YOLO (Ultralytics)

Ultralytics YOLO is an open-source Python computer vision library for developers building applications that analyze images and video on their own hardware. It supports local and edge deployment, including NVIDIA Jetson, Raspberry Pi and mobile phones. A separate hosted platform provides browser-based annotation, cloud GPU training and managed prediction endpoints.

The models detect objects, produce object masks, classify images and estimate body keypoints. They also support semantic segmentation, rotated bounding boxes and depth estimation. Video tracking follows detected objects across frames, with support for detection, segmentation, pose and oriented detection models. These capabilities suit work such as factory inspection, inventory monitoring and robotics perception.

You can train on your own data or use pretrained models. The library supports training, validation and prediction within Python projects, while export formats include ONNX, TensorRT, OpenVINO, CoreML and LiteRT. CPU inference and NVIDIA GPU acceleration provide different deployment options. Integrations include Weights & Biases, Comet ML and Roboflow for experiment tracking and dataset work.

The library uses the AGPL-3.0 license; Ultralytics also offers enterprise licensing with different terms. Hosted training runs on cloud GPUs, while exported models can run on devices you control. The hosted platform includes SAM-assisted annotation, dataset review and versioning, plus training metrics and experiment comparison. Managed endpoints provide request metrics and logs.