app.py

import gradio as gr

import tensorflow as tf
from transformers import GPT2LMHeadModel, GPT2Tokenizer

global tokenizer, model, script_speaker_name, script_responder_name, convo
tokenizer = GPT2Tokenizer.from_pretrained("ethzanalytics/ai-msgbot-gpt2-XL-dialogue")
model = GPT2LMHeadModel.from_pretrained("ethzanalytics/ai-msgbot-gpt2-XL-dialogue", pad_token_id=tokenizer.eos_token_id)

script_speaker_name = "person alpha"

script_responder_name = "person beta"
convo = ''
def output(prompt):
    sentence = convo + '\n' + script_speaker_name + ': ' + prompt + '\n' + script_responder_name + ': '
    input_ids = tokenizer.encode(sentence, return_tensors='pt')
    # generate text until the output length (which includes the context length) reaches 50
    output = model.generate(input_ids, max_length=50, num_beams=5, no_repeat_ngram_size=2, early_stopping=True)
    convo = tokenizer.decode(output[0], skip_special_tokens=True))

iface = gr.Interface(fn=output, inputs="text", outputs="text")
iface.launch()