CREESTL / RAGChatbot

Different variations of questions answering over a document using RAG

Overview

This repository contains examples of RAG systems for answering questions about local .pdf files

Functionality

rag_local_retrieval.py

• Load, embed and store .pdf file in vector DB
• GPT4All local LLM based on mistral-7b model
• Custom prompt
• RetrievalQA chain

rag_api_hf.py

• Load, embed and store .pdf file in vector DB
• Remote HuggingFace LLM based on zephyr-7b model. Access via API
• Custom prompt
• RetrievalQA chain

rag_api_openai_chat.py

• Load, embed and store .pdf file in vector DB
• Remote OpenAI LLM based on GPT3.5-Turbo model. Access via paid API
• Chat memory. Enhanced conversational capabilities compared to chains above
• Custom prompt
• ConversationalRetrievalChain chain