itlab_2024

Build and run ultralytics inference example

1. Clone ultralyrics repository

   git clone https://github.com/ultralytics/ultralytics
   cd ultralytics
   

2. Get a YOLO model

   We aslo will need a YOLO model in ONNX format.

   Create and activate python venv

   python -m venv <venv dir>
   source <venv dir>/bin/activate
   

   Install ultralytics using pip

   pip install ultralytics
   

   Get a YOLO model

   yolo export model=yolov8s.pt imgsz=480,640 format=onnx
   

3. Build and run the example

   cd examples/YOLOv8-CPP-Inference
   cmake -S . -B build
   cd build
   make
   ./Yolov8CPPInference
   

   Note that by default the CMake file will try to import the CUDA library to be used with the OpenCVs dnn (cuDNN) GPU Inference. If your OpenCV build does not use CUDA/cuDNN you can remove that import call and run the example on CPU.

How to build ONNX library on Linux(Ubuntu)

1. Install necessary tools:

sudo apt-get install -y python3-pip
sudo apt-get install -y python3-venv
sudo apt-get install python3-pip python3-dev libprotobuf-dev protobuf-compiler

2. Open directory where you want to save the ONNX library.

3. Open a terminal and execute these commands:

cd onnx
git submodule update --init --recursive
export CMAKE_ARGS='-DONNX_USE_LITE_PROTO=ON -DONNX_USE_PROTOBUF_SHARED_LIBS=ON'

when you use pip, you need to use venv to avoid conflict between package managers (apt and pip) in source(onnx) directory execute these commands:

python3 -m venv .venv
source .venv/bin/activate
pip install -e . -v

4. Run to verify it works.:

 python -c "import onnx"