hushem_5x_deit_tiny_adamax_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: 2.7115
- Accuracy: 0.6585
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 |
|---|---|---|---|---|
| 1.4348 | 1.0 | 28 | 1.3734 | 0.2439 |
| 1.3624 | 2.0 | 56 | 1.4229 | 0.2683 |
| 1.1348 | 3.0 | 84 | 1.1315 | 0.3902 |
| 0.9863 | 4.0 | 112 | 1.0099 | 0.6829 |
| 0.9002 | 5.0 | 140 | 0.8153 | 0.7317 |
| 0.8747 | 6.0 | 168 | 0.8078 | 0.7317 |
| 0.7431 | 7.0 | 196 | 0.8202 | 0.7073 |
| 0.7236 | 8.0 | 224 | 0.6730 | 0.7073 |
| 0.7214 | 9.0 | 252 | 0.7811 | 0.6829 |
| 0.7661 | 10.0 | 280 | 0.8373 | 0.6341 |
| 0.6997 | 11.0 | 308 | 0.7829 | 0.7073 |
| 0.5964 | 12.0 | 336 | 0.9580 | 0.5366 |
| 0.601 | 13.0 | 364 | 0.8593 | 0.6341 |
| 0.4989 | 14.0 | 392 | 0.8291 | 0.7317 |
| 0.484 | 15.0 | 420 | 0.8268 | 0.7317 |
| 0.3579 | 16.0 | 448 | 0.8735 | 0.6585 |
| 0.3201 | 17.0 | 476 | 1.3019 | 0.6341 |
| 0.2054 | 18.0 | 504 | 1.2022 | 0.6829 |
| 0.2162 | 19.0 | 532 | 1.3723 | 0.6098 |
| 0.2359 | 20.0 | 560 | 2.1538 | 0.5854 |
| 0.1213 | 21.0 | 588 | 1.4495 | 0.6829 |
| 0.1657 | 22.0 | 616 | 1.5861 | 0.6341 |
| 0.2091 | 23.0 | 644 | 1.3652 | 0.6585 |
| 0.0692 | 24.0 | 672 | 1.7622 | 0.6585 |
| 0.1092 | 25.0 | 700 | 2.0505 | 0.6585 |
| 0.0584 | 26.0 | 728 | 2.2675 | 0.5610 |
| 0.0661 | 27.0 | 756 | 1.7051 | 0.7073 |
| 0.0353 | 28.0 | 784 | 1.9468 | 0.6585 |
| 0.0164 | 29.0 | 812 | 2.4092 | 0.6341 |
| 0.0019 | 30.0 | 840 | 2.7744 | 0.6585 |
| 0.0033 | 31.0 | 868 | 3.2900 | 0.5610 |
| 0.0105 | 32.0 | 896 | 2.4900 | 0.5854 |
| 0.0008 | 33.0 | 924 | 2.5105 | 0.6341 |
| 0.0047 | 34.0 | 952 | 2.0758 | 0.7073 |
| 0.0004 | 35.0 | 980 | 2.7140 | 0.6585 |
| 0.0 | 36.0 | 1008 | 2.9025 | 0.6585 |
| 0.0013 | 37.0 | 1036 | 2.6654 | 0.6585 |
| 0.0 | 38.0 | 1064 | 2.6558 | 0.6829 |
| 0.0 | 39.0 | 1092 | 2.6667 | 0.6585 |
| 0.0 | 40.0 | 1120 | 2.6779 | 0.6585 |
| 0.0 | 41.0 | 1148 | 2.6850 | 0.6585 |
| 0.0 | 42.0 | 1176 | 2.6917 | 0.6585 |
| 0.0 | 43.0 | 1204 | 2.6986 | 0.6585 |
| 0.0 | 44.0 | 1232 | 2.7032 | 0.6585 |
| 0.0 | 45.0 | 1260 | 2.7065 | 0.6585 |
| 0.0 | 46.0 | 1288 | 2.7090 | 0.6585 |
| 0.0 | 47.0 | 1316 | 2.7107 | 0.6585 |
| 0.0 | 48.0 | 1344 | 2.7115 | 0.6585 |
| 0.0 | 49.0 | 1372 | 2.7115 | 0.6585 |
| 0.0 | 50.0 | 1400 | 2.7115 | 0.6585 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0