Udacity Dog Breed Identification

This is my project submission for the Udacity project "Dog Breed Identification".

The originally provided source code template is linked here: project dog classification.

Installation

Source Code

You need a python 3.X installation and preferably a environment management like conda.

Download or clone this git repository and install the python dependencies using the following command:

# Install uda_dog environment:
conda env create -f requirements.yml

# Remove uda_dog environment:
conda remove --name uda_dog --all

Data

To run the dog app you need to download the following images of dogs and humans:

• Download the dog dataset. Unzip the folder and place it in this project's home directory, at the location /data/dog_images.

• Download the human dataset. Unzip the folder and place it in the home directory, at location /data/lfw.

Documentation

The project is documented in the following two files:

• Capstone Proposal
• Project Report

Run App

To run the "dog app" open and run the jupyter notebook dog_app

Contact

Grzegorz Lippe - @GrzegorzLippe - grzegorz.lippe@eposteo.de