Ask PDF Bot

A Telegram bot that answers questions based on the content of uploaded PDF documents. The bot uses a Retrieval-Augmented Generation (RAG) pipeline powered by the mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated model from Hugging Face.

Features

• PDF Processing: Upload a PDF document, and the bot will process it to create a searchable vector store.
• Question Answering: Ask questions about the content of the uploaded PDF, and the bot will provide accurate and concise answers.
• RAG Pipeline: Combines a vector store (FAISS) with a custom LLM wrapper for context-aware responses.
• Telegram Integration: Easy-to-use bot interface for seamless interaction.

Prerequisites

Before running the bot, ensure you have the following:

1. Python 3.8+: The bot is written in Python.
2. Telegram Bot Token: Obtain a bot token from BotFather.
3. Hugging Face Model: The bot uses the mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated model. Ensure you have access to it on Hugging Face.

Installation

1. Clone the Repository:

   git clone https://github.com/yourusername/telegram-pdf-qa-bot.git
   cd telegram-pdf-qa-bot