Ensure GPU memory in diarization can be cleaned up
Browse files- app.py +27 -8
- cli.py +1 -1
- src/diarization/diarization.py +17 -10
- src/diarization/diarizationContainer.py +77 -0
app.py
CHANGED
@@ -15,6 +15,7 @@ import torch
|
|
15 |
|
16 |
from src.config import VAD_INITIAL_PROMPT_MODE_VALUES, ApplicationConfig, VadInitialPromptMode
|
17 |
from src.diarization.diarization import Diarization
|
|
|
18 |
from src.hooks.progressListener import ProgressListener
|
19 |
from src.hooks.subTaskProgressListener import SubTaskProgressListener
|
20 |
from src.hooks.whisperProgressHook import create_progress_listener_handle
|
@@ -74,7 +75,10 @@ class WhisperTranscriber:
|
|
74 |
self.deleteUploadedFiles = delete_uploaded_files
|
75 |
self.output_dir = output_dir
|
76 |
|
77 |
-
|
|
|
|
|
|
|
78 |
self.app_config = app_config
|
79 |
|
80 |
def set_parallel_devices(self, vad_parallel_devices: str):
|
@@ -88,6 +92,17 @@ class WhisperTranscriber:
|
|
88 |
self.vad_cpu_cores = min(os.cpu_count(), MAX_AUTO_CPU_CORES)
|
89 |
print("[Auto parallel] Using GPU devices " + str(self.parallel_device_list) + " and " + str(self.vad_cpu_cores) + " CPU cores for VAD/transcription.")
|
90 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
91 |
# Entry function for the simple tab
|
92 |
def transcribe_webui_simple(self, modelName, languageName, urlData, multipleFiles, microphoneData, task,
|
93 |
vad, vadMergeWindow, vadMaxMergeSize,
|
@@ -108,9 +123,9 @@ class WhisperTranscriber:
|
|
108 |
vadOptions = VadOptions(vad, vadMergeWindow, vadMaxMergeSize, self.app_config.vad_padding, self.app_config.vad_prompt_window, self.app_config.vad_initial_prompt_mode)
|
109 |
|
110 |
if diarization:
|
111 |
-
self.
|
112 |
else:
|
113 |
-
self.
|
114 |
|
115 |
return self.transcribe_webui(modelName, languageName, urlData, multipleFiles, microphoneData, task, vadOptions,
|
116 |
word_timestamps=word_timestamps, highlight_words=highlight_words, progress=progress)
|
@@ -157,10 +172,10 @@ class WhisperTranscriber:
|
|
157 |
|
158 |
# Set diarization
|
159 |
if diarization:
|
160 |
-
self.
|
161 |
-
|
162 |
else:
|
163 |
-
self.
|
164 |
|
165 |
return self.transcribe_webui(modelName, languageName, urlData, multipleFiles, microphoneData, task, vadOptions,
|
166 |
initial_prompt=initial_prompt, temperature=temperature, best_of=best_of, beam_size=beam_size, patience=patience, length_penalty=length_penalty, suppress_tokens=suppress_tokens,
|
@@ -226,9 +241,9 @@ class WhisperTranscriber:
|
|
226 |
current_progress += source_audio_duration
|
227 |
|
228 |
# Diarization
|
229 |
-
if self.diarization:
|
230 |
print("Diarizing ", source.source_path)
|
231 |
-
diarization_result = list(self.diarization.run(source.source_path))
|
232 |
|
233 |
# Print result
|
234 |
print("Diarization result: ")
|
@@ -494,6 +509,10 @@ class WhisperTranscriber:
|
|
494 |
if (self.cpu_parallel_context is not None):
|
495 |
self.cpu_parallel_context.close()
|
496 |
|
|
|
|
|
|
|
|
|
497 |
|
498 |
def create_ui(app_config: ApplicationConfig):
|
499 |
ui = WhisperTranscriber(app_config.input_audio_max_duration, app_config.vad_process_timeout, app_config.vad_cpu_cores,
|
|
|
15 |
|
16 |
from src.config import VAD_INITIAL_PROMPT_MODE_VALUES, ApplicationConfig, VadInitialPromptMode
|
17 |
from src.diarization.diarization import Diarization
|
18 |
+
from src.diarization.diarizationContainer import DiarizationContainer
|
19 |
from src.hooks.progressListener import ProgressListener
|
20 |
from src.hooks.subTaskProgressListener import SubTaskProgressListener
|
21 |
from src.hooks.whisperProgressHook import create_progress_listener_handle
|
|
|
75 |
self.deleteUploadedFiles = delete_uploaded_files
|
76 |
self.output_dir = output_dir
|
77 |
|
78 |
+
# Support for diarization
|
79 |
+
self.diarization: DiarizationContainer = None
|
80 |
+
# Dictionary with parameters to pass to diarization.run - if None, diarization is not enabled
|
81 |
+
self.diarization_kwargs = None
|
82 |
self.app_config = app_config
|
83 |
|
84 |
def set_parallel_devices(self, vad_parallel_devices: str):
|
|
|
92 |
self.vad_cpu_cores = min(os.cpu_count(), MAX_AUTO_CPU_CORES)
|
93 |
print("[Auto parallel] Using GPU devices " + str(self.parallel_device_list) + " and " + str(self.vad_cpu_cores) + " CPU cores for VAD/transcription.")
