AI Paper Atlas

Research map / Benchmarks / FFHQ 256×256

FFHQ 256×256 leaderboard

16 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on FFHQ 256×256. Best FID: 2.24 (ConPreDiff, NeurIPS 2023).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

FFHQ 256×256 · FID

16 papers, each with its best reported number for its own method; lower is better.

#FID ↓MethodPaperTopic
12.24ConPreDiffImproving Diffusion-Based Image Synthesis with Context Prediction
NeurIPS 2023 · Table: Evaluation results for
Autoregressive and latent generation
22.72Poly-INRPolynomial Implicit Neural Representations for Large Diverse Datasets
CVPR 2023 · Table 2
RNNs and normalization
32.81StyleSwinStyleSwin: Transformer-Based GAN for High-Resolution Image Generation
CVPR 2022 · Main results table
GANs
42.96EqGAN-SA
StyleGAN2
Improving GAN Equilibrium by Raising Spatial Awareness
CVPR 2022 · Table 1
GANs
53.61StyleGAN2 + CoopInit
StyleGAN2
CoopInit: Initializing Generative Adversarial Networks via Cooperative Learning
AAAI 2023 · Table 2
GANs
64.38CIPSImage Generators With Conditionally-Independent Pixel Synthesis
CVPR 2021 · Table: FID on multiple dataset
Autoregressive and latent generation
74.5StyleGAN3-R
StyleGAN2
Alias-Free Generative Adversarial Networks
NeurIPS 2021 · Appendix ablation table
GANs
84.72EG3DBenchmarking and Analyzing 3D-aware Image Synthesis with a Modularized Codebase
NeurIPS 2023 · Methods table
NeRF and radiance fields
95.61DMP
DiT-L/2
Diffusion Model Patching via Mixture-of-Prompts
AAAI 2025 · Table: Comparison of further t
Fast diffusion sampling
105.85Binary Latent Diffusion, 64 stepsBinary Latent Diffusion
CVPR 2023 · Table: unconditional image gen
Autoregressive and latent generation
6 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.