16.3 Prepare model for iOS.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "import keras\n",
    "from keras.models import load_model\n",
    "import coremltools\n",
    "import PIL\n",
    "import requests\n",
    "from io import BytesIO\n",
    "import json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "keras_model = load_model('zoo/09.3 retrained pet recognizer.h5')\n",
    "class_labels = json.load(open('zoo/09.3 pet_labels.json'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 : input_1, <keras.engine.topology.InputLayer object at 0x7f45dfa71290>\n",
      "1 : conv2d_1, <keras.layers.convolutional.Conv2D object at 0x7f45dfa71250>\n",
      "2 : batch_normalization_1, <keras.layers.normalization.BatchNormalization object at 0x7f46584678d0>\n",
      "3 : activation_1, <keras.layers.core.Activation object at 0x7f45dfa71710>\n",
      "4 : conv2d_2, <keras.layers.convolutional.Conv2D object at 0x7f45dbbcff90>\n",
      "5 : batch_normalization_2, <keras.layers.normalization.BatchNormalization object at 0x7f45dfa0cfd0>\n",
      "6 : activation_2, <keras.layers.core.Activation object at 0x7f45dbbfed90>\n",
      "7 : conv2d_3, <keras.layers.convolutional.Conv2D object at 0x7f45dbb97b10>\n",
      "8 : batch_normalization_3, <keras.layers.normalization.BatchNormalization object at 0x7f45dbba9550>\n",
      "9 : activation_3, <keras.layers.core.Activation object at 0x7f45dbb04f90>\n",
      "10 : max_pooling2d_1, <keras.layers.pooling.MaxPooling2D object at 0x7f45dbb71d10>\n",
      "11 : conv2d_4, <keras.layers.convolutional.Conv2D object at 0x7f45dbadf9d0>\n",
      "12 : batch_normalization_4, <keras.layers.normalization.BatchNormalization object at 0x7f45dbb34050>\n",
      "13 : activation_4, <keras.layers.core.Activation object at 0x7f45dbaf6b90>\n",
      "14 : conv2d_5, <keras.layers.convolutional.Conv2D object at 0x7f45dbabe750>\n",
      "15 : batch_normalization_5, <keras.layers.normalization.BatchNormalization object at 0x7f45dba52190>\n",
      "16 : activation_5, <keras.layers.core.Activation object at 0x7f45dba2e910>\n",
      "17 : max_pooling2d_2, <keras.layers.pooling.MaxPooling2D object at 0x7f45db9c4f50>\n",
      "18 : conv2d_9, <keras.layers.convolutional.Conv2D object at 0x7f45db9d7650>\n",
      "19 : batch_normalization_9, <keras.layers.normalization.BatchNormalization object at 0x7f45db99e750>\n",
      "20 : activation_9, <keras.layers.core.Activation object at 0x7f45db99ee50>\n",
      "21 : conv2d_7, <keras.layers.convolutional.Conv2D object at 0x7f45db9627d0>\n",
      "22 : conv2d_10, <keras.layers.convolutional.Conv2D object at 0x7f45db962e10>\n",
      "23 : batch_normalization_7, <keras.layers.normalization.BatchNormalization object at 0x7f45db925ed0>\n",
      "24 : batch_normalization_10, <keras.layers.normalization.BatchNormalization object at 0x7f45db8c0610>\n",
      "25 : activation_7, <keras.layers.core.Activation object at 0x7f45db883f90>\n",
      "26 : activation_10, <keras.layers.core.Activation object at 0x7f45db85b710>\n",
      "27 : average_pooling2d_1, <keras.layers.pooling.AveragePooling2D object at 0x7f45db8095d0>\n",
      "28 : conv2d_6, <keras.layers.convolutional.Conv2D object at 0x7f4658467090>\n",
      "29 : conv2d_8, <keras.layers.convolutional.Conv2D object at 0x7f45db7d2610>\n",
      "30 : conv2d_11, <keras.layers.convolutional.Conv2D object at 0x7f45db7fe310>\n",
      "31 : conv2d_12, <keras.layers.convolutional.Conv2D object at 0x7f45db7fef50>\n",
      "32 : batch_normalization_6, <keras.layers.normalization.BatchNormalization object at 0x7f45db794c90>\n",
      "33 : batch_normalization_8, <keras.layers.normalization.BatchNormalization object at 0x7f45db7acfd0>\n",
      "34 : batch_normalization_11, <keras.layers.normalization.BatchNormalization object at 0x7f45db753cd0>\n",
      "35 : batch_normalization_12, <keras.layers.normalization.BatchNormalization object at 0x7f45db72da10>\n",
      "36 : activation_6, <keras.layers.core.Activation object at 0x7f45db6a1950>\n",
      "37 : activation_8, <keras.layers.core.Activation object at 0x7f45db615890>\n",
      "38 : activation_11, <keras.layers.core.Activation object at 0x7f45db5c1750>\n",
      "39 : activation_12, <keras.layers.core.Activation object at 0x7f45db5c1cd0>\n",
      "40 : mixed0, <keras.layers.merge.Concatenate object at 0x7f45db58a8d0>\n",
      "41 : conv2d_16, <keras.layers.convolutional.Conv2D object at 0x7f45db58a7d0>\n",
      "42 : batch_normalization_16, <keras.layers.normalization.BatchNormalization object at 0x7f45db54dad0>\n",
      "43 : activation_16, <keras.layers.core.Activation object at 0x7f45db54dfd0>\n",
      "44 : conv2d_14, <keras.layers.convolutional.Conv2D object at 0x7f45db54de50>\n",
      "45 : conv2d_17, <keras.layers.convolutional.Conv2D object at 0x7f45db55c790>\n",
      "46 : batch_normalization_14, <keras.layers.normalization.BatchNormalization object at 0x7f45db524850>\n",
      "47 : batch_normalization_17, <keras.layers.normalization.BatchNormalization object at 0x7f45db53bed0>\n",