|
94 |
|
95 |
+
def set_diarization(self, auth_token: str, enable_daemon_process: bool = True, **kwargs):
|
96 |
+
if self.diarization is None:
|
97 |
+
self.diarization = DiarizationContainer(auth_token=auth_token, enable_daemon_process=enable_daemon_process,
|
98 |
+
auto_cleanup_timeout_seconds=self.vad_process_timeout, cache=self.model_cache)
|
99 |
+
# Set parameters
|
100 |
+
self.diarization_kwargs = kwargs
|
101 |
+
|
102 |
+
def unset_diarization(self):
|
103 |
+
self.diarization.cleanup()
|
104 |
+
self.diarization_kwargs = None
|
105 |
+
|
106 |
# Entry function for the simple tab
|
107 |
def transcribe_webui_simple(self, modelName, languageName, urlData, multipleFiles, microphoneData, task,
|
108 |
vad, vadMergeWindow, vadMaxMergeSize,
|
|
|
123 |
vadOptions = VadOptions(vad, vadMergeWindow, vadMaxMergeSize, self.app_config.vad_padding, self.app_config.vad_prompt_window, self.app_config.vad_initial_prompt_mode)
|
124 |
|
125 |
if diarization:
|
126 |
+
self.set_diarization(auth_token=self.app_config.auth_token, num_speakers=diarization_speakers)
|
127 |
else:
|
128 |
+
self.unset_diarization()
|
129 |
|
130 |
return self.transcribe_webui(modelName, languageName, urlData, multipleFiles, microphoneData, task, vadOptions,
|
131 |
word_timestamps=word_timestamps, highlight_words=highlight_words, progress=progress)
|
|
|
172 |
|
173 |
# Set diarization
|
174 |
if diarization:
|
175 |
+
self.set_diarization(auth_token=self.app_config.auth_token, num_speakers=diarization_speakers,
|
176 |
+
min_speakers=diarization_min_speakers, max_speakers=diarization_max_speakers)
|
177 |
else:
|
178 |
+
self.unset_diarization()
|
179 |
|
180 |
return self.transcribe_webui(modelName, languageName, urlData, multipleFiles, microphoneData, task, vadOptions,
|
181 |
initial_prompt=initial_prompt, temperature=temperature, best_of=best_of, beam_size=beam_size, patience=patience, length_penalty=length_penalty, suppress_tokens=suppress_tokens,
|
|
|
241 |
current_progress += source_audio_duration
|
242 |
|
243 |
# Diarization
|
244 |
+
if self.diarization and self.diarization_kwargs:
|
245 |
print("Diarizing ", source.source_path)
|
246 |
+
diarization_result = list(self.diarization.run(source.source_path, **self.diarization_kwargs))
|
247 |
|
248 |
# Print result
|
249 |
print("Diarization result: ")
|
|
|
509 |
if (self.cpu_parallel_context is not None):
|
510 |
self.cpu_parallel_context.close()
|
511 |
|
512 |
+
# Cleanup diarization
|
513 |
+
if (self.diarization is not None):
|
514 |
+
self.diarization.cleanup()
|
515 |
+
self.diarization = None
|
516 |
|
517 |
def create_ui(app_config: ApplicationConfig):
|
518 |
ui = WhisperTranscriber(app_config.input_audio_max_duration, app_config.vad_process_timeout, app_config.vad_cpu_cores,
|
cli.py
CHANGED
@@ -162,7 +162,7 @@ def cli():
|
|
162 |
