Explore the latest advancements in MLflow for productionizing machine learning applications in this keynote presentation from the Data + AI Summit EU 2020. Dive into new features including the Model Registry for model management and review, APIs for automated CI/CD, model schemas to detect data format discrepancies, and integration with model explainability tools. Learn about the challenges of deploying ML applications and how MLOps practices and ML platforms address these issues. Witness a demo on CI/CD and MLOps with MLflow, and gain insights into the PyTorch integration with MLflow for seamless transition from research to production. Discover how these tools and practices can help manage complex ML applications, catch potential failures, and streamline the productionization process.
Taking Machine Learning to Production with MLflow - New Features and Best Practices
Databricks via YouTube
Overview
Syllabus
Introduction
Building ML applications is complex
What are ML platforms
MLflow
Community
MLflow PyTorch
MLflow Tracking
Data Versioning with Delta Lake
Model Schema Tracking
Interpretability
Model Registry
Model Industry
Databricks Model Registry
Tags and Search APIs
Model Registry Webhooks
Model Registry Comments
GA Demo
Webhooks
Webhook Setup
Wrapup
Keynote Presentation
EndtoEnd Workflow
Taught by
Databricks
Related Courses
-
Machine Learning at Scale with MLflow and Apache Spark
-
Simplifying Model Development and Management with MLflow - New Features and Innovations
-
MLflow and Azure Machine Learning - The Power Couple for ML Lifecycle Management
-
MLOps on Azure Databricks with MLflow
2.0 -
Productionalizing Models through CI/CD Design with MLflow
-
State of the Industry - Machine Learning Platforms and MLOps
