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
7.1673	1.0	392	6.4983
6.3087	2.0	784	6.2686
6.1399	3.0	1176	6.1614
6.0374	4.0	1568	6.0855
5.9563	5.0	1960	6.0309
5.8864	6.0	2352	5.9894
5.8227	7.0	2744	5.9568
5.7625	8.0	3136	5.9310
5.7042	9.0	3528	5.9099
5.6474	10.0	3920	5.8963
5.5918	11.0	4312	5.8884
5.538	12.0	4704	5.8837
5.486	13.0	5096	5.8846
5.4363	14.0	5488	5.8874
5.3902	15.0	5880	5.8926
5.3476	16.0	6272	5.9003
5.3098	17.0	6664	5.9072
5.2783	18.0	7056	5.9154
5.2534	19.0	7448	5.9197
5.2349	20.0	7840	5.9222

fpadovani/bnc_sh1_51

bnc_sh1_51