Use Jupyter notebooks to explore Azure Machine Learning service

For your convenience, we have developed a series of Jupyter Python notebooks you can use to explore the Azure Machine Learning service.

Learn how to use the service with the documentation on this site and use these notebooks to customize them to your situation.

Use one of the paths below to run a notebook server with these sample notebooks. Once the server is running, find tutorial notebooks in tutorials folder, or explore different features in how-to-use-azureml folder.

A managed cloud notebook server

It's easy to get started with your own cloud-based notebook server. The sample notebook and the Azure Machine Learning SDK for Python are already installed and configured for you once you create this cloud resource.

  • The samples are available on the notebook webpage.

A Data Science Virtual Machine (DSVM)

The Azure Machine Learning SDK for Python and notebook server are already installed and configured for you on a DSVM.

After you create a DSVM, use these steps on the DSVM to run the notebooks.

  1. Create an Azure Machine Learning service workspace.

  2. Clone the GitHub repository.

    git clone https://github.com/Azure/MachineLearningNotebooks.git
    
  3. Add a workspace configuration file to the cloned directory using either of these methods:

    • In the Azure portal, select Download config.json from the Overview section of your workspace.

    Download config.json

    • Create a new workspace using code in the configuration.ipynb notebook in your cloned directory.
  4. Start the notebook server from your cloned directory.

    jupyter notebook
    

Your own Jupyter Notebook server

Use these steps to create a local Jupyter Notebook server on your computer.

  1. Use the instructions at Create a Azure Machine Learning service workspace to:

    • Create a Miniconda environment
    • Install the Azure Machine Learning SDK for Python
    • Create a workspace
    • Write a workspace configuration file (aml_config/config.json).
  2. Clone the GitHub repository.

    git clone https://github.com/Azure/MachineLearningNotebooks.git
    
  3. Add a workspace configuration file using any of these methods:

    • Copy the aml_config/config.json file you created in step 1 into the cloned directory.

    • In the Azure portal, select Download config.json from the Overview section of your workspace.

    Download config.json

  4. Start the notebook server from your cloned directory.

    jupyter notebook
    

The setup instructions will install the packages you need to run the quickstart and tutorial notebooks. Other sample notebooks may require installation of additional components. For more information about these components, see Install the Azure Machine Learning SDK for Python.

Azure Notebooks

The sample notebooks and the Azure Machine Learning SDK for Python are already installed and configured for you on Azure Notebooks. The installation and future updates are automatically managed via Azure services.

Use the Azure portal to get started with Azure Notebooks. Open your workspace and from the Overview section, select Get Started in Azure Notebooks.

Next steps

Try these tutorials: