Hugging Face's logo

MK-runner
/
CheXLoc

Model card Files
xet
 Community
MK-runner commited on
Commit
41d2e96
·
verified ·
1 Parent(s): 1a26332

Delete checkpoints/mimic-cxr/align/priorrg-v2_2025_12_06_15-best/log_2025_12_06_15.log

Browse files
Files changed (1) hide show
  1. checkpoints/mimic-cxr/align/priorrg-v2_2025_12_06_15-best/log_2025_12_06_15.log +0 -337
checkpoints/mimic-cxr/align/priorrg-v2_2025_12_06_15-best/log_2025_12_06_15.log DELETED
@@ -1,337 +0,0 @@
1
- task: align
2
- phase: finetune
3
- data_name: mimic_cxr
4
- ann_path: /home/server/kang_liu/MIMIC-CXR/PriorRG-V2-mimic-cxr-annotation.json
5
- view_position_path: /home/server/kang_liu/MIMIC-CXR/view-positions-dict.json
6
- images_dir: /home/server/kang_liu/MIMIC-CXR/files
7
- max_length: 100
8
- encoder_max_length: 300
9
- num_workers: 12
10
- epochs: 50
11
- batch_size: 48
12
- num_gpus: 1
13
- patience: 10
14
- save_best_model: True
15
- online_checkpoint: False
16
- ckpt_zoo_dir: /home/server/kang_liu/checkpoints
17
- text_encoder_num_blocks: 6
18
- temporal_fusion_num_blocks: 3
19
- perceiver_num_blocks: 3
20
- num_heads: 8
21
- num_latents: 128
22
- lr: 5e-05
23
- monitor_metric: val_epoch_loss
24
- hidden_size: 768
25
- resume: None
26
- load: None
27
- test_ckpt_path: None
28
- version: priorrg-v2
29
- chexbert_path: /home/server/kang_liu/checkpoints/chexbert.pth
30
- bert_path: /home/server/kang_liu/checkpoints/google-bert/bert-base-uncased
31
- radgraph_path: /home/server/kang_liu/checkpoints/radgraph
32
- cxr_bert_path: /home/server/kang_liu/checkpoints/microsoft/BiomedVLP-CXR-BERT-specialized
33
- rad_dino_path: /home/server/kang_liu/checkpoints/microsoft/rad-dino
34
- llama_path: /home/server/kang_liu/checkpoints/meta-llama/Llama-3.2-3B-Instruct
35
- distilgpt2_path: /home/server/kang_liu/checkpoints/distilbert/distilgpt2
36
- seed: 9233
37
- num_beams: 3
38
- save_period: 1
39
- exp_dir_trial: results/mimic_cxr/align/priorrg-v2_2025_12_06_15
40
- print_step: 500
41
- prior_knowledge_type: prior-image
42
- prior_knowledge_ratio: 0.1
43
- device: cuda
44
- monitor_mode: min
45
- checkpoint_dir: results/mimic_cxr/align/priorrg-v2_2025_12_06_15/checkpoint
46
- time: 2025_12_06_15
47
-
48
- No. of training & validation examples: 240197 & 2113.
49
- Epoch 0, validation step 0/45, loss = 3.97, instance_loss = 3.88, token_loss = 0.09, lr: 5e-05
50
- disease: 3.0 anatomy: 1.5 severity: 2.0 temporal: 1.5 uncertainty: 1.0 negation: 2.0 device: 1.2000000476837158 quality: 0.800000011920929 observation: 2.0
51
- The best model is saved on epoch 0!
52
- ###############################################################
53
- Epoch 0, Validation is over, current val loss:loss = 3.97, instance_loss = 3.87, token_loss = 0.09, lr: 5e-05
54
- best validation loss: loss = 3.96632719039917, instance_loss = 3.8748693466186523, token_loss = 0.09145787358283997, epoch = 0
55
-
56
- Epoch 0, training step 0/5004, loss = 3.97, instance_loss = 3.88, token_loss = 0.09, lr: 5e-05
57
- Epoch 0, training step 500/5004, loss = 2.95, instance_loss = 2.94, token_loss = 0.01, lr: 5e-05
58
- Epoch 0, training step 1000/5004, loss = 2.24, instance_loss = 2.23, token_loss = 0.01, lr: 5e-05
59
- Epoch 0, training step 1500/5004, loss = 1.79, instance_loss = 1.78, token_loss = 0.01, lr: 5e-05
60
- Epoch 0, training step 2000/5004, loss = 1.76, instance_loss = 1.75, token_loss = 0.01, lr: 5e-05
61
- Epoch 0, training step 2500/5004, loss = 1.61, instance_loss = 1.6, token_loss = 0.01, lr: 5e-05
62
- Epoch 0, training step 3000/5004, loss = 1.23, instance_loss = 1.22, token_loss = 0.01, lr: 5e-05
63
- Epoch 0, training step 3500/5004, loss = 1.0, instance_loss = 0.99, token_loss = 0.01, lr: 5e-05
64
- Epoch 0, training step 4000/5004, loss = 1.29, instance_loss = 1.28, token_loss = 0.01, lr: 5e-05
65
- Epoch 0, training step 4500/5004, loss = 1.04, instance_loss = 1.03, token_loss = 0.01, lr: 5e-05
66
- Epoch 0, training step 5000/5004, loss = 0.72, instance_loss = 0.71, token_loss = 0.01, lr: 5e-05
67
- Epoch 0, training step 5003/5004, loss = 0.86, instance_loss = 0.85, token_loss = 0.01, lr: 5e-05
68
- Epoch 0, validation step 0/45, loss = 1.18, instance_loss = 1.17, token_loss = 0.01, lr: 5e-05
69
- Epoch 0, validation step 44/45, loss = 0.02, instance_loss = -0.0, token_loss = 0.02, lr: 5e-05
70
- disease: 2.944469928741455 anatomy: 1.3407914638519287 severity: 1.937269926071167 temporal: 1.3645144701004028 uncertainty: 0.8509482741355896 negation: 1.8442435264587402 device: 1.1249140501022339 quality: 0.6540372967720032 observation: 1.8407591581344604
71
- The best model is saved on epoch 0!
