app.py

from huggingface_hub import InferenceClient

import gradio as gr
import random

client = InferenceClient("Qwen/Qwen2.5-72B-Instruct")

def respond(message, history):
    
    messages = [
        {"role": "system",
        "content": "You are a chatbot that gives restaurants based on dietary restrictions and location, make sure these are real restaurants"}
    ]

    if history:
        messages.extend(history)
        
    messages.append(
        {"role": "user",
        "content": message}
    )
    
    response = client.chat_completion(
        messages, max_tokens = 500, temperature = 1.7, top_p = .3
    )

    return response ["choices"][0]["message"]["content"].strip()
    
chatbot = gr.ChatInterface(respond, type="messages")
chatbot.launch()