      "48 : activation_14, <keras.layers.core.Activation object at 0x7f45db4fefd0>\n",
      "49 : activation_17, <keras.layers.core.Activation object at 0x7f45db471f10>\n",
      "50 : average_pooling2d_2, <keras.layers.pooling.AveragePooling2D object at 0x7f45db45b050>\n",
      "51 : conv2d_13, <keras.layers.convolutional.Conv2D object at 0x7f45db436f50>\n",
      "52 : conv2d_15, <keras.layers.convolutional.Conv2D object at 0x7f45db3d4710>\n",
      "53 : conv2d_18, <keras.layers.convolutional.Conv2D object at 0x7f45db3e6c50>\n",
      "54 : conv2d_19, <keras.layers.convolutional.Conv2D object at 0x7f45db394ed0>\n",
      "55 : batch_normalization_13, <keras.layers.normalization.BatchNormalization object at 0x7f45db3fcf90>\n",
      "56 : batch_normalization_15, <keras.layers.normalization.BatchNormalization object at 0x7f45db3a9c10>\n",
      "57 : batch_normalization_18, <keras.layers.normalization.BatchNormalization object at 0x7f45db358910>\n",
      "58 : batch_normalization_19, <keras.layers.normalization.BatchNormalization object at 0x7f45db3303d0>\n",
      "59 : activation_13, <keras.layers.core.Activation object at 0x7f45db2a8310>\n",
      "60 : activation_15, <keras.layers.core.Activation object at 0x7f45db218250>\n",
      "61 : activation_18, <keras.layers.core.Activation object at 0x7f45db1c7110>\n",
      "62 : activation_19, <keras.layers.core.Activation object at 0x7f45db1c7690>\n",
      "63 : mixed1, <keras.layers.merge.Concatenate object at 0x7f45db18a850>\n",
      "64 : conv2d_23, <keras.layers.convolutional.Conv2D object at 0x7f45db18a8d0>\n",
      "65 : batch_normalization_23, <keras.layers.normalization.BatchNormalization object at 0x7f45db1a0e50>\n",
      "66 : activation_23, <keras.layers.core.Activation object at 0x7f45db1b6990>\n",
      "67 : conv2d_21, <keras.layers.convolutional.Conv2D object at 0x7f45db14c710>\n",
      "68 : conv2d_24, <keras.layers.convolutional.Conv2D object at 0x7f45db160150>\n",
      "69 : batch_normalization_21, <keras.layers.normalization.BatchNormalization object at 0x7f45db126210>\n",
      "70 : batch_normalization_24, <keras.layers.normalization.BatchNormalization object at 0x7f45db13dfd0>\n",
      "71 : activation_21, <keras.layers.core.Activation object at 0x7f45db0e8cd0>\n",
      "72 : activation_24, <keras.layers.core.Activation object at 0x7f45db043a10>\n",
      "73 : average_pooling2d_3, <keras.layers.pooling.AveragePooling2D object at 0x7f45db0728d0>\n",
      "74 : conv2d_20, <keras.layers.convolutional.Conv2D object at 0x7f45dafd1ed0>\n",
      "75 : conv2d_22, <keras.layers.convolutional.Conv2D object at 0x7f45db037910>\n",
      "76 : conv2d_25, <keras.layers.convolutional.Conv2D object at 0x7f45dafe8b50>\n",
      "77 : conv2d_26, <keras.layers.convolutional.Conv2D object at 0x7f45dafe8c90>\n",
      "78 : batch_normalization_20, <keras.layers.normalization.BatchNormalization object at 0x7f45daffefd0>\n",
      "79 : batch_normalization_22, <keras.layers.normalization.BatchNormalization object at 0x7f45daf40f10>\n",
      "80 : batch_normalization_25, <keras.layers.normalization.BatchNormalization object at 0x7f45daf55f90>\n",
      "81 : batch_normalization_26, <keras.layers.normalization.BatchNormalization object at 0x7f45daf19d50>\n",
      "82 : activation_20, <keras.layers.core.Activation object at 0x7f45dae8cc90>\n",
      "83 : activation_22, <keras.layers.core.Activation object at 0x7f45dae00bd0>\n",
      "84 : activation_25, <keras.layers.core.Activation object at 0x7f45dae2ca90>\n",
      "85 : activation_26, <keras.layers.core.Activation object at 0x7f45dae2ca50>\n",
      "86 : mixed2, <keras.layers.merge.Concatenate object at 0x7f45dadf3b10>\n",
      "87 : conv2d_28, <keras.layers.convolutional.Conv2D object at 0x7f45dad8b290>\n",
      "88 : batch_normalization_28, <keras.layers.normalization.BatchNormalization object at 0x7f45dadb3050>\n",
      "89 : activation_28, <keras.layers.core.Activation object at 0x7f45dadb3e50>\n",
      "90 : conv2d_29, <keras.layers.convolutional.Conv2D object at 0x7f45dada1e90>\n",
      "91 : batch_normalization_29, <keras.layers.normalization.BatchNormalization object at 0x7f45dad48ad0>\n",
      "92 : activation_29, <keras.layers.core.Activation object at 0x7f45dad0fb90>\n",
      "93 : conv2d_27, <keras.layers.convolutional.Conv2D object at 0x7f45dacd3fd0>\n",
      "94 : conv2d_30, <keras.layers.convolutional.Conv2D object at 0x7f45dacd3350>\n",
      "95 : batch_normalization_27, <keras.layers.normalization.BatchNormalization object at 0x7f45dacb30d0>\n",
      "96 : batch_normalization_30, <keras.layers.normalization.BatchNormalization object at 0x7f45dac47e90>\n",
      "97 : activation_27, <keras.layers.core.Activation object at 0x7f45dac6fe10>\n",
      "98 : activation_30, <keras.layers.core.Activation object at 0x7f45dabcc8d0>\n",
      "99 : max_pooling2d_3, <keras.layers.pooling.MaxPooling2D object at 0x7f45dabf9790>\n",
      "100 : mixed3, <keras.layers.merge.Concatenate object at 0x7f45dab59ed0>\n",
