Welcome to the wiki for bnpy, practical Bayesian nonparametric machine learning with Python.
Want to dive in and learn by doing? See the Demos page for demos you can run in less than 5 minutes.
If you already have standard Python packages like numpy, scipy, and matplotlib, installing bnpy takes just one command.
git clone https://email@example.com/michaelchughes/bnpy-dev.git
If you need further details, check out the installation instructions to get bnpy and its prerequisites (numpy, scipy, matplotlib) installed on your system.
You'll need to make sure that bnpy is on the
PYTHONPATH, and that you have set
BNPYOUTDIR so bnpy knows where to save results on your local system. See the configuration notes for details.
To learn more about the probabilistic models that bnpy supports, checkout the Resources page.
This page provides academic and practical references at beginner and advanced levels.
Visit the FAQ for frequently asked questions.
Have a new question? Please Contact us.
Code Overview and Documentation
bnpy intends to provide a modular and extensible platform that allows you to easily compare and contrast different models and algorithms on big datasets.
To get started, first look at the Big Picture to understand the big concepts and how everything fits together.
Next, you can visit the Step-by-step Walkthrough, where we discuss the major objects and functions called in the end-to-end execution of a bnpy experiment.
Then, you can dig into the detailed documentation (more coming soon, please email with requests!)
On the Datasets page, you can browse a list of popular datasets that bnpy supports (or will support soon).
Want to report a bug? Read How to Report Issues
Want to run this software on the grid at Brown CS? Read How to Use Brown CS Grid