使用
####


.. figure:: https://img.zhaoweiguo.com/uPic/2024/04/X2d8ra.png

    OpenAI 的 Lilian Weng 将以 LLM 为驱动的 AI Agent，形式化为如下的公式：AI Agent = LLM（大模型） + Planning（规划） + Memory（记忆） + Tools（工具）



Al Agent 的开源项目
===================

* BabyAGI: https://github.com/yoheinakajima/babyagi
* MetaGPT: 
* GPT Researcher: 
* AutoGPT: 
* A1-16Z的 Al-town: 
* 斯坦福的 Al-Town: 
* ChatDev: 




BabyAGI
-------

* Task-driven Autonomous Agent: https://twitter.com/yoheinakajima/status/1640934493489070080?s=20

.. figure:: https://img.zhaoweiguo.com/uPic/2024/04/EMBb4O.png

    Task-driven Autonomous Agent: An agent that leverages @openai's GPT-4, @pinecone vector search, and @LangChainAI framework to autonomously create and perform tasks based on an objective.