/QA-generator-llm

Primary LanguageHTML

Question-Answer-Generation-App-using-Mistral-7B

Question Answer Generation App using Mistral 7B, Langchain, and FastAPI.

https://www.loom.com/share/f9a5a625d79b4f7381a54126e0b740e2?sid=98f1a0dc-f92e-4d8b-9d09-5ece38302ce7

Project Name

Description

This project is a Question-Answer Pair Generator with Zephyr-7B, which is a tool for generating questions and answers based on study materials. It utilizes the LangChain library for natural language processing tasks.

Note to be considered

This application is developed in fast prototyping and so the frontend is first written in HTML CSS and JS then migrated to React and TSX also the backend prompt is tested on openai and then migrated to the LLM and using of LLM on free cloud platform is not easy because it gives a lot of resource limitation and doest't perform aslo there is use of streamlit for now which makes it hard to deploy it in any cloud platform .

Installation Instructions

Frontend

  1. Navigate to the frontend directory in your terminal.
  2. Run npm install to install the necessary dependencies.
  3. Run npm start to start the development server.
  4. Go to http://localhost:8000/ to see you app running.

Backend

  1. Navigate to the backend directory in your terminal.

  2. Ensure you have Python installed on your system.

  3. Run pip install -r requirements.txt to install the required Python dependencies.

  4. Download the AI model from here and place it in the backend directory.

  5. Run python app.py to start the backend server.

  6. In case you want to test only the prompt through Open AI API run python test_openai_app.py start backend server of openai version of the app.

  7. backend server run at http://localhost:8000.

Usage

  1. After starting both the frontend and backend servers, visit http://localhost:3000 in your web browser.
  2. Upload a PDF file containing study material using the provided file upload input.
  3. Click on the "Upload" button to analyze the uploaded file.
  4. Once the analysis is complete, you can view and download the generated question-answer pairs.

Tech Stack

  • Python
  • FastAPI
  • Mistral-7B AI Model
  • React
  • TypeScript
  • Langchain
  • FAISS(similar search)
  • streamlit
  • ChromaDB(vector db)
  • Docker

Credits

This project utilizes the LangChain library for natural language processing tasks.