learning_rate: 0.0001
train_batch_size: 256
eval_batch_size: 256
seed: 42
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
4.8246	1.0	495	3.7829
3.5358	2.0	990	3.4373
3.2822	3.0	1485	3.2929
3.1379	4.0	1980	3.2110
3.0361	5.0	2475	3.1546
2.9554	6.0	2970	3.1169
2.8872	7.0	3465	3.0929
2.8263	8.0	3960	3.0772
2.7708	9.0	4455	3.0658
2.719	10.0	4950	3.0605
2.67	11.0	5445	3.0612
2.6235	12.0	5940	3.0647
2.5783	13.0	6435	3.0716
2.5369	14.0	6930	3.0806
2.5001	15.0	7425	3.0869
2.4663	16.0	7920	3.0978
2.4367	17.0	8415	3.1059
2.4105	18.0	8910	3.1153
2.3888	19.0	9405	3.1227
2.3717	20.0	9900	3.1267

fpadovani/cds_np_42

cds_np_42