Spaces:
Sleeping
Sleeping
File size: 4,357 Bytes
fa1fe99 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 |
import torch
from transformers import pipeline
import os
import gradio as gr
from pydub import AudioSegment
from pytube import YouTube
import timeit
import math
device = "cuda" if torch.cuda.is_available() else "cpu"
pipe = pipeline("automatic-speech-recognition", model="distil-whisper/distil-medium.en", device=device)
def transcribe_speech_local(filepath):
if filepath is None:
return [{"error": "No audio found, please retry."}]
# Split audio into 15-second chunks
audio = AudioSegment.from_file(filepath)
chunk_length_ms = 15000 # 15 seconds in milliseconds
chunks = [audio[i:i + chunk_length_ms] for i in range(0, len(audio), chunk_length_ms)]
print(chunks)
aligned_chunks = []
transcription_time_total = 0
# Transcribe each chunk and measure time
for chunk_id, chunk in enumerate(chunks):
start_time = timeit.default_timer()
chunk.export("temp_chunk.wav", format="wav")
output = pipe("temp_chunk.wav")
transcription_time = timeit.default_timer() - start_time
transcription_time_total += transcription_time
# Calculate start and end times in seconds
start_time_sec = chunk_id * 15
end_time_sec = start_time_sec + len(chunk) / 1000.0
aligned_chunks.append({
"chunk_id": chunk_id,
"chunk_length": len(chunk) / 1000.0,
"text": output["text"],
"start_time": start_time_sec,
"end_time": end_time_sec,
"transcription_time": transcription_time
})
return aligned_chunks
def download_audio_from_youtube(youtube_url):
yt = YouTube(youtube_url)
stream = yt.streams.filter(only_audio=True).first()
output_path = stream.download()
base, ext = os.path.splitext(output_path)
audio_file = base + '.mp3'
os.rename(output_path, audio_file)
return audio_file
def transcribe_speech_from_youtube(youtube_url):
audio_filepath = download_audio_from_youtube(youtube_url)
# Convert to WAV format with 16kHz sample rate if necessary
audio = AudioSegment.from_file(audio_filepath)
audio = audio.set_frame_rate(16000).set_channels(1)
audio.export("converted_audio.wav", format="wav")
audio = AudioSegment.from_file("converted_audio.wav")
# Split audio into 15-second chunks
chunk_length_ms = 15000 # 15 seconds in milliseconds
chunks = [audio[i:i + chunk_length_ms] for i in range(0, len(audio), chunk_length_ms)]
aligned_chunks = []
transcription_time_total = 0
# Transcribe each chunk and measure time
for chunk_id, chunk in enumerate(chunks):
start_time = timeit.default_timer()
chunk.export("temp_chunk.wav", format="wav")
output = pipe("temp_chunk.wav")
transcription_time = timeit.default_timer() - start_time
transcription_time_total += transcription_time
# Calculate start and end times in seconds
start_time_sec = chunk_id * 15
end_time_sec = start_time_sec + len(chunk) / 1000.0
aligned_chunks.append({
"chunk_id": chunk_id,
"chunk_length": len(chunk) / 1000.0,
"text": output["text"],
"start_time": start_time_sec,
"end_time": end_time_sec,
"transcription_time": transcription_time
})
# Clean up temporary files
if os.path.exists("temp_chunk.wav"):
os.remove("temp_chunk.wav")
if os.path.exists("converted_audio.wav"):
os.remove("converted_audio.wav")
if os.path.exists(audio_filepath):
os.remove(audio_filepath)
return aligned_chunks
file_transcribe = gr.Interface(
fn=transcribe_speech_local,
inputs=gr.Audio(sources="upload", type="filepath"),
outputs=gr.JSON(label="Transcription with Time Alignment"),
allow_flagging="never"
)
link_transcribe = gr.Interface(
fn=transcribe_speech_from_youtube,
inputs=gr.Textbox(lines=1, placeholder="Enter YouTube URL here...", label="YouTube URL"),
outputs=gr.JSON(label="Transcription with Time Alignment"),
allow_flagging="never"
)
demo = gr.TabbedInterface(
[file_transcribe, link_transcribe ],
["Local files(mp3/mp4/wav)", "Links"]
)
demo.launch(share=True)
|