← 研踪 PaperTrailCVF · CVPR 2026:cvf:fe0f837dd5e389f9d641873aaf81f9e9 · 记录 #1
Learning Straight Flows: Variational Flow Matching for Efficient Generation
来源日期(仅到年份) 2026
来源记录日期 2026
雷达观测 2026-11-05 02:36
摘要
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 Straight Variational Flow Matching (S-VFM), which integrates a variational latent code representing the "generation overview" into the Flow Matching framework. 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.
论文发现出处
这些页面提供论文线索,不代表同行评审或质量认证。
观测到的版本
平台记录号代表本站观测到的元数据变化,不是作者修订号。来源日期仅到月份或年份时,按该月或该年首日存档。
记录 #1 · 2026 · 最新观测版本
Learning Straight Flows: Variational Flow Matching for Efficient Generation
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 Straight Variational Flow Matching (S-VFM), which integrates a variational latent code representing the "generation overview" into the Flow Matching framework. 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.
查看此版本解读 →查看此版本原文 ↗