A Noise is Worth Diffusion Guidance

Guardado en:
书目详细资料
发表在:arXiv.org (Dec 5, 2024), p. n/a
主要作者: Ahn, Donghoon
其他作者: Kang, Jiwon, Lee, Sanghyun, Min, Jaewon, Kim, Minjae, Jang, Wooseok, Cho, Hyoungwon, Sayak, Paul, Kim, SeonHwa, Cha, Eunju, Jin, Kyong Hwan, Kim, Seungryong
出版:
Cornell University Library, arXiv.org
主题:
在线阅读:Citation/Abstract
Full text outside of ProQuest
标签: 添加标签
没有标签, 成为第一个标记此记录!

MARC

LEADER 00000nab a2200000uu 4500
001 3141682707
003 UK-CbPIL
022 |a 2331-8422 
035 |a 3141682707 
045 0 |b d20241205 
100 1 |a Ahn, Donghoon 
245 1 |a A Noise is Worth Diffusion Guidance 
260 |b Cornell University Library, arXiv.org  |c Dec 5, 2024 
513 |a Working Paper 
520 3 |a Diffusion models excel in generating high-quality images. However, current diffusion models struggle to produce reliable images without guidance methods, such as classifier-free guidance (CFG). Are guidance methods truly necessary? Observing that noise obtained via diffusion inversion can reconstruct high-quality images without guidance, we focus on the initial noise of the denoising pipeline. By mapping Gaussian noise to `guidance-free noise', we uncover that small low-magnitude low-frequency components significantly enhance the denoising process, removing the need for guidance and thus improving both inference throughput and memory. Expanding on this, we propose \ours, a novel method that replaces guidance methods with a single refinement of the initial noise. This refined noise enables high-quality image generation without guidance, within the same diffusion pipeline. Our noise-refining model leverages efficient noise-space learning, achieving rapid convergence and strong performance with just 50K text-image pairs. We validate its effectiveness across diverse metrics and analyze how refined noise can eliminate the need for guidance. See our project page: https://cvlab-kaist.github.io/NoiseRefine/. 
653 |a Random noise 
653 |a Noise generation 
653 |a Image quality 
653 |a Image reconstruction 
653 |a Image enhancement 
653 |a Image processing 
653 |a Noise reduction 
700 1 |a Kang, Jiwon 
700 1 |a Lee, Sanghyun 
700 1 |a Min, Jaewon 
700 1 |a Kim, Minjae 
700 1 |a Jang, Wooseok 
700 1 |a Cho, Hyoungwon 
700 1 |a Sayak, Paul 
700 1 |a Kim, SeonHwa 
700 1 |a Cha, Eunju 
700 1 |a Jin, Kyong Hwan 
700 1 |a Kim, Seungryong 
773 0 |t arXiv.org  |g (Dec 5, 2024), p. n/a 
786 0 |d ProQuest  |t Engineering Database 
856 4 1 |3 Citation/Abstract  |u https://www.proquest.com/docview/3141682707/abstract/embedded/6A8EOT78XXH2IG52?source=fedsrch 
856 4 0 |3 Full text outside of ProQuest  |u http://arxiv.org/abs/2412.03895