      "101 : conv2d_35, <keras.layers.convolutional.Conv2D object at 0x7f45dabbe7d0>\n",
      "102 : batch_normalization_35, <keras.layers.normalization.BatchNormalization object at 0x7f45dabbeed0>\n",
      "103 : activation_35, <keras.layers.core.Activation object at 0x7f45dab71910>\n",
      "104 : conv2d_36, <keras.layers.convolutional.Conv2D object at 0x7f45dab05b10>\n",
      "105 : batch_normalization_36, <keras.layers.normalization.BatchNormalization object at 0x7f45dab18550>\n",
      "106 : activation_36, <keras.layers.core.Activation object at 0x7f45daae1610>\n",
      "107 : conv2d_32, <keras.layers.convolutional.Conv2D object at 0x7f45daaf4c90>\n",
      "108 : conv2d_37, <keras.layers.convolutional.Conv2D object at 0x7f45daaa3a50>\n",
      "109 : batch_normalization_32, <keras.layers.normalization.BatchNormalization object at 0x7f45daa63b10>\n",
      "110 : batch_normalization_37, <keras.layers.normalization.BatchNormalization object at 0x7f45daa12f10>\n",
      "111 : activation_32, <keras.layers.core.Activation object at 0x7f45da9c1890>\n",
      "112 : activation_37, <keras.layers.core.Activation object at 0x7f45da99e350>\n",
      "113 : conv2d_33, <keras.layers.convolutional.Conv2D object at 0x7f45da948210>\n",
      "114 : conv2d_38, <keras.layers.convolutional.Conv2D object at 0x7f45da948790>\n",
      "115 : batch_normalization_33, <keras.layers.normalization.BatchNormalization object at 0x7f45da926d50>\n",
      "116 : batch_normalization_38, <keras.layers.normalization.BatchNormalization object at 0x7f45da8d3f10>\n",
      "117 : activation_33, <keras.layers.core.Activation object at 0x7f45da8e8f90>\n",
      "118 : activation_38, <keras.layers.core.Activation object at 0x7f45da8abd50>\n",
      "119 : average_pooling2d_4, <keras.layers.pooling.AveragePooling2D object at 0x7f45da85cc10>\n",
      "120 : conv2d_31, <keras.layers.convolutional.Conv2D object at 0x7f45da820c50>\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "121 : conv2d_34, <keras.layers.convolutional.Conv2D object at 0x7f45da83c410>\n",
      "122 : conv2d_39, <keras.layers.convolutional.Conv2D object at 0x7f45da7ce950>\n",
      "123 : conv2d_40, <keras.layers.convolutional.Conv2D object at 0x7f45da7cefd0>\n",
      "124 : batch_normalization_31, <keras.layers.normalization.BatchNormalization object at 0x7f45da790f10>\n",
      "125 : batch_normalization_34, <keras.layers.normalization.BatchNormalization object at 0x7f45da7f8fd0>\n",
      "126 : batch_normalization_39, <keras.layers.normalization.BatchNormalization object at 0x7f45da7bd610>\n",
      "127 : batch_normalization_40, <keras.layers.normalization.BatchNormalization object at 0x7f45da72ff50>\n",
      "128 : activation_31, <keras.layers.core.Activation object at 0x7f45da6f5fd0>\n",
      "129 : activation_34, <keras.layers.core.Activation object at 0x7f45da669f10>\n",
      "130 : activation_39, <keras.layers.core.Activation object at 0x7f45da618dd0>\n",
      "131 : activation_40, <keras.layers.core.Activation object at 0x7f45da618d90>\n",
      "132 : mixed4, <keras.layers.merge.Concatenate object at 0x7f45da5dbe50>\n",
      "133 : conv2d_45, <keras.layers.convolutional.Conv2D object at 0x7f45da5f65d0>\n",
      "134 : batch_normalization_45, <keras.layers.normalization.BatchNormalization object at 0x7f45da58cb50>\n",
      "135 : activation_45, <keras.layers.core.Activation object at 0x7f45da58cdd0>\n",
      "136 : conv2d_46, <keras.layers.convolutional.Conv2D object at 0x7f45da5b57d0>\n",
      "137 : batch_normalization_46, <keras.layers.normalization.BatchNormalization object at 0x7f45da5b5e10>\n",
      "138 : activation_46, <keras.layers.core.Activation object at 0x7f45da578ed0>\n",
      "139 : conv2d_42, <keras.layers.convolutional.Conv2D object at 0x7f45da53e910>\n",
      "140 : conv2d_47, <keras.layers.convolutional.Conv2D object at 0x7f45da4d3350>\n",
      "141 : batch_normalization_42, <keras.layers.normalization.BatchNormalization object at 0x7f45da497410>\n",
      "142 : batch_normalization_47, <keras.layers.normalization.BatchNormalization object at 0x7f45da45af10>\n",
      "143 : activation_42, <keras.layers.core.Activation object at 0x7f45da471e50>\n",
      "144 : activation_47, <keras.layers.core.Activation object at 0x7f45da433c10>\n",
      "145 : conv2d_43, <keras.layers.convolutional.Conv2D object at 0x7f45da3e2ad0>\n",
      "146 : conv2d_48, <keras.layers.convolutional.Conv2D object at 0x7f45da3e2a90>\n",
      "147 : batch_normalization_43, <keras.layers.normalization.BatchNormalization object at 0x7f45da355ad0>\n",
      "148 : batch_normalization_48, <keras.layers.normalization.BatchNormalization object at 0x7f45da355e90>\n",
      "149 : activation_43, <keras.layers.core.Activation object at 0x7f45da309b90>\n",
      "150 : activation_48, <keras.layers.core.Activation object at 0x7f45da2e0650>\n",
      "151 : average_pooling2d_5, <keras.layers.pooling.AveragePooling2D object at 0x7f45da28f510>\n",
      "152 : conv2d_41, <keras.layers.convolutional.Conv2D object at 0x7f45da255550>\n",
