.62,29,32
59
+ 58,大会,90.62,29,32
60
+ 59,我国,90.54,67,74
61
+ 60,反对,90.48,19,21
62
+ 61,发展,90.24,111,123
63
+ 62,中共中央,90.00,9,10
64
+ 63,中部,90.00,9,10
65
+ 64,伊斯兰,90.00,9,10
66
+ 65,改革,90.00,18,20
67
+ 66,断交,90.00,9,10
68
+ 67,繁荣,90.00,9,10
69
+ 68,高校,90.00,9,10
70
+ 69,北京,89.66,52,58
71
+ 70,报道,89.56,163,182
72
+ 71,暴雨,89.55,60,67
73
+ 72,不过,89.47,34,38
74
+ 73,越来越,89.47,17,19
75
+ 74,十八,89.19,33,37
76
+ 75,一路,89.09,49,55
77
+ 76,国家,89.05,179,201
78
+ 77,作用,88.89,16,18
79
+ 78,发出,88.89,8,9
80
+ 79,安全,88.81,127,143
81
+ 80,媒体,88.76,79,89
82
+ 81,联合国,88.24,15,17
83
+ 82,打击,88.00,22,25
84
+ 83,高峰,88.00,22,25
85
+ 84,国防部,87.80,36,41
86
+ 85,市委,87.50,7,8
87
+ 86,水平,87.50,7,8
88
+ 87,法院,87.50,14,16
89
+ 88,海外,87.50,7,8
90
+ 89,确保,87.50,7,8
91
+ 90,违法,87.50,21,24
92
+ 91,部长,87.50,14,16
93
+ 92,面对,87.50,7,8
94
+ 93,政府,87.20,109,125
95
+ 94,南方,86.96,20,23
96
+ 95,国务院,86.96,20,23
97
+ 96,访问,86.89,53,61
98
+ 97,保护,86.67,13,15
99
+ 98,强调,86.67,26,30
100
+ 99,欧盟,86.67,13,15
101
+ 100,部门,86.67,39,45
102
+ 101,直播,86.54,90,104
103
+ 102,中方,86.52,77,89
104
+ 103,加强,86.36,38,44
105
+ 104,受伤,86.27,44,51
106
+ 105,北部,85.71,6,7
107
+ 106,江苏,85.71,6,7
108
+ 107,美国,85.46,288,337
109
+ 108,全国,85.19,46,54
110
+ 109,收看,85.19,161,189
111
+ 110,高温,85.19,23,27
112
+ 111,中国,84.58,351,415
113
+ 112,公开,84.21,16,19
114
+ 113,全球,83.67,41,49
115
+ 114,下午,83.33,55,66
116
+ 115,创新,83.33,10,12
117
+ 116,召开,83.33,20,24
118
+ 117,并且,83.33,35,42
119
+ 118,报,83.33,10,12
120
+ 119,江西,83.33,10,12
121
+ 120,深入,83.33,10,12
122
+ 121,资源,83.33,5,6
123
+ 122,工作,82.76,72,87
124
+ 123,城市,82.50,33,40
125
+ 124,新闻,82.41,253,307
126
+ 125,开展,82.14,23,28
127
+ 126,频道,82.14,23,28
128
+ 127,韩国,82.00,82,100
129
+ 128,号,81.99,428,522
130
+ 129,造成,81.93,68,83
131
+ 130,公民,81.82,18,22
132
+ 131,发挥,81.82,9,11
133
+ 132,引发,81.82,36,44
134
+ 133,新一轮,81.82,9,11
135
+ 134,贸易,81.82,9,11
136
+ 135,强降雨,81.58,31,38
137
+ 136,取得,81.48,22,27
138
+ 137,项目,81.48,22,27
139
+ 138,宣布,81.36,48,59
140
+ 139,警方,81.33,61,75
141
+ 140,客户端,81.25,13,16
142
+ 141,全部,80.95,17,21
143
+ 142,发生,80.85,114,141
144
+ 143,首次,80.85,38,47
145
+ 144,未来,80.77,42,52
146
+ 145,半岛,80.30,53,66
147
+ 146,不断,80.00,24,30
148
+ 147,五十,80.00,12,15
149
+ 148,代表,80.00,12,15
150
+ 149,商业,80.00,8,10
151
+ 150,对话,80.00,44,55
152
+ 151,开幕式,80.00,8,10
153
+ 152,比较,80.00,32,40
154
+ 153,秘书长,80.00,8,10
155
+ 154,聚焦,80.00,16,20
156
+ 155,至少,80.00,32,40
157
+ 156,详细,80.00,16,20
158
+ 157,负责人,80.00,12,15
159
+ 158,领导,80.00,12,15
160
+ 159,消息,79.90,159,199
161
+ 160,上午,79.73,59,74
162
+ 161,投票,79.66,47,59
163
+ 162,朝鲜,79.49,62,78
164
+ 163,高度,79.31,23,29
165
+ 164,我们,79.28,658,830
166
+ 165,视频,79.10,53,67
167
+ 166,不同,78.95,15,19
168
+ 167,以来,78.95,45,57
169
+ 168,元,78.95,15,19
170
+ 169,当地,78.86,97,123
171
+ 170,成果,78.79,26,33
172
+ 171,部,78.79,26,33
173
+ 172,严重,78.72,37,47
174
+ 173,昨天,78.63,103,131
175
+ 174,人员,78.57,66,84
176
+ 175,总书记,78.57,11,14
177
+ 176,战略,78.57,22,28
178
+ 177,下面,78.38,29,37
179
+ 178,的,78.31,4437,5666
180
+ 179,产品,78.26,18,23
181
+ 180,湖南,78.26,18,23
182
+ 181,更加,78.12,25,32
183
+ 182,问题,78.12,150,192
184
+ 183,交通,78.05,32,41
185
+ 184,合作,77.87,95,122
186
+ 185,举办,77.78,14,18
187
+ 186,广泛,77.78,7,9
188
+ 187,探索,77.78,7,9
189
+ 188,调整,77.78,7,9
190
+ 189,成为,77.59,45,58
191
+ 190,节目,77.38,65,84
192
+ 191,国际,77.30,109,141
193
+ 192,绑架,77.27,17,22
194
+ 193,讲话,77.27,17,22
195
+ 194,活动,77.11,64,83
196
+ 195,专家,76.92,20,26
197
+ 196,关键,76.92,10,13
198
+ 197,努力,76.92,30,39
199
