Uni-DAD：面向少步少样本图像生成的扩散模型统一蒸馏与适配
Uni-DAD: Unified Distillation and Adaptation of Diffusion Models for Few-step Few-shot Image Generation
把扩散蒸馏与域适配合并成单阶段训练：用双域分布匹配蒸馏（源教师+目标教师）保住源域多样性，再用多头GAN损失在少样本下稳定训练，可直接用于少样本域适配与主体驱动个性化，在少于4步采样下逼近两阶段流程的质量。
原文依据 a dual-domain distribution-matching distillation (DMD) objective that guides the student toward the distributions of the source teacher and a target teacher
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论扩散模型（DMs）能生成高质量图像，但当其被适配到新领域时，采样代价依然高昂。蒸馏后的扩散模型速度更快，但通常仍被限制在教师模型的领域之内。因此，面向新领域的快速高质量生成依赖于两阶段流程：先适配再蒸馏，或先蒸馏再适配。然而，这两种做法都增加了设计复杂度，且常常损害质量或多样性。我们提出 Uni-DAD，一个将扩散模型蒸馏与适配统一起来的单阶段流程。它耦合两种训练信号：(i) 双域分布匹配蒸馏（DMD）目标，引导学生模型逼近源教师模型与目标教师模型的分布；(ii) 多头生成对抗网络（GAN）损失，在多个特征尺度上促进目标域的真实感。源域蒸馏保留了多样化的源知识，而多头 GAN 稳定了训练并减少过拟合，在少样本情形下尤为明显。引入目标教师模型有助于适配到结构上差异更大的领域。我们在两个面向少样本图像生成（FSIG）与主体驱动个性化（SDP）的综合基准上、使用扩散骨干网络对 Uni-DAD 进行了评估。即使在少于 4 步采样的情况下，它也能达到与最优（SoTA）适配方法相当或更好的质量，并常常在质量与多样性上超越两阶段流程。代码：https://github.com/yaramohamadi/uni-DAD 。
英文原文摘要 1479 字符
Diffusion models (DMs) produce high-quality images, yet their sampling remains costly when adapted to new domains. Distilled DMs are faster but typically remain confined within their teacher's domain. Thus, fast and high-quality generation for novel domains relies on two-stage pipelines: Adapt-then-Distill or Distill-then-Adapt. However, both add design complexity and often degrade quality or diversity. We introduce Uni-DAD, a single-stage pipeline that unifies DM distillation and adaptation. It couples two training signals: (i) a dual-domain distribution-matching distillation (DMD) objective that guides the student toward the distributions of the source teacher and a target teacher, and (ii) a multi-head generative adversarial network (GAN) loss that encourages target realism across multiple feature scales. The source domain distillation preserves diverse source knowledge, while the multi-head GAN stabilizes training and reduces overfitting, especially in few-shot regimes. The inclusion of a target teacher facilitates adaptation to more structurally distant domains. We evaluate Uni-DAD on two comprehensive benchmarks for few-shot image generation (FSIG) and subject-driven personalization (SDP) using diffusion backbones. It delivers better or comparable quality to state-of-the-art (SoTA) adaptation methods even with less than 4 sampling steps, and often surpasses two-stage pipelines in quality and diversity. Code: https://github.com/yaramohamadi/uni-DAD .