We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Noise Estimation for Generative Diffusion Models

Abstract: Generative diffusion models have emerged as leading models in speech and image generation. However, in order to perform well with a small number of denoising steps, a costly tuning of the set of noise parameters is needed. In this work, we present a simple and versatile learning scheme that can step-by-step adjust those noise parameters, for any given number of steps, while the previous work needs to retune for each number separately. Furthermore, without modifying the weights of the diffusion model, we are able to significantly improve the synthesis results, for a small number of steps. Our approach comes at a negligible computation cost.
Subjects: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
Cite as: arXiv:2104.02600 [cs.LG]
  (or arXiv:2104.02600v2 [cs.LG] for this version)

Submission history

From: Eliya Nachmani [view email]
[v1] Tue, 6 Apr 2021 15:46:16 GMT (10024kb,D)
[v2] Sun, 12 Sep 2021 07:49:25 GMT (10912kb,D)

Link back to: arXiv, form interface, contact.