learning_rate: 0.0006
train_batch_size: 4
eval_batch_size: 4
seed: 1234
gradient_accumulation_steps: 16
total_train_batch_size: 64
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: cosine
lr_scheduler_warmup_ratio: 0.03
num_epochs: 1

Training Loss	Epoch	Step	Validation Loss
6.8532	0.0320	100	6.7483
6.1754	0.0640	200	6.1243
5.8756	0.0959	300	5.7804
5.5348	0.1279	400	5.5261
5.2918	0.1599	500	5.3721
5.329	0.1919	600	5.2467
5.0479	0.2239	700	5.1346
5.0769	0.2559	800	5.0477
4.9082	0.2878	900	4.9726
4.8851	0.3198	1000	4.9025
4.8578	0.3518	1100	4.8424
4.7683	0.3838	1200	4.7891
4.7845	0.4158	1300	4.7421
4.7651	0.4477	1400	4.6986
4.6101	0.4797	1500	4.6589
4.5814	0.5117	1600	4.6180
4.5607	0.5437	1700	4.5858
4.62	0.5757	1800	4.5545
4.4465	0.6076	1900	4.5254
4.5038	0.6396	2000	4.5018
4.4746	0.6716	2100	4.4765
4.4328	0.7036	2200	4.4544
4.4182	0.7356	2300	4.4368
4.4987	0.7676	2400	4.4215
4.4017	0.7995	2500	4.4085
4.4284	0.8315	2600	4.3983
4.3105	0.8635	2700	4.3901
4.2949	0.8955	2800	4.3846
4.3673	0.9275	2900	4.3812
4.3048	0.9594	3000	4.3796
4.4036	0.9914	3100	4.3793

mtzig/lltransformer-linear-test1

lltransformer-linear-test1