Configure a development environment for Azure Machine Learning

In this article, you learn how to configure a development environment to work with Azure Machine Learning service. Machine Learning service is platform agnostic.

The only requirements for your development environment are Python 3, Anaconda (for isolated environments), and a configuration file that contains your Azure Machine Learning workspace information.

This article focuses on the following environments and tools:

  • Azure Notebooks: A Jupyter Notebooks service that's hosted in the Azure cloud. It's the easiest way to get started, because the Azure Machine Learning SDK is already installed.

  • The Data Science Virtual Machine (DSVM): A pre-configured development or experimentation environment in the Azure cloud that's designed for data science work and can be deployed to either CPU only VM instances or GPU-based instances. Python 3, Conda, Jupyter Notebooks, and the Azure Machine Learning SDK are already installed. The VM comes with popular machine learning and deep learning frameworks, tools, and editors for developing machine learning solutions. It's probably the most complete development environment for machine learning on the Azure platform.

  • The Jupyter Notebook: If you're already using the Jupyter Notebook, the SDK has some extras that you should install.

  • Visual Studio Code: If you use Visual Studio Code, it has some useful extensions that you can install.

  • Azure Databricks: A popular data analytics platform that's based on Apache Spark. Learn how to get the Azure Machine Learning SDK onto your cluster so that you can deploy models.

If you already have a Python 3 environment, or just want the basic steps for installing the SDK, see the Local computer section.


  • An Azure Machine Learning service workspace. To create the workspace, see Create an Azure Machine Learning service workspace.

  • Either the Anaconda or Miniconda package manager.


    Anaconda and Miniconda are not required when you're using Azure Notebooks.

  • On Linux or macOS, you need the bash shell.


    If you're on Linux or macOS and use a shell other than bash (for example, zsh) you might receive errors when you run some commands. To work around this problem, use the bash command to start a new bash shell and run the commands there.

  • On Windows, you need the command prompt or Anaconda prompt (installed by Anaconda and Miniconda).

Azure Notebooks

Azure Notebooks (preview) is an interactive development environment in the Azure cloud. It's the easiest way to get started with Azure Machine Learning development.

  • The Azure Machine Learning SDK is already installed.
  • After you create an Azure Machine Learning service workspace in the Azure portal, you can click a button to automatically configure your Azure Notebook environment to work with the workspace.

To get started developing with Azure Notebooks, see Get started with Azure Machine Learning service.

By default, Azure Notebooks uses a free service tier that is limited to 4GB of memory and 1GB of data. You can, however, remove these limits by attaching a Data Science Virtual Machine instance to the Azure Notebooks project. For more information, see Manage and configure Azure Notebooks projects - Compute tier.

Data Science Virtual Machine

The DSVM is a customized virtual machine (VM) image. It's designed for data science work that's pre-configured with:

  • Packages such as TensorFlow, PyTorch, Scikit-learn, XGBoost, and the Azure Machine Learning SDK
  • Popular data science tools such as Spark Standalone and Drill
  • Azure tools such as the Azure CLI, AzCopy, and Storage Explorer
  • Integrated development environments (IDEs) such as Visual Studio Code and PyCharm
  • Jupyter Notebook Server

The Azure Machine Learning SDK works on either the Ubuntu or Windows version of the DSVM. But if you plan to use the DSVM as a compute target as well, only Ubuntu is supported.

To use the DSVM as a development environment, do the following:

  1. Create a DSVM in either of the following environments:

    • The Azure portal:

    • The Azure CLI:


      • When you use the Azure CLI, you must first sign in to your Azure subscription by using the az login command.

      • When you use the commands in this step, you must provide a resource group name, a name for the VM, a username, and a password.

      • To create an Ubuntu Data Science Virtual Machine, use the following command:

        # create a Ubuntu DSVM in your resource group
        # note you need to be at least a contributor to the resource group in order to execute this command successfully
        # If you need to create a new resource group use: "az group create --name YOUR-RESOURCE-GROUP-NAME --location YOUR-REGION (For example: westus2)"
        az vm create --resource-group YOUR-RESOURCE-GROUP-NAME --name YOUR-VM-NAME --image microsoft-dsvm:linux-data-science-vm-ubuntu:linuxdsvmubuntu:latest --admin-username YOUR-USERNAME --admin-password YOUR-PASSWORD --generate-ssh-keys --authentication-type password
      • To create a Windows Data Science Virtual Machine, use the following command:

