app.py

import gradio as gr
import torch
from transformers import pipeline

# Load the Whisper pipeline
transcriber = pipeline("automatic-speech-recognition", model="openai/whisper-base")  # Choose your Whisper size

def transcribe_audio(audio_file):
    if audio_file is not None:
        text = transcriber(audio_file)["text"]
        return text
    else:
        return "No audio file uploaded"


with gr.Blocks() as demo:
    gr.Markdown("## Audio Transcription with Whisper")
    audio_input = gr.Audio(type="filepath", label="Upload Audio File")
    text_output = gr.Textbox(label="Transcription")
    btn = gr.Button("Transcribe")
    btn.click(transcribe_audio, inputs=audio_input, outputs=text_output, return_timestamps=True)

demo.launch()