YOLOv5

ultralytics/yolov5Reviewed by AI

What it does

YOLOv5 is an image recognition model and set of scripts from the company Ultralytics. It can spot and outline objects in pictures, videos, screen captures and live streams, sort images into categories, and be trained on your own photos. It is aimed at developers, students and researchers. You need Python 3.8 or newer and PyTorch, and you run it by typing commands. Training your own model realistically needs a graphics card. The README itself points new projects to the newer "ultralytics" package.

How easy it is to install

C · install 1 or 2 tools · Windows

Windows C
macOS C
Linux C

The three systems are equally hard and all sit at C, with Linux the best documented (pyproject.toml lists Linux, macOS and Windows, and the only release file is a Linux one): there is no installer, so you install Python and PyTorch, clone the code, run pip install -r requirements.txt and then start detect.py or train.py from a terminal. Training on your own photos also needs a graphics card in practice.

Numbers on GitHub

Stars: 58,247on Oct 31, 2026
Downloads: 135,107,80854,733 this week
Growth: +20stars in 5 days

Stars over time

Facts

Runs on: Windows, macOS, Linux
You need to install: Python, Graphics card
To build from source: Python, Graphics card
Disk space: 1.4 GB–5.8 GB · estimate
Accounts or keys: None
Cost: Free · Free under the AGPL-3.0 license. Ultralytics sells an Enterprise License for companies that cannot follow the AGPL terms.
License: AGPL-3.0 (Copyleft)
Last update: Oct 21, 2026

Risk signals (5)

Attentionstale releases
Last release on 2022-11-22 (1437 days ago)
automatic check
Infocode downloaded at run time
README quick start loads the model with torch.hub.load("ultralytics/yolov5", "yolov5s"), which fetches the repository code from GitHub when it runs; README also says "Models are automatically downloaded from the latest YOLOv5 release".
AI review
Infosuperseded by newer package
README: "the actively maintained `ultralytics` package brings the latest Ultralytics YOLO models together in one place"; requirements.txt also depends on "ultralytics>=8.4.164".
AI review
Infocommercial license upsell
README: "To request an Enterprise License, please complete the form at Ultralytics Licensing" (the code itself is AGPL-3.0, which obliges you to publish your own changes, including when the program is only reached over a network).
AI review
Infodatasets downloaded automatically
README training section: "Both models and datasets are downloaded automatically from the latest YOLOv5 release", and the validation examples note downloads such as "Download ImageNet validation split (6.3GB, 50,000 images)". Commands can pull gigabytes without asking again.
AI review

Evidence is shown in its original language.

What Radar could not check

Whether the install and scripts work on Windows without errors.
Real disk use after installing PyTorch (with or without graphics card support) and downloading model files.
How fast detection runs on a normal computer without a graphics card.
Which files the scripts download from the internet at run time besides the model weights.
Whether the code still matches the last release, which dates from 2022, since commits continue.