PeRFlow: Piecewise Rectified Flow as Universal Plug-and-Play Accelerator

TL;DR

PeRFlow introduces a novel piecewise rectified flow method that accelerates diffusion model sampling by straightening trajectories, enabling fast, transferable, and plug-and-play generative acceleration.

Contribution

The paper proposes PeRFlow, a universal, plug-and-play acceleration method for diffusion models that leverages piecewise linear flows and inherits knowledge from pretrained models.

Findings

Achieves superior few-step generation performance.

Models inherit knowledge from pretrained diffusion models.

Fast convergence and broad transferability.

Abstract

We present Piecewise Rectified Flow (PeRFlow), a flow-based method for accelerating diffusion models. PeRFlow divides the sampling process of generative flows into several time windows and straightens the trajectories in each interval via the reflow operation, thereby approaching piecewise linear flows. PeRFlow achieves superior performance in a few-step generation. Moreover, through dedicated parameterizations, the PeRFlow models inherit knowledge from the pretrained diffusion models. Thus, the training converges fast and the obtained models show advantageous transfer ability, serving as universal plug-and-play accelerators that are compatible with various workflows based on the pre-trained diffusion models. Codes for training and inference are publicly released. https://github.com/magic-research/piecewise-rectified-flow