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
5.2778	1.0	422	4.0582
3.762	2.0	844	3.7296
3.5498	3.0	1266	3.6169
3.4393	4.0	1688	3.5491
3.36	5.0	2110	3.5029
3.2963	6.0	2532	3.4730
3.2408	7.0	2954	3.4512
3.1904	8.0	3376	3.4392
3.1432	9.0	3798	3.4301
3.0976	10.0	4220	3.4267
3.0539	11.0	4642	3.4269
3.0115	12.0	5064	3.4295
2.9707	13.0	5486	3.4376
2.9318	14.0	5908	3.4456
2.8937	15.0	6330	3.4540
2.8591	16.0	6752	3.4638
2.829	17.0	7174	3.4750
2.8032	18.0	7596	3.4843
2.7818	19.0	8018	3.4906
2.7648	20.0	8440	3.4945

fpadovani/candor_o_42

candor_o_42