smids_10x_deit_tiny_sgd_00001_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: 0.9599
- Accuracy: 0.5233
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: 1e-05
- 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 |
|---|---|---|---|---|
| 1.3812 | 1.0 | 750 | 1.2940 | 0.36 |
| 1.3175 | 2.0 | 1500 | 1.2389 | 0.3683 |
| 1.2691 | 3.0 | 2250 | 1.1984 | 0.3733 |
| 1.221 | 4.0 | 3000 | 1.1690 | 0.385 |
| 1.1712 | 5.0 | 3750 | 1.1479 | 0.3967 |
| 1.1485 | 6.0 | 4500 | 1.1321 | 0.4067 |
| 1.0967 | 7.0 | 5250 | 1.1194 | 0.3917 |
| 1.0947 | 8.0 | 6000 | 1.1088 | 0.4017 |
| 1.1331 | 9.0 | 6750 | 1.0995 | 0.405 |
| 1.0758 | 10.0 | 7500 | 1.0911 | 0.4167 |
| 1.0859 | 11.0 | 8250 | 1.0832 | 0.4117 |
| 1.074 | 12.0 | 9000 | 1.0758 | 0.4217 |
| 1.0354 | 13.0 | 9750 | 1.0688 | 0.4233 |
| 1.0611 | 14.0 | 10500 | 1.0620 | 0.4217 |
| 1.0504 | 15.0 | 11250 | 1.0556 | 0.425 |
| 1.0195 | 16.0 | 12000 | 1.0495 | 0.4367 |
| 1.0374 | 17.0 | 12750 | 1.0437 | 0.4383 |
| 1.0062 | 18.0 | 13500 | 1.0380 | 0.4433 |
| 1.0602 | 19.0 | 14250 | 1.0326 | 0.45 |
| 1.024 | 20.0 | 15000 | 1.0275 | 0.4533 |
| 0.9853 | 21.0 | 15750 | 1.0225 | 0.4567 |
| 1.024 | 22.0 | 16500 | 1.0178 | 0.46 |
| 1.0062 | 23.0 | 17250 | 1.0132 | 0.4617 |
| 0.9775 | 24.0 | 18000 | 1.0089 | 0.4683 |
| 0.9615 | 25.0 | 18750 | 1.0048 | 0.4733 |
| 0.9865 | 26.0 | 19500 | 1.0008 | 0.4783 |
| 0.9677 | 27.0 | 20250 | 0.9971 | 0.4867 |
| 0.9698 | 28.0 | 21000 | 0.9935 | 0.4867 |
| 0.9829 | 29.0 | 21750 | 0.9901 | 0.49 |
| 0.9556 | 30.0 | 22500 | 0.9870 | 0.49 |
| 0.963 | 31.0 | 23250 | 0.9840 | 0.4917 |
| 0.9489 | 32.0 | 24000 | 0.9813 | 0.495 |
| 0.9694 | 33.0 | 24750 | 0.9787 | 0.4967 |
| 0.9392 | 34.0 | 25500 | 0.9762 | 0.4967 |
| 0.9586 | 35.0 | 26250 | 0.9740 | 0.5 |
| 0.9291 | 36.0 | 27000 | 0.9720 | 0.5083 |
| 0.9064 | 37.0 | 27750 | 0.9701 | 0.5117 |
| 0.9352 | 38.0 | 28500 | 0.9684 | 0.5117 |
| 0.9164 | 39.0 | 29250 | 0.9668 | 0.5133 |
| 0.9501 | 40.0 | 30000 | 0.9654 | 0.515 |
| 0.8967 | 41.0 | 30750 | 0.9642 | 0.5167 |
| 0.9489 | 42.0 | 31500 | 0.9632 | 0.5167 |
| 0.9594 | 43.0 | 32250 | 0.9623 | 0.52 |
| 0.9042 | 44.0 | 33000 | 0.9616 | 0.5217 |
| 0.9218 | 45.0 | 33750 | 0.9610 | 0.5217 |
| 0.9234 | 46.0 | 34500 | 0.9605 | 0.5217 |
| 0.9392 | 47.0 | 35250 | 0.9602 | 0.5217 |
| 0.9497 | 48.0 | 36000 | 0.9600 | 0.525 |
| 0.9139 | 49.0 | 36750 | 0.9599 | 0.5233 |
| 0.8915 | 50.0 | 37500 | 0.9599 | 0.5233 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2