hushem_1x_deit_tiny_adamax_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: 0.8253
- 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: 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 |
|---|---|---|---|---|
| No log | 1.0 | 6 | 1.4425 | 0.2791 |
| 1.416 | 2.0 | 12 | 1.3728 | 0.3023 |
| 1.416 | 3.0 | 18 | 1.3124 | 0.3488 |
| 1.2388 | 4.0 | 24 | 1.2509 | 0.3721 |
| 1.1051 | 5.0 | 30 | 1.1962 | 0.3488 |
| 1.1051 | 6.0 | 36 | 1.1517 | 0.3721 |
| 0.9682 | 7.0 | 42 | 1.1212 | 0.3721 |
| 0.9682 | 8.0 | 48 | 1.0990 | 0.4186 |
| 0.8769 | 9.0 | 54 | 1.0709 | 0.4884 |
| 0.7643 | 10.0 | 60 | 1.0587 | 0.5116 |
| 0.7643 | 11.0 | 66 | 1.0451 | 0.4884 |
| 0.6717 | 12.0 | 72 | 1.0399 | 0.5581 |
| 0.6717 | 13.0 | 78 | 1.0224 | 0.5349 |
| 0.5988 | 14.0 | 84 | 1.0021 | 0.4884 |
| 0.5291 | 15.0 | 90 | 0.9852 | 0.4884 |
| 0.5291 | 16.0 | 96 | 0.9774 | 0.5116 |
| 0.4581 | 17.0 | 102 | 0.9701 | 0.5116 |
| 0.4581 | 18.0 | 108 | 0.9598 | 0.5116 |
| 0.3895 | 19.0 | 114 | 0.9410 | 0.5814 |
| 0.3415 | 20.0 | 120 | 0.9223 | 0.5581 |
| 0.3415 | 21.0 | 126 | 0.9172 | 0.5349 |
| 0.3044 | 22.0 | 132 | 0.9106 | 0.5349 |
| 0.3044 | 23.0 | 138 | 0.9037 | 0.5581 |
| 0.2632 | 24.0 | 144 | 0.8935 | 0.5581 |
| 0.2425 | 25.0 | 150 | 0.8847 | 0.5814 |
| 0.2425 | 26.0 | 156 | 0.8721 | 0.5581 |
| 0.2102 | 27.0 | 162 | 0.8625 | 0.5581 |
| 0.2102 | 28.0 | 168 | 0.8546 | 0.5581 |
| 0.189 | 29.0 | 174 | 0.8540 | 0.5814 |
| 0.1637 | 30.0 | 180 | 0.8496 | 0.6047 |
| 0.1637 | 31.0 | 186 | 0.8464 | 0.6047 |
| 0.1512 | 32.0 | 192 | 0.8420 | 0.5581 |
| 0.1512 | 33.0 | 198 | 0.8380 | 0.5581 |
| 0.1374 | 34.0 | 204 | 0.8346 | 0.5581 |
| 0.1287 | 35.0 | 210 | 0.8327 | 0.5581 |
| 0.1287 | 36.0 | 216 | 0.8290 | 0.5581 |
| 0.124 | 37.0 | 222 | 0.8276 | 0.5581 |
| 0.124 | 38.0 | 228 | 0.8271 | 0.5581 |
| 0.1186 | 39.0 | 234 | 0.8265 | 0.5581 |
| 0.1159 | 40.0 | 240 | 0.8255 | 0.5581 |
| 0.1159 | 41.0 | 246 | 0.8253 | 0.5581 |
| 0.1139 | 42.0 | 252 | 0.8253 | 0.5581 |
| 0.1139 | 43.0 | 258 | 0.8253 | 0.5581 |
| 0.1142 | 44.0 | 264 | 0.8253 | 0.5581 |
| 0.1107 | 45.0 | 270 | 0.8253 | 0.5581 |
| 0.1107 | 46.0 | 276 | 0.8253 | 0.5581 |
| 0.1118 | 47.0 | 282 | 0.8253 | 0.5581 |
| 0.1118 | 48.0 | 288 | 0.8253 | 0.5581 |
| 0.1159 | 49.0 | 294 | 0.8253 | 0.5581 |
| 0.1095 | 50.0 | 300 | 0.8253 | 0.5581 |
Framework versions
- Transformers 4.35.0
- Pytorch 2.1.0+cu118
- Datasets 2.14.6
- Tokenizers 0.14.1