hushem_5x_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: 1.0019
- Accuracy: 0.5581
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 |
|---|---|---|---|---|
| 1.5059 | 1.0 | 28 | 1.5596 | 0.3023 |
| 1.3905 | 2.0 | 56 | 1.4722 | 0.3023 |
| 1.3398 | 3.0 | 84 | 1.4240 | 0.3023 |
| 1.3159 | 4.0 | 112 | 1.3834 | 0.3488 |
| 1.2162 | 5.0 | 140 | 1.3509 | 0.3721 |
| 1.2552 | 6.0 | 168 | 1.3214 | 0.3721 |
| 1.2492 | 7.0 | 196 | 1.3011 | 0.3721 |
| 1.2051 | 8.0 | 224 | 1.2801 | 0.3953 |
| 1.1592 | 9.0 | 252 | 1.2612 | 0.3953 |
| 1.1549 | 10.0 | 280 | 1.2440 | 0.3953 |
| 1.1103 | 11.0 | 308 | 1.2292 | 0.3953 |
| 1.1084 | 12.0 | 336 | 1.2159 | 0.3953 |
| 1.0301 | 13.0 | 364 | 1.2021 | 0.3953 |
| 1.0455 | 14.0 | 392 | 1.1867 | 0.4419 |
| 1.0462 | 15.0 | 420 | 1.1750 | 0.4419 |
| 1.0114 | 16.0 | 448 | 1.1653 | 0.4419 |
| 0.9862 | 17.0 | 476 | 1.1504 | 0.4651 |
| 0.9557 | 18.0 | 504 | 1.1422 | 0.4651 |
| 0.9635 | 19.0 | 532 | 1.1339 | 0.4651 |
| 0.9341 | 20.0 | 560 | 1.1250 | 0.4419 |
| 0.9142 | 21.0 | 588 | 1.1124 | 0.4651 |
| 0.9058 | 22.0 | 616 | 1.1036 | 0.4651 |
| 0.9242 | 23.0 | 644 | 1.0941 | 0.4884 |
| 0.8383 | 24.0 | 672 | 1.0875 | 0.5349 |
| 0.9017 | 25.0 | 700 | 1.0842 | 0.5116 |
| 0.8456 | 26.0 | 728 | 1.0767 | 0.5349 |
| 0.8584 | 27.0 | 756 | 1.0674 | 0.5349 |
| 0.8615 | 28.0 | 784 | 1.0594 | 0.5349 |
| 0.8302 | 29.0 | 812 | 1.0523 | 0.5581 |
| 0.8218 | 30.0 | 840 | 1.0467 | 0.5581 |
| 0.7666 | 31.0 | 868 | 1.0425 | 0.5581 |
| 0.8374 | 32.0 | 896 | 1.0358 | 0.5581 |
| 0.8102 | 33.0 | 924 | 1.0386 | 0.5581 |
| 0.803 | 34.0 | 952 | 1.0300 | 0.5581 |
| 0.8593 | 35.0 | 980 | 1.0290 | 0.5349 |
| 0.781 | 36.0 | 1008 | 1.0257 | 0.5349 |
| 0.769 | 37.0 | 1036 | 1.0205 | 0.5349 |
| 0.7568 | 38.0 | 1064 | 1.0221 | 0.5349 |
| 0.7521 | 39.0 | 1092 | 1.0159 | 0.5349 |
| 0.7774 | 40.0 | 1120 | 1.0130 | 0.5349 |
| 0.7906 | 41.0 | 1148 | 1.0102 | 0.5349 |
| 0.7478 | 42.0 | 1176 | 1.0092 | 0.5349 |
| 0.7422 | 43.0 | 1204 | 1.0064 | 0.5349 |
| 0.7561 | 44.0 | 1232 | 1.0041 | 0.5349 |
| 0.7139 | 45.0 | 1260 | 1.0022 | 0.5349 |
| 0.7352 | 46.0 | 1288 | 1.0021 | 0.5581 |
| 0.7448 | 47.0 | 1316 | 1.0021 | 0.5581 |
| 0.7592 | 48.0 | 1344 | 1.0020 | 0.5581 |
| 0.7313 | 49.0 | 1372 | 1.0019 | 0.5581 |
| 0.6642 | 50.0 | 1400 | 1.0019 | 0.5581 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0