Neura Market
yolov5 logo

yolov5

Free

YOLOv5 🚀 in PyTorch > ONNX > CoreML > TFLite

Data AnalyticsFreeFree tier
Inputs: image, videoOutputs: image, video, text
Type
Open Source
Company
Ultralytics
Visit Website GitHub (57.3k stars)

About yolov5

Ultralytics YOLO is a family of real-time computer vision models for object detection, instance segmentation, semantic segmentation, classification, pose estimation, oriented bounding boxes, and tracking. It is available through a single Python package and CLI, and supports export to formats such as PyTorch, ONNX, CoreML, and TFLite. The platform is designed for deployment across a range of hardware, from edge devices to cloud APIs, and is built on advancements in deep learning and computer vision, including end-to-end NMS-free inference. YOLOv5 is one of the earlier models in this family, while YOLO26 and YOLO11 are recommended for stable production workloads. The Ultralytics ecosystem also includes a Platform offering data annotation, cloud training, and deployment services, as well as enterprise licensing for commercial use.

Key Features

Real-time object detection, instance segmentation, semantic segmentation, classification, pose estimation, oriented bounding boxes, and tracking
Available as a Python package and CLI with simple two-command setup
Supports export to PyTorch, ONNX, CoreML, and TFLite formats
End-to-end NMS-free inference for optimized performance
Optimized for edge deployment and adaptable to various hardware platforms
Includes built-in multi-object tracking with BoT-SORT and ByteTrack
Offers ready-made vision applications for counting, heatmaps, queue management, security alarms, and workout analysis
Knowledge distillation feature to train smaller models with guidance from a larger teacher model

Pros & Cons

Pros
  • Open-source and free to use under the AGPL-3.0 license
  • Comprehensive documentation and active community (131.6k GitHub stars)
  • Supports a wide range of computer vision tasks in a single framework
  • Optimized for both edge devices and cloud deployment
  • Includes built-in tracking and ready-made vision applications
  • Knowledge distillation feature improves accuracy without extra inference cost
Cons
  • Free tier (AGPL-3.0) has licensing restrictions for commercial use; an enterprise license is required for commercial applications
  • Model performance can vary depending on the specific task and dataset; fine-tuning is often necessary
  • Requires familiarity with Python and deep learning concepts for custom training and deployment
  • Documentation may be extensive, which can be overwhelming for new users

Best For

Object detection and tracking in video surveillance systemsInstance segmentation for autonomous driving and roboticsPose estimation for sports analytics and fitness trackingClassification and segmentation in medical imagingReal-time object counting and heatmap generation for retail or crowd analysisCustom model training for specialized computer vision tasks

Alternatives to yolov5

annotated_deep_learning_paper_implementations

🧑‍🏫 60+ Implementations/tutorials of deep learning papers with side-by-side notes 📝; including transformers (original, xl, switch, feedback, vit, ...), optimizers (adam, adabelief, sophia, ...), ga

d2l-zh

《动手学深度学习》：面向中文读者、能运行、可讨论。中英文版被70多个国家的500多所大学用于教学。

keras

Deep Learning for humans

face_recognition

The world's simplest facial recognition api for Python and the command line

cs-video-courses

List of Computer Science courses with video lectures.

awesome-scalability

The Patterns of Scalable, Reliable, and Performant Large-Scale Systems

FAQ

Is YOLOv5 free to use?
YOLOv5 is open-source under the AGPL-3.0 license, which is free for non-commercial use. For commercial applications, an enterprise license from Ultralytics is required. Details should be verified on the Ultralytics licensing page.
What tasks can YOLO models perform?
Based on available information, YOLO models support object detection, instance segmentation, semantic segmentation, classification, pose estimation, oriented bounding boxes, and tracking.
Can I deploy YOLO on edge devices?
Yes, the models are optimized for edge deployment and can be exported to formats like CoreML and TFLite for use on mobile and embedded devices.
How do I get started with YOLOv5?
The quickstart guide recommends installing the ultralytics package via pip and running a predict command with a pretrained model. Full instructions are available in the Ultralytics documentation.
Does YOLO support video tracking?
Yes, YOLO includes built-in multi-object tracking using BoT-SORT and ByteTrack, which can be applied to video frames.
What is the difference between YOLOv5 and YOLO26?
YOLO26 is a newer model in the Ultralytics YOLO family, recommended for stable production workloads alongside YOLO11. YOLOv5 is an earlier version. The documentation provides details on model comparisons.