facebookresearch/CutLER
A research framework for training object detection and instance segmentation models without human annotations using unsupervised learning techniques.
Not currently ranked — collecting fresh signals.
star history
CutLER is an approach for training object detection and instance segmentation models without human annotations by leveraging cut-and-paste techniques and self-supervised learning. The project includes VideoCutLER, which extends the approach to video instance segmentation. It provides pretrained models, training code, and evaluation pipelines for unsupervised detection tasks across multiple benchmarks.
Frequently asked
- What is facebookresearch/CutLER?
- A research framework for training object detection and instance segmentation models without human annotations using unsupervised learning techniques.
- Is CutLER open source?
- Yes — facebookresearch/CutLER is an open-source project tracked on heatdrop.
- What language is CutLER written in?
- facebookresearch/CutLER is primarily written in Python.
- How popular is CutLER?
- facebookresearch/CutLER has 1.1k stars on GitHub.
- Where can I find CutLER?
- facebookresearch/CutLER is on GitHub at https://github.com/facebookresearch/CutLER.