transcriber.set_auto_parallel(auto_parallel)
|
163 |
|
164 |
if diarization:
|
165 |
-
transcriber.set_diarization(
|
166 |
|
167 |
model = create_whisper_container(whisper_implementation=whisper_implementation, model_name=model_name,
|
168 |
device=device, compute_type=compute_type, download_root=model_dir, models=app_config.models)
|
|
|
162 |
transcriber.set_auto_parallel(auto_parallel)
|
163 |
|
164 |
if diarization:
|
165 |
+
transcriber.set_diarization(auth_token=auth_token, enable_daemon_process=False, num_speakers=num_speakers, min_speakers=min_speakers, max_speakers=max_speakers)
|
166 |
|
167 |
model = create_whisper_container(whisper_implementation=whisper_implementation, model_name=model_name,
|
168 |
device=device, compute_type=compute_type, download_root=model_dir, models=app_config.models)
|
src/diarization/diarization.py
CHANGED
@@ -1,4 +1,5 @@
|
|
1 |
import argparse
|
|
|
2 |
import json
|
3 |
import os
|
4 |
from pathlib import Path
|
@@ -8,9 +9,6 @@ import torch
|
|
8 |
|
9 |
import ffmpeg
|
10 |
|
11 |
-
from src.diarization.transcriptLoader import load_transcript
|
12 |
-
from src.utils import write_srt
|
13 |
-
|
14 |
class DiarizationEntry:
|
15 |
def __init__(self, start, end, speaker):
|
16 |
self.start = start
|
@@ -28,7 +26,7 @@ class DiarizationEntry:
|
|
28 |
}
|
29 |
|
30 |
class Diarization:
|
31 |
-
def __init__(self, auth_token=None
|
32 |
if auth_token is None:
|
33 |
auth_token = os.environ.get("HK_ACCESS_TOKEN")
|
34 |
if auth_token is None:
|
@@ -37,7 +35,6 @@ class Diarization:
|
|
37 |
self.auth_token = auth_token
|
38 |
self.initialized = False
|
39 |
self.pipeline = None
|
40 |
-
self.pipeline_kwargs = kwargs
|
41 |
|
42 |
@staticmethod
|
43 |
def has_libraries():
|
@@ -54,6 +51,7 @@ class Diarization:
|
|
54 |
from pyannote.audio import Pipeline
|
55 |
|
56 |
self.pipeline = Pipeline.from_pretrained("pyannote/speaker-diarization@2.1", use_auth_token=self.auth_token)
|
|
|
57 |
|
58 |
# Load GPU mode if available
|
59 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
@@ -63,7 +61,7 @@ class Diarization:
|
|
63 |
else:
|
64 |
print("Diarization - using CPU")
|
65 |
|
66 |
-
def run(self, audio_file):
|
67 |
self.initialize()
|
68 |
audio_file_obj = Path(audio_file)
|
69 |
|
@@ -78,7 +76,7 @@ class Diarization:
|
|
78 |
except ffmpeg.Error as e:
|
79 |
print(f"Error occurred during audio conversion: {e.stderr}")
|
80 |
|
81 |
-
diarization = self.pipeline(target_file, **
|
82 |
|
83 |
if target_file != audio_file:
|
84 |
# Delete temp file
|
@@ -148,6 +146,9 @@ def _write_file(input_file: str, output_path: str, output_extension: str, file_w
|
|
148 |
print(f"Output saved to {effective_path}")
|
149 |
|
150 |
def main():
|
|
|
|
|
|
|
151 |
parser = argparse.ArgumentParser(description='Add speakers to a SRT file or Whisper JSON file using pyannote/speaker-diarization.')
