learning_rate: 0.0001
train_batch_size: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 32
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 30
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	2	1.8512	0.1579
No log	2.0	4	1.8512	0.1579
No log	3.0	6	1.7512	0.4211
No log	4.0	8	1.6830	0.3684
No log	5.0	10	1.5444	0.6316
No log	6.0	12	1.4469	0.4737
No log	7.0	14	1.3275	0.6316
No log	8.0	16	1.1882	0.8421
No log	9.0	18	1.0719	0.8421
No log	10.0	20	1.0585	0.8421
No log	11.0	22	1.0017	0.8947
No log	12.0	24	0.9294	0.8947
No log	13.0	26	0.8423	0.8947
No log	14.0	28	0.7696	0.8421
No log	15.0	30	0.7280	0.8947
No log	16.0	32	0.6871	0.9474
No log	17.0	34	0.6423	0.9474
No log	18.0	36	0.6102	0.9474
No log	19.0	38	0.5863	0.9474
No log	20.0	40	0.5561	0.9474
No log	21.0	42	0.5284	0.9474
No log	22.0	44	0.5049	0.9474
No log	23.0	46	0.4834	0.9474
No log	24.0	48	0.4601	0.9474
No log	25.0	50	0.4384	0.9474
No log	26.0	52	0.4260	0.9474
No log	27.0	54	0.4083	0.9474
No log	28.0	56	0.3979	0.9474
No log	29.0	58	0.3931	0.9474
No log	30.0	60	0.3852	0.9474

sungsy0323/wks

wks