music_genre_classification_SampleNet.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "-fWomgC-kF5f"
   },
   "source": [
    "**Architecture **\n",
    "\n",
    "<img src=\"http://drive.google.com/uc?export=view&id=12JomC2IswVbNGdE0IIvPpUk8vPjP-MBQ\"  alt=\"artchtecture\">\n",
    "\n",
    "\n",
    "---\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "uN7hQRZsDbgI"
   },
   "source": [
    "(1) Importing dependency"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "3lPxjI5BDAkX",
    "outputId": "88280284-3c51-485b-adfa-4c428507fb92"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import keras\n",
    "from keras.models import Sequential, load_model\n",
    "from keras.layers import Dense, Activation, Dropout, Flatten,\\\n",
    "                         Conv1D, MaxPooling1D\n",
    "from keras.layers.normalization import BatchNormalization\n",
    "from keras.callbacks import ModelCheckpoint, EarlyStopping\n",
    "\n",
    "import pandas as pd\n",
    "import librosa\n",
    "from tqdm import tqdm_notebook as tqdm\n",
    "import os\n",
    "\n",
    "import numpy as np\n",
    "np.random.seed(13)\n",
    "import random\n",
    "random.seed(13)\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "% matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "py5KMVLnDZsC"
   },
   "source": [
    "(2) Get Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "base_data_path = '/data/private/SU/bbchip13/brains_on_beats/magnatagatune/'\n",
    "\n",
    "### Make Y data\n",
    "annotations_filename = base_data_path+'annotations_final_top_50_ver_1.csv'\n",
    "df = pd.read_csv(annotations_filename)\n",
    "y_names = df.columns[2:]\n",
    "y_dict = {info[1][0]:info[1][2:].values.astype(np.int32) for info in df.iterrows()}\n",
    "# y_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "base_wav_path = base_data_path+'split_wav/'\n",
    "train_dir = base_wav_path+'train/'\n",
    "val_dir = base_wav_path+'val/'\n",
    "test_dir = base_wav_path+'test/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def crop_random_5sec(filename):\n",
    "    crop_range = 24\n",
    "    offset = np.random.randint(crop_range)\n",
    "    wav_5sec, sr = librosa.load(filename, offset=offset, duration=6, sr=16000)\n",
    "    wav_5sec = np.expand_dims(wav_5sec, axis=1)\n",
    "    return wav_5sec\n",
    "\n",
    "def load_wavs_and_crop(filenames):\n",
    "    return np.asarray([crop_random_5sec(filename) for filename in tqdm(filenames)])\n",
    "    \n",
    "def find_y_by_filename(filename, y_dict):\n",
    "    basename = os.path.basename(filename)\n",
    "    name = os.path.splitext(basename)[0]\n",
    "    y = y_dict[int(name)]\n",
    "    return y\n",
    "\n",
    "def make_y_by_filenames(filenames, y_dict):\n",
    "    return np.asarray([find_y_by_filename(filename, y_dict) \n",
    "                           for filename in filenames])\n",
    "\n",
    "def make_train_data(filenames, y_dict):\n",
    "    x_train = load_wavs_and_crop(filenames)\n",
    "    y_train = make_y_by_filenames(filenames, y_dict)\n",
    "    return x_train, y_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "bWFB1X_4DWyK",
    "outputId": "4882f32e-10cd-4400-db3b-c5c0cae5f263"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Make train data.......\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "23144e6c31e94be9b4cca511af2c8f09",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=18706), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Make validation data.......\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "469f4a5456174596aaace4390b7de576",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=1825), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "(18706, 96000, 1) (18706, 50) (1825, 96000, 1) (1825, 50)\n"
     ]
    }
   ],
   "source": [
    "print('Make train data.......')\n",
    "x_train_wav_filenames = [train_dir+filename for filename in os.listdir(train_dir)\n",
    "                            if filename.endswith('.wav')]#[:1000]\n",
    "x_train, y_train = make_train_data(x_train_wav_filenames, y_dict)\n",
    "\n",
    "print('Make validation data.......')\n",
    "x_val_wav_filenames = [val_dir+filename for filename in os.listdir(val_dir)\n",
    "                            if filename.endswith('.wav')]#[:200]\n",
    "x_val, y_val = make_train_data(x_val_wav_filenames, y_dict)\n",
    "\n",
    "print(x_train.shape, y_train.shape, x_val.shape, y_val.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "12cS85jvDnfS"
   },
   "source": [
    "(3) Create a sequential model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 985
    },
    "colab_type": "code",
    "id": "fs8Heys2Dm30",
    "outputId": "bad14ede-be9c-4a2f-d052-9d9a29a5e437"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv1d_1 (Conv1D)            (None, 32000, 128)        512       \n",
      "_________________________________________________________________\n",
      "batch_normalization_1 (Batch (None, 32000, 128)        512       \n",
      "_________________________________________________________________\n",
      "activation_1 (Activation)    (None, 32000, 128)        0         \n",
      "_________________________________________________________________\n",
      "conv1d_2 (Conv1D)            (None, 32000, 128)        49280     \n",
      "_________________________________________________________________\n",
      "batch_normalization_2 (Batch (None, 32000, 128)        512       \n",
      "_________________________________________________________________\n",
      "activation_2 (Activation)    (None, 32000, 128)        0         \n",
      "_________________________________________________________________\n",
      "max_pooling1d_1 (MaxPooling1 (None, 10666, 128)        0         \n",
      "_________________________________________________________________\n",
      "conv1d_3 (Conv1D)            (None, 10666, 128)        49280     \n",
      "_________________________________________________________________\n",
      "batch_normalization_3 (Batch (None, 10666, 128)        512       \n",
      "_________________________________________________________________\n",
      "activation_3 (Activation)    (None, 10666, 128)        0         \n",
      "_________________________________________________________________\n",
      "max_pooling1d_2 (MaxPooling1 (None, 3555, 128)         0         \n",
      "_________________________________________________________________\n",
      "conv1d_4 (Conv1D)            (None, 3555, 256)         98560     \n",
      "_________________________________________________________________\n",
      "batch_normalization_4 (Batch (None, 3555, 256)         1024      \n",
      "_________________________________________________________________\n",
      "activation_4 (Activation)    (None, 3555, 256)         0         \n",
      "_________________________________________________________________\n",
      "max_pooling1d_3 (MaxPooling1 (None, 1185, 256)         0         \n",
      "_________________________________________________________________\n",
      "conv1d_5 (Conv1D)            (None, 1185, 256)         196864    \n",
      "_________________________________________________________________\n",
      "batch_normalization_5 (Batch (None, 1185, 256)         1024      \n",
      "_________________________________________________________________\n",
      "activation_5 (Activation)    (None, 1185, 256)         0         \n",
      "_________________________________________________________________\n",
      "max_pooling1d_4 (MaxPooling1 (None, 395, 256)          0         \n",
      "_________________________________________________________________\n",
      "conv1d_6 (Conv1D)            (None, 395, 256)          196864    \n",
      "_________________________________________________________________\n",
      "batch_normalization_6 (Batch (None, 395, 256)          1024      \n",
      "_________________________________________________________________\n",
      "activation_6 (Activation)    (None, 395, 256)          0         \n",
      "_________________________________________________________________\n",
      "max_pooling1d_5 (MaxPooling1 (None, 131, 256)          0         \n",
      "_________________________________________________________________\n",
      "conv1d_7 (Conv1D)            (None, 131, 256)          196864    \n",
      "_________________________________________________________________\n",
      "batch_normalization_7 (Batch (None, 131, 256)          1024      \n",
