docker compose build
docker compose up
pip install -r backend/requirements.txt
pip install -r frontend/requirements.txt
cd backend && python src/app.py
cd .. && streamlit run frontend/src/app.py --server.port 8080
- go to localhost:8080 and have fun :)
- mount volume to save downloaded weights and allow changes to source code
- add customizable categories
- load dataset and export tagging as metadata in coco format