smids_1x_deit_tiny_sgd_0001_fold5
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: 0.8882
- Accuracy: 0.5917
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: 0.0001
- 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 |
|---|---|---|---|---|
| 1.2279 | 1.0 | 75 | 1.2828 | 0.36 |
| 1.1801 | 2.0 | 150 | 1.2204 | 0.3717 |
| 1.1425 | 3.0 | 225 | 1.1764 | 0.3733 |
| 1.1496 | 4.0 | 300 | 1.1467 | 0.3817 |
| 1.0862 | 5.0 | 375 | 1.1259 | 0.39 |
| 1.1317 | 6.0 | 450 | 1.1103 | 0.3983 |
| 1.0711 | 7.0 | 525 | 1.0972 | 0.4083 |
| 1.0717 | 8.0 | 600 | 1.0858 | 0.4167 |
| 1.0458 | 9.0 | 675 | 1.0754 | 0.42 |
| 1.0711 | 10.0 | 750 | 1.0656 | 0.425 |
| 1.0389 | 11.0 | 825 | 1.0563 | 0.4383 |
| 1.0272 | 12.0 | 900 | 1.0476 | 0.4467 |
| 1.0495 | 13.0 | 975 | 1.0393 | 0.4517 |
| 1.0448 | 14.0 | 1050 | 1.0308 | 0.4533 |
| 1.0339 | 15.0 | 1125 | 1.0229 | 0.4583 |
| 0.9744 | 16.0 | 1200 | 1.0150 | 0.4617 |
| 0.9857 | 17.0 | 1275 | 1.0069 | 0.47 |
| 1.0108 | 18.0 | 1350 | 0.9993 | 0.4717 |
| 0.9584 | 19.0 | 1425 | 0.9919 | 0.4717 |
| 0.9977 | 20.0 | 1500 | 0.9844 | 0.485 |
| 0.9787 | 21.0 | 1575 | 0.9775 | 0.49 |
| 0.9724 | 22.0 | 1650 | 0.9707 | 0.5067 |
| 0.9219 | 23.0 | 1725 | 0.9645 | 0.515 |
| 0.923 | 24.0 | 1800 | 0.9585 | 0.525 |
| 0.9224 | 25.0 | 1875 | 0.9527 | 0.5317 |
| 0.9312 | 26.0 | 1950 | 0.9470 | 0.5417 |
| 0.9161 | 27.0 | 2025 | 0.9417 | 0.5433 |
| 0.9574 | 28.0 | 2100 | 0.9369 | 0.5467 |
| 0.9255 | 29.0 | 2175 | 0.9322 | 0.5517 |
| 0.9146 | 30.0 | 2250 | 0.9278 | 0.555 |
| 0.9155 | 31.0 | 2325 | 0.9238 | 0.5617 |
| 0.856 | 32.0 | 2400 | 0.9200 | 0.565 |
| 0.9504 | 33.0 | 2475 | 0.9164 | 0.5717 |
| 0.9096 | 34.0 | 2550 | 0.9130 | 0.5783 |
| 0.8983 | 35.0 | 2625 | 0.9100 | 0.5817 |
| 0.8589 | 36.0 | 2700 | 0.9071 | 0.585 |
| 0.8916 | 37.0 | 2775 | 0.9044 | 0.5817 |
| 0.8984 | 38.0 | 2850 | 0.9020 | 0.585 |
| 0.8824 | 39.0 | 2925 | 0.8998 | 0.5867 |
| 0.8736 | 40.0 | 3000 | 0.8977 | 0.5867 |
| 0.8723 | 41.0 | 3075 | 0.8958 | 0.5883 |
| 0.8965 | 42.0 | 3150 | 0.8942 | 0.59 |
| 0.8854 | 43.0 | 3225 | 0.8928 | 0.59 |
| 0.8622 | 44.0 | 3300 | 0.8915 | 0.5917 |
| 0.8601 | 45.0 | 3375 | 0.8905 | 0.5917 |
| 0.8904 | 46.0 | 3450 | 0.8896 | 0.5917 |
| 0.8654 | 47.0 | 3525 | 0.8890 | 0.5917 |
| 0.8638 | 48.0 | 3600 | 0.8885 | 0.5917 |
| 0.8282 | 49.0 | 3675 | 0.8883 | 0.5917 |
| 0.8485 | 50.0 | 3750 | 0.8882 | 0.5917 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0