learning_rate: 8e-05
train_batch_size: 32
eval_batch_size: 64
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 100
num_epochs: 15

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.5431	1.0	51	0.4347	0.7111	0.3810	1.0	0.5517
0.2821	2.0	102	0.2737	0.8722	0.5849	0.9688	0.7294
0.3530	3.0	153	0.2286	0.95	0.8108	0.9375	0.8696
0.2113	4.0	204	0.2000	0.9667	0.9062	0.9062	0.9062
0.0269	5.0	255	0.1964	0.9667	0.8824	0.9375	0.9091
0.1652	6.0	306	0.2534	0.9722	0.9655	0.875	0.9180
0.0877	7.0	357	0.2379	0.9333	0.7632	0.9062	0.8286
0.0900	8.0	408	0.2605	0.9444	0.7895	0.9375	0.8571
0.0283	9.0	459	0.4335	0.9611	0.9310	0.8438	0.8852
0.1702	10.0	510	0.2988	0.9778	1.0	0.875	0.9333
0.0433	11.0	561	0.4396	0.9667	1.0	0.8125	0.8966
0.0662	12.0	612	0.3019	0.9667	0.9643	0.8438	0.9
0.0142	13.0	663	0.3593	0.9722	1.0	0.8438	0.9153
0.0631	14.0	714	0.3206	0.9722	1.0	0.8438	0.9153
0.0342	15.0	765	0.3031	0.9722	1.0	0.8438	0.9153

Kaspar/anti-scraping_best_model_modernbert

anti-scraping_best_model_modernbert