RelaCtrl: Relevance-Guided Efficient Control for Diffusion Transformers
Ke Cao, Jing Wang, Ao Ma, Jiasong Feng, Zhanjie Zhang, Xuanhua He, Shanyuan Liu, Bo Cheng, Dawei Leng, Yuhui Yin, Jie Zhang
TL;DR
RelaCtrl addresses inefficiency in controllable diffusion transformers by introducing a relevance-guided strategy that allocates control blocks to layers with high ControlNet Relevance Score while replacing heavy copy blocks with the lightweight Two-Dimensional Shuffle Mixer. The approach couples a Relevance-Guided Lightweight Control Block with TDSM to reduce parameters and FLOPs without compromising control fidelity, achieving comparable or superior results to state-of-the-art methods. Experimental results across multiple conditional tasks and models show improved control accuracy and image quality with modest resource overhead, and the method generalizes to Flux. This yields a practical, scalable solution for efficient controllable generation in diffusion transformers.
Abstract
The Diffusion Transformer plays a pivotal role in advancing text-to-image and text-to-video generation, owing primarily to its inherent scalability. However, existing controlled diffusion transformer methods incur significant parameter and computational overheads and suffer from inefficient resource allocation due to their failure to account for the varying relevance of control information across different transformer layers. To address this, we propose the Relevance-Guided Efficient Controllable Generation framework, RelaCtrl, enabling efficient and resource-optimized integration of control signals into the Diffusion Transformer. First, we evaluate the relevance of each layer in the Diffusion Transformer to the control information by assessing the "ControlNet Relevance Score"-i.e., the impact of skipping each control layer on both the quality of generation and the control effectiveness during inference. Based on the strength of the relevance, we then tailor the positioning, parameter scale, and modeling capacity of the control layers to reduce unnecessary parameters and redundant computations. Additionally, to further improve efficiency, we replace the self-attention and FFN in the commonly used copy block with the carefully designed Two-Dimensional Shuffle Mixer (TDSM), enabling efficient implementation of both the token mixer and channel mixer. Both qualitative and quantitative experimental results demonstrate that our approach achieves superior performance with only 15% of the parameters and computational complexity compared to PixArt-delta.