old/DataAugTest.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "gnZoCx3MlMDZ"
   },
   "source": [
    "# AWS-LSTM for Audio-Language Model with Fast.ai\n",
    "by @ste, @aamir7117 ,@zachcaceres\n",
    "\n",
    "\n",
    "### TODO:\n",
    "- full data set\n",
    "- pretrained=False\n",
    "- unfreeze latter layers\n",
    "- true validation holdout (by speaker)\n",
    "- augmentation\n",
    "- make spectrogram a square"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# %%bash\n",
    "# pip install pydub --user\n",
    "# sudo apt-get install ffmpeg\n",
    "# sudo apt-get install sox libsox-dev libsox-fmt-all\n",
    "# git clone https://github.com/pytorch/audio.git torchaudio\n",
    "# cd torchaudio\n",
    "# python setup.py install\n",
    "# %%"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "a-BiaLFumISP"
   },
   "source": [
    "## Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "7O0URnSZl1_8"
   },
   "outputs": [],
   "source": [
    "import glob\n",
    "import matplotlib.pyplot as plt\n",
    "from fastai.text import * \n",
    "from fastai.vision import *\n",
    "from pathlib import Path\n",
    "import pandas as pd\n",
    "import torchaudio\n",
    "from IPython.display import Audio\n",
    "from torchaudio import transforms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 204
    },
    "colab_type": "code",
    "id": "t-wSSSeBm1Wr",
    "outputId": "d116f19c-bdae-4164-b296-2bbd10f6a2c5"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/envs/fastai-audio/lib/python3.7/site-packages/fastai/datasets.py:164: YAMLLoadWarning: calling yaml.load() without Loader=... is deprecated, as the default Loader is unsafe. Please read https://msg.pyyaml.org/load for full details.\n",
      "  with open(fpath, 'r') as yaml_file: return yaml.load(yaml_file)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>label</th>\n",
       "      <th>text</th>\n",
       "      <th>is_valid</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>negative</td>\n",
       "      <td>Un-bleeping-believable! Meg Ryan doesn't even ...</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>positive</td>\n",
       "      <td>This is a extremely well-made film. The acting...</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>negative</td>\n",
       "      <td>Every once in a long while a movie will come a...</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>positive</td>\n",
       "      <td>Name just says it all. I watched this movie wi...</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>negative</td>\n",
       "      <td>This movie succeeds at being one of the most u...</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      label                                               text  is_valid\n",
       "0  negative  Un-bleeping-believable! Meg Ryan doesn't even ...     False\n",
       "1  positive  This is a extremely well-made film. The acting...     False\n",
       "2  negative  Every once in a long while a movie will come a...     False\n",
       "3  positive  Name just says it all. I watched this movie wi...     False\n",
       "4  negative  This movie succeeds at being one of the most u...     False"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "path = untar_data(URLs.IMDB_SAMPLE)\n",
    "df = pd.read_csv(path/'texts.csv')\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "3ON247uFnM2K"
   },
   "outputs": [],
   "source": [
    "# Language model data\n",
    "data_lm = TextLMDataBunch.from_csv(path, 'texts.csv')\n",
    "# Classifier model data\n",
    "data_clas = TextClasDataBunch.from_csv(path, 'texts.csv', vocab=data_lm.train_ds.vocab, bs=32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "GSG9v1XSnTIc"
   },
   "outputs": [],
   "source": [
    "data_lm.save('data_lm_export.pkl')\n",
    "data_clas.save('data_clas_export.pkl')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 462
    },
    "colab_type": "code",
    "id": "26hAYZQjnVjT",
    "outputId": "926f7307-6203-40f2-9f1e-298ca70a5803"
   },
   "outputs": [],
   "source": [
    "data_lm = load_data(path, fname='data_lm_export.pkl')\n",
    "data_clas = load_data(path, fname='data_clas_export.pkl', bs=16)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "jagj8sfImgo7"
   },
   "source": [
    "## Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "by6abtVDmh8i"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/envs/fastai-audio/lib/python3.7/site-packages/fastai/datasets.py:164: YAMLLoadWarning: calling yaml.load() without Loader=... is deprecated, as the default Loader is unsafe. Please read https://msg.pyyaml.org/load for full details.\n",
      "  with open(fpath, 'r') as yaml_file: return yaml.load(yaml_file)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "Total time: 00:12 <p><table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>4.908524</td>\n",
       "      <td>4.108302</td>\n",
       "      <td>0.273274</td>\n",
       "      <td>00:12</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn = language_model_learner(data_lm, AWD_LSTM)\n",