        # create a Windows Server 2016 DSVM in your resource group
        # note you need to be at least a contributor to the resource group in order to execute this command successfully
        az vm create --resource-group YOUR-RESOURCE-GROUP-NAME --name YOUR-VM-NAME --image microsoft-dsvm:dsvm-windows:server-2016:latest --admin-username YOUR-USERNAME --admin-password YOUR-PASSWORD --authentication-type password
  2. The Azure Machine Learning SDK is already installed on the DSVM. To use the Conda environment that contains the SDK, use one of the following commands:

    • For Ubuntu DSVM:

      conda activate py36
    • For Windows DSVM:

      conda activate AzureML
  3. To verify that you can access the SDK and check the version, use the following Python code:

    import azureml.core
  4. To configure the DSVM to use your Azure Machine Learning service workspace, see the Create a workspace configuration file section.

For more information, see Data Science Virtual Machines.

Local computer

When you're using a local computer (which might also be a remote virtual machine), create an Anaconda environment and install the SDK by doing the following:

  1. Download and install Anaconda (Python 3.7 version) if you don't already have it.

  2. Open an Anaconda prompt and create an environment with the following commands:

    Run the following command to create the environment.

    conda create -n myenv python=3.6.5

    Then activate the environment.

    conda activate myenv

    This example creates an environment using python 3.6.5, but any specific subversions can be chosen. SDK compatibility may not be guaranteed with certain major versions (3.5+ is recommended), and it's recommended to try a different version/subversion in your Anaconda environment if you run into errors. It will take several minutes to create the environment while components and packages are downloaded.

  3. Run the following commands in your new environment to enable environment-specific ipython kernels. This will ensure expected kernel and package import behavior when working with Jupyter Notebooks within Anaconda environments:

    conda install notebook ipykernel

    Then run the following command to create the kernel:

    ipython kernel install --user
  4. Use the following commands to install packages:

    This command installs the base Azure Machine Learning SDK with notebook and automl extras. The automl extra is a large install, and can be removed from the brackets if you don't intend to run automated machine learning experiments. The automl extra also includes the Azure Machine Learning Data Prep SDK by default as a dependency.

    pip install azureml-sdk[notebooks,automl]

    Use this command to install the Azure Machine Learning Data Prep SDK on its own:

    pip install azureml-dataprep


    If you get a message that PyYAML can't be uninstalled, use the following command instead:

    pip install --upgrade azureml-sdk[notebooks,automl] azureml-dataprep --ignore-installed PyYAML

    It will take several minutes to install the SDK. See the install guide for more information on installation options.

  5. Install other packages for your machine learning experimentation.

    Use either of the following commands and replace <new package> with the package you want to install. Installing packages via conda install requires that the package is part of the current channels (new channels can be added in Anaconda Cloud).

    conda install <new package>

    Alternatively, you can install packages via pip.

    pip install <new package>

Jupyter Notebooks

Jupyter Notebooks are part of the Jupyter Project. They provide an interactive coding experience where you create documents that mix live code with narrative text and graphics. Jupyter Notebooks are also a great way to share your results with others, because you can save the output of your code sections in the document. You can install Jupyter Notebooks on a variety of platforms.

The procedure in the Local computer section installs necessary components for running Jupyter Notebooks in an Anaconda environment. To enable these components in your Jupyter Notebook environment, do the following:

  1. Open an Anaconda prompt and activate your environment.

    conda activate myenv
  2. Launch the Jupyter Notebook server with the following command:

    jupyter notebook
  3. To verify that Jupyter Notebook can use the SDK, create a New notebook, select Python 3 as your kernel, and then run the following command in a notebook cell:

    import azureml.core
  4. If you encounter issues importing modules and receive a ModuleNotFoundError, ensure your Jupyter kernel is connected to the correct path for your environment by running the following code in a Notebook cell.

    import sys
  5. To configure the Jupyter Notebook to use your Azure Machine Learning service workspace, go to the Create a workspace configuration file section.

Visual Studio Code

Visual Studio Code is a cross platform code editor. It relies on a local Python 3 and Conda installation for Python support, but it provides additional tools for working with AI. It also provides support for selecting the Conda environment from within the code editor.

To use Visual Studio Code for development, do the following:

  1. To learn how to use Visual Studio Code for Python development, see Get started with Python in VSCode.

  2. To select the Conda environment, open VS Code, and then select Ctrl+Shift+P (Linux and Windows) or Command+Shift+P (Mac). The Command Pallet opens.

