Can take any image and categorize the animal it see's as either a dog or cat with 97.3% accuracy.
Used University of Oxford's vgg16 model as a base. With no modifications of the layers, the model can classify accurately any animal (ie Golden Retriever, Bengal Cat, Rottweiler).
The last layer of the learning model was changed to simply return: cat or dog.
Used the Dogs v Cats Redux dataset from Kaggle.
Created to get my feet wet with deep learning (& machine learning).
- Python
- Keras
- Theano
- AWS