wild_cat_classification
This model is a fine-tuned version of vgg19 pre-trained model. It achieves the following results on the evaluation set:
- Loss: 0.0095
- Accuracy: 0.9600
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-02
- train_batch_size: 64
- eval_batch_size: 64
- optimizer: SGD
- num_epochs: 1
Training results
Training Loss | Epoch | Batch |
---|---|---|
0.0620 | 1.0 | 3/37 |
0.0425 | 1.0 | 6/37 |
0.0390 | 1.0 | 9/37 |
0.0258 | 1.0 | 12/37 |
0.0152 | 1.0 | 15/37 |
0.0180 | 1.0 | 18/37 |
0.0225 | 1.0 | 21/37 |
0.0129 | 1.0 | 24/37 |
0.0479 | 1.0 | 27/37 |
0.0231 | 1.0 | 30/37 |
0.0168 | 1.0 | 33/37 |
0.0095 | 1.0 | 36/37 |
Framework versions
- Pytorch 2.1.0+cu121
Evaluation results
- Accuracy on https://www.kaggle.com/datasets/gpiosenka/cats-in-the-wild-image-classificationself-reported0.960