Overview

Added in version 2.0.

This document provides insights into CPU only deployments of NVIDIA AI Enterprise running on VMWare vSphere and serves as a technical resource for understanding system pre-requisites, installation, and configuration.

• Prerequisites

• NVIDIA License System

• Installing Docker and the Docker Utility Engine

• Installing AI and Data Science Applications and Frameworks

Once the above installation is complete, we will cover advanced framework configuration details. This includes recommended startup scripts for NVIDIA AI Enterprise for getting started with AI.

Important

The Following NVIDIA AI Enterprise containers can with CPU only:

• Triton Inference Server

• RAPIDS

• PyTorch

• TensorFlow

NVIDIA AI Enterprise Containers can run in both bare metal and virtualized deployments with VMware vSphere, this document is focused running on VMware vSphere.

For more information regarding GPU accelerated deployments, please see the NVIDIA AI Enterprise Bare Metal, VMware vSphere, & OpenShift on VMware deployment guides.