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
5.8069	1.0	422	4.8892
4.7607	2.0	844	4.7392
4.655	3.0	1266	4.6650
4.59	4.0	1688	4.6228
4.5422	5.0	2110	4.5952
4.5019	6.0	2532	4.5707
4.4653	7.0	2954	4.5549
4.4302	8.0	3376	4.5447
4.3958	9.0	3798	4.5371
4.3612	10.0	4220	4.5343
4.326	11.0	4642	4.5352
4.2902	12.0	5064	4.5400
4.2541	13.0	5486	4.5463
4.2184	14.0	5908	4.5564
4.1836	15.0	6330	4.5667
4.151	16.0	6752	4.5796
4.1208	17.0	7174	4.5917
4.0942	18.0	7596	4.6011
4.0717	19.0	8018	4.6100
4.0552	20.0	8440	4.6150

fpadovani/candor_sh1_13

candor_sh1_13