hushem_5x_deit_base_sgd_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: 1.0680
• Accuracy: 0.5854

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
1.4163	1.0	28	1.3604	0.1951
1.3947	2.0	56	1.3473	0.2195
1.3604	3.0	84	1.3349	0.2683
1.3392	4.0	112	1.3241	0.2927
1.3444	5.0	140	1.3143	0.3171
1.3238	6.0	168	1.3050	0.3415
1.3103	7.0	196	1.2955	0.3659
1.2905	8.0	224	1.2862	0.3902
1.2713	9.0	252	1.2769	0.4146
1.2521	10.0	280	1.2674	0.4390
1.2419	11.0	308	1.2580	0.4390
1.2274	12.0	336	1.2492	0.4634
1.2017	13.0	364	1.2403	0.5122
1.2089	14.0	392	1.2314	0.5366
1.1882	15.0	420	1.2229	0.5366
1.1838	16.0	448	1.2144	0.5610
1.1566	17.0	476	1.2059	0.5610
1.1584	18.0	504	1.1980	0.6098
1.1748	19.0	532	1.1896	0.6098
1.1362	20.0	560	1.1817	0.6098
1.1338	21.0	588	1.1741	0.5854
1.1033	22.0	616	1.1667	0.5854
1.0957	23.0	644	1.1590	0.5854
1.0836	24.0	672	1.1521	0.5854
1.0697	25.0	700	1.1452	0.5610
1.078	26.0	728	1.1389	0.5366
1.0636	27.0	756	1.1332	0.5610
1.0604	28.0	784	1.1274	0.5366
1.0075	29.0	812	1.1217	0.5610
1.0554	30.0	840	1.1163	0.5610
1.0238	31.0	868	1.1110	0.5610
0.9869	32.0	896	1.1060	0.5854
0.9963	33.0	924	1.1019	0.5610
1.0156	34.0	952	1.0973	0.5854
0.9827	35.0	980	1.0931	0.5854
0.9853	36.0	1008	1.0896	0.5854
0.9677	37.0	1036	1.0862	0.5854
0.9703	38.0	1064	1.0831	0.5854
0.9924	39.0	1092	1.0803	0.5854
0.9509	40.0	1120	1.0778	0.5854
0.9744	41.0	1148	1.0755	0.5854
0.957	42.0	1176	1.0735	0.5854
0.958	43.0	1204	1.0718	0.5854
0.965	44.0	1232	1.0705	0.5854
0.9524	45.0	1260	1.0695	0.5854
0.9551	46.0	1288	1.0687	0.5854
0.9588	47.0	1316	1.0682	0.5854
0.9894	48.0	1344	1.0680	0.5854
0.9401	49.0	1372	1.0680	0.5854
0.9662	50.0	1400	1.0680	0.5854

Framework versions

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