  3. Enter Python: Select Interpreter, and then select the Conda environment.

  4. To validate that you can use the SDK, create and then run a new Python file (.py) that contains the following code:

    import azureml.core
  5. To install the Azure Machine Learning extension for Visual Studio Code, see Tools for AI.

    For more information, see Use Azure Machine Learning for Visual Studio Code.

Azure Databricks

Azure Databricks is an Apache Spark-based environment in the Azure cloud. It provides a collaborative Notebook based environment with CPU or GPU based compute cluster.

How Azure Databricks works with Azure Machine Learning service:

  • You can train a model using Spark MLlib and deploy the model to ACI/AKS from within Azure Databricks.
  • You can also use automated machine learning capabilities in a special Azure ML SDK with Azure Databricks.
  • You can use Azure Databricks as a compute target from an Azure Machine Learning pipeline.

Set up your Databricks cluster

Create a Databricks cluster. Some settings apply only if you install the SDK for automated machine learning on Databricks. It will take few minutes to create the cluster.

Use these settings:

Setting Applies to Value
Cluster name always yourclustername
Databricks Runtime always Any non ML runtime (non ML 4.x, 5.x)
Python version always 3
Workers always 2 or higher
Worker node VM types
(determines max # of concurrent iterations)
Automated ML
Memory optimized VM preferred
Enable Autoscaling Automated ML

Wait until the cluster is running before proceeding further.

Install the correct SDK into a Databricks library

Once the cluster is running, create a library to attach the appropriate Azure Machine Learning SDK package to your cluster.

  1. Choose only one option (no other SDK installation are supported)

    SDK package extras Source PyPi Name      
    For Databricks Upload Python Egg or PyPI azureml-sdk[databricks]
    For Databricks -with-
    automated ML capabilities
    Upload Python Egg or PyPI azureml-sdk[automl_databricks]


    No other SDK extras can be installed. Choose only one of the preceding options [databricks] or [automl_databricks].

    • Do not select Attach automatically to all clusters.
    • Select Attach next to your cluster name.
  2. Monitor for errors until status changes to Attached, which may take several minutes. If this step fails, check the following:

    Try restarting your cluster by:

    1. In the left pane, select Clusters.
    2. In the table, select your cluster name.
    3. On the Libraries tab, select Restart.

    Also consider:

    • Some packages, such as psutil, can cause Databricks conflicts during installation. To avoid such errors, install packages by freezing lib version, such as pstuil cryptography==1.5 pyopenssl==16.0.0 ipython==2.2.0.
    • Or, if you have an old SDK version, deselect it from cluster’s installed libs and move to trash. Install the new SDK version and restart the cluster. If there is an issue after this, detach and reattach your cluster.

If install was successful, the imported library should look like one of these:

SDK for Databricks without automated machine learning Azure Machine Learning SDK for Databricks

SDK for Databricks WITH automated machine learning SDK with automated machine learning installed on Databricks

Start exploring

Try it out:

Create a workspace configuration file

The workspace configuration file is a JSON file that tells the SDK how to communicate with your Azure Machine Learning service workspace. The file is named config.json, and it has the following format:

    "subscription_id": "<subscription-id>",
    "resource_group": "<resource-group>",
    "workspace_name": "<workspace-name>"

This JSON file must be in the directory structure that contains your Python scripts or Jupyter Notebooks. It can be in the same directory, a subdirectory named aml_config, or in a parent directory.

To use this file from your code, use ws=Workspace.from_config(). This code loads the information from the file and connects to your workspace.

You can create the configuration file in three ways:

  • Follow the steps in Create an Azure Machine Learning service workspace: A config.json file is created in your Azure Notebooks library. The file contains the configuration information for your workspace. You can download or copy the config.json to other development environments.

  • Create the file manually: With this method, you use a text editor. You can find the values that go into the configuration file by visiting your workspace in the Azure portal. Copy the workspace name, resource group, and subscription ID values and use them in the configuration file.

    Azure portal

  • Create the file programmatically: In the following code snippet, you connect to a workspace by providing the subscription ID, resource group, and workspace name. It then saves the workspace configuration to the file:

    from azureml.core import Workspace
    subscription_id = '<subscription-id>'
    resource_group  = '<resource-group>'
    workspace_name  = '<workspace-name>'
        ws = Workspace(subscription_id = subscription_id, resource_group = resource_group, workspace_name = workspace_name)
        print('Library configuration succeeded')
        print('Workspace not found')

    This code writes the configuration file to the aml_config/config.json file.

Next steps