app.py

from subprocess import Popen
from transformers import AutoProcessor, AutoModelForPreTraining

processor = AutoProcessor.from_pretrained("patrickvonplaten/mms-1b")

model = AutoModelForPreTraining.from_pretrained("patrickvonplaten/mms-1b")

def transcribe(audio):
    Popen('python mms_infer.py --model model --lang eng --audio audio')

gr.Interface(
    title = 'MetaAI (Facebook Research) MMS (Massively Multilingual Speech) ASR', 
    fn=transcribe, 
    inputs=[
        gr.inputs.Audio(source="microphone", type="filepath")
    ],
    outputs=[
        "textbox"
    ],
    live=True).launch()