Goal is to deploy a simple ML workload (object detection) and scale it using kubernetes
# make new venv
python -m venv E:\venvs\ml-kubernetes
# activate venv
E:\venvs\ml-kubernetes\Scripts\activate.bat
cd api && uvicorn server:app --reload
cd api && pytest
- fastapi
- "uvicorn[standard]" # server
- httpx # http client for testing
- pytest # test framework
- pytest-asyncio # for async functions
- opencv # for yolo
- requests # for downloading model
- python-multipart # for file upload handling
- flake8 # code linting
- black # code formatting
cd api && pip freeze > requirements.txt
- learn more about python modules
- how init.py helps?
- different b/w relative and abs imports