72
- ###############################################################
73
- Epoch 0, Validation is over, current val loss:loss = 1.53, instance_loss = 1.51, token_loss = 0.01, lr: 5e-05
74
- best validation loss: loss = 1.525036334991455, instance_loss = 1.514955759048462, token_loss = 0.010080286301672459, epoch = 0
75
-
76
- Epoch 0, Training is over, loss = 1.75, instance_loss = 1.74, token_loss = 0.01, lr: 5e-05
77
- ###############################################################
78
- Epoch 1, training step 0/5004, loss = 0.95, instance_loss = 0.94, token_loss = 0.01, lr: 5e-05
79
- Epoch 1, training step 500/5004, loss = 0.8, instance_loss = 0.79, token_loss = 0.01, lr: 5e-05
80
- Epoch 1, training step 1000/5004, loss = 1.3, instance_loss = 1.29, token_loss = 0.01, lr: 5e-05
81
- Epoch 1, training step 1500/5004, loss = 0.83, instance_loss = 0.81, token_loss = 0.02, lr: 5e-05
82
- Epoch 1, training step 2000/5004, loss = 0.44, instance_loss = 0.43, token_loss = 0.02, lr: 5e-05
83
- Epoch 1, training step 2500/5004, loss = 0.8, instance_loss = 0.78, token_loss = 0.01, lr: 5e-05
84
- Epoch 1, training step 3000/5004, loss = 0.9, instance_loss = 0.89, token_loss = 0.01, lr: 5e-05
85
- Epoch 1, training step 3500/5004, loss = 0.49, instance_loss = 0.47, token_loss = 0.02, lr: 5e-05
86
- Epoch 1, training step 4000/5004, loss = 0.81, instance_loss = 0.79, token_loss = 0.02, lr: 5e-05
87
- Epoch 1, training step 4500/5004, loss = 0.39, instance_loss = 0.38, token_loss = 0.01, lr: 5e-05
88
- Epoch 1, training step 5000/5004, loss = 0.33, instance_loss = 0.32, token_loss = 0.01, lr: 5e-05
89
- Epoch 1, training step 5003/5004, loss = 0.51, instance_loss = 0.5, token_loss = 0.01, lr: 5e-05
90
- Epoch 1, validation step 0/45, loss = 0.97, instance_loss = 0.96, token_loss = 0.01, lr: 5e-05
91
- Epoch 1, validation step 44/45, loss = 0.02, instance_loss = -0.0, token_loss = 0.02, lr: 5e-05
92
- disease: 2.892754077911377 anatomy: 1.195708990097046 severity: 1.8763389587402344 temporal: 1.2444788217544556 uncertainty: 0.7154208421707153 negation: 1.6996718645095825 device: 1.057313084602356 quality: 0.5187632441520691 observation: 1.6974354982376099
93
- The best model is saved on epoch 1!
