These are simple examples of using PyMC with MLflow, taking advantage of the pymc_marketing.mlflow module.

This focuses on logging parameters, metrics, and artifacts to MLflow.

Suggestions or Questions? Comment on this Issue

There are four scripts:

1. Non-PyMC example showing how to log parameters, metrics, and artifacts to MLflow
2. PyMC example which logs some PyMC related metrics to MLflow
3. Logging that and more with pymc_marketing.mlflow module
4. Autologging of Marketing Mix Model with pymc_marketing.mlflow module

Kick them off with make experiments. View with make serve. Clean up with make clean_up.

Use the environment.yml file to create the conda environment. i.e. conda env create -f environment.yml.

There are some helper functions in the utils.py file which help setup mlflow and define some reused PyMC models.

• pymc_marketing.mlflow module
• MLflow Documentation