hushem_40x_deit_base_rms_001_fold5

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

• Loss: 3.5781
• Accuracy: 0.7561

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
1.2271	1.0	220	1.5032	0.3902
0.9545	2.0	440	1.5087	0.3902
0.8667	3.0	660	1.0714	0.4878
0.8154	4.0	880	0.7851	0.6098
0.6309	5.0	1100	1.0215	0.4878
0.5655	6.0	1320	0.8556	0.6098
0.4033	7.0	1540	0.7849	0.7073
0.3567	8.0	1760	1.1431	0.6585
0.3869	9.0	1980	0.7273	0.7561
0.2867	10.0	2200	0.9025	0.6341
0.2933	11.0	2420	1.0767	0.6829
0.2822	12.0	2640	0.9054	0.7561
0.2576	13.0	2860	1.1701	0.7073
0.1424	14.0	3080	1.2265	0.7317
0.1597	15.0	3300	1.2021	0.7317
0.0822	16.0	3520	1.5652	0.7073
0.0859	17.0	3740	1.0512	0.7561
0.1048	18.0	3960	1.9377	0.6341
0.0506	19.0	4180	1.4302	0.7561
0.0595	20.0	4400	1.2065	0.7073
0.1492	21.0	4620	1.7891	0.7073
0.0835	22.0	4840	1.5550	0.7561
0.0475	23.0	5060	1.2142	0.7317
0.0941	24.0	5280	1.4080	0.7073
0.0186	25.0	5500	1.5889	0.7561
0.0776	26.0	5720	1.8453	0.6829
0.0752	27.0	5940	1.5817	0.7805
0.0113	28.0	6160	1.6776	0.7805
0.0011	29.0	6380	2.1296	0.7317
0.0107	30.0	6600	1.9807	0.7073
0.0181	31.0	6820	1.9248	0.7073
0.0106	32.0	7040	2.5784	0.7317
0.0002	33.0	7260	1.8180	0.8049
0.0013	34.0	7480	1.5976	0.8049
0.0031	35.0	7700	1.9747	0.7317
0.0094	36.0	7920	2.4830	0.7317
0.0006	37.0	8140	2.9074	0.7561
0.0049	38.0	8360	2.6503	0.6829
0.0002	39.0	8580	2.4189	0.7561
0.0	40.0	8800	2.4124	0.7561
0.0	41.0	9020	2.5470	0.7561
0.0	42.0	9240	2.6196	0.7805
0.0	43.0	9460	2.7251	0.7805
0.0	44.0	9680	2.9457	0.7805
0.0	45.0	9900	3.1311	0.7805
0.0	46.0	10120	3.2547	0.7805
0.0	47.0	10340	3.3567	0.7317
0.0	48.0	10560	3.5689	0.7561
0.0	49.0	10780	3.5825	0.7561
0.0	50.0	11000	3.5781	0.7561

Framework versions

• Transformers 4.32.1
• Pytorch 2.1.0+cu121
• Datasets 2.12.0
• Tokenizers 0.13.2