RayFlow: Instance-Aware Diffusion Acceleration via Adaptive Flow Trajectories
Huiyang Shao, Xin Xia, Yuhong Yang, Yuxi Ren, Xing Wang, Xuefeng Xiao
2025-03-12
Summary
This paper talks about RayFlow, a new AI tool that speeds up image generation by creating custom paths for each image part, like drawing detailed faces faster than plain backgrounds, without losing quality.
What's the problem?
Current AI image generators are slow because they treat all parts of an image the same way, even simple areas, and speeding them up usually makes images look worse or requires complex retraining.
What's the solution?
RayFlow gives each part of an image its own optimized path to follow during generation and uses a smart time-skipping trick to focus on the most important steps, cutting down the work needed.
Why it matters?
This makes AI image tools faster for artists, game designers, and apps, letting them create high-quality visuals quickly without needing expensive computers.
Abstract
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.