SPFlow

SPFlow is a Python library for building and learning Probabilistic Circuits (PCs) with PyTorch.

Probabilistic Circuits are a flexible and expressive family of probabilistic graphical models that support both exact and approximate inference in linear time for a wide range of queries.

Quick Navigation

• Getting Started — Installation and quick start guide

• Concepts — Core SPFlow concepts and semantics

• Guides — End-to-end tutorials and workflow guides

• API Reference — Complete API documentation

• Paper Zoo — Research implementations and papers (PIC, SOCS, RAT-SPN, etc.)

• FAQ — Frequently asked questions

Indices and tables

• Index

• Module Index

• Search Page