Delete checkpoints/chexpert-plus/report generation/finetune-v1231-gpt2-load-mimic-rg-new-lr-patience5-2026_02_19_20-best/log_2026_02_19_20.log

checkpoints/chexpert-plus/report generation/finetune-v1231-gpt2-load-mimic-rg-new-lr-patience5-2026_02_19_20-best/log_2026_02_19_20.log

@@ -1,718 +0,0 @@
-task:	train-language-model
-phase:	finetune
-data_name:	chexpert-plus
-ann_path:	/home/server/kang_liu/dataset/CheXpert-Plus/chexpert-plus-annotation-previous-image.json
-view_position_path:	/home/server/kang_liu/MIMIC-CXR/view-positions-dict.json
-images_dir:	/home/server/kang_liu/dataset/CheXpert-Plus/PNG
-english_report:	True
-max_length:	100
-encoder_max_length:	300
-num_workers:	10
-epochs:	50
-batch_size:	32
-num_gpus:	1
-patience:	5
-save_best_model:	True
-online_checkpoint:	False
-ckpt_zoo_dir:	/home/server/kang_liu/checkpoints
-text_encoder_num_blocks:	6
-temporal_fusion_num_blocks:	3
-perceiver_num_blocks:	3
-num_heads:	8
-num_latents:	128
-lr:	5e-05
-monitor_metric:	RCB
-hidden_size:	768
-resume:	None
-load:	/home/server/kang_liu/PriorRG-v2/script/results/mimic_cxr/train-language-model/v1207-gpt2_2025_12_07_15/checkpoint/epoch=14-step=56295_temp.ckpt
-test_ckpt_path:	None
-version:	v1231-gpt2-load-mimic-rg-new-lr-patience5
-chexbert_path:	/home/server/kang_liu/checkpoints/chexbert.pth
-bert_path:	/home/server/kang_liu/checkpoints/google-bert/bert-base-uncased
-radgraph_path:	/home/server/kang_liu/checkpoints/radgraph
-cxr_bert_path:	/home/server/kang_liu/checkpoints/microsoft/BiomedVLP-CXR-BERT-specialized
-rad_dino_path:	/home/server/kang_liu/checkpoints/microsoft/rad-dino
-medsiglip_path:	/home/server/kang_liu/checkpoints/google/medsiglip-448
-llama_path:	/home/server/kang_liu/checkpoints/meta-llama/Llama-3.2-3B-Instruct
-distilgpt2_path:	/home/server/kang_liu/checkpoints/distilbert/distilgpt2
-seed:	9233
-num_beams:	10
-save_period:	1
-exp_dir_trial:	results/chexpert-plus/train-language-model/finetune-v1231-gpt2-load-mimic-rg-new-lr-patience5-2026_02_19_20
-print_step:	500
-prior_knowledge_type:	prior-image
-prior_knowledge_ratio:	1
-device:	cuda
-monitor_mode:	max
-checkpoint_dir:	results/chexpert-plus/train-language-model/finetune-v1231-gpt2-load-mimic-rg-new-lr-patience5-2026_02_19_20/checkpoint
-time:	2026_02_19_20
-
-No. of training & validation examples: 59367 & 74.
-Epoch 0, validation step 0/3
-The best model is saved on epoch 0!
-###############################################################
-Epoch 0, Validation is over, current metrics:
-best validation epoch: 0, best val_metrics: RCB = 0.7128899192849105
-F1-Radgraph-partial: 0.16773394889465398
-F1-Radgraph-simple: 0.195410790298365
-F1-Radgraph-complete: 0.11974284925237204
-chexbert_5_micro_f1: 0.47058823529411764
-chexbert_5_macro_f1: 0.44688642674123613
-chexbert_all_micro_p: 0.591304347826087
-chexbert_all_micro_r: 0.40476190476190477
-chexbert_all_micro_f1: 0.48056537102473496
-chexbert_all_macro_p: 0.3918007762545578
-chexbert_all_macro_r: 0.3066650250697324
-chexbert_all_macro_f1: 0.3085027926873921
-chexbert_accuracy: 0.296875
-chen_bleu_1: 0.31230658442289705
-chen_bleu_2: 0.1715793117816511
-chen_bleu_3: 0.1039014558476197
-chen_bleu_4: 0.06459059936552158
-chen_meteor: 0.12491228400522707
-chen_rouge: 0.216659458199761
-chen_num_examples: 55
-RB: 0.23232454826017557
-RC: 0.6482993199193889
-RCB: 0.7128899192849105 
-
-Epoch 0, training step 0/1855, 6.356, lr: 5e-05
-Epoch 0, training step 500/1855, 2.213, lr: 5e-05
-Epoch 0, training step 1000/1855, 1.929, lr: 5e-05
-Epoch 0, training step 1500/1855, 1.7, lr: 5e-05
-Epoch 0, training step 1854/1855, 1.851, lr: 5e-05
-Epoch 0, validation step 0/3
-Epoch 0, validation step 2/3
-The best model is saved on epoch 0!
