From the course: REST APIs, Vectors, and AI in SQL Server 2025
Join today to access over 25,500 courses taught by industry experts.
Understanding embeddings and vector storage - SQL Server Tutorial
From the course: REST APIs, Vectors, and AI in SQL Server 2025
Understanding embeddings and vector storage
- [Instructor] As mentioned before, the backbone of a large language model and putting it to use are embeddings or vectors. You take any kind of content: numerical, text, images, literally anything, and turn that into a vector, a multidimensional space, which describes your content in a numerical way. That part is key to understand. Let's take a very, very simple example. This is a vector representing a cat in five dimensions, so five different numbers. The key part in understanding vectors is similar content results in similar or nearby vectors. The question in the end is how far is a vector from another vector to understand how far it is from a specific item or dataset. If we compare a cat to a dog, for example, we will see the numbers are not identical, but they're very close to each other compared to, for example, a house. Now, imagine that with not five dimensions, but hundreds of different dimensions. This is…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.
Contents
-
-
-
-
-
(Locked)
large language models in a nutshell1m 44s
-
(Locked)
What LLMs can do with SQL: Conversational and generative use cases35s
-
(Locked)
Understanding embeddings and vector storage5m 21s
-
(Locked)
Demo: Distance-based search9m 42s
-
(Locked)
Scaling and optimizing vector workloads through vector indexes1m 9s
-
(Locked)
Demo: Vector indexes2m 39s
-
(Locked)
-
-