6. predict_binary_img_with_CNN.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "먼저 이미지를 배열화 시켜주어야 한다.\n",
    "\n",
    "그리고 여기 이미지에서 훈련 데이터, 테스트 데이터(사실상 검증 데이터)로 나눠준다.\n",
    "\n",
    "이후 numpy 형태로 저장."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\h5py\\__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n",
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cat  :  ./binary_img_data/imgs/Cat\\0.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\10267.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\10537.jpg\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 80000 bytes but only got 0. Skipping tag 64640\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 65536 bytes but only got 0. Skipping tag 3\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 404094976 bytes but only got 0. Skipping tag 5\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 404619264 bytes but only got 0. Skipping tag 5\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 131072 bytes but only got 0. Skipping tag 3\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 425459712 bytes but only got 0. Skipping tag 4\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 1385474 bytes but only got 6833. Skipping tag 513\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 3846701056 bytes but only got 0. Skipping tag 2\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 3300917248 bytes but only got 0. Skipping tag 7\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 196867 bytes but only got 6833. Skipping tag 0\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:756: UserWarning: Corrupt EXIF data.  Expecting to read 12 bytes but only got 8. \n",
      "  warnings.warn(str(msg))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cat  :  ./binary_img_data/imgs/Cat\\10807.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\11077.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\11347.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\11617.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\11888.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\12157.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\12427.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\1465.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\1762.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\2058.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\2355.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\2652.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\2949.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\3245.jpg\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 18350080 bytes but only got 0. Skipping tag 0\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:756: UserWarning: Corrupt EXIF data.  Expecting to read 12 bytes but only got 6. \n",
      "  warnings.warn(str(msg))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cat  :  ./binary_img_data/imgs/Cat\\3530.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\3800.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\4070.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\4340.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\4610.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\4881.jpg\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 307363840 bytes but only got 0. Skipping tag 5\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 307888128 bytes but only got 0. Skipping tag 5\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 328728576 bytes but only got 0. Skipping tag 4\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 1385474 bytes but only got 5357. Skipping tag 513\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 196867 bytes but only got 5357. Skipping tag 0\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cat  :  ./binary_img_data/imgs/Cat\\5150.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\5420.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\5691.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\5961.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\6230.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\6500.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\6772.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\7041.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\7311.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\7582.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\7852.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\8121.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\8392.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\8662.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\8932.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\9201.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\9472.jpg\n",
      "Cat  :  ./binary_img_data/imgs/Cat\\9742.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\0.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\10270.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\10543.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\10816.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\11089.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\11361.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\11634.jpg\n",
      "Dog 1875 번째에서 에러 \n",
      "Dog  :  ./binary_img_data/imgs/Dog\\11907.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\1218.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\12452.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\1498.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\1798.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\2098.jpg\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 262146 bytes but only got 0. Skipping tag 2\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 262151 bytes but only got 0. Skipping tag 56\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:756: UserWarning: Corrupt EXIF data.  Expecting to read 12 bytes but only got 0. \n",
