Overview

Welcome to H2O.ai MLOps. MLOps lets you export your Driverless AI experiments and H2O-3 MOJO models into a shared MLOps project for collaboration, management, deployment, and monitoring.

This document describes how to use MLOps. For information on how to use Driverless AI and H2O-3, refer to the Driverless AI and H2O-3 user guides.

Release Notes

• MLOps Release Notes
  • Version 0.40.0 (Dec 2020)
  • Version 0.31.3 (Dec 2, 2020)
  • Version 0.31.2 (Nov 11, 2020)
  • Version 0.31.1
  • Version 0.31.0 (Oct 21, 2020)
  • Version 0.30.1 (Oct 8, 2020)
  • Version 0.30.0
  • Version 0.22.0 (July 30, 2020)
  • Version 0.21.1 (July 07, 2020)
  • Version 0.21.0 (Jun 12, 2020)
  • Version 0.20.1 (Apr 02, 2020)
  • Version 0.20.0 (Apr 01, 2020)

Overview

• About MLOps

Installation

• Installing MLOps
• Supported Cloud Environments
  • Supported AWS Regions

Understanding the UI

• Requirements and Workflow
  • Requirements
  • Workflow
• Logging In and Navigating the UI
  • Alerts
  • Deployments
  • Models
  • Events

Using MLOps

• Using MLOps
  • Adding a New Project
  • Exporting Experiments from Driverless AI into MLOps
  • Importing Models from H2O-3 into MLOps
  • Model Deployment and Endpoint Security
  • Sharing Projects

MLOps Python Client

• Getting Started
  • Requirements
  • Installing the Client
  • API Notes
  • API Gateway Healthcheck
• Client
  • Connecting to the platform
  • Making API calls
  • Available grpc servers
  • Client API calls and Request/Response objects
  • Accessing a requested entity
  • Empty body calls
  • Authentication
• Examples
  • requirements.txt
  • Who am I
  • Standard Flow
  • Getting experiment’s metadata
  • Getting deployments’ statuses
  • Pump data
  • Upload and Deploy an H2O3 model
  • Create secure endpoints