写点什么

Autogen4j: the Java version of Microsoft AutoGen

作者:HamaWhite
  • 2023-12-23
    河南
  • 本文字数:2470 字

    阅读完需:约 8 分钟

https://github.com/HamaWhiteGG/autogen4j

Java version of Microsoft AutoGen, Enable Next-Gen Large Language Model Applications.

1. What is AutoGen

AutoGen is a framework that enables the development of LLM applications using multiple agents that can converse with each other to solve tasks. AutoGen agents are customizable, conversable, and seamlessly allow human participation. They can operate in various modes that employ combinations of LLMs, human inputs, and tools.


The following example in the autogen4j-example.

2. Quickstart

2.1 Maven Repository

Prerequisites for building:

  • Java 17 or later

  • Unix-like environment (we use Linux, Mac OS X)

  • Maven (we recommend version 3.8.6 and require at least 3.5.4)

<dependency>    <groupId>io.github.hamawhitegg</groupId>    <artifactId>autogen4j-core</artifactId>    <version>0.1.0</version></dependency>
复制代码

2.2 Environment Setup

Using Autogen4j requires OpenAI's APIs, you need to set the environment variable.


export OPENAI_API_KEY=xxx
复制代码

3. Multi-Agent Conversation Framework

Autogen enables the next-gen LLM applications with a generic multi-agent conversation framework. It offers customizable and conversable agents that integrate LLMs, tools, and humans.By automating chat among multiple capable agents, one can easily make them collectively perform tasks autonomously or with human feedback, including tasks that require using tools via code.


Features of this use case include:

  • Multi-agent conversations: AutoGen agents can communicate with each other to solve tasks. This allows for more complex and sophisticated applications than would be possible with a single LLM.

  • Customization: AutoGen agents can be customized to meet the specific needs of an application. This includes the ability to choose the LLMs to use, the types of human input to allow, and the tools to employ.

  • Human participation: AutoGen seamlessly allows human participation. This means that humans can provide input and feedback to the agents as needed.

3.1 Auto Feedback From Code Execution Example

Auto Feedback From Code Execution Example

// create an AssistantAgent named "assistant"var assistant = AssistantAgent.builder()        .name("assistant")        .build();
var codeExecutionConfig = CodeExecutionConfig.builder() .workDir("data/coding") .build();// create a UserProxyAgent instance named "user_proxy"var userProxy = UserProxyAgent.builder() .name("user_proxy") .humanInputMode(NEVER) .maxConsecutiveAutoReply(10) .isTerminationMsg(e -> e.getContent().strip().endsWith("TERMINATE")) .codeExecutionConfig(codeExecutionConfig) .build();
// the assistant receives a message from the user_proxy, which contains the task descriptionuserProxy.initiateChat(assistant, "What date is today? Compare the year-to-date gain for META and TESLA.");
// followup of the previous questionuserProxy.send(assistant, "Plot a chart of their stock price change YTD and save to stock_price_ytd.png.");
复制代码


The figure below shows an example conversation flow with Autogen4j.


After running, you can check the file coding_output.log for the output logs.

The final output is as shown in the following picture.



3.2 Group Chat Example

Group Chat Example

var codeExecutionConfig = CodeExecutionConfig.builder()        .workDir("data/group_chat")        .lastMessagesNumber(2)        .build();
// create a UserProxyAgent instance named "user_proxy"var userProxy = UserProxyAgent.builder() .name("user_proxy") .systemMessage("A human admin.") .humanInputMode(TERMINATE) .codeExecutionConfig(codeExecutionConfig) .build();
// create an AssistantAgent named "coder"var coder = AssistantAgent.builder() .name("coder") .build();
// create an AssistantAgent named "pm"var pm = AssistantAgent.builder() .name("product_manager") .systemMessage("Creative in software product ideas.") .build();
var groupChat = GroupChat.builder() .agents(List.of(userProxy, coder, pm)) .maxRound(12) .build();
// create an GroupChatManager named "manager"var manager = GroupChatManager.builder() .groupChat(groupChat) .build();
userProxy.initiateChat(manager, "Find a latest paper about gpt-4 on arxiv and find its potential applications in software.");
复制代码


After running, you can check the file group_chat_output.log for the output logs.

4. Run Test Cases from Source

git clone https://github.com/HamaWhiteGG/autogen4j.gitcd autogen4j
# export JAVA_HOME=JDK17_INSTALL_HOME && mvn clean testmvn clean test
复制代码


This project uses Spotless to format the code.

If you make any modifications, please remember to format the code using the following command.


# export JAVA_HOME=JDK17_INSTALL_HOME && mvn spotless:applymvn spotless:apply
复制代码

5. Support

Don’t hesitate to ask!

Open an issue if you find a bug or need any help.

用户头像

HamaWhite

关注

还未添加个人签名 2018-01-13 加入

专注分享Flink、Hadoop、Spark等大数据及AI技术,爱好原创

评论

发布
暂无评论
Autogen4j: the Java version of Microsoft AutoGen_agent_HamaWhite_InfoQ写作社区