smids_5x_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: 1.0782
- Accuracy: 0.445
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 |
|---|---|---|---|---|
| 1.4131 | 1.0 | 375 | 1.3423 | 0.3433 |
| 1.3593 | 2.0 | 750 | 1.3099 | 0.3483 |
| 1.3082 | 3.0 | 1125 | 1.2818 | 0.3517 |
| 1.3385 | 4.0 | 1500 | 1.2580 | 0.36 |
| 1.2471 | 5.0 | 1875 | 1.2378 | 0.3633 |
| 1.2728 | 6.0 | 2250 | 1.2206 | 0.3667 |
| 1.2244 | 7.0 | 2625 | 1.2061 | 0.3767 |
| 1.1927 | 8.0 | 3000 | 1.1938 | 0.385 |
| 1.1353 | 9.0 | 3375 | 1.1833 | 0.39 |
| 1.1411 | 10.0 | 3750 | 1.1743 | 0.39 |
| 1.1528 | 11.0 | 4125 | 1.1664 | 0.395 |
| 1.1479 | 12.0 | 4500 | 1.1594 | 0.3917 |
| 1.1757 | 13.0 | 4875 | 1.1532 | 0.3917 |
| 1.1667 | 14.0 | 5250 | 1.1477 | 0.4017 |
| 1.1486 | 15.0 | 5625 | 1.1425 | 0.3967 |
| 1.0937 | 16.0 | 6000 | 1.1378 | 0.4017 |
| 1.1232 | 17.0 | 6375 | 1.1333 | 0.4133 |
| 1.1438 | 18.0 | 6750 | 1.1292 | 0.4183 |
| 1.0814 | 19.0 | 7125 | 1.1253 | 0.42 |
| 1.101 | 20.0 | 7500 | 1.1217 | 0.4183 |
| 1.0634 | 21.0 | 7875 | 1.1182 | 0.42 |
| 1.0937 | 22.0 | 8250 | 1.1150 | 0.4167 |
| 1.107 | 23.0 | 8625 | 1.1120 | 0.4183 |
| 1.1086 | 24.0 | 9000 | 1.1091 | 0.42 |
| 1.0802 | 25.0 | 9375 | 1.1064 | 0.4217 |
| 1.1004 | 26.0 | 9750 | 1.1038 | 0.4233 |
| 1.0865 | 27.0 | 10125 | 1.1014 | 0.4267 |
| 1.0686 | 28.0 | 10500 | 1.0991 | 0.425 |
| 1.0719 | 29.0 | 10875 | 1.0969 | 0.4267 |
| 1.0892 | 30.0 | 11250 | 1.0949 | 0.4267 |
| 1.0865 | 31.0 | 11625 | 1.0931 | 0.4233 |
| 1.1008 | 32.0 | 12000 | 1.0913 | 0.425 |
| 1.0834 | 33.0 | 12375 | 1.0897 | 0.4267 |
| 1.085 | 34.0 | 12750 | 1.0882 | 0.4317 |
| 1.0201 | 35.0 | 13125 | 1.0868 | 0.4367 |
| 1.043 | 36.0 | 13500 | 1.0855 | 0.4367 |
| 1.0791 | 37.0 | 13875 | 1.0844 | 0.4367 |
| 1.0443 | 38.0 | 14250 | 1.0833 | 0.4367 |
| 1.0648 | 39.0 | 14625 | 1.0824 | 0.4383 |
| 1.0415 | 40.0 | 15000 | 1.0816 | 0.4417 |
| 1.025 | 41.0 | 15375 | 1.0808 | 0.4417 |
| 1.0078 | 42.0 | 15750 | 1.0802 | 0.4417 |
| 1.0383 | 43.0 | 16125 | 1.0797 | 0.4433 |
| 1.061 | 44.0 | 16500 | 1.0792 | 0.4433 |
| 1.0733 | 45.0 | 16875 | 1.0789 | 0.4433 |
| 1.039 | 46.0 | 17250 | 1.0786 | 0.4433 |
| 1.091 | 47.0 | 17625 | 1.0784 | 0.445 |
| 1.0592 | 48.0 | 18000 | 1.0783 | 0.445 |
| 1.0783 | 49.0 | 18375 | 1.0782 | 0.445 |
| 1.066 | 50.0 | 18750 | 1.0782 | 0.445 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2