      "  warnings.warn(str(msg))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dog  :  ./binary_img_data/imgs/Dog\\2398.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\2698.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\2984.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\3256.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\3529.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\3801.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\4073.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\4346.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\4619.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\4891.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\5163.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\5436.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\5709.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\5981.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\6253.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\6526.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\6799.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\7070.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\7343.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\7616.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\7889.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\8160.jpg\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 6553600 bytes but only got 0. Skipping tag 49\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 1050744 bytes but only got 4951. Skipping tag 51\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 293339136 bytes but only got 0. Skipping tag 5\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 293863424 bytes but only got 0. Skipping tag 5\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 3368026112 bytes but only got 0. Skipping tag 7\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 134479872 bytes but only got 0. Skipping tag 7\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 295698432 bytes but only got 0. Skipping tag 10\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 296222720 bytes but only got 0. Skipping tag 5\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 14745600 bytes but only got 0. Skipping tag 4\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 25624576 bytes but only got 0. Skipping tag 4\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 317718528 bytes but only got 4956. Skipping tag 4\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 131073 bytes but only got 4952. Skipping tag 0\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 393216 bytes but only got 0. Skipping tag 3\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 287178752 bytes but only got 0. Skipping tag 5\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 287703040 bytes but only got 0. Skipping tag 5\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 524288 bytes but only got 0. Skipping tag 4\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 286654464 bytes but only got 4956. Skipping tag 4\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:756: UserWarning: Corrupt EXIF data.  Expecting to read 12 bytes but only got 2. \n",
      "  warnings.warn(str(msg))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 209715200 bytes but only got 0. Skipping tag 48\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 52428800 bytes but only got 0. Skipping tag 0\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 6468665344 bytes but only got 0. Skipping tag 0\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 1050744 bytes but only got 7027. Skipping tag 48\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 422313984 bytes but only got 0. Skipping tag 5\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 422838272 bytes but only got 0. Skipping tag 5\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 428867584 bytes but only got 0. Skipping tag 10\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 429391872 bytes but only got 0. Skipping tag 5\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 2031616 bytes but only got 0. Skipping tag 3\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 429916160 bytes but only got 0. Skipping tag 5\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 21299200 bytes but only got 0. Skipping tag 4\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 24313856 bytes but only got 0. Skipping tag 4\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 453771264 bytes but only got 7032. Skipping tag 4\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 131073 bytes but only got 7028. Skipping tag 0\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 416415744 bytes but only got 0. Skipping tag 5\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 416940032 bytes but only got 0. Skipping tag 5\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 415825920 bytes but only got 7032. Skipping tag 4\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dog  :  ./binary_img_data/imgs/Dog\\8433.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\8706.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\8979.jpg\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 32 bytes but only got 0. Skipping tag 270\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 5 bytes but only got 0. Skipping tag 271\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 8 bytes but only got 0. Skipping tag 272\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 8 bytes but only got 0. Skipping tag 282\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 8 bytes but only got 0. Skipping tag 283\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 20 bytes but only got 0. Skipping tag 306\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:739: UserWarning: Possibly corrupt EXIF data.  Expecting to read 48 bytes but only got 0. Skipping tag 532\n",
      "  \" Skipping tag %s\" % (size, len(data), tag))\n",
      "C:\\Users\\leesoojin\\Anaconda3\\lib\\site-packages\\PIL\\TiffImagePlugin.py:756: UserWarning: Corrupt EXIF data.  Expecting to read 2 bytes but only got 0. \n",
      "  warnings.warn(str(msg))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dog  :  ./binary_img_data/imgs/Dog\\9251.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\9524.jpg\n",
      "Dog  :  ./binary_img_data/imgs/Dog\\9797.jpg\n"
     ]
    }
   ],
   "source": [
    "from PIL import Image\n",
    "import os, glob, sys, numpy as np\n",
    "from sklearn.model_selection import train_test_split\n",
    "from keras.utils import np_utils\n",
    "\n",
    "img_dir = './binary_img_data/imgs'\n",
    "categories = ['Cat', 'Dog']\n",
    "np_classes = len(categories)\n",
    "\n",
    "image_w = 64\n",
    "image_h = 64\n",
    "\n",
    "\n",
    "pixel = image_h * image_w * 3\n",
    "\n",
    "X = []\n",
    "y = []\n",
    "\n",
