Getting Started | Documentation | Community | Contributing

Pyro is a flexible, scalable deep probabilistic programming library built on PyTorch. Notably, it was designed with these principles in mind: - Universal: Pyro is a universal PPL -- it can represent any computable probability distribution. - Scalable: Pyro scales to large data sets with little overhead compared to hand-written code. - Minimal: Pyro is agile and maintainable. It is implemented with a small core of powerful, composable abstractions. - Flexible: Pyro aims for automation when you want it, control when you need it. This is accomplished through high-level abstractions to express generative and inference models, while allowing experts easy-access to customize inference.

Pyro is in an alpha release. It is developed and used by Uber AI Labs. For more information, check out our blog post.

Installing

Installing a stable Pyro release

First install PyTorch.

Install via pip:

Python 2.7.*:

pip install pyro-ppl

Python 3.5:

pip3 install pyro-ppl

Install from source:

git clone git@github.com:uber/pyro.gitcd pyro
git checkout master  # master is pinned to the latest releasepip install .

Install with extra packages:

pip install pyro-ppl[extras]  # for running examples/tutorials

Installing Pyro dev branch

For recent features you can install Pyro from source.

To install a compatible CPU version of PyTorch on OSX / Linux, you could use the PyTorch install helper script.

bash scripts/install_pytorch.sh

Alternatively, build PyTorch following instructions in the PyTorch README.

git clone --recursive https://github.com/pytorch/pytorchcd pytorch
git checkout 200fb22  # <---- a well-tested commit

On Linux:

python setup.py install

On OSX:

MACOSX_DEPLOYMENT_TARGET=10.9 CC=clang CXX=clang++ python setup.py install

Finally install Pyro

git clone https://github.com/uber/pyrocd pyro
pip install .

Running Pyro from a Docker Container

Refer to the instructions here.