arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2302.06504cs.CV
收藏

预条件得分生成模型

Preconditioned Score-based Generative Models

Fudan University(复旦大学);University of Surrey(萨里大学)

Hengyuan Ma, Xiatian Zhu, Jianfeng Feng, Li Zhang

AI总结:

针对基于得分的生成模型采样缓慢且减少迭代导致性能下降的问题,提出预条件扩散采样方法,通过矩阵预条件缓解病态问题,无需重训练,理论证明保持输出分布并揭示参数与迭代关系,实验显示最高28倍加速且保持合成质量。

摘要:

基于得分的生成模型(SGMs)最近已成为一类有前景的生成模型。然而,其根本限制在于采样过程缓慢,因为需要许多(例如2000次)顺序计算的迭代。一种直观的加速方法是减少采样迭代次数,但这会导致严重的性能下降。我们将此问题归因于采样过程中Langevin动力学和反向扩散的病态问题。基于这一见解,我们提出了一种新颖的预条件扩散采样(PDS)方法,利用矩阵预条件来缓解上述问题。PDS在边际额外计算成本下改变普通SGM的采样过程,且无需重新训练模型。理论上,我们证明PDS保持了SGM的输出分布,不会对原始采样过程引入系统性偏差。我们进一步从理论上揭示了PDS参数与采样迭代次数之间的关系,便于在不同采样迭代次数下估计参数。在各种分辨率和多样性的图像数据集上进行的大量实验验证了PDS能够持续加速现成的SGM,同时保持合成质量。特别是,在更具挑战性的高分辨率(1024x1024)图像生成中,PDS可实现高达28倍的加速。与最新的生成模型(如CLD-SGM和Analytic-DDIM)相比,PDS在CIFAR-10上以1.99的FID分数实现了最佳采样质量。我们的代码公开可用,以促进进一步研究 https://github.com/fudan-zvg/PDS。

英文摘要:

Score-based generative models (SGMs) have recently emerged as a promising class of generative models. However, a fundamental limitation is that their sampling process is slow due to a need for many (e.g., 2000) iterations of sequential computations. An intuitive acceleration method is to reduce the sampling iterations which however causes severe performance degradation. We assault this problem to the ill-conditioned issues of the Langevin dynamics and reverse diffusion in the sampling process. Under this insight, we propose a novel preconditioned diffusion sampling (PDS) method that leverages matrix preconditioning to alleviate the aforementioned problem. PDS alters the sampling process of a vanilla SGM at marginal extra computation cost and without model retraining. Theoretically, we prove that PDS preserves the output distribution of the SGM, with no risk of inducing systematical bias to the original sampling process. We further theoretically reveal a relation between the parameter of PDS and the sampling iterations, easing the parameter estimation under varying sampling iterations. Extensive experiments on various image datasets with a variety of resolutions and diversity validate that our PDS consistently accelerates off-the-shelf SGMs whilst maintaining the synthesis quality. In particular, PDS can accelerate by up to 28x on more challenging high-resolution (1024x1024) image generation. Compared with the latest generative models (e.g., CLD-SGM and Analytic-DDIM), PDS can achieve the best sampling quality on CIFAR-10 at an FID score of 1.99. Our code is publicly available to foster any further research https://github.com/fudan-zvg/PDS.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