app.py

import gradio as gr
import numpy as np
from faster_whisper import WhisperModel

# Load the Whisper model
model = WhisperModel("medium", device="cpu", compute_type="int8")

# Function to handle transcription
def transcribe(audio):
    # Check if the audio is in the correct format (numpy array)
    if isinstance(audio, tuple):
        audio = np.array(audio[1])
    
    # Transcribe the audio
    segments, info = model.transcribe(audio, beam_size=5)
    transcription = ""
    for segment in segments:
        transcription += f"[{segment.start:.2f}s -> {segment.end:.2f}s]: {segment.text}\n"
    return transcription

# Gradio interface for live transcription
audio_input = gr.Audio(type="numpy", label="Speak into the microphone for live transcription")
output_text = gr.Textbox(label="Transcription")

# Create a Gradio interface with live microphone input
demo = gr.Interface(fn=transcribe, inputs=audio_input, outputs=output_text, live=True)

# Launch the Gradio app
demo.launch()