resnet-Alzheimer

This model is a fine-tuned version of microsoft/resnet-50 on Falah/Alzheimer_MRI dataset. It achieves the following results on the evaluation set:

• Loss: 0.0932
• Accuracy: 0.9795

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.002
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• 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
1.0127	1.0	80	0.9888	0.5088
0.9345	2.0	160	0.9422	0.5303
0.8889	3.0	240	0.8724	0.5781
0.8843	4.0	320	0.8536	0.5889
0.8397	5.0	400	0.8354	0.6152
0.8624	6.0	480	0.9221	0.5381
0.7543	7.0	560	0.7568	0.6475
0.6993	8.0	640	0.8830	0.6133
0.7045	9.0	720	0.7373	0.6582
0.6557	10.0	800	0.6076	0.7451
0.5876	11.0	880	0.7281	0.6992
0.5732	12.0	960	0.5769	0.7510
0.4864	13.0	1040	0.4457	0.8311
0.5175	14.0	1120	0.5278	0.7842
0.4865	15.0	1200	0.4164	0.8379
0.4049	16.0	1280	0.4204	0.8301
0.4167	17.0	1360	0.4720	0.8281
0.36	18.0	1440	0.4660	0.8164
0.3195	19.0	1520	0.3064	0.8770
0.3652	20.0	1600	0.2571	0.9121
0.2794	21.0	1680	0.2450	0.9150
0.2704	22.0	1760	0.2391	0.9033
0.2612	23.0	1840	0.2352	0.9277
0.2425	24.0	1920	0.4720	0.8281
0.2567	25.0	2000	0.2296	0.9131
0.2302	26.0	2080	0.3067	0.8945
0.2358	27.0	2160	0.1776	0.9375
0.2173	28.0	2240	0.1596	0.9492
0.1798	29.0	2320	0.1548	0.9414
0.197	30.0	2400	0.1740	0.9570
0.1654	31.0	2480	0.1217	0.9668
0.1896	32.0	2560	0.2552	0.9258
0.1705	33.0	2640	0.1031	0.9727
0.1689	34.0	2720	0.1011	0.9688
0.1439	35.0	2800	0.1175	0.9648
0.1606	36.0	2880	0.1805	0.9443
0.1281	37.0	2960	0.1254	0.9678
0.1518	38.0	3040	0.1184	0.9648
0.1531	39.0	3120	0.0992	0.9736
0.132	40.0	3200	0.0920	0.9775
0.134	41.0	3280	0.1391	0.9639
0.1413	42.0	3360	0.1122	0.9717
0.1097	43.0	3440	0.1171	0.9678
0.1167	44.0	3520	0.1054	0.9766
0.1388	45.0	3600	0.0932	0.9795
0.1221	46.0	3680	0.0946	0.9766
0.1099	47.0	3760	0.1116	0.9756
0.1041	48.0	3840	0.1126	0.9746
0.1025	49.0	3920	0.1114	0.9756
0.0887	50.0	4000	0.1056	0.9756

Framework versions

• Transformers 4.38.2
• Pytorch 2.1.0+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2