RayFlow：基於自適應流軌跡的實例感知擴散加速

RayFlow: Instance-Aware Diffusion Acceleration via Adaptive Flow Trajectories

March 10, 2025

作者: Huiyang Shao, Xin Xia, Yuhong Yang, Yuxi Ren, Xing Wang, Xuefeng Xiao

cs.AI

摘要

擴散模型在多個領域取得了顯著成功。然而，其生成速度緩慢仍然是一個關鍵挑戰。現有的加速方法雖然旨在減少步驟，但往往會犧牲樣本質量、可控性，或引入訓練複雜性。因此，我們提出了RayFlow，這是一種新穎的擴散框架，旨在解決這些限制。與以往方法不同，RayFlow引導每個樣本沿著一條獨特的路徑朝向特定實例的目標分佈。這種方法在最小化採樣步驟的同時，保持了生成的多樣性和穩定性。此外，我們引入了時間採樣器，這是一種重要性採樣技術，通過專注於關鍵時間步來提高訓練效率。大量實驗表明，與現有的加速技術相比，RayFlow在生成高質量圖像方面具有更快的速度、更好的控制和更高的訓練效率。

English

Diffusion models have achieved remarkable success across various domains. However, their slow generation speed remains a critical challenge. Existing acceleration methods, while aiming to reduce steps, often compromise sample quality, controllability, or introduce training complexities. Therefore, we propose RayFlow, a novel diffusion framework that addresses these limitations. Unlike previous methods, RayFlow guides each sample along a unique path towards an instance-specific target distribution. This method minimizes sampling steps while preserving generation diversity and stability. Furthermore, we introduce Time Sampler, an importance sampling technique to enhance training efficiency by focusing on crucial timesteps. Extensive experiments demonstrate RayFlow's superiority in generating high-quality images with improved speed, control, and training efficiency compared to existing acceleration techniques.