ShuLiu1993/PANet

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1347 stars · 278 forks observed · 2026-08-28

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

A PyTorch re-implementation of PANet (Path Aggregation Network), the CVPR 2018 paper that won 1st place in the COCO 2017 Instance Segmentation Challenge. Built on Detectron.pytorch, it provides training and testing code for instance segmentation and object detection models.

Use cases

• run instance segmentation on COCO images
• train an object detection model with path aggregation
• reproduce PANet challenge-winning results
• use PANet as a baseline for segmentation research
• fine-tune a pretrained instance segmentation model
• compare feature pyramid architectures for detection

When to choose

• you need a proven high-accuracy instance segmentation baseline in PyTorch
• you want to reproduce or extend the PANet paper
• you are already using Detectron.pytorch and want PANet features

When to avoid

• you need a maintained library with recent PyTorch support (code targets PyTorch 0.4.0)
• you want a simple out-of-the-box inference tool rather than a research codebase
• you need multi-GPU synchronized batch normalization as in the original paper

Facets

library · maturity maintenance

machine-learning computer-vision image-processing computer-vision deep-learning machine-learning python instance-segmentation object-detection pytorch cvpr-2018 coco panet research-code linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="ShuLiu1993/PANet")