Learning Straight Flows: Variational Flow Matching for Efficient Generation

TL;DR

This paper introduces S-VFM, a novel flow matching method that explicitly enforces straight trajectories for efficient one-step generation, improving training stability and inference speed.

Contribution

S-VFM integrates a variational latent code into flow matching to explicitly promote straight, linear trajectories, addressing limitations of previous methods.

Findings

Achieves competitive results on three benchmarks.

Demonstrates improved training stability.

Offers faster inference compared to existing methods.

Abstract

Flow Matching has limited ability in achieving one-step generation due to its reliance on learned curved trajectories. Previous studies have attempted to address this limitation by either modifying the coupling distribution to prevent interpolant intersections or introducing consistency and mean-velocity modeling to promote straight trajectory learning. However, these approaches often suffer from discrete approximation errors, training instability, and convergence difficulties. To tackle these issues, in the present work, we propose \textbf{S}traight \textbf{V}ariational \textbf{F}low \textbf{M}atching (\textbf{S-VFM}), which integrates a variational latent code representing the ``generation overview'' into the Flow Matching framework. \textbf{S-VFM} explicitly enforces trajectory straightness, ideally producing linear generation paths. The proposed method achieves competitive performance across three challenge benchmarks and demonstrates advantages in both training and inference efficiency compared with existing methods.