+ 198,范围,76.92,20,26
200
+ 199,扫描,76.67,23,30
201
+ 200,中央,76.60,36,47
202
+ 201,重大,76.47,13,17
203
+ 202,马,76.47,52,68
204
+ 203,���括,76.36,42,55
205
+ 204,要求,75.76,25,33
206
+ 205,出现,75.64,59,78
207
+ 206,介绍,75.61,31,41
208
+ 207,全面,75.61,31,41
209
+ 208,另外,75.61,31,41
210
+ 209,非常,75.56,68,90
211
+ 210,二十,75.53,71,94
212
+ 211,万,75.47,40,53
213
+ 212,专项,75.00,6,8
214
+ 213,其他,75.00,36,48
215
+ 214,十二,75.00,15,20
216
+ 215,原因,75.00,12,16
217
+ 216,同时,75.00,30,40
218
+ 217,四川,75.00,6,8
219
+ 218,地区,75.00,111,148
220
+ 219,夫,75.00,30,40
221
+ 220,开放,75.00,12,16
222
+ 221,推动,75.00,18,24
223
+ 222,明天,75.00,18,24
224
+ 223,朗,75.00,117,156
225
+ 224,标志,75.00,9,12
226
+ 225,表达,75.00,6,8
227
+ 226,达成,75.00,15,20
228
+ 227,邀请,75.00,9,12
229
+ 228,银行,75.00,9,12
230
+ 229,随着,75.00,9,12
231
+ 230,黄,75.00,9,12
232
+ 231,不,74.85,128,171
233
+ 232,上海,74.68,59,79
234
+ 233,十五,74.42,32,43
235
+ 234,和平,74.14,43,58
236
+ 235,主权,73.91,17,23
237
+ 236,关于,73.91,17,23
238
+ 237,没有,73.83,79,107
239
+ 238,总统,73.82,141,191
240
+ 239,并,73.77,90,122
241
+ 240,传统,73.68,14,19
242
+ 241,相关,73.68,42,57
243
+ 242,评价,73.68,14,19
244
+ 243,十七,73.53,25,34
245
+ 244,叙利亚,73.47,36,49
246
+ 245,巴黎,73.33,11,15
247
+ 246,感受,73.33,11,15
248
+ 247,文化,73.33,11,15
249
+ 248,状态,73.33,11,15
250
+ 249,讨论,73.08,19,26
251
+ 250,好,72.96,143,196
252
+ 251,会上,72.73,32,44
253
+ 252,伤亡,72.73,8,11
254
+ 253,巴西,72.73,16,22
255
+ 254,资讯,72.73,16,22
256
+ 255,当中,72.46,100,138
257
+ 256,计划,72.41,21,29
258
+ 257,达到,72.41,21,29
259
+ 258,公布,72.34,34,47
260
+ 259,外交,72.22,26,36
261
+ 260,论坛,72.22,26,36
262
+ 261,方,71.96,77,107
263
+ 262,声明,71.74,33,46
264
+ 263,起来,71.74,33,46
265
+ 264,俄罗斯,71.55,83,116
266
+ 265,在,71.49,958,1340
267
+ 266,东部,71.43,15,21
268
+ 267,主席,71.43,50,70
269
+ 268,会晤,71.43,15,21
270
+ 269,共享,71.43,20,28
271
+ 270,寅,71.43,30,42
272
+ 271,最高,71.43,15,21
273
+ 272,环境,71.43,10,14
274
+ 273,风险,71.43,10,14
275
+ 274,情况,71.30,82,115
276
+ 275,解决,71.15,37,52
277
+ 276,十九,70.83,34,48
278
+ 277,原则,70.83,17,24
279
+ 278,如何,70.83,17,24
280
+ 279,民众,70.83,17,24
281
+ 280,领导人,70.83,17,24
282
+ 281,举行,70.73,87,123
283
+ 282,央视,70.65,65,92
284
+ 283,不仅,70.59,12,17
285
+ 284,反,70.37,19,27
286
+ 285,发布会,70.37,19,27
287
+ 286,导致,70.37,19,27
288
+ 287,接受,70.37,19,27
289
+ 288,方式,70.37,19,27
290
+ 289,名,70.34,102,145
291
+ 290,市场,70.27,26,37
292
+ 291,开始,70.10,68,97
293
+ 292,保持,70.00,14,20
294
+ 293,各方,70.00,28,40
295
+ 294,大雨,70.00,7,10
296
+ 295,德国,70.00,21,30
297
+ 296,深化,70.00,7,10
298
+ 297,欢迎,69.84,88,126
299
+ 298,倡议,69.57,16,23
300
+ 299,平台,69.57,16,23
301
+ 300,不少,69.23,18,26
302
+ 301,当前,69.23,9,13
303
+ 302,身份,69.23,9,13
304
+ 303,投资,68.97,20,29
305
+ 304,两岸,68.75,22,32
306
+ 305,社会,68.75,33,48
307
+ 306,移动,68.75,22,32
308
+ 307,持续,68.66,46,67
309
+ 308,提出,68.57,24,35
310
+ 309,八,68.52,37,54
311
+ 310,分子,68.42,13,19
312
+ 311,欧洲,68.42,13,19
313
+ 312,网,68.42,26,38
314
+ 313,共同,68.29,56,82
315
+ 314,重要,68.25,43,63
316
+ 315,所谓,68.18,15,22
317
+ 316,最新,67.92,36,53
318
+ 317,首相,67.74,21,31
319
+ 318,共识,67.65,23,34
320
+ 319,主要,67.39,31,46
321
+ 320,时间,67.23,119,177
322
+ 321,将会,66.97,73,109
323
+ 322,关系,66.95,79,118
324
+ 323,之外,66.67,10,15
325
+ 324,书记,66.67,8,12
326
+ 325,他们,66.67,52,78
327
+ 326,伙伴,66.67,4,6
328
+ 327,儿童,66.67,8,12
329
+ 328,全体,66.67,4,6
330
+ 329,再次,66.67,26,39
331
+ 330,准备,66.67,10,15
332
+ 331,分钟,66.67,10,15
333
+ 332,卫生,66.67,8,12
334
+ 333,友好,66.67,6,9
335
+ 334,友谊,66.67,4,6
336
+ 335,回到,66.67,12,18
337
+ 336,基本,66.67,16,24
