PaperScope
LIVE · 2026-09-29 05:40 UTC

Scaffold Then Internalize: Representation Injection for Diffusion Transformers

Han Fu, Jiacheng Chen, Baoquan Zhao, Weidong Chen, Wei Liu, Qing Li, Xudong Mao

Latestcs.CLcs.LGcs.AIcs.CV
arXiv ID
2609.35292 v1
Category
Submitted
2026-09-28

Abstract

Recent representation alignment (REPA) methods accelerate diffusion transformer training by aligning projections of the transformer's hidden states with representations from pretrained visual encoders. In this work, we explore a reverse and complementary direction to REPA: rather than projecting diffusion representations into the encoder's space, we inject encoder representations into the diffusion transformer, allowing them to actively participate in the denoising process. To this end, we introduce \textit{REPresentation Injection} (REPI), a training framework based on a scaffold-to-internalization strategy, in which projected encoder representations initially serve as a temporary scaffold and are then progressively internalized by the diffusion transformer. REPI outperforms REPA across a wide range of backbones and is highly complementary to it: combining the two yields substantial gains over either alone. Notably, with only 160K training steps, REPI + REPA matches vanilla SiT trained for 7M steps, a speedup of over $43.5\times$. Code will be available at https://jeneveuxpas.github.io/REPI

arXiv abs page · PDF · same-day batch