      "_________________________________________________________________\n",
      "activation_7 (Activation)    (None, 131, 256)          0         \n",
      "_________________________________________________________________\n",
      "max_pooling1d_6 (MaxPooling1 (None, 43, 256)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_8 (Conv1D)            (None, 43, 256)           196864    \n",
      "_________________________________________________________________\n",
      "batch_normalization_8 (Batch (None, 43, 256)           1024      \n",
      "_________________________________________________________________\n",
      "activation_8 (Activation)    (None, 43, 256)           0         \n",
      "_________________________________________________________________\n",
      "max_pooling1d_7 (MaxPooling1 (None, 14, 256)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_9 (Conv1D)            (None, 14, 256)           196864    \n",
      "_________________________________________________________________\n",
      "batch_normalization_9 (Batch (None, 14, 256)           1024      \n",
      "_________________________________________________________________\n",
      "activation_9 (Activation)    (None, 14, 256)           0         \n",
      "_________________________________________________________________\n",
      "max_pooling1d_8 (MaxPooling1 (None, 4, 256)            0         \n",
      "_________________________________________________________________\n",
      "conv1d_10 (Conv1D)           (None, 4, 512)            393728    \n",
      "_________________________________________________________________\n",
      "batch_normalization_10 (Batc (None, 4, 512)            2048      \n",
      "_________________________________________________________________\n",
      "activation_10 (Activation)   (None, 4, 512)            0         \n",
      "_________________________________________________________________\n",
      "max_pooling1d_9 (MaxPooling1 (None, 1, 512)            0         \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 1, 512)            0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 50)                25650     \n",
      "_________________________________________________________________\n",
      "activation_11 (Activation)   (None, 50)                0         \n",
      "=================================================================\n",
      "Total params: 1,611,058\n",
      "Trainable params: 1,606,194\n",
      "Non-trainable params: 4,864\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model=Sequential()\n",
    "\n",
    "# Layer 1\n",
    "model.add(Conv1D (kernel_size=3, filters=128, strides=3, padding='valid',\n",
    "                  kernel_initializer='he_uniform', input_shape=x_train.shape[1:]))                  \n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('relu'))\n",
    "\n",
    "# Layer 2\n",
    "model.add(Conv1D (kernel_size=3, filters=128, padding='same', kernel_initializer='he_uniform'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling1D(pool_size=3, strides=3))\n",
    "\n",
    "# Layer 3\n",
    "model.add(Conv1D (kernel_size=3, filters=128, padding='same', kernel_initializer='he_uniform'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling1D(pool_size=3, strides=3))\n",
    "\n",
    "# Layer 4\n",
    "model.add(Conv1D (kernel_size=3, filters=256, padding='same', kernel_initializer='he_uniform'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling1D(pool_size=3, strides=3))\n",
    "\n",
    "# Layer 5\n",
    "model.add(Conv1D (kernel_size=3, filters=256, padding='same', kernel_initializer='he_uniform'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling1D(pool_size=3, strides=3))\n",
    "\n",
    "# Layer 6\n",
    "model.add(Conv1D (kernel_size=3, filters=256, padding='same', kernel_initializer='he_uniform'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling1D(pool_size=3, strides=3))\n",
    "\n",
    "# Layer 7\n",
    "model.add(Conv1D (kernel_size=3, filters=256, padding='same', kernel_initializer='he_uniform'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling1D(pool_size=3, strides=3))\n",
    "\n",
    "# Layer 8\n",
    "model.add(Conv1D (kernel_size=3, filters=256, padding='same', kernel_initializer='he_uniform'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling1D(pool_size=3, strides=3))\n",
    "\n",
    "# Layer 9\n",
    "model.add(Conv1D (kernel_size=3, filters=256, padding='same', kernel_initializer='he_uniform'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling1D(pool_size=3, strides=3))\n",
    "\n",
    "# Layer 10\n",
    "model.add(Conv1D (kernel_size=3, filters=512, padding='same', kernel_initializer='he_uniform'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling1D(pool_size=3, strides=3))\n",
    "\n",
    "# Layer 11\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Flatten())\n",
    "\n",
    "# Layer 12\n",
    "model.add(Dense(50))\n",
    "model.add(Activation('sigmoid'))\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "RLxfqHNxDuJq"
   },
   "source": [
    "(4) Compile "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "5jPB8IbZDxeJ"
   },
   "outputs": [],
   "source": [
    "model.compile(loss='binary_crossentropy', \n",
    "              optimizer=keras.optimizers.SGD(lr=0.01, momentum=0.9, decay=1e-6, nesterov=True),\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "VUsuRj-7Dzxx"
   },
   "source": [
    "(5) Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_path = 'SampleCNN_check_point/'\n",
    "os.makedirs(model_path, exist_ok=True)\n",
    "model_filename = model_path+'{epoch:02d}-{val_loss:.4f}.hdf5'\n",
    "checkpointer = ModelCheckpoint(filepath = model_filename, monitor = \"val_loss\", verbose=1, save_best_only=True)\n",
    "\n",
    "early_stopping = EarlyStopping(monitor='val_loss', patience=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 404
    },
    "colab_type": "code",
    "id": "ZUVV71K2D2tZ",
    "outputId": "7a454152-003e-4615-acd8-cfdb60ef8170",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 18706 samples, validate on 1825 samples\n",
      "Epoch 1/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1646 - acc: 0.9472 - val_loss: 0.1624 - val_acc: 0.9482\n",
      "\n",
      "Epoch 00001: val_loss improved from inf to 0.16242, saving model to SampleCNN_check_point/01-0.1624.hdf5\n",
      "Epoch 2/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1624 - acc: 0.9475 - val_loss: 0.1620 - val_acc: 0.9477\n",
      "\n",
      "Epoch 00002: val_loss improved from 0.16242 to 0.16204, saving model to SampleCNN_check_point/02-0.1620.hdf5\n",
      "Epoch 3/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1615 - acc: 0.9476 - val_loss: 0.1615 - val_acc: 0.9479\n",
      "\n",
      "Epoch 00003: val_loss improved from 0.16204 to 0.16153, saving model to SampleCNN_check_point/03-0.1615.hdf5\n",
      "Epoch 4/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1600 - acc: 0.9476 - val_loss: 0.1610 - val_acc: 0.9479\n",
      "\n",
      "Epoch 00004: val_loss improved from 0.16153 to 0.16099, saving model to SampleCNN_check_point/04-0.1610.hdf5\n",
      "Epoch 5/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1586 - acc: 0.9478 - val_loss: 0.1621 - val_acc: 0.9483\n",
      "\n",
      "Epoch 00005: val_loss did not improve from 0.16099\n",
      "Epoch 6/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1567 - acc: 0.9481 - val_loss: 0.1621 - val_acc: 0.9476\n",
      "\n",
      "Epoch 00006: val_loss did not improve from 0.16099\n",
      "Epoch 7/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1553 - acc: 0.9483 - val_loss: 0.1626 - val_acc: 0.9481\n",
      "\n",
      "Epoch 00007: val_loss did not improve from 0.16099\n",
      "Epoch 8/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1536 - acc: 0.9485 - val_loss: 0.1595 - val_acc: 0.9482\n",
      "\n",
      "Epoch 00008: val_loss improved from 0.16099 to 0.15949, saving model to SampleCNN_check_point/08-0.1595.hdf5\n",
      "Epoch 9/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1521 - acc: 0.9490 - val_loss: 0.1614 - val_acc: 0.9479\n",
      "\n",
      "Epoch 00009: val_loss did not improve from 0.15949\n",
      "Epoch 10/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1508 - acc: 0.9490 - val_loss: 0.1606 - val_acc: 0.9478\n",
      "\n",
      "Epoch 00010: val_loss did not improve from 0.15949\n",
      "Epoch 11/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1491 - acc: 0.9493 - val_loss: 0.1618 - val_acc: 0.9477\n",
      "\n",
