/redis-vss

Redis Vector Similarity Search Materials

Primary LanguageJupyter Notebook

Rediscover Redis for Vector Similarity Search

Redis Vector Similarity Search Workshop Materials.

Shared Materials

Video Tutorial (FR) Slides Deck

Table of Contents

Pre-requisites

You need to create a Redis Enterprise DB with RedisJSON and RediSearch modules. Then, Use the public endpoint in the notebooks.

To create a Redis Enterprise database, you can use Redis Cloud or you can provision a cluster in your own infrastructure using TerrAmine.

Capture d’écran 2023-10-06 à 15 59 35

Demo 1: Text Vector Search Open In Colab

In this demo, you will learn how to:

  • Create vector embeddings for text,
  • Persist vector integrations in Redis,
  • Create a Secondary Search Index on these Vectors,
  • Find similarity between a new vector (text) and already persisted vectors.
Capture d’écran 2023-11-24 à 23 53 34

Demo 2: Visual Vector Search Open In Colab

In this second demo, you will learn how to:

  • Create vector embeddings for products (by image),
  • Persist vector embeddings in Redis,
  • Create a Secondary Search Index on these Vectors,
  • Find similarity between a new vector (image) and the already persisted vectors.

ezgif-2-9f2d8adeb3

Demo 2bis: Hybrid Search Open In Colab

In this demo, you will learn how to:

  • Create vector embeddings for products (by image),
  • Persist JSON documents containing the vector embeddings and other fields (e.g., tag, location, price...) in Redis,
  • Create a Secondary Search Index on these documents,
  • Find similarity between a new vector (image) and the already persisted vectors.
  • Search for similarity between a new vector (image) and already persisted vectors, pre-filtered by a tag, a location, or a price range.

Capture d’écran 2023-10-02 à 14 40 28

Demo 3: Semantic Search Open In Colab

In this demo, you will learn how to:

  • Create vector embeddings for a private knowledge base (e.g., White papers, blog posts, newsletters...),
  • Persist vector embeddings in Redis,
  • Create a Secondary Search Index on these Vectors,
  • Search semantically (natural language) for the already persisted vectors (relevant resources),
  • Use Redis as a semantic cache.

Semantic_Search

Demo 4: Retieval-Augmented Generation (RAG) on GCP VertexAI Open In Colab

In this last demo, you will learn how to:

  • Create vector embeddings for a private knowledge base (e.g., PDF files, blogs posts, Database),
  • Persist vector embeddings in Redis,
  • Create a Secondary Search Index on these Vectors,
  • Search semantically (natural language) for the already persisted vectors (relevant resources),
  • Use relevant resources as a prompt context for LLM conversation,
  • Generate an augmented response (natural language) using GCP VertexAI models (PaLM),
  • Use Redis as a standard cache,
  • Use Redis as a semantic cache,
  • Use Redis as Q/A history.

RAG

Demo 4bis: Retieval-Augmented Generation (RAG) on AWS Bedrock Open In Colab

In this last demo, you will learn how to:

  • Create vector embeddings for a private knowledge base (e.g., PDF files, blogs posts),
  • Persist vector embeddings in Redis,
  • Create a Secondary Search Index on these Vectors,
  • Search semantically (natural language) for the already persisted vectors (relevant resources),
  • Use relevant resources as a prompt context for LLM conversation,
  • Generate an augmented response (natural language) using AWS Bedrock models (Anthropic Claude 2),
  • Use Redis as a standard cache,
  • Use Redis as a semantic cache,
  • Use Redis as Q/A history.

RAG-AWS