learning_rate: 0.0001
train_batch_size: 256
eval_batch_size: 256
seed: 51
optimizer: Use 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: 500
num_epochs: 20
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss
6.8232	1.0	391	5.7629
5.406	2.0	782	5.1875
5.0087	3.0	1173	4.9331
4.7829	4.0	1564	4.7714
4.6167	5.0	1955	4.6582
4.4815	6.0	2346	4.5697
4.3666	7.0	2737	4.5037
4.2653	8.0	3128	4.4528
4.1745	9.0	3519	4.4176
4.0912	10.0	3910	4.3873
4.017	11.0	4301	4.3680
3.9506	12.0	4692	4.3506
3.8905	13.0	5083	4.3434
3.8364	14.0	5474	4.3391
3.7881	15.0	5865	4.3391
3.745	16.0	6256	4.3355
3.707	17.0	6647	4.3384
3.6743	18.0	7038	4.3387
3.6471	19.0	7429	4.3408
3.6266	20.0	7820	4.3422

fpadovani/bnc_o_51

bnc_o_51