ultralytics 8.4.14

Ultralytics YOLO 🚀 for SOTA object detection, multi-object tracking, instance segmentation, pose estimation and image classification.

Stars: 53371, Watchers: 53371, Forks: 10227, Open Issues: 361

The ultralytics/ultralytics repo was created 3 years ago and the last code push was an hour ago.
The project is extremely popular with a mindblowing 53371 github stars!

How to Install ultralytics

You can install ultralytics using pip

pip install ultralytics

or add it to a project with poetry

poetry add ultralytics

Package Details

Author

None

License

AGPL-3.0

Homepage

None

PyPi:

https://pypi.org/project/ultralytics/

Documentation:

https://docs.ultralytics.com

GitHub Repo:

https://github.com/ultralytics/ultralytics

Classifiers

• Scientific/Engineering
• Scientific/Engineering/Artificial Intelligence
• Scientific/Engineering/Image Recognition
• Software Development

Related Packages

Errors

A list of common ultralytics errors.

Code Examples

Here are some ultralytics code examples and snippets.

GitHub Issues

The ultralytics package has 361 open issues on GitHub

• Add ActionRecognition solution
• YOLOv26 shows lower box recall but higher pose mAP compared to YOLOv8 on the same dataset
• YOLO26n: Dense Same-Category - Missing Det & Duplicate Bboxes
• YOLOv8 shows a decrease in mAP and other metrics in the new version
• PyTorch 2.10 support
• Save close_mosaic as integer in best_hyperparameters.yaml
• Organize hyperparameter tuning task as wandb sweep
• Where are box / cls / DFL losses defined and wired into the training loop in Ultralytics YOLO?
• Overlapping masks
• type of close_mosaic from hyperparameter tuning result is incorrect
• Definition of Rotation Angle in YOLO-OBB
• Feat/support mulit data config via yaml for yoloe
• yolo26 speed up ?
• Batch inference of YOLO11 seg model with NMS in NVIDIA Triton is wrong.
• How can I generate the metrics that can be obtained after k-fold cross-validation training?

See more issues on GitHub