PureCC: Pure Learning for Text-to-Image Concept Customization

TL;DR

PureCC introduces a decoupled learning approach with a dual-branch pipeline and adaptive guidance to enhance personalized concept customization while preserving original model behavior.

Contribution

It proposes a novel decoupled learning objective and dual-branch training pipeline for improved concept customization with minimal impact on the original model.

Findings

Achieves state-of-the-art preservation of original model capabilities

Enables high-fidelity personalized concept customization

Introduces adaptive guidance scale for balanced performance

Abstract

Existing concept customization methods have achieved remarkable outcomes in high-fidelity and multi-concept customization. However, they often neglect the influence on the original model's behavior and capabilities when learning new personalized concepts. To address this issue, we propose PureCC. PureCC introduces a novel decoupled learning objective for concept customization, which combines the implicit guidance of the target concept with the original conditional prediction. This separated form enables PureCC to substantially focus on the original model during training. Moreover, based on this objective, PureCC designs a dual-branch training pipeline that includes a frozen extractor providing purified target concept representations as implicit guidance and a trainable flow model producing the original conditional prediction, jointly achieving pure learning for personalized concepts. Furthermore, PureCC introduces a novel adaptive guidance scale $\lambda^\star$ to dynamically adjust the guidance strength of the target concept, balancing customization fidelity and model preservation. Extensive experiments show that PureCC achieves state-of-the-art performance in preserving the original behavior and capabilities while enabling high-fidelity concept customization. The code is available at https://github.com/lzc-sg/PureCC.