Apache Submarine is a one-stop platform for machine learning, running cloud-native on Both Kubernetes and Cloud. Submarine’s complete platform deployment and YAML files and Docker images from machine learning frameworks like Tensorflow and Pytorch make deployment and use of the entire system very easy, You can run the Submarine machine learning platform on Kubernetes or Cloud with a single Heml command, and Submarine offers Workbench support for multi-user operations, Data scientists and algorithm engineers perform all the work of data processing, algorithm development, job scheduling, model training, and model Serving from a browser. Submarine offers standard Docker images of machine learning frameworks like Tensorflow, Pytorch, Python, and XGBoost, and you can customize and extend them yourself, providing a completely isolated environment for machine learning operations through Docker. With the resource management capabilities of Kubernetes and Cloud, Submarine supports scheduling of a large number of machine learning jobs.

Speakers:

XunLiu: Didi, Senior technical Specialist, Apache Member, Apache Incubator Mentor, Apache Submarine PMC Chair, Apache Hadoop & Zeppelin Committer

Kevin Su, Software Engineer @ Union.ai, Apache Submarine Committer & PMC, LFAI Flyte core Maintainer and Committer