    "learn.fit_one_cycle(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "lr_find(learn)\n",
    "learn.recorder.plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "Total time: 00:37 <p><table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>4.746801</td>\n",
       "      <td>4.010424</td>\n",
       "      <td>0.282708</td>\n",
       "      <td>00:12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>4.632045</td>\n",
       "      <td>3.949140</td>\n",
       "      <td>0.286280</td>\n",
       "      <td>00:12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>4.530967</td>\n",
       "      <td>3.940040</td>\n",
       "      <td>0.286086</td>\n",
       "      <td>00:12</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.fit_one_cycle(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "id": "NuQUUtTAnYG7",
    "outputId": "72a19df7-3ceb-4a26-a4af-8176e498df5f"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'What do you want ? Why do you not win these awards that homage ? Is it ? i must , i am always interested in out of a letter , guilty'"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "learn.predict(\"What do you want\", n_words=30)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## With Timit\n",
    "[dataset](https://github.com/philipperemy/timit)\n",
    "https://stackoverflow.com/questions/44748258/reading-a-wav-file-from-timit-database-in-python"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "timit_dir = Path('./timit')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "(timit_dir/'TRAIN/DR1/MDPK0/').ls()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "similar_and_sil = {\n",
    "    'ao': 'aa',\n",
    "    'hv': 'hh',\n",
    "    'ix': 'ih',\n",
    "    'el': 'l',\n",
    "    'em': 'm',\n",
    "    'en': 'n',\n",
    "    'nx': 'n',\n",
    "    'eng': 'ng',\n",
    "    'ux': 'uw',\n",
    "    'pcl': 'sil',\n",
    "    'tcl': 'sil',\n",
    "    'kcl': 'sil',\n",
    "    'bcl': 'sil',\n",
    "    'dcl': 'sil',\n",
    "    'gcl': 'sil',\n",
    "    'h#':  'sil',\n",
    "    'pau': 'sil',\n",
    "    'epi': 'sil'\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(PosixPath('timit/TRAIN/DR1/MDPK0/SA1.PHN'),\n",
       " PosixPath('timit/TRAIN/DR1/MDPK0/SA1.TXT'))"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sorted((timit_dir/'TRAIN/DR1/MDPK0/').ls())\n",
    "phn, txt, wav = sorted((timit_dir/'TRAIN/DR1/MDPK0/').ls())[:3]\n",
    "phn, txt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "                <audio controls=\"controls\" >\n",
       "                    <source src=\"data:audio/wav;base64,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\" type=\"audio/wav\" />\n",
       "                    Your browser does not support the audio element.\n",
       "                </audio>\n",
       "              "
      ],
      "text/plain": [
       "<IPython.lib.display.Audio object>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sig,sr = torchaudio.load(str(timit_dir/'TRAIN/DR1/MDPK0/SA1.WAV'))\n",
    "Audio(data=sig, rate=sr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "                <audio controls=\"controls\" >\n",
       "                    <source src=\"data:audio/wav;base64,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\" type=\"audio/wav\" />\n",
       "                    Your browser does not support the audio element.\n",
       "                </audio>\n",
       "              "
      ],
      "text/plain": [
       "<IPython.lib.display.Audio object>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Audio(data=sig[-1][11950:15665], rate=sr)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "def process_phn_file(p_file, sig, sr, delimiter=' '):\n",
    "    df = pd.read_csv(p_file, delimiter=delimiter, header=None)\n",
    "    df.columns = ['Start', 'End', 'Phn']\n",
    "    df['Sample'] = df.apply(lambda x : sig[-1][x['Start']: x['End']], axis=1)\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "phns = []\n",
    "final = pd.DataFrame()\n",
    "\n",
    "for phn_file in glob.glob(str(timit_dir/'**/*.PHN'), recursive=True)[:24122]:\n",
    "    sig,sr = torchaudio.load(str(phn_file.replace('PHN', 'WAV')))\n",
    "    df = process_phn_file(phn_file, sig, sr, delimiter=' ')\n",
    "    df['Source'] = phn_file\n",
    "    final = final.append(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Start</th>\n",
       "      <th>End</th>\n",
       "      <th>Phn</th>\n",
       "      <th>Sample</th>\n",
       "      <th>Source</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>2320</td>\n",
       "      <td>h#</td>\n",
       "      <td>[tensor(3.0518e-05), tensor(3.0518e-05), tenso...</td>\n",
       "      <td>transcribe/nbs/AWDLSTM/timit/TRAIN/DR6/MEAL0/S...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2320</td>\n",
       "      <td>4409</td>\n",
       "      <td>s</td>\n",
       "      <td>[tensor(0.0007), tensor(-0.0003), tensor(-0.00...</td>\n",
       "      <td>transcribe/nbs/AWDLSTM/timit/TRAIN/DR6/MEAL0/S...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4409</td>\n",
