aadnk commited on
Commit
74b1efd
1 Parent(s): 883c794

Adding CLI

Browse files

This is similar to the CLI in Whisper, but it also supports
downloading URLs (also playlists), and using a VAD.

Files changed (4) hide show
  1. app.py +64 -49
  2. cli.py +108 -0
  3. src/download.py +14 -7
  4. src/vad.py +3 -1
app.py CHANGED
@@ -53,7 +53,7 @@ class WhisperTranscriber:
53
  self.inputAudioMaxDuration = inputAudioMaxDuration
54
  self.deleteUploadedFiles = deleteUploadedFiles
55
 
56
- def transcribe_file(self, modelName, languageName, urlData, uploadFile, microphoneData, task, vad, vadMergeWindow, vadMaxMergeSize, vadPadding):
57
  try:
58
  source, sourceName = self.__get_source(urlData, uploadFile, microphoneData)
59
 
@@ -67,54 +67,14 @@ class WhisperTranscriber:
67
  model = whisper.load_model(selectedModel)
68
  self.model_cache[selectedModel] = model
69
 
70
- # Callable for processing an audio file
71
- whisperCallable = lambda audio : model.transcribe(audio, language=selectedLanguage, task=task)
72
-
73
- # The results
74
- if (vad == 'silero-vad'):
75
- # Use Silero VAD and include gaps
76
- if (self.vad_model is None):
77
- self.vad_model = VadSileroTranscription()
78
-
79
- process_gaps = VadSileroTranscription(transcribe_non_speech = True,
80
- max_silent_period=vadMergeWindow, max_merge_size=vadMaxMergeSize,
81
- segment_padding_left=vadPadding, segment_padding_right=vadPadding, copy=self.vad_model)
82
- result = process_gaps.transcribe(source, whisperCallable)
83
- elif (vad == 'silero-vad-skip-gaps'):
84
- # Use Silero VAD
85
- if (self.vad_model is None):
86
- self.vad_model = VadSileroTranscription()
87
-
88
- skip_gaps = VadSileroTranscription(transcribe_non_speech = False,
89
- max_silent_period=vadMergeWindow, max_merge_size=vadMaxMergeSize,
90
- segment_padding_left=vadPadding, segment_padding_right=vadPadding, copy=self.vad_model)
91
- result = skip_gaps.transcribe(source, whisperCallable)
92
- elif (vad == 'periodic-vad'):
93
- # Very simple VAD - mark every 5 minutes as speech. This makes it less likely that Whisper enters an infinite loop, but
94
- # it may create a break in the middle of a sentence, causing some artifacts.
95
- periodic_vad = VadPeriodicTranscription(periodic_duration=vadMaxMergeSize)
96
- result = periodic_vad.transcribe(source, whisperCallable)
97
- else:
98
- # Default VAD
99
- result = whisperCallable(source)
100
-
101
- text = result["text"]
102
-
103
- language = result["language"]
104
- languageMaxLineWidth = self.__get_max_line_width(language)
105
-
106
- print("Max line width " + str(languageMaxLineWidth))
107
- vtt = self.__get_subs(result["segments"], "vtt", languageMaxLineWidth)
108
- srt = self.__get_subs(result["segments"], "srt", languageMaxLineWidth)
109
-
110
- # Files that can be downloaded
111
  downloadDirectory = tempfile.mkdtemp()
 
112
  filePrefix = slugify(sourceName, allow_unicode=True)
113
-
114
- download = []
115
- download.append(self.__create_file(srt, downloadDirectory, filePrefix + "-subs.srt"));
116
- download.append(self.__create_file(vtt, downloadDirectory, filePrefix + "-subs.vtt"));
117
- download.append(self.__create_file(text, downloadDirectory, filePrefix + "-transcript.txt"));
118
 
119
  return download, text, vtt
120
 
@@ -127,13 +87,68 @@ class WhisperTranscriber:
127
  except ExceededMaximumDuration as e:
128
  return [], ("[ERROR]: Maximum remote video length is " + str(e.maxDuration) + "s, file was " + str(e.videoDuration) + "s"), "[ERROR]"
129
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
130
  def clear_cache(self):
131
  self.model_cache = dict()
 
