Publications

Detailed Information

Distilling Diffusion Models Into Conditional GANs

Cited 0 time in Web of Science Cited 3 time in Scopus
Authors

Kang, Minguk; Zhang, Richard; Barnes, Connelly; Paris, Sylvain; Kwak, Suha; Park, Jaesik; Shechtman, Eli; Zhu, Jun-Yan; Park, Taesung

Issue Date
2025
Publisher
Springer Science and Business Media Deutschland GmbH
Citation
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol.15086 LNCS, pp.428-447
Abstract
We propose a method to distill a complex multistep diffusion model into a single-step conditional GAN student model, dramatically accelerating inference, while preserving image quality. Our approach interprets diffusion distillation as a paired image-to-image translation task, using noise-to-image pairs of the diffusion models ODE trajectory. For efficient regression loss computation, we propose E-LatentLPIPS, a perceptual loss operating directly in diffusion models latent space, utilizing an ensemble of augmentations. Furthermore, we adapt a diffusion model to construct a multi-scale discriminator with a text alignment loss to build an effective conditional GAN-based formulation. E-LatentLPIPS converges more efficiently than many existing distillation methods, even accounting for dataset construction costs. We demonstrate that our one-step generator outperforms cutting-edge one-step diffusion distillation models – SDXL-Turbo and SDXL-Lightning – on the COCO benchmark.
ISSN
0302-9743
URI
https://hdl.handle.net/10371/216408
DOI
https://doi.org/10.1007/978-3-031-73390-1_25
Files in This Item:
There are no files associated with this item.
Appears in Collections:

Related Researcher

  • College of Engineering
  • Dept. of Computer Science and Engineering
Research Area Computer Graphics, Computer Vision, Machine Learning, Robotics

Altmetrics

Item View & Download Count

  • mendeley

Items in S-Space are protected by copyright, with all rights reserved, unless otherwise indicated.

Share