tutorials/applying_a_pipeline.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Applying a pretrained pipeline\n",
    "\n",
    "In this tutorial, you will learn how to apply `pyannote.audio` pipelines on an audio file.\n",
    "\n",
    "A pipeline takes an audio file as input and returns a labeled temporal segmentation of the audio file. \n",
    "\n",
    "More precisely, it usually applies a pretrained model (= neural network) on the audio file, post-processes the output of the model, and returns its output as a [`pyannote.core.Annotation`](http://pyannote.github.io/pyannote-core/structure.html#annotation) instance. It should become clearer as you keep reading..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Loading pipeline from 🤗 hub\n",
    "\n",
    "A bunch of pretrained pipelines are available on [🤗 Huggingface model hub](https://hf.co/models?other=pyannote-audio-pipeline) and can be listed by looking for the [`pyannote-audio-pipeline`](https://hf.co/models?other=pyannote-audio-pipeline) tag."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['pyannote/overlapped-speech-detection',\n",
       " 'pyannote/speaker-diarization',\n",
       " 'pyannote/speaker-segmentation',\n",
       " 'pyannote/voice-activity-detection']"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from huggingface_hub import HfApi\n",
    "available_pipelines = [p.modelId for p in HfApi().list_models(filter=\"pyannote-audio-pipeline\")]\n",
    "list(filter(lambda p: p.startswith(\"pyannote/\"), available_pipelines))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Official [pyannote.audio](https://github.com/pyannote/pyannote-audio) pipelines (i.e. those under the [`pyannote` organization](https://hf.co/pyannote) umbrella) are open-source, but gated. It means that you have to first accept users conditions on their respective Huggingface page to access the pretrained weights and hyper-parameters. Despite this initial process, those pipelines can perfectly be downloaded for later offline use: keep reading this tutorial until the end to learn how to do that.\n",
    "\n",
    "For instance, to load the speaker diarization pipeline used in this tutorial, you have to visit [hf.co/pyannote/speaker-diarization](https://hf.co/pyannote/speaker-diarization), accept the terms, visit [hf.co/pyannote/segmentation](https://hf.co/pyannote/segmentation) (used internally by the speaker diarization pipeline), accept the terms, and log in using `notebook_login` below:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9466934e67254fe6a7ff67b727a3f3ab",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "VBox(children=(HTML(value='<center> <img\\nsrc=https://huggingface.co/front/assets/huggingface_logo-noborder.sv…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from huggingface_hub import notebook_login\n",
    "notebook_login()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Once authenticated, you can load the pipeline (and the internal models)..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pyannote.audio import Pipeline\n",
    "pipeline = Pipeline.from_pretrained(\"pyannote/speaker-diarization@develop\", use_auth_token=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Processing a file from disk"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "... and apply it to an audio file.  \n",
    "\n",
    "The pipeline will automatically use GPUs when available. \n",
    "On CPU it might take a long while (up to 10x RT)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "ROOT_DIR = \"/Users/hbredin/Development/pyannote/pyannote-audio\"\n",
    "AUDIO_FILE = f\"{ROOT_DIR}/tutorials/assets/sample.wav\"\n",
    "dia = pipeline(AUDIO_FILE)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualizing the output\n",
    "\n",
    "Most pipelines return a [`pyannote.core.Annotation`](http://pyannote.github.io/pyannote-core/structure.html#annotation) instance..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pyannote.core import Annotation\n",
    "assert isinstance(dia, Annotation)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "... whose [API](https://pyannote.github.io/pyannote-core/structure.html#annotation) you can use to print the result:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 6.7  7.1 SPEAKER_02\n",
      " 7.6  8.3 SPEAKER_00\n",
      " 8.3 14.8 SPEAKER_02\n",
      " 9.8 11.0 SPEAKER_00\n",
      "14.3 18.5 SPEAKER_00\n",
      "17.9 21.5 SPEAKER_01\n",
      "21.7 28.6 SPEAKER_00\n",
      "27.8 29.5 SPEAKER_02\n"
     ]
    }
   ],
   "source": [
    "for speech_turn, track, speaker in dia.itertracks(yield_label=True):\n",
