yolov9-onnx 0.4.0

Run YOLOv9 MIT model with ONNX Runtime

ONNX YOLOv9 MIT Object Detection

Python scripts performing object detection using the YOLOv9 MIT model in ONNX.

[!CAUTION] I skipped adding the pad to the input image when resizing, which might affect the accuracy of the model if the input image has a different aspect ratio compared to the input size of the model. Always try to get an input size with a ratio close to the input images you will use.

Requirements

• Check the requirements.txt file.
• For ONNX, if you have a NVIDIA GPU, then install the onnxruntime-gpu, otherwise use the onnxruntime library.

Installation

pip install yolov9-onnx

Or, clone this repository:

git clone https://github.com/ibaiGorordo/ONNX-YOLOv9-MIT-Object-Detection.git
cd ONNX-YOLOv9-MIT-Object-Detection
pip install -r requirements.txt

ONNX Runtime

For Nvidia GPU computers: pip install onnxruntime-gpu

Otherwise: pip install onnxruntime

ONNX model

• If the model file is not found in the models directory, it will be downloaded automatically from the release page.
• Or, for exporting the models with a different input size, use the Google Colab notebook to convert the model:
• Available models:
  • MIT: v9-s_mit.onnx, v9-m_mit.onnx, v9-c_mit.onnx
  • Official: gelan-c.onnx, gelan-e.onnx, yolov9-c.onnx, yolov9-e.onnx

Original YOLOv9 MIT model

The original YOLOv9 MIT model can be found in this repository: YOLOv9 MIT Repository

• The License of the models is MIT license: License

Examples

• Image inference:

python image_object_detection.py

• Webcam inference:

python webcam_object_detection.py

• Video inference: https://youtu.be/X_XVkEqgCUM

python video_object_detection.py

https://github.com/user-attachments/assets/71b3ef97-92ef-4ddb-a62c-5e52922a396d

References:

• YOLOv9 MIT model: https://github.com/WongKinYiu/YOLO
• YOLOv9 model: https://github.com/WongKinYiu/yolov9