NVlabs/ffhq-dataset resource

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

4182 stars · 611 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Flickr-Faces-HQ (FFHQ) is a dataset of 70,000 high-quality 1024x1024 PNG images of human faces, crawled from Flickr and aligned with dlib. It was created by NVIDIA as a benchmark for training generative adversarial networks, notably StyleGAN.

Use cases

• train a GAN to generate realistic human faces
• benchmark generative image models on high-resolution faces
• get a diverse face image dataset for research
• train a diffusion model on face images
• study image dataset bias and licensing metadata

When to choose

• you need large-scale, high-resolution, diverse face images for generative modeling research
• you want per-image license and author metadata for compliant redistribution

When to avoid

• you are building facial recognition technology - explicitly prohibited by the dataset authors
• you need a permissively licensed dataset for commercial products - images are CC BY-NC or similar, non-commercial only

Facets

dataset · maturity stable

image-processing machine-learning data-generation machine-learning computer-vision image-processing deep-learning python cross-platform face-dataset gan generative-models flickr research-dataset stylegan

Member repositories

For agents

markdown · JSON · MCP: product_card(name="NVlabs/ffhq-dataset")