Install DL Workbench on Linux* OS

These instructions show how to run the DL Workbench from Docker Hub* on Linux, whether with a starting script or a docker run command. See Install DL Workbench on Docker Hub to run the DL Workbench on other operating systems.

Prerequisite

To run DL Workbench on Linux, install Docker CE for Linux first.

Run Docker Container with the Starting Script

These steps show you how to use a script that starts or updates the DL Workbench on Ubuntu. If you do not want to use a script, use the alternative instructions.

To quickly start or update the DL Workbench in a Docker container, use one of these options:

Use a single command:

wget https://raw.githubusercontent.com/openvinotoolkit/workbench_aux/master/start_workbench.sh && bash start_workbench.sh

Manually start the script:

  1. Open a terminal and download the script:
    wget https://raw.githubusercontent.com/openvinotoolkit/workbench_aux/master/start_workbench.sh
  2. Make sure you are in the directory that contains the downloaded script. Make the file executable:
    chmod +x start_workbench.sh
  3. Start the DL Workbench. If you are behind a firewall, include your proxy values according to the DL Workbench Command-Line arguments.
    ./start_workbench.sh -IMAGE_NAME openvino/workbench

NOTES:

  • To download the script using the command line, you must have GNU Wget installed.
  • While the single command option is the simplest, it is also limiting as it uses default settings. Run the script with more parameters if you need to specify proxy settings or if you want to use hardware other than a CPU. For greater control, use the steps to manually start the script.

Run the DL Workbench Starting Script with More Parameters

You can run the Docker container in either interactive mode or detached mode.

  • Interactive mode - You are asked for input and see output on the terminal screen. In interactive mode, press Ctrl + C to stop the container.
  • Detached mode - The container runs as a background process. You are not asked for input and do not see output on the terminal screen. In detached mode, stop the container with the docker stop workbench command. To monitor logs in the detached mode, run docker logs workbench.

The parameter set depends on the targets you have on your machine. Choose the combination that suits you and provide appropriate arguments in the starting script:

NOTE: To enable Intel® Vision Accelerator Design with Intel® Movidius™ VPUs in the DL Workbench, install the OpenVINO™ toolkit for Linux* first.

CPU

Choose your preferred mode and run the corresponding commands:

Interactive mode:

  1. Start the DL Workbench. If you are behind a firewall, include your proxy values according to the DL Workbench command-line arguments.
    ./start_workbench.sh -IMAGE_NAME openvino/workbench
  2. Go to http://127.0.0.1:5665 when this link displays in the console log. The link is followed by a generated token. See Enter DL Workbench.

Detached mode:

  1. Start the DL Workbench. If you are behind a firewall, include your proxy values according to the DL Workbench command-line arguments.
    ./start_workbench.sh -IMAGE_NAME openvino/workbench -DETACHED -ASSETS_DIR ~/.workbench
  2. Wait about one minute to provide time for the DL Workbench to set up.
  3. Open the http://127.0.0.1:5665 link .
  4. Enter the token from the ~/.workbench directory. To store the token in a different directory, replace ~/.workbench with the correct directory location. See Enter DL Workbench for details.

You have successfully installed the OpenVINO™ DL Workbench and are ready to work with the models and sample datasets.

CPU and Intel® Processor Graphics (GPU)

Choose your preferred mode and run the corresponding commands:

Interactive mode:

  1. Start the DL Workbench. If you are behind a firewall, include your proxy values according to the DL Workbench command-line arguments.
    ./start_workbench.sh -IMAGE_NAME openvino/workbench -ENABLE_GPU
  2. Go to http://127.0.0.1:5665 when this link displays in the console log. The link is followed by a generated token. See Enter DL Workbench.

Detached mode:

  1. Start the DL Workbench. If you are behind a firewall, include your proxy values according to the DL Workbench command-line arguments.
    ./start_workbench.sh -IMAGE_NAME openvino/workbench -ENABLE_GPU -DETACHED -ASSETS_DIR ~/.workbench
  2. Wait about one minute to provide time for the DL Workbench to set up.
  3. Open the http://127.0.0.1:5665 link.
  4. Enter the token from the ~/.workbench directory. To store the token in a different directory, replace ~/.workbench with the correct directory location. See Enter DL Workbench for details.

You have successfully installed the OpenVINO™ DL Workbench and are ready to work with the models and sample datasets.

CPU and Intel® Neural Compute Stick 2

Choose your preferred mode and run the corresponding commands:

Interactive mode:

  1. Start the DL Workbench. If you are behind a firewall, include your proxy values according to the DL Workbench command-line arguments.
    ./start_workbench.sh -IMAGE_NAME openvino/workbench -ENABLE_MYRIAD
  2. Go to http://127.0.0.1:5665 when this link displays in the console log. The link is followed by a generated token. See Enter DL Workbench.

Detached mode:

  1. Start the DL Workbench. If you are behind a firewall, include your proxy values according to the DL Workbench command-line arguments.
    ./start_workbench.sh -IMAGE_NAME openvino/workbench -ENABLE_MYRIAD -DETACHED -ASSETS_DIR ~/.workbench
  2. Wait about one minute to provide time for the DL Workbench to set up.
  3. Go to http://127.0.0.1:5665.
  4. Enter the token from the ~/.workbench directory. To store the token in a different directory, replace ~/.workbench with the correct directory location. See Enter DL Workbench for details.

You have successfully installed the OpenVINO™ DL Workbench and are ready to work with the models and sample datasets.

CPU and Intel® Vision Accelerator Design with Intel® Movidius™ VPUs

NOTE: You must have the OpenVINO toolkit installed before using these instructions. See the prerequisites. Do not remove the archive you used to install the OpenVINO™ toolkit. The DL Workbench uses this archive.

Step 1. Use the Configuration Guide for Intel® Vision Accelerator Design with Intel® Movidius™ VPUs to set up the dependencies on the host system. This is required to execute hddldaemon, which communicates between the HDDL plugin and the board.

Step 2. In the terminal, run hddldaemon on the host machine:

${HDDL_INSTALL_DIR}/bin/hddldaemon

Step 3. Choose your preferred way to run the Docker container and run the corresponding commands:
Interactive mode:

  1. Start the DL Workbench. If you are behind a firewall, include your proxy values according to the DL Workbench command-line arguments .
    ./start_workbench.sh -IMAGE_NAME openvino/workbench -ENABLE_HDDL
  2. Go to http://127.0.0.1:5665 when this link displays in the console log. The link is followed by a generated token. See Enter DL Workbench.


Detached mode:

  1. Start the DL Workbench. If you are behind a firewall, include your proxy values according to the DL Workbench command-line arguments.
    ./start_workbench.sh -IMAGE_NAME openvino/workbench -ENABLE_HDDL -DETACHED -ASSETS_DIR ~/.workbench
  2. Wait about one minute to provide time for the DL Workbench to set up.
  3. Go to http://127.0.0.1:5665.
  4. Enter the token from the ~/.workbench directory. To store the token in a different directory, replace ~/.workbench with the correct directory location. See Enter DL Workbench for details.

You have successfully installed the OpenVINO™ DL Workbench and are ready to work with the models and sample datasets.

CPU, GPU, and Intel® Neural Compute Stick 2

Choose your preferred way to run the Docker container and run the corresponding commands:

Interactive mode:

  1. Start the DL Workbench. If you are behind a firewall, include your proxy values according to the DL Workbench command-line arguments.
    ./start_workbench.sh -IMAGE_NAME openvino/workbench -ENABLE_MYRIAD
  2. Go to http://127.0.0.1:5665 when this link displays in the console log. The link is followed by a generated token. See Enter DL Workbench.

Detached mode:

  1. Start the DL Workbench. If you are behind a firewall, include your proxy values according to the DL Workbench command-line arguments.
    ./start_workbench.sh -IMAGE_NAME openvino/workbench -ENABLE_MYRIAD -DETACHED -ASSETS_DIR ~/.workbench
  2. Wait about one minute to provide time for the DL Workbench to set up.
  3. Go to http://127.0.0.1:5665.
  4. Enter the token from the ~/.workbench directory. To store the token in a different directory, replace ~/.workbench with the correct directory location. See Enter DL Workbench for details.

CPU, GPU, and Intel® Vision Accelerator Design with Intel® Movidius™ VPUs

NOTE: You must have the OpenVINO toolkit installed before using these instructions. See the prerequisites. Do not remove the archive you used to install the OpenVINO™ toolkit. The DL Workbench uses this archive.

Step 1. Use the Configuration Guide for Intel® Vision Accelerator Design with Intel® Movidius™ VPUs to set up the dependencies on the host system. This is required to execute hddldaemon, which communicates between the HDDL plug-in and the board.

Step 2. In the terminal, run hddldaemon on the host machine:

${HDDL_INSTALL_DIR}/bin/hddldaemon

Step 3. Choose your preferred way to run the Docker container and run the corresponding commands:

Interactive mode:

  1. Start the DL Workbench. If you are behind a firewall, include your proxy values according to the DL Workbench command-line arguments.
    ./start_workbench.sh -IMAGE_NAME openvino/workbench -ENABLE_GPU -ENABLE_HDDL
  2. Go to http://127.0.0.1:5665 when this link displays in the console log. The link is followed by a generated token. See Enter DL Workbench.

Detached mode:

  1. Start the DL Workbench. If you are behind a firewall, include your proxy values according to the DL Workbench command-line arguments.
    ./start_workbench.sh -IMAGE_NAME openvino/workbench -ENABLE_GPU -ENABLE_HDDL -DETACHED -ASSETS_DIR ~/.workbench
  2. Wait about one minute to provide time for the DL Workbench to set up.
  3. Go to http://127.0.0.1:5665.
  4. Enter the token from the ~/.workbench directory. To store the token in a different directory, replace ~/.workbench with the correct directory location. See Enter DL Workbench for details.

You have successfully installed the OpenVINO™ DL Workbench and are ready to work with the models and sample datasets.

Run Docker Container with docker run Command

This section gives you an alternative to run the Docker container if you do not want to use a script.

To start a Docker container with CPU support, run the following:

docker run -p 127.0.0.1:5665:5665 --name workbench -it openvino/workbench:latest

IMPORTANT: If you have run the DL Workbench before, pull the latest Docker image:

docker pull openvino/workbench

For more options, like enabling other devices or changing container settings, see the Parameter Command-Line arguments.

You have completed the instructions to run the DL Workbench in a Docker container on Ubuntu.