Created by exporting the customvision model as tensorflow zip file and sample code.
-- When using a Raspberry Pi with a Raspberry Pi Camera, due to OpenCV (cv2), it requires legacy camera to be enabled in the raspi-config.
cd
python -m venv envtfliteobjectdetection
. envtfliteobjectdetection/bin/activate
git clone https://github.com/NoliAlonso/TFLiteObjectDetection
cd TFLiteObjectDetection
pip install -r requirements.txt
python detect.py
Expected performance is 1 FPS or less. Noticeably faster with the float16 model.