132
 
133
  def __get_source(self, urlData, uploadFile, microphoneData):
134
  if urlData:
135
  # Download from YouTube
136
- source = download_url(urlData, self.inputAudioMaxDuration)
137
  else:
138
  # File input
139
  source = uploadFile if uploadFile is not None else microphoneData
@@ -194,7 +209,7 @@ def create_ui(inputAudioMaxDuration, share=False, server_name: str = None):
194
 
195
  ui_article = "Read the [documentation here](https://huggingface.co/spaces/aadnk/whisper-webui/blob/main/docs/options.md)"
196
 
197
- demo = gr.Interface(fn=ui.transcribe_file, description=ui_description, article=ui_article, inputs=[
198
  gr.Dropdown(choices=["tiny", "base", "small", "medium", "large"], value="medium", label="Model"),
199
  gr.Dropdown(choices=sorted(LANGUAGES), label="Language"),
200
  gr.Text(label="URL (YouTube, etc.)"),
 
53
  self.inputAudioMaxDuration = inputAudioMaxDuration
54
  self.deleteUploadedFiles = deleteUploadedFiles
55
 
56
+ def transcribe_webui(self, modelName, languageName, urlData, uploadFile, microphoneData, task, vad, vadMergeWindow, vadMaxMergeSize, vadPadding):
57
  try:
58
  source, sourceName = self.__get_source(urlData, uploadFile, microphoneData)
59
 
 
67
  model = whisper.load_model(selectedModel)
68
  self.model_cache[selectedModel] = model
69
 
70
+ # Execute whisper
71
+ result = self.transcribe_file(model, source, selectedLanguage, task, vad, vadMergeWindow, vadMaxMergeSize, vadPadding)
72
+
73
+ # Write result
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
74
  downloadDirectory = tempfile.mkdtemp()
75
+
76
  filePrefix = slugify(sourceName, allow_unicode=True)
77
+ download, text, vtt = self.write_result(result, filePrefix, downloadDirectory)
 
 
 
 
78
 
79
  return download, text, vtt
80
 
 
87
  except ExceededMaximumDuration as e:
88
  return [], ("[ERROR]: Maximum remote video length is " + str(e.maxDuration) + "s, file was " + str(e.videoDuration) + "s"), "[ERROR]"
89
 
90
+ def transcribe_file(self, model: whisper.Whisper, audio_path: str, language: str, task: str = None, vad: str = None,
91
+ vadMergeWindow: float = 5, vadMaxMergeSize: float = 150, vadPadding: float = 1, **decodeOptions: dict):
92
+ # Callable for processing an audio file
93
+ whisperCallable = lambda audio : model.transcribe(audio, language=language, task=task, **decodeOptions)
94
+
95
+ # The results
96
+ if (vad == 'silero-vad'):
97
+ # Use Silero VAD and include gaps
98
+ if (self.vad_model is None):
99
+ self.vad_model = VadSileroTranscription()
100
+
101
+ process_gaps = VadSileroTranscription(transcribe_non_speech = True,
102
+ max_silent_period=vadMergeWindow, max_merge_size=vadMaxMergeSize,
103
+ segment_padding_left=vadPadding, segment_padding_right=vadPadding, copy=self.vad_model)
104
+ result = process_gaps.transcribe(audio_path, whisperCallable)
105
+ elif (vad == 'silero-vad-skip-gaps'):
106
+ # Use Silero VAD
107
+ if (self.vad_model is None):
108
+ self.vad_model = VadSileroTranscription()
109
+
110
+ skip_gaps = VadSileroTranscription(transcribe_non_speech = False,
111
+ max_silent_period=vadMergeWindow, max_merge_size=vadMaxMergeSize,
112
+ segment_padding_left=vadPadding, segment_padding_right=vadPadding, copy=self.vad_model)
113
+ result = skip_gaps.transcribe(audio_path, whisperCallable)
114
+ elif (vad == 'periodic-vad'):
115
+ # Very simple VAD - mark every 5 minutes as speech. This makes it less likely that Whisper enters an infinite loop, but
116
+ # it may create a break in the middle of a sentence, causing some artifacts.
117
+ periodic_vad = VadPeriodicTranscription(periodic_duration=vadMaxMergeSize)
118
+ result = periodic_vad.transcribe(audio_path, whisperCallable)
119
+ else:
120
+ # Default VAD
121
+ result = whisperCallable(audio_path)
122
+
123
+ return result
124
+
125
+ def write_result(self, result: dict, source_name: str, output_dir: str):
126
+ if not os.path.exists(output_dir):
127
+ os.makedirs(output_dir)
128
+
129
+ text = result["text"]
130
+ language = result["language"]
131
+ languageMaxLineWidth = self.__get_max_line_width(language)
132
+
133
+ print("Max line width " + str(languageMaxLineWidth))
134
+ vtt = self.__get_subs(result["segments"], "vtt", languageMaxLineWidth)
135
+ srt = self.__get_subs(result["segments"], "srt", languageMaxLineWidth)
136
+
137
+ output_files = []
138
+ output_files.append(self.__create_file(srt, output_dir, source_name + "-subs.srt"));
139
+ output_files.append(self.__create_file(vtt, output_dir, source_name + "-subs.vtt"));
140
+ output_files.append(self.__create_file(text, output_dir, source_name + "-transcript.txt"));
141
+
142
+ return output_files, text, vtt
143
+
144
  def clear_cache(self):
145
  self.model_cache = dict()
146
+ self.vad_model = None
147
 
