Advanced RAG Techniques
A community-maintained collection of more than 42 runnable notebooks, each demonstrating a specific technique for building retrieval-augmented generation systems, from foundational to advanced.
RAG_Techniques is a community-driven collection of more than 42 runnable notebooks, each covering a specific technique for building retrieval-augmented generation (RAG) systems, ranging from foundational to advanced approaches.
Each notebook pairs an explanation of the technique's intuition with working code and references, aimed at helping developers build more accurate, context-rich retrieval systems. Techniques draw on ideas from Plan-and-Solve and RAG research.
The material is distributed as Jupyter notebooks in Python and touches related tools such as LangChain and LlamaIndex.