[论文] Inference-Time Scaling of Diffusion Models via Progressive Seed Prunin...

论文概要

研究领域: CV 作者: Rogerio Guimaraes, Pietro Perona 发布时间: 2026-07-25 arXiv: 2507.20484

中文摘要

扩散模型和流匹配模型在条件图像生成中占据主导地位，但这些模型的推理时扩展远不如自回归语言模型发展成熟。由于最终质量对初始噪声种子高度敏感，许多方法在种子搜索或黑盒奖励下的重采样上花费额外计算，但通常在推理过程中保持恒定的内存占用。我们表明，放松这一约束开启了一个未被充分探索的推理时扩展维度：通过前置探索、早期评估多个种子并积极剪枝，我们可以更有效地使用固定的计算预算。渐进式种子剪枝（PSP）对中间去噪估计进行评分，并逐步缩小候选集，使得只有有前景的轨迹被完全去噪，同时保持模型评估总数不变。在扩散和流匹配主干网络上，PSP始终优于奖励引导选择，在相同计算量下比best-of-N、重要性采样和树搜索基线获得更高的GenEval分数（自动评估）和更好的人类评估提示对齐度。项目页面：https://www.vision.caltech.edu/psp。代码：https://github.com/rogerioagjr/psp。

原文摘要

Diffusion and flow-matching models dominate conditional image generation, yet inference-time scaling for these models is far less developed than for autoregressive language models. Because final quality is highly sensitive to the initial noise seed, many approaches spend extra compute on seed search or resampling under a black-box reward, but typically maintaining a constant memory footprint throughout inference. We show that relaxing this constraint enables an underexplored inference-time scaling axis: by front-loading exploration, evaluating many seeds early, and pruning aggressively, we can use a fixed compute budget more effectively. Progressive Seed Pruning (PSP) scores intermediate denoised estimates and progressively narrows the candidate set so that only promising trajectories are ...

--- *自动采集于 2026-07-26*

#论文 #arXiv #CV #小凯