smids_10x_deit_tiny_rms_0001_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: 2.3327
- Accuracy: 0.8367
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 |
|---|---|---|---|---|
| 0.7568 | 1.0 | 750 | 0.7218 | 0.6717 |
| 0.6894 | 2.0 | 1500 | 0.6565 | 0.6583 |
| 0.6167 | 3.0 | 2250 | 0.6175 | 0.7017 |
| 0.6218 | 4.0 | 3000 | 0.6221 | 0.705 |
| 0.5637 | 5.0 | 3750 | 0.5565 | 0.7667 |
| 0.5037 | 6.0 | 4500 | 0.5481 | 0.7533 |
| 0.441 | 7.0 | 5250 | 0.5584 | 0.7667 |
| 0.5186 | 8.0 | 6000 | 0.5150 | 0.78 |
| 0.4475 | 9.0 | 6750 | 0.5478 | 0.775 |
| 0.4117 | 10.0 | 7500 | 0.5408 | 0.7867 |
| 0.4227 | 11.0 | 8250 | 0.4928 | 0.8017 |
| 0.4314 | 12.0 | 9000 | 0.5195 | 0.785 |
| 0.3606 | 13.0 | 9750 | 0.4998 | 0.8033 |
| 0.3206 | 14.0 | 10500 | 0.5323 | 0.8083 |
| 0.2628 | 15.0 | 11250 | 0.5344 | 0.79 |
| 0.2274 | 16.0 | 12000 | 0.5778 | 0.81 |
| 0.1649 | 17.0 | 12750 | 0.5559 | 0.8283 |
| 0.2197 | 18.0 | 13500 | 0.5838 | 0.8017 |
| 0.242 | 19.0 | 14250 | 0.5757 | 0.83 |
| 0.178 | 20.0 | 15000 | 0.6143 | 0.8183 |
| 0.1596 | 21.0 | 15750 | 0.6500 | 0.8267 |
| 0.1681 | 22.0 | 16500 | 0.7191 | 0.83 |
| 0.1356 | 23.0 | 17250 | 0.6652 | 0.83 |
| 0.154 | 24.0 | 18000 | 0.7572 | 0.835 |
| 0.0973 | 25.0 | 18750 | 0.7886 | 0.8283 |
| 0.1206 | 26.0 | 19500 | 0.9030 | 0.8033 |
| 0.0856 | 27.0 | 20250 | 1.0266 | 0.8083 |
| 0.0629 | 28.0 | 21000 | 0.8154 | 0.8333 |
| 0.0803 | 29.0 | 21750 | 1.0582 | 0.8133 |
| 0.0608 | 30.0 | 22500 | 1.1240 | 0.8317 |
| 0.0468 | 31.0 | 23250 | 1.1197 | 0.8183 |
| 0.0343 | 32.0 | 24000 | 1.2322 | 0.8217 |
| 0.0156 | 33.0 | 24750 | 1.3344 | 0.8367 |
| 0.0192 | 34.0 | 25500 | 1.3961 | 0.8133 |
| 0.0219 | 35.0 | 26250 | 1.5315 | 0.8033 |
| 0.0147 | 36.0 | 27000 | 1.5425 | 0.8233 |
| 0.0123 | 37.0 | 27750 | 1.6413 | 0.835 |
| 0.0089 | 38.0 | 28500 | 1.7045 | 0.8167 |
| 0.0003 | 39.0 | 29250 | 1.6054 | 0.8183 |
| 0.0102 | 40.0 | 30000 | 1.6942 | 0.825 |
| 0.0008 | 41.0 | 30750 | 1.7260 | 0.84 |
| 0.0077 | 42.0 | 31500 | 1.9643 | 0.8217 |
| 0.0048 | 43.0 | 32250 | 2.0335 | 0.825 |
| 0.0015 | 44.0 | 33000 | 2.2512 | 0.8367 |
| 0.0015 | 45.0 | 33750 | 2.1796 | 0.8333 |
| 0.0001 | 46.0 | 34500 | 2.2799 | 0.83 |
| 0.0 | 47.0 | 35250 | 2.2493 | 0.8317 |
| 0.0 | 48.0 | 36000 | 2.3177 | 0.8417 |
| 0.0 | 49.0 | 36750 | 2.3130 | 0.8317 |
| 0.0 | 50.0 | 37500 | 2.3327 | 0.8367 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2