Add Language Detector

silero-vad.ipynb

@@ -312,6 +312,78 @@
     "Audio('no_numbers.wav')"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {
+    "heading_collapsed": true
+   },
+   "source": [
+    "## Language detector"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {
+    "heading_collapsed": true,
+    "hidden": true
+   },
+   "source": [
+    "### Install Dependencies"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "hidden": true
+   },
+   "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 soundfile\n",
+    "\n",
+    "import glob\n",
+    "import torch\n",
+    "torch.set_num_threads(1)\n",
+    "\n",
+    "from IPython.display import Audio\n",
+    "from pprint import pprint\n",
+    "\n",
+    "model, utils = torch.hub.load(repo_or_dir='snakers4/silero-vad',\n",
+    "                              model='silero_lang_detector',\n",
+    "                              force_reload=True)\n",
+    "\n",
+    "(get_language,\n",
+    " read_audio) = utils\n",
+    "\n",
+    "files_dir = torch.hub.get_dir() + '/snakers4_silero-vad_master/files'"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {
+    "heading_collapsed": true,
+    "hidden": true
+   },
+   "source": [
+    "### Full audio"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "hidden": true
+   },
+   "outputs": [],
+   "source": [
+    "wav = read_audio(f'{files_dir}/en.wav')\n",
+    "lang = get_language(wav, model)\n",
+    "print(lang)"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {
@@ -677,6 +749,98 @@
     "           drop_chunks(number_timestamps, wav), 16000) \n",
     "Audio('no_numbers.wav')"
    ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {
+    "heading_collapsed": true
+   },
+   "source": [
+    "## Language detector"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {
+    "heading_collapsed": true,
+    "hidden": true,
+    "id": "bL4kn4KJrlyL"
+   },
+   "source": [
+    "### Install Dependencies"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2020-12-30T17:25:19.107534Z",
+     "start_time": "2020-12-30T17:24:51.853293Z"
+    },
+    "cellView": "form",
+    "hidden": true,
+    "id": "Q4QIfSpprnkI"
+   },
+   "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 soundfile onnxruntime\n",
+    "\n",
+    "import glob\n",
+    "import torch\n",
+    "import onnxruntime\n",
+    "from pprint import pprint\n",
+    "\n",
+    "from IPython.display import Audio\n",
+    "\n",
+    "_, utils = torch.hub.load(repo_or_dir='snakers4/silero-vad',\n",
+    "                              model='silero_lang_detector',\n",
+    "                              force_reload=True)\n",
+    "\n",
+    "(get_language,\n",
+    " read_audio) = utils\n",
+    "\n",
+    "files_dir = torch.hub.get_dir() + '/snakers4_silero-vad_master/files'\n",
+    "\n",
+    "def init_onnx_model(model_path: str):\n",
+    "    return onnxruntime.InferenceSession(model_path)\n",
+    "\n",
+    "def validate_onnx(model, inputs):\n",
+    "    with torch.no_grad():\n",
+    "        ort_inputs = {'input': inputs.cpu().numpy()}\n",
+    "        outs = model.run(None, ort_inputs)\n",
+    "        outs = [torch.Tensor(x) for x in outs]\n",
+    "    return outs"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {
+    "heading_collapsed": true,
+    "hidden": true,
+    "id": "5JHErdB7jsr0"
+   },
+   "source": [
+    "### Full Audio"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "hidden": true
+   },
+   "outputs": [],
+   "source": [
+    "model = init_onnx_model(f'{files_dir}/number_detector.onnx')\n",
+    "wav = read_audio(f'{files_dir}/en.wav')\n",
+    "\n",
+    "lang = get_language(wav, model, run_function=validate_onnx)\n",
+    "print(lang)"
+   ]
   }
  ],
  "metadata": {