Skip to content

API Reference

Package Structure

  • Core

    Base classes, energy models, schedulers, and device management.

  • Samplers

    MCMC, optimization, and flow/diffusion sampling.

  • Losses

    Training objectives for energy-based and flow-based models.

  • Interpolants

    Noise-to-data interpolation schemes.

  • Integrators

    Numerical integrators for SDE, ODE, and Hamiltonian dynamics.

  • Models

    Neural network architectures.

  • Datasets

    Synthetic data generators.

  • Utils

    Visualization and training utilities.

  • CUDA

    Accelerated implementations.

Each module page lists every public class and function with full signatures, parameters, and usage notes. Start from the module that matches your task, or browse the base classes in Core to understand the shared interfaces.