• Project: Use a Pre-trained Image Classifier to Identify Dog Breeds
• Nanodegree program: AI Programming with Python
• Course provider: Udacity

Clone this repository with the tag project-starter-files to get the project starter files without any of the solutions.

Using SSH

git clone --depth 1 --branch project-starter-files [email protected]:jsstrn/dog-breed-image-classifier.git 

Using HTTPS

git clone --depth 1 --branch project-starter-files https://github.com/jsstrn/dog-breed-image-classifier.git

First, install Anaconda. This should also install the latest version of Python 3 for you.

Check the version on Anaconda:

conda --version

Update Anaconda and its dependencies:

conda update conda

Create a new virtual environment:

conda create --name my-new-env

Activate your new environment:

conda activate my-new-env

Check which environment you are currently on. The current environment is indicated by an asterisk *:

conda info --env
base          /some/dir/to/your/env
my-new-env *  /some/dir/to/your/env

Now install the following packages:

conda install pillow
conda install torchvision

To run the program:

python check_images.py

To run image classifier with all three models

sh run_models_batch.sh

To run image classifer against images in the uploaded_images directory:

sh run_models_batch_uploaded.sh

“Programs are meant to be read by humans and only incidentally for computers to execute.”

-- Donald Knuth, The Art of Computer Programming.