learning_rate: 0.0003
train_batch_size: 4
eval_batch_size: 4
seed: 42
gradient_accumulation_steps: 8
total_train_batch_size: 32
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: 100
training_steps: 500
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss
10.0209	0.0008	20	9.9645
9.8202	0.0016	40	9.7292
9.4854	0.0024	60	9.5337
9.4763	0.0032	80	9.5153
9.4127	0.0041	100	9.5089
9.4982	0.0049	120	9.5073
9.4613	0.0057	140	9.5042
9.4643	0.0065	160	9.5022
9.4624	0.0073	180	9.5008
9.4839	0.0081	200	9.4995
9.4669	0.0089	220	9.4988
9.5096	0.0097	240	9.4982
9.4860	0.0105	260	9.4972
9.6281	0.0113	280	9.4966
9.5361	0.0122	300	9.4963
9.5112	0.0130	320	9.4961
9.5311	0.0138	340	9.4954
9.5714	0.0146	360	9.4951
9.4447	0.0154	380	9.4948
9.4350	0.0162	400	9.4945
9.4349	0.0170	420	9.4943
9.6344	0.0178	440	9.4941
9.6377	0.0186	460	9.4939
9.4358	0.0194	480	9.4938
9.4618	0.0203	500	9.4937

ueihieu/llama-2-7b-genwiki-context

llama-2-7b-genwiki-context