-###############################################################
-Epoch 0, Validation is over, current metrics:
-best validation epoch: 0, best val_metrics: RCB = 0.7581469642157972
-F1-Radgraph-partial: 0.2054309579005303
-F1-Radgraph-simple: 0.22580553030912281
-F1-Radgraph-complete: 0.15449605590579343
-chexbert_5_micro_f1: 0.3624161073825503
-chexbert_5_macro_f1: 0.2929185709371468
-chexbert_all_micro_p: 0.4934210526315789
-chexbert_all_micro_r: 0.4166666666666667
-chexbert_all_micro_f1: 0.45180722891566266
-chexbert_all_macro_p: 0.3059278440857388
-chexbert_all_macro_r: 0.29423477524981284
-chexbert_all_macro_f1: 0.28042238839740674
-chexbert_accuracy: 0.24324324324324326
-chen_bleu_1: 0.33518523086873153
-chen_bleu_2: 0.20813963673881455
-chen_bleu_3: 0.14227959672792104
-chen_bleu_4: 0.10090877739960431
-chen_meteor: 0.138550494374325
-chen_rouge: 0.24058342958834694
-chen_num_examples: 62
-RB: 0.3063397353001346
-RC: 0.657238186816193
-RCB: 0.7581469642157972 
-
-Epoch 0, Training is over, training epoch loss = 2.1213018894195557, lr: 5e-05. Epoch training time: 3411.0294868946075 seconds
-###############################################################
-Epoch 1, training step 0/1855, 1.841, lr: 5e-05
-Epoch 1, training step 500/1855, 1.678, lr: 5e-05
-Epoch 1, training step 1000/1855, 1.451, lr: 5e-05
-Epoch 1, training step 1500/1855, 1.654, lr: 5e-05
-Epoch 1, training step 1854/1855, 1.659, lr: 5e-05
-Epoch 1, validation step 0/3
-Epoch 1, validation step 2/3
-The best model is saved on epoch 1!
-###############################################################
-Epoch 1, Validation is over, current metrics:
-best validation epoch: 1, best val_metrics: RCB = 0.7788732702417003
-F1-Radgraph-partial: 0.22222291034779193
-F1-Radgraph-simple: 0.24409066849359246
-F1-Radgraph-complete: 0.16146498670148637
-chexbert_5_micro_f1: 0.3356643356643357
-chexbert_5_macro_f1: 0.26119047619047614
-chexbert_all_micro_p: 0.4691358024691358
-chexbert_all_micro_r: 0.4222222222222222
-chexbert_all_micro_f1: 0.4444444444444444
-chexbert_all_macro_p: 0.30525269579635544
-chexbert_all_macro_r: 0.29563877225531365
-chexbert_all_macro_f1: 0.2848051650730222
-chexbert_accuracy: 0.25675675675675674
-chen_bleu_1: 0.3396723156224136
-chen_bleu_2: 0.21620845938340802
-chen_bleu_3: 0.1519208657945907
-chen_bleu_4: 0.11220591544946396
-chen_meteor: 0.14666781287194652
-chen_rouge: 0.2688448435334641
-chen_num_examples: 62
-RB: 0.33442882579725586
-RC: 0.6666673547922364
-RCB: 0.7788732702417003 
-
-Epoch 1, Training is over, training epoch loss = 1.632585048675537, lr: 5e-05. Epoch training time: 3272.9845497608185 seconds
-###############################################################
-Epoch 2, training step 0/1855, 1.49, lr: 5e-05
-Epoch 2, training step 500/1855, 1.478, lr: 5e-05
-Epoch 2, training step 1000/1855, 1.441, lr: 5e-05
-Epoch 2, training step 1500/1855, 1.486, lr: 5e-05
-Epoch 2, training step 1854/1855, 1.546, lr: 5e-05
-Epoch 2, validation step 0/3
-Epoch 2, validation step 2/3
-The best model is saved on epoch 2!
