smids_5x_deit_small_adamax_0001_fold5
This model is a fine-tuned version of facebook/deit-small-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.8104
- Accuracy: 0.925
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 |
|---|---|---|---|---|
| 0.1649 | 1.0 | 750 | 0.2981 | 0.8767 |
| 0.1219 | 2.0 | 1500 | 0.3351 | 0.9117 |
| 0.088 | 3.0 | 2250 | 0.4113 | 0.9117 |
| 0.0191 | 4.0 | 3000 | 0.5041 | 0.91 |
| 0.0316 | 5.0 | 3750 | 0.6431 | 0.905 |
| 0.0519 | 6.0 | 4500 | 0.6485 | 0.9017 |
| 0.0145 | 7.0 | 5250 | 0.6712 | 0.91 |
| 0.001 | 8.0 | 6000 | 0.7180 | 0.91 |
| 0.0005 | 9.0 | 6750 | 0.6201 | 0.9067 |
| 0.0001 | 10.0 | 7500 | 0.7375 | 0.92 |
| 0.009 | 11.0 | 8250 | 0.7397 | 0.9133 |
| 0.0 | 12.0 | 9000 | 0.7531 | 0.9233 |
| 0.0005 | 13.0 | 9750 | 0.7094 | 0.9167 |
| 0.0 | 14.0 | 10500 | 0.6906 | 0.9217 |
| 0.0 | 15.0 | 11250 | 0.7622 | 0.92 |
| 0.0 | 16.0 | 12000 | 0.7690 | 0.9167 |
| 0.0099 | 17.0 | 12750 | 0.7093 | 0.925 |
| 0.0049 | 18.0 | 13500 | 0.7817 | 0.9167 |
| 0.0 | 19.0 | 14250 | 0.7714 | 0.9183 |
| 0.0 | 20.0 | 15000 | 0.7423 | 0.92 |
| 0.0 | 21.0 | 15750 | 0.7472 | 0.9283 |
| 0.0 | 22.0 | 16500 | 0.8201 | 0.9217 |
| 0.0003 | 23.0 | 17250 | 0.7230 | 0.925 |
| 0.0 | 24.0 | 18000 | 0.7873 | 0.9233 |
| 0.0 | 25.0 | 18750 | 0.7903 | 0.9233 |
| 0.0 | 26.0 | 19500 | 0.7611 | 0.9233 |
| 0.0 | 27.0 | 20250 | 0.7662 | 0.9267 |
| 0.0 | 28.0 | 21000 | 0.7601 | 0.9267 |
| 0.0 | 29.0 | 21750 | 0.7659 | 0.925 |
| 0.0054 | 30.0 | 22500 | 0.7697 | 0.9217 |
| 0.0 | 31.0 | 23250 | 0.7755 | 0.9217 |
| 0.0 | 32.0 | 24000 | 0.7712 | 0.9217 |
| 0.0 | 33.0 | 24750 | 0.7599 | 0.9267 |
| 0.0 | 34.0 | 25500 | 0.7735 | 0.9267 |
| 0.0 | 35.0 | 26250 | 0.7806 | 0.925 |
| 0.0 | 36.0 | 27000 | 0.7835 | 0.9217 |
| 0.0039 | 37.0 | 27750 | 0.7879 | 0.925 |
| 0.0 | 38.0 | 28500 | 0.7885 | 0.9267 |
| 0.0 | 39.0 | 29250 | 0.7918 | 0.925 |
| 0.0 | 40.0 | 30000 | 0.7945 | 0.9267 |
| 0.0 | 41.0 | 30750 | 0.7955 | 0.9267 |
| 0.0 | 42.0 | 31500 | 0.7991 | 0.9233 |
| 0.0 | 43.0 | 32250 | 0.8003 | 0.925 |
| 0.0 | 44.0 | 33000 | 0.8023 | 0.925 |
| 0.0 | 45.0 | 33750 | 0.8041 | 0.925 |
| 0.0 | 46.0 | 34500 | 0.8060 | 0.925 |
| 0.0 | 47.0 | 35250 | 0.8084 | 0.925 |
| 0.0 | 48.0 | 36000 | 0.8088 | 0.925 |
| 0.0 | 49.0 | 36750 | 0.8102 | 0.9267 |
| 0.0 | 50.0 | 37500 | 0.8104 | 0.925 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2