Welcome to Yohou-MLflow's documentation!¶
An MLflow integration for saving and serving Yohou forecasters
-
Get Started in 5 Minutes
Register a forecaster, load it in a later run, observe new data, and forecast.
-
Need Help?
Check a model before deploying, forecast through pyfunc, trust your own estimators, keep the registry small.
-
Learn the Concepts
Why loading is strict about versions, and why a saved model cannot choose what to trust.
-
See It In Action
Explore interactive example notebooks.
Table of Contents¶
Getting Started¶
Run the scheduled forecasting loop against a model registry.
Examples¶
Interactive notebooks demonstrating Yohou-MLflow in practice.
How-to Guides¶
Check a model before deploying, forecast through pyfunc, migrate pickled forecasters.
API Reference¶
Complete API reference - classes, functions, and options.
About Saving Forecasters¶
Why saves are verified, loads are strict about versions, and the file cannot extend trust.
License¶
Yohou-MLflow is open source and licensed under the Apache-2.0 License.