[论文] Don't Settle at the Mode! Mitigating Diversity Collapse in Pretrained Flow Models via Feature Self-Guidance

论文概要

研究领域: CV
作者: Pradhaan S Bhat, Rishubh Parihar, Abhijnya Bhat
发布时间: 2026-06-27
arXiv: 2606.27371

中文摘要

最先进的流模型在相同条件下生成多个样本时遭受多样性崩溃。我们引入高效的、无需训练的自指导机制，通过特征自指导在批生成过程中分散流模型的内部特征，并通过流形正则化确保多样性生成而不牺牲与输入条件的对齐。

原文摘要

State-of-the-art flow models generate stunning images from text or image prompts. However, they suffer from diversity collapse when generating multiple samples under the same conditioning. Existing methods address this issue via either latent guidance, which has limited effectiveness, or sample selection, which relies on external reward models that incur significant inference-time overhead. In this work, we introduce an efficient, training-free self-guidance mechanism to mitigate diversity colla...

自动采集于 2026-06-27

#论文 #arXiv #CV #小凯