app.py

import os
import gradio as gr
from openai import OpenAI


title = "ERNIE 4.5 Turbo: BAIDU's LLM"
description = """
- Official Website: <https://yiyan.baidu.com/> (UI in Chinese)
- API services: [Qianfan Large Model Platform](https://cloud.baidu.com/product-s/qianfan_home) (cloud platform providing LLM services, UI in Chinese)
- [ERNIE 4.5 Turbo Demo](https://huggingface.co/spaces/PaddlePaddle/ernie_4.5_turbo_demo) |  [ERNIE X1 Turbo Demo](https://huggingface.co/spaces/PaddlePaddle/ernie_x1_turbo_demo)
"""


qianfan_api_key = os.getenv("QIANFAN_TOKEN")
qianfan_model = "ernie-4.5-turbo-32k"

client = OpenAI(base_url="https://qianfan.baidubce.com/v2", api_key=qianfan_api_key)


def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
):
    messages = [{"role": "system", "content": system_message}]
    messages.extend(history)
    messages.append({"role": "user", "content": message})

    response = client.chat.completions.create(
        model=qianfan_model,
        messages=messages,
        max_completion_tokens=max_tokens,
        temperature=temperature,
        top_p=top_p,
        stream=True,
    )

    output_message = ""
    for chunk in response:
        token = chunk.choices[0].delta.content
        output_message += token
        yield output_message


demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(value="", label="System message"),
        gr.Slider(minimum=2, maximum=12288, value=2048, step=1, label="Max new tokens"),
        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Temperature"),
        gr.Slider(
            minimum=0.1,
            maximum=1.0,
            value=0.7,
            step=0.05,
            label="Top-p (nucleus sampling)",
        ),
    ],
    title=title,
    description=description,
    type='messages',
    concurrency_limit=50
)

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