smids_5x_deit_small_sgd_0001_fold4
This model is a fine-tuned version of facebook/deit-small-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.4830
- Accuracy: 0.8233
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: 0.0001
- 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 |
|---|---|---|---|---|
| 1.0422 | 1.0 | 375 | 1.0355 | 0.445 |
| 0.9877 | 2.0 | 750 | 0.9987 | 0.5017 |
| 0.9301 | 3.0 | 1125 | 0.9591 | 0.5367 |
| 0.9069 | 4.0 | 1500 | 0.9204 | 0.5917 |
| 0.8815 | 5.0 | 1875 | 0.8838 | 0.6217 |
| 0.8208 | 6.0 | 2250 | 0.8478 | 0.6383 |
| 0.7819 | 7.0 | 2625 | 0.8141 | 0.6817 |
| 0.7955 | 8.0 | 3000 | 0.7823 | 0.7033 |
| 0.7492 | 9.0 | 3375 | 0.7528 | 0.7233 |
| 0.7403 | 10.0 | 3750 | 0.7259 | 0.7317 |
| 0.7047 | 11.0 | 4125 | 0.7009 | 0.745 |
| 0.6669 | 12.0 | 4500 | 0.6790 | 0.76 |
| 0.6557 | 13.0 | 4875 | 0.6594 | 0.7667 |
| 0.6563 | 14.0 | 5250 | 0.6418 | 0.77 |
| 0.5999 | 15.0 | 5625 | 0.6263 | 0.7667 |
| 0.589 | 16.0 | 6000 | 0.6125 | 0.77 |
| 0.5618 | 17.0 | 6375 | 0.5999 | 0.7767 |
| 0.5666 | 18.0 | 6750 | 0.5885 | 0.7817 |
| 0.6067 | 19.0 | 7125 | 0.5784 | 0.7867 |
| 0.5796 | 20.0 | 7500 | 0.5694 | 0.79 |
| 0.547 | 21.0 | 7875 | 0.5612 | 0.7883 |
| 0.5698 | 22.0 | 8250 | 0.5540 | 0.7867 |
| 0.5377 | 23.0 | 8625 | 0.5473 | 0.7917 |
| 0.5508 | 24.0 | 9000 | 0.5411 | 0.7967 |
| 0.5752 | 25.0 | 9375 | 0.5355 | 0.7983 |
| 0.5019 | 26.0 | 9750 | 0.5303 | 0.8 |
| 0.5146 | 27.0 | 10125 | 0.5255 | 0.8017 |
| 0.5114 | 28.0 | 10500 | 0.5210 | 0.8033 |
| 0.4588 | 29.0 | 10875 | 0.5170 | 0.8033 |
| 0.5045 | 30.0 | 11250 | 0.5133 | 0.805 |
| 0.5118 | 31.0 | 11625 | 0.5098 | 0.805 |
| 0.4619 | 32.0 | 12000 | 0.5067 | 0.8083 |
| 0.4796 | 33.0 | 12375 | 0.5037 | 0.81 |
| 0.5217 | 34.0 | 12750 | 0.5011 | 0.81 |
| 0.4423 | 35.0 | 13125 | 0.4986 | 0.8133 |
| 0.4692 | 36.0 | 13500 | 0.4964 | 0.815 |
| 0.4889 | 37.0 | 13875 | 0.4944 | 0.815 |
| 0.487 | 38.0 | 14250 | 0.4925 | 0.82 |
| 0.5206 | 39.0 | 14625 | 0.4909 | 0.82 |
| 0.4988 | 40.0 | 15000 | 0.4894 | 0.82 |
| 0.4485 | 41.0 | 15375 | 0.4881 | 0.8217 |
| 0.4284 | 42.0 | 15750 | 0.4870 | 0.8217 |
| 0.4979 | 43.0 | 16125 | 0.4860 | 0.8217 |
| 0.454 | 44.0 | 16500 | 0.4851 | 0.8217 |
| 0.4865 | 45.0 | 16875 | 0.4845 | 0.8217 |
| 0.4847 | 46.0 | 17250 | 0.4839 | 0.8217 |
| 0.5681 | 47.0 | 17625 | 0.4835 | 0.8217 |
| 0.4795 | 48.0 | 18000 | 0.4832 | 0.8217 |
| 0.4757 | 49.0 | 18375 | 0.4831 | 0.8233 |
| 0.4471 | 50.0 | 18750 | 0.4830 | 0.8233 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2