Datasets › CelebA-HQ

CelebA-HQ

Introduced by Tero Karras et al. in Progressive Growing of GANs for Improved Quality, Stability, and Variation27 Oct 2017 archive 2025-07-28

The CelebA-HQ dataset is a high-quality version of CelebA that consists of 30,000 images at 1024×1024 resolution.

Source: IntroVAE: Introspective Variational Autoencoders for Photographic Image Synthesis

Image Source: https://github.com/tkarras/progressive_growing_of_gans

Benchmarks archive 2025-07-28

All 13 leaderboards whose dataset resolves to this page shown (sort by any header). "First row" is the archive's own first row at snapshot, in the archive's row order; nothing here re-ranks and metric direction is not asserted.

First row (archive order)PaperCode
Image Generation CelebA-HQ 256x256 RDM FID 3.15 Relay Diffusion: Unifying diffusion process across... THUDM/RelayDiffusion 19 Compare
Image Generation CelebA-HQ 1024x1024 StyleSwin FID 4.43 StyleSwin: Transformer-based GAN for High-resolution... microsoft/StyleSwin 10 Compare
Image Generation CelebA-HQ 128x128 U-Net GAN FID 2.03 A U-Net Based Discriminator for Generative Adversarial Networks boschresearch/unetgan +2 7 Compare
Blind Face Restoration CelebA-HQ GPEN FID 31.72 GAN Prior Embedded Network for Blind Face Restoration in the Wild yangxy/GPEN +2 6 Compare
Image Inpainting CelebA-HQ MAT FID 4.86 MAT: Mask-Aware Transformer for Large Hole Image Inpainting fenglinglwb/mat 6 Compare
Image-to-Image Translation CelebA-HQ StarGAN v2 FID 13.73 StarGAN v2: Diverse Image Synthesis for Multiple Domains clovaai/stargan-v2 +13 6 Compare
Image Super-Resolution CelebA-HQ 128x128 TSRGP PSNR 24.09 Trustworthy Image Super-Resolution via Generative Pseudoinverse andreasfloros/trustworthy-super-resolution 4 Compare
Multimodal Unsupervised Image-To-Image Translation CelebA-HQ StarGAN v2 FID 13.73 StarGAN v2: Diverse Image Synthesis for Multiple Domains clovaai/stargan-v2 +13 4 Compare
Image Generation CelebA-HQ 64x64 COCO-GAN FID 4.0 COCO-GAN: Generation by Parts via Conditional Coordinating hubert0527/COCO-GAN 3 Compare
Density Estimation CelebA-HQ 256x256 RNODE Log-likelihood 1.04 How to train your neural ODE: the world of Jacobian and... cfinlay/ffjord-rnode +1 1 Compare
Image Generation CelebA-HQ DDPM FLD 1.4 Normalizing Flow-Based Metric for Image Generation pranavphoenix/FLD 1 Compare
Image Generation CelebA-HQ 512x512 WaveDiff FID 6.40 Wavelet Diffusion Models are fast and scalable Image Generators vinairesearch/wavediff 1 Compare
Image Super-Resolution Celeb-HQ 4x upscaling Edge-informed SR PSNR 28.23 Edge-Informed Single Image Super-Resolution knazeri/edge-informed-sisr +1 1 Compare

Papers archive 2025-07-28

30 shown of 55 papers with a leaderboard row on this dataset's benchmarks, newest first. The archive's own "papers using this dataset" list was never published, so this is the benchmark-backed subset; the archive's count for this dataset is 954. The Syntology column is from Syntology's graph (read 2026-09-24), stated per sample; it is not part of any archive number.

