learning_rate: 5e-05
train_batch_size: 2
eval_batch_size: 2
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 5
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss
3.1243	0.1	20	3.4478
2.873	0.2	40	3.3281
3.1557	0.3	60	3.2113
3.3101	0.4	80	3.0157
3.2964	0.5	100	2.8869
2.6947	0.6	120	2.7128
3.0246	0.7	140	2.6585
2.9896	0.8	160	2.5381
3.0139	0.9	180	2.5066
2.881	1.0	200	2.4570
2.4749	1.1	220	2.2652
2.4013	1.2	240	2.1501
2.3117	1.3	260	2.0454
2.4801	1.4	280	2.0140
2.4018	1.5	300	2.0051
2.7772	1.6	320	1.9326
2.3631	1.7	340	1.8746
2.6293	1.8	360	1.7805
2.0789	1.9	380	1.7009
2.0693	2.0	400	1.6646
1.593	2.1	420	1.5909
1.7124	2.2	440	1.5472
1.7553	2.3	460	1.4950
2.323	2.4	480	1.4694
1.8226	2.5	500	1.4138
1.7175	2.6	520	1.3636
2.0969	2.7	540	1.3204
1.4713	2.8	560	1.2640
1.7953	2.9	580	1.2374
1.886	3.0	600	1.2027
1.4564	3.1	620	1.1724
1.3773	3.2	640	1.1475
1.3743	3.3	660	1.1306
0.9328	3.4	680	1.1071
1.3489	3.5	700	1.0849
1.6036	3.6	720	1.0749
1.2027	3.7	740	1.0523
1.3651	3.8	760	1.0285
2.1274	3.9	780	1.0114
1.4608	4.0	800	0.9799
1.1099	4.1	820	0.9630
0.9091	4.2	840	0.9468
1.1335	4.3	860	0.9308
0.9962	4.4	880	0.9115
1.3564	4.5	900	0.9003
1.334	4.6	920	0.8957
1.3839	4.7	940	0.8893
1.4252	4.8	960	0.8861
1.192	4.9	980	0.8830
1.3856	5.0	1000	0.8822

orionshaw777/gpt2-belle-sft-offcial-1st

gpt2-belle-sft-offcial-1st