learning_rate: 0.001
train_batch_size: 512
eval_batch_size: 512
seed: 42
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: constant
training_steps: 1000

Training Loss	Epoch	Step	Validation Loss
6.0216	0.0270	50	5.7194
5.1598	0.0539	100	4.9726
4.4774	0.0809	150	4.2392
3.8862	0.1078	200	3.8088
3.6419	0.1348	250	3.5390
3.3524	0.1617	300	3.3258
3.2040	0.1887	350	3.1441
3.0148	0.2156	400	2.9984
2.9060	0.2426	450	2.8648
2.7618	0.2695	500	2.7503
2.6942	0.2965	550	2.6675
2.5873	0.3235	600	2.5746
2.5373	0.3504	650	2.5231
2.4670	0.3774	700	2.4693
2.4254	0.4043	750	2.4194
2.3780	0.4313	800	2.3802
2.3325	0.4582	850	2.3369
2.2927	0.4852	900	2.2886
2.2744	0.5121	950	2.2674
2.2442	0.5391	1000	2.2536

w-ahmad/A-mlp-w1a-21L

A-mlp-w1a-21L