Latent Denoising Diffusion GAN: Faster sampling, Higher image quality

TL;DR

This paper introduces the Latent Denoising Diffusion GAN, a model that significantly improves inference speed and image quality by operating in a compressed latent space and employing a weighted learning strategy, outperforming previous diffusion models.

Contribution

It presents a novel latent space approach combined with weighted learning to enhance diffusion model efficiency and output quality, surpassing prior methods like DiffusionGAN and Wavelet Diffusion.

Findings

Achieves state-of-the-art speed among diffusion models.

Shows significant improvements in image quality metrics.

Demonstrates effectiveness across multiple datasets.

Abstract

Diffusion models are emerging as powerful solutions for generating high-fidelity and diverse images, often surpassing GANs under many circumstances. However, their slow inference speed hinders their potential for real-time applications. To address this, DiffusionGAN leveraged a conditional GAN to drastically reduce the denoising steps and speed up inference. Its advancement, Wavelet Diffusion, further accelerated the process by converting data into wavelet space, thus enhancing efficiency. Nonetheless, these models still fall short of GANs in terms of speed and image quality. To bridge these gaps, this paper introduces the Latent Denoising Diffusion GAN, which employs pre-trained autoencoders to compress images into a compact latent space, significantly improving inference speed and image quality. Furthermore, we propose a Weighted Learning strategy to enhance diversity and image quality. Experimental results on the CIFAR-10, CelebA-HQ, and LSUN-Church datasets prove that our model achieves state-of-the-art running speed among diffusion models. Compared to its predecessors, DiffusionGAN and Wavelet Diffusion, our model shows remarkable improvements in all evaluation metrics. Code and pre-trained checkpoints: \url{https://github.com/thanhluantrinh/LDDGAN.git}