94
- ###############################################################
95
- Epoch 1, Validation is over, current val loss:loss = 1.32, instance_loss = 1.31, token_loss = 0.01, lr: 5e-05
96
- best validation loss: loss = 1.3237375020980835, instance_loss = 1.3099178075790405, token_loss = 0.013819640502333641, epoch = 1
97
-
98
- Epoch 1, Training is over, loss = 0.78, instance_loss = 0.77, token_loss = 0.01, lr: 5e-05
99
- ###############################################################
100
- Epoch 2, training step 0/5004, loss = 0.47, instance_loss = 0.46, token_loss = 0.01, lr: 5e-05
101
- Epoch 2, training step 500/5004, loss = 0.74, instance_loss = 0.73, token_loss = 0.01, lr: 5e-05
102
- Epoch 2, training step 1000/5004, loss = 0.52, instance_loss = 0.51, token_loss = 0.01, lr: 5e-05
103
- Epoch 2, training step 1500/5004, loss = 0.46, instance_loss = 0.44, token_loss = 0.01, lr: 5e-05
104
- Epoch 2, training step 2000/5004, loss = 0.51, instance_loss = 0.5, token_loss = 0.01, lr: 5e-05
105
- Epoch 2, training step 2500/5004, loss = 0.69, instance_loss = 0.68, token_loss = 0.01, lr: 5e-05
106
- Epoch 2, training step 3000/5004, loss = 0.6, instance_loss = 0.58, token_loss = 0.02, lr: 5e-05
107
- Epoch 2, training step 3500/5004, loss = 0.53, instance_loss = 0.52, token_loss = 0.01, lr: 5e-05
108
- Epoch 2, training step 4000/5004, loss = 0.51, instance_loss = 0.5, token_loss = 0.01, lr: 5e-05
109
- Epoch 2, training step 4500/5004, loss = 0.39, instance_loss = 0.38, token_loss = 0.01, lr: 5e-05
110
- Epoch 2, training step 5000/5004, loss = 0.61, instance_loss = 0.6, token_loss = 0.01, lr: 5e-05
111
- Epoch 2, training step 5003/5004, loss = 0.37, instance_loss = 0.35, token_loss = 0.02, lr: 5e-05
112
- Epoch 2, validation step 0/45, loss = 0.94, instance_loss = 0.93, token_loss = 0.01, lr: 5e-05
113
- Epoch 2, validation step 44/45, loss = 0.02, instance_loss = -0.0, token_loss = 0.02, lr: 5e-05
114
- disease: 2.8404619693756104 anatomy: 1.0621291399002075 severity: 1.8191426992416382 temporal: 1.1290091276168823 uncertainty: 0.5881428122520447 negation: 1.5674104690551758 device: 0.9880581498146057 quality: 0.3905099034309387 observation: 1.5647062063217163
115
- The best model is saved on epoch 2!
116
- ###############################################################
117
- Epoch 2, Validation is over, current val loss:loss = 1.24, instance_loss = 1.23, token_loss = 0.01, lr: 5e-05
118
- best validation loss: loss = 1.2440471649169922, instance_loss = 1.2295652627944946, token_loss = 0.014481655322015285, epoch = 2
119
-
120
- Epoch 2, Training is over, loss = 0.54, instance_loss = 0.53, token_loss = 0.01, lr: 5e-05
121
- ###############################################################
122
- Epoch 3, training step 0/5004, loss = 0.51, instance_loss = 0.5, token_loss = 0.01, lr: 5e-05
123
- Epoch 3, training step 500/5004, loss = 0.36, instance_loss = 0.34, token_loss = 0.01, lr: 5e-05
124
- Epoch 3, training step 1000/5004, loss = 0.47, instance_loss = 0.46, token_loss = 0.01, lr: 5e-05
125
- Epoch 3, training step 1500/5004, loss = 0.35, instance_loss = 0.34, token_loss = 0.01, lr: 5e-05
126
- Epoch 3, training step 2000/5004, loss = 0.41, instance_loss = 0.39, token_loss = 0.02, lr: 5e-05
127
- Epoch 3, training step 2500/5004, loss = 0.4, instance_loss = 0.39, token_loss = 0.01, lr: 5e-05
128
- Epoch 3, training step 3000/5004, loss = 0.42, instance_loss = 0.41, token_loss = 0.01, lr: 5e-05
129
- Epoch 3, training step 3500/5004, loss = 0.47, instance_loss = 0.46, token_loss = 0.01, lr: 5e-05
130
- Epoch 3, training step 4000/5004, loss = 0.47, instance_loss = 0.46, token_loss = 0.01, lr: 5e-05
131
- Epoch 3, training step 4500/5004, loss = 0.34, instance_loss = 0.32, token_loss = 0.02, lr: 5e-05
132
- Epoch 3, training step 5000/5004, loss = 0.26, instance_loss = 0.25, token_loss = 0.01, lr: 5e-05
133
- Epoch 3, training step 5003/5004, loss = 0.55, instance_loss = 0.54, token_loss = 0.01, lr: 5e-05
134
- Epoch 3, validation step 0/45, loss = 0.91, instance_loss = 0.89, token_loss = 0.01, lr: 5e-05
135
- Epoch 3, validation step 44/45, loss = 0.02, instance_loss = -0.0, token_loss = 0.02, lr: 5e-05
136
- disease: 2.789578914642334 anatomy: 0.9296764731407166 severity: 1.7594966888427734 temporal: 1.0148006677627563 uncertainty: 0.4610097110271454 negation: 1.4360558986663818 device: 0.9202081561088562 quality: 0.2621957063674927 observation: 1.4330401420593262
137
- The best model is saved on epoch 3!