      "153 : conv2d_44, <keras.layers.convolutional.Conv2D object at 0x7f45da255cd0>\n",
      "154 : conv2d_49, <keras.layers.convolutional.Conv2D object at 0x7f45da2012d0>\n",
      "155 : conv2d_50, <keras.layers.convolutional.Conv2D object at 0x7f45da201ed0>\n",
      "156 : batch_normalization_41, <keras.layers.normalization.BatchNormalization object at 0x7f45da216e90>\n",
      "157 : batch_normalization_44, <keras.layers.normalization.BatchNormalization object at 0x7f45da22cf50>\n",
      "158 : batch_normalization_49, <keras.layers.normalization.BatchNormalization object at 0x7f45da1d4ed0>\n",
      "159 : batch_normalization_50, <keras.layers.normalization.BatchNormalization object at 0x7f45da1b3990>\n",
      "160 : activation_41, <keras.layers.core.Activation object at 0x7f45da1248d0>\n",
      "161 : activation_44, <keras.layers.core.Activation object at 0x7f45da098810>\n",
      "162 : activation_49, <keras.layers.core.Activation object at 0x7f45da0476d0>\n",
      "163 : activation_50, <keras.layers.core.Activation object at 0x7f45da047c50>\n",
      "164 : mixed5, <keras.layers.merge.Concatenate object at 0x7f45da00be10>\n",
      "165 : conv2d_55, <keras.layers.convolutional.Conv2D object at 0x7f45da00be90>\n",
      "166 : batch_normalization_55, <keras.layers.normalization.BatchNormalization object at 0x7f45d9fcfa50>\n",
      "167 : activation_55, <keras.layers.core.Activation object at 0x7f45d9fcff50>\n",
      "168 : conv2d_56, <keras.layers.convolutional.Conv2D object at 0x7f45d9fcfed0>\n",
      "169 : batch_normalization_56, <keras.layers.normalization.BatchNormalization object at 0x7f45d9fe2710>\n",
      "170 : activation_56, <keras.layers.core.Activation object at 0x7f45d9fa67d0>\n",
      "171 : conv2d_52, <keras.layers.convolutional.Conv2D object at 0x7f45d9f6bed0>\n",
      "172 : conv2d_57, <keras.layers.convolutional.Conv2D object at 0x7f45d9f6bc10>\n",
      "173 : batch_normalization_52, <keras.layers.normalization.BatchNormalization object at 0x7f45d9f32cd0>\n",
      "174 : batch_normalization_57, <keras.layers.normalization.BatchNormalization object at 0x7f45d9eddd50>\n",
      "175 : activation_52, <keras.layers.core.Activation object at 0x7f45e8649710>\n",
      "176 : activation_57, <keras.layers.core.Activation object at 0x7f45d9e4dd90>\n",
      "177 : conv2d_53, <keras.layers.convolutional.Conv2D object at 0x7f45d9e76cd0>\n",
      "178 : conv2d_58, <keras.layers.convolutional.Conv2D object at 0x7f45d9e76c90>\n",
      "179 : batch_normalization_53, <keras.layers.normalization.BatchNormalization object at 0x7f45d9dea850>\n",
      "180 : batch_normalization_58, <keras.layers.normalization.BatchNormalization object at 0x7f45d9deae10>\n",
      "181 : activation_53, <keras.layers.core.Activation object at 0x7f45d9d95d90>\n",
      "182 : activation_58, <keras.layers.core.Activation object at 0x7f45d9d70850>\n",
      "183 : average_pooling2d_6, <keras.layers.pooling.AveragePooling2D object at 0x7f45d9d22710>\n",
      "184 : conv2d_51, <keras.layers.convolutional.Conv2D object at 0x7f45d9cffe50>\n",
      "185 : conv2d_54, <keras.layers.convolutional.Conv2D object at 0x7f45d9ce7750>\n",
      "186 : conv2d_59, <keras.layers.convolutional.Conv2D object at 0x7f45d9c96990>\n",
      "187 : conv2d_60, <keras.layers.convolutional.Conv2D object at 0x7f45d9c96d50>\n",
      "188 : batch_normalization_51, <keras.layers.normalization.BatchNormalization object at 0x7f45d9cace10>\n",
      "189 : batch_normalization_54, <keras.layers.normalization.BatchNormalization object at 0x7f45d9c40e50>\n",
      "190 : batch_normalization_59, <keras.layers.normalization.BatchNormalization object at 0x7f45d9c52ed0>\n",
      "191 : batch_normalization_60, <keras.layers.normalization.BatchNormalization object at 0x7f45d9bc5b90>\n",
      "192 : activation_51, <keras.layers.core.Activation object at 0x7f45d9bb7ad0>\n",
      "193 : activation_54, <keras.layers.core.Activation object at 0x7f45d9b2ba10>\n",
      "194 : activation_59, <keras.layers.core.Activation object at 0x7f45d9ad98d0>\n",
      "195 : activation_60, <keras.layers.core.Activation object at 0x7f45d9ad9e50>\n",
      "196 : mixed6, <keras.layers.merge.Concatenate object at 0x7f45d9a9ec90>\n",
      "197 : conv2d_65, <keras.layers.convolutional.Conv2D object at 0x7f45d9a9ea10>\n",
      "198 : batch_normalization_65, <keras.layers.normalization.BatchNormalization object at 0x7f45d9a4f650>\n",
      "199 : activation_65, <keras.layers.core.Activation object at 0x7f45d9a4fcd0>\n",
      "200 : conv2d_66, <keras.layers.convolutional.Conv2D object at 0x7f45d9a62c50>\n",
      "201 : batch_normalization_66, <keras.layers.normalization.BatchNormalization object at 0x7f45d9a77910>\n",
      "202 : activation_66, <keras.layers.core.Activation object at 0x7f45d9a3a9d0>\n",
      "203 : conv2d_62, <keras.layers.convolutional.Conv2D object at 0x7f45d99807d0>\n",
      "204 : conv2d_67, <keras.layers.convolutional.Conv2D object at 0x7f45d9980e10>\n",
      "205 : batch_normalization_62, <keras.layers.normalization.BatchNormalization object at 0x7f45d9944ed0>\n",
