This repository provides a collection of Python notebooks demonstrating advanced Retrieval-Augmented Generation (RAG) techniques. It targets developers and researchers looking to enhance Large Language Models (LLMs) with external knowledge, offering practical implementations using Langchain, OpenAI GPTs, and Meta Llama 3. The notebooks cover a spectrum of RAG strategies, from basic query flow to complex agentic and self-correcting approaches, enabling more informed and accurate LLM outputs.

How It Works

The project leverages the Langchain framework to orchestrate RAG pipelines. It explores various components including query transformation, data source routing, diverse indexing methods for vector databases, and advanced retrieval mechanisms like reranking and RAG Fusion. The notebooks detail sophisticated techniques such as Multi-Query Retriever, Self-Reflection-RAG, and Agentic RAG variants (Adaptive, Corrective), culminating in a local Llama 3 8B agentic RAG implementation. This modular approach allows users to build and experiment with increasingly complex and context-aware RAG systems.

Quick Start & Requirements

• Install: pip install -r requirements.txt (specific commands for running notebooks are within the .ipynb files).
• Prerequisites: Python 3.x, Langchain, OpenAI API key, potentially Llama 3 models (local setup details provided in 10_LLAMA_3_Rag_Agent_Local.ipynb).
• Resources: Notebooks require standard Python environments; local Llama 3 execution will necessitate appropriate hardware.
• Links: Notebooks are self-contained; no external quick-start or demo links are provided.

Highlighted Details

• Comprehensive coverage of RAG techniques from basic to advanced agentic flows.
• Practical implementation examples using Langchain.
• Includes notebooks for Multi-Query Retriever, Self-Reflection-RAG, and various Agentic RAG patterns.
• Demonstrates local execution with Llama 3 8B.

Maintenance & Community

• The repository is maintained by NisaarAgharia.
• No specific community channels (Discord, Slack) or roadmap links are provided in the README.

Licensing & Compatibility

• The repository does not explicitly state a license.
• Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

The repository lacks explicit licensing information, which may impact commercial adoption. There are no external links for quick starts or demos, requiring users to navigate the notebooks directly for setup and execution guidance.