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

论文概要

研究领域: CV
作者: Rogerio Guimaraes, Pietro Perona
发布时间: 2026-07-24
arXiv: 2507.19318

中文摘要

扩散模型和流匹配模型主导了条件图像生成，但这些模型的推理时间缩放远不如自回归语言模型发达。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，放宽这一约束可以启用一个未被充分探索的推理时间缩放维度：通过在早期加载探索、评估大量种子并积极剪枝，我们可以更有效地使用固定计算预算。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固定。在扩散和流匹配主干网络上，PSP持续改进了奖励引导选择，在匹配计算量下取得了比best-of-N、重要性采样和树搜索基线更高的GenEval分数（自动评估）和更好的人工评估提示对齐度。

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 2026-07-25

#论文 #arXiv #CV #小凯