338
+ 337,外长,66.67,12,18
339
+ 338,大,66.67,142,213
340
+ 339,天然气,66.67,4,6
341
+ 340,存在,66.67,18,27
342
+ 341,接下来,66.67,48,72
343
+ 342,提高,66.67,8,12
344
+ 343,搜索,66.67,14,21
345
+ 344,日本,66.67,54,81
346
+ 345,时代,66.67,6,9
347
+ 346,最后,66.67,26,39
348
+ 347,案件,66.67,12,18
349
+ 348,网站,66.67,12,18
350
+ 349,转向,66.67,14,21
351
+ 350,防范,66.67,6,9
352
+ 351,高级,66.67,12,18
353
+ 352,来,66.54,338,508
354
+ 353,之后,66.10,78,118
355
+ 354,公司,65.96,31,47
356
+ 355,发言人,65.96,62,94
357
+ 356,左右,65.79,25,38
358
+ 357,推进,65.79,25,38
359
+ 358,受到,65.62,21,32
360
+ 359,正在,65.58,101,154
361
+ 360,内容,65.45,36,55
362
+ 361,交流,65.22,15,23
363
+ 362,有关,65.00,65,100
364
+ 363,居民,64.71,11,17
365
+ 364,采访,64.71,11,17
366
+ 365,极端,64.44,29,45
367
+ 366,首都,64.44,29,45
368
+ 367,分歧,64.29,9,14
369
+ 368,怎么样,64.29,9,14
370
+ 369,抵达,64.29,18,28
371
+ 370,指出,64.29,9,14
372
+ 371,西部,64.29,9,14
373
+ 372,通报,64.29,9,14
374
+ 373,天气,64.20,52,81
375
+ 374,那么,64.15,136,212
376
+ 375,感谢,64.13,59,92
377
+ 376,生活,64.10,25,39
378
+ 377,大家,64.04,57,89
379
+ 378,周年,63.89,23,36
380
+ 379,二十七,63.64,14,22
381
+ 380,做好,63.64,7,11
382
+ 381,愿意,63.64,7,11
383
+ 382,方向,63.64,14,22
384
+ 383,最终,63.64,14,22
385
+ 384,落实,63.64,7,11
386
+ 385,调研,63.64,7,11
387
+ 386,晚上,63.46,33,52
388
+ 387,预警,63.33,19,30
389
+ 388,首先,63.33,38,60
390
+ 389,完成,63.16,12,19
391
+ 390,服务,63.16,12,19
392
+ 391,法,63.16,12,19
393
+ 392,被,63.16,120,190
394
+ 393,选举,63.16,12,19
395
+ 394,更多,62.90,39,62
396
+ 395,了解,62.50,30,48
397
+ 396,优势,62.50,5,8
398
+ 397,做到,62.50,5,8
399
+ 398,协商,62.50,10,16
400
+ 399,广西,62.50,5,8
401
+ 400,河北,62.50,5,8
402
+ 401,浙江,62.50,5,8
403
+ 402,生产,62.50,5,8
404
+ 403,管理,62.50,15,24
405
+ 404,良好,62.50,5,8
406
+ 405,香港,62.45,153,245
407
+ 406,东方,62.34,48,77
408
+ 407,两,62.33,134,215
409
+ 408,行动,62.26,33,53
410
+ 409,就是,62.22,168,270
411
+ 410,当天,62.22,28,45
412
+ 411,过程,62.22,28,45
413
+ 412,死亡,62.07,36,58
414
+ 413,无人机,61.90,13,21
415
+ 414,为,61.79,131,212
416
+ 415,华,61.76,21,34
417
+ 416,组织,61.76,63,102
418
+ 417,产业,61.54,8,13
419
+ 418,湖北,61.54,8,13
420
+ 419,稳定,61.54,24,39
421
+ 420,连续,61.54,8,13
422
+ 421,重点,61.54,8,13
423
+ 422,回归,61.36,27,44
424
+ 423,记者,61.15,85,139
425
+ 424,显示,61.11,22,36
426
+ 425,条件,61.11,11,18
427
+ 426,犯罪,61.11,11,18
428
+ 427,视线,61.11,11,18
429
+ 428,集团,61.11,11,18
430
+ 429,需要,61.11,22,36
431
+ 430,去年,60.87,14,23
432
+ 431,委员,60.87,14,23
433
+ 432,岛,60.87,14,23
434
+ 433,意见,60.71,17,28
435
+ 434,所有,60.71,17,28
436
+ 435,展开,60.61,20,33
437
+ 436,同胞,60.00,9,15
438
+ 437,团队,60.00,9,15
439
+ 438,坚决,60.00,9,15
440
+ 439,大型,60.00,6,10
441
+ 440,大学,60.00,6,10
442
+ 441,推出,60.00,6,10
443
+ 442,文明,60.00,9,15
444
+ 443,有效,60.00,6,10
445
+ 444,灾害,60.00,6,10
446
+ 445,班,60.00,3,5
447
+ 446,规范,60.00,6,10
448
+ 447,默,60.00,6,10
449
+ 448,但是,59.85,79,132
450
+ 449,国内,59.38,19,32
451
+ 450,确认,59.38,19,32
452
+ 451,希望,59.32,35,59
453
+ 452,主义,59.26,16,27
454
+ 453,十四,59.26,16,27
455
+ 454,针对,59.26,16,27
456
+ 455,还有,59.15,42,71
457
+ 456,协议,59.09,13,22
458
+ 457,发现,58.97,23,39
459
+ 458,三十,58.62,17,29
460
+ 459,首,58.62,17,29
461
+ 460,将,58.53,151,258
462
+ 461,负责,58.33,7,12
463
+ 462,影响,57.58,38,66
464
+ 463,看,57.52,153,266
465
+ 464,不要,57.14,4,7
466
+ 465,共和国,57.14,4,7
467
+ 466,完全,57.14,12,21
468
+ 467,工业,57.14,4,7
469
+ 468,文,57.14,36,63
470
+ 469,河南,57.14,4,7
471
+ 470,重视,57.14,4,7
472
+ 471,领域,57.14,16,28
473
+ 472,中,56.79,92,162
474
+ 473,把,56.79,46,81
475
+ 474,或者,56.76,21,37
476
+ 475,英国,56.69,72,127
477
+ 476,届,56.67,17,30
