Chat Models

Overview

Chat models are the engine that power all invokable classes. When a user invokes an Invokable, the task is passed through to the underlying chat model. The chat model executes the task according to the Invokable's persona, history, and configuration, and then passes the result back to the Invokable.

Attributes

BHChatModels are fairly simple — they are meant to be a light wrapper around the underling LLM provider's client:

Attribute type	Description
model str	LLM model, e.g., gpt-3.5-turbo.
**model_config	Additional keyword arguments accepted by the LLM provider's client.

Chat models can be used via the chat or stream methods:

class BHChatModel(BaseModel):
    ...
    def chat(
        self,
        task_message: BHMessage | None,
        temperature: int,
        tools: dict[str, BHTool],
        conversation: list[BHMessage | BHToolMessage],
    ) -> list[BHMessage | BHToolMessage]:
        ...

    def stream(
        self,
        task_message: BHMessage | None,
        temperature: int,
        tools: dict[str, BHTool],
        conversation: list[BHMessage | BHToolMessage],
        printer: Optional["Printer"] = None,
    ) -> list[BHMessage | BHToolMessage]:
        ...

These methods have nearly identical definitions. The stream method has one additional argument printer:

Argument	Description
task_message	User message containing the task. If this message does not have a 'user' role, Beehive will throw an error.
temperature	Temperature setting for underlying LLM.
tools	Any tools to prove the LLM. Note that the language model must support tool calling in order to properly make use of tools.
conversation	List of messages that the LLM should treat as context for the current task.
printer stream only	output.printer.Printer instance. Used to prettify the streamed output.

The supported chat models are listed below.

OpenAI

Beehive's OpenAIModel class provides a thin wrapper over the openai.OpenAI Python client:

from beehive.message import BHMessage, MessageRole
from beehive.models.openai_model import OpenAIModel

openai_chat_model = OpenAIModel(
    model="gpt-3.5-turbo",
    api_key="<your_api_key">,  # keyword argument accepted by openai.OpenAI client
    max_retries=10,  # keyword argument accepted by openai.OpenAI client
)
joke = openai_chat_model.chat(
    input_messages=[
        BHMessage(
            role=MessageRole.USER,
            content="Tell me a joke!"
        )
    ],
    temperature=0,
    tools={},
    conversation=[],
)
print(joke)

This will print the following:

[BHMessage(role=<MessageRole.ASSISTANT: 'assistant'>, content="Why couldn't the bicycle stand up by itself?\n\nBecause it was two tired!", tool_calls=[])]

Warning

You can certainly chat with BHChatModels directly. However, doing so require that you use Beehive-native types. We highly recommend using Invokables instead — these are simpler to work with "out-of-the-box".