Intelligent Q&A System Based on LangChain, FAISS, and LLM

Project Overview

This project implements an intelligent Q&A system based on LangChain, FAISS, and LLM. The system's features include:

• File Search: Supports file uploads, vectorizes the files, and performs semantic search using FAISS to find relevant content.
• LLM-based Q&A: The system utilizes LLM to generate answers based on the retrieved information and context.

Show program

Below are some screenshots demonstrating how the project looks when running:

Installation

Clone the repository

git clone git@github.com:first-coding/langchain-project.git

Install required dependencies

pip install -r requirements.txt

PS: If the installation is slow, you can try using a mirror.

Run the project

python main.py

Notes

1.Due to limited computational resources, I am using a smaller model and running it on a CPU.

2.Remember to initialize the system when uploading files, otherwise errors may occur.

3.Users with inconsistent Python versions may encounter version-related errors.

4.If you want to run the system on a GPU, modify main.py, cls/langchain_application.py, and gpt_service.py, and install the corresponding CUDA version.

Suggestions and Issues

I hope this project is helpful to everyone，certainly，This project is a reproduction and modification based on the LangChain code from GitHub. If you have any suggestions or issues, feel free to discuss them with me through the issues section. Thank you!