|
152 |
parser.add_argument('audio_file', type=str, help='Input audio file')
|
153 |
parser.add_argument('whisper_file', type=str, help='Input Whisper JSON/SRT file')
|
@@ -166,8 +167,8 @@ def main():
|
|
166 |
# Read whisper JSON or SRT file
|
167 |
whisper_result = load_transcript(args.whisper_file)
|
168 |
|
169 |
-
diarization = Diarization(auth_token=args.auth_token
|
170 |
-
diarization_result = list(diarization.run(args.audio_file))
|
171 |
|
172 |
# Print result
|
173 |
print("Diarization result:")
|
@@ -185,4 +186,10 @@ def main():
|
|
185 |
lambda f: write_srt(marked_whisper_result["segments"], f, maxLineWidth=args.max_line_width))
|
186 |
|
187 |
if __name__ == "__main__":
|
188 |
-
main()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
import argparse
|
2 |
+
import gc
|
3 |
import json
|
4 |
import os
|
5 |
from pathlib import Path
|
|
|
9 |
|
10 |
import ffmpeg
|
11 |
|
|
|
|
|
|
|
12 |
class DiarizationEntry:
|
13 |
def __init__(self, start, end, speaker):
|
14 |
self.start = start
|
|
|
26 |
}
|
27 |
|
28 |
class Diarization:
|
29 |
+
def __init__(self, auth_token=None):
|
30 |
if auth_token is None:
|
31 |
auth_token = os.environ.get("HK_ACCESS_TOKEN")
|
32 |
if auth_token is None:
|
|
|
35 |
self.auth_token = auth_token
|
36 |
self.initialized = False
|
37 |
self.pipeline = None
|
|
|
38 |
|
39 |
@staticmethod
|
40 |
def has_libraries():
|
|
|
51 |
from pyannote.audio import Pipeline
|
52 |
|
53 |
self.pipeline = Pipeline.from_pretrained("pyannote/speaker-diarization@2.1", use_auth_token=self.auth_token)
|
54 |
+
self.initialized = True
|
55 |
|
56 |
# Load GPU mode if available
|
57 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
|
|
61 |
else:
|
62 |
print("Diarization - using CPU")
|
63 |
|
64 |
+
def run(self, audio_file, **kwargs):
|
65 |
self.initialize()
|
66 |
audio_file_obj = Path(audio_file)
|
67 |
|
|
|
76 |
except ffmpeg.Error as e:
|
77 |
print(f"Error occurred during audio conversion: {e.stderr}")
|
78 |
|
79 |
+
diarization = self.pipeline(target_file, **kwargs)
|
80 |
|
81 |
if target_file != audio_file:
|
82 |
# Delete temp file
|
|
|
146 |
print(f"Output saved to {effective_path}")
|
147 |
|
148 |
def main():
|
149 |
+
from src.utils import write_srt
|
150 |
+
from src.diarization.transcriptLoader import load_transcript
|
151 |
+
|
152 |
parser = argparse.ArgumentParser(description='Add speakers to a SRT file or Whisper JSON file using pyannote/speaker-diarization.')
|
153 |
parser.add_argument('audio_file', type=str, help='Input audio file')
|
154 |
parser.add_argument('whisper_file', type=str, help='Input Whisper JSON/SRT file')
|
|
|
167 |
# Read whisper JSON or SRT file
|
168 |
whisper_result = load_transcript(args.whisper_file)
|
169 |
|
170 |
+
diarization = Diarization(auth_token=args.auth_token)
|
171 |
+
diarization_result = list(diarization.run(args.audio_file, num_speakers=args.num_speakers, min_speakers=args.min_speakers, max_speakers=args.max_speakers))
|
172 |
|
173 |
# Print result
|
174 |
print("Diarization result:")
|
|
|
186 |
lambda f: write_srt(marked_whisper_result["segments"], f, maxLineWidth=args.max_line_width))
|
187 |
|
188 |
if __name__ == "__main__":
|
189 |
+
main()
|
190 |
+
|
191 |
+
#test = Diarization()
|
192 |
+
#print("Initializing")
|
193 |
+
#test.initialize()
|
194 |
+
|
195 |
+
#input("Press Enter to continue...")
