MLflow - Databricks - 2018 - Databricks

mlFlow

Simplifying the End-to-End
Machine Learning Lifecycle

Sign Up For Updates

Image
Icon section

Track experiments

Quickly run experiments with any ML library, framework, or language, locally or in the cloud, and on any platform.

Automatically keep track of parameters, results, code, and data from each experiment, and reproduce runs.

Interactively explore results in one place, and identify best performing models across multiple users.

Icon section

Share and reuse projects

Package projects with a standard format that integrates with Git and Anaconda and capture dependencies like libraries, parameters, and data.

Share knowledge and quickly test new ideas by running MLflow projects available on GitHub or other file storage systems.

Manage hyperparameter tuning and cross-validation more efficiently by launching multiple runs in parallel and compare results based on inputs.

Image
Image
Icon section

Productionize models faster

Quickly deploy models to any platforms based on your needs, locally or in the cloud, from experimentation to production.

Run batch inference in near real-time and unmatched performance using Apache SparkTM.

Expose models via REST APIs with built-in integration with Docker containers, Azure ML, or Amazon SageMaker.

Coming soon

mlFlow on

Benefit from managed MLflow on Databricks Unified Analytics Platformfor business-critical AI applications.

Ready to get started?