BastinFlorian / LLMs

LLMs use cases and studies

LLMs

This repository contains LLMs use cases and studies

Prerequisites

• Create a virtual environnement:

  • python3.10 -m venv .venv
  • source .venv/bin/activate
  • pip install -r requirements.txt

• Copy the env.template and fill your environment variables

  • cp .env.template .env

Use cases

I. Help Desk from Confluence data

Help desk allows you to create a Question Answering bot in streamlit using your company Confluence data.

Code here

Documentation here

To run the streamlit app run:

cd use_cases/confluence_help_desk
streamlit run streamlit.py

II. Smartphone advisor

Smartphone advisor is a LLM Chain Chat Bot that answer technical questions about tech products.

It uses a Chroma vector store from Youtube video extracted texts.

It follows the course by Andrew NG and Harrison Chase, Langchain CEO.

Code here

Documentation here

You will learn with this use case the following concepts:

• Character Text Splitter vs Recursive Character Text Splitter
• Maximum Marginal Relevance vs Semantic Search
• Chroma DB vs MyScale
• Contextual Compression & Self Query Retriever
• Stuff, Refine, Map Reduce & Map Rerank chain types
• Retrieval QA Chain vs Conversational Retrieval Chain

To run the streamlit app run:

cd use_cases/confluence_help_desk
streamlit run 7_streamlit_app.py