A Few-Step Generative Model on Cumulative Flow Maps
This work provides a general framework to accelerate diffusion- and flow-based generative models for practitioners seeking faster inference without sacrificing quality.
The paper introduces a unified few-step generative modeling framework based on cumulative flow maps, enabling few-step and even one-step generation while preserving synthesis quality across diverse tasks such as image generation, geometric distribution modeling, joint prediction, and SDF generation, with reduced inference cost.
We propose a unified, few-step generative modeling framework based on \emph{cumulative flow maps} for long-range transport in probability space, inspired by flow-map techniques for physical transport and dynamics. At its core is a cumulative-flow abstraction that connects local, instantaneous updates with finite-time transport, enabling generative models to reason about global state transitions. This perspective yields a unified few-step framework built on cumulative transport and \revise{cumulative} parameterization that applies broadly to existing diffusion- and flow-based models without being tied to a specific prediction \revise{instantiation}. Our formulation supports few-step and even one-step generation while preserving synthesis quality, requiring only minimal changes to time embeddings and training objectives, and no increase in model capacity. We demonstrate its effectiveness across diverse tasks, including image generation, geometric distribution modeling, joint prediction, and SDF generation, with reduced inference cost.