138
- ###############################################################
139
- Epoch 3, Validation is over, current val loss:loss = 1.19, instance_loss = 1.17, token_loss = 0.01, lr: 5e-05
140
- best validation loss: loss = 1.188161015510559, instance_loss = 1.1741516590118408, token_loss = 0.014009666629135609, epoch = 3
141
-
142
- Epoch 3, Training is over, loss = 0.42, instance_loss = 0.4, token_loss = 0.01, lr: 5e-05
143
- ###############################################################
144
- Epoch 4, training step 0/5004, loss = 0.33, instance_loss = 0.32, token_loss = 0.01, lr: 5e-05
145
- Epoch 4, training step 500/5004, loss = 0.46, instance_loss = 0.44, token_loss = 0.02, lr: 5e-05
146
- Epoch 4, training step 1000/5004, loss = 0.39, instance_loss = 0.38, token_loss = 0.01, lr: 5e-05
147
- Epoch 4, training step 1500/5004, loss = 0.26, instance_loss = 0.25, token_loss = 0.01, lr: 5e-05
148
- Epoch 4, training step 2000/5004, loss = 0.23, instance_loss = 0.21, token_loss = 0.01, lr: 5e-05
149
- Epoch 4, training step 2500/5004, loss = 0.22, instance_loss = 0.21, token_loss = 0.01, lr: 5e-05
150
- Epoch 4, training step 3000/5004, loss = 0.21, instance_loss = 0.2, token_loss = 0.01, lr: 5e-05
151
- Epoch 4, training step 3500/5004, loss = 0.25, instance_loss = 0.24, token_loss = 0.01, lr: 5e-05
152
- Epoch 4, training step 4000/5004, loss = 0.31, instance_loss = 0.3, token_loss = 0.01, lr: 5e-05
153
- Epoch 4, training step 4500/5004, loss = 0.24, instance_loss = 0.23, token_loss = 0.01, lr: 5e-05
154
- Epoch 4, training step 5000/5004, loss = 0.48, instance_loss = 0.47, token_loss = 0.01, lr: 5e-05
155
- Epoch 4, training step 5003/5004, loss = 0.49, instance_loss = 0.47, token_loss = 0.01, lr: 5e-05
156
- Epoch 4, validation step 0/45, loss = 0.92, instance_loss = 0.9, token_loss = 0.01, lr: 5e-05
157
- Epoch 4, validation step 44/45, loss = 0.02, instance_loss = -0.0, token_loss = 0.02, lr: 5e-05
158
- disease: 2.7376394271850586 anatomy: 0.7984645366668701 severity: 1.695373773574829 temporal: 0.9001869559288025 uncertainty: 0.33462128043174744 negation: 1.3063011169433594 device: 0.8514688014984131 quality: 0.13475215435028076 observation: 1.3024662733078003
159
- The best model is saved on epoch 4!
160
- ###############################################################
161
- Epoch 4, Validation is over, current val loss:loss = 1.19, instance_loss = 1.17, token_loss = 0.01, lr: 5e-05
162
- best validation loss: loss = 1.1859242916107178, instance_loss = 1.1717349290847778, token_loss = 0.014189252629876137, epoch = 4
163
-
164
- Epoch 4, Training is over, loss = 0.34, instance_loss = 0.32, token_loss = 0.01, lr: 5e-05
165
- ###############################################################
166
- Epoch 5, training step 0/5004, loss = 0.44, instance_loss = 0.42, token_loss = 0.01, lr: 5e-05
167
- Epoch 5, training step 500/5004, loss = 0.58, instance_loss = 0.57, token_loss = 0.01, lr: 5e-05
168
- Epoch 5, training step 1000/5004, loss = 0.25, instance_loss = 0.24, token_loss = 0.01, lr: 5e-05
169
- Epoch 5, training step 1500/5004, loss = 0.21, instance_loss = 0.2, token_loss = 0.01, lr: 5e-05
170
- Epoch 5, training step 2000/5004, loss = 0.23, instance_loss = 0.22, token_loss = 0.01, lr: 5e-05
171
- Epoch 5, training step 2500/5004, loss = 0.24, instance_loss = 0.22, token_loss = 0.01, lr: 5e-05
172
- Epoch 5, training step 3000/5004, loss = 0.32, instance_loss = 0.31, token_loss = 0.01, lr: 5e-05
173
- Epoch 5, training step 3500/5004, loss = 0.73, instance_loss = 0.72, token_loss = 0.01, lr: 5e-05
174
- Epoch 5, training step 4000/5004, loss = 0.43, instance_loss = 0.42, token_loss = 0.02, lr: 5e-05
175
- Epoch 5, training step 4500/5004, loss = 0.31, instance_loss = 0.29, token_loss = 0.01, lr: 5e-05
176
- Epoch 5, training step 5000/5004, loss = 0.56, instance_loss = 0.55, token_loss = 0.01, lr: 5e-05
177
- Epoch 5, training step 5003/5004, loss = 0.39, instance_loss = 0.38, token_loss = 0.01, lr: 5e-05
178
- Epoch 5, validation step 0/45, loss = 0.99, instance_loss = 0.98, token_loss = 0.01, lr: 5e-05
179
- Epoch 5, validation step 44/45, loss = 0.02, instance_loss = -0.0, token_loss = 0.02, lr: 5e-05
180
- disease: 2.6872358322143555 anatomy: 0.6663867831230164 severity: 1.632017731666565 temporal: 0.7852123975753784 uncertainty: 0.2067050039768219 negation: 1.1759092807769775 device: 0.7852005958557129 quality: 0.00469777500256896 observation: 1.1713039875030518
181
- The best model is saved on epoch 5!
