deepgeek/silero-vad
2
0
Fork 0
mirror of https://github.com/snakers4/silero-vad.git synced 2026-02-04 09:29:22 +08:00
Code Issues Packages Projects Releases Wiki Activity

use a process specific copy of model

Browse Source
This commit is contained in:
Ziyuan Wang
2023-03-10 16:09:07 +00:00
parent 17903cb41d
commit 55c41abf46
1 changed files with 16 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
examples/parallel_example.ipynb
View File

@@ -14,7 +14,7 @@
"metadata": {},
"outputs": [],
"source": [
"!pip install -q torchaudio\n",
"# !pip install -q torchaudio\n",
"SAMPLING_RATE = 16000\n",
"import torch\n",
"from pprint import pprint\n",
@@ -67,17 +67,27 @@
"metadata": {},
"outputs": [],
"source": [
"import multiprocessing\n",
"\n",
"vad_models = dict()\n",
"\n",
"def init_model(model):\n",
" global vad_model\n",
" vad_model = model\n",
" pid = multiprocessing.current_process().pid\n",
" model, _ = torch.hub.load(repo_or_dir='snakers4/silero-vad',\n",
" model='silero_vad',\n",
" force_reload=False,\n",
" onnx=False)\n",
" vad_models[pid] = model\n",
"\n",
"def vad_process(audio_file: str):\n",
" global vad_model\n",
" \n",
" pid = multiprocessing.current_process().pid\n",
" \n",
" with torch.no_grad():\n",
" wav = read_audio(audio_file, sampling_rate=SAMPLING_RATE)\n",
" return get_speech_timestamps(\n",
" wav,\n",
" vad_model,\n",
" vad_models[pid],\n",
" 0.46, # speech prob threshold\n",
" 16000, # sample rate\n",
" 300, # min speech duration in ms\n",
@@ -132,8 +142,7 @@
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.9.15"
},
"orig_nbformat": 4
}
},
"nbformat": 4,
"nbformat_minor": 2