PhysicsNeMo Diffusion# PhysicsNeMo Diffusion Introduction Framework Components at a Glance Design Philosophy: Layered Customization Core Concepts: DiffusionModel, Predictor, and Denoiser Prediction Types API Reference Model Backbones UNet-Based Architectures Diffusion Transformer (DiT) Adapting Backbones to the DiffusionModel Interface Using External or Custom Models Noise Schedulers Role in Training Role in Inference (Sampling) Three Levels of Customization API Reference Preconditioners Three Approaches How Preconditioners Fit in the Pipeline API Reference Metrics and Losses Training Losses Evaluation Metrics API Reference Samplers and Solvers Generic Sampling Process Sampling Workflow Available Solvers Guidance API Reference Multi-Diffusion Utilities API Reference
