Contents

numpyro 0.15.0

0

Pyro PPL on NumPy

Pyro PPL on NumPy

Stars: 2057, Watchers: 2057, Forks: 220, Open Issues: 74

The pyro-ppl/numpyro repo was created 5 years ago and the last code push was 20 hours ago.
The project is very popular with an impressive 2057 github stars!

How to Install numpyro

You can install numpyro using pip

pip install numpyro

or add it to a project with poetry

poetry add numpyro

Package Details

Author
Uber AI Labs
License
Apache License 2.0
Homepage
https://github.com/pyro-ppl/numpyro
PyPi:
https://pypi.org/project/numpyro/
GitHub Repo:
https://github.com/pyro-ppl/numpyro

Classifiers

No  numpyro  pypi packages just yet.

Errors

A list of common numpyro errors.

Code Examples

Here are some numpyro code examples and snippets.

GitHub Issues

The numpyro package has 74 open issues on GitHub

  • Move contrib.render to infer.inspect
  • Fix HMCECS for likelihoods with multiple plates
  • Add loose strategy for missing plates in MCMC
  • HMC Gibbs Hierarchical Prior Issue
  • Example: Deep Markov Model with steinVI
  • Example: Bayes NN with steinVI
  • User contributed example from epidemiology
  • Add various KL divergences for Gamma/Beta families
  • Add Kumaraswamy and relaxed Bernoulli distributions
  • [FR] raise missing plate warning in MCMC inference
  • [Do Not Merge] Bump Funsor to latest master commit for testing
  • Amortizing overhead in .run
  • Request for Euler Maruyama features in numpyro
  • Add more tests for collapse handler

See more issues on GitHub

Related Packages & Articles

pymc3 3.11.5

Probabilistic Programming in Python: Bayesian Modeling and Probabilistic Machine Learning with Theano

pyro-ppl 1.9.0

A Python library for probabilistic modeling and inference

orbit-ml 1.1.4.9

Orbit is a package for Bayesian time series modeling and inference.

ta 0.11.0

ta is a Python module that provides a technical analysis library, designed to enable feature engineering from financial time series datasets. It is built on the pandas and numpy libraries and offers a wide range of indicators such as volume, volatility, trend, and momentum indicators. ta is designed for Python developers working in the financial sector, making it a valuable asset in the field of Financial Software and Fintech Solutions, particularly for those developing trading algorithms or investment strategies.

pyflux 0.4.15

pyflux is a Python library designed for time series analysis. It includes modern time series models, including ARIMA, GARCH, and state space models. With a flexible set of inference options, pyflux enables a probabilistic approach to time series modeling. Though still in alpha, it's an invaluable tool for statisticians and data scientists working with time series data.

orjson 3.10.3

Fast, correct Python JSON library supporting dataclasses, datetimes, and numpy