This tutorial adresses the following items:
- How to include
kedro-mlflow
capabilities in a Kedro project:- create a new kedro project with updated template
- update an existing kedro project
- Configure mlflow inside a "mlflow initialised" Kedro project
- Version and track objects during execution with mlflow:
- Version parameters inside a Kedro project
- Version data inside a Kedro project
- (COMING in 0.3.0) Version machine learning models inside a Kedro project
- (COMING in 0.3.0) Version metrics inside a Kedro project
- Open mlflow ui with project configuration
- Package and serve a Kedro pipeline
This is a step by step tutorial and it is recommended to read the different chapters above order, but not mandatory.
Some advanced capabilities are adressed in the advanced use section:
- (COMING in 0.3.0) launching a Kedro project directly with mlflow through the
MLProject
file.