hushem_1x_deit_tiny_adamax_001_fold3
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: 0.5538
- Accuracy: 0.8372
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 |
|---|---|---|---|---|
| No log | 1.0 | 6 | 1.3426 | 0.4419 |
| 1.3195 | 2.0 | 12 | 1.0931 | 0.5116 |
| 1.3195 | 3.0 | 18 | 0.8535 | 0.6512 |
| 0.6419 | 4.0 | 24 | 0.9249 | 0.6279 |
| 0.325 | 5.0 | 30 | 0.7057 | 0.7674 |
| 0.325 | 6.0 | 36 | 0.5831 | 0.7674 |
| 0.0848 | 7.0 | 42 | 0.6810 | 0.7907 |
| 0.0848 | 8.0 | 48 | 0.5917 | 0.7674 |
| 0.0193 | 9.0 | 54 | 0.6267 | 0.8140 |
| 0.0077 | 10.0 | 60 | 0.4330 | 0.8372 |
| 0.0077 | 11.0 | 66 | 0.5195 | 0.8372 |
| 0.0032 | 12.0 | 72 | 0.6710 | 0.7907 |
| 0.0032 | 13.0 | 78 | 0.6980 | 0.8372 |
| 0.0012 | 14.0 | 84 | 0.5701 | 0.8372 |
| 0.0006 | 15.0 | 90 | 0.5278 | 0.8605 |
| 0.0006 | 16.0 | 96 | 0.5226 | 0.8372 |
| 0.0005 | 17.0 | 102 | 0.5245 | 0.8605 |
| 0.0005 | 18.0 | 108 | 0.5277 | 0.8605 |
| 0.0004 | 19.0 | 114 | 0.5338 | 0.8372 |
| 0.0003 | 20.0 | 120 | 0.5401 | 0.8372 |
| 0.0003 | 21.0 | 126 | 0.5445 | 0.8372 |
| 0.0003 | 22.0 | 132 | 0.5461 | 0.8372 |
| 0.0003 | 23.0 | 138 | 0.5481 | 0.8372 |
| 0.0003 | 24.0 | 144 | 0.5486 | 0.8372 |
| 0.0003 | 25.0 | 150 | 0.5495 | 0.8372 |
| 0.0003 | 26.0 | 156 | 0.5492 | 0.8372 |
| 0.0002 | 27.0 | 162 | 0.5497 | 0.8372 |
| 0.0002 | 28.0 | 168 | 0.5490 | 0.8372 |
| 0.0002 | 29.0 | 174 | 0.5497 | 0.8372 |
| 0.0002 | 30.0 | 180 | 0.5498 | 0.8372 |
| 0.0002 | 31.0 | 186 | 0.5499 | 0.8372 |
| 0.0002 | 32.0 | 192 | 0.5503 | 0.8372 |
| 0.0002 | 33.0 | 198 | 0.5508 | 0.8372 |
| 0.0002 | 34.0 | 204 | 0.5520 | 0.8372 |
| 0.0002 | 35.0 | 210 | 0.5527 | 0.8372 |
| 0.0002 | 36.0 | 216 | 0.5529 | 0.8372 |
| 0.0002 | 37.0 | 222 | 0.5532 | 0.8372 |
| 0.0002 | 38.0 | 228 | 0.5534 | 0.8372 |
| 0.0002 | 39.0 | 234 | 0.5536 | 0.8372 |
| 0.0002 | 40.0 | 240 | 0.5537 | 0.8372 |
| 0.0002 | 41.0 | 246 | 0.5538 | 0.8372 |
| 0.0002 | 42.0 | 252 | 0.5538 | 0.8372 |
| 0.0002 | 43.0 | 258 | 0.5538 | 0.8372 |
| 0.0002 | 44.0 | 264 | 0.5538 | 0.8372 |
| 0.0002 | 45.0 | 270 | 0.5538 | 0.8372 |
| 0.0002 | 46.0 | 276 | 0.5538 | 0.8372 |
| 0.0002 | 47.0 | 282 | 0.5538 | 0.8372 |
| 0.0002 | 48.0 | 288 | 0.5538 | 0.8372 |
| 0.0002 | 49.0 | 294 | 0.5538 | 0.8372 |
| 0.0002 | 50.0 | 300 | 0.5538 | 0.8372 |
Framework versions
- Transformers 4.35.0
- Pytorch 2.1.0+cu118
- Datasets 2.14.6
- Tokenizers 0.14.1