-###############################################################
-Epoch 2, Validation is over, current metrics:
-best validation epoch: 2, best val_metrics: RCB = 0.8396963896010282
-F1-Radgraph-partial: 0.22296392930797204
-F1-Radgraph-simple: 0.2442942276488967
-F1-Radgraph-complete: 0.1702995502821575
-chexbert_5_micro_f1: 0.4383561643835616
-chexbert_5_macro_f1: 0.36282724947686584
-chexbert_all_micro_p: 0.525
-chexbert_all_micro_r: 0.4666666666666667
-chexbert_all_micro_f1: 0.49411764705882355
-chexbert_all_macro_p: 0.34258893280632413
-chexbert_all_macro_r: 0.31049018650522414
-chexbert_all_macro_f1: 0.30954162956355136
-chexbert_accuracy: 0.3108108108108108
-chen_bleu_1: 0.34954910573236325
-chen_bleu_2: 0.22866069714001924
-chen_bleu_3: 0.16358887079618845
-chen_bleu_4: 0.12261481323423262
-chen_meteor: 0.15030966114946268
-chen_rouge: 0.2736999076715437
-chen_num_examples: 62
-RB: 0.34557874254220466
-RC: 0.7170815763667956
-RCB: 0.8396963896010282 
-
-Epoch 2, Training is over, training epoch loss = 1.4905798435211182, lr: 5e-05. Epoch training time: 3278.290989637375 seconds
-###############################################################
-Epoch 3, training step 0/1855, 1.334, lr: 5e-05
-Epoch 3, training step 500/1855, 1.43, lr: 5e-05
-Epoch 3, training step 1000/1855, 1.391, lr: 5e-05
-Epoch 3, training step 1500/1855, 1.415, lr: 5e-05
-Epoch 3, training step 1854/1855, 1.339, lr: 5e-05
-Epoch 3, validation step 0/3
-Epoch 3, validation step 2/3
-###############################################################
-Epoch 3, Validation is over, current metrics:
-best validation epoch: 2, best val_metrics: RCB = 0.8396963896010282
-F1-Radgraph-partial: 0.22944968456745854
-F1-Radgraph-simple: 0.250578446482917
-F1-Radgraph-complete: 0.16737010117674994
-chexbert_5_micro_f1: 0.36
-chexbert_5_macro_f1: 0.3104300851580692
-chexbert_all_micro_p: 0.4785276073619632
-chexbert_all_micro_r: 0.43333333333333335
-chexbert_all_micro_f1: 0.45481049562682213
-chexbert_all_macro_p: 0.3411169518312375
-chexbert_all_macro_r: 0.33256968178020807
-chexbert_all_macro_f1: 0.312538067970267
-chexbert_accuracy: 0.2972972972972973
-chen_bleu_1: 0.366955215733812
-chen_bleu_2: 0.24056494090065836
-chen_bleu_3: 0.17166744789627206
-chen_bleu_4: 0.1273661835505716
-chen_meteor: 0.15847455872775007
-chen_rouge: 0.2718180280620476
-chen_num_examples: 62
-RB: 0.3568158681180301
-RC: 0.6842601801942807
-RCB: 0.8116263637448522 
-
-Epoch 3, Training is over, training epoch loss = 1.3825280666351318, lr: 5e-05. Epoch training time: 3431.1623187065125 seconds
-###############################################################
-Epoch 4, training step 0/1855, 1.316, lr: 5e-05
-Epoch 4, training step 500/1855, 1.173, lr: 5e-05
-Epoch 4, training step 1000/1855, 1.254, lr: 5e-05
-Epoch 4, training step 1500/1855, 1.568, lr: 5e-05
-Epoch 4, training step 1854/1855, 1.265, lr: 5e-05
-Epoch 4, validation step 0/3
-Epoch 4, validation step 2/3
-###############################################################
-Epoch 4, Validation is over, current metrics:
-best validation epoch: 2, best val_metrics: RCB = 0.8396963896010282
-F1-Radgraph-partial: 0.22571158110052036
-F1-Radgraph-simple: 0.2428537039169649
-F1-Radgraph-complete: 0.1708963462658373
-chexbert_5_micro_f1: 0.40816326530612246
-chexbert_5_macro_f1: 0.3642238713667285
-chexbert_all_micro_p: 0.49079754601226994
-chexbert_all_micro_r: 0.4444444444444444
-chexbert_all_micro_f1: 0.46647230320699706
-chexbert_all_macro_p: 0.3213445681302825
-chexbert_all_macro_r: 0.3084252839891938
-chexbert_all_macro_f1: 0.29984543039626493
-chexbert_accuracy: 0.2972972972972973
-chen_bleu_1: 0.35054308423207614
-chen_bleu_2: 0.2276837717277895
-chen_bleu_3: 0.16244291046065298
