PureCC：面向文本到图像概念定制的纯学习
PureCC: Pure Learning for Text-to-Image Concept Customization
提出解耦学习目标与双分支训练流程，在注入个性化概念的同时显式约束原模型行为漂移，并给出自适应引导尺度 l*；做个性化生成或微调扩散模型时可直接借鉴其「冻结提取器提供隐式引导 + 可训练流模型保留原条件预测」的结构来缓解灾难性遗忘。
原文依据 PureCC introduces a novel decoupled learning objective for concept customization, which combines the implicit guidance of the target concept with the original conditional prediction.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论现有的概念定制方法在高保真与多概念定制方面已取得显著成果。然而，它们往往忽视了在学习新的个性化概念时对原模型行为与能力的影响。为解决这一问题，我们提出 PureCC。PureCC 为概念定制引入了一种新颖的解耦学习目标，它将目标概念的隐式引导与原始条件预测相结合。这种分离的形式使 PureCC 在训练中能够大幅聚焦于原模型。此外，基于该目标，PureCC 设计了一条双分支训练流程，其中包含一个冻结的提取器，用于提供纯净的目标概念表征作为隐式引导，以及一个可训练的流模型，用于产生原始条件预测，二者共同实现对个性化概念的纯学习。进一步地，PureCC 引入了一种新颖的自适应引导尺度 l*，以动态调整目标概念的引导强度，在定制保真度与模型保持之间取得平衡。大量实验表明，PureCC 在保持原模型行为与能力的同时实现高保真概念定制，取得了最优性能。
英文原文摘要 1268 字符
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 l* 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.