Guiding a Diffusion Model by Swapping Its Tokens

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zhang, Weijia, Liu, Yuehao, Guan, Shanyan, Ran, Wu, Ge, Yanhao, Li, Wei, Ma, Chao
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917395196542976
author Zhang, Weijia
Liu, Yuehao
Guan, Shanyan
Ran, Wu
Ge, Yanhao
Li, Wei
Ma, Chao
author_facet Zhang, Weijia
Liu, Yuehao
Guan, Shanyan
Ran, Wu
Ge, Yanhao
Li, Wei
Ma, Chao
contents Classifier-Free Guidance (CFG) is a widely used inference-time technique to boost the image quality of diffusion models. Yet, its reliance on text conditions prevents its use in unconditional generation. We propose a simple method to enable CFG-like guidance for both conditional and unconditional generation. The key idea is to generate a perturbed prediction via simple token swap operations, and use the direction between it and the clean prediction to steer sampling towards higher-fidelity distributions. In practice, we swap pairs of most semantically dissimilar token latents in either spatial or channel dimensions. Unlike existing methods that apply perturbation in a global or less constrained manner, our approach selectively exchanges and recomposes token latents, allowing finer control over perturbation and its influence on generated samples. Experiments on MS-COCO 2014, MS-COCO 2017, and ImageNet datasets demonstrate that the proposed Self-Swap Guidance (SSG), when applied to popular diffusion models, outperforms previous condition-free methods in image fidelity and prompt alignment under different set-ups. Its fine-grained perturbation granularity also improves robustness, reducing side-effects across a wider range of perturbation strengths. Overall, SSG extends CFG to a broader scope of applications including both conditional and unconditional generation, and can be readily inserted into any diffusion model as a plug-in to gain immediate improvements.
format Preprint
id arxiv_https___arxiv_org_abs_2604_08048
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Guiding a Diffusion Model by Swapping Its Tokens
Zhang, Weijia
Liu, Yuehao
Guan, Shanyan
Ran, Wu
Ge, Yanhao
Li, Wei
Ma, Chao
Computer Vision and Pattern Recognition
Classifier-Free Guidance (CFG) is a widely used inference-time technique to boost the image quality of diffusion models. Yet, its reliance on text conditions prevents its use in unconditional generation. We propose a simple method to enable CFG-like guidance for both conditional and unconditional generation. The key idea is to generate a perturbed prediction via simple token swap operations, and use the direction between it and the clean prediction to steer sampling towards higher-fidelity distributions. In practice, we swap pairs of most semantically dissimilar token latents in either spatial or channel dimensions. Unlike existing methods that apply perturbation in a global or less constrained manner, our approach selectively exchanges and recomposes token latents, allowing finer control over perturbation and its influence on generated samples. Experiments on MS-COCO 2014, MS-COCO 2017, and ImageNet datasets demonstrate that the proposed Self-Swap Guidance (SSG), when applied to popular diffusion models, outperforms previous condition-free methods in image fidelity and prompt alignment under different set-ups. Its fine-grained perturbation granularity also improves robustness, reducing side-effects across a wider range of perturbation strengths. Overall, SSG extends CFG to a broader scope of applications including both conditional and unconditional generation, and can be readily inserted into any diffusion model as a plug-in to gain immediate improvements.
title Guiding a Diffusion Model by Swapping Its Tokens
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.08048