NVIDIA/semantic-segmentation

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1828 stars · 389 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

NVIDIA's PyTorch monorepo implementing the paper 'Hierarchical Multi-Scale Attention for Semantic Segmentation', with pretrained weights and training/evaluation scripts for Cityscapes and Mapillary. It is research code accompanying a published paper, last released in 2021.

Use cases

• train a semantic segmentation model on Cityscapes
• run inference with pretrained segmentation weights
• reproduce hierarchical multi-scale attention results
• evaluate segmentation IOU on Mapillary
• dump segmentation output images for inspection

When to choose

• you want to reproduce the paper's state-of-the-art Cityscapes IOU results
• you need a PyTorch baseline for semantic segmentation research
• you want pretrained segmentation weights for Cityscapes or Mapillary

When to avoid

• you need a maintained production segmentation library
• you use recent PyTorch versions, since the code targets PyTorch 1.3
• you need segmentation for datasets other than Cityscapes or Mapillary without adaptation

Facets

library · maturity maintenance

machine-learning deep-learning image-processing computer-vision computer-vision deep-learning machine-learning image-processing python semantic-segmentation pytorch cityscapes mapillary nvidia research-code hierarchical-multi-scale-attention gpu linux docker

Member repositories

For agents

markdown · JSON · MCP: product_card(name="NVIDIA/semantic-segmentation")