Bi-Anchor Interpolation Solver for Accelerating Generative Modeling 文章

ArXiv CS.CV2026-06-19PAPERen作者: Hongxu Chen, Hongxiang Li, Zhen Wang, Long Chen

详细信息

来源站点
ArXiv CS.CV
作者
Hongxu Chen, Hongxiang Li, Zhen Wang, Long Chen
文章类型
PAPER
语言
en
发布日期
2026-06-19

摘要

arXiv:2601.21542v3 Announce Type: replace Abstract: Flow Matching (FM) models have emerged as a leading paradigm for high-fidelity synthesis. However, their reliance on iterative Ordinary Differential Equation (ODE) solving creates a significant latency bottleneck. Existing solutions face a dichotomy: training-free solvers suffer from significant performance degradation at low Neural Function Evaluations (NFEs), while training-based one- or few-steps generation methods incur prohibitive training costs and lack plug-and-play versatility. To bridge this gap, we propose the Bi-Anchor Interpolation Solver (BA-solver). BA-solver retains the versatility of standard training-free solvers while achieving significant acceleration by introducing a lightweight SideNet (1-2% backbone size) alongside the frozen backbone.