hushem_40x_deit_small_rms_001_fold3
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: 3.1579
- Accuracy: 0.7442
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 |
|---|---|---|---|---|
| 1.2419 | 1.0 | 217 | 1.3981 | 0.2791 |
| 1.0235 | 2.0 | 434 | 1.3169 | 0.3953 |
| 0.8369 | 3.0 | 651 | 1.0743 | 0.4884 |
| 0.7963 | 4.0 | 868 | 0.6563 | 0.6977 |
| 0.7399 | 5.0 | 1085 | 1.1403 | 0.4651 |
| 0.591 | 6.0 | 1302 | 0.6390 | 0.7209 |
| 0.4772 | 7.0 | 1519 | 0.8818 | 0.6047 |
| 0.4582 | 8.0 | 1736 | 0.8295 | 0.6744 |
| 0.4273 | 9.0 | 1953 | 1.1233 | 0.4884 |
| 0.3402 | 10.0 | 2170 | 0.8028 | 0.7442 |
| 0.3174 | 11.0 | 2387 | 1.2880 | 0.5581 |
| 0.2909 | 12.0 | 2604 | 1.5844 | 0.6512 |
| 0.2204 | 13.0 | 2821 | 1.1940 | 0.6977 |
| 0.2639 | 14.0 | 3038 | 1.0276 | 0.6279 |
| 0.2085 | 15.0 | 3255 | 1.7122 | 0.6512 |
| 0.1551 | 16.0 | 3472 | 1.0876 | 0.7209 |
| 0.2066 | 17.0 | 3689 | 1.4826 | 0.6279 |
| 0.1259 | 18.0 | 3906 | 1.7194 | 0.6279 |
| 0.1381 | 19.0 | 4123 | 1.1881 | 0.7442 |
| 0.0864 | 20.0 | 4340 | 2.4912 | 0.7209 |
| 0.1059 | 21.0 | 4557 | 1.6650 | 0.6977 |
| 0.0958 | 22.0 | 4774 | 1.6843 | 0.6977 |
| 0.0803 | 23.0 | 4991 | 2.0214 | 0.6279 |
| 0.0716 | 24.0 | 5208 | 2.3668 | 0.6977 |
| 0.0335 | 25.0 | 5425 | 1.8384 | 0.6279 |
| 0.0722 | 26.0 | 5642 | 1.9563 | 0.6744 |
| 0.0543 | 27.0 | 5859 | 2.2739 | 0.6744 |
| 0.024 | 28.0 | 6076 | 1.7616 | 0.6977 |
| 0.0588 | 29.0 | 6293 | 1.9807 | 0.6977 |
| 0.0731 | 30.0 | 6510 | 2.0008 | 0.6279 |
| 0.0315 | 31.0 | 6727 | 2.2264 | 0.7209 |
| 0.0084 | 32.0 | 6944 | 2.2231 | 0.7674 |
| 0.0194 | 33.0 | 7161 | 2.3580 | 0.6977 |
| 0.0559 | 34.0 | 7378 | 2.5423 | 0.7209 |
| 0.0002 | 35.0 | 7595 | 2.6899 | 0.7674 |
| 0.0092 | 36.0 | 7812 | 2.7843 | 0.6744 |
| 0.0002 | 37.0 | 8029 | 2.7034 | 0.7442 |
| 0.016 | 38.0 | 8246 | 2.9844 | 0.7674 |
| 0.0006 | 39.0 | 8463 | 1.9924 | 0.8140 |
| 0.006 | 40.0 | 8680 | 2.8801 | 0.6977 |
| 0.0001 | 41.0 | 8897 | 2.7323 | 0.7674 |
| 0.0001 | 42.0 | 9114 | 3.2030 | 0.6977 |
| 0.0002 | 43.0 | 9331 | 3.6553 | 0.7674 |
| 0.0001 | 44.0 | 9548 | 2.9080 | 0.7209 |
| 0.0001 | 45.0 | 9765 | 2.8393 | 0.7442 |
| 0.0 | 46.0 | 9982 | 2.9525 | 0.7442 |
| 0.0 | 47.0 | 10199 | 3.0057 | 0.7442 |
| 0.0 | 48.0 | 10416 | 3.0880 | 0.7442 |
| 0.0 | 49.0 | 10633 | 3.1339 | 0.7442 |
| 0.0 | 50.0 | 10850 | 3.1579 | 0.7442 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2