app.py

import gradio as gr
from transformers import AutoModelForCausalLM, AutoTokenizer

# Initialize the DialoGPT model and tokenizer
model = AutoModelForCausalLM.from_pretrained("microsoft/DialoGPT-medium")
tokenizer = AutoTokenizer.from_pretrained("gpt2")

chat_history = None

def chat(message):
    global chat_history
    
    # Encode the user's message with the GPT-2 tokenizer
    input_ids = tokenizer.encode(message + tokenizer.eos_token, return_tensors="pt")
    
    # Generate a response from DialoGPT-medium
    response_ids = model.generate(input_ids, max_length=150, pad_token_id=tokenizer.eos_token_id, num_return_sequences=1)
    
    # Decode and return the bot's response
    bot_response = tokenizer.decode(response_ids[0], skip_special_tokens=True)
    
    chat_history = bot_response  # Store the bot's response for reference
    
    return bot_response

# Create and launch the Gradio interface
iface = gr.Interface(
    fn=chat,
    title="UrFriendly Chatbot",
    description="UrFriendly Chatbot is a conversational assistant based on DialoGPT-medium with GPT-2 tokenization. Type or click on one of the examples to get started. Please note that UrFriendly Chatbot is not 100% accurate, so incorrect information may generate. 💬🤗",
    examples=[
        "Howdy!",
        "Tell me a joke.", 
        "Explain quantum computing in simple terms.", 
        "How are you?", 
        "What is an exponent in mathematics?",
        "Does money buy happiness?"
    ],
    inputs="text",
    outputs="text",
    live=True  # Set to True to allow continuous conversation
)

iface.launch()