hushem_1x_deit_base_adamax_001_fold4

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: 2.6471
• Accuracy: 0.5952

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: 0.001
• 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
No log	1.0	6	1.4889	0.2381
1.6928	2.0	12	1.4102	0.2381
1.6928	3.0	18	1.2949	0.4048
1.4466	4.0	24	1.1224	0.4762
1.3026	5.0	30	1.3079	0.3333
1.3026	6.0	36	1.0368	0.6190
1.2798	7.0	42	1.2143	0.3571
1.2798	8.0	48	1.0732	0.5
1.1914	9.0	54	1.0366	0.5476
1.2164	10.0	60	1.0675	0.5
1.2164	11.0	66	1.0841	0.4762
1.2177	12.0	72	1.1092	0.5952
1.2177	13.0	78	0.9871	0.5238
1.183	14.0	84	1.1380	0.4762
1.2146	15.0	90	1.1128	0.5
1.2146	16.0	96	0.9957	0.5952
1.1103	17.0	102	1.0192	0.5952
1.1103	18.0	108	1.1751	0.5
1.0656	19.0	114	1.1301	0.5
1.047	20.0	120	1.1327	0.4048
1.047	21.0	126	1.2359	0.4762
0.8853	22.0	132	1.1524	0.5952
0.8853	23.0	138	1.9551	0.3095
0.7611	24.0	144	1.3513	0.5
0.7727	25.0	150	1.6490	0.5476
0.7727	26.0	156	1.0702	0.4048
0.8546	27.0	162	1.7107	0.3333
0.8546	28.0	168	1.3302	0.4286
0.695	29.0	174	1.1947	0.5714
0.4593	30.0	180	1.8330	0.4762
0.4593	31.0	186	1.6031	0.5952
0.2978	32.0	192	2.1238	0.6190
0.2978	33.0	198	2.3897	0.5476
0.2625	34.0	204	2.1147	0.6190
0.1062	35.0	210	2.6950	0.5
0.1062	36.0	216	2.5016	0.6190
0.0682	37.0	222	2.6327	0.5476
0.0682	38.0	228	2.5000	0.5714
0.0309	39.0	234	2.4431	0.6190
0.019	40.0	240	2.6997	0.5714
0.019	41.0	246	2.6710	0.5952
0.0078	42.0	252	2.6471	0.5952
0.0078	43.0	258	2.6471	0.5952
0.0071	44.0	264	2.6471	0.5952
0.0054	45.0	270	2.6471	0.5952
0.0054	46.0	276	2.6471	0.5952
0.0076	47.0	282	2.6471	0.5952
0.0076	48.0	288	2.6471	0.5952
0.0053	49.0	294	2.6471	0.5952
0.0059	50.0	300	2.6471	0.5952

Framework versions

• Transformers 4.35.1
• Pytorch 2.1.0+cu118
• Datasets 2.14.7
• Tokenizers 0.14.1