Financial Bot - Retrieval-based QA Chatbot

Welcome to the Financial Bot project! This project demonstrates the setup of a retrieval-based question-answering (QA) chatbot that uses the langchain library for handling interactions and retrieval.

Features

• Provides answers to user queries based on a database of financial information.
• Uses a custom prompt template to structure the QA interactions.
• Leverages langchain for language modeling and retrieval.
• Uses chainlit for managing the chatbot interactions.

Getting Started

System Requirements

• The llama-2-7b-chat.ggmlv3.q4_0.bin model can be run on systems with 8GB of RAM.

Installation

1. Clone this repository: git clone https://github.com/yourusername/financial-bot.git
2. Install the required libraries: pip install langchain chainlit

Requirements

To install the required dependencies, you can use the provided requirements.txt file. Simply run:

pip install -r requirements.txt

Usage

1. Configure the DB_faiss_path variable to point to your local FAISS database.
2. Customize the prompt template in the set_custom_prompt function if needed.
3. Run the chatbot using chainlit run app.py -w.