       "      <td>5240</td>\n",
       "      <td>ah</td>\n",
       "      <td>[tensor(0.0009), tensor(0.0015), tensor(0.0011...</td>\n",
       "      <td>transcribe/nbs/AWDLSTM/timit/TRAIN/DR6/MEAL0/S...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5240</td>\n",
       "      <td>6560</td>\n",
       "      <td>m</td>\n",
       "      <td>[tensor(-0.0026), tensor(-0.0023), tensor(-0.0...</td>\n",
       "      <td>transcribe/nbs/AWDLSTM/timit/TRAIN/DR6/MEAL0/S...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6560</td>\n",
       "      <td>6893</td>\n",
       "      <td>w</td>\n",
       "      <td>[tensor(0.0034), tensor(0.0039), tensor(0.0042...</td>\n",
       "      <td>transcribe/nbs/AWDLSTM/timit/TRAIN/DR6/MEAL0/S...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Start   End Phn                                             Sample  \\\n",
       "0      0  2320  h#  [tensor(3.0518e-05), tensor(3.0518e-05), tenso...   \n",
       "1   2320  4409   s  [tensor(0.0007), tensor(-0.0003), tensor(-0.00...   \n",
       "2   4409  5240  ah  [tensor(0.0009), tensor(0.0015), tensor(0.0011...   \n",
       "3   5240  6560   m  [tensor(-0.0026), tensor(-0.0023), tensor(-0.0...   \n",
       "4   6560  6893   w  [tensor(0.0034), tensor(0.0039), tensor(0.0042...   \n",
       "\n",
       "                                              Source  \n",
       "0  transcribe/nbs/AWDLSTM/timit/TRAIN/DR6/MEAL0/S...  \n",
       "1  transcribe/nbs/AWDLSTM/timit/TRAIN/DR6/MEAL0/S...  \n",
       "2  transcribe/nbs/AWDLSTM/timit/TRAIN/DR6/MEAL0/S...  \n",
       "3  transcribe/nbs/AWDLSTM/timit/TRAIN/DR6/MEAL0/S...  \n",
       "4  transcribe/nbs/AWDLSTM/timit/TRAIN/DR6/MEAL0/S...  "
      ]
     },
     "execution_count": 227,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## DataBlock"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "metadata": {},
   "outputs": [],
   "source": [
    "class AudioDataBunch(DataBunch):\n",
    "    def hear_ex(self, rows:int=3, ds_type:DatasetType=DatasetType.Valid, **kwargs):\n",
    "        batch = self.dl(ds_type).dataset[:rows]\n",
    "        self.train_ds.hear_xys(batch.x, batch.y, **kwargs)\n",
    "\n",
    "MIN_SAMPLE_SIZE = 201\n",
    "class AudioItem(ItemBase):\n",
    "    def __init__(self, signal, data=None, **kwargs):\n",
    "        self.signal = signal\n",
    "        self.sr = 16000\n",
    "        self.kwargs = kwargs\n",
    "        if (self.signal.shape[0] < MIN_SAMPLE_SIZE):\n",
    "            self.signal = F.pad(self.signal, (0, MIN_SAMPLE_SIZE - self.signal.shape[0]), value=0)\n",
    "\n",
    "        self.signal = self.signal[None,:] # add channel to signal\n",
    "        self.data = transforms.MelSpectrogram(sr=self.sr, **kwargs)(self.signal)\n",
    "\n",
    "    def __str__(self): return f'Duration: {self.duration} seconds.'\n",
    "    def __len__(self): return self.signal.shape[1]\n",
    "    def _repr_html_(self): return f'{self.__str__()}<br />{self.ipy_audio._repr_html_()}'\n",
    "    \n",
    "    def show(self, title:Optional[str]=None, **kwargs):\n",
    "        \"Show sound on `ax` with `title`, using `cmap` if single-channel, overlaid with optional `y`\"\n",
    "        self.hear(title=title)\n",
    "\n",
    "    def hear(self, title=None):\n",
    "        if title is not None: print(title)\n",
    "        display(self.ipy_audio)\n",
    "\n",
    "    def apply_tfms(self, tfms):\n",
    "        for tfm in tfms:\n",
    "            self.data = tfm(self.data)\n",
    "        return self\n",
    "        \n",
    "    @property\n",
    "    def shape(self):\n",
    "        return self.data.shape\n",
    "\n",
    "    @property\n",
    "    def size(self)->Tuple[int,int]: return self.data.shape[-2:]\n",
    "\n",
    "    @property\n",
    "    def ipy_audio(self): return Audio(data=self.signal, rate=self.sr)\n",
    "\n",
    "    @property\n",
    "    def duration(self): return len(self)/self.sr\n",
    "\n",
    "    @classmethod\n",
    "    def open(cls, fn, **kwargs):\n",
    "        return AudioItem(fn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'AudioDataBunch' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-14-9897e8fe3ee4>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0mAUDIO_EXTENSIONS\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mk\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mk\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mv\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mmimetypes\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtypes_map\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mitems\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstartswith\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'audio/'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m;\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m \u001b[0;32mclass\u001b[0m \u001b[0mAudioList\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mItemList\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      5\u001b[0m     \u001b[0m_bunch\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mAudioDataBunch\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<ipython-input-14-9897e8fe3ee4>\u001b[0m in \u001b[0;36mAudioList\u001b[0;34m()\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0;32mclass\u001b[0m \u001b[0mAudioList\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mItemList\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 5\u001b[0;31m     \u001b[0m_bunch\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mAudioDataBunch\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      6\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m     \u001b[0;31m# TODO: __REPR__\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'AudioDataBunch' is not defined"
