learning_rate: 0.0001
train_batch_size: 16
eval_batch_size: 16
seed: 2025
gradient_accumulation_steps: 2
total_train_batch_size: 32
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 25
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Subset Accuracy	F1 Macro	F1 Micro	Precision Macro	Recall Macro	Roc Auc
0.3261	1.0	767	0.1848	0.0358	0.0273	0.0682	0.1322	0.0165	0.7523
0.1653	2.0	1534	0.1687	0.2000	0.1740	0.3048	0.3058	0.1429	0.8180
0.1422	3.0	2301	0.1700	0.2427	0.2509	0.3398	0.4408	0.1957	0.8239
0.1091	4.0	3068	0.1864	0.3055	0.2801	0.3921	0.4249	0.2469	0.8166
0.0806	5.0	3835	0.2045	0.3021	0.2926	0.3841	0.3993	0.2540	0.8079
0.0542	6.0	4602	0.2328	0.2992	0.2961	0.3785	0.3630	0.2644	0.7991
0.0434	7.0	5369	0.2475	0.2842	0.3076	0.3845	0.3510	0.2955	0.7982
0.0311	8.0	6136	0.2661	0.2919	0.3236	0.3904	0.3442	0.3168	0.7920
0.025	9.0	6903	0.2812	0.2844	0.3201	0.3858	0.3441	0.3085	0.7878
0.0186	10.0	7670	0.3123	0.2826	0.3246	0.3879	0.3432	0.3220	0.7827
0.0152	11.0	8437	0.3156	0.2689	0.3217	0.3774	0.3374	0.3157	0.7828
0.0109	12.0	9204	0.3380	0.2856	0.3127	0.3800	0.3410	0.2994	0.7785
0.0095	13.0	9971	0.3636	0.2684	0.3187	0.3783	0.3385	0.3140	0.7699
0.0063	14.0	10738	0.3684	0.2835	0.3206	0.3807	0.3478	0.3049	0.7755
0.0054	15.0	11505	0.3789	0.2907	0.3225	0.3837	0.3470	0.3052	0.7784

DayCardoso/bert-seq-class-values-no-context_plus

bert-seq-class-values-no-context_plus