pipette
ENEnglish

LEAP: Accelerating ODE-based Diffusion Sampling with Optimal Linear Leaps

Haowei Zheng, Zhenyu Zhou, Jiawei Chen, Can Wang, Defang Chen

Preprint con versión publicadaCódigo disponible

En palabras de los autores

Diffusion models primarily rely on general-purpose ODE-based numerical methods for training-free accelerated sampling, but performance degrades with coarse time discretization. In this paper, we propose optimal linear leaps (LEAP), a specialized method grounded in theoretical and empirical investigations of sampling trajectory geometry. Built upon low-dimensionality and cross-trajectory consistency, LEAP formulates a supervised least-squares optimization problem within the local trajectory subspace, achieving a precise approximation of future states without model evaluations. Extensive experiments demonstrate that LEAP outperforms previous methods across diverse modalities, including text-to-image (Stable Diffusion 3.5, FLUX.1-dev) and image-to-3D (TRELLIS.2) generation. Furthermore, LEAP can be seamlessly integrated into existing solver-based methods, offering a plug-and-play enhancement for modern diffusion models. Code is available at https://github.com/zju-pi/LEAP.

Resultado principalEl resumen no menciona limitaciones.

Apareció: martes, 20 de octubre. arXiv. Preprint con versión publicada.

DOI: 10.1145/3767308.3836106

Versión publicada: 10.1145/3767308.3836106