ahmadsalim / ppx

Probabilistic Programming eXecution protocol (PPX)

PPX

PPX is a cross-platform Probabilistic Programming eXecution protocol and API based on flatbuffers. It is intended as an open interoperability protocol between models and inference engines implemented in different probabilistic programming languages.

Probabilistic programming is about the execution probabilistic models under the control of inference engines, and PPX allows the model and the inference engine to be

• implemented in different programming languages and
• executed in separate processes and on separate machines across networks.

PPX is inspired by ONNX, the Open Neural Network Exchange project allowing interoperability between major deep learning frameworks.

Supported languages

We provide PPX compiled to all programming languages officially supported by flatbuffers. These are:

• C++
• C#
• Dart
• Go
• Java
• JavaScript
• Kotlin
• Lobster
• Lua
• PHP
• Python
• Rust
• Swift
• TypeScript

Build

In order to recompile PPX to all supported languages, you can clone this repository and run the flatbuffers compiler in a Docker container as follows.

Prerequisite: Install Docker for your system.

git clone https://github.com/pyprob/ppx.git
cd ppx
docker build -t ppx .
export UID=$(id -u)
export GID=$(id -g)
docker run --user $UID:$GID --rm -it -v $PWD:/home ppx sh compile.sh

License

PPX is distributed under the BSD License.

Authors

PPX has been developed by

• Atılım Güneş Baydin
• Tuan Anh Le
• Lukas Heinrich
• Bradley Gram-Hansen
• Christian Schroeder de Witt
• Wahid Bhimji
• Kyle Cranmer
• Frank Wood