app.py

import os
import gradio as gr
from huggingface_hub import InferenceClient

# 安全读取 Huggingface Token
HF_TOKEN = os.getenv("HF_TOKEN")
MODEL_NAME = "Qwen/Qwen3-235B-A22B"

client = InferenceClient(model=MODEL_NAME, token=HF_TOKEN)

def chat_fn(message, history, system_message, temperature, top_p, max_tokens, repetition_penalty):
    # 构造消息
    messages = []
    if system_message:
        messages.append({"role": "system", "content": system_message})
    for user_msg, bot_msg in history:
        messages.append({"role": "user", "content": user_msg})
        messages.append({"role": "assistant", "content": bot_msg})
    messages.append({"role": "user", "content": message})

    # 请求模型
    response = client.chat_completion(
        messages=messages,
        temperature=temperature,
        top_p=top_p,
        max_tokens=max_tokens,
        repetition_penalty=repetition_penalty,
    )

    answer = response.choices[0].message["content"]
    history.append((message, answer))
    return history, history

with gr.Blocks(theme=gr.themes.Base(), css="footer {display: none !important}") as demo:
    gr.Markdown("# 🤖 Qwen3 Chatbot")
    with gr.Row():
        with gr.Column(scale=3):
            chatbot = gr.Chatbot()
            message = gr.Textbox(label="Your Message", placeholder="Type something...", scale=4)
            submit = gr.Button("Send", variant="primary")
            clear = gr.Button("Clear")

            state = gr.State([])
        with gr.Column(scale=1):
            system_message = gr.Textbox(label="System Prompt", placeholder="You are a helpful assistant.")
            temperature = gr.Slider(0, 1, value=0.7, step=0.05, label="Temperature")
            top_p = gr.Slider(0.1, 1.0, value=0.9, step=0.05, label="Top-p")
            max_tokens = gr.Slider(64, 2048, value=512, step=64, label="Max Tokens")
            repetition_penalty = gr.Slider(1.0, 2.0, value=1.1, step=0.1, label="Repetition Penalty")

    submit.click(chat_fn, 
                 inputs=[message, state, system_message, temperature, top_p, max_tokens, repetition_penalty], 
                 outputs=[chatbot, state])
    message.submit(chat_fn, 
                   inputs=[message, state, system_message, temperature, top_p, max_tokens, repetition_penalty], 
                   outputs=[chatbot, state])
    clear.click(lambda: ([], []), None, [chatbot, state])

demo.launch()