hushem_40x_deit_tiny_sgd_0001_fold5

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

• Loss: 1.1237
• Accuracy: 0.5122

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: 0.0001
• 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
1.4318	1.0	220	1.6149	0.1707
1.3729	2.0	440	1.5770	0.1951
1.3561	3.0	660	1.5432	0.1707
1.3096	4.0	880	1.5177	0.1463
1.2756	5.0	1100	1.4946	0.1951
1.2485	6.0	1320	1.4738	0.2195
1.1719	7.0	1540	1.4569	0.2195
1.1324	8.0	1760	1.4401	0.2439
1.1522	9.0	1980	1.4251	0.2683
1.1548	10.0	2200	1.4097	0.2439
1.1099	11.0	2420	1.3960	0.2683
1.0799	12.0	2640	1.3821	0.2683
1.072	13.0	2860	1.3689	0.2927
1.0381	14.0	3080	1.3552	0.3171
1.0533	15.0	3300	1.3423	0.2927
1.0294	16.0	3520	1.3293	0.3171
1.004	17.0	3740	1.3169	0.3171
1.0138	18.0	3960	1.3048	0.3171
0.9902	19.0	4180	1.2935	0.3171
0.9047	20.0	4400	1.2817	0.3171
0.9213	21.0	4620	1.2707	0.3415
0.9555	22.0	4840	1.2595	0.3415
0.9607	23.0	5060	1.2491	0.3415
0.9344	24.0	5280	1.2391	0.3415
0.8688	25.0	5500	1.2295	0.3902
0.9175	26.0	5720	1.2208	0.4146
0.887	27.0	5940	1.2120	0.4390
0.905	28.0	6160	1.2036	0.4634
0.8477	29.0	6380	1.1957	0.4878
0.8486	30.0	6600	1.1887	0.4878
0.9203	31.0	6820	1.1822	0.4878
0.8893	32.0	7040	1.1760	0.4878
0.8469	33.0	7260	1.1702	0.4878
0.7935	34.0	7480	1.1645	0.4878
0.7904	35.0	7700	1.1593	0.4878
0.7994	36.0	7920	1.1544	0.5122
0.8205	37.0	8140	1.1499	0.5122
0.8696	38.0	8360	1.1458	0.5122
0.8262	39.0	8580	1.1421	0.5122
0.7584	40.0	8800	1.1388	0.5122
0.8457	41.0	9020	1.1358	0.5122
0.8307	42.0	9240	1.1331	0.5122
0.8183	43.0	9460	1.1307	0.5122
0.7718	44.0	9680	1.1287	0.5122
0.7855	45.0	9900	1.1271	0.5122
0.7875	46.0	10120	1.1258	0.5122
0.8109	47.0	10340	1.1248	0.5122
0.7297	48.0	10560	1.1241	0.5122
0.7352	49.0	10780	1.1238	0.5122
0.7935	50.0	11000	1.1237	0.5122

Framework versions

• Transformers 4.32.1
• Pytorch 2.1.1+cu121
• Datasets 2.12.0
• Tokenizers 0.13.2