smids_3x_deit_tiny_sgd_00001_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: 1.1190
- Accuracy: 0.405
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 |
|---|---|---|---|---|
| 1.3567 | 1.0 | 225 | 1.3548 | 0.3467 |
| 1.2325 | 2.0 | 450 | 1.3329 | 0.345 |
| 1.3433 | 3.0 | 675 | 1.3131 | 0.3467 |
| 1.2405 | 4.0 | 900 | 1.2949 | 0.3467 |
| 1.3135 | 5.0 | 1125 | 1.2785 | 0.3533 |
| 1.2715 | 6.0 | 1350 | 1.2636 | 0.36 |
| 1.2599 | 7.0 | 1575 | 1.2502 | 0.3583 |
| 1.2399 | 8.0 | 1800 | 1.2380 | 0.365 |
| 1.2555 | 9.0 | 2025 | 1.2270 | 0.3667 |
| 1.2073 | 10.0 | 2250 | 1.2172 | 0.38 |
| 1.2054 | 11.0 | 2475 | 1.2082 | 0.3767 |
| 1.2574 | 12.0 | 2700 | 1.2001 | 0.3817 |
| 1.1461 | 13.0 | 2925 | 1.1929 | 0.385 |
| 1.1651 | 14.0 | 3150 | 1.1863 | 0.3883 |
| 1.2048 | 15.0 | 3375 | 1.1803 | 0.39 |
| 1.1932 | 16.0 | 3600 | 1.1750 | 0.395 |
| 1.1333 | 17.0 | 3825 | 1.1701 | 0.3967 |
| 1.1551 | 18.0 | 4050 | 1.1656 | 0.3917 |
| 1.1804 | 19.0 | 4275 | 1.1615 | 0.3933 |
| 1.1749 | 20.0 | 4500 | 1.1578 | 0.39 |
| 1.1596 | 21.0 | 4725 | 1.1543 | 0.39 |
| 1.1144 | 22.0 | 4950 | 1.1511 | 0.39 |
| 1.1024 | 23.0 | 5175 | 1.1481 | 0.395 |
| 1.1547 | 24.0 | 5400 | 1.1454 | 0.3967 |
| 1.1295 | 25.0 | 5625 | 1.1429 | 0.395 |
| 1.1235 | 26.0 | 5850 | 1.1405 | 0.3983 |
| 1.0919 | 27.0 | 6075 | 1.1383 | 0.3983 |
| 1.0892 | 28.0 | 6300 | 1.1364 | 0.4033 |
| 1.1395 | 29.0 | 6525 | 1.1345 | 0.4017 |
| 1.1334 | 30.0 | 6750 | 1.1328 | 0.4 |
| 1.0988 | 31.0 | 6975 | 1.1312 | 0.405 |
| 1.1324 | 32.0 | 7200 | 1.1297 | 0.4067 |
| 1.1023 | 33.0 | 7425 | 1.1284 | 0.41 |
| 1.1357 | 34.0 | 7650 | 1.1271 | 0.4083 |
| 1.1037 | 35.0 | 7875 | 1.1259 | 0.405 |
| 1.1207 | 36.0 | 8100 | 1.1249 | 0.4033 |
| 1.1299 | 37.0 | 8325 | 1.1239 | 0.4033 |
| 1.1551 | 38.0 | 8550 | 1.1231 | 0.405 |
| 1.1268 | 39.0 | 8775 | 1.1223 | 0.4067 |
| 1.1415 | 40.0 | 9000 | 1.1216 | 0.405 |
| 1.1269 | 41.0 | 9225 | 1.1210 | 0.4067 |
| 1.1595 | 42.0 | 9450 | 1.1205 | 0.4067 |
| 1.1387 | 43.0 | 9675 | 1.1201 | 0.4067 |
| 1.0676 | 44.0 | 9900 | 1.1198 | 0.4067 |
| 1.1227 | 45.0 | 10125 | 1.1195 | 0.4067 |
| 1.0985 | 46.0 | 10350 | 1.1193 | 0.4067 |
| 1.0755 | 47.0 | 10575 | 1.1191 | 0.405 |
| 1.0965 | 48.0 | 10800 | 1.1190 | 0.405 |
| 1.1228 | 49.0 | 11025 | 1.1190 | 0.405 |
| 1.0829 | 50.0 | 11250 | 1.1190 | 0.405 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2