• Create Azure Compute https://docs.microsoft.com/en-us/azure/machine-learning/quickstart-create-resources
• Create a new Compute Instance & star this compute instance
• Keep track of the following information:
  • Subscription Id
  • Resource Name
  • Workspace Name (compute workspace created)
  • Compute Name (compute instance created)
• Install Python & dependencies (https://docs.microsoft.com/en-us/python/api/overview/azure/ml/install?view=azure-ml-py)
  • Pip install azureml-core azureml-widgets
• Create a workspace folder as follows
  • code>hello>hello.py

    print("Hello World!")

  • config.json

    {
        "subscription_id": "...",
        "resource_group": "...",
        "workspace_name": "..."
    }

  • sample.ipynb

    from azureml.core import Workspace, Experiment, ScriptRunConfig
    from azureml.widgets import RunDetails
    ws = Workspace.from_config()
    experiment = Experiment(workspace=ws, name='day1-experiment-hello')
    config = ScriptRunConfig(source_directory='./code/hello', script='hello.py', compute_target='cpucluster')
    run = experiment.submit(config)
    RunDetails(run).show()

• Run the notebook cell, output should be as follows:

Notes:

• If you have trouble authenticating, then check out https://github.com/Azure/MachineLearningNotebooks/blob/master/how-to-use-azureml/training/train-on-local/train-on-local.ipynb
• If testing against Azure ML Remote Compute, then launch VS Code from the compute instance, create a notebook as documented above.