Detectron2 is Facebook AI Research's next generation library that provides state-of-the-art detection and segmentation algorithms.
It supports a number of computer vision research projects and production applications in Facebook.
Features
Supports object detection and segmentation tasks
Includes new capabilities such as panoptic segmentation, Densepose, Cascade R-CNN, rotated bounding boxes, PointRend, DeepLab, ViTDet, MViTv2 etc.
Models can be exported to TorchScript format or Caffe2 format for deployment
Trains much faster than its predecessor
Use Cases
Building computer vision research projects
Production applications in Facebook
Suited For
Researchers
Developers
Data scientists
FAQ
What is Detectron2?
Detectron2 is a platform for object detection, segmentation and other visual recognition tasks.
What algorithms does Detectron2 support?
Detectron2 supports a range of algorithms including panoptic segmentation, Densepose, Cascade R-CNN, rotated bounding boxes, PointRend, DeepLab, ViTDet, MViTv2, etc.
Can models trained with Detectron2 be deployed?
Yes, models can be exported to TorchScript format or Caffe2 format for deployment.
How does Detectron2 compare to its predecessor, Detectron?
Detectron2 is the successor of Detectron and offers faster training and improved capabilities.