import torch
import gradio as gr
import yt_dlp as youtube_dl
from transformers import pipeline
from transformers.pipelines.audio_utils import ffmpeg_read
import tempfile
import os
import time
# Model and setup
MODEL_NAME = "openai/whisper-large-v3"
BATCH_SIZE = 8
YT_LENGTH_LIMIT_S = 3600 # 1-hour limit for YouTube files
device = 0 if torch.cuda.is_available() else "cpu"
pipe = pipeline(
task="automatic-speech-recognition",
model=MODEL_NAME,
chunk_length_s=30,
device=device,
)
# Function to transcribe audio
def transcribe(inputs, task):
if inputs is None:
raise gr.Error("No audio file submitted! Please upload or record an audio file.")
text = pipe(inputs, batch_size=BATCH_SIZE, generate_kwargs={"task": task}, return_timestamps=True)["text"]
return text
# YouTube video processing functions
def _return_yt_html_embed(yt_url):
video_id = yt_url.split("?v=")[-1]
return f'
'
def download_yt_audio(yt_url, filename):
# [ ... existing code for download_yt_audio ... ]
def yt_transcribe(yt_url, task):
html_embed_str = _return_yt_html_embed(yt_url)
with tempfile.TemporaryDirectory() as tmpdirname:
filepath = os.path.join(tmpdirname, "video.mp4")
download_yt_audio(yt_url, filepath)
with open(filepath, "rb") as f:
inputs = f.read()
inputs = ffmpeg_read(inputs, pipe.feature_extractor.sampling_rate)
inputs = {"array": inputs, "sampling_rate": pipe.feature_extractor.sampling_rate}
text = pipe(inputs, batch_size=BATCH_SIZE, generate_kwargs={"task": task}, return_timestamps=True)["text"]
return html_embed_str, text
# Gradio interfaces
mf_transcribe = gr.Interface(
fn=transcribe,
inputs=[
gr.inputs.Audio(source="microphone", type="filepath", optional=True),
gr.inputs.Radio(["transcribe", "translate"], label="Task", default="transcribe"),
],
outputs="text",
layout="horizontal",
theme="huggingface",
title="Whisper Large V3: Transcribe Audio",
description="Transcribe long-form microphone or audio inputs with the click of a button!"
)
file_transcribe = gr.Interface(
fn=transcribe,
inputs=[
gr.inputs.Audio(source="upload", type="filepath", optional=True, label="Audio file"),
gr.inputs.Radio(["transcribe", "translate"], label="Task", default="transcribe"),
],
outputs="text",
layout="horizontal",
theme="NoCrypt/miku@1.2.1",
title="Whisper Large V3: Transcribe Audio",
description="Transcribe long-form microphone or audio inputs with the click of a button!"
)
yt_transcribe = gr.Interface(
fn=yt_transcribe,
inputs=[
gr.inputs.Textbox(lines=1, placeholder="Paste the URL to a YouTube video here", label="YouTube URL"),
gr.inputs.Radio(["transcribe", "translate"], label="Task", default="transcribe")
],
outputs=["html", "text"],
layout="horizontal",
theme="NoCrypt/miku@1.2.1",
title="Whisper Large V3: Transcribe YouTube",
description="Transcribe long-form YouTube videos with the click of a button!"
)
# Main Gradio application
with gr.Blocks(theme="NoCrypt/miku@1.2.1") as demo:
gr.HTML("