148
  def __get_source(self, urlData, uploadFile, microphoneData):
149
  if urlData:
150
  # Download from YouTube
151
+ source = download_url(urlData, self.inputAudioMaxDuration)[0]
152
  else:
153
  # File input
154
  source = uploadFile if uploadFile is not None else microphoneData
 
209
 
210
  ui_article = "Read the [documentation here](https://huggingface.co/spaces/aadnk/whisper-webui/blob/main/docs/options.md)"
211
 
212
+ demo = gr.Interface(fn=ui.transcribe_webui, description=ui_description, article=ui_article, inputs=[
213
  gr.Dropdown(choices=["tiny", "base", "small", "medium", "large"], value="medium", label="Model"),
214
  gr.Dropdown(choices=sorted(LANGUAGES), label="Language"),
215
  gr.Text(label="URL (YouTube, etc.)"),
cli.py ADDED
@@ -0,0 +1,108 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import argparse
2
+ import os
3
+ import pathlib
4
+ from urllib.parse import urlparse
5
+ import warnings
6
+ import numpy as np
7
+
8
+ import whisper
9
+
10
+ import torch
11
+ from app import LANGUAGES, WhisperTranscriber
12
+ from src.download import download_url
13
+
14
+ from src.utils import optional_float, optional_int, str2bool
15
+
16
+
17
+ def cli():
18
+ parser = argparse.ArgumentParser(formatter_class=argparse.ArgumentDefaultsHelpFormatter)
19
+ parser.add_argument("audio", nargs="+", type=str, help="audio file(s) to transcribe")
20
+ parser.add_argument("--model", default="small", choices=["tiny", "base", "small", "medium", "large"], help="name of the Whisper model to use")
21
+ parser.add_argument("--model_dir", type=str, default=None, help="the path to save model files; uses ~/.cache/whisper by default")
22
+ parser.add_argument("--device", default="cuda" if torch.cuda.is_available() else "cpu", help="device to use for PyTorch inference")
23
+ parser.add_argument("--output_dir", "-o", type=str, default=".", help="directory to save the outputs")
24
+ parser.add_argument("--verbose", type=str2bool, default=True, help="whether to print out the progress and debug messages")
25
+
26
+ parser.add_argument("--task", type=str, default="transcribe", choices=["transcribe", "translate"], help="whether to perform X->X speech recognition ('transcribe') or X->English translation ('translate')")
27
+ parser.add_argument("--language", type=str, default=None, choices=sorted(LANGUAGES), help="language spoken in the audio, specify None to perform language detection")
28
+
29
+ parser.add_argument("--vad", type=str, default="none", choices=["none", "silero-vad", "silero-vad-skip-gaps", "periodic-vad"], help="The voice activity detection algorithm to use")
30
+ parser.add_argument("--vad_merge_window", type=optional_float, default=5, help="The window size (in seconds) to merge voice segments")
31
+ parser.add_argument("--vad_max_merge_size", type=optional_float, default=150, help="The maximum size (in seconds) of a voice segment")
32
+ parser.add_argument("--vad_padding", type=optional_float, default=1, help="The padding (in seconds) to add to each voice segment")
33
+
34
+ parser.add_argument("--temperature", type=float, default=0, help="temperature to use for sampling")
35
+ parser.add_argument("--best_of", type=optional_int, default=5, help="number of candidates when sampling with non-zero temperature")
36
+ parser.add_argument("--beam_size", type=optional_int, default=5, help="number of beams in beam search, only applicable when temperature is zero")
37
+ parser.add_argument("--patience", type=float, default=None, help="optional patience value to use in beam decoding, as in https://arxiv.org/abs/2204.05424, the default (1.0) is equivalent to conventional beam search")
38
+ parser.add_argument("--length_penalty", type=float, default=None, help="optional token length penalty coefficient (alpha) as in https://arxiv.org/abs/1609.08144, uses simple lengt normalization by default")
39
+
40
