Geek Out Time: Play with AutoGen

Microsoft has open-sourced AutoGen (available at https://github.com/microsoft/autogen ), a framework that enables developers to build next-generation LLM applications. It achieves this by offering a high-level abstraction for multi-agent collaboration, teachability, and personalization.

I experimented with this framework to establish a straightforward group chat involving user proxy agents, a product manager, and a software engineer, aiming to create a simple page

First, start by installing AutoGen and configuring the OpenAI keys,

pip install pyautogen import autogen # config_list_gpt4 = autogen. config_list_from_json ( # "OAI_CONFIG_LIST" , # filter_dict={ # "model" : [ "gpt-4" , "gpt-4-0314" , "gpt4" , "gpt-4-32k" , "gpt-4-32k-0314" , "gpt-4-32k-v0314" ], # }, #) config_list_gpt35 = autogen. config_list_from_json ( "OAI_CONFIG_LIST" , filter_dict={ "model" : { "gpt-3.5-turbo" , "gpt-3.5-turbo-16k" , "gpt-3.5-turbo-0301" , "chatgpt-35-turbo-0301" , "gpt-35-turbo-v0301" , }, }, ) I cannot use the API key to access GPT-4; therefore, I have reverted to using gpt-3.5-turbo. The API key is configured in the OAI_CONFIG_LIST file.

This is an excerpt — the full article continues on Medium.