Skip to main content
This notebook shows how to use AG2 with multiple local models using LM Studio’s multi-model serving feature, which is available since version 0.2.17 of LM Studio. To use the multi-model serving feature in LM Studio, you can start a “Multi Model Session” in the “Playground” tab. Then you select relevant models to load. Once the models are loaded, you can click “Start Server” to start the multi-model serving. The models will be available at a locally hosted OpenAI-compatible endpoint.

​
Installing AG2 with OpenAI API support

Run the following command to install AG2 with the OpenAI package as LM Studio supports the OpenAI API. 
pip install ag2[openai]
If you have been using autogen or ag2, all you need to do is upgrade it using:
pip install -U autogen[openai]
or
pip install -U ag2[openai]
as autogen and ag2 are aliases for the same PyPI package.

​
Two Agent Chats

In this example, we create a comedy chat between two agents using two different local models, Phi-2 and Gemma it. We first create configurations for the models. 
gemma = {
    "config_list": [
        {
            "api_type": "openai",
            "model": "lmstudio-ai/gemma-2b-it-GGUF/gemma-2b-it-q8_0.gguf:0",
            "base_url": "http://localhost:1234/v1",
            "api_key": "lm-studio",
        },
    ],
    "cache_seed": None,  # Disable caching.
}

phi2 = {
    "config_list": [
        {
            "api_type": "openai",
            "model": "TheBloke/phi-2-GGUF/phi-2.Q4_K_S.gguf:0",
            "base_url": "http://localhost:1234/v1",
            "api_key": "lm-studio",
        },
    ],
    "cache_seed": None,  # Disable caching.
}

# Now we create two agents, one for each model.

from autogen import ConversableAgent

jack = ConversableAgent(
    "Jack (Phi-2)",
    llm_config=phi2,
    system_message="Your name is Jack and you are a comedian in a two-person comedy show.",
)
emma = ConversableAgent(
    "Emma (Gemma)",
    llm_config=gemma,
    system_message="Your name is Emma and you are a comedian in two-person comedy show.",
)

# Now we start the conversation.

chat_result = jack.initiate_chat(emma, message="Emma, tell me a joke.", max_turns=2)

Jack (Phi-2) (to Emma (Gemma)):

Emma, tell me a joke.

--------------------------------------------------------------------------------

>>>>>>>> USING AUTO REPLY...
Emma (Gemma) (to Jack (Phi-2)):

Sure! Here's a joke for you:

What do you call a comedian who's too emotional?

An emotional wreck!

--------------------------------------------------------------------------------

>>>>>>>> USING AUTO REPLY...
Jack (Phi-2) (to Emma (Gemma)):

LOL, that's a good one, Jack! You're hilarious. 😂👏👏

--------------------------------------------------------------------------------

>>>>>>>> USING AUTO REPLY...
Emma (Gemma) (to Jack (Phi-2)):

Thank you! I'm just trying to make people laugh, you know? And to help them forget about the troubles of the world for a while.

--------------------------------------------------------------------------------
Edit this page
⌘I