convnext-base-3e-5-wd-1e-8-raug
This model is a fine-tuned version of facebook/convnextv2-base-22k-384 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.2296
- Accuracy: 0.9458
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: 3e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.6237 | 1.0 | 1099 | 0.3587 | 0.8994 |
| 0.4599 | 2.0 | 2198 | 0.2743 | 0.9213 |
| 0.359 | 3.0 | 3297 | 0.2579 | 0.9252 |
| 0.3047 | 4.0 | 4396 | 0.2404 | 0.9388 |
| 0.2869 | 5.0 | 5495 | 0.2348 | 0.9408 |
| 0.2468 | 6.0 | 6594 | 0.2276 | 0.9455 |
| 0.2098 | 7.0 | 7693 | 0.2303 | 0.9471 |
| 0.1944 | 8.0 | 8792 | 0.2244 | 0.9495 |
| 0.1739 | 9.0 | 9891 | 0.2247 | 0.9507 |
| 0.1508 | 10.0 | 10990 | 0.2243 | 0.9487 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.1.2
- Datasets 2.18.0
- Tokenizers 0.15.2