pyprobml
Python code for "Probabilistic Machine learning" book by Kevin Murphy
What is pyprobml?
Python code for "Probabilistic Machine learning" book by Kevin Murphy
Overview of pyprobml
What it is
pyprobml is a collection of Python 3 code that reproduces the figures from the books Probabilistic Machine Learning: An Introduction (book 1) and Probabilistic Machine Learning: Advanced Topics (book 2) by Kevin Murphy. The code relies on standard Python libraries such as NumPy, SciPy, Matplotlib, and scikit-learn. Some parts, especially in book 2, use JAX, while book 1 also uses TensorFlow 2 and a small amount of PyTorch. A separate utility package, probml-utils, provides shared helper functions. As of September 2022, the repository is in maintenance mode.
Key features
- Reproduces figures from two comprehensive probabilistic machine learning textbooks.
- Uses widely adopted Python libraries (NumPy, SciPy, Matplotlib, scikit-learn).
- Incorporates JAX, TensorFlow 2, and PyTorch for advanced topics.
- Includes a shared utility package (probml-utils) for common notebook functions.
- Notebooks can be run on Google Colab or locally with minimal setup.
- Supports cloud computing environments such as Lightning AI Studios.
- Licensed under MIT, making it freely available for modification and distribution.
Use cases
- Reproducing figures and examples from the Probabilistic Machine Learning books for study or teaching.
- Learning probabilistic machine learning concepts through hands-on code.
- Building upon existing implementations for research or experimentation.
- Using as a reference for implementing probabilistic models in Python.
Getting started
To run the notebooks locally, first clone the repository with a shallow clone to save space: git clone --depth 1 https://github.com/probml/pyprobml.git Then install the required Python packages from the local requirements file: pip install -r requirements.txt Before running notebooks, you may need to set environment variables to control figure saving. See the repository README for details.
License and pricing
The project is open source under the MIT license. It is free to use, modify, and distribute.
Based on the project's public README and repository metadata.