|
|
import gradio as gr
|
|
|
from transformers import pipeline
|
|
|
import numpy as np
|
|
|
import librosa
|
|
|
import pandas as pd
|
|
|
|
|
|
|
|
|
MODEL_NAME = "openai/whisper-tiny"
|
|
|
BATCH_SIZE = 8
|
|
|
|
|
|
|
|
|
pipe = pipeline(
|
|
|
task="automatic-speech-recognition",
|
|
|
model=MODEL_NAME,
|
|
|
chunk_length_s=30,
|
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
|
|
def format_output_to_list(data):
|
|
|
formatted_list = "\n".join([f"{item['timestamp'][0]}s - {item['timestamp'][1]}s \t : {item['text']}" for item in data])
|
|
|
return formatted_list
|
|
|
|
|
|
def transcribe(inputs, task):
|
|
|
if inputs is None:
|
|
|
raise gr.Error("No audio file submitted! Please upload or record an audio file before submitting your request.")
|
|
|
|
|
|
output = pipe(inputs, batch_size=BATCH_SIZE, return_timestamps="word", generate_kwargs={"task": task})
|
|
|
text = output['text']
|
|
|
timestamps = format_output_to_list(output['chunks'])
|
|
|
return [text, timestamps]
|
|
|
|
|
|
examples = [
|
|
|
["arabic_english_audios/audios/arabic_audio_1.wav"],
|
|
|
["arabic_english_audios/audios/arabic_audio_2.wav"],
|
|
|
["arabic_english_audios/audios/arabic_audio_3.wav"],
|
|
|
["arabic_english_audios/audios/arabic_audio_4.wav"],
|
|
|
["arabic_english_audios/audios/arabic_hate_audio_1.mp3"],
|
|
|
["arabic_english_audios/audios/arabic_hate_audio_2.mp3"],
|
|
|
["arabic_english_audios/audios/arabic_hate_audio_3.mp3"],
|
|
|
["arabic_english_audios/audios/english_audio_1.wav"],
|
|
|
["arabic_english_audios/audios/english_audio_2.mp3"],
|
|
|
["arabic_english_audios/audios/english_audio_3.mp3"],
|
|
|
["arabic_english_audios/audios/english_audio_4.mp3"],
|
|
|
["arabic_english_audios/audios/english_audio_5.mp3"],
|
|
|
["arabic_english_audios/audios/english_audio_6.wav"]
|
|
|
]
|
|
|
|
|
|
with gr.Blocks(theme=gr.themes.Default()) as demo:
|
|
|
gr.HTML("<h1 style='text-align: center;'>Transcribe Audio with Timestamps using whisper-large-v3</h1>")
|
|
|
gr.Markdown("")
|
|
|
with gr.Row():
|
|
|
with gr.Column():
|
|
|
audio_input = gr.Audio(sources=["upload", 'microphone'], type="filepath", label="Audio file")
|
|
|
task = gr.Radio(["transcribe", "translate"], label="Task")
|
|
|
with gr.Row():
|
|
|
clear_button = gr.ClearButton(value="Clear")
|
|
|
submit_button = gr.Button("Submit", variant="primary", )
|
|
|
|
|
|
with gr.Column():
|
|
|
transcript_output = gr.Text(label="Transcript")
|
|
|
timestamp_output = gr.Text(label="Timestamp")
|
|
|
|
|
|
examples = gr.Examples(examples, inputs=audio_input, outputs=[transcript_output, timestamp_output], fn=transcribe, examples_per_page=20)
|
|
|
|
|
|
submit_button.click(fn=transcribe, inputs=audio_input, outputs=[transcript_output, timestamp_output])
|
|
|
clear_button.add([audio_input, transcript_output, timestamp_output])
|
|
|
|
|
|
|
|
|
if __name__ == "__main__":
|
|
|
demo.launch()
|
|
|
|
