Intel® DevCloud Container Playground

Developer Guide

  • BETA
  • 04/06/2022
  • Public Content
Contents

Access OpenVINO™ utilities

One of the common use-cases in the container development workflow is to onboard models trained in popular frameworks such as Caffe, Tensorflow, MXNet or ONNX before building or running your containers. Your development environment comes installed with OpenVINO™ tools such as Model Optimizer, Accuracy Checker, Post-Training Optimization Tool (POT) or downloader and convertor utilities from the OpenVINO™ Model Zoo.
Above tools can be accessed using the installed Python virtual environment from the terminal.

Navigate to CLI

Use the top navigation menu to access the
Coding Environment
 to open the JupyterLab interface in a new browser tab. Use the
+
 button from the Jupyterlab file browser to open a
Terminal
 from the Launcher.
If you not able to access the JupyterLab interface, make sure to
Allow Pop-ups
 in your browser.

Activate virtual environment

In a new JupyterLab terminal,
Activate
 the
ov2021.4-venv
 python3 virtual environment.
source /opt/ov2021.4-venv/bin/activate
Your terminal session will reflect the name of your activated virtual environment and the shell will begin with
(ov20201.4-venv)[build@cliservice-.....
 .

Access OpenVINO™ Tools

Use the short-hand names listed in the openvino-dev python package. For example, the model downloader tool can be accessed with below command:
omz_downloader --print_all
For best practices, always
Deactivate
 your virtual environment after use with the
deactivate
 command.
  • Use
    pip install <package name> --user
     to install python packages in the virtual environment.
  • For more information on the all the capabilites and short-hand names of the tools, refer to the openvino-dev python package.

Product and Performance Information

1

Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex.