HackerPapers

arXiv:2105.05233cs.LG

Diffusion Models Beat GANs on Image Synthesis

Prafulla Dhariwal, Alex Nichol

Abstract

We show that diffusion models can achieve image sample quality superior to the current state-of-the-art generative models. We achieve this on unconditional image synthesis by finding a better architecture through a series of ablations. For conditional image synthesis, we further improve sample quality with classifier guidance: a simple, compute-efficient method for trading off diversity for fidelity using gradients from a classifier. We achieve an FID of 2.97 on ImageNet 128$\times$128, 4.59 on ImageNet 256$\times$256, and 7.72 on ImageNet 512$\times$512, and we match BigGAN-deep even with as few as 25 forward passes per sample, all while maintaining better coverage of the distribution. Finally, we find that classifier guidance combines well with upsampling diffusion models, further improving FID to 3.94 on ImageNet 256$\times$256 and 3.85 on ImageNet 512$\times$512. We release our code at https://github.com/openai/guided-diffusion

arXiv comments
Added compute requirements, ImageNet 256$\times$256 upsampling FID and samples, DDIM guided sampler, fixed typos
Journal reference
DOI
Subjects
cs.LGMachine Learning cs.AIArtificial Intelligence cs.CVComputer Vision and Pattern Recognition stat.MLMachine Learning (Statistics)

On Hacker News

Linked by 3 identifiable commenters in 4 comments across 4 threads, from Nov 2022 to Jun 2025. Found in 1 HN submission from May 2021 to May 2021. Open submission discussions.

Comments linking this paper, per year.

Discussed in comments4 threads

  1. Normalizing Flows Are Capable Generative Models 1 comment

    • As far as I'm aware, this is the largest Normalizing Flow that exists, and I think they undermined their work by not mentioning this... Their ImageNet model (4_1024_8_8_0.05[0]) is ~820M while AFHQ is ~472M. Prior to that there is DenseFlow[1] and MaCow[2], which are both <200M…

      godelskiView on HNParent context

  2. The Path to StyleGan2 – Implementing the Progressive Growing GAN 1 comment

    • > The rule in ML There's definitely attempts to revive things (in the general sense, not just GANs), but most successes appear to come from large labs dedicating equal computational resources to the older models and often by changing names. This can make things more confusing…

      godelskiView on HNParent context

  3. Diffusion Models 1 comment

    • The original Dickstein 2015 paper [1] formulated diffusion as maximizing (a lower bound of) log-likelihood of generating the distribution, so there was some theory. But my understanding is that the breakthrough was empirical results from Ho [2] and Nichol [3] showing diffusion…

      reasonablekloutView on HNParent context

  4. Ten Years of Image Synthesis 1 comment

    • "It was abandoned without any evidence whatsoever that diffusion model etc is required.", diffusion models beats GAN models in quality on all datasets it was tested on, I suggest to read (again probably): "Diffusion Models Beat GANs on Image Synthesis" [1], the difference in…

      GaggiXView on HNParent context

Submitted to Hacker News

Connected through shared HN threads, not inferred similarity or a recommendation.