smids_10x_deit_tiny_adamax_001_fold5
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: 0.8586
- Accuracy: 0.915
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: 0.001
- 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 |
|---|---|---|---|---|
| 0.3782 | 1.0 | 750 | 0.3344 | 0.8667 |
| 0.2904 | 2.0 | 1500 | 0.3574 | 0.8483 |
| 0.2048 | 3.0 | 2250 | 0.3230 | 0.8817 |
| 0.2 | 4.0 | 3000 | 0.3479 | 0.8933 |
| 0.2233 | 5.0 | 3750 | 0.3431 | 0.8883 |
| 0.1334 | 6.0 | 4500 | 0.3350 | 0.9017 |
| 0.1268 | 7.0 | 5250 | 0.3335 | 0.8967 |
| 0.077 | 8.0 | 6000 | 0.4549 | 0.8883 |
| 0.0723 | 9.0 | 6750 | 0.3771 | 0.9067 |
| 0.0426 | 10.0 | 7500 | 0.4455 | 0.9017 |
| 0.0977 | 11.0 | 8250 | 0.4334 | 0.9067 |
| 0.0237 | 12.0 | 9000 | 0.5437 | 0.9 |
| 0.0358 | 13.0 | 9750 | 0.5148 | 0.885 |
| 0.0032 | 14.0 | 10500 | 0.6045 | 0.9083 |
| 0.0293 | 15.0 | 11250 | 0.6394 | 0.8933 |
| 0.0156 | 16.0 | 12000 | 0.6836 | 0.89 |
| 0.0548 | 17.0 | 12750 | 0.5770 | 0.9017 |
| 0.0127 | 18.0 | 13500 | 0.6663 | 0.8983 |
| 0.0203 | 19.0 | 14250 | 0.6791 | 0.905 |
| 0.0154 | 20.0 | 15000 | 0.6990 | 0.905 |
| 0.0128 | 21.0 | 15750 | 0.7251 | 0.9017 |
| 0.0003 | 22.0 | 16500 | 0.7324 | 0.8933 |
| 0.0024 | 23.0 | 17250 | 0.7123 | 0.9017 |
| 0.0015 | 24.0 | 18000 | 0.6502 | 0.9133 |
| 0.0109 | 25.0 | 18750 | 0.6676 | 0.9117 |
| 0.0004 | 26.0 | 19500 | 0.6984 | 0.9033 |
| 0.0105 | 27.0 | 20250 | 0.8181 | 0.8967 |
| 0.0029 | 28.0 | 21000 | 0.7764 | 0.9 |
| 0.0304 | 29.0 | 21750 | 0.7986 | 0.8967 |
| 0.008 | 30.0 | 22500 | 0.8233 | 0.895 |
| 0.0008 | 31.0 | 23250 | 0.8494 | 0.9033 |
| 0.0 | 32.0 | 24000 | 0.8041 | 0.91 |
| 0.0 | 33.0 | 24750 | 0.8842 | 0.9167 |
| 0.0 | 34.0 | 25500 | 0.7437 | 0.9233 |
| 0.0 | 35.0 | 26250 | 0.7405 | 0.925 |
| 0.0 | 36.0 | 27000 | 0.7962 | 0.9083 |
| 0.0059 | 37.0 | 27750 | 0.7867 | 0.9233 |
| 0.0 | 38.0 | 28500 | 0.8151 | 0.92 |
| 0.0 | 39.0 | 29250 | 0.8010 | 0.91 |
| 0.0 | 40.0 | 30000 | 0.8483 | 0.9133 |
| 0.0 | 41.0 | 30750 | 0.8225 | 0.9167 |
| 0.0 | 42.0 | 31500 | 0.8207 | 0.9167 |
| 0.0 | 43.0 | 32250 | 0.8290 | 0.915 |
| 0.0 | 44.0 | 33000 | 0.8408 | 0.915 |
| 0.0 | 45.0 | 33750 | 0.8374 | 0.9183 |
| 0.0 | 46.0 | 34500 | 0.8446 | 0.9167 |
| 0.0 | 47.0 | 35250 | 0.8518 | 0.915 |
| 0.0 | 48.0 | 36000 | 0.8526 | 0.915 |
| 0.0 | 49.0 | 36750 | 0.8568 | 0.9167 |
| 0.0 | 50.0 | 37500 | 0.8586 | 0.915 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2