      "Epoch 00011: val_loss did not improve from 0.15949\n",
      "Epoch 12/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1479 - acc: 0.9495 - val_loss: 0.1624 - val_acc: 0.9481\n",
      "\n",
      "Epoch 00012: val_loss did not improve from 0.15949\n",
      "Epoch 13/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1465 - acc: 0.9499 - val_loss: 0.1603 - val_acc: 0.9477\n",
      "\n",
      "Epoch 00013: val_loss did not improve from 0.15949\n",
      "Epoch 14/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1454 - acc: 0.9501 - val_loss: 0.1651 - val_acc: 0.9475\n",
      "\n",
      "Epoch 00014: val_loss did not improve from 0.15949\n",
      "Epoch 15/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1439 - acc: 0.9502 - val_loss: 0.1575 - val_acc: 0.9477\n",
      "\n",
      "Epoch 00015: val_loss improved from 0.15949 to 0.15746, saving model to SampleCNN_check_point/15-0.1575.hdf5\n",
      "Epoch 16/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1425 - acc: 0.9506 - val_loss: 0.1645 - val_acc: 0.9477\n",
      "\n",
      "Epoch 00016: val_loss did not improve from 0.15746\n",
      "Epoch 17/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1408 - acc: 0.9510 - val_loss: 0.1594 - val_acc: 0.9482\n",
      "\n",
      "Epoch 00017: val_loss did not improve from 0.15746\n",
      "Epoch 18/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1392 - acc: 0.9512 - val_loss: 0.1623 - val_acc: 0.9463\n",
      "\n",
      "Epoch 00018: val_loss did not improve from 0.15746\n",
      "Epoch 19/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1381 - acc: 0.9515 - val_loss: 0.1651 - val_acc: 0.9477\n",
      "\n",
      "Epoch 00019: val_loss did not improve from 0.15746\n",
      "Epoch 20/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1365 - acc: 0.9521 - val_loss: 0.1652 - val_acc: 0.9472\n",
      "\n",
      "Epoch 00020: val_loss did not improve from 0.15746\n",
      "Epoch 21/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1353 - acc: 0.9524 - val_loss: 0.1613 - val_acc: 0.9476\n",
      "\n",
      "Epoch 00021: val_loss did not improve from 0.15746\n",
      "Epoch 22/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1341 - acc: 0.9526 - val_loss: 0.1603 - val_acc: 0.9473\n",
      "\n",
      "Epoch 00022: val_loss did not improve from 0.15746\n",
      "Epoch 23/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1332 - acc: 0.9527 - val_loss: 0.1624 - val_acc: 0.9470\n",
      "\n",
      "Epoch 00023: val_loss did not improve from 0.15746\n",
      "Epoch 24/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1317 - acc: 0.9531 - val_loss: 0.1649 - val_acc: 0.9456\n",
      "\n",
      "Epoch 00024: val_loss did not improve from 0.15746\n",
      "Epoch 25/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1304 - acc: 0.9535 - val_loss: 0.1718 - val_acc: 0.9463\n",
      "\n",
      "Epoch 00025: val_loss did not improve from 0.15746\n",
      "Epoch 26/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1290 - acc: 0.9537 - val_loss: 0.1635 - val_acc: 0.9467\n",
      "\n",
      "Epoch 00026: val_loss did not improve from 0.15746\n",
      "Epoch 27/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1276 - acc: 0.9543 - val_loss: 0.1624 - val_acc: 0.9486\n",
      "\n",
      "Epoch 00027: val_loss did not improve from 0.15746\n",
      "Epoch 28/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1267 - acc: 0.9545 - val_loss: 0.1769 - val_acc: 0.9444\n",
      "\n",
      "Epoch 00028: val_loss did not improve from 0.15746\n",
      "Epoch 29/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1253 - acc: 0.9549 - val_loss: 0.1722 - val_acc: 0.9444\n",
      "\n",
      "Epoch 00029: val_loss did not improve from 0.15746\n",
      "Epoch 30/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1247 - acc: 0.9551 - val_loss: 0.1715 - val_acc: 0.9481\n",
      "\n",
      "Epoch 00030: val_loss did not improve from 0.15746\n",
      "Epoch 31/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1228 - acc: 0.9555 - val_loss: 0.1628 - val_acc: 0.9473\n",
      "\n",
      "Epoch 00031: val_loss did not improve from 0.15746\n",
      "Epoch 32/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1215 - acc: 0.9559 - val_loss: 0.1684 - val_acc: 0.9460\n",
      "\n",
      "Epoch 00032: val_loss did not improve from 0.15746\n",
      "Epoch 33/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1210 - acc: 0.9560 - val_loss: 0.1711 - val_acc: 0.9479\n",
      "\n",
      "Epoch 00033: val_loss did not improve from 0.15746\n",
      "Epoch 34/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1198 - acc: 0.9562 - val_loss: 0.1631 - val_acc: 0.9474\n",
      "\n",
      "Epoch 00034: val_loss did not improve from 0.15746\n",
      "Epoch 35/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1179 - acc: 0.9568 - val_loss: 0.1745 - val_acc: 0.9483\n",
      "\n",
      "Epoch 00035: val_loss did not improve from 0.15746\n",
      "Epoch 36/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1171 - acc: 0.9571 - val_loss: 0.2086 - val_acc: 0.9224\n",
      "\n",
      "Epoch 00036: val_loss did not improve from 0.15746\n",
      "Epoch 37/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1170 - acc: 0.9570 - val_loss: 0.1762 - val_acc: 0.9459\n",
      "\n",
      "Epoch 00037: val_loss did not improve from 0.15746\n",
      "Epoch 38/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1154 - acc: 0.9573 - val_loss: 0.1785 - val_acc: 0.9472\n",
      "\n",
      "Epoch 00038: val_loss did not improve from 0.15746\n",
      "Epoch 39/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1146 - acc: 0.9578 - val_loss: 0.1699 - val_acc: 0.9470\n",
      "\n",
      "Epoch 00039: val_loss did not improve from 0.15746\n",
      "Epoch 40/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1128 - acc: 0.9583 - val_loss: 0.2033 - val_acc: 0.9440\n",
      "\n",
      "Epoch 00040: val_loss did not improve from 0.15746\n",
      "Epoch 41/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1125 - acc: 0.9582 - val_loss: 0.1694 - val_acc: 0.9471\n",
      "\n",
      "Epoch 00041: val_loss did not improve from 0.15746\n",
      "Epoch 42/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1116 - acc: 0.9588 - val_loss: 0.1759 - val_acc: 0.9475\n",
      "\n",
      "Epoch 00042: val_loss did not improve from 0.15746\n",
      "Epoch 43/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1102 - acc: 0.9590 - val_loss: 0.1754 - val_acc: 0.9404\n",
      "\n",
      "Epoch 00043: val_loss did not improve from 0.15746\n",
      "Epoch 44/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1099 - acc: 0.9591 - val_loss: 0.1778 - val_acc: 0.9469\n",
      "\n",
      "Epoch 00044: val_loss did not improve from 0.15746\n",
      "Epoch 45/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1085 - acc: 0.9595 - val_loss: 0.1707 - val_acc: 0.9453\n",
      "\n",
      "Epoch 00045: val_loss did not improve from 0.15746\n",
      "Epoch 46/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1083 - acc: 0.9596 - val_loss: 0.1850 - val_acc: 0.9481\n",
      "\n",
      "Epoch 00046: val_loss did not improve from 0.15746\n",
      "Epoch 47/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1073 - acc: 0.9598 - val_loss: 0.1849 - val_acc: 0.9444\n",
      "\n",
      "Epoch 00047: val_loss did not improve from 0.15746\n",
      "Epoch 48/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1059 - acc: 0.9605 - val_loss: 0.1744 - val_acc: 0.9478\n",
      "\n",
      "Epoch 00048: val_loss did not improve from 0.15746\n",
      "Epoch 49/1000\n",
      "18706/18706 [==============================] - 252s 13ms/step - loss: 0.1052 - acc: 0.9606 - val_loss: 0.1732 - val_acc: 0.9461\n",
      "\n",
      "Epoch 00049: val_loss did not improve from 0.15746\n",
      "Epoch 50/1000\n",
      "18706/18706 [==============================] - 251s 13ms/step - loss: 0.1043 - acc: 0.9609 - val_loss: 0.1770 - val_acc: 0.9451\n",
      "\n",
      "Epoch 00050: val_loss did not improve from 0.15746\n",
      "Epoch 51/1000\n",
      "18706/18706 [==============================] - 251s 13ms/step - loss: 0.1039 - acc: 0.9611 - val_loss: 0.1795 - val_acc: 0.9442\n",
      "\n",
      "Epoch 00051: val_loss did not improve from 0.15746\n",
      "Epoch 52/1000\n",
      "18706/18706 [==============================] - 251s 13ms/step - loss: 0.1026 - acc: 0.9615 - val_loss: 0.1851 - val_acc: 0.9465\n",
      "\n",
      "Epoch 00052: val_loss did not improve from 0.15746\n",
      "Epoch 53/1000\n",
