PuLID：透過對比對齊的方式進行純粹且快速的ID自定義

PuLID: Pure and Lightning ID Customization via Contrastive Alignment

April 24, 2024

作者: Zinan Guo, Yanze Wu, Zhuowei Chen, Lang Chen, Qian He

cs.AI

摘要

我們提出了一種全新的無調整ID定制方法，稱為Pure and Lightning ID customization (PuLID)，專為文本到圖像生成而設。通過將Lightning T2I分支與標準擴散分支結合，PuLID引入了對比對齊損失和準確的ID損失，從而最小化對原始模型的干擾，確保高度ID保真度。實驗表明，PuLID在ID保真度和可編輯性方面均表現優異。PuLID的另一個吸引人之處在於，在ID插入前後，圖像元素（例如背景、燈光、構圖和風格）被保持盡可能一致。代碼和模型將可在以下網址獲得：https://github.com/ToTheBeginning/PuLID

English

We propose Pure and Lightning ID customization (PuLID), a novel tuning-free ID customization method for text-to-image generation. By incorporating a Lightning T2I branch with a standard diffusion one, PuLID introduces both contrastive alignment loss and accurate ID loss, minimizing disruption to the original model and ensuring high ID fidelity. Experiments show that PuLID achieves superior performance in both ID fidelity and editability. Another attractive property of PuLID is that the image elements (e.g., background, lighting, composition, and style) before and after the ID insertion are kept as consistent as possible. Codes and models will be available at https://github.com/ToTheBeginning/PuLID