learning_rate: 5e-05
train_batch_size: 4
eval_batch_size: 4
seed: 2025
gradient_accumulation_steps: 4
total_train_batch_size: 16
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: 20
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Subset Accuracy	F1 Macro	F1 Micro	Precision Macro	Recall Macro	Roc Auc
1.1783	0.5002	767	0.1854	0.0679	0.0755	0.1256	0.3030	0.0466	0.7610
0.6868	1.0	1534	0.1661	0.1378	0.1564	0.2358	0.5075	0.1038	0.8281
0.6248	1.5002	2301	0.1613	0.2025	0.2269	0.3192	0.5102	0.1713	0.8416
0.6101	2.0	3068	0.1561	0.2364	0.2349	0.3546	0.5593	0.1735	0.8521
0.4834	2.5002	3835	0.1669	0.2506	0.2784	0.3696	0.5196	0.2150	0.8427
0.4572	3.0	4602	0.1663	0.2645	0.2952	0.3769	0.5287	0.2266	0.8445
0.2812	3.5002	5369	0.2168	0.2803	0.3263	0.3868	0.3978	0.2925	0.8280
0.2386	4.0	6136	0.2064	0.2894	0.3236	0.4024	0.4280	0.2785	0.8263
0.1431	4.5002	6903	0.2515	0.2842	0.3292	0.3992	0.4010	0.2898	0.8179
0.123	5.0	7670	0.2675	0.2736	0.3156	0.3847	0.3955	0.2744	0.8145
0.0905	5.5002	8437	0.3188	0.2851	0.3272	0.3999	0.3789	0.2947	0.8143
0.0678	6.0	9204	0.3120	0.2762	0.3217	0.3860	0.4000	0.2736	0.8112
0.0544	6.5002	9971	0.3340	0.2847	0.3330	0.4064	0.4040	0.2937	0.8085
0.0395	7.0	10738	0.3480	0.2811	0.3341	0.3964	0.3944	0.2954	0.8096
0.0257	7.5002	11505	0.3684	0.2865	0.3275	0.4046	0.4225	0.2841	0.8056
0.0291	8.0	12272	0.3824	0.2821	0.3210	0.3994	0.4053	0.2835	0.8038
0.0173	8.5002	13039	0.3921	0.2821	0.3194	0.3945	0.4329	0.2679	0.8053
0.0157	9.0	13806	0.4061	0.2829	0.3307	0.4065	0.3905	0.2954	0.8035
0.0123	9.5002	14573	0.4170	0.2778	0.3294	0.3990	0.4078	0.2869	0.8058

DayCardoso/modern-bert-seq-class-values-no-context

modern-bert-seq-class-values-no-context