Skip to content

Latest commit

 

History

History
87 lines (59 loc) · 8.32 KB

Lab01B.md

File metadata and controls

87 lines (59 loc) · 8.32 KB

Lab 1B: Working with Azure Machine Learning Tools

In this lab, you will explore various tools for working with an Azure Machine Learning workspace.

Before You Start

Before starting this lab, you must have created an Azure Machine Learning workspace by following the instructions in the previous lab.

Task 1: Use the Azure ML SDK in a Compute Instance

You can perform most asset management tasks to set up your environment in the Studio interface, but it's also important to be able to script configuration tasks to make them easier to repeat and automate.

  1. In Azure Machine Learning studio, on the Compute page for your workspace, view the Compute Instances tab, and if necessary, click Refresh periodically until the compute instance you created in the previous lab has started. Then click its Jupyter link to open Jupyter Notebooks on the VM.

  2. In the notebook environment, create a new Terminal. This will open a new tab with a command shell.

  3. The Azure Machine Learning SDK is already installed in the compute instance image, but it's worth ensuring you have the latest version, with the optional packages you'll need in this course; so enter the following command to update the SDK packages:

    pip install --upgrade azureml-sdk[notebooks,automl,explain]

    More Information: For more details about installing the Azure ML SDK and its optional components, see the Azure ML SDK Documentation.

  4. Next, run the following commands to change the current directory to the Users directory, and retrieve the notebooks you will use in the labs for this course:

    cd Users
git clone https://github.com/MicrosoftLearning/DP100

  5. After the command has completed, close the terminal tab and view the home page in your Jupyter notebook file explorer. Then open the Users folder - it should contain an DP100 folder, containing the files you will use in the rest of this lab.

  6. In the Users/DP100 folder, open the 01B - Intro to the Azure ML SDK.ipynb notebook. Then read the notes in the notebook, running each code cell in turn.

Task 2: Set Up a Visual Studio Online Environment

Compute instances in Azure Machine Learning provide an easy to manage Python environment for working with Azure ML without the need to manage your own Python installation. However, sometimes you may want to use your own graphical Python development environment. In this course, we'll use Visual Studio Online to simplify installation, but the principles of using the Azure Machine Learning SDK are the same in any Python environment.

  1. In a new browser tab, navigate to https://online.visualstudio.com, and click Get Started.

  2. Sign into Visual Studio Online using the same Microsoft credentials you used to sign into Azure.

  3. Create a new environment with the following settings, creating a billing plan in your Azure subscription first if prompted:

    • Environment Name: A unique name of your choice
    • Git Repository: MicrosoftLearning/DP100
    • Instance Type: Standard (Linux)
    • suspend idle environment after: 30 Minutes

  4. Wait for your environment to be created, and then click its name to connect to it.

    Visual Studio Online is a hosted instance of Visual Studio Code that you can use in a web browser. Visual Studio Code is a general code editing environment, with support for various programming langaues through the installation of extensions. To work with Python, you'll need the Microsoft Python extension, which was installed for you along with some commonly used Python packages when you created this environment from the DP100 repo.

    The hosted Visual Studio Code environment includes three installations of Python (versions 2.7.13, 3.5.3, and 3.8.0). You will use the Python 3.5.3 virtual environment. In your own installation, you are responsible for installing Python, creating virtual environments, and installing the packages you need. In this lab, most of the general Python configuration has been done for you, but you need to install the Azure Machine Learning SDK.

  5. In the Visual Studio Online environment, wait for the contents of the DP100 repo to be loaded, and then in the Application Menu (☰), on the View menu, click Command Palette (or press CTRL+SHIFT+P). Then in the Palette, enter the command Python: Create Terminal. This opens a Python terminal pane at the bottom of the Visual Studio Online interface. If prompted, select the Python 3.5.3 interpreter.

  6. In the terminal pane, enter the following command to change to the directory where the Python 3.5.3 virtual environment is defined:

    cd /usr/bin

  7. Now install the Azure Machine Learning SDK (with the optional notebooks extra package) using this command:

    sudo pip install azureml-sdk[notebooks]

  8. Close the Terminal pane.

Task 3: Use the Azure ML SDK in Visual Studio Online

Now that you have a Python development environment, you can use the Azure Machine Learning SDK in it. First, you need to get the configuration information required to connect to your Azure Machine Learning workspace.

  1. In a new browser tab, open the Azure portal at https://portal.azure.com, signing in if necessary.
  2. Open the Azure Machine Learning workspace resource you created in the previous lab, and on its Overview page, click Download config.json and download the file to your local computer.
  3. Open the downloaded config.json file in a text editor, and copy it's contents to the clipboard. This file contains the configuration information necessary to connect to your workspace.
  4. In Visual Studio Online, create a new file named config.json in the root folder of your VS Online workspace.
  5. Paste the copied configuration information into the new config.json file in your Visual Studio Online workspace, and save it.
  6. In Visual Studio Online, open the 01B - Intro to the Azure ML SDK.ipynb notebook - this will be loaded in the Jupyter Notebook interface within Visual Studio Online. It may take a while to load the first time the Jupyter Notebooks interface is used, and you may briefly see two panes - one containing the JSON representation of the notebook, and the other containing the notebook visual interface.
  7. When the notebook has loaded, check at the bottom left of the Visual Studio Online interface to verify the current Python virtual environment. This should have changed to Python 3.5.3 based on the configuration settings in the repo. If not, change it.
  8. Read the notes in the notebook, running each code cell in turn, just as you did in the Azure Machine Learning Notebook VM Jupyter environment.

Task 4: Use the Visual Studio Code Azure Machine Learning Extension

If you plan to work with Azure Machine Learning in Visual Studio Online (or a local installation of Visual Studio Code), the Azure Machine Learning extension can help make it easier to work with resources in your workspace without needing to switch between your code development environment and the Azure Machine Learning studio web interface.

  1. In Visual Studio Online, click the Extensions tab (⊞), and search for "Azure Machine Learning". Then install the Azure Machine Learning extension from Microsoft. After the extension has installed, click the Reload Required button to reload the environment with the extension.
  2. In Visual Studio Online, click the Azure tab (Δ) and in the Azure Machine Learning section, expand your subscription and your Azure Machine Learning workspace.
  3. Expand Compute and verify that the aml-cluster compute resource you created in your workspace is listed along with a local compute resource, which in this case represents the Visual Studio Online hosted environment - you can run Azure Machine Learning code experiments on local compute as well as on compute resources defined in the workspace.
  4. Close the Visual Studio Online browser tab.

Note: If you intend to continue straight to the next exercise, leave your compute instance running. If you're taking a break, you might want to close all Jupyter tabs and Stop your compute instance to avoid incurring unnecessary costs.