I’m looking on gradio examples of ASR (https://gradio.app/real-time-speech-recognition/) ,and how they handle the input from the microphone.
I run this example:
import gradio as gr import time def transcribe(audio): time.sleep(5) return audio inf = gr.Interface( fn=transcribe, inputs=[ gr.Audio(source="microphone", type="filepath", streaming=True) ], outputs=[ "textbox" ], live=True) if __name__ == "__main__": inf.launch(share=True, debug=False)
I run the app from my PC and connect from another PC.
It seems that the microphone is opened on the server and not on client, because all the wav files are quiet.
What do I need to change in order to run this app and open the microphone on the client side ?