RAG (Retrieval Augmented Generation)

It's actually a search problem.

LLM for Devs

Collection thumbnail

Recently added

Don't naive RAG do hybrid search instead (Pinecone Weaviate or pgvector + full text search & rerank)
Lesson 1

Don't naive RAG do hybrid search instead (Pinecone Weaviate or pgvector + full text search & rerank)

Improving search engine performance requires moving beyond simple similarity searches; hybrid approaches, combining techniques like BM25 with vector embeddings and careful pre-processing, are crucial for handling complex data and achieving superior results.

42mJan 29, 2025
Free

All lessons

Don't naive RAG do hybrid search instead (Pinecone Weaviate or pgvector + full text search & rerank)
Lesson 1

Don't naive RAG do hybrid search instead (Pinecone Weaviate or pgvector + full text search & rerank)

Improving search engine performance requires moving beyond simple similarity searches; hybrid approaches, combining techniques like BM25 with vector embeddings and careful pre-processing, are crucial for handling complex data and achieving superior results.

42mJan 29, 2025
Free