478
+ 477,机构,56.52,13,23
479
+ 478,比如,56.52,13,23
480
+ 479,带,56.41,44,78
481
+ 480,一下,56.25,54,96
482
+ 481,基础,56.25,9,16
483
+ 482,科学,56.25,9,16
484
+ 483,多,56.19,127,226
485
+ 484,联合,56.10,23,41
486
+ 485,会见,56.00,14,25
487
+ 486,继续,56.00,70,125
488
+ 487,和,55.98,295,527
489
+ 488,作为,55.88,38,68
490
+ 489,军方,55.88,19,34
491
+ 490,后,55.88,19,34
492
+ 491,不能,55.56,15,27
493
+ 492,二十六,55.56,10,18
494
+ 493,人民,55.56,10,18
495
+ 494,企业,55.56,20,36
496
+ 495,健康,55.56,5,9
497
+ 496,先进,55.56,5,9
498
+ 497,嘉宾,55.56,5,9
499
+ 498,增加,55.56,10,18
500
+ 499,工程,55.56,5,9
501
+ 500,开发,55.56,5,9
502
+ 501,朋友,55.56,10,18
503
+ 502,设备,55.56,5,9
504
+ 503,迎来,55.56,15,27
505
+ 504,选择,55.56,10,18
506
+ 505,凌晨,55.26,21,38
507
+ 506,特别,55.26,21,38
508
+ 507,特,55.06,147,267
509
+ 508,沙特,54.84,17,31
510
+ 509,嫌疑人,54.55,12,22
511
+ 510,海上,54.55,6,11
512
+ 511,连线,54.17,13,24
513
+ 512,看到,54.12,46,85
514
+ 513,五,53.97,68,126
515
+ 514,具体,53.85,14,26
516
+ 515,出席,53.85,21,39
517
+ 516,十三,53.85,7,13
518
+ 517,亿,53.66,22,41
519
+ 518,一系列,53.33,8,15
520
+ 519,之下,53.33,8,15
521
+ 520,十六,53.33,8,15
522
+ 521,挑战,53.33,8,15
523
+ 522,提升,53.33,8,15
524
+ 523,能够,53.33,48,90
525
+ 524,山,52.94,9,17
526
+ 525,对,52.80,179,339
527
+ 526,而,52.71,107,203
528
+ 527,数据,52.63,10,19
529
+ 528,赴,52.63,10,19
530
+ 529,会,52.61,131,249
531
+ 530,建立,52.38,11,21
532
+ 531,普京,52.38,11,21
533
+ 532,称,52.38,33,63
534
+ 533,维,52.38,11,21
535
+ 534,第,52.24,105,201
536
+ 535,份,52.17,12,23
537
+ 536,措施,52.17,12,23
538
+ 537,很多,52.05,38,73
539
+ 538,二十一,52.00,13,25
540
+ 539,坚持,52.00,13,25
541
+ 540,遭遇,52.00,13,25
542
+ 541,百,51.85,14,27
543
+ 542,莹,51.72,15,29
544
+ 543,金融,51.72,15,29
545
+ 544,看看,51.67,31,60
546
+ 545,信息,51.61,16,31
547
+ 546,参与,51.61,32,62
548
+ 547,来自,51.35,19,37
549
+ 548,韩,51.06,24,47
550
+ 549,正式,50.88,29,57
551
+ 550,还,50.65,78,154
552
+ 551,通过,50.56,45,89
553
+ 552,进行,50.52,98,194
554
+ 553,是,50.38,528,1048
555
+ 554,上升,50.00,5,10
556
+ 555,严格,50.00,4,8
557
+ 556,主体,50.00,2,4
558
+ 557,事务,50.00,6,12
559
+ 558,二十四,50.00,15,30
560
+ 559,于,50.00,35,70
561
+ 560,以下,50.00,1,2
562
+ 561,会议,50.00,22,44
563
+ 562,共有,50.00,6,12
564
+ 563,周,50.00,6,12
565
+ 564,国,50.00,56,112
566
+ 565,威胁,50.00,9,18
567
+ 566,孩子,50.00,26,52
568
+ 567,开启,50.00,4,8
569
+ 568,强烈,50.00,12,24
570
+ 569,必须,50.00,6,12
571
+ 570,意识,50.00,2,4
572
+ 571,改善,50.00,7,14
573
+ 572,放,50.00,1,2
574
+ 573,日前,50.00,12,24
575
+ 574,春,50.00,13,26
576
+ 575,普遍,50.00,5,10
577
+ 576,更好,50.00,4,8
578
+ 577,本月,50.00,8,16
579
+ 578,机制,50.00,11,22
580
+ 579,而且,50.00,20,40
581
+ 580,证,50.00,1,2
582
+ 581,贡献,50.00,4,8
583
+ 582,资本,50.00,4,8
584
+ 583,过去,50.00,14,28
585
+ 584,龙,50.00,10,20
586
+ 585,进入,49.35,38,77
587
+ 586,来说,48.72,19,39
588
+ 587,高,48.48,16,33
589
+ 588,今天,48.39,225,465
590
+ 589,预计,48.39,15,31
591
+ 590,与,48.04,86,179
592
+ 591,发,48.00,12,25
593
+ 592,地方,47.92,23,48
594
+ 593,决定,47.83,11,23
595
+ 594,省,47.83,11,23
596
+ 595,上,47.71,125,262
597
+ 596,外,47.62,10,21
598
+ 597,王,47.62,10,21
599
+ 598,还是,47.62,20,42
600
+ 599,到,47.46,131,276
601
+ 600,今年,47.44,37,78
602
+ 601,积极,47.37,9,19
603
+ 602,一,47.09,389,826
604
+ 603,分,47.06,8,17
605
+ 604,年,47.06,72,153
606
+ 605,应对,47.06,8,17
607
+ 606,小时,46.88,15,32
608
+ 607,产生,46.67,7,15
609
+ 608,促进,46.67,7,15
610
+ 609,兰,46.67,14,30
611
+ 610,按照,46.67,7,15
612
+ 611,再,46.54,101,217
613
+ 612,其中,46.43,26,56
614
+ 613,行为,46.43,26,56
615
+ 614,实现,46.34,19,41
616
+ 615,稍后,46.30,25,54
617
+ 616,以上,46.15,6,13
618
