learning_rate: 0.001
train_batch_size: 32
eval_batch_size: 64
seed: 63
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
6.0721	0.9999	1507	4.5216	0.2803
4.1353	1.9998	3014	4.0893	0.3122
3.94	2.9998	4521	3.8747	0.3281
3.7064	3.9997	6028	3.7568	0.3374
3.6181	4.9996	7535	3.6738	0.3454
3.5108	5.9995	9042	3.6136	0.3513
3.4484	6.9994	10549	3.5901	0.3528
3.395	8.0	12057	3.5658	0.3555
3.3481	8.9999	13564	3.5571	0.3568
3.326	9.9998	15071	3.5395	0.3584
3.2846	10.9998	16578	3.5257	0.3602
3.2805	11.9997	18085	3.5198	0.3609
3.2426	12.9996	19592	3.5322	0.3596
3.2487	13.9995	21099	3.5246	0.3614
3.2137	14.9994	22606	3.5166	0.3618
3.2257	16.0	24114	3.5283	0.3608
3.1932	16.9999	25621	3.5097	0.3632
3.2104	17.9998	27128	3.5262	0.3601
3.1805	18.9998	28635	3.5176	0.3619
3.1995	19.9983	30140	3.5107	0.3634

qing-yao/no-2postverbal_seed-63_1e-3

noditrans_cf_seed-63_1e-3