改进可控生成：通过 x0 监督实现更快训练与更优性能
Improving Controllable Generation: Faster Training and Better Performance via x0-Supervision
把可控扩散/流模型的训练目标从噪声预测损失换成对干净目标图的 x0 监督（或等价的损失重加权），可直接套用到 ControlNet 类布局控制、inpainting、参考图条件生成等训练流程中，在不改网络结构的前提下加速收敛并提升条件精度。
原文依据 direct supervision on the clean target image, dubbed x_0-supervision, or an equivalent re-weighting of the diffusion loss, yields faster convergence
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论文本到图像（T2I）扩散/流模型近期在视觉保真度与文本对齐方面取得了显著进展。然而，当用户需要精确控制图像布局时，这些模型仍然存在局限，因为仅靠自然语言无法可靠地表达布局。可控生成方法通过为初始 T2I 模型增加更容易描述场景的额外条件来增强模型。已有工作直接沿用初始网络的损失来训练增强后的网络。这种做法乍看之下很自然，但在某些情况下会导致收敛前需要很长的训练时间。在本工作中，我们通过细致分析可控扩散模型的去噪动力学，重新审视其训练目标。我们表明，对干净目标图像施加直接监督（称为 x0 监督），或对扩散损失做等价的重新加权，能够带来更快的收敛。在多种控制设置上的实验表明，根据我们提出的新指标（平均收敛曲线下面积，mAUCC），我们的形式化方法可将收敛速度提升至多 2 倍，同时改善视觉质量与条件精度。代码见 https://github.com/CEA-LIST/x0-supervision。
英文原文摘要 1230 字符
Text-to-Image (T2I) diffusion/flow models have recently achieved remarkable progress in visual fidelity and text alignment. However, they remain limited when users need to precisely control image layouts, something that natural language alone cannot reliably express. Controllable generation methods augment the initial T2I model with additional conditions that more easily describe the scene. Prior works straightforwardly train the augmented network with the same loss as the initial network. Although natural at first glance, this can lead to very long training times in some cases before convergence. In this work, we revisit the training objective of controllable diffusion models through a detailed analysis of their denoising dynamics. We show that direct supervision on the clean target image, dubbed x_0-supervision, or an equivalent re-weighting of the diffusion loss, yields faster convergence. Experiments on multiple control settings demonstrate that our formulation accelerates convergence by up to 2xaccording to our novel metric (mean Area Under the Convergence Curve - mAUCC), while also improving both visual quality and conditioning accuracy. Our code is available at https://github.com/CEA-LIST/x0-supervision.