DateSamples run Syntology
Trustworthy Image Super-Resolution via Generative Pseudoinverse 1 1 18 May 2025 not harvested
Normalizing Flow-Based Metric for Image Generation 1 1 2 Oct 2024 not harvested
Learning Latent Space Hierarchical EBM Diffusion Models 0 1 22 May 2024 not harvested
DDMI: Domain-Agnostic Latent Diffusion Models for Synthesizing High-Quality Implicit Neural Representations 1 1 23 Jan 2024 ran 17 of 21 samples (4 unverified)
Bellman Optimal Stepsize Straightening of Flow-Matching Models 1 1 27 Dec 2023 ran 8 of 14 samples (6 unverified; 14 pointer-only for licence)
Learning Energy-based Model via Dual-MCMC Teaching 0 1 5 Dec 2023 not harvested
A High-Quality Robust Diffusion Framework for Corrupted Dataset 1 1 28 Nov 2023 ran 5 of 12 samples (7 unverified)
Relay Diffusion: Unifying diffusion process across resolutions for image synthesis 1 1 4 Sep 2023 ran 12 of 17 samples (5 unverified)
Flow Matching in Latent Space 1 1 17 Jul 2023 ran 1 of 1 samples (0 unverified; 1 pointer-only for licence)
WavePaint: Resource-efficient Token-mixer for Self-supervised Inpainting 1 1 1 Jul 2023 not harvested
Learning Joint Latent Space EBM Prior Model for Multi-layer Generator 0 1 10 Jun 2023 not harvested
Waving Goodbye to Low-Res: A Diffusion-Wavelet Approach for Image Super-Resolution 1 1 4 Apr 2023 ran 8 of 10 samples (2 unverified)
Implicit Diffusion Models for Continuous Super-Resolution 1 1 29 Mar 2023 ran 5 of 10 samples (5 unverified; 10 pointer-only for licence)
Wavelet Diffusion Models are fast and scalable Image Generators 1 3 29 Nov 2022 not harvested
EGSDE: Unpaired Image-to-Image Translation via Energy-Guided Stochastic Differential Equations 1 1 14 Jul 2022 not harvested
Hypercomplex Image-to-Image Translation 1 2 4 May 2022 not harvested
MAT: Mask-Aware Transformer for Large Hole Image Inpainting 1 1 29 Mar 2022 not harvested
Efficient-VDVAE: Less is more 1 1 25 Mar 2022 not harvested
Polarity Sampling: Quality and Diversity Control of Pre-Trained Generative Networks via Singular Values 1 1 3 Mar 2022 not harvested
StyleSwin: Transformer-based GAN for High-resolution Image Generation 1 2 20 Dec 2021 ran 5 of 5 samples (0 unverified; 3 pointer-only for licence)
High-Resolution Image Synthesis with Latent Diffusion Models 41 1 20 Dec 2021 ran 19 of 28 samples (9 unverified; 5 pointer-only for licence)
Tackling the Generative Learning Trilemma with Denoising Diffusion GANs 5 1 15 Dec 2021 ran 1 of 12 samples (11 unverified)
Resolution-robust Large Mask Inpainting with Fourier Convolutions 8 1 15 Sep 2021 ran 0 of 4 samples (4 unverified)
Image-to-Image Translation with Low Resolution Conditioning 1 1 23 Jul 2021 not harvested
Improved Transformer for High-Resolution GANs 1 1 14 Jun 2021 ran 3 of 4 samples (1 unverified)
Soft Truncation: A Universal Training Technique of Score-based Diffusion Model for High Precision Score Estimation 1 1 10 Jun 2021 ran 4 of 12 samples (8 unverified)
Score-based Generative Modeling in Latent Space 1 1 10 Jun 2021 ran 2 of 5 samples (3 unverified; 5 pointer-only for licence)
GAN Prior Embedded Network for Blind Face Restoration in the Wild 3 6 13 May 2021 ran 16 of 21 samples (5 unverified; 21 pointer-only for licence)
Image Inpainting by End-to-End Cascaded Refinement with Mask Awareness 1 1 28 Apr 2021 not harvested
Quaternion Generative Adversarial Networks 3 1 19 Apr 2021 not harvested

The full list of 55 is in the JSON twin.

Dataset loaders archive 2025-07-28

2 loaders as listed in the archive; links are outbound and not re-checked here.

Tasks archive 2025-07-28

License archive 2025-07-28

CC BY-NC 4

Modalities archive 2025-07-28

Languages archive 2025-07-28

No language tagged.

Variants archive 2025-07-28

  • CelebA-HQ 512x512
  • CelebA-HQ 256x256
  • CelebA-HQ 1024x1024
  • Celeb-HQ 4x upscaling
  • CelebA-HQ 64x64
  • CelebA-HQ 128x128
  • CelebA-HQ

7 variant names, as the archive lists them.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections