Prerequisite topics and steps
This section describes the prerequisite topics and steps for setting up our Vertex AI Workbench environment.
Custom Jupyter kernels and package dependency management
When we run our code in a Jupyter Notebook (such as a Vertex AI Workbench Notebook), the environment in which our code executes is referred to as a kernel. Vertex AI Workbench instances come with various kernels already installed for popular tools and frameworks such as TensorFlow and PyTorch, which we will cover in more depth in this chapter.
However, we can also create custom kernels if we want to define isolated environments with specific packages installed. This is a good practice to follow when using packages that are in preview mode, for example, as they may have very specific dependency requirements. We will use one such library, called bigframes, which I will describe in detail in this chapter. As a prerequisite, I will outline how to create a custom Jupyter kernel and explain...
