app.py

import gradio as gr 
from transformers import pipeline
# from huggingsound import SpeechRecognitionModel
from scipy.io.wavfile import write
from huggingsound import SpeechRecognitionModel

model = SpeechRecognitionModel("jonatasgrosman/wav2vec2-large-xlsr-53-chinese-zh-cn")
# audio_paths = ["/path/to/file.mp3", "/path/to/another_file.wav"]

# transcriptions = model.transcribe(audio_paths)

def func(input_audio):
    print(input_audio )
    write('./audio.wav', input_audio[0], input_audio[1])
    out_txt = model.transcribe(['./audio.wav'])
    return out_txt[0]['transcription']

inp =[gr.Audio(source='microphone')]
out = [gr.Textbox()]
demo = gr.Interface(func,
            #  gr.Audio(source='microphone'), 
            inputs = inp,
            outputs = out)
            #  examples=examples)

demo.launch()