app.py

import gradio as gr
import spaces

from transformers import pipeline

if torch.cuda.is_available():
  print("Running on GPU")
else:
  print("Running on CPU")

pipe = pipeline(task="question-answering", model="decodingchris/distilbert-base-uncased-finetuned-squad-v2")

@spaces.GPU
def get_answer(context, question):
  result = pipe(context=context, question=question)
  return result['answer']

demo = gr.Interface(fn=get_answer, inputs=["textarea", "text"], outputs=gr.Textbox(label="answer"))

demo.launch()