learning_rate: 5e-05
train_batch_size: 64
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: 1000
num_epochs: 2
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss
4.0115	0.0558	200	3.8299
3.8477	0.1117	400	3.6582
3.7729	0.1675	600	3.5849
3.7034	0.2233	800	3.5290
3.6939	0.2792	1000	3.5021
3.6489	0.3350	1200	3.4778
3.6077	0.3908	1400	3.4516
3.6035	0.4467	1600	3.4380
3.5752	0.5025	1800	3.4236
3.5582	0.5583	2000	3.4045
3.55	0.6142	2200	3.3969
3.5716	0.6700	2400	3.3876
3.5457	0.7259	2600	3.3816
3.5142	0.7817	2800	3.3731
3.5076	0.8375	3000	3.3619
3.5049	0.8934	3200	3.3641
3.5096	0.9492	3400	3.3531
3.4916	1.0050	3600	3.3495
3.5046	1.0609	3800	3.3467
3.5012	1.1167	4000	3.3414
3.4795	1.1725	4200	3.3315
3.4746	1.2284	4400	3.3363
3.4401	1.2842	4600	3.3295
3.4625	1.3400	4800	3.3322
3.4851	1.3959	5000	3.3239
3.4795	1.4517	5200	3.3224
3.4653	1.5075	5400	3.3143
3.4749	1.5634	5600	3.3119
3.4574	1.6192	5800	3.3112
3.4566	1.6750	6000	3.3134
3.4627	1.7309	6200	3.3128
3.465	1.7867	6400	3.3046
3.4601	1.8425	6600	3.2984
3.4565	1.8984	6800	3.2979
3.4504	1.9542	7000	3.3070

opttt/AISecurityFinal

AISecurityFinal