-chen_bleu_4: 0.12382016436943244
-chen_meteor: 0.15011542268010536
-chen_rouge: 0.2688574375927487
-chen_num_examples: 62
-RB: 0.3495317454699528
-RC: 0.6921838843075174
-RCB: 0.8160040486769499 
-
-Epoch 4, Training is over, training epoch loss = 1.2729856967926025, lr: 5e-05. Epoch training time: 3313.8814165592194 seconds
-###############################################################
-Epoch 5, training step 0/1855, 1.068, lr: 5e-05
-Epoch 5, training step 500/1855, 1.295, lr: 5e-05
-Epoch 5, training step 1000/1855, 1.153, lr: 5e-05
-Epoch 5, training step 1500/1855, 1.106, lr: 5e-05
-Epoch 5, training step 1854/1855, 1.166, lr: 5e-05
-Epoch 5, validation step 0/3
-Epoch 5, validation step 2/3
-###############################################################
-Epoch 5, Validation is over, current metrics:
-best validation epoch: 2, best val_metrics: RCB = 0.8396963896010282
-F1-Radgraph-partial: 0.2104520181119516
-F1-Radgraph-simple: 0.23583850351928914
-F1-Radgraph-complete: 0.16197533557860835
-chexbert_5_micro_f1: 0.4025974025974026
-chexbert_5_macro_f1: 0.3745169082125604
-chexbert_all_micro_p: 0.5151515151515151
-chexbert_all_micro_r: 0.4722222222222222
-chexbert_all_micro_f1: 0.4927536231884058
-chexbert_all_macro_p: 0.37518949584723976
-chexbert_all_macro_r: 0.30578177044342464
-chexbert_all_macro_f1: 0.3185596609509653
-chexbert_accuracy: 0.25675675675675674
-chen_bleu_1: 0.348775169262384
-chen_bleu_2: 0.22759996928856854
-chen_bleu_3: 0.16008565100408526
-chen_bleu_4: 0.11783655790035427
-chen_meteor: 0.1556787550671602
-chen_rouge: 0.2626379098597724
-chen_num_examples: 62
-RB: 0.3282885760123059
-RC: 0.7032056413003575
-RCB: 0.8210421992007118 
-
-Epoch 5, Training is over, training epoch loss = 1.1555395126342773, lr: 5e-05. Epoch training time: 3336.533994913101 seconds
-###############################################################
-Epoch 6, training step 0/1855, 1.211, lr: 5e-05
-Epoch 6, training step 500/1855, 1.021, lr: 5e-05
-Epoch 6, training step 1000/1855, 1.122, lr: 5e-05
-Epoch 6, training step 1500/1855, 0.966, lr: 5e-05
-Epoch 6, training step 1854/1855, 0.972, lr: 5e-05
-Epoch 6, validation step 0/3
-Epoch 6, validation step 2/3
-The best model is saved on epoch 6!
-###############################################################
-Epoch 6, Validation is over, current metrics:
-best validation epoch: 6, best val_metrics: RCB = 0.8454387528294205
-F1-Radgraph-partial: 0.21377494003215572
-F1-Radgraph-simple: 0.23321029232454
-F1-Radgraph-complete: 0.1618291767596106
-chexbert_5_micro_f1: 0.43312101910828027
-chexbert_5_macro_f1: 0.3985867074102368
-chexbert_all_micro_p: 0.5141242937853108
-chexbert_all_micro_r: 0.5055555555555555
-chexbert_all_micro_f1: 0.5098039215686274
-chexbert_all_macro_p: 0.3394181279549943
-chexbert_all_macro_r: 0.3291657173236121
-chexbert_all_macro_f1: 0.32207675057696417
-chexbert_accuracy: 0.28378378378378377
-chen_bleu_1: 0.351007708408334
-chen_bleu_2: 0.23260142904459505
-chen_bleu_3: 0.16488975891178295
-chen_bleu_4: 0.12185989122863732
-chen_meteor: 0.1556922512226225
-chen_rouge: 0.2690694086877414
-chen_num_examples: 62
-RB: 0.335634831260793
-RC: 0.7235788616007831
-RCB: 0.8454387528294205 
-
-Epoch 6, Training is over, training epoch loss = 1.0451730489730835, lr: 5e-05. Epoch training time: 3565.302356004715 seconds
-###############################################################
-Epoch 7, training step 0/1855, 0.937, lr: 5e-05
-Epoch 7, training step 500/1855, 0.984, lr: 5e-05
-Epoch 7, training step 1000/1855, 0.946, lr: 5e-05
-Epoch 7, training step 1500/1855, 1.003, lr: 5e-05
-Epoch 7, training step 1854/1855, 0.954, lr: 5e-05
-Epoch 7, validation step 0/3
-Epoch 7, validation step 2/3
-The best model is saved on epoch 7!
