Dockerfile Setup

We provide a base Dockerfile in this repository. A Dockerfile is a text document that contains all the commands a user could call on the command line to assemble an image by installing software and configuring the OS. When you build the Docker image, these commands are run one after the other to create your Docker image.

In the example Dockerfile, when you run ./rcc.sh build, your Docker image:

1. Starts with an Ubuntu 18.04 Operating System
2. Updates the Ubuntu installer and installs tree, vim, curl, python3, python3-pip, & git
3. Uses pip3 to install several common python data science packages
4. Creates the work directory in the root folder
5. Uses git to clone this repository into the root folder

You will be responsible for updating this Dockerfile to include all of the software and libraries necessary to train and run your model for running against the development fold and holdout corpora. You can follow the examples here for installing OS packages or any other software you will need (provided that it is open source).

References

• Dockerfile reference
• Ubuntu apt-get reference