hushem_40x_deit_tiny_adamax_001_fold1

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: 2.2164
• Accuracy: 0.7778

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
0.1419	1.0	215	1.1495	0.6667
0.2513	2.0	430	1.6987	0.6667
0.0735	3.0	645	1.5261	0.7556
0.0999	4.0	860	1.1841	0.7556
0.0224	5.0	1075	1.9099	0.7556
0.0347	6.0	1290	1.5424	0.7778
0.0153	7.0	1505	2.1022	0.7111
0.0692	8.0	1720	2.1279	0.6667
0.032	9.0	1935	1.4685	0.8
0.012	10.0	2150	1.8478	0.7333
0.0122	11.0	2365	2.0002	0.7333
0.0062	12.0	2580	1.8733	0.8
0.0592	13.0	2795	1.7430	0.7778
0.0	14.0	3010	1.4594	0.8
0.0388	15.0	3225	2.5836	0.6889
0.0003	16.0	3440	2.4594	0.6889
0.0001	17.0	3655	1.4002	0.8222
0.0004	18.0	3870	2.8500	0.6667
0.0028	19.0	4085	2.8669	0.6889
0.0	20.0	4300	1.9979	0.7556
0.0	21.0	4515	1.9762	0.7778
0.0	22.0	4730	1.9694	0.7778
0.0	23.0	4945	1.9654	0.7778
0.0	24.0	5160	1.9624	0.7778
0.0	25.0	5375	1.9616	0.7778
0.0	26.0	5590	1.9605	0.8
0.0	27.0	5805	1.9607	0.8
0.0	28.0	6020	1.9631	0.7778
0.0	29.0	6235	1.9640	0.7778
0.0	30.0	6450	1.9713	0.7778
0.0	31.0	6665	1.9756	0.7778
0.0	32.0	6880	1.9854	0.7778
0.0	33.0	7095	1.9940	0.7778
0.0	34.0	7310	2.0033	0.7778
0.0	35.0	7525	2.0146	0.7778
0.0	36.0	7740	2.0257	0.7778
0.0	37.0	7955	2.0418	0.7778
0.0	38.0	8170	2.0556	0.7778
0.0	39.0	8385	2.0718	0.7778
0.0	40.0	8600	2.0881	0.7778
0.0	41.0	8815	2.1047	0.7778
0.0	42.0	9030	2.1211	0.7778
0.0	43.0	9245	2.1375	0.7778
0.0	44.0	9460	2.1536	0.7778
0.0	45.0	9675	2.1694	0.7778
0.0	46.0	9890	2.1838	0.7778
0.0	47.0	10105	2.1973	0.7778
0.0	48.0	10320	2.2074	0.7778
0.0	49.0	10535	2.2141	0.7778
0.0	50.0	10750	2.2164	0.7778

Framework versions

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