English

FreeLoRA: Enabling Training-Free LoRA Fusion for Autoregressive Multi-Subject Personalization

Computer Vision and Pattern Recognition 2025-07-03 v1

Abstract

Subject-driven image generation plays a crucial role in applications such as virtual try-on and poster design. Existing approaches typically fine-tune pretrained generative models or apply LoRA-based adaptations for individual subjects. However, these methods struggle with multi-subject personalization, as combining independently adapted modules often requires complex re-tuning or joint optimization. We present FreeLoRA, a simple and generalizable framework that enables training-free fusion of subject-specific LoRA modules for multi-subject personalization. Each LoRA module is adapted on a few images of a specific subject using a Full Token Tuning strategy, where it is applied across all tokens in the prompt to encourage weakly supervised token-content alignment. At inference, we adopt Subject-Aware Inference, activating each module only on its corresponding subject tokens. This enables training-free fusion of multiple personalized subjects within a single image, while mitigating overfitting and mutual interference between subjects. Extensive experiments show that FreeLoRA achieves strong performance in both subject fidelity and prompt consistency.

Keywords

Cite

@article{arxiv.2507.01792,
  title  = {FreeLoRA: Enabling Training-Free LoRA Fusion for Autoregressive Multi-Subject Personalization},
  author = {Peng Zheng and Ye Wang and Rui Ma and Zuxuan Wu},
  journal= {arXiv preprint arXiv:2507.01792},
  year   = {2025}
}
R2 v1 2026-07-01T03:43:23.888Z