pymc-mlflow
Installation
SKILL.md
PyMC experiment tracking with MLflow
Record how a fit was produced, preserve its inference output, and retrieve it
without changing the consuming project's modeling environment. pymc-mlflow
is this skill's name; the optional integration module is
pymc_marketing.mlflow, not a separate pymc-mlflow package.
Choose the integration
| Need | Reference |
|---|---|
| Track ordinary PyMC models without installing PyMC-Marketing | Manual tracking and DataTree artifacts |
| Use maintained PyMC-Marketing autologging or track CLV fits | Autologging and domain-model tracking |
| Save/restore MMM models or log a prediction wrapper to MLflow | MMM persistence and registry preparation |