pyro-ppl/pyro

Deep universal probabilistic programming with Python and PyTorch

73
/ 100
Verified

Supports automatic inference through variational inference and Markov chain Monte Carlo samplers, with composable effect handlers enabling custom inference algorithms without modifying model code. Built on PyTorch's autograd for efficient gradient-based inference at scale. Integrates with NumPyro for JAX-based inference and provides specialized modules for time-series modeling, causal inference, and experimental design optimization.

8,989 stars and 1,262,278 monthly downloads. Used by 10 other packages. No commits in the last 6 months. Available on PyPI.

Stale 6m
Maintenance 2 / 25
Adoption 25 / 25
Maturity 25 / 25
Community 21 / 25

How are scores calculated?

Stars

8,989

Forks

1,005

Language

Python

License

Apache-2.0

Last pushed

Jul 09, 2025

Monthly downloads

1,262,278

Commits (30d)

0

Dependencies

5

Reverse dependents

10

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/pyro-ppl/pyro"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.