发表机构
Peking University; Melon Group; Tsinghua University; Alibaba Group（北京大学; Melon集团; 清华大学; 阿里巴巴集团）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
CrossDistill提出轨迹级混合蒸馏框架，通过噪声区间划分与交叉耦合平衡质量与多样性，在文本到视频扩散模型上扩展了少步质量-多样性前沿。
AI 中文摘要
少步蒸馏加速了扩散模型，但必须在多样性与保真度之间取得平衡：基于轨迹的蒸馏保留了模式覆盖，而分布匹配使样本更锐利但可能降低多样性。我们表明，这种张力可以以依赖于噪声区间的方式加以利用：高噪声步骤主要决定全局模式，而低噪声步骤细化局部细节。我们提出CrossDistill，一种轨迹级混合蒸馏框架，它在交叉点处分割采样轨迹，在高噪声区间应用轨迹保持目标，在低噪声区间应用分布匹配目标，并通过交叉状态耦合这两个阶段。与损失级混合相比，并与训练时两阶段方案互补，CrossDistill沿噪声轴显式分配互补目标，从而在局部统计被锐化之前保留全局分支。CrossDistill是一种噪声级调度策略：PCM和DMD是可插拔的实例，而噪声划分、交叉耦合和目标排序是关键设计要素。在文本到视频扩散模型上的实验以及定性的图像到视频结果表明，CrossDistill扩展了少步质量-多样性前沿，在保持种子级变异的同时实现了有竞争力的视觉保真度。
英文摘要
Few-step distillation accelerates diffusion models but must balance diversity and fidelity: trajectory-based distillation preserves mode coverage, while distribution matching sharpens samples but can reduce diversity. We show that this tension can be exploited in a noise-regime-dependent way: high-noise steps largely determine global modes, whereas low-noise steps refine local details. We propose CrossDistill, a trajectory-level hybrid distillation framework that splits the sampling trajectory at a crossover point, applies a trajectory-preserving objective on the high-noise interval and a distribution-matching objective on the low-noise interval, and couples the two stages through the crossover state. In contrast to loss-level mixing, and complementarily to training-time two-stage recipes, CrossDistill explicitly assigns complementary objectives along the noise axis, so that global branching is preserved before local statistics are sharpened. CrossDistill is a noise-level scheduling policy: PCM and DMD are plug-in instantiations, while the noise partition, crossover coupling, and objective ordering are the key design elements. Experiments on text-to-video diffusion models and qualitative image-to-video results show that CrossDistill expands the few-step quality-diversity frontier, retaining seed-level variation while achieving competitive visual fidelity.
