CVDec 5, 2023

Orthogonal Adaptation for Modular Customization of Diffusion Models

arXiv:2312.02432v374 citationsh-index: 30CVPR
AI Analysis

This addresses the scalability challenge in text-to-image model customization for users needing to generate images with countless concepts, representing a significant but incremental advance over existing methods.

The paper tackles the problem of efficiently merging independently fine-tuned diffusion models for multiple concepts to enable scalable customization without compromising fidelity or adding computational costs, achieving superior performance in efficiency and identity preservation compared to baselines.

Customization techniques for text-to-image models have paved the way for a wide range of previously unattainable applications, enabling the generation of specific concepts across diverse contexts and styles. While existing methods facilitate high-fidelity customization for individual concepts or a limited, pre-defined set of them, they fall short of achieving scalability, where a single model can seamlessly render countless concepts. In this paper, we address a new problem called Modular Customization, with the goal of efficiently merging customized models that were fine-tuned independently for individual concepts. This allows the merged model to jointly synthesize concepts in one image without compromising fidelity or incurring any additional computational costs. To address this problem, we introduce Orthogonal Adaptation, a method designed to encourage the customized models, which do not have access to each other during fine-tuning, to have orthogonal residual weights. This ensures that during inference time, the customized models can be summed with minimal interference. Our proposed method is both simple and versatile, applicable to nearly all optimizable weights in the model architecture. Through an extensive set of quantitative and qualitative evaluations, our method consistently outperforms relevant baselines in terms of efficiency and identity preservation, demonstrating a significant leap toward scalable customization of diffusion models.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes