smids_5x_deit_tiny_sgd_0001_fold3
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: 0.4818
- Accuracy: 0.8283
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 |
|---|---|---|---|---|
| 1.1768 | 1.0 | 375 | 1.1673 | 0.3967 |
| 1.0999 | 2.0 | 750 | 1.1000 | 0.4267 |
| 1.0421 | 3.0 | 1125 | 1.0522 | 0.4717 |
| 0.9929 | 4.0 | 1500 | 1.0115 | 0.5 |
| 0.9542 | 5.0 | 1875 | 0.9748 | 0.5233 |
| 0.8817 | 6.0 | 2250 | 0.9389 | 0.5417 |
| 0.8788 | 7.0 | 2625 | 0.9041 | 0.57 |
| 0.8581 | 8.0 | 3000 | 0.8691 | 0.59 |
| 0.7544 | 9.0 | 3375 | 0.8356 | 0.6033 |
| 0.7744 | 10.0 | 3750 | 0.8037 | 0.6317 |
| 0.7141 | 11.0 | 4125 | 0.7743 | 0.66 |
| 0.7231 | 12.0 | 4500 | 0.7462 | 0.6733 |
| 0.7108 | 13.0 | 4875 | 0.7217 | 0.7033 |
| 0.6884 | 14.0 | 5250 | 0.6985 | 0.7217 |
| 0.6146 | 15.0 | 5625 | 0.6770 | 0.7333 |
| 0.573 | 16.0 | 6000 | 0.6579 | 0.7467 |
| 0.6638 | 17.0 | 6375 | 0.6415 | 0.76 |
| 0.653 | 18.0 | 6750 | 0.6259 | 0.775 |
| 0.5464 | 19.0 | 7125 | 0.6119 | 0.7833 |
| 0.5679 | 20.0 | 7500 | 0.6000 | 0.79 |
| 0.5005 | 21.0 | 7875 | 0.5883 | 0.795 |
| 0.5248 | 22.0 | 8250 | 0.5790 | 0.8017 |
| 0.5865 | 23.0 | 8625 | 0.5701 | 0.8017 |
| 0.5913 | 24.0 | 9000 | 0.5617 | 0.8033 |
| 0.542 | 25.0 | 9375 | 0.5539 | 0.8033 |
| 0.5193 | 26.0 | 9750 | 0.5470 | 0.8033 |
| 0.5655 | 27.0 | 10125 | 0.5403 | 0.805 |
| 0.59 | 28.0 | 10500 | 0.5340 | 0.8117 |
| 0.4649 | 29.0 | 10875 | 0.5287 | 0.815 |
| 0.4871 | 30.0 | 11250 | 0.5240 | 0.8133 |
| 0.5615 | 31.0 | 11625 | 0.5192 | 0.815 |
| 0.5652 | 32.0 | 12000 | 0.5148 | 0.8167 |
| 0.5595 | 33.0 | 12375 | 0.5110 | 0.8183 |
| 0.4782 | 34.0 | 12750 | 0.5072 | 0.8167 |
| 0.4 | 35.0 | 13125 | 0.5039 | 0.8183 |
| 0.427 | 36.0 | 13500 | 0.5006 | 0.82 |
| 0.4646 | 37.0 | 13875 | 0.4979 | 0.82 |
| 0.4598 | 38.0 | 14250 | 0.4954 | 0.8217 |
| 0.4848 | 39.0 | 14625 | 0.4932 | 0.8233 |
| 0.4928 | 40.0 | 15000 | 0.4911 | 0.8267 |
| 0.4256 | 41.0 | 15375 | 0.4892 | 0.8267 |
| 0.4494 | 42.0 | 15750 | 0.4875 | 0.8267 |
| 0.5087 | 43.0 | 16125 | 0.4862 | 0.8267 |
| 0.4788 | 44.0 | 16500 | 0.4849 | 0.8267 |
| 0.4663 | 45.0 | 16875 | 0.4839 | 0.8267 |
| 0.4385 | 46.0 | 17250 | 0.4831 | 0.8267 |
| 0.4878 | 47.0 | 17625 | 0.4825 | 0.8283 |
| 0.4959 | 48.0 | 18000 | 0.4821 | 0.8283 |
| 0.4851 | 49.0 | 18375 | 0.4819 | 0.8283 |
| 0.4569 | 50.0 | 18750 | 0.4818 | 0.8283 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2