    "    print(f\"{speech_turn.start:4.1f} {speech_turn.end:4.1f} {speaker}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If you happen to be running this example in a _Jupyter notebook_, `dia` can be [visualized directly](http://pyannote.github.io/pyannote-core/visualization.html):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<pyannote.core.annotation.Annotation at 0x285261280>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# we visualize [0, 30] time range\n",
    "from pyannote.core import notebook, Segment\n",
    "notebook.crop = Segment(0, 30)\n",
    "dia"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "When available, the reference annotation can be visualized too, for comparison:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<pyannote.core.annotation.Annotation at 0x288b56df0>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from pyannote.database.util import load_rttm\n",
    "REFERENCE = f\"{ROOT_DIR}/tutorials/assets/sample.rttm\"\n",
    "reference = load_rttm(REFERENCE)[\"sample\"]\n",
    "\n",
    "# map hypothesized and reference speakers for visualization purposes\n",
    "pipeline.optimal_mapping(dia, reference)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Processing a file from memory\n",
    "\n",
    "In case the audio file is not stored on disk, pipelines can also process audio provided as a `{\"waveform\": ..., \"sample_rate\": ...}` dictionary. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "type(waveform)=<class 'torch.Tensor'>\n",
      "waveform.shape=torch.Size([1, 480000])\n",
      "waveform.dtype=torch.float32\n"
     ]
    }
   ],
   "source": [
    "import torchaudio\n",
    "waveform, sample_rate = torchaudio.load(AUDIO_FILE)\n",
    "\n",
    "print(f\"{type(waveform)=}\")\n",
    "print(f\"{waveform.shape=}\")\n",
    "print(f\"{waveform.dtype=}\")\n",
    "\n",
    "audio_in_memory = {\"waveform\": waveform, \"sample_rate\": sample_rate}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<pyannote.core.annotation.Annotation at 0x288d4db80>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vad = Pipeline.from_pretrained(\"pyannote/voice-activity-detection\", use_auth_token=True)\n",
    "vad(audio_in_memory)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Offline use\n",
    "\n",
    "Gating models and pipelines allows [me](https://herve.niderb.fr) to know a bit more about `pyannote.audio` user base and eventually help me write grant proposals to make `pyannote.audio` even better. Please fill this form as precisely as possible. \n",
    "\n",
    "For instance, before gating `pyannote/speaker-diarization`, I had no idea that so many people were relying on it in production. Hint: sponsors are more than welcome! maintaining open source libraries is time consuming.\n",
    "\n",
    "That being said: this whole authentication process does not prevent you from using official `pyannote.audio` models and pipelines offline (i.e. without going through the authentication process in every `docker run ...` or whatever you are using in production).\n",
    "\n",
    "* Step 1: download `config.yaml` of [`pyannote/voice-activity-detection`](https://hf.co/pyannote/voice-activity-detection) pipeline\n",
    "\n",
    "![](assets/download-pipeline.png)\n",
    "\n",
    "* Step 2: download the `pytorch_model.bin` model\n",
    "\n",
    "![](assets/download-model.png)\n",
    "\n",
    "* Step 3: edit `config.yaml` to point to the local model\n",
    "\n",
    "```diff\n",
    "pipeline:\n",
    "  name: pyannote.audio.pipelines.VoiceActivityDetection\n",
    "  params:\n",
    "-    segmentation: pyannote/segmentation@Interspeech2021\n",
    "+    segmentation: pytorch_model.bin\n",
    "\n",
    "params:\n",
    "  min_duration_off: 0.09791355693027545\n",
    "  min_duration_on: 0.05537587440407595\n",
    "  offset: 0.4806866463041527\n",
    "  onset: 0.8104268538848918\n",
    "```\n",
    "\n",
    "* Step 4: load the pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<pyannote.core.annotation.Annotation at 0x288d1e940>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# look ma: no hands!\n",
    "offline_vad = Pipeline.from_pretrained(\"config.yaml\")\n",
    "offline_vad(audio_in_memory)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# just checking output is the same\n",
    "assert (vad(audio_in_memory) == offline_vad(audio_in_memory))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.9.13 ('pyannote-mps')",
   "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": "3.9.13"
  },
  "vscode": {
   "interpreter": {
    "hash": "36a3a48a52702f18671693adf589423ec3f7db45d50f6ee539f1b0696bb58d43"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}