This project provides a Rust implementation of the YOLO v11 object detection model, enabling inference on images to identify objects along with their bounding boxes, labels, and confidence scores. It utilizes the YOLO v11 model in ONNX format and leverages the ort library for ONNX Runtime integration. The implementation is inspired by the YOLOv8 example from the ort repository.
See docs.rs for the latest documentation.
- Object Detection: Detects objects within an image and provides their bounding boxes, labels, and confidence scores.
- ONNX Model Integration: Employs the YOLO v11 model in ONNX format for efficient inference.
- Rust Implementation: Written entirely in Rust, ensuring performance and safety.
- ONNX Runtime: Utilizes the
ortlibrary for executing the ONNX model.
On a MacBook Pro (2024) with M3 Max, it tooks about 57ms to inferring an image with the YOLO11x model.
yolo-cli inference logs on MacBook Pro (2024) with M3 Max
2025-07-11T13:15:07.344683Z INFO example_yolo_gui: Inference took 56.944167ms
2025-07-11T13:15:07.344765Z INFO example_yolo_gui: Found entity "person" with confidence 0.91 at (23.53, 325.31) - (127.34, 480.19)
2025-07-11T13:15:07.344813Z INFO example_yolo_gui: Found entity "person" with confidence 0.91 at (268.38, 285.00) - (349.62, 480.00)
2025-07-11T13:15:07.344843Z INFO example_yolo_gui: Found entity "person" with confidence 0.89 at (106.31, 373.31) - (238.69, 480.19)
2025-07-11T13:15:07.344865Z INFO example_yolo_gui: Found entity "baseball glove" with confidence 0.75 at (210.94, 409.50) - (240.06, 453.00)
2025-07-11T13:15:07.344875Z INFO example_yolo_gui: Found entity "person" with confidence 0.66 at (20.17, 276.28) - (64.45, 364.97)
2025-07-11T13:15:07.344890Z INFO example_yolo_gui: Found entity "baseball bat" with confidence 0.50 at (222.94, 372.84) - (275.81, 381.66)
- yolo-cli: Command-line interface for running YOLO inference on images.
This project is inspired by the YOLOv8 example from the ort repository.
This project is dual-licensed under the MIT License and Apache-2.0 LICENSE. See the LICENSE file for details.