app.py

import gradio as gr
import random
from huggingface_hub import InferenceClient

client = InferenceClient("Qwen/Qwen2.5-7B-Instruct-1M")

def respond(message, history):
    messages = [{"role": "system", "content": "You are a friendly chatbot."}]    
    if history:
        messages.extend(history)
        messages.append({"role": "user", "content": message})
    response = client.chat_completion(messages, max_tokens = 100)
    return response["choices"][0]["message"]["content"].strip()

def echo(message, history):
    return message
    
def yes_no(message, history):
    responses = ["Yes", "No"]
    return random.choice(responses)
    
chatbot = gr.ChatInterface(respond, type="messages")
chatbot.launch()