run_demo.py

from transformers import pipeline
import gradio as gr

pipe = pipeline(model="bhuang/wav2vec2-xls-r-1b-cv9-fr")


def transcribe(audio):
    # text = pipe(audio, chunk_length_s=30, stride_length_s=5)["text"]
    text = pipe(audio)["text"]
    return text


iface = gr.Interface(
    fn=transcribe,
    inputs=gr.Audio(source="microphone", type="filepath", label="Record something..."),
    outputs="text",
    title="Speech-to-Text in French",
    description="Realtime demo for French automatic speech recognition.",
    allow_flagging="never",
)

iface.launch()