发表机构
Tsinghua University; Fudan University; Peking University; Zhejiang University; DeepSeek-AI; ByteDance Seed; University of California, Berkeley; BAAI（清华大学; 复旦大学; 北京大学; 浙江大学; 深度求索人工智能; 字节跳动Seed; 加利福尼亚大学伯克利分校; 北京智源人工智能研究院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
提出统一分布训练框架及MGFlow方法，通过高斯混合建模与最优传输匹配，实现一步式视觉生成，在ImageNet和文本到图像任务上超越基线。
AI 中文摘要
分布训练通过在冻结的表示空间中匹配真实特征与生成特征，为一步式视觉生成提供集体监督。我们引入了一个统一的理论框架，该框架将分布建模与匹配差异分离，并通过Wasserstein梯度流将全局目标与逐点特征更新联系起来。在此框架下，FD-Loss和高斯核漂移分别通过高斯最优传输和基于核密度的KL匹配得以恢复。该框架催生了MGFlow，它使用高斯混合模型以可调节的粒度（介于全局矩和基于样本的表示之间）对特征分布进行建模。MGFlow支持最优传输和基于分数的匹配，并将质量约束的样本分配与成对分量更新相结合，以解决混合表达力本身无法解决的模式坍缩问题。在ImageNet 256×256上，MGFlow显著超越了FD-Loss基线，在pMF-H上取得了1.45 FDr^6的最先进结果，在JiT-H上取得了1.64的结果。对于文本到图像生成，MGFlow将FLUX.2 [klein] 4B后训练为一步式生成器，在GenEval和PickScore上均优于原始的四步模型。项目页面：此https URL
英文摘要
Distributional training provides collective supervision for one-step visual generation by matching real and generated features in frozen representation spaces. We introduce a unified theoretical framework that separates distribution modeling from matching discrepancy and connects global objectives to pointwise feature updates through Wasserstein gradient flow. Under this framework, FD-Loss and Gaussian-kernel Drifting are recovered through Gaussian optimal transport and kernel-density-based KL matching, respectively. The framework motivates MGFlow, which models feature distributions with Gaussian mixtures at an adjustable granularity between global moments and sample-based representations. MGFlow supports both optimal transport and score-based matching, and couples mass-constrained sample assignment with paired component updates to address mode collapse that mixture expressivity alone does not resolve. On ImageNet $256\times256$, MGFlow substantially surpasses the FD-Loss baseline, achieving state-of-the-art results with 1.45 $\mathrm{FDr}^6$ on pMF-H and 1.64 on JiT-H. For text-to-image generation, MGFlow post-trains FLUX.2 [klein] 4B into a one-step generator that outperforms the original four-step model on both GenEval and PickScore.
CommentsProject page: https://shihaoyang0423.github.io/MGFlow-website/