+ 617,具有,46.15,6,13
619
+ 618,制造,46.15,6,13
620
+ 619,千,46.15,6,13
621
+ 620,形势,46.15,12,26
622
+ 621,沟通,46.15,6,13
623
+ 622,现场,46.05,35,76
624
+ 623,认为,45.95,17,37
625
+ 624,美,45.86,72,157
626
+ 625,北约,45.83,11,24
627
+ 626,政策,45.71,16,35
628
+ 627,专业,45.45,5,11
629
+ 628,公共,45.45,5,11
630
+ 629,再见,45.45,10,22
631
+ 630,改变,45.45,10,22
632
+ 631,结果,45.45,15,33
633
+ 632,综合,45.45,10,22
634
+ 633,责任,45.45,10,22
635
+ 634,带来,45.24,19,42
636
+ 635,二,45.12,37,82
637
+ 636,三,45.05,41,91
638
+ 637,主题,45.00,9,20
639
+ 638,参加,45.00,9,20
640
+ 639,我,45.00,117,260
641
+ 640,播出,45.00,9,20
642
+ 641,根据,45.00,18,40
643
+ 642,有,44.59,198,444
644
+ 643,位于,44.44,8,18
645
+ 644,区,44.44,12,27
646
+ 645,四十,44.44,4,9
647
+ 646,处理,44.44,12,27
648
+ 647,安,44.44,12,27
649
+ 648,抓,44.44,4,9
650
+ 649,数,44.44,4,9
651
+ 650,监管,44.44,4,9
652
+ 651,世界,44.07,26,59
653
+ 652,以后,43.59,17,39
654
+ 653,回应,43.48,10,23
655
+ 654,话题,43.48,30,69
656
+ 655,新,43.41,56,129
657
+ 656,科技,43.33,13,30
658
+ 657,获得,43.33,13,30
659
+ 658,加,43.24,16,37
660
+ 659,印度,43.18,19,44
661
+ 660,下载,42.86,9,21
662
+ 661,体制,42.86,3,7
663
+ 662,作出,42.86,12,28
664
+ 663,即将,42.86,6,14
665
+ 664,应急,42.86,3,7
666
+ 665,拥有,42.86,3,7
667
+ 666,桥,42.86,3,7
668
+ 667,片,42.86,3,7
669
+ 668,真正,42.86,6,14
670
+ 669,近年来,42.86,3,7
671
+ 670,说,42.62,101,237
672
+ 671,经济,42.59,23,54
673
+ 672,啊,42.54,171,402
674
+ 673,至,42.42,14,33
675
+ 674,事故,42.31,11,26
676
+ 675,注意,42.11,8,19
677
+ 676,一道,41.67,5,12
678
+ 677,以及,41.67,30,72
679
+ 678,双,41.67,5,12
680
+ 679,局,41.67,15,36
681
+ 680,意味着,41.67,5,12
682
+ 681,意外,41.67,5,12
683
+ 682,承诺,41.67,5,12
684
+ 683,月,41.51,22,53
685
+ 684,总理,41.46,17,41
686
+ 685,比,41.46,17,41
687
+ 686,就,41.36,158,382
688
+ 687,什么,41.27,26,63
689
+ 688,教育,41.18,7,17
690
+ 689,无,41.18,7,17
691
+ 690,是不是,41.18,7,17
692
+ 691,期间,41.18,7,17
693
+ 692,访,41.03,16,39
694
+ 693,这,40.91,207,506
695
+ 694,已经,40.80,102,250
696
+ 695,降雨,40.62,13,32
697
+ 696,们,40.54,15,37
698
+ 697,祖国,40.54,15,37
699
+ 698,系统,40.54,15,37
700
+ 699,等,40.51,64,158
701
+ 700,最,40.16,51,127
702
+ 701,三天,40.00,4,10
703
+ 702,中午,40.00,4,10
704
+ 703,中心,40.00,16,40
705
+ 704,人士,40.00,6,15
706
+ 705,价格,40.00,4,10
707
+ 706,使用,40.00,8,20
708
+ 707,党,40.00,10,25
709
+ 708,六,40.00,22,55
710
+ 709,办,40.00,6,15
711
+ 710,医院,40.00,10,25
712
+ 711,协调,40.00,4,10
713
+ 712,就是说,40.00,4,10
714
+ 713,巴,40.00,10,25
715
+ 714,引起,40.00,4,10
716
+ 715,执行,40.00,8,20
717
+ 716,指导,40.00,4,10
718
+ 717,提供,40.00,18,45
719
+ 718,机关,40.00,8,20
720
+ 719,涉嫌,40.00,6,15
721
+ 720,这里,40.00,28,70
722
+ 721,了,39.84,506,1270
723
+ 722,从,39.74,62,156
724
+ 723,也,39.47,210,532
725
+ 724,对于,39.47,30,76
726
+ 725,一起,39.39,26,66
727
+ 726,局势,39.39,13,33
728
+ 727,不会,39.29,11,28
729
+ 728,斯,39.22,20,51
730
+ 729,尔,39.18,38,97
731
+ 730,会谈,39.13,9,23
732
+ 731,相信,39.13,9,23
733
+ 732,通,39.13,9,23
734
+ 733,巨大,38.89,7,18
735
+ 734,此,38.71,24,62
736
+ 735,本,38.64,17,44
737
+ 736,人,38.49,122,317
738
+ 737,之前,38.46,10,26
739
+ 738,岁,38.46,10,26
740
+ 739,成功,38.46,15,39
741
+ 740,所以,38.18,21,55
742
+ 741,话,37.74,20,53
743
+ 742,专机,37.50,3,8
744
+ 743,亚,37.50,9,24
745
+ 744,利益,37.50,6,16
746
+ 745,区域,37.50,6,16
747
+ 746,咱们,37.50,3,8
748
+ 747,大约,37.50,6,16
749
+ 748,气温,37.50,6,16
750
+ 749,研究,37.50,6,16
751
+ 750,突破,37.50,3,8
752
+ 751,自主,37.50,3,8
753
+ 752,门,37.50,9,24
754
+ 753,阶段,37.50,6,16
755
+ 754,陆续,37.50,3,8
756
+ 755,一次,37.31,25,67
757
+ 756,袭击,37.17,42,113
758
+ 757,可以,36.99,64,173
759
+ 758,人数,36.84,7,19
760