-###############################################################
-Epoch 7, Validation is over, current metrics:
-best validation epoch: 7, best val_metrics: RCB = 0.8633592256171855
-F1-Radgraph-partial: 0.20452139989293763
-F1-Radgraph-simple: 0.2225077461545921
-F1-Radgraph-complete: 0.15282023981687623
-chexbert_5_micro_f1: 0.46060606060606063
-chexbert_5_macro_f1: 0.43863187722572033
-chexbert_all_micro_p: 0.518324607329843
-chexbert_all_micro_r: 0.55
-chexbert_all_micro_f1: 0.5336927223719676
-chexbert_all_macro_p: 0.36302998996847924
-chexbert_all_macro_r: 0.39041299135284097
-chexbert_all_macro_f1: 0.362561033487406
-chexbert_accuracy: 0.2972972972972973
-chen_bleu_1: 0.3660310705182532
-chen_bleu_2: 0.23923196199594268
-chen_bleu_3: 0.16965656704017723
-chen_bleu_4: 0.1251451033522802
-chen_meteor: 0.15896443100790422
-chen_rouge: 0.2745260003662961
-chen_num_examples: 62
-RB: 0.32966650324521785
-RC: 0.7382141222649052
-RCB: 0.8633592256171855 
-
-Epoch 7, Training is over, training epoch loss = 0.9423288702964783, lr: 5e-05. Epoch training time: 3317.7727003097534 seconds
-###############################################################
-Epoch 8, training step 0/1855, 0.904, lr: 5e-05
-Epoch 8, training step 500/1855, 0.773, lr: 5e-05
-Epoch 8, training step 1000/1855, 1.004, lr: 5e-05
-Epoch 8, training step 1500/1855, 0.951, lr: 5e-05
-Epoch 8, training step 1854/1855, 0.859, lr: 5e-05
-Epoch 8, validation step 0/3
-Epoch 8, validation step 2/3
-###############################################################
-Epoch 8, Validation is over, current metrics:
-best validation epoch: 7, best val_metrics: RCB = 0.8633592256171855
-F1-Radgraph-partial: 0.2095601755353686
-F1-Radgraph-simple: 0.23786859077395142
-F1-Radgraph-complete: 0.15973690788597364
-chexbert_5_micro_f1: 0.4276729559748428
-chexbert_5_macro_f1: 0.3902937642068077
-chexbert_all_micro_p: 0.5251396648044693
-chexbert_all_micro_r: 0.5222222222222223
-chexbert_all_micro_f1: 0.5236768802228412
-chexbert_all_macro_p: 0.4142318510982106
-chexbert_all_macro_r: 0.3639725938222179
-chexbert_all_macro_f1: 0.36840241974138727
-chexbert_accuracy: 0.28378378378378377
-chen_bleu_1: 0.3446166507338335
-chen_bleu_2: 0.2211358837799951
-chen_bleu_3: 0.15569240627806558
-chen_bleu_4: 0.11316636258764776
-chen_meteor: 0.1536064844415331
-chen_rouge: 0.2641348602085754
-chen_num_examples: 62
-RB: 0.3227265381230163
-RC: 0.7332370557582097
-RCB: 0.8464034183458575 
-
-Epoch 8, Training is over, training epoch loss = 0.8553245067596436, lr: 5e-05. Epoch training time: 3165.634534597397 seconds
-###############################################################
-Epoch 9, training step 0/1855, 0.86, lr: 5e-05
-Epoch 9, training step 500/1855, 0.744, lr: 5e-05
-Epoch 9, training step 1000/1855, 0.934, lr: 5e-05
-Epoch 9, training step 1500/1855, 0.823, lr: 5e-05
-Epoch 9, training step 1854/1855, 0.722, lr: 5e-05
-Epoch 9, validation step 0/3
-Epoch 9, validation step 2/3
-The best model is saved on epoch 9!
-###############################################################
-Epoch 9, Validation is over, current metrics:
-best validation epoch: 9, best val_metrics: RCB = 0.8796999875058473
-F1-Radgraph-partial: 0.19972788342299275
-F1-Radgraph-simple: 0.23219854869722778
-F1-Radgraph-complete: 0.152644218918469
-chexbert_5_micro_f1: 0.4888888888888889
-chexbert_5_macro_f1: 0.46671447030495444
-chexbert_all_micro_p: 0.5380710659898477
-chexbert_all_micro_r: 0.5888888888888889
-chexbert_all_micro_f1: 0.5623342175066313
-chexbert_all_macro_p: 0.4014816373190757
-chexbert_all_macro_r: 0.4218285432195206
-chexbert_all_macro_f1: 0.4011878473251877
-chexbert_accuracy: 0.2702702702702703
-chen_bleu_1: 0.354796183588271
-chen_bleu_2: 0.22913221248992194
-chen_bleu_3: 0.16177322277471629
