NVlabs/SPADE

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

7717 stars · 970 forks observed · 2026-08-28

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

Official PyTorch implementation of SPADE (GauGAN), a CVPR 2019 method for synthesizing photorealistic images from semantic segmentation maps using spatially-adaptive normalization. Released by NVIDIA for academic research under CC BY-NC-SA 4.0, with a newer reimplementation available in the Imaginaire repository.

Use cases

• generate photorealistic images from segmentation maps
• turn a rough sketch layout into a realistic landscape image
• reproduce GauGAN research results on COCO-Stuff or Cityscapes
• train a conditional GAN for semantic image synthesis
• control style and content of generated images via layout input

When to choose

• you need the original reference implementation of SPADE for research or paper reproduction
• you want to synthesize images conditioned on semantic layouts in PyTorch
• you are studying spatially-adaptive normalization techniques

When to avoid

• you need commercial use - the license is non-commercial only
• you want the most performant modern implementation - use NVlabs/imaginaire instead
• you lack multi-GPU hardware for training at paper-scale

Facets

library · maturity maintenance

machine-learning deep-learning image-processing deep-learning computer-vision image-processing artificial-intelligence python gan image-synthesis gaugan pytorch semantic-segmentation research-code cvpr-2019 non-commercial gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="NVlabs/SPADE")