Using RAG to create external memory for LLM

This project enhances an LLM with external memory using RAG techniques. It processes data (txt, pdf, pptx, docx) into a Qdrant vector database, retrieves context for prompts, and interacts with the meta/llama3-70b-instruct model via a FastAPI backend. A Streamlit web app provides a simple interface for queries and responses. Built with Poetry, it extends LLM capabilities with external knowledge.