smids_3x_deit_tiny_rms_0001_fold2
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: 1.0507
- Accuracy: 0.8885
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 |
|---|---|---|---|---|
| 0.3654 | 1.0 | 225 | 0.3820 | 0.8419 |
| 0.2564 | 2.0 | 450 | 0.3888 | 0.8502 |
| 0.2078 | 3.0 | 675 | 0.3453 | 0.8669 |
| 0.1779 | 4.0 | 900 | 0.3342 | 0.8785 |
| 0.0923 | 5.0 | 1125 | 0.4468 | 0.8702 |
| 0.1133 | 6.0 | 1350 | 0.4712 | 0.8885 |
| 0.1075 | 7.0 | 1575 | 0.5119 | 0.8785 |
| 0.0546 | 8.0 | 1800 | 0.5949 | 0.8852 |
| 0.107 | 9.0 | 2025 | 0.6816 | 0.8619 |
| 0.0417 | 10.0 | 2250 | 0.6436 | 0.8918 |
| 0.0434 | 11.0 | 2475 | 0.6287 | 0.8918 |
| 0.0459 | 12.0 | 2700 | 0.7263 | 0.8802 |
| 0.01 | 13.0 | 2925 | 1.0463 | 0.8586 |
| 0.0718 | 14.0 | 3150 | 0.7632 | 0.8686 |
| 0.0139 | 15.0 | 3375 | 0.8074 | 0.8752 |
| 0.0175 | 16.0 | 3600 | 0.9064 | 0.8819 |
| 0.0398 | 17.0 | 3825 | 0.8900 | 0.8569 |
| 0.0628 | 18.0 | 4050 | 0.8666 | 0.8769 |
| 0.0021 | 19.0 | 4275 | 1.0191 | 0.8636 |
| 0.0486 | 20.0 | 4500 | 0.9743 | 0.8619 |
| 0.0363 | 21.0 | 4725 | 0.8658 | 0.8636 |
| 0.0187 | 22.0 | 4950 | 0.8042 | 0.8802 |
| 0.061 | 23.0 | 5175 | 0.9235 | 0.8735 |
| 0.0107 | 24.0 | 5400 | 0.9113 | 0.8752 |
| 0.0112 | 25.0 | 5625 | 1.0185 | 0.8785 |
| 0.0001 | 26.0 | 5850 | 0.9687 | 0.8636 |
| 0.0001 | 27.0 | 6075 | 0.8990 | 0.8686 |
| 0.0 | 28.0 | 6300 | 0.8022 | 0.8735 |
| 0.0001 | 29.0 | 6525 | 0.9932 | 0.8752 |
| 0.0056 | 30.0 | 6750 | 0.9438 | 0.8785 |
| 0.0025 | 31.0 | 6975 | 0.8626 | 0.8719 |
| 0.0001 | 32.0 | 7200 | 0.8253 | 0.8835 |
| 0.0037 | 33.0 | 7425 | 0.8896 | 0.8952 |
| 0.0001 | 34.0 | 7650 | 0.8932 | 0.8785 |
| 0.0037 | 35.0 | 7875 | 0.9625 | 0.8885 |
| 0.0037 | 36.0 | 8100 | 0.9054 | 0.8869 |
| 0.0 | 37.0 | 8325 | 0.9088 | 0.8802 |
| 0.0 | 38.0 | 8550 | 1.0141 | 0.8719 |
| 0.0067 | 39.0 | 8775 | 1.0333 | 0.8902 |
| 0.0 | 40.0 | 9000 | 0.9904 | 0.8802 |
| 0.0007 | 41.0 | 9225 | 1.0454 | 0.8802 |
| 0.0 | 42.0 | 9450 | 1.0162 | 0.8835 |
| 0.0 | 43.0 | 9675 | 1.0365 | 0.8819 |
| 0.0 | 44.0 | 9900 | 1.0455 | 0.8819 |
| 0.0 | 45.0 | 10125 | 1.0251 | 0.8852 |
| 0.0 | 46.0 | 10350 | 1.0400 | 0.8902 |
| 0.0 | 47.0 | 10575 | 1.0402 | 0.8869 |
| 0.0 | 48.0 | 10800 | 1.0455 | 0.8852 |
| 0.0025 | 49.0 | 11025 | 1.0501 | 0.8885 |
| 0.0025 | 50.0 | 11250 | 1.0507 | 0.8885 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2