app.py

import torch
import torchaudio
from transformers import pipeline
import streamlit as st

# Load the pre-trained wav2vec 2.0 model
model = pipeline('voice-recognition', model='facebook/wav2vec2-base-960h')

def recognize_voice(audio_file):
    waveform, sample_rate = torchaudio.load(audio_file)
    transcript = model(waveform.numpy()[0], sample_rate=sample_rate)
    return transcript[0]['sentence']

def main():
    st.title("Voice Recognition App")

    # Upload audio file
    audio_file = st.file_uploader("Upload an audio file", type=["mp3", "wav"])

    if audio_file is not None:
        st.audio(audio_file, format='audio/wav')

        if st.button("Recognize Voice"):
            try:
                # Perform voice recognition
                result = recognize_voice(audio_file)
                st.success(f"Recognition Result: {result}")
            except Exception as e:
                st.error(f"Error: {e}")

if __name__ == "__main__":
    main()