vit-base-brain-mri-dementia-detection

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1089
• Accuracy: 0.9789

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: 0.0002
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.8826	0.3125	100	0.9027	0.575
0.8908	0.625	200	0.8484	0.5984
0.8229	0.9375	300	0.7514	0.6695
0.5299	1.25	400	0.6798	0.7164
0.5207	1.5625	500	0.6466	0.7375
0.4967	1.875	600	0.6303	0.7461
0.3977	2.1875	700	0.7240	0.7719
0.2744	2.5	800	0.3544	0.8734
0.4271	2.8125	900	0.3037	0.8938
0.2484	3.125	1000	0.4111	0.8602
0.0797	3.4375	1100	0.3782	0.8953
0.0662	3.75	1200	0.3096	0.9172
0.0894	4.0625	1300	0.2818	0.9289
0.1005	4.375	1400	0.2164	0.9469
0.0997	4.6875	1500	0.3378	0.9109
0.0715	5.0	1600	0.3627	0.9133
0.0567	5.3125	1700	0.3061	0.9234
0.0558	5.625	1800	0.2393	0.9461
0.0061	5.9375	1900	0.1738	0.9586
0.0449	6.25	2000	0.2094	0.9492
0.0073	6.5625	2100	0.1834	0.9539
0.0425	6.875	2200	0.2847	0.9266
0.0397	7.1875	2300	0.4031	0.9125
0.0284	7.5	2400	0.2995	0.9406
0.0158	7.8125	2500	0.1909	0.9664
0.006	8.125	2600	0.3524	0.9297
0.0017	8.4375	2700	0.1908	0.9617
0.0026	8.75	2800	0.1787	0.9625
0.001	9.0625	2900	0.1329	0.9688
0.0497	9.375	3000	0.1878	0.9594
0.09	9.6875	3100	0.1754	0.9648
0.0046	10.0	3200	0.1584	0.9672
0.0006	10.3125	3300	0.2008	0.9648
0.0008	10.625	3400	0.1272	0.975
0.028	10.9375	3500	0.1453	0.9766
0.0005	11.25	3600	0.1256	0.975
0.0005	11.5625	3700	0.1089	0.9789
0.0004	11.875	3800	0.1098	0.9781
0.0003	12.1875	3900	0.1779	0.9625
0.0163	12.5	4000	0.2500	0.9539
0.0003	12.8125	4100	0.1556	0.9734
0.0003	13.125	4200	0.1205	0.9742
0.0002	13.4375	4300	0.1543	0.9719
0.0002	13.75	4400	0.1548	0.975
0.0003	14.0625	4500	0.1497	0.975
0.0002	14.375	4600	0.2317	0.9641
0.0003	14.6875	4700	0.1418	0.9781
0.0002	15.0	4800	0.1537	0.9734
0.0002	15.3125	4900	0.1426	0.9781
0.0002	15.625	5000	0.1253	0.9820
0.0002	15.9375	5100	0.1128	0.9836
0.0002	16.25	5200	0.1246	0.9805
0.0002	16.5625	5300	0.1137	0.9828
0.0001	16.875	5400	0.1101	0.9844
0.0001	17.1875	5500	0.1112	0.9844
0.0001	17.5	5600	0.1121	0.9844
0.0001	17.8125	5700	0.1129	0.9836
0.0001	18.125	5800	0.1135	0.9844
0.0001	18.4375	5900	0.1140	0.9844
0.0001	18.75	6000	0.1146	0.9844
0.0001	19.0625	6100	0.1150	0.9844
0.0001	19.375	6200	0.1153	0.9844
0.0001	19.6875	6300	0.1155	0.9844
0.0001	20.0	6400	0.1155	0.9844

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1