     ]
    }
   ],
   "source": [
    "# These are valid file extensions for audio files\n",
    "AUDIO_EXTENSIONS = set(k for k,v in mimetypes.types_map.items() if v.startswith('audio/'));\n",
    "\n",
    "class AudioList(ItemList):\n",
    "    _bunch = AudioDataBunch\n",
    "    \n",
    "    # TODO: __REPR__    \n",
    "    def __init__(self, *args, **kwargs):\n",
    "        super().__init__(*args, **kwargs)\n",
    "            \n",
    "    def get(self, i): return AudioItem.open(self.items[i])\n",
    "    \n",
    "    def reconstruct(self, t:Tensor): return Image(t.transpose(1,2))\n",
    "\n",
    "    def hear_xys(self, xs, ys, **kwargs):\n",
    "        for x, y in zip(xs, ys): x.hear(title=y, **kwargs)\n",
    "\n",
    "    @classmethod\n",
    "    def from_folder(cls, path:PathOrStr='.', extensions:Collection[str]=None, **kwargs)->ItemList:\n",
    "        extensions = ifnone(extensions, AUDIO_EXTENSIONS)\n",
    "        return super().from_folder(path=path, extensions=extensions, **kwargs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Padding To Max"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'AudioList' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-22-41676bd81e76>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mAudioList\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfrom_df\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfinal\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpath\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'.'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcols\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'Sample'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'AudioList' is not defined"
     ]
    }
   ],
   "source": [
    "data = AudioList.from_df(final, path='.', cols=['Sample'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def _pad_to_max(mx, value, t):\n",
    "    \"\"\"Pad tensor with `value` until it reaches length `mx`\"\"\"\n",
    "    if t.shape[1] == mx: return t\n",
    "    return F.pad(t, (0,0, 0,mx-t.shape[1]), value=value)\n",
    "\n",
    "# Get longest audio clip\n",
    "max_len_of_audio = sorted([a.data.shape[1] for a in data])[-1]\n",
    "\n",
    "# # Get a transform function that uses the max length of our audio list and pads with silence (0)\n",
    "pad_to_max = partial(_pad_to_max, max_len_of_audio, 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Databunch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = AudioList.from_df(final, path='.', cols=['Sample']).split_by_rand_pct(0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Normal datablock setup from our AudioList from above.\n",
    "data = (data\n",
    "    .label_from_df('Phn')\n",
    "    .transform([[pad_to_max], [pad_to_max]])\n",
    "    .databunch())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "data.get(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "Image.show(data.get(0).data.sha, figsize=(10,10))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "yJCcrRGUpa7-"
   },
   "source": [
    "## CNN Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 234,
   "metadata": {},
   "outputs": [],
   "source": [
    "newlayer = nn.Conv2d(1, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)\n",
    "newlayer.cuda(); # Our layer should use cuda, since the rest of the model will."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 235,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn = cnn_learner(data, models.resnet18, metrics=accuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 236,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Replace our first layer\n",
    "learn.model[0][0] = newlayer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "lr_find(learn)\n",
    "learn.recorder.plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 238,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "    <div>\n",
       "        <style>\n",
       "            /* Turns off some styling */\n",
       "            progress {\n",
       "                /* gets rid of default border in Firefox and Opera. */\n",
       "                border: none;\n",
       "                /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
       "                background-size: auto;\n",
       "            }\n",
       "            .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
