learning_rate: 0.001
train_batch_size: 32
eval_batch_size: 64
seed: 42
gradient_accumulation_steps: 8
total_train_batch_size: 256
optimizer: Use OptimizerNames.ADAMW_TORCH 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: 32000
num_epochs: 20.0
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Accuracy
6.2835	1.0	1731	4.8659	0.2345
4.3226	2.0	3462	4.5072	0.2571
3.9025	3.0	5193	4.3039	0.2727
3.7604	4.0	6924	4.1970	0.2799
3.5741	5.0	8655	4.1415	0.2859
3.469	6.0	10386	4.0908	0.2912
3.3965	7.0	12117	4.0822	0.2911
3.3436	8.0	13848	4.0801	0.2913
3.3009	9.0	15579	4.0623	0.2940
3.2774	10.0	17310	4.0713	0.2937
3.2586	11.0	19041	4.0397	0.2959
3.211	12.0	20772	4.0307	0.2963
3.2057	13.0	22503	4.0445	0.2955
3.2042	14.0	24234	4.0426	0.2953
3.1711	15.0	25965	4.0267	0.2969
3.1575	16.0	27696	4.0514	0.2945
3.1697	17.0	29427	4.0238	0.2976
3.1769	18.0	31158	4.0275	0.2970
3.1431	19.0	32889	4.0065	0.3000
3.0097	20.0	34620	3.9945	0.3029

qing-yao/long_first_nodatives_seed-42_1e-3

long_first_nodatives_seed-42_1e-3