hushem_1x_deit_small_adamax_00001_fold4

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

• Loss: 0.7508
• Accuracy: 0.6667

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: 1e-05
• 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	1.3350	0.3571
1.346	2.0	12	1.2810	0.3810
1.346	3.0	18	1.2346	0.4048
1.107	4.0	24	1.1917	0.4048
0.9637	5.0	30	1.1623	0.3571
0.9637	6.0	36	1.1357	0.4048
0.8241	7.0	42	1.1137	0.4286
0.8241	8.0	48	1.0906	0.4286
0.6746	9.0	54	1.0721	0.4286
0.594	10.0	60	1.0502	0.4286
0.594	11.0	66	1.0303	0.4286
0.4897	12.0	72	1.0072	0.4524
0.4897	13.0	78	0.9837	0.4762
0.4223	14.0	84	0.9800	0.4762
0.3482	15.0	90	0.9580	0.5
0.3482	16.0	96	0.9315	0.5238
0.2808	17.0	102	0.9182	0.5238
0.2808	18.0	108	0.9032	0.5714
0.2441	19.0	114	0.8918	0.6190
0.2119	20.0	120	0.8729	0.6190
0.2119	21.0	126	0.8574	0.6190
0.1699	22.0	132	0.8454	0.6190
0.1699	23.0	138	0.8308	0.6190
0.1443	24.0	144	0.8166	0.6190
0.1255	25.0	150	0.8066	0.6905
0.1255	26.0	156	0.8088	0.6905
0.1078	27.0	162	0.7901	0.6905
0.1078	28.0	168	0.7892	0.6667
0.094	29.0	174	0.7900	0.6667
0.0785	30.0	180	0.7761	0.6667
0.0785	31.0	186	0.7673	0.6667
0.071	32.0	192	0.7632	0.6667
0.071	33.0	198	0.7572	0.6667
0.066	34.0	204	0.7549	0.6667
0.0595	35.0	210	0.7582	0.6667
0.0595	36.0	216	0.7573	0.6667
0.0553	37.0	222	0.7569	0.6667
0.0553	38.0	228	0.7526	0.6667
0.0524	39.0	234	0.7502	0.6667
0.0501	40.0	240	0.7502	0.6667
0.0501	41.0	246	0.7508	0.6667
0.0507	42.0	252	0.7508	0.6667
0.0507	43.0	258	0.7508	0.6667
0.0466	44.0	264	0.7508	0.6667
0.0501	45.0	270	0.7508	0.6667
0.0501	46.0	276	0.7508	0.6667
0.0512	47.0	282	0.7508	0.6667
0.0512	48.0	288	0.7508	0.6667
0.0478	49.0	294	0.7508	0.6667
0.0501	50.0	300	0.7508	0.6667

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1