182
- ###############################################################
183
- Epoch 5, Validation is over, current val loss:loss = 1.18, instance_loss = 1.16, token_loss = 0.01, lr: 5e-05
184
- best validation loss: loss = 1.1759480237960815, instance_loss = 1.161513090133667, token_loss = 0.014434901997447014, epoch = 5
185
-
186
- Epoch 5, Training is over, loss = 0.29, instance_loss = 0.27, token_loss = 0.01, lr: 5e-05
187
- ###############################################################
188
- Epoch 6, training step 0/5004, loss = 0.3, instance_loss = 0.28, token_loss = 0.01, lr: 5e-05
189
- Epoch 6, training step 500/5004, loss = 0.26, instance_loss = 0.24, token_loss = 0.01, lr: 5e-05
190
- Epoch 6, training step 1000/5004, loss = 0.07, instance_loss = 0.06, token_loss = 0.01, lr: 5e-05
191
- Epoch 6, training step 1500/5004, loss = 0.19, instance_loss = 0.18, token_loss = 0.02, lr: 5e-05
192
- Epoch 6, training step 2000/5004, loss = 0.21, instance_loss = 0.19, token_loss = 0.01, lr: 5e-05
193
- Epoch 6, training step 2500/5004, loss = 0.18, instance_loss = 0.16, token_loss = 0.01, lr: 5e-05
194
- Epoch 6, training step 3000/5004, loss = 0.37, instance_loss = 0.35, token_loss = 0.01, lr: 5e-05
195
- Epoch 6, training step 3500/5004, loss = 0.16, instance_loss = 0.15, token_loss = 0.01, lr: 5e-05
196
- Epoch 6, training step 4000/5004, loss = 0.29, instance_loss = 0.28, token_loss = 0.01, lr: 5e-05
197
- Epoch 6, training step 4500/5004, loss = 0.25, instance_loss = 0.24, token_loss = 0.01, lr: 5e-05
198
- Epoch 6, training step 5000/5004, loss = 0.3, instance_loss = 0.29, token_loss = 0.02, lr: 5e-05
199
- Epoch 6, training step 5003/5004, loss = 0.31, instance_loss = 0.29, token_loss = 0.01, lr: 5e-05
200
- Epoch 6, validation step 0/45, loss = 0.91, instance_loss = 0.89, token_loss = 0.01, lr: 5e-05
201
- Epoch 6, validation step 44/45, loss = 0.03, instance_loss = -0.0, token_loss = 0.03, lr: 5e-05
202
- disease: 2.6368274688720703 anatomy: 0.5341441035270691 severity: 1.5682138204574585 temporal: 0.6700339317321777 uncertainty: 0.07771336287260056 negation: 1.045549750328064 device: 0.717060923576355 quality: -0.00016657088417559862 observation: 1.0402244329452515
203
- The best model is saved on epoch 6!
204
- ###############################################################
205
- Epoch 6, Validation is over, current val loss:loss = 1.15, instance_loss = 1.14, token_loss = 0.01, lr: 5e-05
206
- best validation loss: loss = 1.1538242101669312, instance_loss = 1.139663577079773, token_loss = 0.014160425402224064, epoch = 6
207
-
208
- Epoch 6, Training is over, loss = 0.25, instance_loss = 0.23, token_loss = 0.01, lr: 5e-05
209
- ###############################################################
210
- Epoch 7, training step 0/5004, loss = 0.19, instance_loss = 0.17, token_loss = 0.01, lr: 5e-05
211
- Epoch 7, training step 500/5004, loss = 0.21, instance_loss = 0.19, token_loss = 0.01, lr: 5e-05
212
- Epoch 7, training step 1000/5004, loss = 0.45, instance_loss = 0.44, token_loss = 0.01, lr: 5e-05
213
- Epoch 7, training step 1500/5004, loss = 0.23, instance_loss = 0.21, token_loss = 0.01, lr: 5e-05
214
- Epoch 7, training step 2000/5004, loss = 0.28, instance_loss = 0.27, token_loss = 0.01, lr: 5e-05
215
- Epoch 7, training step 2500/5004, loss = 0.26, instance_loss = 0.25, token_loss = 0.01, lr: 5e-05
216
- Epoch 7, training step 3000/5004, loss = 0.21, instance_loss = 0.2, token_loss = 0.01, lr: 5e-05
217
- Epoch 7, training step 3500/5004, loss = 0.23, instance_loss = 0.22, token_loss = 0.01, lr: 5e-05
218
- Epoch 7, training step 4000/5004, loss = 0.23, instance_loss = 0.22, token_loss = 0.01, lr: 5e-05
219
- Epoch 7, training step 4500/5004, loss = 0.32, instance_loss = 0.31, token_loss = 0.01, lr: 5e-05
220
- Epoch 7, training step 5000/5004, loss = 0.13, instance_loss = 0.11, token_loss = 0.01, lr: 5e-05
221
- Epoch 7, training step 5003/5004, loss = 0.14, instance_loss = 0.13, token_loss = 0.01, lr: 5e-05
222
- Epoch 7, validation step 0/45, loss = 0.99, instance_loss = 0.98, token_loss = 0.01, lr: 5e-05
223
- Epoch 7, validation step 44/45, loss = 0.02, instance_loss = -0.0, token_loss = 0.02, lr: 5e-05
224
- disease: 2.5825963020324707 anatomy: 0.40084049105644226 severity: 1.5035829544067383 temporal: 0.5538555979728699 uncertainty: -0.00016670355398673564 negation: 0.9143282175064087 device: 0.6500915884971619 quality: -6.409887282643467e-05 observation: 0.9085214138031006
225
- The best model is saved on epoch 7!
