smids_3x_deit_small_sgd_0001_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: 0.5879
- Accuracy: 0.775
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 |
|---|---|---|---|---|
| 1.0612 | 1.0 | 225 | 1.0496 | 0.4583 |
| 1.0331 | 2.0 | 450 | 1.0241 | 0.4917 |
| 0.9976 | 3.0 | 675 | 0.9994 | 0.5317 |
| 0.9646 | 4.0 | 900 | 0.9754 | 0.5567 |
| 0.9242 | 5.0 | 1125 | 0.9523 | 0.5833 |
| 0.9274 | 6.0 | 1350 | 0.9296 | 0.6183 |
| 0.9164 | 7.0 | 1575 | 0.9074 | 0.6367 |
| 0.9203 | 8.0 | 1800 | 0.8860 | 0.6517 |
| 0.8456 | 9.0 | 2025 | 0.8654 | 0.6783 |
| 0.8517 | 10.0 | 2250 | 0.8458 | 0.6767 |
| 0.8446 | 11.0 | 2475 | 0.8273 | 0.685 |
| 0.8321 | 12.0 | 2700 | 0.8097 | 0.6933 |
| 0.8204 | 13.0 | 2925 | 0.7928 | 0.695 |
| 0.8011 | 14.0 | 3150 | 0.7770 | 0.7017 |
| 0.737 | 15.0 | 3375 | 0.7621 | 0.7017 |
| 0.7399 | 16.0 | 3600 | 0.7486 | 0.7067 |
| 0.7567 | 17.0 | 3825 | 0.7359 | 0.715 |
| 0.7583 | 18.0 | 4050 | 0.7243 | 0.7167 |
| 0.7119 | 19.0 | 4275 | 0.7132 | 0.7233 |
| 0.6839 | 20.0 | 4500 | 0.7031 | 0.7317 |
| 0.6897 | 21.0 | 4725 | 0.6934 | 0.7317 |
| 0.6996 | 22.0 | 4950 | 0.6842 | 0.7333 |
| 0.6814 | 23.0 | 5175 | 0.6758 | 0.7417 |
| 0.6885 | 24.0 | 5400 | 0.6680 | 0.7433 |
| 0.6315 | 25.0 | 5625 | 0.6607 | 0.7417 |
| 0.6519 | 26.0 | 5850 | 0.6539 | 0.7417 |
| 0.6951 | 27.0 | 6075 | 0.6475 | 0.7467 |
| 0.6243 | 28.0 | 6300 | 0.6416 | 0.75 |
| 0.6218 | 29.0 | 6525 | 0.6361 | 0.7533 |
| 0.5941 | 30.0 | 6750 | 0.6309 | 0.7533 |
| 0.5704 | 31.0 | 6975 | 0.6263 | 0.755 |
| 0.5836 | 32.0 | 7200 | 0.6219 | 0.7583 |
| 0.6485 | 33.0 | 7425 | 0.6178 | 0.76 |
| 0.5854 | 34.0 | 7650 | 0.6142 | 0.76 |
| 0.5905 | 35.0 | 7875 | 0.6108 | 0.7617 |
| 0.5617 | 36.0 | 8100 | 0.6076 | 0.7633 |
| 0.5964 | 37.0 | 8325 | 0.6047 | 0.7683 |
| 0.5721 | 38.0 | 8550 | 0.6021 | 0.7683 |
| 0.5681 | 39.0 | 8775 | 0.5996 | 0.7683 |
| 0.5364 | 40.0 | 9000 | 0.5974 | 0.7683 |
| 0.5643 | 41.0 | 9225 | 0.5955 | 0.7683 |
| 0.6152 | 42.0 | 9450 | 0.5938 | 0.77 |
| 0.5824 | 43.0 | 9675 | 0.5924 | 0.7717 |
| 0.627 | 44.0 | 9900 | 0.5911 | 0.7733 |
| 0.5753 | 45.0 | 10125 | 0.5900 | 0.7733 |
| 0.5992 | 46.0 | 10350 | 0.5892 | 0.7733 |
| 0.6048 | 47.0 | 10575 | 0.5886 | 0.775 |
| 0.5934 | 48.0 | 10800 | 0.5882 | 0.775 |
| 0.5665 | 49.0 | 11025 | 0.5880 | 0.775 |
| 0.5873 | 50.0 | 11250 | 0.5879 | 0.775 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2