Udacity project: Building a Convolutional Neural Network to determine the dog breed of a dog in a given image
To use the interactive jupyter notebook version of this project:
If you already have Anaconda, you already have installed Jupyter Notebook. If not, see install instructions here: http://jupyter.readthedocs.io/en/latest/install.html
Clone the DogBreeds repository and navigate to the directory where the Jupyter notebook is located.
Download the dog dataset. Unzip the dog dataset folder and place it in the cloned repo at location path/to/dog-project/dogImages.
Download the human dataset. Unzip the human dataset folder and place it in the repo at location path/to/dog-project/lfw.
Download the VGG-16 bottleneck features. Place it in the repo at location path/to/dog-project/bottleneck_features.
Via TensorFlow with GPU support on your local machine or AWS, set up an environment with keras using a TensorFlow backend.
Open the Jupyter (IPython) notebook with the jupyter notebook dog_app.ipynb command.
If you are running the project on your local machine (and not using AWS), before running code, change the kernel to match the dog-project environment by using the drop-down menu (Kernel > Change kernel > dog-project).
Udacity provided the project design, starter code, and the majority of the information in this README.