+ parser.add_argument("--suppress_tokens", type=str, default="-1", help="comma-separated list of token ids to suppress during sampling; '-1' will suppress most special characters except common punctuations")
41
+ parser.add_argument("--initial_prompt", type=str, default=None, help="optional text to provide as a prompt for the first window.")
42
+ parser.add_argument("--condition_on_previous_text", type=str2bool, default=True, help="if True, provide the previous output of the model as a prompt for the next window; disabling may make the text inconsistent across windows, but the model becomes less prone to getting stuck in a failure loop")
43
+ parser.add_argument("--fp16", type=str2bool, default=True, help="whether to perform inference in fp16; True by default")
44
+
45
+ parser.add_argument("--temperature_increment_on_fallback", type=optional_float, default=0.2, help="temperature to increase when falling back when the decoding fails to meet either of the thresholds below")
46
+ parser.add_argument("--compression_ratio_threshold", type=optional_float, default=2.4, help="if the gzip compression ratio is higher than this value, treat the decoding as failed")
47
+ parser.add_argument("--logprob_threshold", type=optional_float, default=-1.0, help="if the average log probability is lower than this value, treat the decoding as failed")
48
+ parser.add_argument("--no_speech_threshold", type=optional_float, default=0.6, help="if the probability of the <|nospeech|> token is higher than this value AND the decoding has failed due to `logprob_threshold`, consider the segment as silence")
49
+
50
+ args = parser.parse_args().__dict__
51
+ model_name: str = args.pop("model")
52
+ model_dir: str = args.pop("model_dir")
53
+ output_dir: str = args.pop("output_dir")
54
+ device: str = args.pop("device")
55
+ os.makedirs(output_dir, exist_ok=True)
56
+
57
+ if model_name.endswith(".en") and args["language"] not in {"en", "English"}:
58
+ warnings.warn(f"{model_name} is an English-only model but receipted '{args['language']}'; using English instead.")
59
+ args["language"] = "en"
60
+
61
+ temperature = args.pop("temperature")
62
+ temperature_increment_on_fallback = args.pop("temperature_increment_on_fallback")
63
+ if temperature_increment_on_fallback is not None:
64
+ temperature = tuple(np.arange(temperature, 1.0 + 1e-6, temperature_increment_on_fallback))
65
+ else:
66
+ temperature = [temperature]
67
+
68
+ vad = args.pop("vad")
69
+ vad_merge_window = args.pop("vad_merge_window")
70
+ vad_max_merge_size = args.pop("vad_max_merge_size")
71
+ vad_padding = args.pop("vad_padding")
72
+
73
+ model = whisper.load_model(model_name, device=device, download_root=model_dir)
74
+ transcriber = WhisperTranscriber(deleteUploadedFiles=False)
75
+
76
+ for audio_path in args.pop("audio"):
77
+ sources = []
78
+
79
+ # Detect URL and download the audio
80
+ if (uri_validator(audio_path)):
81
+ # Download from YouTube/URL directly
82
+ for source_path in download_url(audio_path, maxDuration=-1, destinationDirectory=output_dir, playlistItems=None):
83
+ source_name = os.path.basename(source_path)
84
+ sources.append({ "path": source_path, "name": source_name })
85
+ else:
86
+ sources.append({ "path": audio_path, "name": os.path.basename(audio_path) })
87
+
88
+ for source in sources:
89
+ source_path = source["path"]
90
+ source_name = source["name"]
91
+
92
+ result = transcriber.transcribe_file(model, source_path, temperature=temperature,
93
+ vad=vad, vadMergeWindow=vad_merge_window, vadMaxMergeSize=vad_max_merge_size,
94
+ vadPadding=vad_padding, **args)
95
+
96
+ transcriber.write_result(result, source_name, output_dir)
97
+
98
+ transcriber.clear_cache()
99
+
100
+ def uri_validator(x):
101
+ try:
102
+ result = urlparse(x)
103
+ return all([result.scheme, result.netloc])
104
+ except:
105
+ return False
106
+
107
+ if __name__ == '__main__':
108
+ cli()
src/download.py CHANGED
@@ -1,4 +1,5 @@
1
  from tempfile import mkdtemp
 
