In the preceding chapter, you learned to load and split documents into manageable chunks. To make those chunks useful for retrieval, a computer must be able to compare them based on their semantic meaning, not just their keywords. This chapter focuses on text embeddings, which are numerical representations of text that capture this meaning. By converting text into vectors, we can use mathematical calculations to find the most relevant information for a user's query.

This chapter shows you how to generate these vector representations using Kerb's embedding module. You will learn about vector similarity metrics, such as cosine similarity, which is used to measure how closely related two pieces of text are. The formula for cosine similarity between two vectors $A$ and $B$ is:

$$\text{similarity} = \cos(\theta) = \frac{A \cdot B}{\|A\| \|B\|}$$

With this foundation, you will build a semantic search function to find relevant document chunks. The chapter concludes with guidance on how to select an appropriate embedding model based on factors like performance, size, and your specific use case.