-chen_bleu_4: 0.11763788657622337
-chen_meteor: 0.1551498980852446
-chen_rouge: 0.26971463386849304
-chen_num_examples: 62
-RB: 0.3173657699992161
-RC: 0.762062100929624
-RCB: 0.8796999875058473 
-
-Epoch 9, Training is over, training epoch loss = 0.7797667980194092, lr: 5e-05. Epoch training time: 3303.1085772514343 seconds
-###############################################################
-Epoch 10, training step 0/1855, 0.765, lr: 5e-05
-Epoch 10, training step 500/1855, 0.641, lr: 5e-05
-Epoch 10, training step 1000/1855, 0.684, lr: 5e-05
-Epoch 10, training step 1500/1855, 0.795, lr: 5e-05
-Epoch 10, training step 1854/1855, 0.836, lr: 5e-05
-Epoch 10, validation step 0/3
-Epoch 10, validation step 2/3
-###############################################################
-Epoch 10, Validation is over, current metrics:
-best validation epoch: 9, best val_metrics: RCB = 0.8796999875058473
-F1-Radgraph-partial: 0.1992894482296047
-F1-Radgraph-simple: 0.22519366622545012
-F1-Radgraph-complete: 0.14662290945656212
-chexbert_5_micro_f1: 0.4161849710982659
-chexbert_5_macro_f1: 0.3804093567251462
-chexbert_all_micro_p: 0.4717948717948718
-chexbert_all_micro_r: 0.5111111111111111
-chexbert_all_micro_f1: 0.49066666666666664
-chexbert_all_macro_p: 0.32401906275330644
-chexbert_all_macro_r: 0.34128074515292567
-chexbert_all_macro_f1: 0.32003761693611316
-chexbert_accuracy: 0.20270270270270271
-chen_bleu_1: 0.35126010727716905
-chen_bleu_2: 0.22330124084144842
-chen_bleu_3: 0.15632463281573358
-chen_bleu_4: 0.1142193367543019
-chen_meteor: 0.15430710584101837
-chen_rouge: 0.2537626840525148
-chen_num_examples: 62
-RB: 0.3135087849839066
-RC: 0.6899561148962714
-RCB: 0.8041754516505732 
-
-Epoch 10, Training is over, training epoch loss = 0.7102962136268616, lr: 5e-05. Epoch training time: 3169.5202717781067 seconds
-###############################################################
-Epoch 11, training step 0/1855, 0.729, lr: 5e-05
-Epoch 11, training step 500/1855, 0.568, lr: 5e-05
-Epoch 11, training step 1000/1855, 0.713, lr: 5e-05
-Epoch 11, training step 1500/1855, 0.644, lr: 5e-05
-Epoch 11, training step 1854/1855, 0.653, lr: 5e-05
-Epoch 11, validation step 0/3
-Epoch 11, validation step 2/3
-###############################################################
-Epoch 11, Validation is over, current metrics:
-best validation epoch: 9, best val_metrics: RCB = 0.8796999875058473
-F1-Radgraph-partial: 0.21475044799880041
-F1-Radgraph-simple: 0.2364068314955351
-F1-Radgraph-complete: 0.1621375710198932
-chexbert_5_micro_f1: 0.44047619047619047
-chexbert_5_macro_f1: 0.43222111222111226
-chexbert_all_micro_p: 0.4972972972972973
-chexbert_all_micro_r: 0.5111111111111111
-chexbert_all_micro_f1: 0.5041095890410959
-chexbert_all_macro_p: 0.3322562358276645
-chexbert_all_macro_r: 0.3310574596664822
-chexbert_all_macro_f1: 0.32393524395404094
-chexbert_accuracy: 0.28378378378378377
-chen_bleu_1: 0.3567271295314524
-chen_bleu_2: 0.23296514394745924
-chen_bleu_3: 0.16453050794947513
-chen_bleu_4: 0.121202788764822
-chen_meteor: 0.15688277135102827
-chen_rouge: 0.2690422702986062
-chen_num_examples: 62
-RB: 0.33595323676362243
-RC: 0.7188600370398963
-RCB: 0.8400628258047184 
-
-Epoch 11, Training is over, training epoch loss = 0.6462647914886475, lr: 5e-05. Epoch training time: 3382.2880659103394 seconds
-###############################################################
-Epoch 12, training step 0/1855, 0.571, lr: 5e-05
-Epoch 12, training step 500/1855, 0.454, lr: 5e-05
-Epoch 12, training step 1000/1855, 0.573, lr: 5e-05
-Epoch 12, training step 1500/1855, 0.683, lr: 5e-05
-Epoch 12, training step 1854/1855, 0.59, lr: 5e-05
-Epoch 12, validation step 0/3
-Epoch 12, validation step 2/3
-The best model is saved on epoch 12!
