app.py

import gradio as gr
import time

from transformers import pipeline
p = pipeline("automatic-speech-recognition")
# Define the transcription function
def transcribe(audio_path, state):
    try:
        # Transcribe the audio file
        result = speech_to_text_pipeline(audio_path)
        transcription = result['text']
    except Exception as e:
        transcription = f"An error occurred: {str(e)}"
    return transcription, state

# Create the Gradio interface
gr.Interface(
    fn=transcribe,
    inputs=[
        gr.Audio(type="filepath"),  # Removed 'live' from here
        "state"
    ],
    outputs=[
        "textbox",  # Display the transcribed text
        "state"
    ],
    live=True  # Specify 'live' interaction here
).launch()