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
4.4895	1.0	496	3.2871
3.0404	2.0	992	2.9813
2.812	3.0	1488	2.8688
2.6919	4.0	1984	2.8078
2.6081	5.0	2480	2.7706
2.5419	6.0	2976	2.7438
2.485	7.0	3472	2.7265
2.4341	8.0	3968	2.7137
2.3864	9.0	4464	2.7111
2.3415	10.0	4960	2.7089
2.2983	11.0	5456	2.7111
2.257	12.0	5952	2.7166
2.2176	13.0	6448	2.7254
2.1798	14.0	6944	2.7353
2.1436	15.0	7440	2.7440
2.1115	16.0	7936	2.7541
2.0833	17.0	8432	2.7650
2.0588	18.0	8928	2.7714
2.0385	19.0	9424	2.7802
2.0226	20.0	9920	2.7836

fpadovani/cds_o_51

cds_o_51