Skip to content

Source paper

Decoupled DMD: CFG Augmentation as the Spear, Distribution Matching as the Shield

Dongyang Liu, Peng Gao, David Liu and 8 more

Dongyang Liu, Peng Gao, David Liu, Ruoyi Du, Zhen Li, Qilong Wu, Xin Jin, Sihan Cao, Shifeng Zhang, Hongsheng Li, Steven Hoi

arXiv 2511.22677 · 202527 upvotes on Hugging Face

Research ideas from this paper (3)

Generated by AI. The paper’s authors did not write these.

AI summary

This paper reveals that the success of Distribution Matching Distillation (DMD) in few-step generation is driven by two decoupled mechanisms: CFG Augmentation (CA), which acts as the 'engine' for conversion, and Distribution Matching (DM), which acts as a stabilizing 'regularizer'. By identifying this division of labor, the authors propose a Decoupled-Hybrid re-noising schedule that focuses the CA engine on unresolved noise levels while keeping the DM regularizer global, leading to state-of-the-art results in 4-step and 8-step image synthesis.

How it works (AI summary)

The researchers perform a mathematical decomposition of the DMD gradient into two distinct terms: a CA term (the difference between conditional and unconditional scores) and a DM term (the difference between real teacher and auxiliary student scores). This decomposition allows them to implement independent re-noising noise schedules (tau) for each component, optimizing the distillation process by tailoring the noise level at which each 'force' is applied to the student model.