app.py

import gradio as gr
import whisper

# Initialize the Whisper model
model = whisper.load_model("large")

def transcribe(audio_file):
    # Load audio file
    audio = whisper.load_audio(audio_file.name)
    # Pad or trim audio to the expected length
    audio = whisper.pad_or_trim(audio)
    # Generate a mel spectrogram
    mel = whisper.log_mel_spectrogram(audio).to(model.device)
    # Options for decoding the spectrogram
    options = whisper.DecodingOptions()
    # Perform the transcription
    result = whisper.decode(model, mel, options)
    return result.text

# Create the Gradio interface
iface = gr.Interface(
    fn=transcribe,
    inputs=gr.Audio(source="upload", type="file", label="Upload your audio file"),
    outputs="text",
    title="Whisper ASR",
    description="Upload an audio file and it will be transcribed using OpenAI's Whisper model."
)

# Launch the app
if __name__ == "__main__":
    iface.launch()