learning_rate: 5e-05
train_batch_size: 4
eval_batch_size: 4
seed: 42
optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 40
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss
2.2238	1.0	941	0.4982
0.5068	2.0	1882	0.4804
0.4874	3.0	2823	0.4810
0.4821	4.0	3764	0.4825
0.4811	5.0	4705	0.4766
0.4801	6.0	5646	0.4801
0.4743	7.0	6587	0.4825
0.4732	8.0	7528	0.4769
0.4717	9.0	8469	0.4731
0.4675	10.0	9410	0.4779
0.4667	11.0	10351	0.4795
0.4637	12.0	11292	0.4822
0.4575	13.0	12233	0.4817
0.4505	14.0	13174	0.4884
0.4452	15.0	14115	0.4935
0.4413	16.0	15056	0.4993
0.4386	17.0	15997	0.4943
0.4398	18.0	16938	0.5011
0.4263	19.0	17879	0.5114
0.4293	20.0	18820	0.4993
0.4255	21.0	19761	0.5084
0.4245	22.0	20702	0.5193
0.4153	23.0	21643	0.5032
0.4164	24.0	22584	0.5182
0.4012	25.0	23525	0.5239
0.3999	26.0	24466	0.5245
0.3969	27.0	25407	0.5293
0.3988	28.0	26348	0.5221
0.4033	29.0	27289	0.5394
0.3918	30.0	28230	0.5299
0.3841	31.0	29171	0.5396
0.3805	32.0	30112	0.5426
0.3733	33.0	31053	0.5435
0.384	34.0	31994	0.5532
0.3762	35.0	32935	0.5476
0.371	36.0	33876	0.5485
0.3629	37.0	34817	0.5529
0.3675	38.0	35758	0.5512
0.3671	39.0	36699	0.5543
0.3596	40.0	37640	0.5548

ngdangkhanh/results_lora

results_lora