hushem_40x_deit_base_sgd_0001_fold5

This model is a fine-tuned version of facebook/deit-base-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 1.1678
• Accuracy: 0.5122

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0001
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.3902	1.0	220	1.3668	0.2683
1.375	2.0	440	1.3610	0.2927
1.3643	3.0	660	1.3560	0.2683
1.3352	4.0	880	1.3513	0.2683
1.343	5.0	1100	1.3466	0.2683
1.2985	6.0	1320	1.3416	0.2683
1.3152	7.0	1540	1.3365	0.2927
1.2618	8.0	1760	1.3311	0.3171
1.2728	9.0	1980	1.3254	0.3415
1.2604	10.0	2200	1.3195	0.3415
1.2446	11.0	2420	1.3136	0.3415
1.2322	12.0	2640	1.3076	0.3902
1.2519	13.0	2860	1.3017	0.4146
1.2115	14.0	3080	1.2958	0.4146
1.2112	15.0	3300	1.2899	0.4390
1.1892	16.0	3520	1.2841	0.4390
1.1942	17.0	3740	1.2784	0.4390
1.2008	18.0	3960	1.2727	0.4390
1.1853	19.0	4180	1.2671	0.4390
1.1573	20.0	4400	1.2615	0.4634
1.1577	21.0	4620	1.2560	0.4634
1.1317	22.0	4840	1.2506	0.4634
1.1597	23.0	5060	1.2453	0.4878
1.1283	24.0	5280	1.2401	0.4878
1.1168	25.0	5500	1.2349	0.4634
1.142	26.0	5720	1.2300	0.4634
1.1324	27.0	5940	1.2251	0.4634
1.1074	28.0	6160	1.2203	0.4634
1.107	29.0	6380	1.2157	0.4634
1.098	30.0	6600	1.2113	0.4634
1.1034	31.0	6820	1.2071	0.4634
1.0941	32.0	7040	1.2031	0.4634
1.0839	33.0	7260	1.1993	0.4634
1.0528	34.0	7480	1.1956	0.4634
1.0292	35.0	7700	1.1922	0.4634
1.0585	36.0	7920	1.1890	0.4634
1.0434	37.0	8140	1.1859	0.4634
1.0597	38.0	8360	1.1831	0.4634
1.0626	39.0	8580	1.1805	0.4634
1.0375	40.0	8800	1.1782	0.4634
1.0422	41.0	9020	1.1761	0.4634
1.0304	42.0	9240	1.1742	0.4634
1.0373	43.0	9460	1.1726	0.4878
1.0134	44.0	9680	1.1712	0.4878
1.0323	45.0	9900	1.1701	0.4878
1.0327	46.0	10120	1.1692	0.5122
1.0599	47.0	10340	1.1685	0.5122
1.0079	48.0	10560	1.1681	0.5122
1.0145	49.0	10780	1.1679	0.5122
1.0358	50.0	11000	1.1678	0.5122

Framework versions

• Transformers 4.32.1
• Pytorch 2.1.0+cu121
• Datasets 2.12.0
• Tokenizers 0.13.2