Discover how Weaviate enables RAG applications! You'll build a RAG workflow end-to-end by-hand to get familiar with the Retrieval-Augmentation-Generation steps. This understanding will enable robust and optimized RAG workflows in Chapter 2 using Weaviate.

From RAGs to Riches!

Text generation with LLMs

Creating text embeddings

Comparing text embeddings

Putting together the "R", "A", "G"

RAG Fundamentals with Weaviate

Although coding out RAG workflows by-hand is fun, you may be missing out on optimizations provided by tools like Weaviate. In this chapter, you'll embed, store, retrieve, and generate responses all using Weaviate!

RAG Workflows with Weaviate

PDF text extraction and chunking

Getting set up with Weaviate

Queries in Weaviate

RAG from top to bottom

End-to-End RAG with Weaviate

In the last chapter, you used the text content from the PDF documents to build your document chunks, but left the image content behind. This results in a lot of lost context that might be useful for retrieval and generation! In this chapter, you'll use ColPali multi-modal models to embed and generate text and images to provide more context for your model responses.

RAG with text and images

Preparing for Multi-Modal RAG

Multi-Modal RAG with Weaviate

Congratulations!

Multi-Modal RAG

Learn how to go from simple LLM calls to multi-modal RAG workflows with Weaviate! You'll learn how to process PDF documents to extract key text content like paragraphs, headings, and tables. You'll embed and store this data for retrieval with Weaviate. Finally, you'll craft effective retrieval prompts to pass to generative models. To cap this all off, you'll treat PDFs as images to allow you to capture context lost from images and plots. You'll use the ColPali multi-modal embedding model with a multi-modal generative model from OpenAI to begin having conversations with images and documents!

Working with the OpenAI API

Master RAG with Weaviate! Embed text and images for retrieval, and experiment with vector, BM25, and hybrid search.

Text generation with LLMs

End-to-End RAG with Weaviate

Instruções do exercício

Exercício interativo prático