226
- ###############################################################
227
- Epoch 7, Validation is over, current val loss:loss = 1.15, instance_loss = 1.14, token_loss = 0.01, lr: 5e-05
228
- best validation loss: loss = 1.1501336097717285, instance_loss = 1.1362653970718384, token_loss = 0.013867913745343685, epoch = 7
229
-
230
- Epoch 7, Training is over, loss = 0.22, instance_loss = 0.21, token_loss = 0.01, lr: 5e-05
231
- ###############################################################
232
- Epoch 8, training step 0/5004, loss = 0.14, instance_loss = 0.12, token_loss = 0.01, lr: 5e-05
233
- Epoch 8, training step 500/5004, loss = 0.3, instance_loss = 0.29, token_loss = 0.01, lr: 5e-05
234
- Epoch 8, training step 1000/5004, loss = 0.23, instance_loss = 0.22, token_loss = 0.01, lr: 5e-05
235
- Epoch 8, training step 1500/5004, loss = 0.18, instance_loss = 0.17, token_loss = 0.01, lr: 5e-05
236
- Epoch 8, training step 2000/5004, loss = 0.23, instance_loss = 0.22, token_loss = 0.01, lr: 5e-05
237
- Epoch 8, training step 2500/5004, loss = 0.17, instance_loss = 0.16, token_loss = 0.01, lr: 5e-05
238
- Epoch 8, training step 3000/5004, loss = 0.28, instance_loss = 0.27, token_loss = 0.01, lr: 5e-05
239
- Epoch 8, training step 3500/5004, loss = 0.13, instance_loss = 0.12, token_loss = 0.01, lr: 5e-05
240
- Epoch 8, training step 4000/5004, loss = 0.1, instance_loss = 0.09, token_loss = 0.01, lr: 5e-05
241
- Epoch 8, training step 4500/5004, loss = 0.15, instance_loss = 0.13, token_loss = 0.01, lr: 5e-05
242
- Epoch 8, training step 5000/5004, loss = 0.14, instance_loss = 0.12, token_loss = 0.02, lr: 5e-05
243
- Epoch 8, training step 5003/5004, loss = 0.24, instance_loss = 0.22, token_loss = 0.01, lr: 5e-05
244
- Epoch 8, validation step 0/45, loss = 0.99, instance_loss = 0.97, token_loss = 0.01, lr: 5e-05
245
- Epoch 8, validation step 44/45, loss = 0.02, instance_loss = -0.0, token_loss = 0.02, lr: 5e-05
246
- ###############################################################
247
- Epoch 8, Validation is over, current val loss:loss = 1.16, instance_loss = 1.15, token_loss = 0.01, lr: 5e-05
248
- best validation loss: loss = 1.1501336097717285, instance_loss = 1.1362653970718384, token_loss = 0.013867913745343685, epoch = 7
249
-
250
- Epoch 8, Training is over, loss = 0.2, instance_loss = 0.19, token_loss = 0.01, lr: 5e-05
251
- ###############################################################
252
- Epoch 9, training step 0/5004, loss = 0.1, instance_loss = 0.09, token_loss = 0.01, lr: 5e-05
253
- Epoch 9, training step 500/5004, loss = 0.24, instance_loss = 0.23, token_loss = 0.01, lr: 5e-05
254
- Epoch 9, training step 1000/5004, loss = 0.39, instance_loss = 0.37, token_loss = 0.01, lr: 5e-05
255
- Epoch 9, training step 1500/5004, loss = 0.13, instance_loss = 0.12, token_loss = 0.01, lr: 5e-05
256
- Epoch 9, training step 2000/5004, loss = 0.23, instance_loss = 0.21, token_loss = 0.01, lr: 5e-05
257
- Epoch 9, training step 2500/5004, loss = 0.25, instance_loss = 0.23, token_loss = 0.01, lr: 5e-05
258
- Epoch 9, training step 3000/5004, loss = 0.18, instance_loss = 0.17, token_loss = 0.01, lr: 5e-05
259
- Epoch 9, training step 3500/5004, loss = 0.09, instance_loss = 0.08, token_loss = 0.01, lr: 5e-05
260
- Epoch 9, training step 4000/5004, loss = 0.21, instance_loss = 0.2, token_loss = 0.01, lr: 5e-05
261
- Epoch 9, training step 4500/5004, loss = 0.17, instance_loss = 0.15, token_loss = 0.01, lr: 5e-05
262
- Epoch 9, training step 5000/5004, loss = 0.25, instance_loss = 0.23, token_loss = 0.01, lr: 5e-05
263
- Epoch 9, training step 5003/5004, loss = 0.27, instance_loss = 0.26, token_loss = 0.01, lr: 5e-05
264
- Epoch 9, validation step 0/45, loss = 0.9, instance_loss = 0.89, token_loss = 0.01, lr: 5e-05
265
- Epoch 9, validation step 44/45, loss = 0.02, instance_loss = -0.0, token_loss = 0.02, lr: 5e-05
266
- ###############################################################
267
- Epoch 9, Validation is over, current val loss:loss = 1.15, instance_loss = 1.14, token_loss = 0.01, lr: 5e-05
