tommanzur/autogen_groupchat_rag

This project demonstrates a group chat system powered by Retrieval Augmented Generation (RAG), utilizing the `autogen` library. It showcases how conversational agents, powered by llms, tools, or human inputs, can perform tasks collectively through automated chat.

Group Chat with Retrieval Augmented Generation (RAG)

Overview

This project demonstrates a group chat system powered by Retrieval Augmented Generation (RAG), utilizing the autogen library. It showcases how conversational agents, powered by large language models, tools, or human inputs, can perform tasks collectively through automated chat. The framework facilitates tool use and human participation in multi-agent conversations, offering a dynamic and flexible approach to interactive tasks.

Features

• Multi-agent conversation system with support for automated agents and human input.
• Integration of retrieval-augmented generation for enriched conversation responses.
• Flexible agent configuration, including support for custom termination conditions and auto-replies.
• Extensible architecture for adding new agents or modifying existing ones.

Installation

Before you can run this project, you need to install some dependencies. You can install the required packages via pip:

pip install pyautogen

Configuration

Set up your API endpoint by using the config_list_from_json function. This function loads a list of configurations from an environment variable or a json file. Configure your agents and their interactions within the group chat system. Refer to the notebook for detailed examples on setting up agents like "Boss" and "Boss_Assistant".

Usage

To start using this group chat system, follow the instructions laid out in the notebook:

• Configure the agents and their roles within the conversation.
• Initialize the group chat and define the conversation logic.
• Start the chat and interact with the system through the predefined agents.