This requires an older tensorflow version so we have to use pyenv

1. Install pyenv https://github.com/pyenv/pyenv
2. pyenv install 3.7
3. pyenv global 3.7
4. Run python -m venv .env
5. Run source .env/bin/activate
6. Run pip install -r requirements.txt
7. Run wget https://github.com/matterport/Mask_RCNN/releases/download/v2.0/mask_rcnn_coco.h5
8. In another directory run git clone https://github.com/matterport/Mask_RCNN.git and copy the mrcnn folder to the root folder of this project

Development

1. maturin develop
2. PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=python python3 model.py

Release build

1. maturin build --release
2. pip install <path>/<to>/<wheel>.whl --force-reinstall
3. PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=python python3 -O model.py

You may be prompted to install a library related to graphics, like xcb in order to display the image. Follow the instructions.

Ref: https://blog.paperspace.com/mask-r-cnn-in-tensorflow-2-0/