convnext-nano-3e-4-augment
This model is a fine-tuned version of facebook/convnextv2-nano-22k-384 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.2872
- Accuracy: 0.9280
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: 0.0003
- train_batch_size: 64
- eval_batch_size: 64
- 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 |
|---|---|---|---|---|
| 0.8119 | 1.0 | 275 | 0.5637 | 0.8270 |
| 0.5829 | 2.0 | 550 | 0.5016 | 0.8497 |
| 0.4623 | 3.0 | 825 | 0.4494 | 0.8755 |
| 0.359 | 4.0 | 1100 | 0.3809 | 0.8887 |
| 0.2881 | 5.0 | 1375 | 0.3742 | 0.8998 |
| 0.2302 | 6.0 | 1650 | 0.3402 | 0.9113 |
| 0.1827 | 7.0 | 1925 | 0.3150 | 0.9121 |
| 0.1466 | 8.0 | 2200 | 0.3012 | 0.9229 |
| 0.1223 | 9.0 | 2475 | 0.2996 | 0.9249 |
| 0.1332 | 10.0 | 2750 | 0.2948 | 0.9249 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.1.2
- Datasets 2.18.0
- Tokenizers 0.15.2