app.py

import os
from huggingface_hub import 
InferenceClient
client = InferenceClient(
    model="Qwen/Qwen2.5-7B-Instruct",
    token=os.getenv("HF_TOKEN")
)
def respond(message, history):
    messages = [{"role": "system", "content": "you are a friendly chatbot"}]


    if history:
        messages.extend(history)

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

        
        response = client.chat_completion (
            messages,
            max_tokens=100
        )
    return random.choices[0].messages.content.strip()

    chatbot = gr.ChatInerfae(respond)
    chatbot.launh()