app.py

""" Star Coder2 chat demo """

from typing import List, Tuple, Union
import gradio as gr
from huggingface_hub import InferenceClient


# HF InferenceClient
client = InferenceClient("HuggingFaceH4/starchat2-15b-v0.1")


def chat(
    message: str,
    history: List[Tuple[str, str]],
    system_message: str,
    max_tokens: Union[int, None],
    temperature: Union[float, None],
    top_p: Union[float, None],
):
    """Code assistant"""

    # Chat history
    messages = [{"role": "system", "content": system_message}]

    for val in history:
        if val[0]:
            messages.append({"role": "user", "content": val[0]})

        if val[1]:
            messages.append({"role": "assistant", "content": val[1]})

    # Add user message
    messages.append({"role": "user", "content": message})

    llm_message = client.chat_completion(
        messages,
        max_tokens=max_tokens,
        temperature=temperature,
        top_p=top_p,
    )

    # Add chatbot message
    messages.append(
        {
            "role": "assistant",
            "content": llm_message.choices[0].message.content,
        }
    )

    yield llm_message.choices[0].message.content


# UI
star_coder = gr.ChatInterface(
    chat,
    title="Star Coder2",
    theme="soft",
    description="Star Coder 2 is a code assistant trained on 15B parameters and 100K lines of code. "
    "It is a powerful code assistant that can help you write and debug code, as well as explain code to others.",
    additional_inputs=[
        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p"),
    ],
)


if __name__ == "__main__":
    star_coder.launch()