In this project, I build a convolutional neural network (CNN) that can classify the breed of dog from any user-supplied image. If the image is of a human and not a dog, the algorithm will provide an estimate of the dog breed that is most resembling.
The code is written in Python 3 and Keras with Tensorflow backend all presented in Jupyter Notebook. I used AWS EC2 gpu instance for training the model.
If you'd like to use this notebook to do any re-training on the dataset you can grab it at the links below.
wget https://s3-us-west-1.amazonaws.com/udacity-aind/dog-project/dogImages.zip; unzip dogImages.zip; rm dogImages.zip
wget http://vis-www.cs.umass.edu/lfw/lfw.tgz; tar -xvzf lfw.tgz; rm lfw.tgz