hushem_40x_deit_tiny_adamax_001_fold4
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: 0.5391
- Accuracy: 0.9524
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 |
|---|---|---|---|---|
| 0.2113 | 1.0 | 219 | 0.5669 | 0.8571 |
| 0.1247 | 2.0 | 438 | 0.3852 | 0.8571 |
| 0.0797 | 3.0 | 657 | 0.3243 | 0.8810 |
| 0.079 | 4.0 | 876 | 0.2935 | 0.9286 |
| 0.1755 | 5.0 | 1095 | 0.3153 | 0.8810 |
| 0.1228 | 6.0 | 1314 | 0.4983 | 0.9048 |
| 0.047 | 7.0 | 1533 | 0.4737 | 0.9048 |
| 0.0236 | 8.0 | 1752 | 0.2530 | 0.9286 |
| 0.0027 | 9.0 | 1971 | 0.9366 | 0.8810 |
| 0.0257 | 10.0 | 2190 | 0.8815 | 0.8810 |
| 0.032 | 11.0 | 2409 | 0.7642 | 0.9048 |
| 0.0025 | 12.0 | 2628 | 0.6321 | 0.9286 |
| 0.0 | 13.0 | 2847 | 0.4805 | 0.9048 |
| 0.0406 | 14.0 | 3066 | 0.7911 | 0.9286 |
| 0.0286 | 15.0 | 3285 | 0.2463 | 0.9048 |
| 0.0029 | 16.0 | 3504 | 0.0537 | 0.9762 |
| 0.0065 | 17.0 | 3723 | 0.3008 | 0.9286 |
| 0.0001 | 18.0 | 3942 | 0.8021 | 0.8810 |
| 0.0 | 19.0 | 4161 | 0.3160 | 0.9762 |
| 0.0084 | 20.0 | 4380 | 1.2037 | 0.8333 |
| 0.0 | 21.0 | 4599 | 0.5426 | 0.9286 |
| 0.0001 | 22.0 | 4818 | 0.3468 | 0.9524 |
| 0.0204 | 23.0 | 5037 | 0.7324 | 0.9286 |
| 0.0 | 24.0 | 5256 | 0.8099 | 0.9048 |
| 0.0 | 25.0 | 5475 | 1.1998 | 0.8810 |
| 0.0 | 26.0 | 5694 | 0.5294 | 0.9524 |
| 0.0 | 27.0 | 5913 | 0.5383 | 0.9524 |
| 0.0 | 28.0 | 6132 | 0.5204 | 0.9524 |
| 0.0 | 29.0 | 6351 | 0.5193 | 0.9524 |
| 0.0 | 30.0 | 6570 | 0.5189 | 0.9524 |
| 0.0 | 31.0 | 6789 | 0.5187 | 0.9524 |
| 0.0 | 32.0 | 7008 | 0.5190 | 0.9524 |
| 0.0 | 33.0 | 7227 | 0.5187 | 0.9524 |
| 0.0 | 34.0 | 7446 | 0.5193 | 0.9524 |
| 0.0 | 35.0 | 7665 | 0.5201 | 0.9524 |
| 0.0 | 36.0 | 7884 | 0.5213 | 0.9524 |
| 0.0 | 37.0 | 8103 | 0.5225 | 0.9524 |
| 0.0 | 38.0 | 8322 | 0.5239 | 0.9524 |
| 0.0 | 39.0 | 8541 | 0.5256 | 0.9524 |
| 0.0 | 40.0 | 8760 | 0.5271 | 0.9524 |
| 0.0 | 41.0 | 8979 | 0.5287 | 0.9524 |
| 0.0 | 42.0 | 9198 | 0.5302 | 0.9524 |
| 0.0 | 43.0 | 9417 | 0.5318 | 0.9524 |
| 0.0 | 44.0 | 9636 | 0.5333 | 0.9524 |
| 0.0 | 45.0 | 9855 | 0.5348 | 0.9524 |
| 0.0 | 46.0 | 10074 | 0.5359 | 0.9524 |
| 0.0 | 47.0 | 10293 | 0.5372 | 0.9524 |
| 0.0 | 48.0 | 10512 | 0.5381 | 0.9524 |
| 0.0 | 49.0 | 10731 | 0.5389 | 0.9524 |
| 0.0 | 50.0 | 10950 | 0.5391 | 0.9524 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2