Overview
Learn advanced techniques for managing the complete machine learning lifecycle using MLflow in this comprehensive tutorial. Explore data preparation, artifact handling, parameter tuning, and hyperparameter optimization through randomized search. Discover the benefits of MLflow's Community Edition and gain insights into selecting models for production. Master the MLflow Client, understand Model URI concepts, and implement User-Defined Functions (UDFs) for efficient model deployment. Address challenges related to data silos, define effective experiment paths, and learn strategies for evaluating current models. Engage in hands-on exercises to reinforce your understanding of MLflow models and participate in a Q&A session to clarify any remaining doubts.
Syllabus
Intro
Data Preparation
Artifacts
Parameters
HyperParameter Optimization
Randomized Search
Community Edition
What model should we put to production
MLflow Client
Model URI
UDF
Running the model in production
Data silos
Defining the experiment path
Evaluating the current model
Exercise
MLflow Model
Questions
Taught by
Databricks