Skip to content

Latest commit

 

History

History
49 lines (37 loc) · 2.16 KB

README.md

File metadata and controls

49 lines (37 loc) · 2.16 KB
Raw

Reproducing "Data Analysis - A Bayesian Tutorial" Examples with Pyro

Pyro Logo

Description

This project is part of my journey to learn Pyro, a universal probabilistic programming language. I'm working through examples from the book "Data Analysis - A Bayesian Tutorial" (Second edition) by D. S. Sivia and J. Skilling, implementing solutions using Pyro instead of the mathematical approaches in the book.

Key Features

  • Implementation of Bayesian analysis examples using Pyro
  • Focus on examples with extractable data from the book
  • Comparative analysis between mathematical solutions and Pyro implementations

How to Use

  1. Clone this repository
  2. Install the environment using Miniconda: 
    conda env create -f environment.yml
conda activate pyro
  3. Open the Jupyter notebook in the repository

Alternatively, click the badge below to launch this project in a Binder environment in your browser.

Binder

Notebooks

  1. Main Notebook - Implementation of book examples using Pyro
  2. Bayesian Linear Regression - An extension to the book's content, implementing and testing a Bayesian Linear Regression model on simulated data.

Technologies Used

  • Python 3.x
  • Pyro 1.4.0
  • Jupyter Notebook
  • PyTorch
  • Matplotlib

Future Work

  • Implement more examples from the book
  • Implement all examples in NumPyro

How to Contribute

This is a personal learning project, but suggestions and discussions are welcome! Feel free to open an issue or submit a pull request.

Status

This project under sporadic development. Content and implementations may change as I progress through the book and deepen my understanding of Pyro.

Resources