hushem_40x_deit_tiny_rms_0001_fold3

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

• Loss: 1.8398
• Accuracy: 0.8140

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
0.1282	1.0	217	0.8975	0.8372
0.0209	2.0	434	0.8245	0.7674
0.0218	3.0	651	0.3670	0.9302
0.0205	4.0	868	1.2586	0.8372
0.0008	5.0	1085	0.8797	0.7907
0.0606	6.0	1302	1.1984	0.8605
0.1313	7.0	1519	1.3827	0.8372
0.0283	8.0	1736	0.8068	0.8605
0.0037	9.0	1953	1.0055	0.8837
0.0058	10.0	2170	1.7904	0.8140
0.0074	11.0	2387	1.3591	0.8140
0.0197	12.0	2604	1.3843	0.8605
0.0	13.0	2821	1.1075	0.8837
0.0155	14.0	3038	1.0442	0.8837
0.0002	15.0	3255	1.5088	0.8605
0.0288	16.0	3472	0.6806	0.8605
0.0057	17.0	3689	0.9450	0.8837
0.0	18.0	3906	1.1935	0.8372
0.0	19.0	4123	1.2605	0.8605
0.0	20.0	4340	1.0286	0.8140
0.0001	21.0	4557	0.9245	0.8605
0.0039	22.0	4774	1.3627	0.8372
0.0	23.0	4991	1.4994	0.8605
0.0001	24.0	5208	1.2134	0.7907
0.0001	25.0	5425	1.0301	0.8372
0.0	26.0	5642	1.0457	0.8837
0.0	27.0	5859	1.2728	0.8140
0.0	28.0	6076	1.0821	0.8837
0.0	29.0	6293	1.1243	0.8837
0.0	30.0	6510	1.1728	0.8837
0.0	31.0	6727	1.2386	0.8605
0.0	32.0	6944	1.3089	0.8605
0.0	33.0	7161	1.3713	0.8605
0.0	34.0	7378	1.4458	0.8605
0.0	35.0	7595	1.5096	0.8605
0.0	36.0	7812	1.5439	0.8605
0.0	37.0	8029	1.5992	0.8605
0.0	38.0	8246	1.6228	0.8605
0.0	39.0	8463	1.6686	0.8372
0.0	40.0	8680	1.7133	0.8372
0.0	41.0	8897	1.7502	0.8372
0.0	42.0	9114	1.7750	0.8372
0.0	43.0	9331	1.7947	0.8372
0.0	44.0	9548	1.8093	0.8372
0.0	45.0	9765	1.8201	0.8372
0.0	46.0	9982	1.8280	0.8372
0.0	47.0	10199	1.8337	0.8372
0.0	48.0	10416	1.8373	0.8372
0.0	49.0	10633	1.8394	0.8372
0.0	50.0	10850	1.8398	0.8140

Framework versions

• Transformers 4.32.1
• Pytorch 2.1.1+cu121
• Datasets 2.12.0
• Tokenizers 0.13.2