268
- best validation loss: loss = 1.1501336097717285, instance_loss = 1.1362653970718384, token_loss = 0.013867913745343685, epoch = 7
269
-
270
- Epoch 9, Training is over, loss = 0.18, instance_loss = 0.17, token_loss = 0.01, lr: 5e-05
271
- ###############################################################
272
- Epoch 10, training step 0/5004, loss = 0.17, instance_loss = 0.16, token_loss = 0.01, lr: 5e-05
273
- Epoch 10, training step 500/5004, loss = 0.23, instance_loss = 0.22, token_loss = 0.01, lr: 5e-05
274
- Epoch 10, training step 1000/5004, loss = 0.27, instance_loss = 0.26, token_loss = 0.01, lr: 5e-05
275
- Epoch 10, training step 1500/5004, loss = 0.11, instance_loss = 0.09, token_loss = 0.01, lr: 5e-05
276
- Epoch 10, training step 2000/5004, loss = 0.07, instance_loss = 0.05, token_loss = 0.01, lr: 5e-05
277
- Epoch 10, training step 2500/5004, loss = 0.25, instance_loss = 0.24, token_loss = 0.01, lr: 5e-05
278
- Epoch 10, training step 3000/5004, loss = 0.21, instance_loss = 0.19, token_loss = 0.01, lr: 5e-05
279
- Epoch 10, training step 3500/5004, loss = 0.14, instance_loss = 0.13, token_loss = 0.01, lr: 5e-05
280
- Epoch 10, training step 4000/5004, loss = 0.33, instance_loss = 0.32, token_loss = 0.01, lr: 5e-05
281
- Epoch 10, training step 4500/5004, loss = 0.27, instance_loss = 0.25, token_loss = 0.01, lr: 5e-05
282
- Epoch 10, training step 5000/5004, loss = 0.23, instance_loss = 0.22, token_loss = 0.01, lr: 5e-05
283
- Epoch 10, training step 5003/5004, loss = 0.2, instance_loss = 0.19, token_loss = 0.01, lr: 5e-05
284
- Epoch 10, validation step 0/45, loss = 0.98, instance_loss = 0.97, token_loss = 0.01, lr: 5e-05
285
- Epoch 10, validation step 44/45, loss = 0.02, instance_loss = -0.0, token_loss = 0.02, lr: 5e-05
286
- ###############################################################
287
- Epoch 10, Validation is over, current val loss:loss = 1.16, instance_loss = 1.15, token_loss = 0.01, lr: 5e-05
288
- best validation loss: loss = 1.1501336097717285, instance_loss = 1.1362653970718384, token_loss = 0.013867913745343685, epoch = 7
289
-
290
- Epoch 10, Training is over, loss = 0.17, instance_loss = 0.16, token_loss = 0.01, lr: 5e-05
291
- ###############################################################
292
- Epoch 11, training step 0/5004, loss = 0.11, instance_loss = 0.1, token_loss = 0.01, lr: 5e-05
293
- Epoch 11, training step 500/5004, loss = 0.09, instance_loss = 0.08, token_loss = 0.01, lr: 5e-05
294
- Epoch 11, training step 1000/5004, loss = 0.11, instance_loss = 0.1, token_loss = 0.01, lr: 5e-05
295
- Epoch 11, training step 1500/5004, loss = 0.1, instance_loss = 0.09, token_loss = 0.01, lr: 5e-05
296
- Epoch 11, training step 2000/5004, loss = 0.05, instance_loss = 0.04, token_loss = 0.01, lr: 5e-05
297
- Epoch 11, training step 2500/5004, loss = 0.4, instance_loss = 0.39, token_loss = 0.01, lr: 5e-05
298
- Epoch 11, training step 3000/5004, loss = 0.21, instance_loss = 0.2, token_loss = 0.01, lr: 5e-05
299
- Epoch 11, training step 3500/5004, loss = 0.27, instance_loss = 0.26, token_loss = 0.01, lr: 5e-05
300
- Epoch 11, training step 4000/5004, loss = 0.11, instance_loss = 0.1, token_loss = 0.01, lr: 5e-05
301
- Epoch 11, training step 4500/5004, loss = 0.09, instance_loss = 0.08, token_loss = 0.01, lr: 5e-05
302
- Epoch 11, training step 5000/5004, loss = 0.12, instance_loss = 0.11, token_loss = 0.01, lr: 5e-05
303
- Epoch 11, training step 5003/5004, loss = 0.12, instance_loss = 0.11, token_loss = 0.01, lr: 5e-05
304
- Epoch 11, validation step 0/45, loss = 0.9, instance_loss = 0.89, token_loss = 0.01, lr: 5e-05
305
- Epoch 11, validation step 44/45, loss = 0.02, instance_loss = -0.0, token_loss = 0.02, lr: 5e-05
306
- disease: 2.3677098751068115 anatomy: -0.0014437580248340964 severity: 1.2428560256958008 temporal: 0.07819056510925293 uncertainty: -0.0015763232950121164 negation: 0.3747429847717285 device: 0.36465129256248474 quality: -0.03154028579592705 observation: 0.3640725612640381
307
- The best model is saved on epoch 11!
