Salvato in:
| Autori principali: | , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2512.08785 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911310409629696 |
|---|---|
| author | Hao, Yiming Xu, Mutian Ye, Chongjie Qin, Jie Lu, Shunlin Qin, Yipeng Han, Xiaoguang |
| author_facet | Hao, Yiming Xu, Mutian Ye, Chongjie Qin, Jie Lu, Shunlin Qin, Yipeng Han, Xiaoguang |
| contents | Personalizing visual generative models to meet specific user needs has gained increasing attention, yet current methods like Low-Rank Adaptation (LoRA) remain impractical due to their demand for task-specific data and lengthy optimization. While a few hypernetwork-based approaches attempt to predict adaptation weights directly, they struggle to map fine-grained user prompts to complex LoRA distributions, limiting their practical applicability. To bridge this gap, we propose LoFA, a general framework that efficiently predicts personalized priors for fast model adaptation. We first identify a key property of LoRA: structured distribution patterns emerge in the relative changes between LoRA and base model parameters. Building on this, we design a two-stage hypernetwork: first predicting relative distribution patterns that capture key adaptation regions, then using these to guide final LoRA weight prediction. Extensive experiments demonstrate that our method consistently predicts high-quality personalized priors within seconds, across multiple tasks and user prompts, even outperforming conventional LoRA that requires hours of processing. Project page: https://jaeger416.github.io/lofa/. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_08785 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | LoFA: Learning to Predict Personalized Priors for Fast Adaptation of Visual Generative Models Hao, Yiming Xu, Mutian Ye, Chongjie Qin, Jie Lu, Shunlin Qin, Yipeng Han, Xiaoguang Computer Vision and Pattern Recognition Personalizing visual generative models to meet specific user needs has gained increasing attention, yet current methods like Low-Rank Adaptation (LoRA) remain impractical due to their demand for task-specific data and lengthy optimization. While a few hypernetwork-based approaches attempt to predict adaptation weights directly, they struggle to map fine-grained user prompts to complex LoRA distributions, limiting their practical applicability. To bridge this gap, we propose LoFA, a general framework that efficiently predicts personalized priors for fast model adaptation. We first identify a key property of LoRA: structured distribution patterns emerge in the relative changes between LoRA and base model parameters. Building on this, we design a two-stage hypernetwork: first predicting relative distribution patterns that capture key adaptation regions, then using these to guide final LoRA weight prediction. Extensive experiments demonstrate that our method consistently predicts high-quality personalized priors within seconds, across multiple tasks and user prompts, even outperforming conventional LoRA that requires hours of processing. Project page: https://jaeger416.github.io/lofa/. |
| title | LoFA: Learning to Predict Personalized Priors for Fast Adaptation of Visual Generative Models |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.08785 |