hushem_5x_deit_small_adamax_00001_fold5
This model is a fine-tuned version of facebook/deit-small-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.6350
- Accuracy: 0.8049
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: 1e-05
- 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 |
|---|---|---|---|---|
| 1.307 | 1.0 | 28 | 1.2048 | 0.4878 |
| 1.0055 | 2.0 | 56 | 1.0344 | 0.5366 |
| 0.7917 | 3.0 | 84 | 0.8814 | 0.6829 |
| 0.5612 | 4.0 | 112 | 0.7794 | 0.6585 |
| 0.4121 | 5.0 | 140 | 0.6731 | 0.7805 |
| 0.3453 | 6.0 | 168 | 0.6198 | 0.7561 |
| 0.2136 | 7.0 | 196 | 0.5552 | 0.7805 |
| 0.1402 | 8.0 | 224 | 0.5538 | 0.7805 |
| 0.1098 | 9.0 | 252 | 0.5179 | 0.8049 |
| 0.0661 | 10.0 | 280 | 0.4716 | 0.8293 |
| 0.0459 | 11.0 | 308 | 0.4940 | 0.8049 |
| 0.0201 | 12.0 | 336 | 0.4943 | 0.7805 |
| 0.0128 | 13.0 | 364 | 0.4835 | 0.8049 |
| 0.013 | 14.0 | 392 | 0.5177 | 0.8049 |
| 0.005 | 15.0 | 420 | 0.5313 | 0.7805 |
| 0.0049 | 16.0 | 448 | 0.5255 | 0.8293 |
| 0.0033 | 17.0 | 476 | 0.5525 | 0.8049 |
| 0.0027 | 18.0 | 504 | 0.5486 | 0.8049 |
| 0.0024 | 19.0 | 532 | 0.5501 | 0.8049 |
| 0.0021 | 20.0 | 560 | 0.5689 | 0.8049 |
| 0.0017 | 21.0 | 588 | 0.5750 | 0.8049 |
| 0.0016 | 22.0 | 616 | 0.5752 | 0.8049 |
| 0.0015 | 23.0 | 644 | 0.5846 | 0.8049 |
| 0.0013 | 24.0 | 672 | 0.5888 | 0.8049 |
| 0.0012 | 25.0 | 700 | 0.5919 | 0.8049 |
| 0.0012 | 26.0 | 728 | 0.5956 | 0.8049 |
| 0.0011 | 27.0 | 756 | 0.5988 | 0.8049 |
| 0.0011 | 28.0 | 784 | 0.6017 | 0.8049 |
| 0.001 | 29.0 | 812 | 0.6080 | 0.8049 |
| 0.0009 | 30.0 | 840 | 0.6107 | 0.8049 |
| 0.0009 | 31.0 | 868 | 0.6102 | 0.8049 |
| 0.0008 | 32.0 | 896 | 0.6145 | 0.8049 |
| 0.0008 | 33.0 | 924 | 0.6168 | 0.8049 |
| 0.0008 | 34.0 | 952 | 0.6219 | 0.8049 |
| 0.0008 | 35.0 | 980 | 0.6219 | 0.8049 |
| 0.0008 | 36.0 | 1008 | 0.6245 | 0.8049 |
| 0.0007 | 37.0 | 1036 | 0.6250 | 0.8049 |
| 0.0007 | 38.0 | 1064 | 0.6281 | 0.8049 |
| 0.0007 | 39.0 | 1092 | 0.6275 | 0.8049 |
| 0.0006 | 40.0 | 1120 | 0.6308 | 0.8049 |
| 0.0007 | 41.0 | 1148 | 0.6308 | 0.8049 |
| 0.0006 | 42.0 | 1176 | 0.6332 | 0.8049 |
| 0.0006 | 43.0 | 1204 | 0.6344 | 0.8049 |
| 0.0006 | 44.0 | 1232 | 0.6352 | 0.8049 |
| 0.0006 | 45.0 | 1260 | 0.6342 | 0.8049 |
| 0.0006 | 46.0 | 1288 | 0.6344 | 0.8049 |
| 0.0006 | 47.0 | 1316 | 0.6347 | 0.8049 |
| 0.0006 | 48.0 | 1344 | 0.6350 | 0.8049 |
| 0.0006 | 49.0 | 1372 | 0.6350 | 0.8049 |
| 0.0006 | 50.0 | 1400 | 0.6350 | 0.8049 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0