分布匹配蒸馏中的谱幅纯化用于扩散蒸馏
Spectral Amplitude Purification in Distribution Matching for Diffusion Distillation
浏览论文内容
中文总结 AI 辅助
针对分布匹配蒸馏中低频主导导致细节恢复慢的问题，提出谱幅纯化方法SAP-DMD，通过自适应抑制幅值谱尾部，在PixArt-α、SD3和SD3.5上加速收敛并提升2步和4步采样的生成质量。
中文摘要 AI 辅助
分布匹配蒸馏（DMD）能够仅用几步实现高质量的扩散采样，但其优化动态仍受粗粒度、低频信号主导，延迟了细粒度细节的恢复。我们发现在DMD方向误差中，谱幅在低频处显著集中，占主导地位的低频分量压倒了较弱的中频和高频信号。为解决此问题，我们提出了用于分布匹配蒸馏的谱幅纯化（SAP-DMD），这是一种即插即用的方法，可自适应地调节DMD方向场的幅值谱。通过抑制幅值谱的主导尾部，SAP-DMD减少了低频主导性，并促进了细结构和纹理的更有效恢复。在PixArt-α、SD3和SD3.5上的实验表明，SAP-DMD在2步和4步采样下均加速了训练收敛并提高了生成质量。
英文摘要
Distribution Matching Distillation (DMD) enables high-quality diffusion sampling in only a few steps, but its optimization dynamics remain dominated by coarse, low-frequency signals, delaying the recovery of fine-grained details. We identify a pronounced concentration of spectral amplitudes at low frequencies in the DMD directional error, where dominant low-frequency components overwhelm weaker mid- and high-frequency signals. To address this issue, we propose Spectral Amplitude Purification for Distribution Matching Distillation (SAP-DMD), a plug-and-play approach that adaptively modulates the amplitude spectrum of the DMD directional field. By suppressing the dominant tail of the amplitude spectrum, SAP-DMD reduces low-frequency dominance and promotes more effective recovery of fine structures and textures. Experiments on PixArt-$α$, SD3, and SD3.5 demonstrate that SAP-DMD accelerates training convergence and improves generation quality under both 2-step and 4-step sampling.
发表机构
- Zhejiang University（浙江大学）
- University of California, Berkeley（加州大学伯克利分校）
机构由 AI 辅助整理，请以论文原文为准。
