hushem_5x_deit_small_rms_001_fold4
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: 1.6694
- Accuracy: 0.5952
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 |
|---|---|---|---|---|
| 2.1638 | 1.0 | 28 | 1.7503 | 0.2381 |
| 1.4446 | 2.0 | 56 | 1.5611 | 0.2619 |
| 1.4481 | 3.0 | 84 | 1.4312 | 0.2381 |
| 1.3982 | 4.0 | 112 | 1.3919 | 0.2619 |
| 1.3867 | 5.0 | 140 | 1.4053 | 0.2619 |
| 1.382 | 6.0 | 168 | 1.3617 | 0.2619 |
| 1.2911 | 7.0 | 196 | 1.5439 | 0.4048 |
| 1.1486 | 8.0 | 224 | 1.1564 | 0.4286 |
| 1.0554 | 9.0 | 252 | 1.0568 | 0.4762 |
| 1.0402 | 10.0 | 280 | 0.8946 | 0.6190 |
| 0.9192 | 11.0 | 308 | 0.7214 | 0.7381 |
| 1.0116 | 12.0 | 336 | 0.8931 | 0.6905 |
| 0.9735 | 13.0 | 364 | 0.8359 | 0.6905 |
| 0.9105 | 14.0 | 392 | 0.6761 | 0.7619 |
| 0.8218 | 15.0 | 420 | 0.6339 | 0.7857 |
| 0.8745 | 16.0 | 448 | 0.7396 | 0.7619 |
| 0.8355 | 17.0 | 476 | 0.7738 | 0.7381 |
| 0.8644 | 18.0 | 504 | 0.6532 | 0.7619 |
| 0.8014 | 19.0 | 532 | 0.7016 | 0.7381 |
| 0.8685 | 20.0 | 560 | 0.7175 | 0.7381 |
| 0.7709 | 21.0 | 588 | 0.6588 | 0.7619 |
| 0.778 | 22.0 | 616 | 0.8635 | 0.7381 |
| 0.8232 | 23.0 | 644 | 0.6385 | 0.7143 |
| 0.891 | 24.0 | 672 | 0.7133 | 0.6667 |
| 0.714 | 25.0 | 700 | 0.6807 | 0.6905 |
| 0.6766 | 26.0 | 728 | 0.9128 | 0.6429 |
| 0.734 | 27.0 | 756 | 0.7515 | 0.6905 |
| 0.7087 | 28.0 | 784 | 0.6378 | 0.6905 |
| 0.6295 | 29.0 | 812 | 0.9113 | 0.6667 |
| 0.6414 | 30.0 | 840 | 0.9201 | 0.6190 |
| 0.6359 | 31.0 | 868 | 0.7354 | 0.7143 |
| 0.6485 | 32.0 | 896 | 0.6558 | 0.6429 |
| 0.6242 | 33.0 | 924 | 0.7790 | 0.6429 |
| 0.647 | 34.0 | 952 | 1.0490 | 0.5952 |
| 0.6524 | 35.0 | 980 | 0.7508 | 0.6667 |
| 0.5325 | 36.0 | 1008 | 0.9344 | 0.6667 |
| 0.476 | 37.0 | 1036 | 1.0580 | 0.5952 |
| 0.4941 | 38.0 | 1064 | 0.9380 | 0.7143 |
| 0.4232 | 39.0 | 1092 | 1.0384 | 0.5476 |
| 0.4302 | 40.0 | 1120 | 1.0844 | 0.6190 |
| 0.4057 | 41.0 | 1148 | 1.3995 | 0.5952 |
| 0.3483 | 42.0 | 1176 | 1.4823 | 0.5476 |
| 0.3043 | 43.0 | 1204 | 1.2186 | 0.6667 |
| 0.2598 | 44.0 | 1232 | 1.3028 | 0.5952 |
| 0.2113 | 45.0 | 1260 | 1.5042 | 0.6190 |
| 0.2104 | 46.0 | 1288 | 1.6174 | 0.5952 |
| 0.1769 | 47.0 | 1316 | 1.5011 | 0.6429 |
| 0.1341 | 48.0 | 1344 | 1.6784 | 0.5714 |
| 0.1239 | 49.0 | 1372 | 1.6694 | 0.5952 |
| 0.1545 | 50.0 | 1400 | 1.6694 | 0.5952 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0