GimmBO: Interactive Generative Image Model Merging via Bayesian Optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Liu, Chenxi, Ling, Selena, Jacobson, Alec
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917223808892928
author Liu, Chenxi
Ling, Selena
Jacobson, Alec
author_facet Liu, Chenxi
Ling, Selena
Jacobson, Alec
contents Fine-tuning-based adaptation is widely used to customize diffusion-based image generation, leading to large collections of community-created adapters that capture diverse subjects and styles. Adapters derived from the same base model can be merged with weights, enabling the synthesis of new visual results within a vast and continuous design space. To explore this space, current workflows rely on manual slider-based tuning, an approach that scales poorly and makes weight selection difficult, even when the candidate set is limited to 20-30 adapters. We propose GimmBO to support interactive exploration of adapter merging for image generation through Preferential Bayesian Optimization (PBO). Motivated by observations from real-world usage, including sparsity and constrained weight ranges, we introduce a two-stage BO backend that improves sampling efficiency and convergence in high-dimensional spaces. We evaluate our approach with simulated users and a user study, demonstrating improved convergence, high success rates, and consistent gains over BO and line-search baselines, and further show the flexibility of the framework through several extensions.
format Preprint
id arxiv_https___arxiv_org_abs_2601_18585
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle GimmBO: Interactive Generative Image Model Merging via Bayesian Optimization
Liu, Chenxi
Ling, Selena
Jacobson, Alec
Computer Vision and Pattern Recognition
Graphics
I.3.6; I.4.9
Fine-tuning-based adaptation is widely used to customize diffusion-based image generation, leading to large collections of community-created adapters that capture diverse subjects and styles. Adapters derived from the same base model can be merged with weights, enabling the synthesis of new visual results within a vast and continuous design space. To explore this space, current workflows rely on manual slider-based tuning, an approach that scales poorly and makes weight selection difficult, even when the candidate set is limited to 20-30 adapters. We propose GimmBO to support interactive exploration of adapter merging for image generation through Preferential Bayesian Optimization (PBO). Motivated by observations from real-world usage, including sparsity and constrained weight ranges, we introduce a two-stage BO backend that improves sampling efficiency and convergence in high-dimensional spaces. We evaluate our approach with simulated users and a user study, demonstrating improved convergence, high success rates, and consistent gains over BO and line-search baselines, and further show the flexibility of the framework through several extensions.
title GimmBO: Interactive Generative Image Model Merging via Bayesian Optimization
topic Computer Vision and Pattern Recognition
Graphics
I.3.6; I.4.9
url https://arxiv.org/abs/2601.18585