hushem_40x_deit_tiny_adamax_00001_fold4
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: 1.0865
- Accuracy: 0.9048
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: 1e-05
- 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 |
|---|---|---|---|---|
| 0.5149 | 1.0 | 219 | 0.5408 | 0.7857 |
| 0.1388 | 2.0 | 438 | 0.2883 | 0.8571 |
| 0.0532 | 3.0 | 657 | 0.2259 | 0.9048 |
| 0.0146 | 4.0 | 876 | 0.3103 | 0.8810 |
| 0.0044 | 5.0 | 1095 | 0.2128 | 0.9048 |
| 0.001 | 6.0 | 1314 | 0.4066 | 0.8571 |
| 0.0004 | 7.0 | 1533 | 0.5492 | 0.8571 |
| 0.0003 | 8.0 | 1752 | 0.5191 | 0.8571 |
| 0.0002 | 9.0 | 1971 | 0.5554 | 0.8571 |
| 0.0002 | 10.0 | 2190 | 0.6021 | 0.8571 |
| 0.0001 | 11.0 | 2409 | 0.6325 | 0.8571 |
| 0.0001 | 12.0 | 2628 | 0.5941 | 0.8810 |
| 0.0001 | 13.0 | 2847 | 0.6178 | 0.8810 |
| 0.0 | 14.0 | 3066 | 0.6345 | 0.8810 |
| 0.0 | 15.0 | 3285 | 0.6789 | 0.8810 |
| 0.0 | 16.0 | 3504 | 0.6912 | 0.8810 |
| 0.0 | 17.0 | 3723 | 0.6975 | 0.8810 |
| 0.0 | 18.0 | 3942 | 0.7160 | 0.8810 |
| 0.0 | 19.0 | 4161 | 0.7194 | 0.8810 |
| 0.0 | 20.0 | 4380 | 0.7354 | 0.8810 |
| 0.0 | 21.0 | 4599 | 0.7292 | 0.9048 |
| 0.0 | 22.0 | 4818 | 0.7594 | 0.9048 |
| 0.0 | 23.0 | 5037 | 0.7524 | 0.9048 |
| 0.0 | 24.0 | 5256 | 0.7681 | 0.9048 |
| 0.0 | 25.0 | 5475 | 0.7964 | 0.9048 |
| 0.0 | 26.0 | 5694 | 0.8348 | 0.9048 |
| 0.0 | 27.0 | 5913 | 0.8454 | 0.9048 |
| 0.0 | 28.0 | 6132 | 0.8650 | 0.9048 |
| 0.0 | 29.0 | 6351 | 0.8560 | 0.9048 |
| 0.0 | 30.0 | 6570 | 0.8777 | 0.9048 |
| 0.0 | 31.0 | 6789 | 0.8901 | 0.9048 |
| 0.0 | 32.0 | 7008 | 0.9135 | 0.9048 |
| 0.0 | 33.0 | 7227 | 0.9102 | 0.9048 |
| 0.0 | 34.0 | 7446 | 0.9561 | 0.9048 |
| 0.0 | 35.0 | 7665 | 0.9681 | 0.9048 |
| 0.0 | 36.0 | 7884 | 0.9813 | 0.9048 |
| 0.0 | 37.0 | 8103 | 0.9769 | 0.9048 |
| 0.0 | 38.0 | 8322 | 1.0135 | 0.9048 |
| 0.0 | 39.0 | 8541 | 1.0218 | 0.9048 |
| 0.0 | 40.0 | 8760 | 1.0098 | 0.9048 |
| 0.0 | 41.0 | 8979 | 1.0382 | 0.9048 |
| 0.0 | 42.0 | 9198 | 1.0217 | 0.9048 |
| 0.0 | 43.0 | 9417 | 1.0481 | 0.9048 |
| 0.0 | 44.0 | 9636 | 1.0751 | 0.9048 |
| 0.0 | 45.0 | 9855 | 1.0579 | 0.9048 |
| 0.0 | 46.0 | 10074 | 1.0662 | 0.9048 |
| 0.0 | 47.0 | 10293 | 1.0827 | 0.9048 |
| 0.0 | 48.0 | 10512 | 1.0853 | 0.9048 |
| 0.0 | 49.0 | 10731 | 1.0917 | 0.9048 |
| 0.0 | 50.0 | 10950 | 1.0865 | 0.9048 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2