Converting your project to Hyrax

The basic flow for using Hyrax in your machine learning project is to:

1. Write code that connects your data and model to hyrax in a notebook.

2. Put working code into a small Python package.

3. Reuse that package across many Hyrax notebooks.

The pages below walk you through making custom dataset and model classes for your data in your project as well as how to distribute working datasets and models via a python package which can be used by your collaborators. If you are deciding what to implement first, see Required inputs for the minimum pieces Hyrax needs from users.

• Dataset class reference
• Dataset class notebook example
• Model class reference
• Model class notebook example
• External package setup

If you want a concrete example package, see external_hyrax_example.