A CNN to distinguish between cats and dogs.
Training set of 2000 images, 1000 images of cats and dogs each.
100 epochs
With Augmentation
Gives a training accuracy of 89% and a validation accuracy of 87.1%
Without Augmentation
Give a training accuracy of 99% and validation accuracy of 76%
50 epochs
Without Augmentation
Gives a training accuracy of 85% and validation accuracy of 84%
Trained on Google Colab
Cats_vs_Dogs_model2.ipynb (50 epochs)