app.py

import gradio as gr

from transformers import pipeline

s2t=gr.Interface.load('huggingface/facebook/s2t-medium-librispeech-asr')


def out(audio1,audio2):
  if (audio1==None) and (audio2==None):
    return "no audio","no audio"
  elif audio1==None:

    x=s2t(audio2)
    return x, grammar(x)
  else:
    x=s2t(audio1)
    return x, grammar(x)

iface = gr.Interface(
  fn=out, 
  title="DIFF-SVC Cloud Rendering",
  description="",
  inputs=[gr.inputs.Audio(source="upload", type="filepath", label=None, optional=True)],                                           
)

iface.launch(enable_queue=True,show_error=True)