Skip to content

Latest commit

 

History

History
 
 

PanopticSeg

English | 简体中文

Panoptic Segmentation Toolkit Based on PaddleSeg

Contents

Introduction

Panoptic segmentation is an image parsing task that unifies the typically distinct tasks of semantic segmentation (assign a class label to each pixel) and instance segmentation (detect and segment each object instance). Built on top of PaddleSeg, this toolkit aims to facilitate the training, evaluation, and deployment of panoptic segmentation models.

  • High-Performance Models: This toolkit provides state-of-the-art panoptic segmentation models that can be used out of the box.
  • High Efficiency: This toolkit supports multi-process asynchronous I/O, multi-card parallel training, evaluation, and other acceleration strategies, combined with the memory optimization function of PaddlePaddle, all these allowing developers to train panoptic segmentation models at a lower cost.
  • Complete Flow: This toolkit supports a complete worflow from model designing to model deployment. With the help of this toolkit, developers can easily complete all tasks.

  • The pictures above are based on images from the Cityscapes and MS COCO datasets and the results obtained by models trained with this toolkit.

Update Notes

  • 2022.12
    • Add Mask2Former and Panoptic-DeepLab models.

Models

Tutorials

Community

  • If you have any questions, suggestions or feature requests, please do not hesitate to create an issue in GitHub Issues.
  • Please scan the following QR code and join PaddleSeg WeChat group to communicate with us.