-###############################################################
-Epoch 12, Validation is over, current metrics:
-best validation epoch: 12, best val_metrics: RCB = 0.8904913842627782
-F1-Radgraph-partial: 0.21346428762310224
-F1-Radgraph-simple: 0.2374007763791543
-F1-Radgraph-complete: 0.16172873674084237
-chexbert_5_micro_f1: 0.5384615384615384
-chexbert_5_macro_f1: 0.5265643648763854
-chexbert_all_micro_p: 0.5868263473053892
-chexbert_all_micro_r: 0.5444444444444444
-chexbert_all_micro_f1: 0.5648414985590778
-chexbert_all_macro_p: 0.4335233860443944
-chexbert_all_macro_r: 0.4097051611713266
-chexbert_all_macro_f1: 0.411902970631918
-chexbert_accuracy: 0.32432432432432434
-chen_bleu_1: 0.34976570841803306
-chen_bleu_2: 0.2203744448483044
-chen_bleu_3: 0.15471346102479588
-chen_bleu_4: 0.11218559808059807
-chen_meteor: 0.14848611892190736
-chen_rouge: 0.26404931819454563
-chen_num_examples: 62
-RB: 0.3256498857037003
-RC: 0.77830578618218
-RCB: 0.8904913842627782 
-
-Epoch 12, Training is over, training epoch loss = 0.58570396900177, lr: 5e-05. Epoch training time: 3258.8526549339294 seconds
-###############################################################
-Epoch 13, training step 0/1855, 0.343, lr: 5e-05
-Epoch 13, training step 500/1855, 0.565, lr: 5e-05
-Epoch 13, training step 1000/1855, 0.55, lr: 5e-05
-Epoch 13, training step 1500/1855, 0.49, lr: 5e-05
-Epoch 13, training step 1854/1855, 0.581, lr: 5e-05
-Epoch 13, validation step 0/3
-Epoch 13, validation step 2/3
-###############################################################
-Epoch 13, Validation is over, current metrics:
-best validation epoch: 12, best val_metrics: RCB = 0.8904913842627782
-F1-Radgraph-partial: 0.22166534962809928
-F1-Radgraph-simple: 0.24854847264707447
-F1-Radgraph-complete: 0.17090046632299813
-chexbert_5_micro_f1: 0.43312101910828027
-chexbert_5_macro_f1: 0.4127351908461353
-chexbert_all_micro_p: 0.5089820359281437
-chexbert_all_micro_r: 0.4722222222222222
-chexbert_all_micro_f1: 0.4899135446685879
-chexbert_all_macro_p: 0.33600460386174674
-chexbert_all_macro_r: 0.30813901637209906
-chexbert_all_macro_f1: 0.3088396094481447
-chexbert_accuracy: 0.20270270270270271
-chen_bleu_1: 0.3533298794495646
-chen_bleu_2: 0.2280113596958507
-chen_bleu_3: 0.16303476434759134
-chen_bleu_4: 0.11990811636716939
-chen_meteor: 0.15642998552986945
-chen_rouge: 0.2692314678017771
-chen_num_examples: 62
-RB: 0.34157346599526867
-RC: 0.7115788942966872
-RCB: 0.8314870106638566 
-
-Epoch 13, Training is over, training epoch loss = 0.5284979939460754, lr: 5e-05. Epoch training time: 3251.961864709854 seconds
-###############################################################
-Epoch 14, training step 0/1855, 0.446, lr: 5e-05
-Epoch 14, training step 500/1855, 0.475, lr: 5e-05
-Epoch 14, training step 1000/1855, 0.426, lr: 5e-05
-Epoch 14, training step 1500/1855, 0.546, lr: 5e-05
-Epoch 14, training step 1854/1855, 0.512, lr: 5e-05
-Epoch 14, validation step 0/3
-Epoch 14, validation step 2/3
-###############################################################
-Epoch 14, Validation is over, current metrics:
-best validation epoch: 12, best val_metrics: RCB = 0.8904913842627782
-F1-Radgraph-partial: 0.20269397917153106
-F1-Radgraph-simple: 0.22783225810271748
-F1-Radgraph-complete: 0.14776831330955664
-chexbert_5_micro_f1: 0.4585987261146497
-chexbert_5_macro_f1: 0.4558594006164339
-chexbert_all_micro_p: 0.5151515151515151
-chexbert_all_micro_r: 0.4722222222222222
-chexbert_all_micro_f1: 0.4927536231884058
-chexbert_all_macro_p: 0.39138136414673835
-chexbert_all_macro_r: 0.3223175362649047
-chexbert_all_macro_f1: 0.34305695252989243
-chexbert_accuracy: 0.36486486486486486
-chen_bleu_1: 0.3532023096465505
-chen_bleu_2: 0.22729672383113422
-chen_bleu_3: 0.1605252335513812
-chen_bleu_4: 0.11766215044534757
-chen_meteor: 0.15275664310351603
-chen_rouge: 0.2697377782778773
-chen_num_examples: 62
-RB: 0.32035612961687865
-RC: 0.6954476023599369
-RCB: 0.8131097528052844 
-
-Epoch 14, Training is over, training epoch loss = 0.47589635848999023, lr: 5e-05. Epoch training time: 3524.114292860031 seconds
-###############################################################
-Epoch 15, training step 0/1855, 0.388, lr: 5e-05
-Epoch 15, training step 500/1855, 0.38, lr: 5e-05
-Epoch 15, training step 1000/1855, 0.548, lr: 5e-05
-Epoch 15, training step 1500/1855, 0.406, lr: 5e-05
