Now that you have a workspace, let's take a look at how to use it from a Python notebook. We'll modify some deep learning code so we save the trained model to the Azure Machine Learning service workspace as the output of an experiment, and track metrics for the model.
First, we need to install the azureml package as follows:
pip install --upgrade azureml-sdk[notebooks]
The azureml package contains the necessary components to run experiments. In order for it to work, you'll need to create a file called config.json in the root of your machine learning project. If you're working with the sample code for this chapter, you can modify the config.json file in the azure-ml-service folder. It contains the following content: ...
