hushem_5x_deit_tiny_sgd_001_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: 1.0480
- 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: 0.001
- 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 |
|---|---|---|---|---|
| 1.4966 | 1.0 | 28 | 1.5748 | 0.2439 |
| 1.363 | 2.0 | 56 | 1.4510 | 0.2927 |
| 1.3445 | 3.0 | 84 | 1.3731 | 0.3902 |
| 1.2909 | 4.0 | 112 | 1.3148 | 0.3902 |
| 1.2782 | 5.0 | 140 | 1.2775 | 0.4146 |
| 1.2431 | 6.0 | 168 | 1.2527 | 0.4146 |
| 1.1698 | 7.0 | 196 | 1.2349 | 0.4634 |
| 1.1766 | 8.0 | 224 | 1.2144 | 0.4634 |
| 1.17 | 9.0 | 252 | 1.1948 | 0.4634 |
| 1.1062 | 10.0 | 280 | 1.1764 | 0.4390 |
| 1.0601 | 11.0 | 308 | 1.1840 | 0.4634 |
| 1.0566 | 12.0 | 336 | 1.1703 | 0.4634 |
| 1.0478 | 13.0 | 364 | 1.1443 | 0.4634 |
| 1.0482 | 14.0 | 392 | 1.1542 | 0.4634 |
| 1.0161 | 15.0 | 420 | 1.1465 | 0.4634 |
| 1.0335 | 16.0 | 448 | 1.1434 | 0.4634 |
| 0.9719 | 17.0 | 476 | 1.1475 | 0.4634 |
| 0.9588 | 18.0 | 504 | 1.1439 | 0.4634 |
| 1.0081 | 19.0 | 532 | 1.1431 | 0.4634 |
| 0.973 | 20.0 | 560 | 1.1304 | 0.4878 |
| 0.94 | 21.0 | 588 | 1.1093 | 0.4878 |
| 0.8982 | 22.0 | 616 | 1.1184 | 0.4878 |
| 0.9204 | 23.0 | 644 | 1.1332 | 0.4634 |
| 0.8435 | 24.0 | 672 | 1.1088 | 0.4878 |
| 0.8736 | 25.0 | 700 | 1.0913 | 0.4878 |
| 0.846 | 26.0 | 728 | 1.0897 | 0.4878 |
| 0.8446 | 27.0 | 756 | 1.0809 | 0.4878 |
| 0.8745 | 28.0 | 784 | 1.0794 | 0.4878 |
| 0.8251 | 29.0 | 812 | 1.0765 | 0.5122 |
| 0.8547 | 30.0 | 840 | 1.0870 | 0.4878 |
| 0.7939 | 31.0 | 868 | 1.0770 | 0.4878 |
| 0.7828 | 32.0 | 896 | 1.0780 | 0.4878 |
| 0.8106 | 33.0 | 924 | 1.0700 | 0.5122 |
| 0.784 | 34.0 | 952 | 1.0593 | 0.5122 |
| 0.7795 | 35.0 | 980 | 1.0615 | 0.4878 |
| 0.8007 | 36.0 | 1008 | 1.0592 | 0.4878 |
| 0.726 | 37.0 | 1036 | 1.0594 | 0.4878 |
| 0.7657 | 38.0 | 1064 | 1.0523 | 0.4878 |
| 0.7942 | 39.0 | 1092 | 1.0544 | 0.4878 |
| 0.7485 | 40.0 | 1120 | 1.0497 | 0.5122 |
| 0.7752 | 41.0 | 1148 | 1.0549 | 0.5122 |
| 0.7115 | 42.0 | 1176 | 1.0535 | 0.4878 |
| 0.7477 | 43.0 | 1204 | 1.0497 | 0.5122 |
| 0.769 | 44.0 | 1232 | 1.0484 | 0.5122 |
| 0.7292 | 45.0 | 1260 | 1.0496 | 0.5122 |
| 0.7475 | 46.0 | 1288 | 1.0482 | 0.5122 |
| 0.7629 | 47.0 | 1316 | 1.0480 | 0.5122 |
| 0.8 | 48.0 | 1344 | 1.0480 | 0.5122 |
| 0.7301 | 49.0 | 1372 | 1.0480 | 0.5122 |
| 0.738 | 50.0 | 1400 | 1.0480 | 0.5122 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0