learning_rate: 0.0001
train_batch_size: 256
eval_batch_size: 256
seed: 13
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
4.5857	1.0	496	3.5241
3.3228	2.0	992	3.3001
3.1656	3.0	1488	3.2203
3.078	4.0	1984	3.1739
3.0125	5.0	2480	3.1434
2.9601	6.0	2976	3.1195
2.9126	7.0	3472	3.1073
2.868	8.0	3968	3.1036
2.8271	9.0	4464	3.0964
2.787	10.0	4960	3.0980
2.7475	11.0	5456	3.0985
2.71	12.0	5952	3.1066
2.6721	13.0	6448	3.1152
2.6369	14.0	6944	3.1243
2.6025	15.0	7440	3.1367
2.5712	16.0	7936	3.1456
2.5427	17.0	8432	3.1547
2.519	18.0	8928	3.1654
2.4987	19.0	9424	3.1713
2.4831	20.0	9920	3.1763

fpadovani/cds_w_13

cds_w_13