app.py

import transformers
import gradio as gr
import torch

from transformers import GPT2LMHeadModel, GPT2Tokenizer

tokenizer = GPT2Tokenizer.from_pretrained('dennis-fast/DialoGPT-ElonMusk')
model = GPT2LMHeadModel.from_pretrained('dennis-fast/DialoGPT-ElonMusk')

def chat(message, token_response):
    token_message = tokenizer.encode(message + tokenizer.eos_token, return_tensors='pt')
    token_response = model.generate(token_message, max_length=1000, pad_token_id=tokenizer.eos_token_id)
    response = tokenizer.decode(token_response[:, token_message.shape[-1]:][0], skip_special_tokens=True)
    return response, token_response

gr.Interface(fn=predict,
             title="Chat with Elon",
             theme="default",
             css=".footer {display:none !important}",
             allow_screenshot=False,
             allow_flagging='never'
             inputs=["User", "state"],
             outputs=["Elon", "state"]).launch()