hushem_5x_deit_tiny_sgd_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: 1.4406
• Accuracy: 0.2558

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
1.5476	1.0	28	1.6881	0.2791
1.5009	2.0	56	1.6672	0.2791
1.493	3.0	84	1.6491	0.2791
1.4757	4.0	112	1.6326	0.2791
1.4183	5.0	140	1.6168	0.2791
1.4727	6.0	168	1.6027	0.2791
1.5064	7.0	196	1.5899	0.2791
1.4575	8.0	224	1.5786	0.2791
1.4566	9.0	252	1.5679	0.3023
1.4332	10.0	280	1.5586	0.3023
1.4461	11.0	308	1.5502	0.3023
1.4527	12.0	336	1.5422	0.3023
1.4102	13.0	364	1.5344	0.3023
1.4234	14.0	392	1.5271	0.3023
1.4638	15.0	420	1.5205	0.3023
1.4171	16.0	448	1.5148	0.3023
1.3787	17.0	476	1.5087	0.2791
1.4195	18.0	504	1.5032	0.2791
1.3909	19.0	532	1.4981	0.3256
1.4469	20.0	560	1.4935	0.3023
1.382	21.0	588	1.4891	0.3023
1.3548	22.0	616	1.4852	0.3023
1.4115	23.0	644	1.4815	0.3023
1.3595	24.0	672	1.4779	0.2791
1.4648	25.0	700	1.4744	0.2791
1.3584	26.0	728	1.4712	0.2791
1.3694	27.0	756	1.4682	0.2791
1.3704	28.0	784	1.4656	0.2791
1.3747	29.0	812	1.4631	0.2791
1.3528	30.0	840	1.4609	0.2791
1.3372	31.0	868	1.4586	0.2791
1.3782	32.0	896	1.4565	0.2791
1.3746	33.0	924	1.4545	0.2791
1.3597	34.0	952	1.4525	0.2791
1.3491	35.0	980	1.4509	0.2791
1.3872	36.0	1008	1.4493	0.2791
1.3595	37.0	1036	1.4478	0.2791
1.3401	38.0	1064	1.4465	0.2791
1.3573	39.0	1092	1.4454	0.2791
1.3488	40.0	1120	1.4444	0.2791
1.3842	41.0	1148	1.4435	0.2791
1.3433	42.0	1176	1.4428	0.2558
1.3592	43.0	1204	1.4421	0.2558
1.3773	44.0	1232	1.4415	0.2558
1.3285	45.0	1260	1.4411	0.2558
1.3374	46.0	1288	1.4408	0.2558
1.3383	47.0	1316	1.4407	0.2558
1.3567	48.0	1344	1.4406	0.2558
1.3494	49.0	1372	1.4406	0.2558
1.2617	50.0	1400	1.4406	0.2558

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0