https://github.com/jax-ml/bayeux

State of the art inference for your bayesian models.
https://github.com/jax-ml/bayeux

Last synced: 26 days ago
JSON representation

State of the art inference for your bayesian models.

• Host: GitHub
• URL: https://github.com/jax-ml/bayeux
• Owner: jax-ml
• License: apache-2.0
• Created: 2023-11-02T16:52:57.000Z (8 months ago)
• Default Branch: main
• Last Pushed: 2024-05-17T14:18:25.000Z (about 1 month ago)
• Last Synced: 2024-05-17T18:44:55.384Z (about 1 month ago)
• Language: Python
• Homepage: https://jax-ml.github.io/bayeux/
• Size: 1.51 MB
• Stars: 99
• Watchers: 3
• Forks: 4
• Open Issues: 2
• Metadata Files:
  • Readme: README.md
  • Changelog: CHANGELOG.md
  • Contributing: CONTRIBUTING.md
  • License: LICENSE
  • Code of conduct: CODE_OF_CONDUCT.md

Lists

• awesome-stars - jax-ml/bayeux - State of the art inference for your bayesian models. (Python)

README

        
# Bayeux

*Stitching together models and samplers*

[![Unittests](https://github.com/jax-ml/bayeux/actions/workflows/pytest_and_autopublish.yml/badge.svg)](https://github.com/jax-ml/bayeux/actions/workflows/pytest_and_autopublish.yml)

[![PyPI version](https://badge.fury.io/py/bayeux_ml.svg)](https://badge.fury.io/py/bayeux_ml)

`bayeux` lets you write a probabilistic model in JAX and immediately have access to state-of-the-art inference methods. The API aims to be **simple**, **self descriptive**, and **helpful**. Simply provide a log density function (which doesn't even have to be normalized), along with a single point (specified as a [pytree](https://jax.readthedocs.io/en/latest/pytrees.html)) where that log density is finite. Then let `bayeux` do the rest!

## Installation

```bash

pip install bayeux-ml

```

## Quickstart

We define a model by providing a log density in JAX. This could be defined using a probabilistic programming language (PPL) like [numpyro](examples/numpyro_and_bayeux), [PyMC](examples/pymc_and_bayeux), [TFP](examples/tfp_and_bayeux), distrax, oryx, coix, or directly in JAX.

```python

import bayeux as bx

import jax

normal_density = bx.Model(

  log_density=lambda x: -x*x,

  test_point=1.)

seed = jax.random.key(0)

opt_results = normal_density.optimize.optax_adam(seed=seed)

# OR!

idata = normal_density.mcmc.numpyro_nuts(seed=seed)

# OR!

surrogate_posterior, loss = normal_density.vi.tfp_factored_surrogate_posterior(seed=seed)

```

## Read more

* [Defining models](inference)

* [Inspecting models](inspecting)

* [Testing and debugging](debug_mode)

* Also see `bayeux` integration with [numpyro](examples/numpyro_and_bayeux), [PyMC](examples/pymc_and_bayeux), and [TFP](examples/tfp_and_bayeux)!

*This is not an officially supported Google product.*