datawhalechina/all-in-rag

Covers comprehensive RAG system development from data ingestion through vector embeddings (including multimodal support via image-text retrieval), hybrid search strategies combining dense and sparse retrieval, and advanced techniques like Text2SQL and query rewriting. The curriculum integrates practical implementations using vector databases like Milvus, LLM-based generation, and system evaluation methodologies, culminating in production-ready projects demonstrating Graph RAG architectures and optimization patterns.

4,659 stars. Actively maintained with 2 commits in the last 30 days.