      "206 : batch_normalization_67, <keras.layers.normalization.BatchNormalization object at 0x7f45d9970cd0>\n",
      "207 : activation_62, <keras.layers.core.Activation object at 0x7f45d991ef90>\n",
      "208 : activation_67, <keras.layers.core.Activation object at 0x7f45d98f7710>\n",
      "209 : conv2d_63, <keras.layers.convolutional.Conv2D object at 0x7f45d98a75d0>\n",
      "210 : conv2d_68, <keras.layers.convolutional.Conv2D object at 0x7f45d98a7b50>\n",
      "211 : batch_normalization_63, <keras.layers.normalization.BatchNormalization object at 0x7f45d9805d10>\n",
      "212 : batch_normalization_68, <keras.layers.normalization.BatchNormalization object at 0x7f45d981dd10>\n",
      "213 : activation_63, <keras.layers.core.Activation object at 0x7f45d97c8690>\n",
      "214 : activation_68, <keras.layers.core.Activation object at 0x7f45d97b8fd0>\n",
      "215 : average_pooling2d_7, <keras.layers.pooling.AveragePooling2D object at 0x7f45d97a2110>\n",
      "216 : conv2d_61, <keras.layers.convolutional.Conv2D object at 0x7f45d977cf50>\n",
      "217 : conv2d_64, <keras.layers.convolutional.Conv2D object at 0x7f45d97197d0>\n",
      "218 : conv2d_69, <keras.layers.convolutional.Conv2D object at 0x7f45d96c2d50>\n",
      "219 : conv2d_70, <keras.layers.convolutional.Conv2D object at 0x7f45d972bd90>\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "220 : batch_normalization_61, <keras.layers.normalization.BatchNormalization object at 0x7f45d96dcd10>\n",
      "221 : batch_normalization_64, <keras.layers.normalization.BatchNormalization object at 0x7f45d96f1cd0>\n",
      "222 : batch_normalization_69, <keras.layers.normalization.BatchNormalization object at 0x7f45d969c9d0>\n",
      "223 : batch_normalization_70, <keras.layers.normalization.BatchNormalization object at 0x7f45d9678490>\n",
      "224 : activation_61, <keras.layers.core.Activation object at 0x7f45d95eb3d0>\n",
      "225 : activation_64, <keras.layers.core.Activation object at 0x7f45d955f310>\n",
      "226 : activation_69, <keras.layers.core.Activation object at 0x7f45d950d1d0>\n",
      "227 : activation_70, <keras.layers.core.Activation object at 0x7f45d950d750>\n",
      "228 : mixed7, <keras.layers.merge.Concatenate object at 0x7f45d94d5910>\n",
      "229 : conv2d_73, <keras.layers.convolutional.Conv2D object at 0x7f45d94d5990>\n",
      "230 : batch_normalization_73, <keras.layers.normalization.BatchNormalization object at 0x7f45d94ebf10>\n",
      "231 : activation_73, <keras.layers.core.Activation object at 0x7f45d9480a50>\n",
      "232 : conv2d_74, <keras.layers.convolutional.Conv2D object at 0x7f45d94957d0>\n",
      "233 : batch_normalization_74, <keras.layers.normalization.BatchNormalization object at 0x7f45d94a8210>\n",
      "234 : activation_74, <keras.layers.core.Activation object at 0x7f45d946f2d0>\n",
      "235 : conv2d_71, <keras.layers.convolutional.Conv2D object at 0x7f45d941bfd0>\n",
      "236 : conv2d_75, <keras.layers.convolutional.Conv2D object at 0x7f45d942d710>\n",
      "237 : batch_normalization_71, <keras.layers.normalization.BatchNormalization object at 0x7f45d93f67d0>\n",
      "238 : batch_normalization_75, <keras.layers.normalization.BatchNormalization object at 0x7f45d9390e50>\n",
      "239 : activation_71, <keras.layers.core.Activation object at 0x7f45d9352f50>\n",
      "240 : activation_75, <keras.layers.core.Activation object at 0x7f45d9314fd0>\n",
      "241 : conv2d_72, <keras.layers.convolutional.Conv2D object at 0x7f45d92c4e90>\n",
      "242 : conv2d_76, <keras.layers.convolutional.Conv2D object at 0x7f45d92c4e50>\n",
      "243 : batch_normalization_72, <keras.layers.normalization.BatchNormalization object at 0x7f45d92b8a10>\n",
      "244 : batch_normalization_76, <keras.layers.normalization.BatchNormalization object at 0x7f45d92b8fd0>\n",
      "245 : activation_72, <keras.layers.core.Activation object at 0x7f45d9260cd0>\n",
      "246 : activation_76, <keras.layers.core.Activation object at 0x7f45d923ca10>\n",
      "247 : max_pooling2d_4, <keras.layers.pooling.MaxPooling2D object at 0x7f45d91ea8d0>\n",
      "248 : mixed8, <keras.layers.merge.Concatenate object at 0x7f45d914aed0>\n",
      "249 : conv2d_81, <keras.layers.convolutional.Conv2D object at 0x7f45d91b1910>\n",
      "250 : batch_normalization_81, <keras.layers.normalization.BatchNormalization object at 0x7f45d91b1c90>\n",
      "251 : activation_81, <keras.layers.core.Activation object at 0x7f45d91769d0>\n",
      "252 : conv2d_78, <keras.layers.convolutional.Conv2D object at 0x7f45d9176d90>\n",
      "253 : conv2d_82, <keras.layers.convolutional.Conv2D object at 0x7f45d9108690>\n",
      "254 : batch_normalization_78, <keras.layers.normalization.BatchNormalization object at 0x7f45d90d3750>\n",
      "255 : batch_normalization_82, <keras.layers.normalization.BatchNormalization object at 0x7f45d90e8dd0>\n",
      "256 : activation_78, <keras.layers.core.Activation object at 0x7f45d90ac890>\n",
