smids_10x_deit_small_sgd_00001_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: 0.9331
- Accuracy: 0.5933
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 |
|---|---|---|---|---|
| 1.0687 | 1.0 | 750 | 1.0724 | 0.425 |
| 1.0435 | 2.0 | 1500 | 1.0669 | 0.42 |
| 1.0439 | 3.0 | 2250 | 1.0614 | 0.4283 |
| 1.0595 | 4.0 | 3000 | 1.0559 | 0.435 |
| 1.0216 | 5.0 | 3750 | 1.0506 | 0.4383 |
| 1.0179 | 6.0 | 4500 | 1.0454 | 0.4467 |
| 1.0048 | 7.0 | 5250 | 1.0402 | 0.4517 |
| 1.0171 | 8.0 | 6000 | 1.0351 | 0.4533 |
| 1.0075 | 9.0 | 6750 | 1.0302 | 0.4567 |
| 0.9942 | 10.0 | 7500 | 1.0255 | 0.4683 |
| 0.9968 | 11.0 | 8250 | 1.0209 | 0.4783 |
| 0.9853 | 12.0 | 9000 | 1.0163 | 0.485 |
| 0.9829 | 13.0 | 9750 | 1.0118 | 0.4933 |
| 0.9676 | 14.0 | 10500 | 1.0075 | 0.4933 |
| 0.9869 | 15.0 | 11250 | 1.0033 | 0.5 |
| 0.9385 | 16.0 | 12000 | 0.9992 | 0.5117 |
| 0.9422 | 17.0 | 12750 | 0.9953 | 0.5167 |
| 0.9475 | 18.0 | 13500 | 0.9914 | 0.5233 |
| 0.9706 | 19.0 | 14250 | 0.9876 | 0.5267 |
| 0.9823 | 20.0 | 15000 | 0.9840 | 0.53 |
| 0.9281 | 21.0 | 15750 | 0.9805 | 0.535 |
| 0.9429 | 22.0 | 16500 | 0.9770 | 0.54 |
| 0.9545 | 23.0 | 17250 | 0.9738 | 0.545 |
| 0.9266 | 24.0 | 18000 | 0.9706 | 0.545 |
| 0.943 | 25.0 | 18750 | 0.9675 | 0.545 |
| 0.9362 | 26.0 | 19500 | 0.9646 | 0.55 |
| 0.9017 | 27.0 | 20250 | 0.9618 | 0.5517 |
| 0.9415 | 28.0 | 21000 | 0.9592 | 0.555 |
| 0.9141 | 29.0 | 21750 | 0.9566 | 0.555 |
| 0.9329 | 30.0 | 22500 | 0.9543 | 0.5567 |
| 0.931 | 31.0 | 23250 | 0.9520 | 0.5617 |
| 0.9115 | 32.0 | 24000 | 0.9498 | 0.5633 |
| 0.9251 | 33.0 | 24750 | 0.9478 | 0.565 |
| 0.8996 | 34.0 | 25500 | 0.9460 | 0.5717 |
| 0.9232 | 35.0 | 26250 | 0.9442 | 0.5717 |
| 0.8817 | 36.0 | 27000 | 0.9427 | 0.5717 |
| 0.8794 | 37.0 | 27750 | 0.9412 | 0.575 |
| 0.8813 | 38.0 | 28500 | 0.9398 | 0.5767 |
| 0.8952 | 39.0 | 29250 | 0.9386 | 0.58 |
| 0.8846 | 40.0 | 30000 | 0.9375 | 0.5817 |
| 0.8967 | 41.0 | 30750 | 0.9366 | 0.5867 |
| 0.9065 | 42.0 | 31500 | 0.9358 | 0.5883 |
| 0.9123 | 43.0 | 32250 | 0.9351 | 0.59 |
| 0.8878 | 44.0 | 33000 | 0.9345 | 0.59 |
| 0.8772 | 45.0 | 33750 | 0.9340 | 0.59 |
| 0.9035 | 46.0 | 34500 | 0.9336 | 0.5933 |
| 0.9152 | 47.0 | 35250 | 0.9334 | 0.5933 |
| 0.8837 | 48.0 | 36000 | 0.9332 | 0.5933 |
| 0.8879 | 49.0 | 36750 | 0.9331 | 0.5933 |
| 0.8918 | 50.0 | 37500 | 0.9331 | 0.5933 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2