Overview

Web app for Udacity's Dog Breed Classifier Project.

From a Convolutional Neural Network trained on 133 different dog breeds it:

• If a dog image is selected:
  • Determine which dog breed it most likely is.
• If a human face image is selected:
  • Determine which dog breed is most resembling.

In both cases the app displays:

• The five most likely dog breeds with associated probabilities
• An image of the dog breed with the highest probability alongside the selected image

Instructions

1. Clone the repository.

git clone https://github.com/o1sa/dog-project-app.git
cd dog-project-app

2. Create and activate a new virtual environment.

python3 -m virtualenv venv
source venv/bin/activate

3. Download the dependencies using pip.

pip install -r requirements.txt

4. Start the app locally

• either by

python dog-project.py

• or using the flask run command

export FLASK_DEBUG=1
export FLASK_ENV=development
export FLASK_APP=dog-project.py
flask run

5. Open a web browser and navigate to the specified url.