smids_1x_deit_tiny_sgd_001_fold3
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: 0.3753
- Accuracy: 0.8433
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 |
|---|---|---|---|---|
| 1.1 | 1.0 | 75 | 1.0685 | 0.46 |
| 0.9367 | 2.0 | 150 | 0.9646 | 0.5417 |
| 0.8514 | 3.0 | 225 | 0.8770 | 0.6067 |
| 0.8379 | 4.0 | 300 | 0.8018 | 0.6683 |
| 0.6855 | 5.0 | 375 | 0.7375 | 0.715 |
| 0.6563 | 6.0 | 450 | 0.6796 | 0.7433 |
| 0.6233 | 7.0 | 525 | 0.6334 | 0.7633 |
| 0.5422 | 8.0 | 600 | 0.5966 | 0.7817 |
| 0.5439 | 9.0 | 675 | 0.5698 | 0.795 |
| 0.5302 | 10.0 | 750 | 0.5521 | 0.795 |
| 0.503 | 11.0 | 825 | 0.5277 | 0.8083 |
| 0.4486 | 12.0 | 900 | 0.5133 | 0.8083 |
| 0.4811 | 13.0 | 975 | 0.4988 | 0.8133 |
| 0.4014 | 14.0 | 1050 | 0.4857 | 0.815 |
| 0.3711 | 15.0 | 1125 | 0.4759 | 0.8217 |
| 0.4033 | 16.0 | 1200 | 0.4684 | 0.8167 |
| 0.4414 | 17.0 | 1275 | 0.4593 | 0.8233 |
| 0.3659 | 18.0 | 1350 | 0.4514 | 0.8233 |
| 0.3656 | 19.0 | 1425 | 0.4448 | 0.8233 |
| 0.4021 | 20.0 | 1500 | 0.4388 | 0.825 |
| 0.4189 | 21.0 | 1575 | 0.4325 | 0.8267 |
| 0.3524 | 22.0 | 1650 | 0.4276 | 0.8283 |
| 0.356 | 23.0 | 1725 | 0.4217 | 0.8283 |
| 0.3419 | 24.0 | 1800 | 0.4174 | 0.835 |
| 0.3528 | 25.0 | 1875 | 0.4128 | 0.83 |
| 0.3819 | 26.0 | 1950 | 0.4110 | 0.8233 |
| 0.3203 | 27.0 | 2025 | 0.4057 | 0.835 |
| 0.3243 | 28.0 | 2100 | 0.4027 | 0.8333 |
| 0.3594 | 29.0 | 2175 | 0.3998 | 0.835 |
| 0.3291 | 30.0 | 2250 | 0.3976 | 0.84 |
| 0.3185 | 31.0 | 2325 | 0.3939 | 0.84 |
| 0.3337 | 32.0 | 2400 | 0.3916 | 0.8383 |
| 0.3127 | 33.0 | 2475 | 0.3899 | 0.8367 |
| 0.3111 | 34.0 | 2550 | 0.3887 | 0.835 |
| 0.3206 | 35.0 | 2625 | 0.3862 | 0.8383 |
| 0.3237 | 36.0 | 2700 | 0.3848 | 0.845 |
| 0.2644 | 37.0 | 2775 | 0.3830 | 0.85 |
| 0.3198 | 38.0 | 2850 | 0.3820 | 0.845 |
| 0.3031 | 39.0 | 2925 | 0.3811 | 0.8417 |
| 0.2823 | 40.0 | 3000 | 0.3801 | 0.8433 |
| 0.2812 | 41.0 | 3075 | 0.3793 | 0.8417 |
| 0.3147 | 42.0 | 3150 | 0.3783 | 0.8417 |
| 0.2862 | 43.0 | 3225 | 0.3776 | 0.845 |
| 0.2684 | 44.0 | 3300 | 0.3773 | 0.8433 |
| 0.306 | 45.0 | 3375 | 0.3765 | 0.8467 |
| 0.2738 | 46.0 | 3450 | 0.3761 | 0.845 |
| 0.2493 | 47.0 | 3525 | 0.3758 | 0.8433 |
| 0.2906 | 48.0 | 3600 | 0.3756 | 0.8433 |
| 0.2971 | 49.0 | 3675 | 0.3754 | 0.8433 |
| 0.2597 | 50.0 | 3750 | 0.3753 | 0.8433 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0