swin-original-10
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: 0.0443
- Accuracy: 0.9916
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: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 5.6259 | 1.0 | 247 | 3.3200 | 0.3994 |
| 1.9734 | 2.0 | 494 | 0.5108 | 0.9370 |
| 0.6166 | 3.0 | 741 | 0.2288 | 0.9749 |
| 0.4348 | 4.0 | 988 | 0.1149 | 0.9858 |
| 0.2823 | 5.0 | 1235 | 0.0760 | 0.9899 |
| 0.2351 | 6.0 | 1482 | 0.0618 | 0.9906 |
| 0.1889 | 7.0 | 1729 | 0.0550 | 0.9894 |
| 0.1681 | 8.0 | 1976 | 0.0505 | 0.9901 |
| 0.144 | 9.0 | 2223 | 0.0446 | 0.9919 |
| 0.1248 | 10.0 | 2470 | 0.0443 | 0.9916 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.2.2+cu121
- Datasets 3.0.0
- Tokenizers 0.15.2
- Downloads last month
- 6
Model tree for genofog/swin-original-10
Base model
microsoft/swin-tiny-patch4-window7-224