hushem_40x_deit_tiny_sgd_001_fold5

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: 0.5908
• Accuracy: 0.8293

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
1.1602	1.0	220	1.4008	0.2927
0.8954	2.0	440	1.2274	0.3902
0.7859	3.0	660	1.0703	0.5366
0.6718	4.0	880	0.9455	0.6098
0.5505	5.0	1100	0.8399	0.6341
0.4372	6.0	1320	0.7728	0.7073
0.3616	7.0	1540	0.7172	0.7317
0.291	8.0	1760	0.7018	0.7317
0.2597	9.0	1980	0.6678	0.7317
0.2339	10.0	2200	0.6575	0.7073
0.2227	11.0	2420	0.6389	0.7073
0.179	12.0	2640	0.6500	0.7073
0.1598	13.0	2860	0.6290	0.7073
0.1448	14.0	3080	0.6491	0.6585
0.1209	15.0	3300	0.6174	0.7073
0.1192	16.0	3520	0.6084	0.7073
0.1037	17.0	3740	0.6013	0.7317
0.0848	18.0	3960	0.5985	0.7073
0.1048	19.0	4180	0.5896	0.7317
0.0665	20.0	4400	0.6043	0.7073
0.0723	21.0	4620	0.5932	0.7561
0.0444	22.0	4840	0.5749	0.8049
0.0448	23.0	5060	0.5862	0.7805
0.0396	24.0	5280	0.5758	0.8049
0.0378	25.0	5500	0.5566	0.8293
0.0428	26.0	5720	0.5740	0.8293
0.0345	27.0	5940	0.5631	0.8049
0.0515	28.0	6160	0.5844	0.8049
0.0324	29.0	6380	0.5872	0.8293
0.0292	30.0	6600	0.5789	0.8293
0.0208	31.0	6820	0.5688	0.8293
0.0421	32.0	7040	0.5703	0.8293
0.0246	33.0	7260	0.5663	0.8293
0.0318	34.0	7480	0.5726	0.8293
0.0151	35.0	7700	0.5751	0.8293
0.0169	36.0	7920	0.5772	0.8293
0.017	37.0	8140	0.5665	0.8293
0.0393	38.0	8360	0.5815	0.8293
0.0218	39.0	8580	0.5765	0.8293
0.0156	40.0	8800	0.5742	0.8293
0.0183	41.0	9020	0.5956	0.8293
0.0155	42.0	9240	0.5886	0.8293
0.0134	43.0	9460	0.5775	0.8293
0.0186	44.0	9680	0.5921	0.8293
0.0177	45.0	9900	0.5863	0.8293
0.0115	46.0	10120	0.5918	0.8293
0.0196	47.0	10340	0.5892	0.8293
0.0172	48.0	10560	0.5892	0.8293
0.0129	49.0	10780	0.5910	0.8293
0.0197	50.0	11000	0.5908	0.8293

Framework versions

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