      "257 : activation_82, <keras.layers.core.Activation object at 0x7f45d906ef50>\n",
      "258 : conv2d_79, <keras.layers.convolutional.Conv2D object at 0x7f45d901de10>\n",
      "259 : conv2d_80, <keras.layers.convolutional.Conv2D object at 0x7f45d901ddd0>\n",
      "260 : conv2d_83, <keras.layers.convolutional.Conv2D object at 0x7f45d8f90990>\n",
      "261 : conv2d_84, <keras.layers.convolutional.Conv2D object at 0x7f45d8f90f50>\n",
      "262 : average_pooling2d_8, <keras.layers.pooling.AveragePooling2D object at 0x7f45d8f52fd0>\n",
      "263 : conv2d_77, <keras.layers.convolutional.Conv2D object at 0x7f45d8f67510>\n",
      "264 : batch_normalization_79, <keras.layers.normalization.BatchNormalization object at 0x7f45d8f67f50>\n",
      "265 : batch_normalization_80, <keras.layers.normalization.BatchNormalization object at 0x7f45d8f7f8d0>\n",
      "266 : batch_normalization_83, <keras.layers.normalization.BatchNormalization object at 0x7f45d8f15b50>\n",
      "267 : batch_normalization_84, <keras.layers.normalization.BatchNormalization object at 0x7f45d8eef610>\n",
      "268 : conv2d_85, <keras.layers.convolutional.Conv2D object at 0x7f45d8e62550>\n",
      "269 : batch_normalization_77, <keras.layers.normalization.BatchNormalization object at 0x7f45d8dd7490>\n",
      "270 : activation_79, <keras.layers.core.Activation object at 0x7f45d8d64e10>\n",
      "271 : activation_80, <keras.layers.core.Activation object at 0x7f45d8d7cb90>\n",
      "272 : activation_83, <keras.layers.core.Activation object at 0x7f45d8d0d5d0>\n",
      "273 : activation_84, <keras.layers.core.Activation object at 0x7f45d8d0db50>\n",
      "274 : batch_normalization_85, <keras.layers.normalization.BatchNormalization object at 0x7f45d8cd1d10>\n",
      "275 : activation_77, <keras.layers.core.Activation object at 0x7f45d8cd1d90>\n",
      "276 : mixed9_0, <keras.layers.merge.Concatenate object at 0x7f45d8c85f10>\n",
      "277 : concatenate_1, <keras.layers.merge.Concatenate object at 0x7f45d8c855d0>\n",
      "278 : activation_85, <keras.layers.core.Activation object at 0x7f45d8c984d0>\n",
      "279 : mixed9, <keras.layers.merge.Concatenate object at 0x7f45d8c5e710>\n",
      "280 : conv2d_90, <keras.layers.convolutional.Conv2D object at 0x7f45d8c44fd0>\n",
      "281 : batch_normalization_90, <keras.layers.normalization.BatchNormalization object at 0x7f45d8c72d90>\n",
      "282 : activation_90, <keras.layers.core.Activation object at 0x7f45d8c0b910>\n",
      "283 : conv2d_87, <keras.layers.convolutional.Conv2D object at 0x7f45d8c20690>\n",
      "284 : conv2d_91, <keras.layers.convolutional.Conv2D object at 0x7f45d8c340d0>\n",
      "285 : batch_normalization_87, <keras.layers.normalization.BatchNormalization object at 0x7f45d8bfd190>\n",
      "286 : batch_normalization_91, <keras.layers.normalization.BatchNormalization object at 0x7f45d8b95f50>\n",
      "287 : activation_87, <keras.layers.core.Activation object at 0x7f45d8bbced0>\n",
      "288 : activation_91, <keras.layers.core.Activation object at 0x7f45d8b16990>\n",
      "289 : conv2d_88, <keras.layers.convolutional.Conv2D object at 0x7f45d8ac7850>\n",
      "290 : conv2d_89, <keras.layers.convolutional.Conv2D object at 0x7f45d8ac7dd0>\n",
      "291 : conv2d_92, <keras.layers.convolutional.Conv2D object at 0x7f45d8aa6f90>\n",
      "292 : conv2d_93, <keras.layers.convolutional.Conv2D object at 0x7f45d8abac10>\n",
      "293 : average_pooling2d_9, <keras.layers.pooling.AveragePooling2D object at 0x7f45d8a4fcd0>\n",
      "294 : conv2d_86, <keras.layers.convolutional.Conv2D object at 0x7f45d8a12d50>\n",
      "295 : batch_normalization_88, <keras.layers.normalization.BatchNormalization object at 0x7f45d8a00dd0>\n",
      "296 : batch_normalization_89, <keras.layers.normalization.BatchNormalization object at 0x7f45d8a2c110>\n",
      "297 : batch_normalization_92, <keras.layers.normalization.BatchNormalization object at 0x7f45d89c2f90>\n",
      "298 : batch_normalization_93, <keras.layers.normalization.BatchNormalization object at 0x7f45d8984f10>\n",
      "299 : conv2d_94, <keras.layers.convolutional.Conv2D object at 0x7f45d8978f50>\n",
      "300 : batch_normalization_86, <keras.layers.normalization.BatchNormalization object at 0x7f45d88e9e90>\n",
      "301 : activation_88, <keras.layers.core.Activation object at 0x7f45d880dd50>\n",
      "302 : activation_89, <keras.layers.core.Activation object at 0x7f45d8824fd0>\n",
      "303 : activation_92, <keras.layers.core.Activation object at 0x7f45d8824350>\n",
      "304 : activation_93, <keras.layers.core.Activation object at 0x7f45d8839590>\n",
      "305 : batch_normalization_94, <keras.layers.normalization.BatchNormalization object at 0x7f45d8785750>\n",
      "306 : activation_86, <keras.layers.core.Activation object at 0x7f45d87857d0>\n",
      "307 : mixed9_1, <keras.layers.merge.Concatenate object at 0x7f45d8799f50>\n",
      "308 : concatenate_2, <keras.layers.merge.Concatenate object at 0x7f45d87ab950>\n",
      "309 : activation_94, <keras.layers.core.Activation object at 0x7f45d87abed0>\n",
