LeMiCa: Lexicographic Minimax Path Caching for Efficient Diffusion-Based Video Generation

NeurIPSSpotlight2025

Authors: Huanlin Gao, Ping Chen, Fuyuan Shi, Chao Tan, Zhaoxiang Liu, Fang Zhao, Kai Wang, Shiguo Lian
Venue: NeurIPS 2025
Track: Spotlight

TL;DR

We present LeMiCa, a training-free and efficient acceleration framework for diffusion-based video generation…

Opening excerpt from the authors’ abstract. source

Read the paper

OpenReview

Topics

diffusion efficient graph video

← All NeurIPS 2025 Spotlight papers · Browse the whole archive