Saved in:
Bibliographic Details
Main Authors: Liu, Chang, Shah, Viraj, Cui, Aiyu, Lazebnik, Svetlana
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2412.04465
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918127353200640
author Liu, Chang
Shah, Viraj
Cui, Aiyu
Lazebnik, Svetlana
author_facet Liu, Chang
Shah, Viraj
Cui, Aiyu
Lazebnik, Svetlana
contents This paper introduces UnZipLoRA, a method for decomposing an image into its constituent subject and style, represented as two distinct LoRAs (Low-Rank Adaptations). Unlike existing personalization techniques that focus on either subject or style in isolation, or require separate training sets for each, UnZipLoRA disentangles these elements from a single image by training both the LoRAs simultaneously. UnZipLoRA ensures that the resulting LoRAs are compatible, i.e., they can be seamlessly combined using direct addition. UnZipLoRA enables independent manipulation and recontextualization of subject and style, including generating variations of each, applying the extracted style to new subjects, and recombining them to reconstruct the original image or create novel variations. To address the challenge of subject and style entanglement, UnZipLoRA employs a novel prompt separation technique, as well as column and block separation strategies to accurately preserve the characteristics of subject and style, and ensure compatibility between the learned LoRAs. Evaluation with human studies and quantitative metrics demonstrates UnZipLoRA's effectiveness compared to other state-of-the-art methods, including DreamBooth-LoRA, Inspiration Tree, and B-LoRA.
format Preprint
id arxiv_https___arxiv_org_abs_2412_04465
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle UnZipLoRA: Separating Content and Style from a Single Image
Liu, Chang
Shah, Viraj
Cui, Aiyu
Lazebnik, Svetlana
Computer Vision and Pattern Recognition
This paper introduces UnZipLoRA, a method for decomposing an image into its constituent subject and style, represented as two distinct LoRAs (Low-Rank Adaptations). Unlike existing personalization techniques that focus on either subject or style in isolation, or require separate training sets for each, UnZipLoRA disentangles these elements from a single image by training both the LoRAs simultaneously. UnZipLoRA ensures that the resulting LoRAs are compatible, i.e., they can be seamlessly combined using direct addition. UnZipLoRA enables independent manipulation and recontextualization of subject and style, including generating variations of each, applying the extracted style to new subjects, and recombining them to reconstruct the original image or create novel variations. To address the challenge of subject and style entanglement, UnZipLoRA employs a novel prompt separation technique, as well as column and block separation strategies to accurately preserve the characteristics of subject and style, and ensure compatibility between the learned LoRAs. Evaluation with human studies and quantitative metrics demonstrates UnZipLoRA's effectiveness compared to other state-of-the-art methods, including DreamBooth-LoRA, Inspiration Tree, and B-LoRA.
title UnZipLoRA: Separating Content and Style from a Single Image
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2412.04465