       "                background: #F44336;\n",
       "            }\n",
       "        </style>\n",
       "      <progress value='3' class='' max='5', style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
       "      60.00% [3/5 24:26<16:17]\n",
       "    </div>\n",
       "    \n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>2.250237</td>\n",
       "      <td>2.254788</td>\n",
       "      <td>0.342357</td>\n",
       "      <td>08:08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>1.943848</td>\n",
       "      <td>2.010880</td>\n",
       "      <td>0.370069</td>\n",
       "      <td>08:08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>1.777995</td>\n",
       "      <td>2.039773</td>\n",
       "      <td>0.386527</td>\n",
       "      <td>08:09</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table><p>\n",
       "\n",
       "    <div>\n",
       "        <style>\n",
       "            /* Turns off some styling */\n",
       "            progress {\n",
       "                /* gets rid of default border in Firefox and Opera. */\n",
       "                border: none;\n",
       "                /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
       "                background-size: auto;\n",
       "            }\n",
       "            .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
       "                background: #F44336;\n",
       "            }\n",
       "        </style>\n",
       "      <progress value='0' class='progress-bar-interrupted' max='3015', style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
       "      Interrupted\n",
       "    </div>\n",
       "    "
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Traceback (most recent call last):\n",
      "  File \"/opt/anaconda3/envs/fastai-audio/lib/python3.7/multiprocessing/queues.py\", line 242, in _feed\n",
      "    send_bytes(obj)\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-238-b62783a18d94>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mlearn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit_one_cycle\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m5\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mslice\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m1e-2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/opt/anaconda3/envs/fastai-audio/lib/python3.7/site-packages/fastai/train.py\u001b[0m in \u001b[0;36mfit_one_cycle\u001b[0;34m(learn, cyc_len, max_lr, moms, div_factor, pct_start, final_div, wd, callbacks, tot_epochs, start_epoch)\u001b[0m\n\u001b[1;32m     20\u001b[0m     callbacks.append(OneCycleScheduler(learn, max_lr, moms=moms, div_factor=div_factor, pct_start=pct_start,\n\u001b[1;32m     21\u001b[0m                                        final_div=final_div, tot_epochs=tot_epochs, start_epoch=start_epoch))\n\u001b[0;32m---> 22\u001b[0;31m     \u001b[0mlearn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcyc_len\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmax_lr\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwd\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mwd\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcallbacks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mcallbacks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     23\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     24\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mlr_find\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlearn\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mLearner\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstart_lr\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mFloats\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1e-7\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mend_lr\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mFloats\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m10\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnum_it\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mint\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstop_div\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mbool\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwd\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mfloat\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/fastai-audio/lib/python3.7/site-packages/fastai/basic_train.