+ 759,罗,36.84,14,38
761
+ 760,轮,36.84,7,19
762
+ 761,因为,36.47,31,85
763
+ 762,全,36.36,4,11
764
+ 763,少,36.36,4,11
765
+ 764,展示,36.36,4,11
766
+ 765,每年,36.36,4,11
767
+ 766,生命,36.36,4,11
768
+ 767,军事,36.11,13,36
769
+ 768,阿,36.11,13,36
770
+ 769,海,36.00,9,25
771
+ 770,据,35.90,28,78
772
+ 771,之,35.71,30,84
773
+ 772,技术,35.71,10,28
774
+ 773,火灾,35.71,10,28
775
+ 774,现在,35.67,61,171
776
+ 775,驻,35.56,16,45
777
+ 776,化,35.48,11,31
778
+ 777,铁路,35.48,11,31
779
+ 778,前,35.44,28,79
780
+ 779,然后,35.38,23,65
781
+ 780,公里,35.29,6,17
782
+ 781,危险,35.29,6,17
783
+ 782,这个,35.27,176,499
784
+ 783,当时,35.00,7,20
785
+ 784,事件,34.58,37,107
786
+ 785,支持,34.48,10,29
787
+ 786,同,34.43,21,61
788
+ 787,九,34.38,11,32
789
+ 788,政治,34.38,11,32
790
+ 789,里面,34.38,11,32
791
+ 790,核,34.29,12,35
792
+ 791,到了,33.93,19,56
793
+ 792,四,33.90,20,59
794
+ 793,近日,33.87,21,62
795
+ 794,七十,33.33,4,12
796
+ 795,中俄,33.33,2,6
797
+ 796,乘,33.33,2,6
798
+ 797,任何,33.33,11,33
799
+ 798,位置,33.33,4,12
800
+ 799,作,33.33,1,3
801
+ 800,信心,33.33,4,12
802
+ 801,刚刚,33.33,7,21
803
+ 802,别,33.33,3,9
804
+ 803,制度,33.33,5,15
805
+ 804,医疗,33.33,2,6
806
+ 805,历史,33.33,14,42
807
+ 806,各界,33.33,3,9
808
+ 807,多次,33.33,6,18
809
+ 808,市,33.33,7,21
810
+ 809,式,33.33,7,21
811
+ 810,得到,33.33,6,18
812
+ 811,成立,33.33,4,12
813
+ 812,截止,33.33,5,15
814
+ 813,所在,33.33,5,15
815
+ 814,找到,33.33,4,12
816
+ 815,控制,33.33,4,12
817
+ 816,时候,33.33,30,90
818
+ 817,本次,33.33,3,9
819
+ 818,点,33.33,33,99
820
+ 819,等等,33.33,5,15
821
+ 820,约,33.33,7,21
822
+ 821,著名,33.33,3,9
823
+ 822,请,33.33,10,30
824
+ 823,走,33.33,4,12
825
+ 824,转,33.33,3,9
826
+ 825,您,33.02,104,315
827
+ 826,个,32.72,71,217
828
+ 827,觉得,32.65,16,49
829
+ 828,一种,32.43,12,37
830
+ 829,拉,32.43,24,74
831
+ 830,进展,32.26,10,31
832
+ 831,吧,32.14,9,28
833
+ 832,要,32.04,58,181
834
+ 833,汽车,31.82,7,22
835
+ 834,小,31.71,13,41
836
+ 835,嗯,31.63,31,98
837
+ 836,启动,31.58,6,19
838
+ 837,精神,31.58,6,19
839
+ 838,十,31.51,23,73
840
+ 839,可能,31.46,28,89
841
+ 840,天,31.37,32,102
842
+ 841,起,31.25,15,48
843
+ 842,怎么,30.91,17,55
844
+ 843,一直,30.77,12,39
845
+ 844,为了,30.77,12,39
846
+ 845,性,30.77,12,39
847
+ 846,新型,30.77,4,13
848
+ 847,老,30.77,12,39
849
+ 848,但,30.56,11,36
850
+ 849,最近,30.56,11,36
851
+ 850,实施,30.43,14,46
852
+ 851,迪,30.43,7,23
853
+ 852,次,30.09,34,113
854
+ 853,整个,30.00,6,20
855
+ 854,段,30.00,18,60
856
+ 855,集中,30.00,3,10
857
+ 856,下,29.89,26,87
858
+ 857,七,29.73,11,37
859
+ 858,路,29.63,8,27
860
+ 859,感觉,29.41,5,17
861
+ 860,项,29.41,5,17
862
+ 861,仍然,29.03,9,31
863
+ 862,三年,28.57,2,7
864
+ 863,侧,28.57,2,7
865
+ 864,力量,28.57,4,14
866
+ 865,型,28.57,2,7
867
+ 866,机场,28.57,8,28
868
+ 867,活力,28.57,2,7
869
+ 868,短,28.57,4,14
870
+ 869,空间,28.57,2,7
871
+ 870,贵州,28.57,2,7
872
+ 871,车,28.57,10,35
873
+ 872,场,28.30,15,53
874
+ 873,那,28.19,42,149
875
+ 874,近,28.00,7,25
876
+ 875,卡,27.78,5,18
877
+ 876,回,27.59,8,29
878
+ 877,塔,27.59,8,29
879
+ 878,结束,27.59,8,29
880
+ 879,智能,27.27,3,11
881
+ 880,祝贺,27.27,3,11
882
+ 881,统一,27.27,3,11
883
+ 882,联系,27.27,3,11
884
+ 883,能源,27.27,3,11
885
+ 884,这么,27.27,6,22
886
+ 885,长,27.27,12,44
887
+ 886,除了,27.27,6,22
888
+ 887,都,27.23,52,191
889
+ 888,自己,27.06,23,85
890
+ 889,军,27.03,10,37
891
+ 890,升级,26.67,4,15
892
+ 891,说是,26.67,4,15
893
+ 892,时,26.56,17,64
894
+ 893,规定,26.32,5,19
895
+ 894,应该,26.03,19,73
896
+ 895,以,25.88,22,85
897
+ 896,克,25.81,8,31
898
+ 897,一些,25.71,27,105
899
+ 898,一个,25.07,92,367
900
+ 899,一样,25.00,6,24
901
+ 900,京,25.00,1,4
902
+ 901,住,25.00,2,8
903
+ 902,做,25.00,13,52
904
