Yolov9 Inference

Jose Sarmiento | josedanielsarmiento219@gmail.com

Resumen

The next repository aims to provide a basic c++ script using std 17 over, to do it and consider the speed The code use OpenCv backend (version 4.8.1), to do it the pytorch model might convert to an dnn supported by Opencv, the framework select is Onnx [Slim]

Steps

1. Download de model you want

   • gelan-c
   • gelan-e
   • yolov9-c
   • yolov9-e

    python download_model.py  --model {MODEL_SELECTED}

2. Download de yolov9 repository to convert the Scripts

    git clone https://github.com/WongKinYiu/yolov9

• It recommended to create a virtual environment to do it, Use miniforge | anaconda | miniconda

    conda create --name yolov9 python=3.11 -y
    conda activate yolov9
    pip install -r ./yolov9/requirements.txt

3. Export the model, note the model can export in the follow format, See more in the export script

   • PyTorch
   • TorchScript
   • ONNX
   • OpenVINO
   • TensorRT
   • CoreML
   • TensorFlow SavedModel
   • TensorFlow GraphDef
   • TensorFlow Lite
   • TensorFlow Edge TPU
   • TensorFlow.js
   • PaddlePaddle

    pip install onnx

    pip install onnxruntime

If you have Gpu, download the backend supported that

    pip install onnxruntime-gpu

Export, depending of your hardware, configure your device (gpu or cpu), view the oficial export file

    python export.py --weights ../yolov9-c.pt --include onnx 

4. Compile

Use Cmake file

    mkdir build and cd build

    cmake .. && make

5. Execute

    ./yolov9_inference <Source> <Model path>

5.1 Example

./yolov9_inference 1 ../yolov9-c.onnx