!pip install transformers
!pip install torch

from transformers import pipeline
import gradio as gr

modelo = pipeline("automatic-speech-recognition", model="facebook/wav2vec2-large-xlsr-53-spanish")

def transcribe(audio):
  text = modelo(audio)["text"]
  return text
    
interface = gr.Interface(
    fn=transcribe,
    inputs=gr.Audio(type="filepath"),
    outputs=["textbox"]
)
interface.launch()