      "18706/18706 [==============================] - 251s 13ms/step - loss: 0.1019 - acc: 0.9617 - val_loss: 0.1791 - val_acc: 0.9457\n",
      "\n",
      "Epoch 00053: val_loss did not improve from 0.15746\n",
      "Epoch 54/1000\n",
      "18706/18706 [==============================] - 251s 13ms/step - loss: 0.1002 - acc: 0.9620 - val_loss: 0.1789 - val_acc: 0.9457\n",
      "\n",
      "Epoch 00054: val_loss did not improve from 0.15746\n",
      "Epoch 55/1000\n",
      "18706/18706 [==============================] - 251s 13ms/step - loss: 0.1006 - acc: 0.9621 - val_loss: 0.1778 - val_acc: 0.9451\n",
      "\n",
      "Epoch 00055: val_loss did not improve from 0.15746\n",
      "Epoch 56/1000\n",
      "18706/18706 [==============================] - 251s 13ms/step - loss: 0.0995 - acc: 0.9623 - val_loss: 0.1803 - val_acc: 0.9404\n",
      "\n",
      "Epoch 00056: val_loss did not improve from 0.15746\n",
      "Epoch 57/1000\n",
      "18706/18706 [==============================] - 251s 13ms/step - loss: 0.0983 - acc: 0.9628 - val_loss: 0.1880 - val_acc: 0.9440\n",
      "\n",
      "Epoch 00057: val_loss did not improve from 0.15746\n",
      "Epoch 58/1000\n",
      "18706/18706 [==============================] - 251s 13ms/step - loss: 0.0976 - acc: 0.9632 - val_loss: 0.1783 - val_acc: 0.9446\n",
      "\n",
      "Epoch 00058: val_loss did not improve from 0.15746\n",
      "Epoch 59/1000\n",
      "18706/18706 [==============================] - 251s 13ms/step - loss: 0.0969 - acc: 0.9632 - val_loss: 0.1899 - val_acc: 0.9452\n",
      "\n",
      "Epoch 00059: val_loss did not improve from 0.15746\n",
      "Epoch 60/1000\n",
      "18706/18706 [==============================] - 251s 13ms/step - loss: 0.0960 - acc: 0.9635 - val_loss: 0.1753 - val_acc: 0.9463\n",
      "\n",
      "Epoch 00060: val_loss did not improve from 0.15746\n",
      "Epoch 61/1000\n",
      "18706/18706 [==============================] - 251s 13ms/step - loss: 0.0951 - acc: 0.9639 - val_loss: 0.1803 - val_acc: 0.9453\n",
      "\n",
      "Epoch 00061: val_loss did not improve from 0.15746\n",
      "Epoch 62/1000\n",
      "18706/18706 [==============================] - 251s 13ms/step - loss: 0.0947 - acc: 0.9641 - val_loss: 0.1805 - val_acc: 0.9464\n",
      "\n",
      "Epoch 00062: val_loss did not improve from 0.15746\n",
      "Epoch 63/1000\n",
      "18706/18706 [==============================] - 251s 13ms/step - loss: 0.0933 - acc: 0.9644 - val_loss: 0.1849 - val_acc: 0.9451\n",
      "\n",
      "Epoch 00063: val_loss did not improve from 0.15746\n",
      "Epoch 64/1000\n",
      "18706/18706 [==============================] - 251s 13ms/step - loss: 0.0932 - acc: 0.9645 - val_loss: 0.2163 - val_acc: 0.9477\n",
      "\n",
      "Epoch 00064: val_loss did not improve from 0.15746\n",
      "Epoch 65/1000\n",
      "18706/18706 [==============================] - 251s 13ms/step - loss: 0.0925 - acc: 0.9647 - val_loss: 0.1867 - val_acc: 0.9452\n",
      "\n",
      "Epoch 00065: val_loss did not improve from 0.15746\n"
     ]
    }
   ],
   "source": [
    "hist = model.fit(x_train, y_train, batch_size=64, epochs=1000, validation_data=[x_val, y_val],\n",
    "                 shuffle=True, callbacks = [checkpointer, early_stopping])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, loss_ax = plt.subplots()\n",
    "loss_ax.plot(hist.history['loss'], 'y', label='train loss')\n",
    "loss_ax.plot(hist.history['val_loss'], 'r', label='val loss')\n",
    "loss_ax.set_xlabel('epoch')\n",
    "loss_ax.set_ylabel('loss')\n",
    "loss_ax.legend(loc='upper left')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "(6) Test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_patch_data(wav, sr, i):\n",
    "    patch_data = wav[i*sr:(i+6)*sr]\n",
    "    patch_data = np.expand_dims(patch_data, axis=1)\n",
    "    return patch_data\n",
    "\n",
    "def make_patch_by_filename(filename):\n",
    "    wav, sr = librosa.load(filename, sr=16000)\n",
    "    patch_range = 24\n",
    "    patch_list = [make_patch_data(wav, sr, i) for i in range(patch_range)]\n",
    "    return np.asarray([patch_list])\n",
    "\n",
    "def load_test_wavs(filenames):\n",
    "    patch_arr = np.asarray([make_patch_by_filename(filename) for filename in tqdm(filenames)])\n",
    "    result_arr = patch_arr.reshape(-1, *patch_arr.shape[-2:])\n",
    "    return result_arr\n",
    "#    return np.concatenate(patch_list, axis=None)\n",
    "\n",
    "def make_test_data(filenames, y_dict):\n",
    "    x_test = load_test_wavs(filenames)\n",
    "    repeat_size = x_test.shape[0] / len(filenames)\n",
    "    y_test = np.repeat(make_y_by_filenames(filenames, y_dict), repeat_size, axis=0)\n",
    "    return x_test, y_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Make test data.......\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f6e30f8406384eb5baff8b5738b15c43",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=500), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "(12000, 96000, 1) (12000, 50)\n"
     ]
    }
   ],
   "source": [
    "print('Make test data.......')\n",
    "x_test_wav_filenames = [test_dir+filename for filename in os.listdir(test_dir)\n",
    "                            if filename.endswith('.wav')]\n",
    "x_test_wav_filenames = random.sample(x_test_wav_filenames, 500)\n",
    "x_test, y_test = make_test_data(x_test_wav_filenames, y_dict)\n",
    "\n",
    "print(x_test.shape, y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "model = load_model(model_path+'37-0.0856.hdf5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "DefsBoWAFKdi",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "12000/12000 [==============================] - 11s 932us/step\n",
      "Loss: 0.15847472915053368 Accuracy: 0.9489050016403199\n"
     ]
    }
   ],
   "source": [
    "[loss, accuracy] = model.evaluate(x_test, y_test)\n",
    "print('Loss:', loss, 'Accuracy:', accuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def sort_list(list1, list2): \n",
    "    pairs = zip(list2, list1)\n",
    "    result = [(x, y) for y, x in sorted(pairs, reverse=True)]\n",
    "    x_list = [x for x, _ in result]\n",
    "    y_list = [y for _, y in result]\n",
    "\n",
    "    return x_list, y_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cello : 0.9958333333333333\n",
      "harp : 0.988\n",
      "country : 0.988\n",
      "classic : 0.988\n",
      "voice : 0.984\n",
      "female voice : 0.98\n",
      "solo : 0.978\n",
      "no voice : 0.976\n",
      "male voice : 0.976\n",
      "man : 0.9739166666666667\n",
      "harpsichord : 0.9738333333333333\n",
      "violin : 0.97325\n",
      "weird : 0.972\n",
      "new age : 0.97\n",
      "beats : 0.96975\n",
      "female vocal : 0.968\n",
      "flute : 0.9664166666666667\n",
      "indian : 0.966\n",
      "metal : 0.9649166666666666\n",
      "sitar : 0.9648333333333333\n",
      "dance : 0.9641666666666666\n",
      "choir : 0.9641666666666666\n",
      "choral : 0.9638333333333333\n",
      "male vocal : 0.9618333333333333\n",
      "soft : 0.958\n",
      "no vocals : 0.958\n",
      "no vocal : 0.956\n",
      "male : 0.95275\n",
      "quiet : 0.952\n",
      "woman : 0.948\n",
      "pop : 0.948\n",
      "strings : 0.9478333333333333\n",
      "opera : 0.9394166666666667\n",
      "singing : 0.938\n",
      "loud : 0.93775\n",
      "female : 0.9359166666666666\n",
      "rock : 0.9335833333333333\n",
      "synth : 0.932\n",
      "vocals : 0.93\n",
      "beat : 0.9278333333333333\n",
      "fast : 0.9143333333333333\n",
      "vocal : 0.914\n",
      "piano : 0.911\n",
      "drums : 0.9086666666666666\n",
      "techno : 0.9026666666666666\n",
      "electronic : 0.9010833333333333\n",
      "classical : 0.9005833333333333\n",
      "ambient : 0.8996666666666666\n",
      "slow : 0.8820833333333333\n",
      "guitar : 0.8453333333333334\n"
     ]
    }
   ],
   "source": [
    "pred = model.predict(x_test)\n",
    "accu_table = (pred > 0.5) == y_test\n",
    "accu_arr = accu_table.sum(axis=0) / y_test.shape[0]\n",
    "\n",
    "tag_list, accu_list = sort_list(y_names, accu_arr)\n",
    "for tag, accu in zip(tag_list, accu_list):\n",
    "    print(tag, ':', accu)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (15, 6))\n",
    "x = np.arange(len(accu_list))\n",
    "plt.bar(x, accu_list)\n",
    "plt.xticks(x, tag_list, rotation='vertical')\n",
    "plt.xlabel('Tag') # , fontsize=18)\n",
    "plt.ylabel('AUC') # , fontsize=16)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rock : 0.7137292890501068\n",
