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",
    "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": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['pyannote/overlapped-speech-detection',\n",
       " 'pyannote/speaker-diarization',\n",
       " 'pyannote/voice-activity-detection']"
      ]
     },
     "execution_count": 1,
     "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",
    "available_pipelines"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's load the speaker diarization pipeline..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pyannote.audio import Pipeline\n",
    "pipeline = Pipeline.from_pretrained(\"pyannote/speaker-diarization\")"
   ]
  },
  {
   "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": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "ROOT_DIR = \"/Users/bredin/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": 4,
   "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": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 6.7  7.1 SPEAKER_01\n",
      " 7.5  8.2 SPEAKER_00\n",
      " 8.2 10.0 SPEAKER_01\n",
      " 9.9 11.0 SPEAKER_00\n",
      "10.6 14.7 SPEAKER_01\n",
      "14.3 17.9 SPEAKER_00\n",
      "18.0 21.5 SPEAKER_01\n",
      "18.1 18.6 SPEAKER_00\n",
      "21.7 28.5 SPEAKER_00\n",
      "27.8 29.7 SPEAKER_01\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": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<pyannote.core.annotation.Annotation at 0x134c7f1f0>"
      ]
     },
     "execution_count": 6,
     "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": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<pyannote.core.annotation.Annotation at 0x1342d12b0>"
      ]
     },
     "execution_count": 7,
     "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": 8,
   "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": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<pyannote.core.annotation.Annotation at 0x142cc3220>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vad = Pipeline.from_pretrained(\"pyannote/voice-activity-detection\")\n",
    "vad(audio_in_memory)"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "41379f2c2a4eb17f5ac9a1f5014f4b793a0ead0b6469d8877f81a91eb030f53e"
  },
  "kernelspec": {
   "display_name": "Python 3.8.2 64-bit ('pyannote': conda)",
   "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": "3.8.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}