Classical MLOps With UV

Tutorial site: https://thomasbury.github.io/mlops-uv/

acebet is a compact MLOps tutorial for a tabular tennis match predictor. The repo covers prepared-data validation, feature assembly, model training, FastAPI serving, tests, Docker packaging, and optional MLflow tracking without adding infrastructure that the code does not actually implement.

The canonical documentation source lives in docs/ and is built by Zensical using zensical.toml. For the full walkthrough, start there.

Quickstart

uv sync
just data-check
just feature-state
just train
just test
just serve

Then open http://127.0.0.1:8000/docs for OpenAPI.

Docs Workflow

just docs-serve
just docs-build

Use those commands to preview or build the Zensical site from docs/.

Read More

• docs/getting-started.md for local setup and first run
• docs/classical-mlops-tutorial.md for the end-to-end workflow
• docs/uv/uv-workflows-for-mlops.md for repo command conventions
• docs/user-guide/training-and-inference.md for artifacts and feature-contract details
• docs/api/sync-endpoints.md for the current API surface

GitHub Star History