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.
First install PyTorch.
Install via pip:
Python 2.7.*:
pip install pyro-pplPython 3.5:
pip3 install pyro-ppl
Install from source:
git clone [email protected]:uber/pyro.git
cd pyro
pip install .Note that to use Pyro's dev branch, you will need to build PyTorch's master branch.
Refer to the README for more details
on how to build PyTorch from source. Otherwise, to use the latest version of Pyro that
is compatible with PyTorch release, use the master branch of Pyro.
Refer to the instructions here.