      "310 : mixed10, <keras.layers.merge.Concatenate object at 0x7f45d876ea90>\n",
      "311 : pool_2d, <keras.layers.pooling.GlobalAveragePooling2D object at 0x7f45d876ea10>\n",
      "312 : dense, <keras.layers.core.Dense object at 0x7f45d871d390>\n",
      "313 : dense__activation__, <keras.layers.core.Activation object at 0x7f45ce9ace50>\n",
      "314 : dense_1, <keras.layers.core.Dense object at 0x7f45d871da90>\n",
      "315 : dense_1__activation__, <keras.layers.core.Activation object at 0x7f45ce9acf50>\n"
     ]
    },
    {
     "ename": "TypeError",
     "evalue": "Class labels must be of type Integer or String. (not <type 'unicode'>)",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-15-ee435c6b002a>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      4\u001b[0m                                                     \u001b[0mclass_labels\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mclass_labels\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m                                                     \u001b[0mis_bgr\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mTrue\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 6\u001b[0;31m                                                     image_scale=1/255.)\n\u001b[0m",
      "\u001b[0;32m/home/douwe/proj/notebooks/venv2/local/lib/python2.7/site-packages/coremltools/converters/keras/_keras_converter.pyc\u001b[0m in \u001b[0;36mconvert\u001b[0;34m(model, input_names, output_names, image_input_names, is_bgr, red_bias, green_bias, blue_bias, gray_bias, image_scale, class_labels, predicted_feature_name, model_precision, predicted_probabilities_output, add_custom_layers, custom_conversion_functions)\u001b[0m\n\u001b[1;32m    720\u001b[0m                       \u001b[0mpredicted_probabilities_output\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    721\u001b[0m                       \u001b[0madd_custom_layers\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 722\u001b[0;31m                       custom_conversion_functions=custom_conversion_functions)\n\u001b[0m\u001b[1;32m    723\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    724\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0m_MLModel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mspec\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/douwe/proj/notebooks/venv2/local/lib/python2.7/site-packages/coremltools/converters/keras/_keras_converter.pyc\u001b[0m in \u001b[0;36mconvertToSpec\u001b[0;34m(model, input_names, output_names, image_input_names, is_bgr, red_bias, green_bias, blue_bias, gray_bias, image_scale, class_labels, predicted_feature_name, model_precision, predicted_probabilities_output, add_custom_layers, custom_conversion_functions)\u001b[0m\n\u001b[1;32m    525\u001b[0m                                            \u001b[0mpredicted_probabilities_output\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mpredicted_probabilities_output\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    526\u001b[0m                                            \u001b[0madd_custom_layers\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0madd_custom_layers\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 527\u001b[0;31m                                            custom_conversion_functions=custom_conversion_functions)\n\u001b[0m\u001b[1;32m    528\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    529\u001b[0m         raise RuntimeError(\n",
      "\u001b[0;32m/home/douwe/proj/notebooks/venv2/local/lib/python2.7/site-packages/coremltools/converters/keras/_keras2_converter.pyc\u001b[0m in \u001b[0;36m_convert\u001b[0;34m(model, input_names, output_names, image_input_names, is_bgr, red_bias, green_bias, blue_bias, gray_bias, image_scale, class_labels, predicted_feature_name, predicted_probabilities_output, add_custom_layers, custom_conversion_functions)\u001b[0m\n\u001b[1;32m    314\u001b[0m                                      prediction_blob = predicted_probabilities_output)\n\u001b[1;32m    315\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 316\u001b[0;31m             \u001b[0mbuilder\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_class_labels\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclasses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    317\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    318\u001b[0m     \u001b[0;31m# Set pre-processing paramsters\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/douwe/proj/notebooks/venv2/local/lib/python2.7/site-packages/coremltools/models/neural_network.pyc\u001b[0m in \u001b[0;36mset_class_labels\u001b[0;34m(self, class_labels, predicted_feature_name, prediction_blob)\u001b[0m\n\u001b[1;32m    269\u001b[0m         \u001b[0mclass_type\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclass_labels\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    270\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mclass_type\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32min\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mint\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 