supervision

We write your reusable computer vision tools. 💜

What is supervision?

We write your reusable computer vision tools. 💜

Overview of supervision

What it is

Supervision is an open-source Python package for computer vision workflows. The README describes it as a toolkit that provides building blocks from data loading to real-time zone counting, so users can focus on applications around their models. It is model agnostic and supports classification, detection, and segmentation models. It requires Python 3.10 or later.

Key features

  • Model-agnostic connectors for libraries such as Ultralytics, Transformers, MMDetection, and Inference; integrations such as rfdetr return sv.Detections directly.
  • Customizable annotators for composing visualizations, including BoxAnnotator.
  • Dataset utilities to load, split, merge, save, and convert datasets.
  • Supported dataset formats include YOLO, Pascal VOC, and COCO.
  • On-demand image loading when iterating over datasets.
  • Tutorials cover dwell time analysis and speed estimation with vehicle tracking.

Use cases

The README presents workflows such as real-time zone counting, dwell time analysis, and speed estimation and vehicle tracking. It also documents dataset preparation for detection tasks, including loading COCO data, splitting datasets, merging datasets with different classes, and converting between YOLO, Pascal VOC, and COCO. Annotators can be used to visualize detections from supported models. These are examples stated in the README.

Getting started

Install in a Python>=3.10 environment:bashpip install supervisionOptional dependencies for the README model example:bashpip install pillow rfdetrThe README also provides quickstart examples for models, annotators, and datasets. See the repository README for full examples.

License and pricing

Supervision is released under the MIT license. Metadata lists pricing as open source; no paid pricing is stated in the README or metadata.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/roboflowsupervision.svg)](https://yunodb.com/tools/roboflowsupervision?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/roboflowsupervision?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/roboflowsupervision.svg" alt="Featured on YunoDB" height="20"></a>