+ 903,力,25.00,2,8
905
+ 904,多年,25.00,2,8
906
+ 905,奇,25.00,1,4
907
+ 906,实际,25.00,2,8
908
+ 907,家,25.00,7,28
909
+ 908,开,25.00,5,20
910
+ 909,整治,25.00,1,4
911
+ 910,暨,25.00,2,8
912
+ 911,有所,25.00,2,8
913
+ 912,梦,25.00,1,4
914
+ 913,正常,25.00,2,8
915
+ 914,签署,25.00,4,16
916
+ 915,自,25.00,5,20
917
+ 916,议会,25.00,3,12
918
+ 917,试,25.00,1,4
919
+ 918,违反,25.00,4,16
920
+ 919,金砖,25.00,2,8
921
+ 920,长期,25.00,3,12
922
+ 921,面临,25.00,2,8
923
+ 922,很,24.32,18,74
924
+ 923,台,24.19,15,62
925
+ 924,立即,24.00,6,25
926
+ 925,由于,23.81,5,21
927
+ 926,体现,23.53,4,17
928
+ 927,公安,23.53,4,17
929
+ 928,各地,23.53,4,17
930
+ 929,吗,23.53,4,17
931
+ 930,太,23.53,4,17
932
+ 931,那个,23.53,12,51
933
+ 932,去,23.38,18,77
934
+ 933,出,23.33,14,60
935
+ 934,水,23.08,6,26
936
+ 935,钱,23.08,3,13
937
+ 936,任务,22.86,8,35
938
+ 937,强,22.73,5,22
939
+ 938,此次,22.73,5,22
940
+ 939,涉及,22.73,5,22
941
+ 940,经过,22.73,5,22
942
+ 941,这样,22.33,23,103
943
+ 942,动力,22.22,2,9
944
+ 943,只有,22.22,4,18
945
+ 944,听,22.22,4,18
946
+ 945,快,22.22,2,9
947
+ 946,线,22.22,4,18
948
+ 947,融合,22.22,2,9
949
+ 948,辆,22.22,4,18
950
+ 949,先,21.88,7,32
951
+ 950,所,21.82,12,55
952
+ 951,一定,21.62,8,37
953
+ 952,一点,21.62,8,37
954
+ 953,几,21.57,11,51
955
+ 954,让,21.50,23,107
956
+ 955,座,21.43,6,28
957
+ 956,级,21.43,3,14
958
+ 957,呃,21.30,23,108
959
+ 958,呢,21.14,163,771
960
+ 959,向,21.11,19,90
961
+ 960,受,21.05,4,19
962
+ 961,梅,20.83,5,24
963
+ 962,用,20.83,10,48
964
+ 963,各位,20.69,6,29
965
+ 964,哈,20.69,6,29
966
+ 965,与会,20.00,2,10
967
+ 966,习,20.00,1,5
968
+ 967,予以,20.00,2,10
969
+ 968,关心,20.00,2,10
970
+ 969,喜欢,20.00,2,10
971
+ 970,地,20.00,28,140
972
+ 971,形成,20.00,2,10
973
+ 972,改,20.00,1,5
974
+ 973,权,20.00,1,5
975
+ 974,核心,20.00,2,10
976
+ 975,森,20.00,4,20
977
+ 976,记录,20.00,3,15
978
+ 977,该,20.00,5,25
979
+ 978,这些,20.00,12,60
980
+ 979,由,19.15,9,47
981
+ 980,及,19.05,4,21
982
+ 981,能力,19.05,4,21
983
+ 982,实践,18.75,3,16
984
+ 983,如果,18.37,9,49
985
+ 984,人类,18.18,2,11
986
+ 985,地址,18.18,4,22
987
+ 986,林,18.18,4,22
988
+ 987,每,18.18,4,22
989
+ 988,系列,18.18,2,11
990
+ 989,遭,18.18,2,11
991
+ 990,这种,18.06,13,72
992
+ 991,知道,17.86,5,28
993
+ 992,又,17.81,13,73
994
+ 993,他,17.46,33,189
995
+ 994,纳,17.39,4,23
996
+ 995,位,16.92,11,65
997
+ 996,条,16.91,23,136
998
+ 997,东,16.67,3,18
999
+ 998,庆祝,16.67,4,24
1000
+ 999,愿,16.67,2,12
1001
+ 1000,有的,16.67,2,12
1002
+ 1001,着,16.67,10,60
1003
+ 1002,统计,16.67,1,6
1004
+ 1003,者,16.67,6,36
1005
+ 1004,认识,16.67,1,6
1006
+ 1005,其实,16.13,5,31
1007
+ 1006,德,16.07,9,56
1008
+ 1007,此前,15.62,5,32
1009
+ 1008,及时,15.38,2,13
1010
+ 1009,坚定,15.38,2,13
1011
+ 1010,利用,15.00,3,20
1012
+ 1011,期待,14.81,4,27
1013
+ 1012,都是,14.55,8,55
1014
+ 1013,义,14.29,2,14
1015
+ 1014,仪式,14.29,4,28
1016
+ 1015,件,14.29,3,21
1017
+ 1016,出来,14.29,3,21
1018
+ 1017,各国,14.29,2,14
1019
+ 1018,埃及,14.29,1,7
1020
+ 1019,往,14.29,1,7
1021
+ 1020,提示,14.29,1,7
1022
+ 1021,曾经,14.29,4,28
1023
+ 1022,电视,14.29,1,7
1024
+ 1023,站在,14.29,1,7
1025
+ 1024,观众,14.29,4,28
1026
+ 1025,设施,14.29,1,7
1027
+ 1026,货币,14.29,1,7
1028
+ 1027,越,14.29,1,7
1029
+ 1028,过,14.00,7,50
1030
+ 1029,米,13.73,7,51
1031
+ 1030,甚至,13.64,3,22
1032
+ 1031,张,13.33,4,30
1033
+ 1032,它,13.18,17,129
1034
+ 1033,各,12.50,2,16
1035
+ 1034,套,12.50,1,8
1036
+ 1035,批,12.50,1,8
1037
+ 1036,日,12.50,5,40
1038
+ 1037,资金,12.50,1,8
1039
+ 1038,远,12.50,1,8
1040
+ 1039,雷,12.50,3,24
1041
+ 1040,之间,12.00,3,25
1042
+ 1041,大概,12.00,3,25
1043
+ 1042,内,11.94,8,67
1044
+ 1043,做出,11.76,2,17
1045
+ 1044,先生,11.76,2,17
1046
+ 1045,半,11.76,2,17
1047
