Machine Learning Operator & Controller for Kubernetes

Introduction

Kubernetes offers the facility of extending it's API through the concept of 'Operators' (Introducing Operators: Putting Operational Knowledge into Software). This repository contains the resources and code to deploy an Polyaxon native CRDs using a native Operator for Kubernetes.

This project is a Kubernetes controller that manages and watches Customer Resource Definitions (CRDs) that define primitives to handle, operate and reconcile operations like: builds, jobs, experiments, distributed training, notebooks, tensorboards, kubeflow integrations, ...

Kubeflow, Ray, and Dask operators

This Operator extends natively Kubeflow-Operators (TFJob/PytorchJob/MXNet/XGBoost/MPI/Paddle), Dask Operator, Ray Operator.