You have an Azure Machine Learning workspace. You connect to a terminal session from the Notebooks page in Azure Machine Learning studio. You plan to…

Adding a new Jupyter kernel to an Azure Machine Learning compute instance is an additive process. The standard procedure involves creating a new environment (for example, using Conda), activating it, installing the ipykernel package, and then registering the new environment as a kernel for Jupyter. This process allows multiple kernels to coexist. Deleting the default Python 3.8 - AzureML kernel is not a prerequisite and is an unrelated, unnecessary action for the goal of adding a new kernel.

A. Yes: This is incorrect because deleting an existing kernel is not a required step to add a new one. Jupyter is designed to support multiple, coexisting kernels.

1. Microsoft Learn. (2023). Add new Jupyter kernels to a compute instance. This document outlines the correct procedure for adding a new kernel. The steps include creating a new Conda environment

activating it

and installing ipykernel. The procedure does not mention deleting any existing kernels as a prerequisite. The entire process is additive.

Reference: In the main body of the document

under the steps to add a kernel

the commands shown are conda create

conda activate

pip install ipykernel

and python -m ipykernel install. None of these steps involve deletion.

2. Microsoft Learn. (2023). Customize a compute instance with a script. This document describes how to automate the customization of a compute instance

including the installation of new Jupyter kernels. The example scripts demonstrate the creation of new Conda environments and kernels without removing the default ones

reinforcing that deletion is not a required step.

Reference: The section "Example startup scripts" provides code for creating new Conda environments and kernels

which can be run on instance creation. This automation adds to

rather than replaces

the existing configuration.