How to Run Jupyter Notebook in Azure with Azure ML

To run Jupyter Notebook in Azure with Azure ML, you'll need to create a new Azure Machine Learning workspace. This is a crucial step, as it allows you to manage your machine learning resources and configure your environment.

Azure ML workspaces can be created through the Azure portal or using the Azure CLI. You can also use the Azure SDKs for Python or other languages to create a workspace programmatically.

By creating a workspace, you'll be able to access the Azure ML service and start running Jupyter Notebooks. This is where the magic happens, and you'll be able to write and execute code, visualize data, and collaborate with others.

Before you can start using Jupyter Notebooks in an Azure Quantum workspace, you need to have an active Azure account with a subscription. If you don't have an Azure account, you can register for free and sign up for a pay-as-you-go subscription.

To get started with Azure Quantum, you'll need to create a workspace. This is a straightforward process that involves a few clicks.

To be able to use Jupyter Notebooks in an Azure Quantum workspace, you'll need to have the following prerequisites:

To set up Jupyter Notebook in Azure, start by selecting your Azure Quantum workspace in the Azure portal. From there, select Notebooks and then My notebooks, where you can add a new notebook by typing a File Name and selecting Create file.

The first cell of the notebook is populated automatically with the connection string to the Azure Quantum workspace. You can add a code or markdown text cell by selecting + Code or + Markdown.

A notebook auto-saves every 30 seconds, or you can manually save by pressing CTRL + S. This ensures that your work is always up to date and easily accessible.

To run a notebook, find and open the file you want to execute, such as app_jupyter.ipynb. Make sure the running compute and the correct kernel are selected, then go over the first block of code and press Shift and Enter.

To upload notebooks to Azure, you need to select your Azure Quantum workspace in the Azure portal. Select Notebooks, then My notebooks, and finally Upload new. You can drag and drop your notebook files or select Browse for Files to upload them. If you want to overwrite existing files, select Overwrite if already exists.

Make sure to only upload notebooks from trusted sources, as running untrusted notebooks can pose risks. You can find your uploaded notebooks under My notebooks.

Here are the steps to upload notebooks in a concise format:

After uploading your notebooks, you can find them under My notebooks. If you're uploading a notebook from another Azure Quantum workspace, you may need to update the connection information before running the notebook.

Running Jupyter Notebooks in Azure can be a bit tricky, but don't worry, I've got the lowdown. One major limitation is that the free Jupyter Notebook instance will only live for up to 12 hours, after which it will be deleted or recreated, and you'll need to reinstall any custom packages.

This can be frustrating, especially if you're in the middle of a project. Another thing to keep in mind is that Azure Quantum may need to delete your Jupyter Notebook instance to apply security patches or roll out emergency changes. They'll try not to interrupt your workflow, but it's something to be aware of.

Notebook instances also come with some pre-set specs: 2 vCPU and 4 GB of memory. If you're working with large datasets, this might not be enough, but it's a good starting point.

Idle kernels are terminated after 30 minutes, so make sure to keep an eye on your work and rerun any cells that are still running. The location of your storage account and workspace can also affect performance, so try to create them in a region close to where you'll be using your notebooks.

If you need to open more than 10 notebooks concurrently in one region per subscription, you'll need to deploy workspaces across multiple regions. Similarly, a single user can only open up to five notebooks concurrently in one region per subscription.

To get started with Azure, you'll need to create a machine learning workspace. You can set this up using Azure Machine Learning for the Fast.ai practical deep learning for coders course.

First, select Go to resource to find your new machine learning workspace. Select the workspace to begin.

For the course notebooks, you'll need to expose your dashboard using Voilà library. Add a new application setting with WEBSITE_PORT as 8866 to do this. This will make your dashboard accessible.

To create an Azure Machine Learning workspace, you need an Azure subscription, which can be obtained from your organization if they have an agreement with Microsoft, or by creating a free account.

Reach out to your organization to get a subscription if they have one, or create a free account before proceeding.

In the Azure portal, click on the + Create a resource button in the upper-left corner and search for Machine Learning.

Select the Machine Learning option and then click Create to begin the setup process.

Each resource in Azure needs to be created inside a logical collection called a Resource Group, so select Create new under Resource Group if you don't have one already.

You'll need to choose a unique name for your machine learning workspace.

Select a region that's closest to you or your users.

For the sake of this example, we'll leave the rest of the fields as defaults.

If you want to configure your workspace further, you can follow the instructions from Azure documentation.

When you're finished configuring your workspace, click Review + Create to review the settings.

Review the settings and if everything looks good, click Create to start the creation process.

To create an Azure Web App, you'll need to choose Azure Container Registry for Image Source and select your Docker image that you pushed to your Azure Container Registry.

The next step is to set up your app's configuration. You'll need to add a New Application Setting, which is an environmental variable that's passed to the Docker image.

Add WEBSITE_PORT with a value of 8866 to expose your dashboard, which will be running using the Voilà library. This variable is accessible to your code inside the Docker container.

To access files on your Blob storage, you can go to the /ml_blob_storage folder or any path you choose as the Mount path.