From the course: LLM Foundations: Vector Databases for Caching and Retrieval Augmented Generation (RAG)
Unlock the full course today
Join today to access over 23,400 courses taught by industry experts.
Indexes in Milvus
From the course: LLM Foundations: Vector Databases for Caching and Retrieval Augmented Generation (RAG)
Indexes in Milvus
In this video, we will review indexing capabilities in Milvus. What are Milvus indexes? Indexes play a key role in any database product to help speed up data retrieval. Milvus also has indexes to help with the same. In Milvus, we can create indexes on either scalar or vector fields. Scalar indexes work similar to RDBMS indexes. There can only be one index per field. Also, there is no support for composite indexes in Milvus. Indexes help organize vectors using an ANN metric type like L2 or IP. The index is set up in such a way that will help search using these metrics efficiently. Vector indexes are a prerequisite to perform ANN searches on vector fields. These indexes must be created before any such search. Milvus supports several index types which can be chosen based on the use case. Here is a list of popular index types in Milvus. This list is not exhaustive. Each index type has a specific internal organization. FLAT indexes are suitable for small dataset and have a 100% recall…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.
Contents
-
-
-
-
Introduction to Milvus DB1m 53s
-
(Locked)
Milvus architecture2m 37s
-
(Locked)
Collections in Milvus3m 35s
-
(Locked)
Partitions in Milvus1m 15s
-
(Locked)
Indexes in Milvus1m 50s
-
(Locked)
Managing data in Milvus1m 38s
-
(Locked)
Query and search in Milvus4m 5s
-
(Locked)
Set up Milvus and exercise files4m 55s
-
-
-
-
-
-
-