308
- ###############################################################
309
- Epoch 11, Validation is over, current val loss:loss = 1.15, instance_loss = 1.13, token_loss = 0.01, lr: 5e-05
310
- best validation loss: loss = 1.1451746225357056, instance_loss = 1.1345816850662231, token_loss = 0.01059301383793354, epoch = 11
311
-
312
- Epoch 11, Training is over, loss = 0.16, instance_loss = 0.15, token_loss = 0.01, lr: 5e-05
313
- ###############################################################
314
- Epoch 12, training step 0/5004, loss = 0.14, instance_loss = 0.13, token_loss = 0.01, lr: 5e-05
315
- Epoch 12, training step 500/5004, loss = 0.23, instance_loss = 0.22, token_loss = 0.01, lr: 5e-05
316
- Epoch 12, training step 1000/5004, loss = 0.09, instance_loss = 0.08, token_loss = 0.01, lr: 5e-05
317
- Epoch 12, training step 1500/5004, loss = 0.15, instance_loss = 0.14, token_loss = 0.01, lr: 5e-05
318
- Epoch 12, training step 2000/5004, loss = 0.19, instance_loss = 0.18, token_loss = 0.01, lr: 5e-05
319
- Epoch 12, training step 2500/5004, loss = 0.21, instance_loss = 0.2, token_loss = 0.01, lr: 5e-05
320
- Epoch 12, training step 3000/5004, loss = 0.15, instance_loss = 0.14, token_loss = 0.01, lr: 5e-05
321
- Epoch 12, training step 3500/5004, loss = 0.07, instance_loss = 0.06, token_loss = 0.01, lr: 5e-05
322
- Epoch 12, training step 4000/5004, loss = 0.3, instance_loss = 0.29, token_loss = 0.01, lr: 5e-05
323
- Epoch 12, training step 4500/5004, loss = 0.11, instance_loss = 0.1, token_loss = 0.01, lr: 5e-05
324
- Epoch 12, training step 5000/5004, loss = 0.1, instance_loss = 0.09, token_loss = 0.01, lr: 5e-05
325
- Epoch 12, training step 5003/5004, loss = 0.18, instance_loss = 0.17, token_loss = 0.01, lr: 5e-05
326
- Epoch 12, validation step 0/45, loss = 1.01, instance_loss = 1.0, token_loss = 0.01, lr: 5e-05
327
- Epoch 12, validation step 44/45, loss = 0.02, instance_loss = -0.0, token_loss = 0.02, lr: 5e-05
328
- ###############################################################
329
- Epoch 12, Validation is over, current val loss:loss = 1.17, instance_loss = 1.16, token_loss = 0.01, lr: 5e-05
330
- best validation loss: loss = 1.1451746225357056, instance_loss = 1.1345816850662231, token_loss = 0.01059301383793354, epoch = 11
331
-
332
- Epoch 12, Training is over, loss = 0.15, instance_loss = 0.14, token_loss = 0.01, lr: 5e-05
333
- ###############################################################
334
- Epoch 13, training step 0/5004, loss = 0.09, instance_loss = 0.08, token_loss = 0.01, lr: 5e-05
335
- Epoch 13, training step 500/5004, loss = 0.1, instance_loss = 0.09, token_loss = 0.01, lr: 5e-05
336
- Epoch 13, training step 1000/5004, loss = 0.05, instance_loss = 0.04, token_loss = 0.01, lr: 5e-05
337
- Epoch 13, training step 1500/5004, loss = 0.18, instance_loss = 0.17, token_loss = 0.01, lr: 5e-05