When working with Azure ML Studio notebooks, it is important to ensure that the correct Python kernel is being used. In some cases, you may need to switch from the default kernel to a specific version. In this article, we will explore three different ways to solve the problem of using a Python 3.6 kernel instead of a 3.8 kernel in Azure ML Studio notebooks.

Option 1: Changing the Kernel in the Notebook Settings

The first option is to change the kernel directly in the notebook settings. To do this, follow these steps:

1. Open the Azure ML Studio notebook.
2. Click on the “Kernel” menu.
3. Select “Change Kernel” from the dropdown menu.
4. Choose the desired Python version (in this case, 3.6).
5. Save the changes and restart the kernel.

This method is straightforward and does not require any coding. However, it is important to note that the available kernel versions may vary depending on your Azure ML Studio configuration.

Option 2: Specifying the Kernel in the Notebook Metadata

If you want to ensure that a specific kernel is always used for a particular notebook, you can specify it in the notebook metadata. Here’s how:

1. Open the Azure ML Studio notebook.
2. Click on the “Edit Metadata” button (represented by a pencil icon).
3. In the metadata JSON, locate the “kernelspec” field.
4. Change the value of “name” to the desired kernel (e.g., “python3.6”).
5. Save the changes.

By specifying the kernel in the notebook metadata, you ensure that the correct version is used every time the notebook is opened, regardless of the default kernel settings.

Option 3: Using a Magic Command

If you prefer to change the kernel within the notebook itself, you can use a magic command. Magic commands are special commands in Jupyter notebooks that provide additional functionality. To switch to a specific kernel, use the following magic command:

%%python3.6

Place this command at the beginning of your notebook to switch to the Python 3.6 kernel. Any code executed after this command will use the specified kernel.

Conclusion

All three options provide a solution to the problem of using a Python 3.6 kernel instead of a 3.8 kernel in Azure ML Studio notebooks. The best option depends on your specific requirements and preferences. If you only need to change the kernel for a single notebook, option 1 or 2 may be more suitable. However, if you want to switch kernels within the notebook itself, option 3 using magic commands is the way to go.