Data Versioning and Reproducible ML with DVC and MLflow

Learn how to enhance machine learning development with data versioning and reproducible workflows in this 27-minute tutorial from Databricks. Explore the capabilities of Data Version Control (DVC) and MLflow to manage datasets, track models, and improve reproducibility in ML projects. Discover how these tools integrate with Git to overcome limitations in storing large files and tracking model artifacts. Follow along with a sample ML project to implement best practices for versioning data, tracking experiments, and packaging models for deployment. Gain insights into configuring remote storage, managing existing data, and creating revisions to streamline your ML development process.

Intro
Agenda
Data Science Project
Data is Everywhere
Why is it important
What are the solutions
DVC and MLflow
DVC
MLflow
Git and DVC
Demo
Configure Remote Storage
Copy Existing Data
Tracking Existing Data
Revision
Conclusion