swin-tiny-patch4-window7-224-ve-U11-b-60
This model is a fine-tuned version of microsoft/swin-tiny-patch4-window7-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.7230
- Accuracy: 0.8043
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: 5e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 60
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| No log | 0.92 | 6 | 1.3859 | 0.1304 |
| 1.3858 | 2.0 | 13 | 1.3818 | 0.2609 |
| 1.3858 | 2.92 | 19 | 1.3723 | 0.2609 |
| 1.3747 | 4.0 | 26 | 1.3355 | 0.2174 |
| 1.3001 | 4.92 | 32 | 1.2625 | 0.3696 |
| 1.3001 | 6.0 | 39 | 1.1306 | 0.4565 |
| 1.141 | 6.92 | 45 | 1.0510 | 0.4783 |
| 0.9784 | 8.0 | 52 | 0.9585 | 0.5435 |
| 0.9784 | 8.92 | 58 | 0.9895 | 0.4783 |
| 0.8533 | 10.0 | 65 | 0.9512 | 0.5 |
| 0.7564 | 10.92 | 71 | 0.9522 | 0.5217 |
| 0.7564 | 12.0 | 78 | 0.9144 | 0.5 |
| 0.6735 | 12.92 | 84 | 0.9070 | 0.6087 |
| 0.5919 | 14.0 | 91 | 0.7915 | 0.6522 |
| 0.5919 | 14.92 | 97 | 0.7989 | 0.6522 |
| 0.504 | 16.0 | 104 | 0.9510 | 0.6522 |
| 0.4422 | 16.92 | 110 | 0.8196 | 0.6739 |
| 0.4422 | 18.0 | 117 | 0.6629 | 0.7609 |
| 0.4031 | 18.92 | 123 | 0.8767 | 0.6522 |
| 0.3752 | 20.0 | 130 | 0.8253 | 0.6739 |
| 0.3752 | 20.92 | 136 | 0.7183 | 0.7391 |
| 0.3424 | 22.0 | 143 | 0.8852 | 0.6739 |
| 0.3424 | 22.92 | 149 | 0.7360 | 0.7391 |
| 0.3293 | 24.0 | 156 | 0.7230 | 0.8043 |
| 0.2822 | 24.92 | 162 | 0.8271 | 0.6957 |
| 0.2822 | 26.0 | 169 | 0.7443 | 0.8043 |
| 0.2623 | 26.92 | 175 | 0.9371 | 0.6739 |
| 0.2807 | 28.0 | 182 | 0.7392 | 0.7391 |
| 0.2807 | 28.92 | 188 | 0.8754 | 0.6739 |
| 0.223 | 30.0 | 195 | 0.7146 | 0.7826 |
| 0.2185 | 30.92 | 201 | 0.7702 | 0.7391 |
| 0.2185 | 32.0 | 208 | 0.7330 | 0.7174 |
| 0.2157 | 32.92 | 214 | 0.8817 | 0.6957 |
| 0.2011 | 34.0 | 221 | 0.7460 | 0.7174 |
| 0.2011 | 34.92 | 227 | 0.9663 | 0.6739 |
| 0.2204 | 36.0 | 234 | 0.8056 | 0.7174 |
| 0.1856 | 36.92 | 240 | 0.7799 | 0.7174 |
| 0.1856 | 38.0 | 247 | 0.8410 | 0.6957 |
| 0.1678 | 38.92 | 253 | 0.7334 | 0.7391 |
| 0.1682 | 40.0 | 260 | 0.8508 | 0.6957 |
| 0.1682 | 40.92 | 266 | 0.8106 | 0.6957 |
| 0.1638 | 42.0 | 273 | 0.8403 | 0.7174 |
| 0.1638 | 42.92 | 279 | 0.9157 | 0.6957 |
| 0.1573 | 44.0 | 286 | 0.9271 | 0.7391 |
| 0.1476 | 44.92 | 292 | 0.9167 | 0.7174 |
| 0.1476 | 46.0 | 299 | 0.9309 | 0.7174 |
| 0.1466 | 46.92 | 305 | 0.8236 | 0.7826 |
| 0.1457 | 48.0 | 312 | 0.8835 | 0.7826 |
| 0.1457 | 48.92 | 318 | 0.9162 | 0.7391 |
| 0.1625 | 50.0 | 325 | 0.8969 | 0.7391 |
| 0.1163 | 50.92 | 331 | 0.9183 | 0.7391 |
| 0.1163 | 52.0 | 338 | 0.9173 | 0.7391 |
| 0.1375 | 52.92 | 344 | 0.8886 | 0.7609 |
| 0.1379 | 54.0 | 351 | 0.8771 | 0.7391 |
| 0.1379 | 54.92 | 357 | 0.8857 | 0.7391 |
| 0.1321 | 55.38 | 360 | 0.8884 | 0.7391 |
Framework versions
- Transformers 4.36.2
- Pytorch 2.1.2+cu118
- Datasets 2.16.1
- Tokenizers 0.15.0