      "techno : 0.3844716922129295\n",
      "choir : 0.28306567473582306\n",
      "choral : 0.15258780101915417\n",
      "guitar : 0.1491639065584629\n",
      "beat : 0.14205272168490826\n",
      "fast : 0.11639155151393937\n",
      "metal : 0.11277793647882636\n",
      "classical : 0.10341338775621474\n",
      "opera : 0.10040830939203071\n",
      "loud : 0.08364302200903186\n",
      "flute : 0.06737639314883205\n",
      "dance : 0.06355310544889495\n",
      "harpsichord : 0.06269382797171767\n",
      "piano : 0.06024797703873937\n",
      "ambient : 0.036078680767535176\n",
      "drums : 0.031376237938513754\n",
      "male : 0.019140233356065836\n",
      "sitar : 0.013986402014301404\n",
      "indian : 0.00943492229277628\n",
      "beats : 0.006208658129888053\n",
      "violin : 0.005396478455258042\n",
      "female : 0.00477129303089517\n",
      "electronic : 0.0037449438172712447\n",
      "quiet : 0.0030656935808953253\n",
      "slow : 0.0014332403820726332\n",
      "woman : 0.0\n",
      "weird : 0.0\n",
      "voice : 0.0\n",
      "vocals : 0.0\n",
      "vocal : 0.0\n",
      "synth : 0.0\n",
      "strings : 0.0\n",
      "solo : 0.0\n",
      "soft : 0.0\n",
      "singing : 0.0\n",
      "pop : 0.0\n",
      "no voice : 0.0\n",
      "no vocals : 0.0\n",
      "no vocal : 0.0\n",
      "new age : 0.0\n",
      "man : 0.0\n",
      "male voice : 0.0\n",
      "male vocal : 0.0\n",
      "harp : 0.0\n",
      "female voice : 0.0\n",
      "female vocal : 0.0\n",
      "country : 0.0\n",
      "classic : 0.0\n",
      "cello : 0.0\n"
     ]
    }
   ],
   "source": [
    "pred = model.predict(x_test)\n",
    "true_positive = (pred > 0.5) & y_test\n",
    "precision_arr = true_positive.sum(axis=0) / (pred.sum(axis=0) + keras.backend.epsilon())\n",
    "\n",
    "tag_list, precision_list = sort_list(y_names, precision_arr)\n",
    "for tag, accu in zip(tag_list, precision_list):\n",
    "    print(tag, ':', accu)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (15, 6))\n",
    "x = np.arange(len(precision_list))\n",
    "plt.bar(x, precision_list)\n",
    "plt.xticks(x, tag_list, rotation='vertical')\n",
    "plt.xlabel('Tag') # , fontsize=18)\n",
    "plt.ylabel('AUC') # , fontsize=16)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA34AAAG5CAYAAADRW+YxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3Xv87dWc+PHXu3MKSTc1ScXphmlQqBGacZ+JKJKISL8I45K7k0FuQ5H7zESkKYWiKMolUSnE6aYbIxWVKKYbReL9+2Ot3dnn2/e79+ez93ef7/l+zuv5eOzHd9/W/qy9v3t/Pp/3Wu+1VmQmkiRJkqTuWmWuKyBJkiRJmiwDP0mSJEnqOAM/SZIkSeo4Az9JkiRJ6jgDP0mSJEnqOAM/SZIkSeo4Az9JkiRJ6jgDP0nS2CLiD32Xv0XEbX23nz/G6/4wIvYc8PiDIiL7tvWbiDgxIh7fYhsvi4hvj1HHHet77v8Mvjjq60mSNAkL57oCkqT5LzPX6F2PiCuBF2fmyMFUS3/tbT8iNgSeD3wtIvbJzC8spzpcnplbDHtSRCzMzDuWR4UkSepnj58kaeIiYkFEvC0iLo+I30XE0RGxdn3snhHxhYj4v4i4MSLOjoh1IuKDwHbAp2sv2geHbSczr83Mg4H3AR/o2/7bI+KKiLglIi6KiJ3q/Q8DPgI8rtdjWO9/ZkRcEBE3R8SvIuItI77vAyPicxFxTETcAjx30GdRy+xTt3l9RLyx9mLuUB/7QkS8te+5O0bEZX23N4mIE+rrXh4RL5tSl6Mj4vP1c/hJRGzT9/iivrK/i4gPRsTq9TPYsu95G0fErf11liSt+Az8JEnLwxuAfwF2ADYG/gJ8uD72YkoGykbAesArgdsz8/XAjym9h2vU200dD2wcEZvW2z8DHg2sBRwEfCEi1svM84DXAKfVbdynPv9m4HnA2sAzgDdExI4jvG+AZwFH1G0fx4DPoi8QfU59bBHlMxkqIhYAJwPfB+4L7Ai8JSIe2/e0ZwKfqe/r1LotImJV4OvApcD9gE2A4zLzVuBLQH+67fOBkzLzxuYfgSRprhn4SZKWh5cBizPz15n5J+CdwHMiIiiBz/rA5pl5R2b+ODP/OOb2fl3/rguQmcfU3sC/ZeZngWuAR8xUODNPzcyL6/PPBY4FHjvT84FNa29l77Jz32OnZ+bJ9bVuY/Bn8WxKwPWDzPwz8BaaH6t3AO6emQdl5u2Z+b/A4cBz+57zncw8JTP/CnwW2Kav7JrAWzLz1sy8LTO/Xx87ghLs9exZy0qS5hHH+EmSJqoGNJsAJ0dE9j20CnBv4DDgPsCXImIN4EjgbTU4GdVG9e//1TrsA+xH6c0CWIMBPWkR8RjgvcBWwGrA3Rgc7FwxYIzfVX2vO+yzuG//8zPzpoi4acB2+90fWBQR/T1xC4D+sZa/6bt+K+VzoNbpisz82zSvewawICIeBdwGbEjpHZQkzSMGfpKkicrMjIhrgF0z85wZnvZ24O0RsRnwTeBi4GggZ3j+MM8Ers7MKyLiAcDHgScAP8rMv0XET4HoVXGa8scC/wF8JjP/FBGfYPRj5p2vP+yziIhrKUFY7/ZalBTRnj8Cq/fdvk/f9auAn2bmQ0ao41WUoHGVqcFfrfORlJ6+PwFfyMy/jLANSdIcMtVTkrQ8fAI4MCI2AYiIv4uIp9frT4qIrSJiFcrYujuAXvDxW2CzphuJiPtExGuB/YHF9e416utdD6xSJzzp7537LbBJHefW65VbA/h9DfoeTUnBnC0zfhaUgHPXiHhkRNwNeA9LPwuA84GnRcTaEbER8Kq+x86sr/eaiLh7RCyMiIdGxMMb1OlM4Bbg3XVCl3vU991zJLA7sEe9LkmaZwz8JEnLw/spKYffqbNbfh/oBSQbASdQAo+LKBOUHFMf+zDwwoi4ISLeP8NrL6gzcv4RuAB4IrBLZh4NUMfofQJYAlwLbFqv93wDuBK4LiKuzsykjMM7uNb1TcBsrss342dRJ5t5PWVClauBXwG/6yv7GeCyev/XgM/3Hqi9cE+lTGLzS0qgewhL0zln1Fd2677tPrPv8V9QJsi5JTN/NMJ7liTNsSjHN0mStCKqS0zslplnznE9Pgdckpnvmct6SJJGY4+fJEkaKCK2AJ5GmSVUkjQPGfhJkqQZ1RTb84B3ZeY1c10fSdJoTPWUJEmSpI6zx0+SJEmSOs7AT5IkSZI6bl4v4L7eeuvlokWL5roakiRJkjQnzjnnnN9l5vrDnjevA79FixaxZMmS4U+UJEmSpA6KiF82eZ6pnpIkSZLUcQZ+kiRJktRxBn6SJEmS1HEGfpIkSZLUcQZ+kiRJktRxBn6SJEmS1HEGfpIkSZLUcQZ+kiRJktRxBn6SJEmS1HEGfpIkSZLUcQZ+kiRJktRxBn6SJEmS1HEGfpIkSZLUcQvnugLSfLVo8Umty1x54E4TqIkkSZI0mD1+kiRJktRxBn6SJEmS1HEGfpIkSZLUcQZ+kiRJktRxBn6SJEmS1HEGfpIkSZLUcQZ+kiRJktRxBn6SJEmS1HEGfpIkSZLUcQZ+kiRJktRxBn6SJEmS1HEGfpIkSZLUcQZ+kiRJktRxBn6SJEmS1HEGfpIkSZLUcQZ+kiRJktRxBn6SJEmS1HEGfpIkSZLUcQZ+kiRJktRxBn6SJEmS1HEGfpIkSZLUcQZ+kiRJktRxEw38IuK1EXFxRFwUEZ+PiLtHxKYRcXZEXBYRx0TEavW5d6u3L6uPL5pk3SRJkiRpZTGxwC8iNgJeDWybmQ8GFgDPBQ4CPpyZWwA3APvUIvsAN9T7P1yfJ0mSJEka08Ll8Pr3iIi/AKsD1wJPAJ5XHz8CeAdwCLBLvQ7wJeA/IyIyMydcR0mSpBXaosUntXr+lQfuNKGaSJqvJtbjl5nXAAcDv6IEfDcB5wA3ZuYd9WlXAxvV6xsBV9Wyd9Tn33tS9ZMkSZKklcUkUz3XofTibQrcF7gnsOMsvO6+EbEkIpZcf/31476cJEmSJHXeJCd3eRJwRWZen5l/AY4HHgOsHRG9FNONgWvq9WuATQDq42sBv5/6opl5aGZum5nbrr/++hOsviRJkiR1wyQDv18B20fE6hERwBOBS4DvArvV5+wFnFCvn1hvUx//juP7JEmSJGl8kxzjdzZlkpZzgQvrtg4F3gy8LiIuo4zhO6wWOQy4d73/dcDiSdVNkiRJklYmE53VMzMPAA6YcvflwD9O89w/Ac+eZH0kSZIkaWU00QXcJUmSJElzz8BPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6buFcV0Aa16LFJ7V6/pUH7jShmkiSJEkrJnv8JEmSJKnjDPwkSZIkqeMM/CRJkiSp4wz8JEmSJKnjDPwkSZIkqeMM/CRJkiSp4wz8JEmSJKnjDPwkSZIkqeMM/CRJkiSp4wz8JEmSJKnjDPwkSZIkqeMM/CRJkiSp4wz8JEmSJKnjDPwkSZIkqeMM/CRJkiSp4wz8JEmSJKnjDPwkSZIkqeMM/CRJkiSp4wz8JEmSJKnjDPwkSZIkqeMM/CRJkiSp4wz8JEmSJKnjDPwkSZIkqeMM/CRJkiSp4wz8JEmSJKnjDPwkSZIkqeMM/CRJkiSp4wz8JEmSJKnjDPwkSZIkqeMM/CRJkiSp4yYa+EXE2hHxpYj4aURcGhGPioh1I+KUiPh5/btOfW5ExMci4rKI+ElEPHySdZMkSZKklcWke/w+CnwjMx8EbA1cCiwGTs3MLYFT622ApwBb1su+wCETrpskSZIkrRQmFvhFxFrAPwOHAWTm7Zl5I7ALcER92hHAM+r1XYAjs/ghsHZEbDip+kmSJEnSymKSPX6bAtcDh0fEeRHx6Yi4J7BBZl5bn/MbYIN6fSPgqr7yV9f7JEmSJEljmGTgtxB4OHBIZj4M+CNL0zoByMwEss2LRsS+EbEkIpZcf/31s1ZZSZIkSeqqhRN87auBqzPz7Hr7S5TA77cRsWFmXltTOa+rj18DbNJXfuN63zIy81DgUIBtt922VdAoSZIkddGixSe1LnPlgTtNoCZaUU2sxy8zfwNcFREPrHc9EbgEOBHYq963F3BCvX4i8MI6u+f2wE19KaGSJEmSpBFNsscP4FXA0RGxGnA5sDcl2Dw2IvYBfgnsXp97MvBU4DLg1vpcSZIkSdKYJhr4Zeb5wLbTPPTEaZ6bwCsmWR9JkiRJWhlNeh0/SZIkSdIcM/CTJEmSpI4z8JMkSZKkjjPwkyRJkqSOM/CTJEmSpI4z8JMkSZKkjjPwkyRJkqSOM/CTJEmSpI6b6ALukiRJ0vK2aPFJrctceeBOE6iJtOKwx0+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOm7hXFdAkrTiWbT4pFbPv/LAnSZUE0mSNBvs8ZMkSZKkjrPHT1JjbXuBwJ4gSZKkFYE9fpIkSZLUcQZ+kiRJktRxBn6SJEmS1HEGfpIkSZLUcU7uIkmaVS4F0Y6TJkmSlgd7/CRJkiSp4wz8JEmSJKnjDPwkSZIkqeMajfGLiNcNejwzPzQ71ZEkSZIkzbamk7tsC2wHnFhvPx34EfDzSVRKkiRJkjR7mgZ+GwMPz8xbACLiHcBJmbnnpComSZIkSZodTcf4bQDc3nf79nqfJEmSJGkF17TH70jgRxHx5Xr7GcARk6mSJEmSJGk2NQr8MvM/IuLrwD/Vu/bOzPMmVy1JkiRJ0mxp2uMHsDpwc2YeHhHrR8SmmXnFpComSZLUJYsWn9Tq+VceuNOEaiJpZdRojF9EHAC8Gdi/3rUqcNSkKiVJkiRJmj1NJ3d5JrAz8EeAzPw1cK9JVUqSJEmSNHuaBn63Z2YCCRAR95xclSRJkiRJs6lp4HdsRHwSWDsiXgJ8G/jU5KolSZIkSZotTWf1PDgingzcDDwQeHtmnjLRmkmSJEmSZsXQwC8iFgDfzszHAwZ7kiRJkjTPDE31zMy/An+LiLWWQ30kSZIkSbOs6Tp+fwAujIhTqDN7AmTmqydSK0mSJEnSrGka+B1fL5IkSZKkeWZg4BcR98vMX2XmEcurQpIkSZKk2TWsx+8rwMMBIuK4zHzW5KskSZKkFcGixSe1LnPlgTtNoCaSxjVscpfou77ZJCsiSZIkSZqMYYFfznBdkiRJkjRPDEv13Doibqb0/N2jXqfezsxcc6K1kyRJkiSNbWDgl5kLlldFJEmSJEmTMXQBd0mSJEnS/GbgJ0mSJEkdZ+AnSZIkSR1n4CdJkiRJHTfxwC8iFkTEeRHxtXp704g4OyIui4hjImK1ev/d6u3L6uOLJl03SZIkSVoZLI8ev/2AS/tuHwR8ODO3AG4A9qn37wPcUO//cH2eJEmSJGlMEw38ImJjYCfg0/V2AE8AvlSfcgTwjHp9l3qb+vgT6/MlSZIkSWOYdI/fR4A3AX+rt+8N3JiZd9TbVwMb1esbAVcB1Mdvqs+XJEmSJI1hYoFfRDwNuC4zz5nl1903IpZExJLrr79+Nl9akiRJkjppkj1+jwF2jogrgS9QUjw/CqwdEQvrczYGrqnXrwE2AaiPrwX8fuqLZuahmbltZm67/vrrT7D6kiRJktQNC4c/ZTSZuT+wP0BEPA54Q2Y+PyK+COxGCQb3Ak6oRU6st39QH/9OZuak6idJTS1afFKr51954E4TqokkSdJo5mIdvzcDr4uIyyhj+A6r9x8G3Lve/zpg8RzUTZIkSZI6Z2I9fv0y8zTgtHr9cuAfp3nOn4BnL4/6SJIkSdLKZC56/CRJkiRJy5GBnyRJkiR13HJJ9ZRWVE7aIUmSpJWBPX6SJEmS1HH2+EmSNCazByRJKzp7/CRJkiSp4wz8JEmSJKnjDPwkSZIkqeMM/CRJkiSp4wz8JEmSJKnjDPwkSZIkqeMM/CRJkiSp4wz8JEmSJKnjDPwkSZIkqeMM/CRJkiSp4xbOdQUkSZIkzV+LFp/U6vlXHrjThGqiQezxkyRJkqSOM/CTJEmSpI4z8JMkSZKkjjPwkyRJkqSOc3IXSfOCA8clSZJGZ4+fJEmSJHWcgZ8kSZIkdZyBnyRJkiR1nIGfJEmSJHWcgZ8kSZIkdZyBnyRJkiR1nIGfJEmSJHWcgZ8kSZIkdZyBnyRJkiR1nIGfJEmSJHWcgZ8kSZIkddzCua6AJEmS1BWLFp/U6vlXHrjThGoiLcseP0mSJEnqOAM/SZIkSeo4Az9JkiRJ6jgDP0mSJEnqOCd30axwILMkSZK04rLHT5IkSZI6zsBPkiRJkjrOVE9JWkGZQi1JkmaLPX6SJEmS1HEGfpIkSZLUcaZ6SpIkdZhp45LAHj9JkiRJ6jwDP0mSJEnqOAM/SZIkSeo4Az9JkiRJ6jgnd5EkSdIKx0lpVg7+n5cfAz9pjrijkyRJ0vJi4CdJklqz8UqS5hfH+EmSJElSxxn4SZIkSVLHGfhJkiRJUscZ+EmSJElSxzm5iyRJKyknaJGklYc9fpIkSZLUcQZ+kiRJktRxBn6SJEmS1HEGfpIkSZLUcQZ+kiRJktRxBn6SJEmS1HETC/wiYpOI+G5EXBIRF0fEfvX+dSPilIj4ef27Tr0/IuJjEXFZRPwkIh4+qbpJkiRJ0spkkj1+dwCvz8ytgO2BV0TEVsBi4NTM3BI4td4GeAqwZb3sCxwywbpJkiRJ0kpjYoFfZl6bmefW67cAlwIbAbsAR9SnHQE8o17fBTgyix8Ca0fEhpOqnyRJkiStLJbLGL+IWAQ8DDgb2CAzr60P/QbYoF7fCLiqr9jV9T5JkiRJ0hgmHvhFxBrAccBrMvPm/scyM4Fs+Xr7RsSSiFhy/fXXz2JNJUmSJKmbFk7yxSNiVUrQd3RmHl/v/m1EbJiZ19ZUzuvq/dcAm/QV37jet4zMPBQ4FGDbbbdtFTRKkqSV26LFJ7V6/pUH7jShmkjS8jXJWT0DOAy4NDM/1PfQicBe9fpewAl997+wzu65PXBTX0qoJEmSJGlEk+zxewzwAuDCiDi/3vcW4EDg2IjYB/glsHt97GTgqcBlwK3A3hOsmyRpBWRvjCRJkzGxwC8zzwRihoefOM3zE3jFpOozX7Q96QFPfCRJkrRyssGwueUyq6ckSZIkae4Y+EmSJElSxxn4SZIkSVLHGfhJkiRJUsdNdB0/SZIkrbyceENacdjjJ0mSJEkdZ+AnSZIkSR1nqqfuZDqGJEmSVhYr27mvgZ8kSZK0klvZgqCVkYGfNA+5c5ak5a/tvhfc/0pacTjGT5IkSZI6zh4/SZIkqY+ZNeoie/wkSZIkqePs8ZO03NiCKkmSNDfs8ZMkSZKkjrPHTysEe4I0Sc7EJ0mSVnYGfpIkaV6xsVCS2jPVU5IkSZI6zh4/SZK0XJl+LUnLn4GfJKkzxkkBNH1QktRlpnpKkiRJUscZ+EmSJElSx5nqKUkTZPqgJElaEdjjJ0mSJEkdZ4+ftJKxB0qSJGnlY+A3AZ5YS5IkSVqRmOopSZIkSR1nj58kdZALZEuSpH72+EmSJElSxxn4SZIkSVLHmerZMU4sI0mSJGkqe/wkSZIkqeMM/CRJkiSp4wz8JEmSJKnjDPwkSZIkqeMM/CRJkiSp4wz8JEmSJKnjDPwkSZIkqeMM/CRJkiSp4wz8JEmSJKnjDPwkSZIkqeMM/CRJkiSp4xbOdQUkSdLoFi0+qdXzrzxwpwnVRJK0IrPHT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6zsBPkiRJkjrOwE+SJEmSOs7AT5IkSZI6buFcV0CSJEkSLFp8UqvnX3ngThOqibrIHj9JkiRJ6jgDP0mSJEnquBUq8IuIHSPiZxFxWUQsnuv6SJIkSVIXrDCBX0QsAP4LeAqwFbBHRGw1t7WSJEmSpPlvhQn8gH8ELsvMyzPzduALwC5zXCdJkiRJmvdWpMBvI+CqvttX1/skSZIkSWOIzJzrOgAQEbsBO2bmi+vtFwCPzMxXTnnevsC+9eYDgZ8t14qOZz3gd3NUfj6Wnctt+57nz7Z9z/Nn277n5Vd2Lre9MtZ7ZXzPc7lt3/P82fbK+J7nwv0zc/2hz8rMFeICPAr4Zt/t/YH957pes/wel8xV+flYdr7We2V8z/O13ivje56v9fY9z59tr4z1Xhnf83yt98r4nudrvefre16RLytSquePgS0jYtOIWA14LnDiHNdJkiRJkua9hXNdgZ7MvCMiXgl8E1gAfCYzL57jakmSJEnSvLfCBH4AmXkycPJc12OCDp3D8vOx7Fxu2/c8f7bte54/2/Y9L7+yc7ntlbHeK+N7nstt+57nz7ZXxve8wlphJneRJEmSJE3GijTGT5IkSZI0AQZ+kiRJktRxBn4TFhH3j4gn1ev3iIh7zXWdpJ6I2CEi9q7X14+ITee6TpLml4h4TETcs17fMyI+FBH3n+t6TUJErDvoMtf1k6RBDPwmKCJeAnwJ+GS9a2PgK8thu6tExKNHKLfroMsYdVlzlLKzUX55iIgFEXH0GOWf3eS+Ia+xc0QcXC9Pb1jmAODNlDUzAVYFjmqxzbs1uW+GshtExGER8fV6e6uI2KfptufKOCe4EbF6RLwtIj5Vb28ZEU+bZH1nW0SsExEPnfA2bomIm6e53BIRNzcoP69PzCPioCb3zVD2+IjYKSKW97H9EODWiNgaeD3wC+DINi8QERtFxKMj4p97l5bl71P3g0+PiPu0KdvSOcCS+nfqZckoLxgRa0TEGrNWw2bbvEdEPHB5brNv2wsi4r4Rcb/eZTlsc7rj1Qq/P5hL43RcTNeI3LRhOSL2a3LfNM9ZEBHfbVbDlZeTu0xQRJwP/CNwdmY+rN53YWY+ZEi51w16PDM/1GDb5/W22aK+hw/ebP6/hq/zOeBlwF8p6zOuCXw0Mz8wifKz9Hk9gHLyskFmPrie3O6cme9pWOczgSdk5u1Nnj+l7LmZ+fBh9w0o/z7K96wXfO4B/Dgz3zKk3PnAw4Bz+76fP8nMRif249S7BnyHA/+emVtHxELgvGG/jb7yjwHeAdyfMjtxUL6jmzUouwHwXuC+mfmUiNgKeFRmHtag7E+ArYGHAv8DfBrYPTMf26DsMZSTwxfW79jqwPczc5sh5b4KzLijzsydG2z7vcD7M/PGensd4PWZ+dYGZU8DdqZ8zucA1wFnZebA391sbHsUEXEF5fOKaR4e+B2Zjc+6vs76lEaVrYC795V/QoOy0/2uGv0u60na3sD2wBeBwzPzZw3rvCtwEPB3lM+u95sa2vDWq3NEvB24JjMPa7kPOwh4DnAJZb9P3XbTz/vFwNuB79R6PxZ4V2Z+pkHZzYCPAo8C/gb8AHhtZl7eZNvjiIiHUALkdSn1vh7YKzMvalD2TOB04HuU3+MtLbb7dOBgYLXM3DQitqF8XjN+3rNxnK2v8yrgAOC3lM+7Fm983Hk0sIi+Wekzc2gjQ0ScBDwjM/9Sb28IfC0zH9Gg7LjnBx+b5u6bKAuEnzCk7FqUY90/1btOp/yvbmqw3VWBlwO9RpTTgU/0PoMhZV8C7Ausm5mbR8SWtewTh5Wt5afbj53T8POermyjc9qIOBXYtcnnM6XcrOz754MVajmHDvpzZt4eUc4/6sltk0i716ryQGA7li5k/3TgRw23fWpEPAs4PhtG95m5d8PXHmarzLw5Ip4PfB1YTDlhbBT4jVB+NtJnPwW8kdo7m5k/qQFoox07cDlwVkScCPyxd+egg2FEPAV4KrDRlAPDmsAdLeq+E7BNZv6tvu4RwHnAwMAPuD0zMyKylrtnk43V1vSNgHtExMNYeoK9JrB6wzqvl5nHRsT+cOc6nn8dVqjPYcBrKd+LNuWgBGyHA/9eb/8vcEx9zWHuqJ/ZLsB/1hPcpj2Vm2fmcyJiD4DMvDV6O4fBDq5/dwXuw9Je2T0oJ09NPKW/ISAzb4iIpwJNgq+16u/xxcCRmXlADYCbGmfbRMTfsWzw9KtBz8/McdKVZ+OzhtIIcwzlt/kyYC/KSf2MIuLlwL8Bm035fO8FnNVko5n5beDb9WRxj3r9Ksr+7aghJ3zvB56emZc22dYUt9Tf8p7AP0fpcVy1RflnAA/MzD+PsG0o++6HZebvASLi3sD3gaGBH/A54L+AZ9bbzwU+DzxyWMHaiLEly34/z2hR708Cr8vM79bXexxlCvkmGTsvoAQDzwI+EBF/Br6Xma9tUPYdlMbC02qdz2/QGzNbw1T2o/yvf9+2YER8FtgcOJ++BgKa9S5/BTg2InYDNqGcV72h4abHPT+4O/AgSmMMlP/ZFcDWEfH4zHzNgLKfAS4Cdq+3X0A5fjXJwjqE8jv8776yhwAvblD2FdSOC4DM/HndFw8UEQ8C/gFYK5bNFFuTvt/JDGX3AJ4HbFrPpXruBfxfgzoD/AG4MCJOYdlzsVcPKXfwkMc7w8Bvsk6PiLdQTpCfTDmof3VYocx8J0BEnAE8vNeSFxHvAE5quO2XAq8D7oiIP9Gi9bZuayfKj7f/gPauhttetbY0PYNycvyXXnAxRvkZn9z7vMa0emb+aMp22gRfv6iXVWh+gPw1JTVoZ0oA03MLJahpY22W7hjXaljm2Ij4JLB2bd37f5QD3DD/CryIkrrcH9jewvBgs+eP9eTTzkQgAAAgAElEQVSsF3RuT2kBbeqmzPx6i+f3GyfoHOcE9/aIuAdL3/PmwNAT3cw8vT7/g5m5bd9DX42IpqllCyLibr0T61qPRmm5wMLaOr47S4PlNkbadkTsDHwQuC+ll/H+wKWU/VIjbU/MZ+mzBrh3bRTYr77m6RHx4yFlPkdp6HofpbGr55bMbHrS0wt69qSc5J1HCUJ3oASfjxtQ9LcjBn1QeuueB+yTmb+JkrrXtKEPSsPZqjT4Pczg95T9T88t9b4mVs/Mz/bdPioi3jisUG0I2Y+yHzyf0sv6A2Bor26fe/aCPoDMPK1pA1xmXlGP7bfXy+OBv2+43b9k5k1TjncDj9GzdJwFuIp2+/p+21IahlunqmXmpyJiNUoAuAh4aWZ+v2Hxcc8PHgo8JjP/ChARh1B6ancALhxSdvPMfFbf7XdGydZpYrvM3Lrv9nci4oKGZUftuHgg8DTKOUn/sJNbgJcMKXsucC2wHmXf31+2aWPj8fXSSm/fvzIw8JusxcA+lB/2SymL03+6RfkNKDv0ntvrfUNl5sitcxHxCUrPzeMp9d2N5j2NUFrFrgQuAM6IMgZq6NicIeWbpDXcnfJ5Tw1Ym6So/q6eiPdOynej7IAaGeWgmJkXABfUlsOFwP2apmVN8T7gvCi57UFJ61g8uAhk5sG1QeJmys767Zl5SoNyRwBHRMSzMvO4EeoLpVHiRGDziDgLWJ/yPRsoInrpH9+NiA9QdvB3nixm5rkNtj1O0DnOCe4BwDeATaKMCX0MJYBu6p4RsVkvBa220Dc6SaSc/J8aS9O59waOaFj2XcA3gTMz88dRUuN+3qLeo2773ZST6W9n5sMi4vGUgKaRMU/Mx/msAXo9a9fWRrRfU9L5ZlRTk26KiLcCv8nMP9ceoIdGxJFZU2UHiYgvU37Ln6X03vX2Ycc0CFyXRElH/grL/qaGnkRl5m/oawSqvbJtxvjdCpwfJU2rf9vDWul7LgPOjogTKL/rXYCfRE1PHJKG+PWIWAx8oZZ9DnBy1PFfA4Lu/SgZOT/MzMfXno73Nqxvz+UR8TbK/wvK97tRimlE/AL4HaXB4DDgVb2sjwYujojnURpltgReTekhbbLdcY6zUN7faVFSL/v/101SRS+i9MQ3PjbHsimqAdyPuj+IiO0bbnes8wNgHWANlh5n7klJofxr7akd5LaI2CEzz6zbfgxwW8Pt/jUiNs/MX9Sym9E8Q2bUjosTgBMi4lGZ+YOG2+r5fJaU8V+MGojV85OR1d/D+7hrmv7QYSTzhWP8Jqi23P2pr5VnAXC3zLy1Yfl/p7Syf7ne9QzgmMx8X8PyI6WhRB1P0vd3DeDrmflPw8oOeM2FmdmohSwiNs3MK/puB7BFZg482YyILwI/pZyYvwt4PnBpZjYZFLwZS1NsbqCkYeyZmVc2rPP6wJu468GwyZie1uMtpnmNDSknIQA/qidiw8psClybmX+qt+9BGcNwZcNt3o2SsrKIZcdbNOoZri2ID6QcjH82JA2tV2bQwO1s+Hk/HPg48GDKicT6wG6Z2SZ9cSQ14Nye8p5/mJm/a1F2R8p39PJa/v6UVutvNiz/FKA3PuOUpuVmwyjbjoglmbltbaF+WGb+LSIumNKCPaj8hSw9Md+md2KemUNTpGb4rPfNzG813PbTKC36m1C+a2sC78zMEwcWLGXPp/RsLKI0Fp4A/ENmPnVIuVWAt2TDcUfTlJ9ujHcOOqGPiFuYvhegbYbJXtPd3/QkLspEVTMa1DAXZUzogKLTn/BFxI8zc7v6/3pkDdQvzsy2PdLvpPT8QPnOvCMzb2hQdr9abhPKce904IzeSf6QsqtTeu//hfK/+ibw7t6xYEjZkY+ztfy0/6smjad1/78NpSG6P2gcNDZx5O9G32uMe36wDyW1/TSWNs6+l5JS/I7MnLGHuZ4PHEHJ5AlKZs+LasPxsO0+kZIW2r8f27u/l3lA2VUoAf6d35HMbJIR1CvfuoEgIi6ifC7vpqTWLmNQI1REHJuZu9f9/l32Sdl8DOmZlEbaD1N6LPcGVsnMtzcpPx8Y+E1QRPwQeFJm/qHeXgP4VmY2nnGznqT2Aq4zMvO8huWmbe1ueGJ8dmY+stZ/V0rKzMWZucWQcrM1+HukQcFRB//2BayrUsY8bN9ku/U17kn5kTceKF/LfYsypucN9I3pycw3Nyh7DqUX4rRsNwnQwIkThvV+1R6AR2edkCZKGsxZmbndoHJ95b9BacFcZpxdZn5wxkJLy0538n0TcGFmXteg/J29MYPuG1C+ddBZy013onsTJWX39YO2HxHPBL5Te3aIiLWBx2Vm45l+a7D9oHrzpzn6mKjGagPBq7hrgD/Rwe4R8W1KY9f7KKk/11FSlxrtP0c9Ma8nPNtTvtetP+vawPfqzPxwk+dPU743UcqbgNsy8+PRfGKD1pN6zXf18z4oM5uO15qt7X6ZclL4Gsr++wZg1WEB+gTqsUatxxuAjTNzwYS3N/ZxdoxtTzuJ1qi9QyNsf6Tzg1p2Q8qYOSiTr/26Zfk1ATKzTfZU75jRm731Zy32Y/tl5keH3TegfOsGgojYoT5vd5bObdEzrBFqw8y8NmaYZTszf9mw3udk5iP6z8GanH/OJ6Z6Ttbde0EfQGb+oba0DVUPZhdn5oMoec9tjZOG8rV6UvqBuu2kWYrqWIO/Y4xBwVXv5P3GiHgw8BvKDHWDtjltsBo1r71psMpoY3p6Wo+3qAYFWMnwlLaF2TcLaZZ8/tUabLdn48zcscXz++1DmUWv1/L4OMqJ9qYR8a5cdszNdL4ETA18vwg0mTHsFcDRmXlxvb1OROyRmf89pCjAR4CrKelVQZkIYnPK7+QzDB5DdUBm9nrvycwba2t0o8AvIl445a6tI2LgjHYRcWZm7jBNwNqmR+YrlFSyr7J0Fr4m9R1327tQUppeSzkZWItyAtHU1XU/9hXglIi4ARh68K89i/9VA6im42H6y/81yiQFIwV+wF9q+ReydIxM03GkrSf1iog3Zeb7I+LjTN9S3jTdkmg5EU9fuZHTq+rn/ZimdZxm21NnPjwN+OSwxqDM7E0G847aE7UWJZW7zbYfQAnYFrFso0qTBtoPUnr81qCkML+d0mM4qMxszFzY+jg7pQ4jZ8eME+BFmezj2bns7MJfyMx/bVB2NmYmXoUywdNCYIuI2CIHZGDN4rnJI1j6/dpm2DGjz16U2W77vWia+2ayRWY+OyJ2ycwjogxpGfj9zJLOemaUbI8mk631l722/v1llJm7+zOghjYm9/lzbfz7eUS8EriG8hvrDAO/yfpjRDy81/MSEY+gYW52PZj9LCLu1/TgOcWfMvNPEUGUiRV+Gg3X7MnMd9erx0XE1ygB7NAxUDn+4O9xBgUDHFp3yG+ltBatQTkYDjJbM5W1HtPTZ6TxFpn5+JFqutT1EbFz1vSzKDNVNk49BL4fEQ/JzGGD06ezEPj7zPxt3fYGlDFBjwTOYOmYl2XMQuMAwEsy8796N7LMMvkSls58NsjOuWyq4aERcX5mvjnKeIhBpltbrc0+uL8n9u6U1MlzGTCWKjN3qH/H+Z7/KTOnm458oFnY9t+xNBX5iKipyDSctGPME/PWAdQUZ0XEf1KyAPpnlmvSiLc3JWvgP7JM4LEpM/wepjHKpF69CV1GWoMOIMafiOdwlqZXPZ6aXtWiCudHmQXwiyz7eTeZ5GGkmQ+jjA2+ODNvyczTa4/Mw6izIDb0ReATlIbVtrMT/4ASjLSZbXY2Zi4c5Tjbrzfj7dNoOONtT/3MP06ZxGY1YAHwx4YNWOtn3zjZut9vGrCOOzNxb7mSi+lbwoJyrJvJ2OcmMcIsqDE7M2vCeA0En42IVzPaMhS7UzotTqPs/z4eEW/MzC813PZ+lDkuXk1JOX08pRGuM0z1nKCI2I4yYPzXlC/gfYDnZOY5AwsuLX8G5UDyI5Y9mDVZt2vkNJS+HpH+1q2mPSIj5XZPKT/KoOA5FdOP6XlHZg4dDB1jjLeo5UdqsY4yWP1oyslaUGZbe2FmXtZwu5cAW1DGO/yZpSeZTdYbuyQzt+q7HZSTqK1iQLpaDU6fQZkJtf+gdAul9XZowBxlDMBDeyf0tXf9J9lgbE5E/IBycto7iOxGmY59+xoAzrgmX0R8BriRMnU8lOmy183MFw3b7gyvtzblPQ/tda3/66uzb8IQytIMTSYMeR5lrPC3aD+RTu81FlCCtv5ejYENWjFGKnIsmzHRWu2lvCdl5r5RZkWebgxNNunVmI+ijMN8AlMm4snMRkudxJjpVTHC+MT+uk9pzJn2vmnKnUeZdbu3H1mFsi5bo7ULa5mxUshqwH3nyXGT481c6/tf37k2ZdS07AZll1CyLL5IGQf7QuABmbl/k+0Cz+ztd6KkBH65yf8ryvIq2+WyMxMvaXLMqM//GeWYM/HU/CnbvZSWs6DWz2VTppldmHKcbDpXw4uB4yjHmsOpDQSZ+YkGZT9NaYzpjfF9AfDXzBy6DEXdFz2518tXe5i/Pez33Ff+2Zn5xWH3zWf2+E1QlhnwHsSy+dWNxhJVbxtj29O1djed/n6cHhEordM/pUz7f2dud8OyUGaofAUtA8dx0jHGDVaBG2qv6E2UFiKiYfpRlsl+/p3RpsqHEVuss0wCsH2UMSJkX1pyQ09p+fx+p9Xe5P51jU6LMoZixmAkx5sxrOcblFkOP1lvv5TmPUHPp6S6/Del5fSHwJ71ROCVQ8q+ivKbPqbePoUS/I3qj5QDdBPHAdtGxBaUSQpOoKSrNhmP9BDKd+oJLNta3SiIiRkWbKacEAwyciryuBkTY/aQjtUbH2OkPUbEqTllgeXp7pvy+Kyk/2Xm7yNilYhYJTO/GxEfaVCuZ6z0qhxvDdpRZz6M/hPqLCnCbc+pvhoR/0aZwK2/UWVoz0pEvI8yZuzoeter635x6JI6Y37Hxk17HCc7hsy8LCIWZJk07/AagA8N/CjH1zMj4nRKQ84/URYob2KcWZFhhOVKYvpF3++UzVKwW8+CmmUs3C8pQzFGlpm94UGnA21nxBxnGYpVctnUzt/TLntgf5aelwy6b94y8JuAiHhCZn4n7jqBxQOi5Fc3WmOkpo+MlKscEZ/NzBf0Xqd3H+UEbpgFEXHnQa22nrcZ+9U6t3uKUQPHcdIxxg1WP85dx5xNd9+dIuIjmfmamU68Gp5wQcudZETsmZlHxZQxBNFy7ECWXPodgC0z8/Dastb0ZO0VlGCvFxwfCRxXv3NNTpovi5JauYhle5GaBOpvpgR7L6+3T6HhMitZJm95+gwPnzmk7B9psMzGTKZ8TxZQ0p2ObVj8b1nWK3wm8PGsE4Y0LPtsYLP+IKylURdsHjcVeR1KGnXjjImIeFCWtPhpf7dNeznrfvu9wH0z8ykRsRXwqGw2bqV12mNtuFodWK+eiPcGDK8JbDRke7OR/ndjbUA6Azg6Iq6j7zNvYKz0qjEb7t5IWR6mNzHTIspnPszlUdLRDqm3/42GSzH06c1m2j+DYdLsRHknYJusSzhExBGUdRubrKU6TmrtWGmPwHsiYi3g9SzNjmm6bu2ttfHn/Ih4PyWgaVTvzPxG/V33JqF5TTacVTkzD6q9fr0GlHdnu1mRR1mupFFm2BDrAZfUfWCjWVB7Ytmx2atRAtemabXj7gPHWYbiGxHxTcqMqVCXZ2lQ36dQGkI3mhJ0r0m7NRtXeAZ+k/FY4DtMf4KYNFxcMsbLVV4mBaEGb01TSsbpEYExB38zeuA4ziLVI20zIh5FmeJ5/SmB1JqUk/NBeuN2xj3xaruT7K1HNu5kPAdQ0m0eSDmRWBU4iqXB3IxqgPcllqZMtnUC5f/zbVqOjaknSoew9IStsRrcvoQWAecsBvj935M7gF9m5tUNy/YmDNmL9hOGXEQZd9tmgHy/URdsfhkliOhlH1xFs4arnlEyJl5H6QWYbuKkxr2cwP9QfhO9Xvz/pfT0NjnpuUdmnlob335Jydo4h8HjqF5KSeu/L+WEsRf43Qz856CNZd+EGfWk+gH1ZpsMlXEn4knK/vD+LP1eforhvcI94zTcnUVZO/aJlGyDb1LGzw3zMuBjLA14vk3zHiQAMrNpj/1M1mbpuKu1WpQb5TvWM85xlsz8Wr16Z3ZMCy+gHFdfSfmubUJpQGzq0SxNjQX42kxPnCozv07zrKmpTuSus1QO294yPYoRsXo2XAqszztaPr9/+3eeH0RpFd6FpUFzE//D6PvANzBaYwyZ+cYo47N75yGHZt+kagP8mjLOeWeWDbpvoXnDxLxg4DcBmXlA/TtO+gmUH8x2U3OVGXCyHBH7U1r87hERvWl/g7L4+6ENtztyj0g17uDvUQPHcdIxRt3mapT3t5BlA6mbGbIgedaxnjn+VNStdpKZ+cn6d9zJeJ5JGYN6bn29X0dEo2AyxhukD7B6NlgqY8o2Z2Odn1ECzlkJ8KfJAGiziPo4E4asDfw0yiy1rVqNq5EWbB43FXmUjInM3Lf+HXfipPUy89i6P6b2tjb9vrROe8wyxfpHI+JVmfnxUSocZeznEcCVlGPGJhGxVzZY+5UxJ+Kh7LvfCFxIi5lj+4yTZXIkZX/dm9TseZTfxrMHFarfpeeOUNdl1OPN1JTLJrMuvo8yLOK7cOfacE0zCsZJrR3pOBuzMHtsLp2S/zbK+oeNRcSBlH1BLzV2v4h4dMPU2F2Bgyjf86DlmN+pQVwbtXH5MMr/534RsTVl/dZ/a7DdWVnmojbUfqU29jb9jo2zD7w3ZZ3dRZQx/Y+iReNhZh5HGd7QWJZ1ES+o+46FwP0y82dtXmPeyEwvE7pQWuA+RGlFWEJpRV6rRfkLp9xeZep9A8q+b8y6r0YZ2/NgyqQwy/NzezElTeuxlJPG64CXNSz7FMoJ9sHAvy6Pbdby9x/j/V5Rt7nMpUX5Z1N6GB9KCbC/Tpl0YFi5jSljS66rl+MoSzQ03e6P6t9z6997UgZ/Nym7hDIxzHmUoG/vNt9Z4D3AU1t+zhv2/lfTXRq+xvmj/p9r+dXqb6r174qyttEvKSdZR9bvzW4j1GEdykQDTZ//2OkuLcofMN2lQblx958jf16UtMO3UlqLoUxu87QW2z6NcvLS+21sT5l8Y1CZz9a/b6Kc5G1MaTE/Hti+xbYfTQleXti7NCx3DiUlt3f7AcA5DcsuAVab8j3/cYs6n9n2ezylfG9fdEb9ba1Hw30ocEmT+6Z5zlj7z/oaB1CWtPlt/V//BvhSi/IbUnoodgbu06LcdlO+Y8dR1rpsWn5HWh5ngafXv3tNd2n4Gk+jHDP+jxKs3wLc3LDsTyjjv3q3e5N6NSl7GWUW6rbfy2Pr3wvr9pe5NHyNsyk9m+f13XdRw7K7UhoIbxrh89q177IbcCBlLeim7/00Wu4D+/9X9e8O9fexE3D2pN9z73sK/Ay4ot7eBjix7f9+Rb7MeQW6fKk703dS8vU3qzv541uU/wAl7eRF9fJ1ykK1g8o8qP59+HSXhtt9HOWE6XTKgfQK4J9b1Pu9wNp9t9cB3jPX/48J/Y8/Uv9+laXpHHdeGr7GvfsuG1FStt7Vog4j7SQpPbl7U1q3Ftbv2CkttvsGSorU5ZT0xx8Ar2pYdkl/3ev181ps+xZKz8BtI+7cN6CcRDwN+LsW5VoHnH1lx/1dXdBfV2B94IKGZU+jNA6sW7d7NvChUd7HqBfKieYaLZ4/7v5znM/rGEoAdlG9vTotgv66vz2LcvJxFiXNaWCwDVxCSdW8oO4z1+2/NNzuZylLwfw3pUf948DHGpa9y4nodPfNUPYun03Tz7o+94mUrJI96DvhbFG+13D3zyxtuHtpw7JH0RdYU5aUObJBubH2n/U1LqQ06F5Qb28w7DWY4dhOu2P8tpSg9dxahwub/q/76vl0Wu4/x71QArCHUmekb1n2J/2/o/q7avr9PmvE+s5GY+PZ9W9/4Nd0PzZSwFrLHt53+RQlA63NsbK3D7yx6T6wr+x59e/7gOdNff+Tes+1/DmURsf+z7tRh8t8uZjqOVmbZ2Z//vk7I+L8poVztFzl6caoZN/1JmNUPgj8S9Zu7iiLzH6e5mMEx13zptWg4JiFRaqjTI3/Qu46dmtY+snYaXx510kvPtJivAUsTTncCfhUZp4UEe9pUG79zDy87/b/RMRrGm6TzDw4Ip5MCbweSJmq+ZSGxUcepF+3PfL4xDHHzu4HvCUi/kxJD26T8jPu72qc2crWysybo0yxfWRmHlAnKxhqFgb5P5jyO1m33v4dpRfq4iFFx9p/Mt7ntXlmPqeOiyQzb63jXBrJzHMj4rGU30XQbLzcJ4BTKUFub5xe9v1tMuHHtrScvr3PkijTqB9Vbz+f5mv7jTsRz97Agyjfrf6ZXxuNh6d8v55F2X/30uo2aFj2EZQ1SXuzv94P+FkvJTxnTgEfa/9Z3ZZlNtA7oqwDeB2ld2eQmcaf9jQ5xo+cWjvq/nOmMc492Sx1/CpKY0yr73f97R7M6KmxSyLiGOArLJuuPvD7mX2Lirep7xRXRcSjgYyydNN+NB+/+tvMbDNJ3Z1y/KFKl1AaF26lNMx+hRL8NXFNlDkmngwcFBF3o/m+e+T3XP0lM2+asrsfZX+6wjLwm6zbImKHzDwT7pzev9EC7j3ZMlc56xgVysQV36gne2+jtL68e+aSy1g1+3KbM/N/6w6nqbEGf9NyUHDOziLVJ1Om5m91IMxZGKcXy84guArl5K3Nb3PUneTvI2JPls5+tQfNx+QAUAO9psFevxdQ6jjSIP2IOI7yffhG1lntWmg9drZnzO/YuL+rkWYrqxZGxIaU9MdWy4bk+IP8D6Wsdfjd+hqPo7QgP3pIuXH3n+N8XrfX/VbWbW9Ou6nY706Z5XGH+hrfi4hP5IC1OTPzY8DHIuKQzHz5TM8bovX07X1eTpltt9fY9T2aL+Ez7kQ822XmA4c/bUYnUHpXz6HF/6kaug7mDMbef1ICirUpv4dzgD8wZGKZrONPawA26jH++l6QPoJR95+9xtFdKd/RXgPDHpRU1ybeBJwcZUmGNuOFMyLeSNlv9cb8vjkzf9Nwu2tSAph/6X9Zmk/UN7VRGsr3dQllKYxBs8G+jLKE0EaUsZjfouxbBm2vN6N864A1ZmcZCVg6dva99XajsbPV7tR04sy8sR673jiowDjveYqLo6xduyDKsievpmRRdIYLuE9QRGxDaX3szbZ1AyWXvWlL+8gDiqMujhpluv13U3a6b8/MRzYo+xlK8NPf8rsgmy/A/mZKGkivNXRvStrj+xuW/3Fmbhd9C3nHkMWx63PGWaT63Gyx8O405a9g+gHrTdZF+m5f2TsokyscnJmNWseiLAC/IyUd4ed1J/mQzPzWkHL3p6SCPapu//vAq3P4wtpj97COKyKeRPlebU9ZX+fwbDgQO/oWia63e6lWDxlQZuxp/sf9XdXX6M8A+F6DDIBeuWdTZrk8MzP/LcrMrx+Y0qPWWP9vs8FzR10ge6z9Z32NUT+vJ1MyFLainGg9BnhRZp7WsPyxlFbu3v/6eZT09yYnPSOr+5JtgNbTt8/S9keaiCfKRCEfyMxLRtzuRZn54FHKjmrU/eeA11sErNni/GCcY/wTKQHX1OUFhp4cj7L/nFJ+SWZuO+y+Gcp+ixIcL9NAmw0mKouy3MV/ZuaPm9RzNkXEu4GrKWunBmVSoM0pqbYvz8zHDSh7BGXpiRvq7XWADw46bsTSiXemk0PKXk0J7teh7HOnFm40UU1EXJKZWw27b7b0vedelkS/ge95yuusTnn/vSD/m5TlO9o2KK2wDPwmKOoiozWFg8y8eViZKeUvowyIbt1t3Tsxi7LI64WZ+bmmJ2u1x+gVlNZqqC2/bb74EbEj8KR685RsseZNRJxG6f05JTMfHmUGyIMy87FDyp1P6S1bRGnZPwH4h8wcukh1RLyWckD5Gi0X0q3l79138+6UVq11M3NoumZEvJ5ld1bL/CiHtWaOKiLWz8zrJ/Hay0uU9aD2oOyor6K0nB+VA9LqoqSWbs2yPUE/yQGzhEbEoZm5bz2xniozc2h61Wz8ruZCLLseaa9H+rGZ2WiB34jojSXqpUXvCTwiM585pNxY+89xRMRRlDFBt1HGjJ2dDdf8quWX60lP3zam3UcOykiIWZjttv4OD2DpVPmnU8YpN5qJLyIupZwIX0HZ//YakRot5xARh1LWp7ywyfNnwzj7z1lqSBrnGH8UJbX2YvpSa5ucHI+y/5xS/lJgp14vV5QZhk/OzL9vUHbkAD8ifkqZUOyXlDUmG3/HYrx1Imdq/Do/M7cZ1gg23f+0TcNbWxFxCeXc7euUcenL5jw2Pyc6ihJo/7DefiTwisxsvD7nKGqgvF+vwb9JoDyl/LaU84lFLM28arwvmg9M9ZysKyLiG5Q0xe+MUH6cXOVxcqR3Av5rzIDjPMp4jazX23gdZXKUzSLiLMqkDAOXRqjGWaT6dsq4hX9n6clP03E1447TewQl/eQEyk726ZQW+zbT9Y/irIi4kvL9PK5Jz+hU9cSll852Zma2/V+PrAbbe1JSys6jjFvZgTJL3OMGFL2O0hPT60EeOna2Bn2rAG/NzLNGrPJIv6tpelbvfIjmGQDjnLj0r0fa65HepUG5nv9HmaTleGraY71vmLH2nzHeFOyHAf9E2X9uThkbdEaWZROaODcitp9y0tN0vNzIBgV4A+xX//4PJd296dqQ/T5DSTPdvd5+ASXjY9cZSyxrpHTLvmB1IbB3lCVtWgeOIxpn/9k/Fv8uWRM0G6c3zjF+nNTa1vvPKV5LWd7lcsr7vT9l+agmTo6If8kh2Swz+NcRyvSMs04klDHtu7M0HXY3oJf2Paz3ZZWIWKevx29dGp67jxgETR1rfOfL0eCcqO83uSpLx84m5f/80yb1HtND+3+LWeaYaBMkH02ZuO4iRltaZsWXK/HcxaMAABeCSURBVMAMM129UGaC251ywnMlZSHdHRqU681q9lHKQaX1TGd127sCW9bbG1ImlmhS9nBKq9hnKbN2LWz5vseadp5yYvoGytix4ym53XdvUO7s+lldBGxa72s67fHllHVnRv1f98+sti0lL7/pzFtnAPfqu30v4Izl9B39R8qU+ZdTejv3bFH27ZSUm3fWywWUwKhJ2c3HrPeXKYPH96fOnNb32JIhZQ+gtHR/jzLGcIMW22088+g0Zcf6XY35eX2Rkg72C0pg/C3gow3KLQBeO8Z2F1DSlkcpO9L+s6/8uLO7LaCkEu9f/28/bVH2UspJw5V1//e3el+r2RNHqPMtlHE1N1NOLP9K8+nbx/ldTDer51hLnzTc7rSzJdJi1sQxtz/y/rOWvwfw+ro/O54SFA091tWy4x7jtxrxPY/8Pel7jbtReg23Bu7WolxvNuc/1eutZnMe4//cm2WyN4P2qsAPW5TfjDLz9++A6+v1Lf5/e/ceJVlV3XH8+wN5iIiooMQ1AQGJBBEQxYgBRcFnAN+6kKBiiIAajA9WjKDioBLjJEHJyhgCEpyIOkYEBB8YYIBheDnhpQSiUWOCYRnlsWAEwbjzx741Xd3T3XXr3qq6dat+n7VmMV3M6Xtmuu6pc+45e+/i57/omEYmnbutGL9PLn5/RD/97vXaAm2XV/y3avqevAl4bNfXj6OPrJzULC3Thl8+6jkixZOWTwGHR8TGPf5s5fPZg6JMOvEy8hjHfuSxy6NKtr0JeFHMCf6OHjE9Xe1XkhOXTqHVUvExyuyfx5C1Zr5QHCF5fUR8osQ1LwZeGRG/LNPHedpXjtOTdDv5lKqTDGcz8gOmTrKDvkjahpzA9Hx/drW5HdgzioQVymQYN5bptzI4fwlwPTmBuCL6OKal2YkNTiQX3B+NEkekur7HHuT7+zVkbOhBPZogaRmZfOHcqDB41rmv6ug6FtaJC9qEjHnrmaRF0nUR8ewa176mzHV6fI/S42dXm6si4vd7/8l5215C1qW8mnx/ro4exd/ntN+BjJHZv3jpCjKtOVA7y1/ZPqxPxBMRZbMXVr0vrgaOj9mJeJZFyePAbVdl/CzazfdZ95iIeP3Creqre7S2+B59v0+62j6XDTNolyla34jOGCjpCjKxyp1k7chSJ4IGcP3dmNkFvjRKxsIWc7EDYvZu4eVRMh6zjSS9CfgA+bATMuzmYxGxYuFWs9pXjn9tCx/1HLIi5uIN5FGW7zBzFGZBUT+Nbm0R8bCkb5CLmUcCryRrJZVRJ406wO4xOxbmsuLc+aKKwfA4WD9RfHSZRV9hHVla4DJm3+xlM1hdyIZxevtL2iIieqWg/xxwXRELBflv/Y8lr1tZETv1KmYCzb9KPsEu66fk7mznyMpmZNaxniLi+cpyDvuQxzIvkrRlRDyu5LVPjIiVysQGB5HHdJeTNbjK+hn5Af4L8jhgGUeTx7R+LelB+js+WPe+qqMT83iPsrzCnZT/O18l6W/J0wfrOi/2sci+QdIF5Adxd/syiST6Hj81mOxuN5NHsHcns+/dI+nqiCibVbTzcz2XfI+sIEutnFayfW3Fg4nzJH2Y8mnrodp9cSxwtjLWD4pEPH1cs3UGMH5Cxc+6AaiaybRblfcJklaQ/143MlOKKMjPwTLtD2UmlnRVRFxY9to1nF7MKU4kw1C2JJNllVI8/P5jNlzslnqIX8xtqrwv/gq4WtKsRVCF79MaEfE5Sd9hZqH86rIL5ULd0jJjzzt+Q1Sc/78BWElmtVy3eIsN2tcKUq1KUmdH4gCyVs9K4OKI+HXJ9nWDvysFBSuTwhxKDqxryQ+mqyLiPSWuOe8kJcpnsDqHPOJ5ATnRO5icPD4Z+HL0yGhaxMqt3x2IEcTKKTORngesjIhFU4jPaXcaORBuTy7cvl18/SLyKWjPuJ5iwbZ/8WtrchJwZUR8YdGGM+3rJDZ4O7mA2JZcjKzs84Ohkrr3Vc1rH0WWhdmDPOa1JZkB8DMl2nYS2nQ+LDqL3TJxSAudYOh5cqHq+KkBZXcrvtejycLc7wO2i4hSZWmUNRL37fRZ0qPIkwhDTRCgGol46twXajART1Oqjp9zvkcjCTDqqDt+FruNlWpNSvoL8jOns0N6GHm0/8/7/V4lrzff3GH9w90oGa8taQ15cmAtM4tdIst1DVXV3cJpJen2UZ62aoJ3/IZrj5ofgHWDVKt6E/l0/+iolnGwUvD3AIKCKxepjoizi6OK20fJsgBzLAH2jiKNefGU/SLyyeRaYNGFX7F7UvqY4oDsVOXDl5kkFWvJp9wdq/r4HquK9qeQGd0e6rMPdRIb/DaZHrufYuBAHgGMiAN7vbaAuvdVZRFxRvHbyymZsKhr0jN3Nxv6K2h7RsxJiFMcBeyl0vjZOTGx0IOzMt9D0jvJhxLPJI9tf5acuJUluiZ4xe9LF4CvoU4insr3BfUTmbVR1fGzW9Xi8U2q8z6BerUmXw7sFUXt1uIev4GMwx2GTg3Tp5ILzk7tw04CtrK2KPvgewgeB6yLiLMkbStpx4j4UUN9aYM1knab5AWyF35DNICnnpWzOdUREYfV/BaPIo8W3UVOBMoWvzy45nUrF6mWdAhZB2lTYEdlDbGlUb7+1ROYXTj4YTLo/QFJY5muv+qkpewuaA/bkLXRngccJ+k35I5I2eMzfRd47ajydFiZFXMLYJtiAdGZxG9FFtYtc92691XfFnhivV6PJ9ZzJz1zs86WdRoZg9nrtbl9qzt+1nlwtjkZs7W24o7sWcC1c45vn1nh+/SlTphAzV2TXcnx+x3AmZIuBL4YRczfJBrAog8Gc+RypAawu7YNcKukqrUmtybnFjBT43MooqgPWMT27R0R9xVfn0Q+2C3rQkkvj4ivD76XCyseQD+LHMPPIh+q/xMztU1tQ88hw34qx7+OOy/8xlv3+WyRKYDH/nx2MVh+RDPB35dL6hn8HfUTHiwli22ujojrlUWqy5ZEOImMz1hV9OXGon1ZnycneucXXx8CnFMc8ZrIJ0eSDiazjO1AjiWl492KxdoPyafHS4Dnkh9KpUQm4Tm36+v/odoT5LKOBv4UeBK5U9lJbX0fuYgZV53F27zHHhdrWHfSI2lf8ue67ZwF6FZkxsxhq/zgLCKW1blwRPx1cfS8U7PxyBEd315Cvh/XF60ndz2rlGgorbgfVwIrNZOI53JG83NurQF85rXRSTXafpwslbKKHM+eR3/xq1U9kSz51PFQ8VpZ7wI+UDwEfpg+Y8NreBXwDIrTRBHx0+L4ui2sdQ9j+uWF3xiL+kGqTasU/F1VRHyZmUxORBaIfU3J5g9HxL3SrLlx6RouEXGyMmlHZ8J1TER0jkQeXvb7tMypZDrxW/p9+l0s+m4DVpNJWY6scNxzZCLrt31K0oeAU4sjxR8kd60qxfeMQtfirfKxR6pPejYlYwkfwcwCFDKLYZm6nHU1mtigoePbZwHnkH9XyDqXZ5FHoodKFRLx2PSJarUmOw4mj13fTR5j/rOIuHMQ/eqhVgK2iGhqsfVQRISkgPWxxraIaXgY4+QuQ6TMcHYSM0k7LiePD97bo91WxcRy3gyHEXHXfK8PUp14twaTZ1QuUi3pTDJ97/vJxeJxwCYRccxwets8Sb9DLrqeGBG7Fzu0h0bER0u2vww4sBNv0ee1N6rSrmmaKYewH7nbuYxMklIqm2jNONLK5kt800cynBPI+7l70vOliDil5LV3qPJhWnX8nPM9piqxgaQbI2KvXq8N4bo/pkYiszaqO35OG0mrI2I/SfcxT9H6Mrtfkl7ATFKwncn33BXFg7mhUoUEbJJ2jYjbirYbiD7KD1Uh6X3ALuSDn1OAtwLnxAizC9v48cJviCR9hQxk7sREHUHWPVs066GkCyPi4OKM8XwD5FBrx3THu0VE3/FuykyLX6oR/F1J8WT/NrIW0lJyp+3fIuJdi7RZERFHSPoAGZv4YvLf+VvAyVHUqJtEylp6xwN/31kASPpuROxesv0+5OLncmbHavTMdNbWSZPqZROtdV/V7Hetek5VJj1dbb8NvG7ObuMXI+IlPdpVGj+nmbL+4FnMZFQ+jNxNL5N8qM51txpATGar1B0/rRpJG5Mxxy8g6/Y+EBG7Ntur+Uk6PSLeppnMyNA1p4uSmZFrXP8TwL8we15zUDSXaMbGgBd+Q9TU09e6JK0ln5Kv6vpAu6XsJLEpqlCkWlkz6SDgG+QHySyj2F1tiqTrI2Kf7oVLP+9PZdH7+4Fb6DoW2zle2KNtKydNRdKKO8gnqHsDD5AlLPYs0bax+0o1i9rWvHal3ca2jp9NUhaOPw3Yl5xgrgH+JCL+q9GOTaC646f1r3iw8SjyeP2VZDz/zxZv1TxJrwe+OSdE4OQR7Pj9a0TsPee1m2OCEpVY/xzjN1wPSNovisxmyhTmZQsAd/78jRGxTtIfkoPFqRHxkx5N65ov3q0NTwiqFKn+DHnEcydmyhTATPKOoe6uNuznknam+NlKei39JUh5Uo2F2hYRcd2c99jQ69kNQOVsojR4X0Wz8cK/kbR9Z9wqFidl/t61xs8ptRR485yd3WXkES8brLrjp/XvZrIExu7AveRn/dURMe7jwokRsbIIEXgheU8uB0qFCPRL0rHA24GdNLuk1aOBq+ZvZdPCC7/hOgb4XBGrIjIF8Vv6aL8c2FPSnsB7gTOAFcDzB9zPub4n6Y3AxpJ2IePdypZkaNLpxTGyD5L1drYEPrRYg4j4NPBpScsj4tgR9HGcvAM4HdhV0h3Aj8hkEGV9XdKLI+LiCtdu5aQp6mUTbfS+KhZ6TcS4nQCsLnZ5RR4ZfVuJdnXHz2m0R2fRB3liQaOp/TqN6o6f1qeIeDeAMjPlW8hjzdsBmzXYrTI6NT3/APiHiLhI0jDDGs4hTzGdwuysp/dN8ikmK8dHPUdA0lbQf12qzja9MpPgHRFx5nxb94MmaQtysjY18W7TTJnpa6Mo0vX30e4+8thN3ymqlaUyTifT/d9NMWmKiB/31/v2mOb7StI2ZH0kgGsi4ud9tK00fk6jurGcNa5bOxFPW1UdP61/kt5JvseeSWb1vJIM57i0yX71UidEwGzQvPAbIkmbkRkin0zX7mpELC3Z/nLgm+Qxnf3J8gg3jXus3aipXpHqqTPIf69iYrkLs7Oolk7X7UnT5FOebT0c2CkilkraHtguIhYtAl93/JxGTcVyTlMiHn/eNEeZpfJKYG1EtCE0AFj/0O+lZEKw7xchAk+veFrGrBYf9Ryu88lz6GvpynrYhzeQGSrfGhF3FhOmTw6wf7NI+hqLxN6MIvtgRZWLVE+pgdQUknQUWZh2CXAjuaOzBuiZQXDupL4T9zaJk/oW31eD8ndk8p8XkjFo9wFfITPzLabu+Dl1Gozl3DkiumumfkTSSLM6j5ALYDckIpY13YcqaoYImA2Ud/yGaBBZCiVtBzybnDheH0MsVqoswLugfnZymqAFilRHiTp+1j9Jt5CT92siYi9JuwIfL/OUX9I3mZnUd+IfiIiyRcVbo+33VV1dR9a7sx/e1OuYUxuyvFqSdDVw/JxEPMsiYt9me2ZmZt284zdcayQ9PSJuqdK42FH5EHApuZN1mqSlEfHZQXayYwImoHt0Fn0AEXG3ExssTDUK3hcejIgHJSFps8hCtU8t2XZJRLy0zy630gTcV3U9rKy91Unksy1d5T8WUWv8tJGaukQ8Axg/zcxGzgu/4doPeIuyEPuvmEl+UbaGyvHAMyLiFwCSHk8epRvKwq+jyDh4CrAbsz/Qxr20wUaSHjsnsYHf4wtbQRa8fwldBe/7aP/fkrYGzgO+Lelu4D9Ltp26SX2L76u6Pg18FXiCpI8BrwVOLNGu7vhpIxIRN5EZqKcpEU/d8dPMbOR81HOIinpVG4iIUpNjSWvIDG0PFV9vShZ/fu7gejnvdVcDHwb+BjgEOJJMwLFoaYSmNZXYoK1UoeD9It/r+cBjyCK1D5X487cCTyGzeU7FpL6t99UgFMeADyR/zpdERM8Jct3x00ZnGhPxDHL8NDMbFe+GDNEAJig/AK6VdD55TOoVwM2drGJDzB72yIi4RJKKv8NJktbSoyZe0xouUt1GVQrez6vCccaXVblOy7Xyvqqq2HHv+Bnwhe7/16uelBd4rTKNiXgGNn6amY2KF37j7T+KXx3nF/8ddlaxX0naCPh+UTfnDrIY+thrsEh1G/Vd8H5QpnRS39r7qqK1zGTa7T5a0vl60o+4TpOpidnt0tj4aWZWlY96jqkiGcInIuJ9DVx7HzJWYWvgZPII319GxDWj7ovZpJjW+6pY7B4O7NhVx++3IuLahrtmAyLpdOC0aYrZNTNrIy/8xpikq50O24alSMzyJjaMyzmuqT7Z5JG0nKKOX0T8brFLcnFE9KrjZy0xpTG7Hj/NrHV81HO83SjpAjJZybrOixFx7sJN6pP0LOAEYAdmf6BN7If4lPo6cA1wC+XS61sNU3xf/V6njh+sL7OyadOdsoGaxphdj59m1jpe+I23zYFfMJOsBDI2ZqgLP+DzZCkJf6BNts0j4j1Nd2KKTOt9VbWOn7XElMbsevw0s9bxUU/bgKTVEbFf0/2w4ZL0buB+4EK6MvH1yrZo1UzrfSXpcOANwN7A2RR1/CLiy4s2NBtjHj/NrI288BtjkjYH/gh4GrMLPr91yNc9EDgMuITZH2jD3mm0EZL0DuBjwD3MZF2MKSgo3ohpvq+q1PEzG2ceP82sjXzUc7ytAG4DXgIsJTPjjWLCdCSwK7AJM0eyRnHE1EbrvcBTIuLnTXdkSkztfRURt5Fjmdmk8PhpZq3jhd94e0pEvE7SKyLibEnnAFeO4Lr7RMRTR3Ada9YPgF823Ykp4vvKbHJ4/DSz1vHCb7w9XPz3Hkm7A3cCTxjBdddI2q0ohm6Tax2ZOfYyZh89dDry4fB9ZTY5PH6aWet44TfeTi9qXp0IXABsCXxwBNd9DvmBNjU1mabUecUvGw3fV2aTw+OnmbWOk7uMMUmbAa8hC8RuUrwcEbF0yNfdYb7XpzRl90ST9Ehg+4i4vem+TDrfV2aTxeOnmbWNd/zG2/nAvcBauo6SDJsnotNB0iHAMmBTYEdJewFLI+LQZns2mXxfmU0Oj59m1kbe8Rtjkr4bEbs33Q+bTJLWAi8EVkXEM4rX/J4zM+vB46eZtdFGTXfAFrVG0tOb7oRNrIcj4t45r/1m3j9pZmbdPH6aWev4qOcYknQLWd/rEcCRkn6Ik0HY4H1P0huBjSXtAhwHrGm4T2ZmbeDx08xax0c9x9BCSSA6HCtkgyBpC+AE4MXFS98CPhoRDzbXKzOz8efx08zayAs/sykjaUVEHCHpXRHxqab7Y2bWFh4/zazNvPAzmzKSbgUOAr4BHEAeIV4vIu5qoFtmZmPP46eZtZlj/Mymz2eAS4CdyFIh3ROXKF43M7MNefw0s9byjp/ZlJK0PCKObbofZmZt4/HTzNrICz8zMzMzM7MJ5zp+ZmZmZmZmE84LPzMzMzMzswnn5C5mZmbzkPR4MpEHwHbA/wH/W3z97Ih4qJGOmZmZVeAYPzMzsx4knQTcHxHLmu6LmZlZFT7qaWZm1idJX5O0VtL3JB3V9frRkv5d0rWSzpB0apP9NDMz6/BRTzMzs/69OSLukrQF8B1JXwG2BN4P7A2sA1YB1zXXRTMzsxle+JmZmfXv3ZIOLX6/BNgZeDJwaUTcDSDpn4Htm+memZnZbF74mZmZ9UHSQcDzgOdExAOSVgObN9wtMzOzRTnGz8zMrD+PAe4qFn1PA/YpXr8OeIGkrSVtAry6sR6amZnN4R0/MzOz/lwEvE3SrcDtwLUAEfETSZ8ErgfuKv7fvY310szMrIvLOZiZmQ2IpC0j4v5ix+98YHlEfK3pfpmZmfmop5mZ2eCcLOkG4GZyx+/ChvtjZmYGeMfPzMzMzMxs4nnHz8zMzMzMbMJ54WdmZmZmZjbhvPAzMzMzMzObcF74mZmZmZmZTTgv/MzMzMzMzCacF35mZmZmZmYT7v8BoyUVuJpXC6cAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "temp_train_filenames = [train_dir+filename for filename in os.listdir(train_dir)\n",
    "                            if filename.endswith('.wav')]\n",
    "temp_y_train = make_y_by_filenames(temp_train_filenames, y_dict)\n",
    "temp_y_train_freq = temp_y_train.sum(axis=0)\n",
    "\n",
    "plt.figure(figsize = (15, 6))\n",
    "x = np.arange(len(y_names))\n",
    "plt.bar(x, temp_y_train_freq)\n",
    "plt.xticks(x, y_names, rotation='vertical')\n",
    "plt.xlabel('Tag') # , fontsize=18)\n",
    "plt.ylabel('Freq') # , fontsize=16)\n",
    "plt.title('Train Data Frequency')\n",
    "plt.show()\n",
    "\n",
    "\n",
    "temp_val_filenames = [train_dir+filename for filename in os.listdir(val_dir)\n",
    "                            if filename.endswith('.wav')]\n",
    "temp_y_val = make_y_by_filenames(temp_val_filenames, y_dict)\n",
    "temp_y_val_freq = temp_y_val.sum(axis=0)\n",
    "\n",
    "plt.figure(figsize = (15, 6))\n",
    "x = np.arange(len(y_names))\n",
    "plt.bar(x, temp_y_val_freq)\n",
    "plt.xticks(x, y_names, rotation='vertical')\n",
    "plt.xlabel('Tag') # , fontsize=18)\n",
    "plt.ylabel('Freq') # , fontsize=16)\n",
    "plt.title('Validation Data Frequency')\n",
    "plt.show()\n",
    "\n",
    "\n",
    "temp_test_filenames = [train_dir+filename for filename in os.listdir(test_dir)\n",
    "                            if filename.endswith('.wav')]\n",
    "temp_y_test = make_y_by_filenames(temp_test_filenames, y_dict)\n",
    "temp_y_test_freq = temp_y_test.sum(axis=0)\n",
    "\n",
    "plt.figure(figsize = (15, 6))\n",
    "x = np.arange(len(y_names))\n",
    "plt.bar(x, temp_y_test_freq)\n",
    "plt.xticks(x, y_names, rotation='vertical')\n",
    "plt.xlabel('Tag') # , fontsize=18)\n",
    "plt.ylabel('Freq') # , fontsize=16)\n",
    "plt.title('Test Data Frequency')\n",
    "plt.show()\n",
    "\n",
    "\n",
    "temp_y_test = make_y_by_filenames(x_test_wav_filenames, y_dict)\n",
    "temp_y_test_freq = temp_y_test.sum(axis=0)\n",
    "\n",
    "plt.figure(figsize = (15, 6))\n",
    "x = np.arange(len(y_names))\n",
    "plt.bar(x, temp_y_test_freq)\n",
    "plt.xticks(x, y_names, rotation='vertical')\n",
    "plt.xlabel('Tag') # , fontsize=18)\n",
    "plt.ylabel('Freq') # , fontsize=16)\n",
    "plt.title('Current Test Data Frequency (random_n == 500)')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "brains_on_beats_model_test",
   "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}