GPU ACCELERATED YOLOv5-OpenVINO

Note: OpenVINO supports Intel GPUs that is its for use on devices that are equipped with Intel Computing units like Lenovo and others with intel iris(x) and others

# Car Object Detection Using YOLO and OpenVINO accelerated by GPUs (mainly for video inference)

Step 1: Install the Required Libraries

Clone the repository, install dependencies and cd to this local directory for commands in Step 2.

python -m venv venv
source ./venv/bin/activate
pip install --upgrade pip
pip install -r requirements.txt

Note

• Make sure you've downloaded the video we're using for testing from here: https://youtu.be/MNn9qKG2UFI?si=Pt6RE8dt17OV67ne

Use the commands below to run the

Convert the yolov5s.pt to onnx and then convert the onnx model to OpenVINO IR format
use the command to convert *.pt model to .onnx format
python export.py --weights path/to/yolov5s.pt --data_type FP16/FP32/INT8 --include onnx

# Quick run
Loading OpenVINO model on GPU device

python inference-gpu.py --input_path "path/to/sample_video.mp4" --target_device "GPU" --media_type "video" --threshold 0.4(video inferencing on GPU)
python inference-gpu.py --input_path "path/to/sample_image.jpg" --target_device "GPU" --media_type "image" --threshold 0.4(image inferencing on GPU)

Loading OpenVINO model on CPU device
python inference-gpu.py --input_path "path/to/sample_video.mp4" --target_device "CPU" --media_type "video" --threshold 0.4(video inferencing on CPU)
python inference-gpu.py --input_path "C:\Users\Moses\Downloads\Video\sample_video.mp4" --target_device "GPU" --media_type "image" --threshold 0.4(image inferencing on CPU) 

For Running with Native yolov5 use these commands below

# If you want to save results
python main.py --source "path/to/video.mp4" --save-img --view-img

# If you want to run model on CPU
python main.py --source "path/to/video.mp4" --save-img --view-img --device cpu

# If you want to change model file
python main.py --source "path/to/video.mp4" --save-img --weights "path/to/model.pt"

# If you want to detect specific class (first class and third class)
python main.py --source "path/to/video.mp4" --classes 0 2 --weights "path/to/model.pt"

# If you don't want to save results
python main.py --source "path/to/video.mp4" --view-img

Usage Options

• --source: Specifies the path to the video file you want to run inference on.
• --device: Specifies the device cpu or 0
• --save-img: Flag to save the detection results as images.
• --weights: Specifies a different YOLOv8 model file (e.g., yolov8n.pt, yolov8s.pt, yolov8m.pt, yolov8l.pt, yolov8x.pt).
• --classes: Specifies the class to be detected
• --line-thickness: Specifies the bounding box thickness
• --region-thickness: Specifies the region boxes thickness
• --track-thickness: Specifies the track line thickness

Feel free to report any issue about this