silero-vad/silero-vad.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "heading_collapsed": true,
    "id": "62A6F_072Fwq"
   },
   "source": [
    "## Install Dependencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "hidden": true,
    "id": "5w5AkskZ2Fwr"
   },
   "outputs": [],
   "source": [
    "#@title Install and Import Dependencies\n",
    "\n",
    "# this assumes that you have a relevant version of PyTorch installed\n",
    "!pip install -q torchaudio\n",
    "\n",
    "SAMPLING_RATE = 16000\n",
    "\n",
    "import torch\n",
    "torch.set_num_threads(1)\n",
    "\n",
    "from IPython.display import Audio\n",
    "from pprint import pprint\n",
    "# download example\n",
    "torch.hub.download_url_to_file('https://models.silero.ai/vad_models/en.wav', 'en_example.wav')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "pSifus5IilRp"
   },
   "outputs": [],
   "source": [
    "USE_PIP = True # download model using pip package or torch.hub\n",
    "USE_ONNX = False # change this to True if you want to test onnx model\n",
    "if USE_ONNX:\n",
    "    !pip install -q onnxruntime\n",
    "if USE_PIP:\n",
    "  !pip install -q silero-vad\n",
    "  from silero_vad import (load_silero_vad,\n",
    "                          read_audio,\n",
    "                          get_speech_timestamps,\n",
    "                          save_audio,\n",
    "                          VADIterator,\n",
    "                          collect_chunks)\n",
    "  model = load_silero_vad(onnx=USE_ONNX)\n",
    "else:\n",
    "  model, utils = torch.hub.load(repo_or_dir='snakers4/silero-vad',\n",
    "                                model='silero_vad',\n",
    "                                force_reload=True,\n",
    "                                onnx=USE_ONNX)\n",
    "\n",
    "  (get_speech_timestamps,\n",
    "  save_audio,\n",
    "  read_audio,\n",
    "  VADIterator,\n",
    "  collect_chunks) = utils"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "fXbbaUO3jsrw"
   },
   "source": [
    "## Speech timestapms from full audio"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "aI_eydBPjsrx"
   },
   "outputs": [],
   "source": [
    "wav = read_audio('en_example.wav', sampling_rate=SAMPLING_RATE)\n",
    "# get speech timestamps from full audio file\n",
    "speech_timestamps = get_speech_timestamps(wav, model, sampling_rate=SAMPLING_RATE)\n",
    "pprint(speech_timestamps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "OuEobLchjsry"
   },
   "outputs": [],
   "source": [
    "# merge all speech chunks to one audio\n",
    "save_audio('only_speech.wav',\n",
    "           collect_chunks(speech_timestamps, wav), sampling_rate=SAMPLING_RATE)\n",
    "Audio('only_speech.wav')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "zeO1xCqxUC6w"
   },
   "source": [
    "## Entire audio inference"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "LjZBcsaTT7Mk"
   },
   "outputs": [],
   "source": [
    "wav = read_audio('en_example.wav', sampling_rate=SAMPLING_RATE)\n",
    "# audio is being splitted into 31.25 ms long pieces\n",
    "# so output length equals ceil(input_length * 31.25 / SAMPLING_RATE)\n",
    "predicts = model.audio_forward(wav, sr=SAMPLING_RATE)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "iDKQbVr8jsry"
   },
   "source": [
    "## Stream imitation example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "q-lql_2Wjsry"
   },
   "outputs": [],
   "source": [
    "## using VADIterator class\n",
    "\n",
    "vad_iterator = VADIterator(model, sampling_rate=SAMPLING_RATE)\n",
    "wav = read_audio(f'en_example.wav', sampling_rate=SAMPLING_RATE)\n",
    "\n",
    "window_size_samples = 512 if SAMPLING_RATE == 16000 else 256\n",
    "for i in range(0, len(wav), window_size_samples):\n",
    "    chunk = wav[i: i+ window_size_samples]\n",
    "    if len(chunk) < window_size_samples:\n",
    "      break\n",
    "    speech_dict = vad_iterator(chunk, return_seconds=True)\n",
    "    if speech_dict:\n",
    "        print(speech_dict, end=' ')\n",
    "vad_iterator.reset_states() # reset model states after each audio"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "BX3UgwwB2Fwv"
   },
   "outputs": [],
   "source": [
    "## just probabilities\n",
    "\n",
    "wav = read_audio('en_example.wav', sampling_rate=SAMPLING_RATE)\n",
    "speech_probs = []\n",
    "window_size_samples = 512 if SAMPLING_RATE == 16000 else 256\n",
    "for i in range(0, len(wav), window_size_samples):\n",
    "    chunk = wav[i: i+ window_size_samples]\n",
    "    if len(chunk) < window_size_samples:\n",
    "      break\n",
    "    speech_prob = model(chunk, SAMPLING_RATE).item()\n",
    "    speech_probs.append(speech_prob)\n",
    "vad_iterator.reset_states() # reset model states after each audio\n",
    "\n",
    "print(speech_probs[:10]) # first 10 chunks predicts"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "name": "silero-vad.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}