Repository navigation
Final version of ML app with Docker Compose
This release marks the completion of the IPUM course (MLOps intro).
The application serves predictions from a trained ML model (Iris dataset) through a FastAPI server, fully containerized using Docker and Docker Compose.
Key features:
Dependency and environment management via uv and .env files
Secure secrets handling with sops and GPG
Unit and integration tests with pytest
ML model loading and /predict endpoint
FastAPI app serving /docs and /predict
Containerization using Dockerfile and docker-compose.yaml
Verified to run via docker compose up
Usage:
After cloning the repo, run: docker compose up.
Then visit http://localhost:8000/docs for the API interface.