Spaces:
Runtime error
Runtime error
File size: 7,500 Bytes
bbe4238 884c49d 70814d8 0c8b1e1 70814d8 0c8b1e1 67bd832 4ce3fcd e4b4acb 4ce3fcd 0c8b1e1 e4b4acb 0c8b1e1 4f7fe11 0c8b1e1 c1a800b 70814d8 c1a800b 70814d8 0c8b1e1 70814d8 4f7fe11 70814d8 4f7fe11 70814d8 4f7fe11 70814d8 4f7fe11 70814d8 4f7fe11 70814d8 b21acb5 70814d8 0c8b1e1 bbe4238 70814d8 bbe4238 70814d8 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 |
import gradio as gr
from whisperplus.pipelines.whisper import SpeechToTextPipeline
from whisperplus.pipelines.whisper_diarize import ASRDiarizationPipeline
from whisperplus.utils.download_utils import download_and_convert_to_mp3
from whisperplus.utils.text_utils import format_speech_to_dialogue
import subprocess
def install_package(package):
subprocess.check_call(['pip', 'install', package, '--no-build-isolation'])
# Then install flash-attn
install_package('flash-attn')
def youtube_url_to_text(url, model_id, language_choice):
"""
Main function that downloads and converts a video to MP3 format, performs speech-to-text conversion using
a specified model, and returns the transcript along with the video path.
Args:
url (str): The URL of the video to download and convert.
model_id (str): The ID of the speech-to-text model to use.
language_choice (str): The language choice for the speech-to-text conversion.
Returns:
transcript (str): The transcript of the speech-to-text conversion.
video_path (str): The path of the downloaded video.
"""
video_path = download_and_convert_to_mp3(url)
output = SpeechToTextPipeline(model_id)
print(video_path)
transcript = output(audio_path=video_path, language=language_choice)
return transcript, video_path
def speaker_diarization(url, model_id, num_speakers, min_speaker, max_speaker):
"""
Main function that downloads and converts a video to MP3 format, performs speech-to-text conversion using
a specified model, and returns the transcript along with the video path.
Args:
url (str): The URL of the video to download and convert.
model_id (str): The ID of the speech-to-text model to use.
language_choice (str): The language choice for the speech-to-text conversion.
Returns:
transcript (str): The transcript of the speech-to-text conversion.
video_path (str): The path of the downloaded video.
"""
pipeline = ASRDiarizationPipeline.from_pretrained(
asr_model=model_id,
diarizer_model="pyannote/speaker-diarization",
chunk_length_s=30,
device="cuda",
)
audio_path = download_and_convert_to_mp3(url)
output_text = pipeline(
audio_path, num_speakers=num_speakers, min_speaker=min_speaker, max_speaker=max_speaker)
dialogue = format_speech_to_dialogue(output_text)
return dialogue, audio_path
def youtube_url_to_text_app():
with gr.Blocks():
with gr.Row():
with gr.Column():
youtube_url_path = gr.Text(placeholder="Enter Youtube URL", label="Youtube URL")
language_choice = gr.Dropdown(
choices=[
"English",
"Turkish",
"Spanish",
"French",
"Chinese",
"Japanese",
"Korean",
],
value="Turkish",
label="Language",
)
whisper_model_id = gr.Dropdown(
choices=[
"openai/whisper-large-v3",
"openai/whisper-large",
"openai/whisper-medium",
"openai/whisper-base",
"openai/whisper-small",
"openai/whisper-tiny",
],
value="openai/whisper-large-v3",
label="Whisper Model",
)
whisperplus_in_predict = gr.Button(value="Generator")
with gr.Column():
output_text = gr.Textbox(label="Output Text")
output_audio = gr.Audio(label="Output Audio")
whisperplus_in_predict.click(
fn=youtube_url_to_text,
inputs=[
youtube_url_path,
whisper_model_id,
language_choice,
],
outputs=[output_text, output_audio],
)
gr.Examples(
examples=[
[
"https://www.youtube.com/watch?v=di3rHkEZuUw",
"distil-whisper/distil-large-v3",
"English",
],
],
fn=youtube_url_to_text,
inputs=[
youtube_url_path,
whisper_model_id,
language_choice,
],
outputs=[output_text, output_audio],
cache_examples=True,
)
def speaker_diarization_app():
with gr.Blocks():
with gr.Row():
with gr.Column():
youtube_url_path = gr.Text(placeholder="Enter Youtube URL", label="Youtube URL")
whisper_model_id = gr.Dropdown(
choices=[
"openai/whisper-large-v3",
"distil-whisper/distil-large-v3",
"distil-whisper/distil-large-v2",
],
value="distil-whisper/distil-large-v3",
label="Whisper Model",
)
num_speakers = gr.Number(value=2, label="Number of Speakers")
min_speaker = gr.Number(value=1, label="Minimum Number of Speakers")
max_speaker = gr.Number(value=2, label="Maximum Number of Speakers")
whisperplus_in_predict = gr.Button(value="Generator")
with gr.Column():
output_text = gr.Textbox(label="Output Text")
output_audio = gr.Audio(label="Output Audio")
whisperplus_in_predict.click(
fn=speaker_diarization,
inputs=[
youtube_url_path,
whisper_model_id,
num_speakers,
min_speaker,
max_speaker,
],
outputs=[output_text, output_audio],
)
gr.Examples(
examples=[
[
"https://www.youtube.com/shorts/o8PgLUgte2k",
"distil-whisper/distil-large-v3",
2,
1,
2,
],
],
fn=speaker_diarization,
inputs=[
youtube_url_path,
whisper_model_id,
num_speakers,
min_speaker,
max_speaker,
],
outputs=[output_text, output_audio],
cache_examples=False,
)
gradio_app = gr.Blocks()
with gradio_app:
gr.HTML(
"""
<h1 style='text-align: center'>
WhisperPlus: Advancing Speech-to-Text Processing 🚀
</h1>
""")
gr.HTML(
"""
<h3 style='text-align: center'>
Follow me for more!
<a href='https://twitter.com/kadirnar_ai' target='_blank'>Twitter</a> | <a href='https://github.com/kadirnar' target='_blank'>Github</a> | <a href='https://www.linkedin.com/in/kadir-nar/' target='_blank'>Linkedin</a> | <a href='https://www.huggingface.co/kadirnar/' target='_blank'>HuggingFace</a>
</h3>
""")
with gr.Row():
with gr.Column():
with gr.Tab(label="Youtube URL to Text"):
youtube_url_to_text_app()
with gr.Tab(label="Speaker Diarization"):
speaker_diarization_app()
gradio_app.launch(debug=True) |