MLflow

30-minute quick start workshop ..

MLflow

Get hands-on experience with MLflow's core features in record time.

By the end of this workshop, you'll have:

• Tracked your first ML experiment

• Logged parameters, metrics, and models

• Used the MLflow UI to compare experiments

• Save and loaded models

Introduction

Machine Learning Models in Pentaho Data Catalog

Pentaho Data Catalog's ML Models feature integrates machine learning workflows into your data cataloging ecosystem. You can connect to ML model servers, import components and metadata, and organize models, experiments, versions, and runs within a structured hierarchy alongside your enterprise data.

Local Management Capabilities

The feature allows you to add ML model servers and build complete model hierarchies locally without requiring live connections to MLflow or other tracking servers. This is particularly useful for managing legacy models, internal-only components, or maintaining records of decommissioned ML assets that no longer exist on their original servers.

Lifecycle and Migration Support

You can create, edit, and remove ML components to keep your catalog clean and current. The system also supports importing and exporting ML model hierarchies, enabling you to reuse, back up, and migrate ML assets across different environments while maintaining governance standards and business policy alignment.

x

First Tab

Second Tab

x