import gradio as gr
from transformers import PreTrainedTokenizerFast, BartForConditionalGeneration

model_name = 'ainize/kobart-news'

tokenizer = PreTrainedTokenizerFast.from_pretrained(model_name)
model = BartForConditionalGeneration.from_pretrained(model_name)

def summ(txt):
  input_ids = tokenizer.encode(txt, return_tensors="pt")
  summary_text_ids = model.generate(
    input_ids=input_ids,
    bos_token_id=model.config.bos_token_id,
    eos_token_id=model.config.eos_token_id,
    length_penalty=2.0,
    max_length=142,
    min_length=56,
    num_beams=4,
  )
  return tokenizer.decode(summary_text_ids[0], skip_special_tokens=True)

interface = gr.Interface(summ, [gr.Textbox(label = 'original text')],
                        [gr.Textbox(label = 'summary')])

interface.launch(share = True)