2
  from yt_dlp import YoutubeDL
3
 
4
  import yt_dlp
@@ -13,25 +14,28 @@ class FilenameCollectorPP(PostProcessor):
13
  self.filenames.append(information["filepath"])
14
  return [], information
15
 
16
- def download_url(url: str, maxDuration: int = None):
17
  try:
18
- return _perform_download(url, maxDuration=maxDuration)
19
  except yt_dlp.utils.DownloadError as e:
20
  # In case of an OS error, try again with a different output template
21
  if e.msg and e.msg.find("[Errno 36] File name too long") >= 0:
22
  return _perform_download(url, maxDuration=maxDuration, outputTemplate="%(title).10s %(id)s.%(ext)s")
23
  pass
24
 
25
- def _perform_download(url: str, maxDuration: int = None, outputTemplate: str = None):
26
- destinationDirectory = mkdtemp()
 
 
27
 
28
  ydl_opts = {
29
  "format": "bestaudio/best",
30
- 'playlist_items': '1',
31
  'paths': {
32
  'home': destinationDirectory
33
  }
34
  }
 
 
35
 
36
  # Add output template if specified
37
  if outputTemplate:
@@ -53,8 +57,11 @@ def _perform_download(url: str, maxDuration: int = None, outputTemplate: str = N
53
  if len(filename_collector.filenames) <= 0:
54
  raise Exception("Cannot download " + url)
55
 
56
- result = filename_collector.filenames[0]
57
- print("Downloaded " + result)
 
 
 
58
 
59
  return result
60
 
 
1
  from tempfile import mkdtemp
2
+ from typing import List
3
  from yt_dlp import YoutubeDL
4
 
5
  import yt_dlp
 
14
  self.filenames.append(information["filepath"])
15
  return [], information
16
 
17
+ def download_url(url: str, maxDuration: int = None, destinationDirectory: str = None, playlistItems: str = "1") -> List[str]:
18
  try:
19
+ return _perform_download(url, maxDuration=maxDuration, outputTemplate=None, destinationDirectory=destinationDirectory, playlistItems=playlistItems)
20
  except yt_dlp.utils.DownloadError as e:
21
  # In case of an OS error, try again with a different output template
22
  if e.msg and e.msg.find("[Errno 36] File name too long") >= 0:
23
  return _perform_download(url, maxDuration=maxDuration, outputTemplate="%(title).10s %(id)s.%(ext)s")
24
  pass
25
 
26
+ def _perform_download(url: str, maxDuration: int = None, outputTemplate: str = None, destinationDirectory: str = None, playlistItems: str = "1"):
27
+ # Create a temporary directory to store the downloaded files
28
+ if destinationDirectory is None:
29
+ destinationDirectory = mkdtemp()
30
 
31
  ydl_opts = {
32
  "format": "bestaudio/best",
 
33
  'paths': {
34
  'home': destinationDirectory
35
  }
36
  }
37
+ if (playlistItems):
38
+ ydl_opts['playlist_items'] = playlistItems
39
 
40
  # Add output template if specified
41
  if outputTemplate:
 
57
  if len(filename_collector.filenames) <= 0:
58
  raise Exception("Cannot download " + url)
59
 
60
+ result = []
61
+
62
+ for filename in filename_collector.filenames:
63
+ result.append(filename)
64
+ print("Downloaded " + filename)
65
 
66
  return result
67
 
src/vad.py CHANGED
@@ -188,7 +188,9 @@ class AbstractTranscription(ABC):
188
 
189
  result.append(current_segment)
190
 
191
- last_segment = result[-1]
 
 
192
 
193
  # Also include total duration if specified
194
  if (total_duration is not None):
 
188
 
189
  result.append(current_segment)
190
 
191
+ # Add last segment
192
+ last_segment = segments[-1]
193
+ result.append(last_segment)
194
 
195
  # Also include total duration if specified
196
  if (total_duration is not None):