hushem_40x_deit_base_rms_001_fold4
This model is a fine-tuned version of facebook/deit-base-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 1.7808
- Accuracy: 0.8095
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 |
|---|---|---|---|---|
| 1.4095 | 1.0 | 219 | 1.4091 | 0.2381 |
| 1.3846 | 2.0 | 438 | 1.3865 | 0.2381 |
| 1.2802 | 3.0 | 657 | 1.3372 | 0.2381 |
| 1.1537 | 4.0 | 876 | 1.4032 | 0.2619 |
| 1.177 | 5.0 | 1095 | 1.3147 | 0.4286 |
| 1.1719 | 6.0 | 1314 | 0.9703 | 0.6667 |
| 1.0403 | 7.0 | 1533 | 1.2271 | 0.4762 |
| 0.9188 | 8.0 | 1752 | 0.9431 | 0.5714 |
| 0.8565 | 9.0 | 1971 | 1.0056 | 0.5952 |
| 0.8519 | 10.0 | 2190 | 0.7845 | 0.6429 |
| 0.7519 | 11.0 | 2409 | 0.7049 | 0.6905 |
| 0.8514 | 12.0 | 2628 | 0.6628 | 0.7857 |
| 0.8808 | 13.0 | 2847 | 0.8006 | 0.7381 |
| 0.796 | 14.0 | 3066 | 0.7332 | 0.6905 |
| 0.7213 | 15.0 | 3285 | 0.7486 | 0.6905 |
| 0.663 | 16.0 | 3504 | 0.4390 | 0.7857 |
| 0.5845 | 17.0 | 3723 | 0.9856 | 0.5952 |
| 0.5228 | 18.0 | 3942 | 0.6588 | 0.7381 |
| 0.5581 | 19.0 | 4161 | 0.6093 | 0.8571 |
| 0.518 | 20.0 | 4380 | 0.5316 | 0.6905 |
| 0.5058 | 21.0 | 4599 | 0.7052 | 0.7381 |
| 0.453 | 22.0 | 4818 | 0.6155 | 0.7143 |
| 0.4128 | 23.0 | 5037 | 0.7141 | 0.7381 |
| 0.44 | 24.0 | 5256 | 0.6896 | 0.7619 |
| 0.3933 | 25.0 | 5475 | 0.6353 | 0.7619 |
| 0.3648 | 26.0 | 5694 | 0.7225 | 0.8095 |
| 0.2677 | 27.0 | 5913 | 0.6987 | 0.8810 |
| 0.3023 | 28.0 | 6132 | 0.8143 | 0.8333 |
| 0.332 | 29.0 | 6351 | 0.8300 | 0.8333 |
| 0.2772 | 30.0 | 6570 | 0.6339 | 0.7619 |
| 0.1878 | 31.0 | 6789 | 0.6694 | 0.8333 |
| 0.2152 | 32.0 | 7008 | 0.7930 | 0.7619 |
| 0.2378 | 33.0 | 7227 | 0.7856 | 0.7619 |
| 0.1874 | 34.0 | 7446 | 0.6614 | 0.8571 |
| 0.2043 | 35.0 | 7665 | 0.7218 | 0.8095 |
| 0.122 | 36.0 | 7884 | 1.0415 | 0.8333 |
| 0.1837 | 37.0 | 8103 | 1.2016 | 0.7381 |
| 0.1148 | 38.0 | 8322 | 0.8289 | 0.7857 |
| 0.0825 | 39.0 | 8541 | 1.4711 | 0.7381 |
| 0.0828 | 40.0 | 8760 | 0.9405 | 0.8810 |
| 0.0736 | 41.0 | 8979 | 1.4104 | 0.8810 |
| 0.0864 | 42.0 | 9198 | 1.1297 | 0.8333 |
| 0.0176 | 43.0 | 9417 | 1.2293 | 0.7857 |
| 0.0392 | 44.0 | 9636 | 1.3878 | 0.8095 |
| 0.0272 | 45.0 | 9855 | 1.2021 | 0.8571 |
| 0.0125 | 46.0 | 10074 | 2.3102 | 0.7619 |
| 0.0149 | 47.0 | 10293 | 1.8621 | 0.7857 |
| 0.0032 | 48.0 | 10512 | 1.7899 | 0.8333 |
| 0.0016 | 49.0 | 10731 | 1.9528 | 0.8095 |
| 0.0001 | 50.0 | 10950 | 1.7808 | 0.8095 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2