AutoLoRA: AutoGuidance Meets Low-Rank Adaptation for Diffusion Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kasymov, Artur, Sendera, Marcin, Stypułkowski, Michał, Zięba, Maciej, Spurek, Przemysław
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914965419458560
author Kasymov, Artur
Sendera, Marcin
Stypułkowski, Michał
Zięba, Maciej
Spurek, Przemysław
author_facet Kasymov, Artur
Sendera, Marcin
Stypułkowski, Michał
Zięba, Maciej
Spurek, Przemysław
contents Low-rank adaptation (LoRA) is a fine-tuning technique that can be applied to conditional generative diffusion models. LoRA utilizes a small number of context examples to adapt the model to a specific domain, character, style, or concept. However, due to the limited data utilized during training, the fine-tuned model performance is often characterized by strong context bias and a low degree of variability in the generated images. To solve this issue, we introduce AutoLoRA, a novel guidance technique for diffusion models fine-tuned with the LoRA approach. Inspired by other guidance techniques, AutoLoRA searches for a trade-off between consistency in the domain represented by LoRA weights and sample diversity from the base conditional diffusion model. Moreover, we show that incorporating classifier-free guidance for both LoRA fine-tuned and base models leads to generating samples with higher diversity and better quality. The experimental results for several fine-tuned LoRA domains show superiority over existing guidance techniques on selected metrics.
format Preprint
id arxiv_https___arxiv_org_abs_2410_03941
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle AutoLoRA: AutoGuidance Meets Low-Rank Adaptation for Diffusion Models
Kasymov, Artur
Sendera, Marcin
Stypułkowski, Michał
Zięba, Maciej
Spurek, Przemysław
Computer Vision and Pattern Recognition
Low-rank adaptation (LoRA) is a fine-tuning technique that can be applied to conditional generative diffusion models. LoRA utilizes a small number of context examples to adapt the model to a specific domain, character, style, or concept. However, due to the limited data utilized during training, the fine-tuned model performance is often characterized by strong context bias and a low degree of variability in the generated images. To solve this issue, we introduce AutoLoRA, a novel guidance technique for diffusion models fine-tuned with the LoRA approach. Inspired by other guidance techniques, AutoLoRA searches for a trade-off between consistency in the domain represented by LoRA weights and sample diversity from the base conditional diffusion model. Moreover, we show that incorporating classifier-free guidance for both LoRA fine-tuned and base models leads to generating samples with higher diversity and better quality. The experimental results for several fine-tuned LoRA domains show superiority over existing guidance techniques on selected metrics.
title AutoLoRA: AutoGuidance Meets Low-Rank Adaptation for Diffusion Models
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2410.03941