vit-base-patch16-224-in21k-cards-base-classifier-defects-finder
This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.0683
- Accuracy: 0.999
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: 5e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.4892 | 0.9929 | 70 | 1.3366 | 0.859 |
| 0.4362 | 2.0 | 141 | 0.4142 | 0.971 |
| 0.231 | 2.9929 | 211 | 0.2250 | 0.988 |
| 0.1654 | 4.0 | 282 | 0.1687 | 0.982 |
| 0.1289 | 4.9929 | 352 | 0.1322 | 0.991 |
| 0.0999 | 6.0 | 423 | 0.1184 | 0.988 |
| 0.0824 | 6.9929 | 493 | 0.0852 | 0.996 |
| 0.0789 | 8.0 | 564 | 0.0809 | 0.998 |
| 0.07 | 8.9929 | 634 | 0.0723 | 0.997 |
| 0.067 | 9.9291 | 700 | 0.0683 | 0.999 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.0.1+cu117
- Datasets 2.19.2
- Tokenizers 0.19.1