This is just a minimal introduction to Docker. The goal is to jump-start moving forward from research to production. Docker containers establish a very lightweight virtualization solution that may support the reproducibility and portability of ML model development and deployment.

1. Install the Docker Engine.

   • On Linux distributions adopting the systemd service manager, which is the majority, one may then need to start Docker by running sudo systemctl start docker.
   • Try sudo -i or, better yet, sudo -s, if you find continuous sudo-ing annoying. At the very least, the commands below have to be run with root privileges.
   • You may want to familiarize yourself with some Docker commands, for example by consulting docker help.

2. Make sure that everything works properly:

   docker run hello-world
   

3. Build an image from the shipped dockerfile:

   docker build -t my-image .
   

4. Create and run a container from the image:

   docker run -p 8888:8888 my-image