hushem_5x_deit_tiny_rms_001_fold4
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: 1.2679
- Accuracy: 0.6905
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 |
|---|---|---|---|---|
| 2.1506 | 1.0 | 28 | 2.1514 | 0.2381 |
| 1.4805 | 2.0 | 56 | 1.6187 | 0.2619 |
| 1.4792 | 3.0 | 84 | 1.5112 | 0.2619 |
| 1.5148 | 4.0 | 112 | 1.3546 | 0.3095 |
| 1.3804 | 5.0 | 140 | 1.3723 | 0.4286 |
| 1.4296 | 6.0 | 168 | 1.1490 | 0.4048 |
| 1.1847 | 7.0 | 196 | 1.3299 | 0.4524 |
| 1.1564 | 8.0 | 224 | 1.0799 | 0.4762 |
| 1.0992 | 9.0 | 252 | 1.1631 | 0.5 |
| 1.0863 | 10.0 | 280 | 1.1300 | 0.4524 |
| 1.0126 | 11.0 | 308 | 0.9131 | 0.5 |
| 1.0272 | 12.0 | 336 | 0.9239 | 0.5 |
| 0.9747 | 13.0 | 364 | 0.9521 | 0.6667 |
| 0.9219 | 14.0 | 392 | 0.8729 | 0.7619 |
| 0.8522 | 15.0 | 420 | 0.6286 | 0.7381 |
| 0.8968 | 16.0 | 448 | 0.8515 | 0.6429 |
| 0.8266 | 17.0 | 476 | 0.8301 | 0.6429 |
| 0.8581 | 18.0 | 504 | 1.0046 | 0.5476 |
| 0.8265 | 19.0 | 532 | 0.8082 | 0.6429 |
| 0.8594 | 20.0 | 560 | 0.8196 | 0.6190 |
| 0.7439 | 21.0 | 588 | 0.7591 | 0.6190 |
| 0.7899 | 22.0 | 616 | 0.8303 | 0.5952 |
| 0.8223 | 23.0 | 644 | 0.6299 | 0.7143 |
| 0.8203 | 24.0 | 672 | 0.7361 | 0.7143 |
| 0.7414 | 25.0 | 700 | 0.7251 | 0.7143 |
| 0.6879 | 26.0 | 728 | 0.8771 | 0.6905 |
| 0.8008 | 27.0 | 756 | 0.8469 | 0.5714 |
| 0.7402 | 28.0 | 784 | 0.6058 | 0.7857 |
| 0.7223 | 29.0 | 812 | 0.8210 | 0.6905 |
| 0.7302 | 30.0 | 840 | 0.8614 | 0.7143 |
| 0.7098 | 31.0 | 868 | 0.9312 | 0.7143 |
| 0.7044 | 32.0 | 896 | 0.8159 | 0.7143 |
| 0.7096 | 33.0 | 924 | 0.9197 | 0.6905 |
| 0.6854 | 34.0 | 952 | 0.8631 | 0.6190 |
| 0.7442 | 35.0 | 980 | 0.8324 | 0.6667 |
| 0.6271 | 36.0 | 1008 | 0.8632 | 0.7381 |
| 0.6052 | 37.0 | 1036 | 0.8753 | 0.7143 |
| 0.6189 | 38.0 | 1064 | 1.0917 | 0.7381 |
| 0.5817 | 39.0 | 1092 | 0.9635 | 0.6429 |
| 0.5324 | 40.0 | 1120 | 1.0245 | 0.6667 |
| 0.5312 | 41.0 | 1148 | 1.1733 | 0.6905 |
| 0.5538 | 42.0 | 1176 | 1.0809 | 0.7143 |
| 0.4355 | 43.0 | 1204 | 1.0395 | 0.6667 |
| 0.3909 | 44.0 | 1232 | 1.1631 | 0.6667 |
| 0.301 | 45.0 | 1260 | 1.2110 | 0.6667 |
| 0.3678 | 46.0 | 1288 | 1.2357 | 0.6905 |
| 0.3355 | 47.0 | 1316 | 1.2487 | 0.7143 |
| 0.2983 | 48.0 | 1344 | 1.2713 | 0.6905 |
| 0.2527 | 49.0 | 1372 | 1.2679 | 0.6905 |
| 0.2761 | 50.0 | 1400 | 1.2679 | 0.6905 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0