-Epoch 15, training step 1854/1855, 0.521, lr: 5e-05
-Epoch 15, validation step 0/3
-Epoch 15, validation step 2/3
-###############################################################
-Epoch 15, Validation is over, current metrics:
-best validation epoch: 12, best val_metrics: RCB = 0.8904913842627782
-F1-Radgraph-partial: 0.20006528397694634
-F1-Radgraph-simple: 0.22319933977464823
-F1-Radgraph-complete: 0.1399975799819895
-chexbert_5_micro_f1: 0.5061728395061729
-chexbert_5_macro_f1: 0.4589031910404815
-chexbert_all_micro_p: 0.5657142857142857
-chexbert_all_micro_r: 0.55
-chexbert_all_micro_f1: 0.5577464788732395
-chexbert_all_macro_p: 0.41968119155619155
-chexbert_all_macro_r: 0.3984954811270601
-chexbert_all_macro_f1: 0.3991504070565822
-chexbert_accuracy: 0.35135135135135137
-chen_bleu_1: 0.32904273021927166
-chen_bleu_2: 0.20596314513558867
-chen_bleu_3: 0.1425729504051587
-chen_bleu_4: 0.1031871891594472
-chen_meteor: 0.14752424331917288
-chen_rouge: 0.25716151594150544
-chen_num_examples: 62
-RB: 0.30325247313639353
-RC: 0.7578117628501858
-RCB: 0.8609989520096331 
-
-Epoch 15, Training is over, training epoch loss = 0.4276135265827179, lr: 5e-05. Epoch training time: 3805.063766479492 seconds
-###############################################################
-Epoch 16, training step 0/1855, 0.478, lr: 5e-05
-Epoch 16, training step 500/1855, 0.386, lr: 5e-05
-Epoch 16, training step 1000/1855, 0.415, lr: 5e-05
-Epoch 16, training step 1500/1855, 0.279, lr: 5e-05
-Epoch 16, training step 1854/1855, 0.455, lr: 5e-05
-Epoch 16, validation step 0/3
-Epoch 16, validation step 2/3
-###############################################################
-Epoch 16, Validation is over, current metrics:
-best validation epoch: 12, best val_metrics: RCB = 0.8904913842627782
-F1-Radgraph-partial: 0.19500457787528733
-F1-Radgraph-simple: 0.22367288609052008
-F1-Radgraph-complete: 0.14385065291721189
-chexbert_5_micro_f1: 0.4157303370786517
-chexbert_5_macro_f1: 0.4061367331222726
-chexbert_all_micro_p: 0.46875
-chexbert_all_micro_r: 0.5
-chexbert_all_micro_f1: 0.4838709677419355
-chexbert_all_macro_p: 0.30834223907532177
-chexbert_all_macro_r: 0.3300599984810511
-chexbert_all_macro_f1: 0.31018189528041745
-chexbert_accuracy: 0.22972972972972974
-chen_bleu_1: 0.3459906355606208
-chen_bleu_2: 0.2196180823678822
-chen_bleu_3: 0.1521887431354563
-chen_bleu_4: 0.10987164409445438
-chen_meteor: 0.1495198277824665
-chen_rouge: 0.25902060035850816
-chen_num_examples: 62
-RB: 0.30487622196974173
-RC: 0.6788755456172229
-RCB: 0.7887471897116772 
-
-Epoch 16, Training is over, training epoch loss = 0.38339459896087646, lr: 5e-05. Epoch training time: 3775.7138345241547 seconds
-###############################################################
-Epoch 17, training step 0/1855, 0.365, lr: 5e-05
-Epoch 17, training step 500/1855, 0.319, lr: 5e-05
-Epoch 17, training step 1000/1855, 0.321, lr: 5e-05
-Epoch 17, training step 1500/1855, 0.334, lr: 5e-05
-Epoch 17, training step 1854/1855, 0.33, lr: 5e-05
-Epoch 17, validation step 0/3
-Epoch 17, validation step 2/3
-###############################################################
-Epoch 17, Validation is over, current metrics:
-best validation epoch: 12, best val_metrics: RCB = 0.8904913842627782
-F1-Radgraph-partial: 0.1825730540526958
-F1-Radgraph-simple: 0.2008214815642524
-F1-Radgraph-complete: 0.1249465253922876
-chexbert_5_micro_f1: 0.4024390243902439
-chexbert_5_macro_f1: 0.37110771581359814
-chexbert_all_micro_p: 0.49707602339181284
-chexbert_all_micro_r: 0.4722222222222222
-chexbert_all_micro_f1: 0.4843304843304843
-chexbert_all_macro_p: 0.3515516890516891
-chexbert_all_macro_r: 0.32245847302238284
-chexbert_all_macro_f1: 0.33082034217465656
-chexbert_accuracy: 0.25675675675675674
-chen_bleu_1: 0.3395538894554275
-chen_bleu_2: 0.21285520719549617
-chen_bleu_3: 0.14764517273851474
-chen_bleu_4: 0.10717229564275034
-chen_meteor: 0.14400912008824707
-chen_rouge: 0.25011718923236953
-chen_num_examples: 62
-RB: 0.28974534969544613
-RC: 0.6669035383831801
-RCB: 0.7740758340259304 
-
-Epoch 17, Training is over, training epoch loss = 0.34514665603637695, lr: 5e-05. Epoch training time: 3546.3713114261627 seconds
-###############################################################
-Epoch 18, training step 0/1855, 0.283, lr: 5e-05
-Epoch 18, training step 500/1855, 0.333, lr: 5e-05
-Epoch 18, training step 1000/1855, 0.323, lr: 5e-05