aniketmaurya / discord-llm-bot

Fun project: LLM powered RAG Discord Bot that works seamlessly on CPU
31 stars 4 forks source link

Discord-LLM-Bot

LLM powered Discord Bot that works seamlessly on CPU. Powered by LanceDB and Llama.cpp.

This Discord bot is designed to helps answer questions based on a knowledge base (vector db). By default, it contains knowledge of PyTorch Lightning and Lit-GPT.

This is a personal project created for educational purposes. It is not affiliated with or endorsed by the official Lightning AI team.

graph LR
    A((User Query)) --> B((Convert to Embedding))
    B --> C((Find Similar Document<br>from Vector Database))
    C --> D((Use Retrieved Document<br>as Context to Answer Question<br>using Mistral 7B LLM))

Getting Started

To get started with Discord-LLM-Bot, follow these steps:

  1. Invite the Bot to Your Server:

    • Create a Discord bot and install to your server. Learn more
    • Make sure you have the necessary permissions to add the bot to your server.
  2. Run the bot:

    • Add bot token to environment variable as LEARNER_BOT_TOKEN
    • Run the Python script: python src/app.py
  3. Usage:

    • Once the bot is in your server, you can start asking questions by mentioning the bot (e.g., @discord-llm-bot How does Lit-GPT work?).
    • The bot will reply with helpful information and explanations.

Example Usage

Here are some examples of how you can use Discord-LLM-Bot:

Contributing

If you want to contribute to this project or have suggestions for new features, feel free to open an issue or create a pull request. All contributions are welcome from the community.

License

This project is licensed under the MIT License. See the LICENSE file for more details.

Enjoy using Discord-LLM-Bot for learning and exploring PyTorch Lightning and Lit-GPT!