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
4.8895	1.0	495	3.9068
3.6476	2.0	990	3.5451
3.3809	3.0	1485	3.3822
3.2288	4.0	1980	3.2913
3.1213	5.0	2475	3.2308
3.0375	6.0	2970	3.1927
2.9663	7.0	3465	3.1607
2.9042	8.0	3960	3.1398
2.8473	9.0	4455	3.1283
2.7963	10.0	4950	3.1195
2.7467	11.0	5445	3.1206
2.7012	12.0	5940	3.1164
2.6561	13.0	6435	3.1189
2.615	14.0	6930	3.1275
2.5788	15.0	7425	3.1326
2.5452	16.0	7920	3.1433
2.5148	17.0	8415	3.1499
2.4904	18.0	8910	3.1584
2.468	19.0	9405	3.1654
2.4511	20.0	9900	3.1664

fpadovani/cds_sh1_42

cds_sh1_42