|
src/diarization/diarizationContainer.py
ADDED
@@ -0,0 +1,77 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from typing import List
|
2 |
+
from src.diarization.diarization import Diarization, DiarizationEntry
|
3 |
+
from src.modelCache import GLOBAL_MODEL_CACHE, ModelCache
|
4 |
+
from src.vadParallel import ParallelContext
|
5 |
+
|
6 |
+
class DiarizationContainer:
|
7 |
+
def __init__(self, auth_token: str = None, enable_daemon_process: bool = True, auto_cleanup_timeout_seconds=60, cache: ModelCache = None):
|
8 |
+
self.auth_token = auth_token
|
9 |
+
self.enable_daemon_process = enable_daemon_process
|
10 |
+
self.auto_cleanup_timeout_seconds = auto_cleanup_timeout_seconds
|
11 |
+
self.diarization_context: ParallelContext = None
|
12 |
+
self.cache = cache
|
13 |
+
self.model = None
|
14 |
+
|
15 |
+
def run(self, audio_file, **kwargs):
|
16 |
+
# Create parallel context if needed
|
17 |
+
if self.diarization_context is None and self.enable_daemon_process:
|
18 |
+
# Number of processes is set to 1 as we mainly use this in order to clean up GPU memory
|
19 |
+
self.diarization_context = ParallelContext(num_processes=1)
|
20 |
+
|
21 |
+
# Run directly
|
22 |
+
if self.diarization_context is None:
|
23 |
+
return self.execute(audio_file, **kwargs)
|
24 |
+
|
25 |
+
# Otherwise run in a separate process
|
26 |
+
pool = self.diarization_context.get_pool()
|
27 |
+
|
28 |
+
try:
|
29 |
+
result = pool.apply(self.execute, (audio_file,), kwargs)
|
30 |
+
return result
|
31 |
+
finally:
|
32 |
+
self.diarization_context.return_pool(pool)
|
33 |
+
|
34 |
+
def mark_speakers(self, diarization_result: List[DiarizationEntry], whisper_result: dict):
|
35 |
+
if self.model is not None:
|
36 |
+
return self.model.mark_speakers(diarization_result, whisper_result)
|
37 |
+
|
38 |
+
# Create a new diarization model (calling mark_speakers will not initialize pyannote.audio)
|
39 |
+
model = Diarization(self.auth_token)
|
40 |
+
return model.mark_speakers(diarization_result, whisper_result)
|
41 |
+
|
42 |
+
def get_model(self):
|
43 |
+
# Lazy load the model
|
44 |
+
if (self.model is None):
|
45 |
+
if self.cache:
|
46 |
+
print("Loading diarization model from cache")
|
47 |
+
self.model = self.cache.get("diarization", lambda : Diarization(self.auth_token))
|
48 |
+
else:
|
49 |
+
print("Loading diarization model")
|
50 |
+
self.model = Diarization(self.auth_token)
|
51 |
+
return self.model
|
52 |
+
|
53 |
+
def execute(self, audio_file, **kwargs):
|
54 |
+
model = self.get_model()
|
55 |
+
|
56 |
+
# We must use list() here to force the iterator to run, as generators are not picklable
|
57 |
+
result = list(model.run(audio_file, **kwargs))
|
58 |
+
return result
|
59 |
+
|
60 |
+
def cleanup(self):
|
61 |
+
if self.diarization_context is not None:
|
62 |
+
self.diarization_context.close()
|
63 |
+
|
64 |
+
def __getstate__(self):
|
65 |
+
return {
|
66 |
+
"auth_token": self.auth_token,
|
67 |
+
"enable_daemon_process": self.enable_daemon_process,
|
68 |
+
"auto_cleanup_timeout_seconds": self.auto_cleanup_timeout_seconds
|
69 |
+
}
|
70 |
+
|
71 |
+
def __setstate__(self, state):
|
72 |
+
self.auth_token = state["auth_token"]
|
73 |
+
self.enable_daemon_process = state["enable_daemon_process"]
|
74 |
+
self.auto_cleanup_timeout_seconds = state["auto_cleanup_timeout_seconds"]
|
75 |
+
self.diarization_context = None
|
76 |
+
self.cache = GLOBAL_MODEL_CACHE
|
77 |
+
self.model = None
|