smids_5x_deit_tiny_rms_0001_fold5
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: 0.9895
- Accuracy: 0.9067
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 |
|---|---|---|---|---|
| 0.3055 | 1.0 | 375 | 0.4440 | 0.825 |
| 0.2222 | 2.0 | 750 | 0.3224 | 0.8867 |
| 0.2186 | 3.0 | 1125 | 0.3702 | 0.8883 |
| 0.1592 | 4.0 | 1500 | 0.4759 | 0.85 |
| 0.0922 | 5.0 | 1875 | 0.4560 | 0.8767 |
| 0.0977 | 6.0 | 2250 | 0.5531 | 0.875 |
| 0.0567 | 7.0 | 2625 | 0.5054 | 0.8883 |
| 0.0612 | 8.0 | 3000 | 0.5016 | 0.9067 |
| 0.0471 | 9.0 | 3375 | 0.6558 | 0.895 |
| 0.0783 | 10.0 | 3750 | 0.7144 | 0.89 |
| 0.0337 | 11.0 | 4125 | 0.7483 | 0.8833 |
| 0.0522 | 12.0 | 4500 | 0.6408 | 0.8967 |
| 0.0122 | 13.0 | 4875 | 0.5578 | 0.8917 |
| 0.0456 | 14.0 | 5250 | 0.6886 | 0.9 |
| 0.0505 | 15.0 | 5625 | 0.6222 | 0.9067 |
| 0.0186 | 16.0 | 6000 | 0.7341 | 0.8867 |
| 0.0232 | 17.0 | 6375 | 0.6650 | 0.9083 |
| 0.0384 | 18.0 | 6750 | 0.6731 | 0.9133 |
| 0.0134 | 19.0 | 7125 | 0.7917 | 0.8883 |
| 0.0197 | 20.0 | 7500 | 0.7544 | 0.9033 |
| 0.006 | 21.0 | 7875 | 0.7694 | 0.8983 |
| 0.084 | 22.0 | 8250 | 0.7873 | 0.8917 |
| 0.0405 | 23.0 | 8625 | 0.7521 | 0.8967 |
| 0.0002 | 24.0 | 9000 | 0.9409 | 0.8883 |
| 0.0009 | 25.0 | 9375 | 0.8364 | 0.8967 |
| 0.0273 | 26.0 | 9750 | 0.7668 | 0.8933 |
| 0.001 | 27.0 | 10125 | 0.7995 | 0.88 |
| 0.0262 | 28.0 | 10500 | 0.8060 | 0.8883 |
| 0.0003 | 29.0 | 10875 | 0.7588 | 0.9083 |
| 0.0189 | 30.0 | 11250 | 0.9019 | 0.8867 |
| 0.0003 | 31.0 | 11625 | 1.0397 | 0.8867 |
| 0.0 | 32.0 | 12000 | 0.9253 | 0.895 |
| 0.0002 | 33.0 | 12375 | 0.8619 | 0.905 |
| 0.0003 | 34.0 | 12750 | 0.9328 | 0.9 |
| 0.0 | 35.0 | 13125 | 0.9364 | 0.905 |
| 0.0002 | 36.0 | 13500 | 0.9470 | 0.8967 |
| 0.0001 | 37.0 | 13875 | 0.9360 | 0.9033 |
| 0.0033 | 38.0 | 14250 | 1.0063 | 0.9033 |
| 0.0 | 39.0 | 14625 | 0.9618 | 0.9017 |
| 0.0 | 40.0 | 15000 | 0.9713 | 0.9083 |
| 0.0 | 41.0 | 15375 | 0.9440 | 0.9083 |
| 0.0 | 42.0 | 15750 | 0.9330 | 0.91 |
| 0.0 | 43.0 | 16125 | 0.9519 | 0.9083 |
| 0.0 | 44.0 | 16500 | 0.9407 | 0.905 |
| 0.0 | 45.0 | 16875 | 0.9804 | 0.9033 |
| 0.0 | 46.0 | 17250 | 0.9891 | 0.9033 |
| 0.0031 | 47.0 | 17625 | 0.9794 | 0.9033 |
| 0.0 | 48.0 | 18000 | 0.9842 | 0.9033 |
| 0.0 | 49.0 | 18375 | 0.9888 | 0.9067 |
| 0.0021 | 50.0 | 18750 | 0.9895 | 0.9067 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2