Setup

Setup#

This course uses the following tools and libraries, all of which are free:

  • Python

  • Jupyter

  • NumPy

  • Matplotlib

  • Scikit-learn

  • PyTorch

Installing each of these tools locally is certainly possible using pip and/or conda, but it is typically easier to run the lessons in a Docker container. Docker is a tool for creating, distributing, and running virtual machines with various software and configurations pre-installed. Installation instructions for Docker can be found at Docker’s Getting Started page.

Starting Jupyter using docker run#

Once Docker has been installed, all of the material in this course can be run using a single publicly-available docker image named quay.io/jupyter/pytorch-notebook:python-3.12. This Docker image already contains all of the tools above, including a ready installation of Jupyter. We can start the Jupyter instance using the following command in a terminal:

docker run --rm -it -p8888:8888 quay.io/jupyter/pytorch-notebook:python-3.12

The first time you run this command, it will likely need to pull (download) the image. This can take several minutes depending on your internet connection, but it should save the image locally so that subsequent commands don’t require as much time. Once the Docker image has been pulled, Docker will start up the embedded virtual machine (a Docker “container”). During startup, the virtual machine will print several messages to the screen; eventually, it should print something like the following:

    To access the server, open this file in a browser:
        file:///home/jovyan/.local/share/jupyter/runtime/jpserver-7-open.html
    Or copy and paste one of these URLs:
        http://localhost:8888/lab?token=b53e9f7c290d48eb3cc9886a288bbcca8ae5f9d61025d54e
        http://127.0.0.1:8888/lab?token=b53e9f7c290d48eb3cc9886a288bbcca8ae5f9d61025d54e

The precise message may be slightly different, but the important part is that you find a web address that starts with http://127.0.0.1/ and that ends with a long token of numbers and the letters a–f. Copy the last line of this message (starting with the http: and ending with the entire token) into your web browser navigation bar, and it should load the Jupyter instance embedded in the Docker container.

All of the lessons in this course can be completed using the Jupyter instance loaded using these instructions.

Starting Jupyter using docker compose#

An even easier way to get started with the lessons is to use the docker compose command along with the docker-compose.yml file in this Jupyter Book’s GitHub repository. To do this, you must have Docker installed and running, and you must have git installed; see the Software Carpentry git lesson for help installing git and getting accustomed to its basic usage. In the command line enter the following commands:

  1. git clone https://github.com/noahbenson/AI-ABCs

  2. cd AI-ABCs

  3. docker compose up

The final command should start the same docker image, but it will include a directory lessons that contains the lesson notebooks for the course.

Contents