# OrthoFuse: Training-free Riemannian Fusion of Orthogonal Style-Concept Adapters for Diffusion Models

Source: [OrthoFuse: Training-free Riemannian Fusion of Orthogonal Style-Concept Adapters for Diffusion Models](<https://www.sota2.com/research/paper/69d5e1287d0172140ec75b9c>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-09T05:00:04.534Z

Authors: Ali Aliev, Kamil Garifullin, Nikolay Yudin, Vera Soboleva, Alexander Molozhavenko, Ivan Oseledets, Aibek Alanov, Maxim Rakhuba

Published: 2026-04-06T21:24:21.000Z

## Abstract

In a rapidly growing field of model training there is a constant practical interest in parameter-efficient fine-tuning and various techniques that use a small amount of training data to adapt the model to a narrow task. However, there is an open question: how to combine several adapters tuned for different tasks into one which is able to yield adequate results on both tasks? Specifically, merging subject and style adapters for generative models remains unresolved. In this paper we seek to show that in the case of orthogonal fine-tuning (OFT), we can use structured orthogonal parametrization and its geometric properties to get the formulas for training-free adapter merging. In particular, we derive the structure of the manifold formed by the recently proposed Group-and-Shuffle ($\\mathcal{GS}$) orthogonal matrices, and obtain efficient formulas for the geodesics approximation between two points. Additionally, we propose a $\\text{spectra restoration}$ transform that restores spectral properties of the merged adapter for higher-quality fusion. We conduct experiments in subject-driven generation tasks showing that our technique to merge two $\\mathcal{GS}$ orthogonal matrices is capable of uniting concept and style features of different adapters. To the best of our knowledge, this is the first training-free method for merging multiplicative orthogonal adapters. Code is available via the $\\href{https://github.com/ControlGenAI/OrthoFuse}{link}$.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Subject-driven Stylized Image Generation on DreamBooth (StyleDrop/K-LoRA) (test)](<https://www.sota2.com/research/sota/subject-driven-stylized-image-generation-on-dreambooth-styledrop-k-lora-test.md>): Subject-driven Stylized Image Generation; DreamBooth (concepts) + StyleDrop and K-LoRA (styles) (test); Style Similarity: 61.

- [Subject-driven Stylized Image Generation on Concept-Style Pairs User Study (test)](<https://www.sota2.com/research/sota/subject-driven-stylized-image-generation-on-concept-style-pairs-user-study-test.md>): Subject-driven Stylized Image Generation; Concept-Style Pairs User Study (test); Concept Preservation: 54.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2604.05183>)

[PDF](<https://arxiv.org/pdf/2604.05183.pdf>)