+ 1046,距离,11.76,2,17
1048
+ 1047,里,11.54,6,52
1049
+ 1048,事业,11.11,1,9
1050
+ 1049,互信,11.11,1,9
1051
+ 1050,取消,11.11,1,9
1052
+ 1051,块,11.11,1,9
1053
+ 1052,基地,11.11,1,9
1054
+ 1053,实际上,11.11,2,18
1055
+ 1054,尤其,11.11,1,9
1056
+ 1055,机遇,11.11,1,9
1057
+ 1056,李,11.11,1,9
1058
+ 1057,自由,11.11,1,9
1059
+ 1058,近期,11.11,2,18
1060
+ 1059,进,11.11,1,9
1061
+ 1060,靠,11.11,1,9
1062
+ 1061,给,11.01,12,109
1063
+ 1062,个��,10.81,4,37
1064
+ 1063,更,10.81,4,37
1065
+ 1064,站,10.71,3,28
1066
+ 1065,想,10.53,4,38
1067
+ 1066,港,10.34,3,29
1068
+ 1067,实验,10.00,1,10
1069
+ 1068,度,10.00,1,10
1070
+ 1069,真的,10.00,1,10
1071
+ 1070,福,10.00,1,10
1072
+ 1071,尼,9.68,3,31
1073
+ 1072,已,9.52,2,21
1074
+ 1073,形式,9.09,1,11
1075
+ 1074,叫,8.70,2,23
1076
+ 1075,才,8.70,2,23
1077
+ 1076,明,8.70,2,23
1078
+ 1077,成,8.33,1,12
1079
+ 1078,应,8.00,2,25
1080
+ 1079,利,7.69,1,13
1081
+ 1080,只是,7.69,1,13
1082
+ 1081,像,7.32,3,41
1083
+ 1082,嘛,7.14,1,14
1084
+ 1083,因,7.14,1,14
1085
+ 1084,处,7.14,1,14
1086
+ 1085,密切,7.14,1,14
1087
+ 1086,东西,6.67,1,15
1088
+ 1087,发射,6.67,4,60
1089
+ 1088,行,6.67,1,15
1090
+ 1089,议题,6.67,1,15
1091
+ 1090,飞,6.25,1,16
1092
+ 1091,科,5.88,1,17
1093
+ 1092,英,5.88,1,17
1094
+ 1093,西,5.88,1,17
1095
+ 1094,没,5.56,1,18
1096
+ 1095,讲,5.26,1,19
1097
+ 1096,俄,4.84,3,62
1098
+ 1097,跟,4.65,2,43
1099
+ 1098,总,4.17,1,24
1100
+ 1099,其,4.00,1,25
1101
+ 1100,你,3.88,4,103
1102
+ 1101,组,3.33,1,30
1103
+ 1102,事,2.38,1,42
1104
+ 1103,却,2.33,1,43
1105
+ 1104,得,2.13,1,47
1106
+ 1105,C,0.00,0,9
1107
+ 1106,一切,0.00,0,7
1108
+ 1107,一块,0.00,0,8
1109
+ 1108,一致,0.00,0,15
1110
+ 1109,世,0.00,0,14
1111
+ 1110,世纪,0.00,0,4
1112
+ 1111,丽,0.00,0,3
1113
+ 1112,互,0.00,0,1
1114
+ 1113,代,0.00,0,5
1115
+ 1114,任,0.00,0,21
1116
+ 1115,伊,0.00,0,7
1117
+ 1116,低,0.00,0,5
1118
+ 1117,体,0.00,0,8
1119
+ 1118,使,0.00,0,15
1120
+ 1119,依,0.00,0,4
1121
+ 1120,停,0.00,0,8
1122
+ 1121,勇,0.00,0,5
1123
+ 1122,县,0.00,0,7
1124
+ 1123,只,0.00,0,23
1125
+ 1124,可,0.00,0,16
1126
+ 1125,吃,0.00,0,15
1127
+ 1126,合,0.00,0,7
1128
+ 1127,同意,0.00,0,9
1129
+ 1128,呀,0.00,0,5
1130
+ 1129,唉,0.00,0,15
1131
+ 1130,商,0.00,0,7
1132
+ 1131,团结,0.00,0,10
1133
+ 1132,城,0.00,0,11
1134
+ 1133,基,0.00,0,10
1135
+ 1134,夜,0.00,0,10
1136
+ 1135,她,0.00,0,16
1137
+ 1136,始终,0.00,0,9
1138
+ 1137,学,0.00,0,7
1139
+ 1138,尊重,0.00,0,7
1140
+ 1139,干,0.00,0,6
1141
+ 1140,年轻,0.00,0,9
1142
+ 1141,建,0.00,0,4
1143
+ 1142,当,0.00,0,14
1144
+ 1143,心,0.00,0,9
1145
+ 1144,志,0.00,0,2
1146
+ 1145,息,0.00,0,2
1147
+ 1146,意义,0.00,0,11
1148
+ 1147,感,0.00,0,5
1149
+ 1148,户,0.00,0,6
1150
+ 1149,打,0.00,0,24
1151
+ 1150,支,0.00,0,15
1152
+ 1151,施,0.00,0,4
1153
+ 1152,既,0.00,0,4
1154
+ 1153,期,0.00,0,4
1155
+ 1154,机,0.00,0,29
1156
+ 1155,梦想,0.00,0,7
1157
+ 1156,楼,0.00,0,7
1158
+ 1157,正,0.00,0,19
1159
+ 1158,江,0.00,0,8
1160
+ 1159,消,0.00,0,3
1161
+ 1160,深,0.00,0,7
1162
+ 1161,清,0.00,0,5
1163
+ 1162,热,0.00,0,11
1164
+ 1163,爱,0.00,0,13
1165
+ 1164,电,0.00,0,2
1166
+ 1165,界,0.00,0,9
1167
+ 1166,直接,0.00,0,17
1168
+ 1167,确定,0.00,0,11
1169
+ 1168,种,0.00,0,2
1170
+ 1169,策,0.00,0,3
1171
+ 1170,类,0.00,0,6
1172
+ 1171,经,0.00,0,19
1173
+ 1172,背景,0.00,0,4
1174
+ 1173,能,0.00,0,60
1175
+ 1174,节,0.00,0,9
1176
+ 1175,蒙,0.00,0,23
1177
+ 1176,行政,0.00,0,9
1178
+ 1177,试验,0.00,0,6
1179
+ 1178,过来,0.00,0,7
1180
+ 1179,运营,0.00,0,8
1181
+ 1180,连,0.00,0,9
1182
+ 1181,采,0.00,0,1
1183
+ 1182,重,0.00,0,9
1184
+ 1183,量,0.00,0,15
1185
+ 1184,钟,0.00,0,13
sha256.txt ADDED
@@ -0,0 +1 @@
 
 
1
+ 2ef301608cbdc841d2f1ba0255bef65f65c3b94cfcb9d4991f50167d80a15ef3 best_model.pth