Skip to content
RobothecaBeta
Log in

Ultralytics YOLO

  • Software
  • Perception
  • Library

Ultralytics YOLO is a family of computer vision models and a Python package for detection, segmentation, pose, classification, depth and tracking tasks. It offers CLI and Python interfaces.

Source on GitHub: github.com/ultralytics/ultralytics (opens in a new tab). Facts read at commit 501a597 (opens in a new tab).

Licenses

Code
AGPL-3.0-only (opens in a new tab)

Read by Robotheca from LICENSE (opens in a new tab) (AGPL-3.0-only) at commit 501a597 (opens in a new tab), on 16 Oct 2026.

Commercial use
Yes

Copyleft: Modified source must be published, also when it runs as a service.

What it does

  • Performs object detection, instance and semantic segmentation, pose estimation, classification, depth estimation and oriented object detection.
  • Can track detected objects across video frames.
  • Provides a command-line interface (yolo command) and a Python API for predicting, training, validating and exporting models.
  • Pretrained YOLO26 models download automatically on first use; models can be exported to formats such as ONNX.
  • Reports benchmarks on COCO, Cityscapes, ImageNet and NYU Depth V2, with CPU ONNX and GPU TensorRT speeds.
  • A YOLO27 family is previewed but not yet available.

Written by Robotheca from the project's README (opens in a new tab) at commit 501a597.

Getting started

Difficulty
Beginner
Hardware it needs, as the project states it
device=0), with benchmark speeds measured on an Amazon EC2 P4d instance, a Tesla T4 and an RTX3090.

Maintenance

Status
Active: a commit in the last 90 days
Newest commit
16 Oct 2026 on main
Latest release
v8.4.178 - 8.4.178 Print the Platform-resolved model link after training starts (#26699) (16 Oct 2026)

Checked on GitHub on 16 Oct 2026.

Links

Robots on Robotheca that use it

No robot on Robotheca is linked to it yet. Links come from a robot's robot.yaml, from its dependency files read as text, or from Robotheca's editors.

Robotheca lists software as a reference: nothing is hosted or run here, and every fact links to where it was read.