This project implements a multi-agent system powered by RAG (Retrieval-Augmented Generation) for collaborative reasoning and evaluation.

📦 Project Structure

├── rag_system_build.py # Build and initialize the RAG knowledge base

├── agent_*.py # Individual agent scripts (API and URL need to be configured)

├── evaluation.py # Evaluation script for the final results

└── README.md

🚀 Usage

1️⃣ Build the RAG System

Run the following script to construct the RAG knowledge base and initialize the retriever: ```bash python rag_system_build.py ```

2️⃣ Configure Agents

Open each `agent_xxx.py` file and fill in your own API key, base URL, or any other required parameters: ```python API_KEY = "your_api_key_here" BASE_URL = "https://your-model-url.com" ```

3️⃣ Run the Agents

Execute the agents one by one to complete their reasoning tasks:

4️⃣ Evaluate Results

Finally, run the evaluation script to assess the outputs from all agents: ```bash python evaluation.py ```