Gen Inference and Learning Template Library (GenTL)

Work-in-progress template inference and learning library in C++ based on probabilistic programming principles

Documentation (work-in-progress)

This is currently a header-only library, and the source code consists of .h files under include/gentl/.

Dependencies

C++ compiler

Most of the library uses C++17 features, which are supported by recent clang++ and g++ versions.

Currently, include/gentl/learning.h uses C++20 features, which requires g++ version 11 or above, and clang++ version 11 or above.

CMake

CMake 3.7 or above. See Installing CMake.

Doxygen

This is only required for generating documentation.

See Doxygen Installation.

Testing

Use CMake to build and test. Note that you can configure the C++ compiler to use with either the CXX environment variable or by setting the CMAKE_CXX_COMPILER variable, e.g.:

cmake -S . -B build -DCMAKE_CXX_COMPILER=g++-11
cmake --build build
cmake --build build --target test

Generating documentation

cmake --build --target docs

Note that there is a Github Action that generates and deploys documentation to https://github.com/OpenGen/gentl-docs/ when there is a push to the main branch of this repository.

References

Gen: A General-Purpose Probabilistic Programming System with Programmable Inference. Cusumano-Towner, M. F.; Saad, F. A.; Lew, A.; and Mansinghka, V. K. In Proceedings of the 40th ACM SIGPLAN Conference on Programming Language Design and Implementation (PLDI ‘19). (pdf) (bibtex)

Marco Cusumano-Towner's PhD thesis.