hushem_40x_deit_tiny_f4
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: 0.1343
- Accuracy: 0.9524
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: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- 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 |
|---|---|---|---|---|
| 0.1784 | 1.0 | 109 | 0.2924 | 0.9048 |
| 0.0478 | 2.0 | 218 | 0.2362 | 0.8810 |
| 0.0048 | 2.99 | 327 | 0.2393 | 0.9286 |
| 0.0107 | 4.0 | 437 | 0.2679 | 0.8810 |
| 0.0008 | 5.0 | 546 | 0.1124 | 0.9524 |
| 0.0001 | 6.0 | 655 | 0.4513 | 0.9048 |
| 0.0 | 6.99 | 764 | 0.0770 | 0.9524 |
| 0.0 | 8.0 | 874 | 0.1185 | 0.9524 |
| 0.0 | 9.0 | 983 | 0.1295 | 0.9524 |
| 0.0 | 9.98 | 1090 | 0.1343 | 0.9524 |
Framework versions
- Transformers 4.35.0
- Pytorch 2.1.0+cu118
- Datasets 2.14.6
- Tokenizers 0.14.1