smids_3x_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: 0.2981
- Accuracy: 0.885
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 |
|---|---|---|---|---|
| 0.8586 | 1.0 | 225 | 0.8598 | 0.645 |
| 0.6573 | 2.0 | 450 | 0.6819 | 0.745 |
| 0.5284 | 3.0 | 675 | 0.5681 | 0.79 |
| 0.4678 | 4.0 | 900 | 0.5034 | 0.82 |
| 0.4555 | 5.0 | 1125 | 0.4634 | 0.835 |
| 0.3848 | 6.0 | 1350 | 0.4333 | 0.8467 |
| 0.3445 | 7.0 | 1575 | 0.4134 | 0.85 |
| 0.3821 | 8.0 | 1800 | 0.3974 | 0.8533 |
| 0.3848 | 9.0 | 2025 | 0.3892 | 0.86 |
| 0.352 | 10.0 | 2250 | 0.3850 | 0.8483 |
| 0.3444 | 11.0 | 2475 | 0.3683 | 0.86 |
| 0.3393 | 12.0 | 2700 | 0.3678 | 0.8583 |
| 0.3465 | 13.0 | 2925 | 0.3528 | 0.865 |
| 0.3329 | 14.0 | 3150 | 0.3471 | 0.865 |
| 0.2839 | 15.0 | 3375 | 0.3437 | 0.865 |
| 0.3134 | 16.0 | 3600 | 0.3375 | 0.865 |
| 0.3254 | 17.0 | 3825 | 0.3357 | 0.865 |
| 0.2941 | 18.0 | 4050 | 0.3300 | 0.8717 |
| 0.2779 | 19.0 | 4275 | 0.3259 | 0.8767 |
| 0.2907 | 20.0 | 4500 | 0.3243 | 0.88 |
| 0.2541 | 21.0 | 4725 | 0.3245 | 0.875 |
| 0.2729 | 22.0 | 4950 | 0.3234 | 0.8717 |
| 0.2394 | 23.0 | 5175 | 0.3164 | 0.8867 |
| 0.24 | 24.0 | 5400 | 0.3164 | 0.8767 |
| 0.2295 | 25.0 | 5625 | 0.3132 | 0.8783 |
| 0.2317 | 26.0 | 5850 | 0.3134 | 0.8767 |
| 0.2077 | 27.0 | 6075 | 0.3127 | 0.875 |
| 0.2292 | 28.0 | 6300 | 0.3093 | 0.885 |
| 0.2441 | 29.0 | 6525 | 0.3091 | 0.885 |
| 0.2145 | 30.0 | 6750 | 0.3086 | 0.8767 |
| 0.2324 | 31.0 | 6975 | 0.3048 | 0.8833 |
| 0.2227 | 32.0 | 7200 | 0.3048 | 0.8817 |
| 0.245 | 33.0 | 7425 | 0.3041 | 0.8833 |
| 0.214 | 34.0 | 7650 | 0.3030 | 0.8833 |
| 0.1813 | 35.0 | 7875 | 0.3036 | 0.8817 |
| 0.2664 | 36.0 | 8100 | 0.3024 | 0.8833 |
| 0.1941 | 37.0 | 8325 | 0.3004 | 0.8883 |
| 0.2199 | 38.0 | 8550 | 0.3022 | 0.8833 |
| 0.2069 | 39.0 | 8775 | 0.2984 | 0.885 |
| 0.1869 | 40.0 | 9000 | 0.2998 | 0.8833 |
| 0.2101 | 41.0 | 9225 | 0.2995 | 0.8833 |
| 0.2258 | 42.0 | 9450 | 0.2989 | 0.8883 |
| 0.2128 | 43.0 | 9675 | 0.2987 | 0.8867 |
| 0.2283 | 44.0 | 9900 | 0.2982 | 0.8867 |
| 0.2441 | 45.0 | 10125 | 0.2992 | 0.8833 |
| 0.1971 | 46.0 | 10350 | 0.2984 | 0.885 |
| 0.221 | 47.0 | 10575 | 0.2978 | 0.8867 |
| 0.2053 | 48.0 | 10800 | 0.2980 | 0.885 |
| 0.2465 | 49.0 | 11025 | 0.2980 | 0.885 |
| 0.2188 | 50.0 | 11250 | 0.2981 | 0.885 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2