smids_10x_deit_small_sgd_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: 0.2811
- Accuracy: 0.9083
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 |
|---|---|---|---|---|
| 0.545 | 1.0 | 750 | 0.5587 | 0.785 |
| 0.4133 | 2.0 | 1500 | 0.4211 | 0.8467 |
| 0.358 | 3.0 | 2250 | 0.3782 | 0.8633 |
| 0.3237 | 4.0 | 3000 | 0.3490 | 0.87 |
| 0.3443 | 5.0 | 3750 | 0.3305 | 0.8767 |
| 0.2928 | 6.0 | 4500 | 0.3200 | 0.8817 |
| 0.2686 | 7.0 | 5250 | 0.3122 | 0.8867 |
| 0.2534 | 8.0 | 6000 | 0.3123 | 0.885 |
| 0.2251 | 9.0 | 6750 | 0.2946 | 0.8933 |
| 0.1954 | 10.0 | 7500 | 0.2908 | 0.9 |
| 0.2504 | 11.0 | 8250 | 0.2911 | 0.8967 |
| 0.2172 | 12.0 | 9000 | 0.2849 | 0.905 |
| 0.2089 | 13.0 | 9750 | 0.2810 | 0.905 |
| 0.2631 | 14.0 | 10500 | 0.2804 | 0.905 |
| 0.2076 | 15.0 | 11250 | 0.2751 | 0.915 |
| 0.1833 | 16.0 | 12000 | 0.2763 | 0.9067 |
| 0.2051 | 17.0 | 12750 | 0.2775 | 0.905 |
| 0.1927 | 18.0 | 13500 | 0.2752 | 0.9083 |
| 0.1896 | 19.0 | 14250 | 0.2722 | 0.9117 |
| 0.193 | 20.0 | 15000 | 0.2720 | 0.905 |
| 0.1978 | 21.0 | 15750 | 0.2723 | 0.905 |
| 0.193 | 22.0 | 16500 | 0.2691 | 0.91 |
| 0.1867 | 23.0 | 17250 | 0.2706 | 0.9133 |
| 0.1588 | 24.0 | 18000 | 0.2753 | 0.9083 |
| 0.1896 | 25.0 | 18750 | 0.2771 | 0.8983 |
| 0.1697 | 26.0 | 19500 | 0.2708 | 0.9133 |
| 0.1259 | 27.0 | 20250 | 0.2702 | 0.9117 |
| 0.152 | 28.0 | 21000 | 0.2731 | 0.9083 |
| 0.1891 | 29.0 | 21750 | 0.2747 | 0.9117 |
| 0.1716 | 30.0 | 22500 | 0.2723 | 0.9083 |
| 0.1252 | 31.0 | 23250 | 0.2778 | 0.905 |
| 0.1227 | 32.0 | 24000 | 0.2742 | 0.9083 |
| 0.166 | 33.0 | 24750 | 0.2738 | 0.9017 |
| 0.1299 | 34.0 | 25500 | 0.2772 | 0.9083 |
| 0.1287 | 35.0 | 26250 | 0.2752 | 0.91 |
| 0.1172 | 36.0 | 27000 | 0.2784 | 0.9033 |
| 0.1292 | 37.0 | 27750 | 0.2763 | 0.9033 |
| 0.1686 | 38.0 | 28500 | 0.2772 | 0.9067 |
| 0.1469 | 39.0 | 29250 | 0.2777 | 0.9067 |
| 0.1673 | 40.0 | 30000 | 0.2785 | 0.9083 |
| 0.1244 | 41.0 | 30750 | 0.2779 | 0.9067 |
| 0.149 | 42.0 | 31500 | 0.2782 | 0.9067 |
| 0.1031 | 43.0 | 32250 | 0.2799 | 0.905 |
| 0.1374 | 44.0 | 33000 | 0.2832 | 0.9067 |
| 0.1179 | 45.0 | 33750 | 0.2818 | 0.905 |
| 0.1282 | 46.0 | 34500 | 0.2810 | 0.905 |
| 0.1603 | 47.0 | 35250 | 0.2819 | 0.9067 |
| 0.1237 | 48.0 | 36000 | 0.2811 | 0.9083 |
| 0.1333 | 49.0 | 36750 | 0.2808 | 0.9067 |
| 0.1344 | 50.0 | 37500 | 0.2811 | 0.9083 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2