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, epochs, lr, wd, callbacks)\u001b[0m\n\u001b[1;32m    194\u001b[0m         \u001b[0mcallbacks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mcb\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mcb\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcallback_fns\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mlistify\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcallbacks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    195\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mdefaults\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mextra_callbacks\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mcallbacks\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0mdefaults\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mextra_callbacks\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 196\u001b[0;31m         \u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mepochs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmetrics\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmetrics\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcallbacks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcallbacks\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mcallbacks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    197\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    198\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mcreate_opt\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlr\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mFloats\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwd\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mFloats\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0.\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m->\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/fastai-audio/lib/python3.7/site-packages/fastai/basic_train.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(epochs, learn, callbacks, metrics)\u001b[0m\n\u001b[1;32m     98\u001b[0m             \u001b[0;32mfor\u001b[0m \u001b[0mxb\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0myb\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mprogress_bar\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlearn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain_dl\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mparent\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mpbar\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     99\u001b[0m                 \u001b[0mxb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0myb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcb_handler\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mon_batch_begin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mxb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0myb\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 100\u001b[0;31m                 \u001b[0mloss\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mloss_batch\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlearn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mxb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0myb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlearn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mloss_func\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlearn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mopt\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcb_handler\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    101\u001b[0m                 \u001b[0;32mif\u001b[0m \u001b[0mcb_handler\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mon_batch_end\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mloss\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;32mbreak\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    102\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/fastai-audio/lib/python3.7/site-packages/fastai/basic_train.py\u001b[0m in \u001b[0;36mloss_batch\u001b[0;34m(model, xb, yb, loss_func, opt, cb_handler)\u001b[0m\n\u001b[1;32m     35\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mcb_handler\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mon_step_end\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m     \u001b[0mopt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mzero_grad\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     36\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 37\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0mloss\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdetach\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcpu\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     38\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     39\u001b[0m def get_preds(model:nn.Module, dl:DataLoader, pbar:Optional[PBar]=None, cb_handler:Optional[CallbackHandler]=None,\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  File \"/opt/anaconda3/envs/fastai-audio/lib/python3.7/multiprocessing/connection.py\", line 200, in send_bytes\n",
      "    self._send_bytes(m[offset:offset + size])\n",
      "  File \"/opt/anaconda3/envs/fastai-audio/lib/python3.7/multiprocessing/connection.py\", line 404, in _send_bytes\n",
      "    self._send(header + buf)\n",
      "  File \"/opt/anaconda3/envs/fastai-audio/lib/python3.7/multiprocessing/connection.py\", line 368, in _send\n",
      "    n = write(self._handle, buf)\n",
      "BrokenPipeError: [Errno 32] Broken pipe\n"
     ]
    }
   ],
   "source": [
    "learn.fit_one_cycle(5, slice(1e-2))"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "name": "Audio AWD-LSTM",
   "provenance": [],
   "version": "0.3.2"
  },
  "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}