hushem_5x_deit_base_sgd_001_fold4

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: 1.1129
• Accuracy: 0.5476

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
1.4194	1.0	28	1.4029	0.1667
1.3727	2.0	56	1.3926	0.1667
1.3518	3.0	84	1.3828	0.1905
1.3497	4.0	112	1.3744	0.1905
1.3305	5.0	140	1.3656	0.2381
1.2982	6.0	168	1.3571	0.2619
1.2935	7.0	196	1.3493	0.2619
1.2822	8.0	224	1.3412	0.2857
1.2479	9.0	252	1.3334	0.3095
1.2377	10.0	280	1.3251	0.3333
1.2239	11.0	308	1.3164	0.3095
1.2188	12.0	336	1.3079	0.3095
1.2085	13.0	364	1.2995	0.3095
1.1996	14.0	392	1.2913	0.3333
1.1629	15.0	420	1.2831	0.3810
1.1847	16.0	448	1.2748	0.4048
1.1633	17.0	476	1.2664	0.4048
1.1189	18.0	504	1.2580	0.4048
1.1351	19.0	532	1.2497	0.4286
1.1255	20.0	560	1.2415	0.4286
1.0873	21.0	588	1.2339	0.4524
1.0919	22.0	616	1.2256	0.4524
1.0941	23.0	644	1.2182	0.4524
1.0702	24.0	672	1.2106	0.4762
1.0623	25.0	700	1.2031	0.4762
1.03	26.0	728	1.1958	0.4762
1.0368	27.0	756	1.1889	0.4762
1.0301	28.0	784	1.1818	0.4762
1.0348	29.0	812	1.1756	0.5
1.0312	30.0	840	1.1689	0.5
1.0387	31.0	868	1.1627	0.5238
1.0057	32.0	896	1.1569	0.5
1.0052	33.0	924	1.1517	0.5
0.9799	34.0	952	1.1467	0.5
1.0233	35.0	980	1.1420	0.5
0.9665	36.0	1008	1.1375	0.5
0.976	37.0	1036	1.1338	0.5
0.9856	38.0	1064	1.1303	0.5
0.9687	39.0	1092	1.1270	0.5
0.9553	40.0	1120	1.1240	0.5238
0.9573	41.0	1148	1.1217	0.5476
0.9651	42.0	1176	1.1193	0.5476
0.9476	43.0	1204	1.1173	0.5476
0.938	44.0	1232	1.1158	0.5476
0.9255	45.0	1260	1.1146	0.5476
0.9569	46.0	1288	1.1137	0.5476
0.9503	47.0	1316	1.1131	0.5476
0.9115	48.0	1344	1.1129	0.5476
0.9482	49.0	1372	1.1129	0.5476
0.9401	50.0	1400	1.1129	0.5476

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0