Python LLM Q&A Bot with Long-Term Memory

This is a Jupyter Notebook showcasing a Large Language Model GPT chat bot with the ability to retain long-term memory by utilizing a vector database.

Tools and Frameworks Used In This Project

• Language: Python
• ChromaDB - An open-source vector database you can host on your machine.
  • You can also use other vector database like Pinecone, Milvus, and etc. But for simplicity's sake, I will use Chroma's transient storage to store the data.
• LangChain - A framework that makes it much easier to work with LLMs and vector database. It basically acts as a bridge between language models and vector database. It also provides many functionalities that help prompt generation and query processing.
• OpenAI - This will be our LLM bot that convert the data queried from the database into proper sentences.

pip install -q openai langchain python-dotenv tiktoken chromadb wikipedia