hushem_40x_deit_tiny_sgd_00001_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: 1.4228
- Accuracy: 0.2791
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 |
|---|---|---|---|---|
| 1.4808 | 1.0 | 217 | 1.4816 | 0.2791 |
| 1.4823 | 2.0 | 434 | 1.4791 | 0.2791 |
| 1.4134 | 3.0 | 651 | 1.4766 | 0.2791 |
| 1.4759 | 4.0 | 868 | 1.4742 | 0.2791 |
| 1.4883 | 5.0 | 1085 | 1.4718 | 0.2791 |
| 1.4518 | 6.0 | 1302 | 1.4695 | 0.3023 |
| 1.4499 | 7.0 | 1519 | 1.4671 | 0.2791 |
| 1.4363 | 8.0 | 1736 | 1.4648 | 0.2791 |
| 1.4639 | 9.0 | 1953 | 1.4626 | 0.2791 |
| 1.447 | 10.0 | 2170 | 1.4604 | 0.2791 |
| 1.4636 | 11.0 | 2387 | 1.4583 | 0.3023 |
| 1.4249 | 12.0 | 2604 | 1.4562 | 0.3023 |
| 1.4551 | 13.0 | 2821 | 1.4542 | 0.3023 |
| 1.4299 | 14.0 | 3038 | 1.4523 | 0.2791 |
| 1.4254 | 15.0 | 3255 | 1.4505 | 0.2791 |
| 1.3712 | 16.0 | 3472 | 1.4487 | 0.2791 |
| 1.4294 | 17.0 | 3689 | 1.4469 | 0.2791 |
| 1.3982 | 18.0 | 3906 | 1.4452 | 0.2791 |
| 1.39 | 19.0 | 4123 | 1.4437 | 0.2791 |
| 1.3895 | 20.0 | 4340 | 1.4422 | 0.2791 |
| 1.3897 | 21.0 | 4557 | 1.4407 | 0.2791 |
| 1.381 | 22.0 | 4774 | 1.4393 | 0.2791 |
| 1.3878 | 23.0 | 4991 | 1.4380 | 0.2791 |
| 1.4255 | 24.0 | 5208 | 1.4367 | 0.2791 |
| 1.397 | 25.0 | 5425 | 1.4355 | 0.2791 |
| 1.3946 | 26.0 | 5642 | 1.4344 | 0.2791 |
| 1.4141 | 27.0 | 5859 | 1.4334 | 0.2791 |
| 1.391 | 28.0 | 6076 | 1.4324 | 0.2791 |
| 1.3772 | 29.0 | 6293 | 1.4314 | 0.2791 |
| 1.4053 | 30.0 | 6510 | 1.4305 | 0.2791 |
| 1.3414 | 31.0 | 6727 | 1.4297 | 0.2791 |
| 1.368 | 32.0 | 6944 | 1.4288 | 0.2791 |
| 1.3993 | 33.0 | 7161 | 1.4281 | 0.2791 |
| 1.3039 | 34.0 | 7378 | 1.4274 | 0.2791 |
| 1.3467 | 35.0 | 7595 | 1.4268 | 0.2791 |
| 1.3754 | 36.0 | 7812 | 1.4262 | 0.2791 |
| 1.3681 | 37.0 | 8029 | 1.4257 | 0.2791 |
| 1.3927 | 38.0 | 8246 | 1.4252 | 0.2791 |
| 1.3307 | 39.0 | 8463 | 1.4248 | 0.2791 |
| 1.3625 | 40.0 | 8680 | 1.4244 | 0.2791 |
| 1.3775 | 41.0 | 8897 | 1.4240 | 0.2791 |
| 1.3411 | 42.0 | 9114 | 1.4237 | 0.2791 |
| 1.3645 | 43.0 | 9331 | 1.4235 | 0.2791 |
| 1.3775 | 44.0 | 9548 | 1.4233 | 0.2791 |
| 1.3259 | 45.0 | 9765 | 1.4231 | 0.2791 |
| 1.3653 | 46.0 | 9982 | 1.4230 | 0.2791 |
| 1.3629 | 47.0 | 10199 | 1.4229 | 0.2791 |
| 1.3538 | 48.0 | 10416 | 1.4229 | 0.2791 |
| 1.3676 | 49.0 | 10633 | 1.4228 | 0.2791 |
| 1.357 | 50.0 | 10850 | 1.4228 | 0.2791 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2