Skip to content

daisukeiot/OpenVINO-Toolkit-Setup

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Introduction

Collection of scripts to set up OpenVINO environment

  • Physical Device
  • Containers

Motivation

As a developer, we want to spend our time building applications with AI, however :

  • Setting up environment takes time
  • Depending on OS, you have to know different ticks and tricks
  • Containerization is one way to solve but building container is not that simple

For physical device set up, write scripts so we can just clone the repo and run the scripts.

For containerization, instead of building container from scratch, you can create base image and just add your application.

This saves time and disk space on your development machine.

Assumptions

  • All instruction assumes you clone this repo to your home directory

    • ~/ on Linux
    • C:\ on Windows
  • To make container images smaller, all docker commands uses --squash experimental feature

    Enable Experimental feature by modifying /etc/docker/daemon.json

    Create daemon.json if it does not exist

    {
      "experimental": true
    }

    [!WARNING]
    --squash option did not work with Docker Desktop v19.03.8

Physical Device Setup

These instructions cover how to setup OpenVINO Toolkit in silent mode, including libraries and tools.
Once you set up your host machine, you can :

  • Compile OpenVINO samples
  • Develop OpenVINO Application
  • Run OpenVINO application on the host machine
  • Host containerized OpenVINO application (or container), including IoT Edge Module

Currently tested with following Operating Systems

Running Object Detection Python App

Please follow this instruction to run Object Detection App

Container Setup

This section covers multi-stage container build scripts and techniques.

Currently tested with Ubuntu 16.04 nd 18.04 on UP2

  • Ubuntu 18.04
  • Raspbian (RP3 and RP4) : To be added in the future
  • Windows 10 : To be added in the future

Running Containerized Object Detection Python App

Please follow this instruction to run containerized Object Detection App