numpyro
Probabilistic programming with NumPy powered by JAX for autograd and JIT compilation to GPU/TPU/CPU.
Decision gist · record as of 2026-08-14
Yes, if you need probabilistic programming with GPU/TPU acceleration and are comfortable with an actively-developed library that may change its API. The low install friction, active maintenance, permissive license, and strong maintenance signal (recent release, 2736 stars) make it a solid choice for Bayesian inference workflows. Not recommended if you require API stability or are new to probabilistic programming.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Requires JAX and jaxlib, which have their own system-level dependencies (CUDA/cuDNN for GPU support); Python 3.11 or later.
- Installation is straightforward with low friction; the package is actively maintained with a recent release and carries 5 runtime dependencies including jax and jaxlib.
- The codebase is under active development, so API stability is not guaranteed.
License · maintenance · safety
permissive license (permissive) — Licensed under Apache 2.0 (permissive), allowing free use, modification, and distribution with minimal restrictions.
last release 2026-05-02 (104 days) · last repo commit 2026-08-13 · 2,736 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 777,898 downloads/mo, #5,087 on PyPI
Alternatives
Verify before relying
pip install numpyro
import numpyro
import numpyro.distributions as dist
from numpyro.infer import MCMC, NUTS
from jax import random
def model(data):
mu = numpyro.sample('mu', dist.Normal(0, 1))
numpyro.sample('obs', dist.Normal(mu, 1), obs=data)
kernel = NUTS(model)
mcmc = MCMC(kernel, num_warmup=500, num_samples=1000)
mcmc.run(random.key(0), data)- Whether the API brittleness mentioned in the description excerpt affects common use cases or only edge cases.
- Performance characteristics compared to other probabilistic programming frameworks for typical model sizes.
What it is and what it does
NumPyro is a lightweight probabilistic programming library that brings Pyro's API to JAX, enabling fast Bayesian inference through automatic differentiation and JIT compilation. It provides MCMC samplers (including the No-U-Turn Sampler and Hamiltonian Monte Carlo variants), variational inference with flexible guides, a comprehensive distribution library, and effect handlers for building custom inference algorithms.
The library is designed for users building hierarchical Bayesian models, performing posterior inference, or exploring probabilistic programming on modern hardware. It relies on jax, jaxlib, numpy, multipledispatch, and tqdm. The package is actively developed but explicitly warns of potential API changes as the design evolves.
Use it for
- Run MCMC inference on hierarchical Bayesian models with GPU acceleration via JAX JIT compilation.
- Implement variational inference for models with discrete and continuous latent variables using ADVI.
- Build custom inference algorithms by composing effect handlers and Pyro primitives.
- Perform Bayesian data analysis with standard distributions and constraints similar to PyTorch's API.
- Accelerate Hamiltonian Monte Carlo by compiling the entire verlet integrator and tree-building stage.
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes, if you need probabilistic programming with GPU/TPU acceleration and are comfortable with an actively-developed library that may change its API.
The low install friction, active maintenance, permissive license, and strong maintenance signal (recent release, 2736 stars) make it a solid choice for Bayesian inference workflows. Not recommended if you require API stability or are new to probabilistic programming.
Install
numpyro on PyPI
Before you install
Installation is straightforward with low friction; the package is actively maintained with a recent release and carries 5 runtime dependencies including jax and jaxlib. The codebase is under active development, so API stability is not guaranteed.
Requires JAX and jaxlib, which have their own system-level dependencies (CUDA/cuDNN for GPU support); Python 3.11 or later.
License in practice
Licensed under Apache 2.0 (permissive), allowing free use, modification, and distribution with minimal restrictions.
Quickstart
pip install numpyro
import numpyro
import numpyro.distributions as dist
from numpyro.infer import MCMC, NUTS
from jax import random
def model(data):
mu = numpyro.sample('mu', dist.Normal(0, 1))
numpyro.sample('obs', dist.Normal(mu, 1), obs=data)
kernel = NUTS(model)
mcmc = MCMC(kernel, num_warmup=500, num_samples=1000)
mcmc.run(random.key(0), data)Verify before relying
- Whether the API brittleness mentioned in the description excerpt affects common use cases or only edge cases.
- Performance characteristics compared to other probabilistic programming frameworks for typical model sizes.
Package facts
| License | permissive license permissive |
| Python support | Supports the current Python release >=3.11 |
| Install friction | Low. Pure-Python wheel |
| Runtime dependencies | 5 packagesjaxjaxlibmultipledispatchnumpytqdm |
| Maintenance | Actively maintained 104 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 777,898 / month, #5,087 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
| Classifiers | Intended Audience :: DevelopersIntended Audience :: EducationIntended Audience :: Science/ResearchLicense :: OSI Approved :: Apache Software LicenseOperating System :: MacOS :: MacOS XOperating System :: POSIX :: LinuxProgramming Language :: Python :: 3.11Programming Language :: Python :: 3.12Programming Language :: Python :: 3.13Programming Language :: Python :: 3.14 |
Evidence: numpyro-0.21.0-py3-none-any.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “probabilistic programming bayesian inference”
- numpyroNumPyro is a probabilistic programming library that uses JAX for…
- pyro-pplPyro is a deep probabilistic programming library built on PyTorch…
- pymc3PyMC3 is a Python package for Bayesian statistical modeling and…
Give your agent the search over MCP, or paste the wish link into any chat.
More Artificial Intelligence packages
LiteLLM provides a unified Python interface to call 100+ LLM providers (OpenAI, Anthropic, Gemini, Bedrock, Azure, and others) using OpenAI-compatible API format, available as both a Python SDK and a self-hosted AI Gateway proxy server.
Install it if you need to work with multiple LLM providers or want to centralize LLM routing in your organization.
Client library and CLI tool for downloading, uploading, and managing models, datasets, and repositories on the Hugging Face Hub platform.
Install it if you work with Hugging Face Hub models or datasets.
LangChain provides a framework for building agents and LLM-powered applications by composing language models, tools, and memory through a unified API that abstracts over multiple model providers.
hf-xet provides chunk-based deduplication and efficient file transfer for the Hugging Face Hub, enabling faster uploads and downloads of large files with local disk caching.
Tokenizers converts raw text into token sequences for NLP models, with support for training custom vocabularies and using pre-built tokenizers (BPE, WordPiece) optimized for speed via Rust.
Transformers provides a unified framework for loading, fine-tuning, and running state-of-the-art pretrained models across text, vision, audio, video, and multimodal tasks using PyTorch, JAX, or TensorFlow.
Install it if you need to run or train any transformer-based model for NLP, vision, audio, or multimodal tasks.
See also blackjax · pyro-ppl · pyro-api · pymc3 · pystan · distrax · tensorflow-probability · nutpie · pymc · emcee