Paper ID: 2408.06740

DiffLoRA: Generating Personalized Low-Rank Adaptation Weights with Diffusion

Yujia Wu, Yiming Shi, Jiwei Wei, Chengwei Sun, Yuyang Zhou, Yang Yang, Heng Tao Shen

Personalized text-to-image generation has gained significant attention for its capability to generate high-fidelity portraits of specific identities conditioned on user-defined prompts. Existing methods typically involve test-time fine-tuning or instead incorporating an additional pre-trained branch. However, these approaches struggle to simultaneously address the demands of efficiency, identity fidelity, and preserving the model's original generative capabilities. In this paper, we propose DiffLoRA, a novel approach that leverages diffusion models as a hypernetwork to predict personalized low-rank adaptation (LoRA) weights based on the reference images. By integrating these LoRA weights into the text-to-image model, DiffLoRA achieves personalization during inference without further training. Additionally, we propose an identity-oriented LoRA weight construction pipeline to facilitate the training of DiffLoRA. By utilizing the dataset produced by this pipeline, our DiffLoRA consistently generates high-performance and accurate LoRA weights. Extensive evaluations demonstrate the effectiveness of our method, achieving both time efficiency and maintaining identity fidelity throughout the personalization process.

Submitted: Aug 13, 2024