app.py

import gradio as gr
from transformers import pipeline
import torchaudio
import time

# Load Whisper ASR model
transcriber = pipeline(model="openai/whisper-base")

# Load summarization model
summarization_model = pipeline("summarization")

def translate_audio(audio):
    # Step 1: Transcribe audio to text
    transcription = transcriber(audio)
    print('transcription', transcription)

    # Step 2: Translate text to Hindi
    summary = summarization_model(transcription['text'])
    print('summary', summary)

    return transcription['text'], summary[0]['summary_text']


# Create Gradio interface
with gr.Blocks() as iface:
    gr.Markdown("# Audio Translator, Summarizer")

    with gr.Row():
        audio_input = gr.Audio(type="filepath", label="Upload Audio")
        transcription_output = gr.Textbox(
            label="Transcribed Text", 
            info="Initial text")
        translation_output = gr.Textbox(
            label="Summary", 
            info="Meeting minute")

    translate_button = gr.Button("Translate Audio")

    translate_button.click(
        translate_audio,
        inputs=[audio_input],
        outputs=[transcription_output, translation_output]
    )

# Launch the app
iface.launch(share=True)  # 'share=True' to get a public link