zhangxiaosong18/FreeAnchor
A PyTorch implementation of a one-stage object detection method that learns to match anchors for visual object detection.
Not currently ranked — collecting fresh signals.
star history
FreeAnchor is an implementation of a NeurIPS 2019 paper on object detection. It provides an alternative to the IoU-based anchor matching strategy used in traditional one-stage detectors by learning to match anchors for each object. The repository builds on maskrcnn-benchmark and includes pre-trained models with multi-scale training and testing support on the COCO dataset.
Frequently asked
- What is zhangxiaosong18/FreeAnchor?
- A PyTorch implementation of a one-stage object detection method that learns to match anchors for visual object detection.
- Is FreeAnchor open source?
- Yes — zhangxiaosong18/FreeAnchor is open source, released under the MIT license.
- What language is FreeAnchor written in?
- zhangxiaosong18/FreeAnchor is primarily written in Python.
- How popular is FreeAnchor?
- zhangxiaosong18/FreeAnchor has 669 stars on GitHub.
- Where can I find FreeAnchor?
- zhangxiaosong18/FreeAnchor is on GitHub at https://github.com/zhangxiaosong18/FreeAnchor.