One-Step Image Translation with Text-to-Image Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Parmar, Gaurav, Park, Taesung, Narasimhan, Srinivasa, Zhu, Jun-Yan
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910371237855232
author Parmar, Gaurav
Park, Taesung
Narasimhan, Srinivasa
Zhu, Jun-Yan
author_facet Parmar, Gaurav
Park, Taesung
Narasimhan, Srinivasa
Zhu, Jun-Yan
contents In this work, we address two limitations of existing conditional diffusion models: their slow inference speed due to the iterative denoising process and their reliance on paired data for model fine-tuning. To tackle these issues, we introduce a general method for adapting a single-step diffusion model to new tasks and domains through adversarial learning objectives. Specifically, we consolidate various modules of the vanilla latent diffusion model into a single end-to-end generator network with small trainable weights, enhancing its ability to preserve the input image structure while reducing overfitting. We demonstrate that, for unpaired settings, our model CycleGAN-Turbo outperforms existing GAN-based and diffusion-based methods for various scene translation tasks, such as day-to-night conversion and adding/removing weather effects like fog, snow, and rain. We extend our method to paired settings, where our model pix2pix-Turbo is on par with recent works like Control-Net for Sketch2Photo and Edge2Image, but with a single-step inference. This work suggests that single-step diffusion models can serve as strong backbones for a range of GAN learning objectives. Our code and models are available at https://github.com/GaParmar/img2img-turbo.
format Preprint
id arxiv_https___arxiv_org_abs_2403_12036
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle One-Step Image Translation with Text-to-Image Models
Parmar, Gaurav
Park, Taesung
Narasimhan, Srinivasa
Zhu, Jun-Yan
Computer Vision and Pattern Recognition
Graphics
Machine Learning
In this work, we address two limitations of existing conditional diffusion models: their slow inference speed due to the iterative denoising process and their reliance on paired data for model fine-tuning. To tackle these issues, we introduce a general method for adapting a single-step diffusion model to new tasks and domains through adversarial learning objectives. Specifically, we consolidate various modules of the vanilla latent diffusion model into a single end-to-end generator network with small trainable weights, enhancing its ability to preserve the input image structure while reducing overfitting. We demonstrate that, for unpaired settings, our model CycleGAN-Turbo outperforms existing GAN-based and diffusion-based methods for various scene translation tasks, such as day-to-night conversion and adding/removing weather effects like fog, snow, and rain. We extend our method to paired settings, where our model pix2pix-Turbo is on par with recent works like Control-Net for Sketch2Photo and Edge2Image, but with a single-step inference. This work suggests that single-step diffusion models can serve as strong backbones for a range of GAN learning objectives. Our code and models are available at https://github.com/GaParmar/img2img-turbo.
title One-Step Image Translation with Text-to-Image Models
topic Computer Vision and Pattern Recognition
Graphics
Machine Learning
url https://arxiv.org/abs/2403.12036