hushem_40x_deit_small_sgd_0001_fold4
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: 1.1108
- Accuracy: 0.5238
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 |
|---|---|---|---|---|
| 1.763 | 1.0 | 219 | 1.5785 | 0.2857 |
| 1.5719 | 2.0 | 438 | 1.5003 | 0.2857 |
| 1.452 | 3.0 | 657 | 1.4620 | 0.2143 |
| 1.4006 | 4.0 | 876 | 1.4368 | 0.2143 |
| 1.3854 | 5.0 | 1095 | 1.4164 | 0.2143 |
| 1.3041 | 6.0 | 1314 | 1.3988 | 0.2381 |
| 1.296 | 7.0 | 1533 | 1.3830 | 0.2619 |
| 1.276 | 8.0 | 1752 | 1.3685 | 0.2381 |
| 1.2474 | 9.0 | 1971 | 1.3546 | 0.2381 |
| 1.2128 | 10.0 | 2190 | 1.3420 | 0.2381 |
| 1.2113 | 11.0 | 2409 | 1.3297 | 0.2381 |
| 1.2121 | 12.0 | 2628 | 1.3176 | 0.2619 |
| 1.1861 | 13.0 | 2847 | 1.3062 | 0.2619 |
| 1.1756 | 14.0 | 3066 | 1.2946 | 0.3095 |
| 1.1431 | 15.0 | 3285 | 1.2837 | 0.3571 |
| 1.1487 | 16.0 | 3504 | 1.2730 | 0.3095 |
| 1.1705 | 17.0 | 3723 | 1.2625 | 0.3095 |
| 1.1482 | 18.0 | 3942 | 1.2522 | 0.2857 |
| 1.1037 | 19.0 | 4161 | 1.2421 | 0.3095 |
| 1.0872 | 20.0 | 4380 | 1.2325 | 0.3810 |
| 1.1026 | 21.0 | 4599 | 1.2229 | 0.4048 |
| 1.0517 | 22.0 | 4818 | 1.2135 | 0.4048 |
| 1.0226 | 23.0 | 5037 | 1.2052 | 0.4286 |
| 1.0485 | 24.0 | 5256 | 1.1974 | 0.4286 |
| 1.0319 | 25.0 | 5475 | 1.1896 | 0.4286 |
| 0.9983 | 26.0 | 5694 | 1.1821 | 0.4286 |
| 1.0014 | 27.0 | 5913 | 1.1755 | 0.4048 |
| 1.0162 | 28.0 | 6132 | 1.1694 | 0.4048 |
| 0.986 | 29.0 | 6351 | 1.1635 | 0.4048 |
| 0.9747 | 30.0 | 6570 | 1.1582 | 0.4286 |
| 0.9811 | 31.0 | 6789 | 1.1532 | 0.4286 |
| 0.9907 | 32.0 | 7008 | 1.1482 | 0.4286 |
| 0.9904 | 33.0 | 7227 | 1.1437 | 0.4286 |
| 0.9293 | 34.0 | 7446 | 1.1399 | 0.4524 |
| 0.9752 | 35.0 | 7665 | 1.1362 | 0.4524 |
| 0.9789 | 36.0 | 7884 | 1.1326 | 0.4762 |
| 0.9516 | 37.0 | 8103 | 1.1293 | 0.5 |
| 0.9703 | 38.0 | 8322 | 1.1262 | 0.5 |
| 0.8944 | 39.0 | 8541 | 1.1236 | 0.5238 |
| 0.9388 | 40.0 | 8760 | 1.1213 | 0.5238 |
| 0.9573 | 41.0 | 8979 | 1.1191 | 0.5238 |
| 0.9441 | 42.0 | 9198 | 1.1172 | 0.5238 |
| 0.9438 | 43.0 | 9417 | 1.1156 | 0.5238 |
| 0.9221 | 44.0 | 9636 | 1.1141 | 0.5238 |
| 0.9079 | 45.0 | 9855 | 1.1130 | 0.5238 |
| 0.962 | 46.0 | 10074 | 1.1121 | 0.5238 |
| 0.9464 | 47.0 | 10293 | 1.1114 | 0.5238 |
| 0.9323 | 48.0 | 10512 | 1.1110 | 0.5238 |
| 0.9581 | 49.0 | 10731 | 1.1108 | 0.5238 |
| 0.942 | 50.0 | 10950 | 1.1108 | 0.5238 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2