hushem_1x_deit_tiny_rms_lr001_fold4

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: 1.1726
• Accuracy: 0.4524

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	5.8408	0.2619
4.7138	2.0	12	1.8632	0.2381
4.7138	3.0	18	1.9369	0.2619
1.8439	4.0	24	1.7584	0.2381
1.6449	5.0	30	1.4723	0.2619
1.6449	6.0	36	1.7187	0.2381
1.5171	7.0	42	1.4960	0.2381
1.5171	8.0	48	1.3962	0.2619
1.4701	9.0	54	1.4942	0.2619
1.4652	10.0	60	1.3642	0.2381
1.4652	11.0	66	1.4490	0.2619
1.4547	12.0	72	1.1912	0.4524
1.4547	13.0	78	1.4737	0.2857
1.3944	14.0	84	1.2170	0.4286
1.3536	15.0	90	1.3540	0.2381
1.3536	16.0	96	1.0819	0.6190
1.2835	17.0	102	1.1640	0.4286
1.2835	18.0	108	1.2309	0.3333
1.306	19.0	114	1.3288	0.2857
1.2522	20.0	120	1.4561	0.2857
1.2522	21.0	126	1.0774	0.4762
1.2491	22.0	132	1.1807	0.4286
1.2491	23.0	138	1.1668	0.3810
1.1882	24.0	144	1.2075	0.4286
1.2028	25.0	150	1.2635	0.3333
1.2028	26.0	156	1.1653	0.3810
1.1822	27.0	162	1.1741	0.4048
1.1822	28.0	168	1.4014	0.2619
1.1086	29.0	174	1.0259	0.5476
1.1111	30.0	180	1.1225	0.5238
1.1111	31.0	186	1.1813	0.5
1.0458	32.0	192	1.1678	0.4286
1.0458	33.0	198	1.1915	0.4048
1.1348	34.0	204	1.3148	0.5
0.9776	35.0	210	1.0082	0.5238
0.9776	36.0	216	0.9144	0.6190
0.9456	37.0	222	1.0677	0.4762
0.9456	38.0	228	1.0695	0.5238
0.8714	39.0	234	1.1982	0.4762
0.8643	40.0	240	1.1143	0.4048
0.8643	41.0	246	1.1270	0.4524
0.7971	42.0	252	1.1726	0.4524
0.7971	43.0	258	1.1726	0.4524
0.7662	44.0	264	1.1726	0.4524
0.7801	45.0	270	1.1726	0.4524
0.7801	46.0	276	1.1726	0.4524
0.7773	47.0	282	1.1726	0.4524
0.7773	48.0	288	1.1726	0.4524
0.7728	49.0	294	1.1726	0.4524
0.7828	50.0	300	1.1726	0.4524

Framework versions

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