swin-tiny-patch4-window7-224-finetuned-rsna-2018
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: 0.5745
- Accuracy: 0.7410
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: 30
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.6448 | 0.9940 | 83 | 0.6735 | 0.6737 |
| 0.736 | 2.0 | 167 | 0.6969 | 0.6557 |
| 0.6895 | 2.9940 | 250 | 0.6265 | 0.6916 |
| 0.6631 | 4.0 | 334 | 0.6275 | 0.7156 |
| 0.6725 | 4.9940 | 417 | 0.6311 | 0.7126 |
| 0.6778 | 6.0 | 501 | 0.6194 | 0.7066 |
| 0.6734 | 6.9940 | 584 | 0.6024 | 0.7141 |
| 0.6231 | 8.0 | 668 | 0.6082 | 0.7231 |
| 0.6164 | 8.9940 | 751 | 0.5846 | 0.7171 |
| 0.6261 | 10.0 | 835 | 0.5682 | 0.7380 |
| 0.6153 | 10.9940 | 918 | 0.6007 | 0.7186 |
| 0.6046 | 12.0 | 1002 | 0.5745 | 0.7410 |
| 0.5679 | 12.9940 | 1085 | 0.5957 | 0.7231 |
| 0.6027 | 14.0 | 1169 | 0.5884 | 0.7216 |
| 0.6249 | 14.9940 | 1252 | 0.5808 | 0.7365 |
| 0.6059 | 16.0 | 1336 | 0.5699 | 0.7350 |
| 0.5776 | 16.9940 | 1419 | 0.5770 | 0.7320 |
| 0.5903 | 18.0 | 1503 | 0.5806 | 0.7216 |
| 0.5633 | 18.9940 | 1586 | 0.5768 | 0.7380 |
| 0.5544 | 20.0 | 1670 | 0.5830 | 0.7350 |
| 0.5515 | 20.9940 | 1753 | 0.5966 | 0.7260 |
| 0.5249 | 22.0 | 1837 | 0.6079 | 0.7335 |
| 0.5212 | 22.9940 | 1920 | 0.5972 | 0.7246 |
| 0.5268 | 24.0 | 2004 | 0.5922 | 0.7231 |
| 0.5406 | 24.9940 | 2087 | 0.6100 | 0.7350 |
| 0.5257 | 26.0 | 2171 | 0.6004 | 0.7305 |
| 0.5152 | 26.9940 | 2254 | 0.6092 | 0.7320 |
| 0.4858 | 28.0 | 2338 | 0.6100 | 0.7231 |
| 0.5412 | 28.9940 | 2421 | 0.6116 | 0.7350 |
| 0.4972 | 29.8204 | 2490 | 0.6120 | 0.7290 |
Framework versions
- Transformers 4.43.3
- Pytorch 2.4.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1