271\u001b[0;31m             \u001b[0;32mraise\u001b[0m \u001b[0mTypeError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Class labels must be of type Integer or String. (not %s)\"\u001b[0m \u001b[0;34m%\u001b[0m \u001b[0mclass_type\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    272\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    273\u001b[0m         \u001b[0mspec\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdescription\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredictedProbabilitiesName\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mprobOutput\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mTypeError\u001b[0m: Class labels must be of type Integer or String. (not <type 'unicode'>)"
     ]
    }
   ],
   "source": [
    "coreml_model = coremltools.converters.keras.convert(keras_model,\n",
    "                                                    image_input_names=\"input_1\", \n",
    "                                                    input_names='input_1',\n",
    "                                                    class_labels=class_labels,\n",
    "                                                    is_bgr=True,\n",
    "                                                    image_scale=1/255.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "coreml_model.save('zoo/PetRecognizer3.mlmodel')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "IOPub data rate exceeded.\n",
      "The notebook server will temporarily stop sending output\n",
      "to the client in order to avoid crashing it.\n",
      "To change this limit, set the config variable\n",
      "`--NotebookApp.iopub_data_rate_limit`.\n"
     ]
    }
   ],
   "source": [
    "img_url = 'https://upload.wikimedia.org/wikipedia/commons/9/93/Golden_Retriever_Carlos_%2810581910556%29.jpg'\n",
    "response = requests.get(img_url)\n",
    "img = PIL.Image.open(BytesIO(response.content))\n",
    "img"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['__class__',\n",
       " '__delattr__',\n",
       " '__dict__',\n",
       " '__doc__',\n",
       " '__format__',\n",
       " '__getattribute__',\n",
       " '__hash__',\n",
       " '__init__',\n",
       " '__module__',\n",
       " '__new__',\n",
       " '__proxy__',\n",
       " '__reduce__',\n",
       " '__reduce_ex__',\n",
       " '__repr__',\n",
       " '__setattr__',\n",
       " '__sizeof__',\n",
       " '__str__',\n",
       " '__subclasshook__',\n",
       " '__weakref__',\n",
       " '_input_description',\n",
       " '_output_description',\n",
       " '_spec',\n",
       " 'author',\n",
       " 'get_spec',\n",
       " 'input_description',\n",
       " 'license',\n",
       " 'output_description',\n",
       " 'predict',\n",
       " 'save',\n",
       " 'short_description',\n",
       " 'user_defined_metadata']"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dir(coreml_model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['__class__',\n",
       " '__contains__',\n",
       " '__delattr__',\n",
       " '__dict__',\n",
       " '__doc__',\n",
       " '__format__',\n",
       " '__getattribute__',\n",
       " '__getitem__',\n",
       " '__hash__',\n",
       " '__init__',\n",
       " '__iter__',\n",
       " '__len__',\n",
       " '__module__',\n",
       " '__new__',\n",
       " '__reduce__',\n",
       " '__reduce_ex__',\n",
       " '__repr__',\n",
       " '__setattr__',\n",
       " '__setitem__',\n",
       " '__sizeof__',\n",
       " '__str__',\n",
       " '__subclasshook__',\n",
       " '__weakref__',\n",
       " '_fd_spec']"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dir(coreml_model.input_description)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[u'Egyptian_Mau',\n",
       " u'shiba_inu',\n",
       " u'Bombay',\n",
       " u'Siamese',\n",
       " u'yorkshire_terrier',\n",
       " u'staffordshire_bull_terrier',\n",
       " u'pomeranian',\n",
       " u'havanese',\n",
       " u'chihuahua',\n",
       " u'Bengal',\n",
       " u'basset_hound',\n",
       " u'keeshond',\n",
       " u'Sphynx',\n",
       " u'pug',\n",
       " u'Birman',\n",
       " u'english_cocker_spaniel',\n",
       " u'scottish_terrier',\n",
       " u'beagle',\n",
       " u'Russian_Blue',\n",
       " u'Maine_Coon',\n",
       " u'german_shorthaired',\n",
       " u'miniature_pinscher',\n",
       " u'Abyssinian',\n",
       " u'British_Shorthair',\n",
       " u'saint_bernard',\n",
       " u'leonberger',\n",
       " u'english_setter',\n",
       " u'Persian',\n",
       " u'Ragdoll',\n",
       " u'great_pyrenees',\n",
       " u'american_pit_bull_terrier',\n",
       " u'japanese_chin',\n",
       " u'newfoundland',\n",
       " u'samoyed',\n",
       " u'american_bulldog',\n",
       " u'wheaten_terrier',\n",
       " u'boxer']"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python2.7",
   "language": "python",
   "name": "python2.7"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}