smids_10x_deit_tiny_sgd_0001_fold4
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: 0.4161
- Accuracy: 0.8417
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 |
|---|---|---|---|---|
| 1.0688 | 1.0 | 750 | 1.0895 | 0.3983 |
| 0.9867 | 2.0 | 1500 | 0.9975 | 0.4833 |
| 0.9278 | 3.0 | 2250 | 0.9175 | 0.54 |
| 0.8528 | 4.0 | 3000 | 0.8474 | 0.5783 |
| 0.714 | 5.0 | 3750 | 0.7808 | 0.635 |
| 0.6866 | 6.0 | 4500 | 0.7206 | 0.6917 |
| 0.6101 | 7.0 | 5250 | 0.6718 | 0.7167 |
| 0.6452 | 8.0 | 6000 | 0.6299 | 0.7333 |
| 0.6113 | 9.0 | 6750 | 0.5989 | 0.7467 |
| 0.5055 | 10.0 | 7500 | 0.5743 | 0.7633 |
| 0.4983 | 11.0 | 8250 | 0.5553 | 0.7717 |
| 0.5538 | 12.0 | 9000 | 0.5377 | 0.7733 |
| 0.4959 | 13.0 | 9750 | 0.5236 | 0.7833 |
| 0.4737 | 14.0 | 10500 | 0.5129 | 0.7867 |
| 0.4376 | 15.0 | 11250 | 0.5024 | 0.7967 |
| 0.3926 | 16.0 | 12000 | 0.4941 | 0.8033 |
| 0.397 | 17.0 | 12750 | 0.4866 | 0.805 |
| 0.4304 | 18.0 | 13500 | 0.4793 | 0.8067 |
| 0.4526 | 19.0 | 14250 | 0.4737 | 0.81 |
| 0.4267 | 20.0 | 15000 | 0.4680 | 0.81 |
| 0.3746 | 21.0 | 15750 | 0.4626 | 0.8183 |
| 0.4237 | 22.0 | 16500 | 0.4581 | 0.815 |
| 0.4022 | 23.0 | 17250 | 0.4540 | 0.82 |
| 0.465 | 24.0 | 18000 | 0.4503 | 0.8233 |
| 0.3585 | 25.0 | 18750 | 0.4464 | 0.8267 |
| 0.3671 | 26.0 | 19500 | 0.4431 | 0.8267 |
| 0.3889 | 27.0 | 20250 | 0.4400 | 0.8283 |
| 0.3836 | 28.0 | 21000 | 0.4372 | 0.83 |
| 0.3751 | 29.0 | 21750 | 0.4351 | 0.83 |
| 0.3772 | 30.0 | 22500 | 0.4334 | 0.8333 |
| 0.3959 | 31.0 | 23250 | 0.4312 | 0.8333 |
| 0.3701 | 32.0 | 24000 | 0.4290 | 0.8317 |
| 0.3441 | 33.0 | 24750 | 0.4274 | 0.8317 |
| 0.371 | 34.0 | 25500 | 0.4262 | 0.8333 |
| 0.327 | 35.0 | 26250 | 0.4246 | 0.8333 |
| 0.3799 | 36.0 | 27000 | 0.4233 | 0.8367 |
| 0.3186 | 37.0 | 27750 | 0.4226 | 0.835 |
| 0.3955 | 38.0 | 28500 | 0.4215 | 0.835 |
| 0.4171 | 39.0 | 29250 | 0.4206 | 0.835 |
| 0.4116 | 40.0 | 30000 | 0.4196 | 0.8367 |
| 0.369 | 41.0 | 30750 | 0.4189 | 0.8383 |
| 0.3461 | 42.0 | 31500 | 0.4184 | 0.8383 |
| 0.3837 | 43.0 | 32250 | 0.4178 | 0.8417 |
| 0.3565 | 44.0 | 33000 | 0.4174 | 0.8417 |
| 0.3745 | 45.0 | 33750 | 0.4170 | 0.8417 |
| 0.3413 | 46.0 | 34500 | 0.4167 | 0.8417 |
| 0.301 | 47.0 | 35250 | 0.4164 | 0.8417 |
| 0.3105 | 48.0 | 36000 | 0.4162 | 0.8417 |
| 0.3511 | 49.0 | 36750 | 0.4161 | 0.8417 |
| 0.3221 | 50.0 | 37500 | 0.4161 | 0.8417 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2