app.py

import gradio as gr
import random
from huggingface_hub import InferenceClient

client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")

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 = ""
    
    for message in client.chat_completion(
        messages,
        max_tokens=500,
        temperature=0.2,
        top_p=0.9,
        stream=True
    ):
        token = message.choices[0].delta.content
        response += token
        yield response
    
    
    return response['choices'][0]['message']['content'].strip()


def yes_no(message, history):
    response = ['Yes', 'No', '100% yes', 'I think not', 'Possibly', 'Never']
    return random.choice(response)

def echo(message, history):
	return message

print("Hello, world!")

chatbot = gr.ChatInterface(yes_no, type="messages", title="Magic Eight Ball",
                           description="This magic eight ball answers your questions!",
                           theme=gr.themes.Soft())

chatbot.launch()