    "for idx, cat in enumerate(categories):\n",
    "    img_dir_detail = img_dir + \"/\" + cat\n",
    "    files = glob.glob(img_dir_detail+\"/*.jpg\")\n",
    "\n",
    "\n",
    "    for i, f in enumerate(files):\n",
    "        try:\n",
    "            img = Image.open(f)\n",
    "            img = img.convert(\"RGB\")\n",
    "            img = img.resize((image_w, image_h))\n",
    "            data = np.asarray(img)\n",
    "            #Y는 0 아니면 1이니까 idx값으로 넣는다.\n",
    "            X.append(data)\n",
    "            y.append(idx)\n",
    "            if i % 300 == 0:\n",
    "                print(cat, \" : \", f)\n",
    "        except:\n",
    "            print(cat, str(i)+\" 번째에서 에러 \")\n",
    "X = np.array(X)\n",
    "Y = np.array(y)\n",
    "\n",
    "\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.1)\n",
    "\n",
    "xy = (X_train, X_test, Y_train, Y_test)\n",
    "np.save(\"./numpy_data/binary_image_data.npy\", xy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "이제 훈련을 시작한다. 저장한 numpy 값을 가지고 와서 컨볼루션 망을 구축한다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(22056, 64, 64, 3)\n",
      "22056\n",
      "[11094 10962]\n",
      "[1191 1260]\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Conv2D, MaxPooling2D, Dense, Flatten, Dropout\n",
    "from keras.callbacks import EarlyStopping, ModelCheckpoint\n",
    "import matplotlib.pyplot as plt\n",
    "import keras.backend.tensorflow_backend as K\n",
    "\n",
    "X_train, X_test, y_train, y_test = np.load('./numpy_data/binary_image_data.npy')\n",
    "print(X_train.shape)\n",
    "print(X_train.shape[0])\n",
    "print(np.bincount(y_train))\n",
    "print(np.bincount(y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "image_w = 64\n",
    "image_h = 64\n",
    "X_train = X_train.astype('float32') / 255\n",
    "X_test = X_test.astype('float32') / 255\n",
    "\n",
    "\n",
    "with K.tf_ops.device('/device:GPU:0'):\n",
    "    model = Sequential()\n",
    "    model.add(Conv2D(32, (3,3), padding=\"same\", input_shape=X_train.shape[1:], activation=\"relu\"))\n",
    "    model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "    model.add(Conv2D(32, (3,3), padding=\"same\", activation=\"relu\"))\n",
    "    model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "\n",
    "    model.add(Conv2D(64, (3,3), padding=\"same\", activation=\"relu\"))\n",
    "    model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "    model.add(Dropout(0.25))\n",
    "\n",
    "    model.add(Conv2D(64, (3,3), padding=\"same\", activation=\"relu\"))\n",
    "    model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "    model.add(Dropout(0.25))\n",
    "\n",
    "    model.add(Flatten())\n",
    "    model.add(Dense(256, activation=\"relu\"))\n",
    "    model.add(Dropout(0.5))\n",
    "    model.add(Dense(1, activation=\"sigmoid\"))\n",
    "    \n",
    "    model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "    \n",
    "    model_dir = './model'\n",
    "    if not os.path.exists(model_dir):\n",
    "        os.mkdir(model_dir)\n",
    "    model_path = model_dir + \"/dog_cat_classify.model\"\n",
    "    \n",
    "    checkpoint = ModelCheckpoint(filepath=model_path, monitor='val_loss', verbose=1, save_best_only=True)\n",
    "    early_stopping = EarlyStopping(monitor='val_loss', patience=7)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_1 (Conv2D)            (None, 64, 64, 32)        896       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 32, 32, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_2 (Conv2D)            (None, 32, 32, 32)        9248      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2 (None, 16, 16, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_3 (Conv2D)            (None, 16, 16, 64)        18496     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_3 (MaxPooling2 (None, 8, 8, 64)          0         \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 8, 8, 64)          0         \n",
      "_________________________________________________________________\n",
      "conv2d_4 (Conv2D)            (None, 8, 8, 64)          36928     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_4 (MaxPooling2 (None, 4, 4, 64)          0         \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 4, 4, 64)          0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 1024)              0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 256)               262400    \n",
      "_________________________________________________________________\n",
      "dropout_3 (Dropout)          (None, 256)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 1)                 257       \n",
      "=================================================================\n",
      "Total params: 328,225\n",
      "Trainable params: 328,225\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 18747 samples, validate on 3309 samples\n",
      "Epoch 1/100\n",
      "18747/18747 [==============================] - 9s 480us/step - loss: 0.6535 - acc: 0.5976 - val_loss: 0.5828 - val_acc: 0.6878\n",
      "\n",
      "Epoch 00001: val_loss improved from inf to 0.58284, saving model to ./model/dog_cat_classify.model\n",
      "Epoch 2/100\n",
      "18747/18747 [==============================] - 5s 258us/step - loss: 0.5458 - acc: 0.7248 - val_loss: 0.4519 - val_acc: 0.7875\n",
      "\n",
      "Epoch 00002: val_loss improved from 0.58284 to 0.45193, saving model to ./model/dog_cat_classify.model\n",
      "Epoch 3/100\n",
      "18747/18747 [==============================] - 5s 257us/step - loss: 0.4688 - acc: 0.7803 - val_loss: 0.4365 - val_acc: 0.8005\n",
      "\n",
      "Epoch 00003: val_loss improved from 0.45193 to 0.43645, saving model to ./model/dog_cat_classify.model\n",
      "Epoch 4/100\n",
      "18747/18747 [==============================] - 5s 256us/step - loss: 0.4099 - acc: 0.8158 - val_loss: 0.3824 - val_acc: 0.8290\n",
      "\n",
      "Epoch 00004: val_loss improved from 0.43645 to 0.38235, saving model to ./model/dog_cat_classify.model\n",
      "Epoch 5/100\n",
      "18747/18747 [==============================] - 5s 253us/step - loss: 0.3696 - acc: 0.8354 - val_loss: 0.3514 - val_acc: 0.8471\n",
      "\n",
      "Epoch 00005: val_loss improved from 0.38235 to 0.35142, saving model to ./model/dog_cat_classify.model\n",
      "Epoch 6/100\n",
      "18747/18747 [==============================] - 5s 253us/step - loss: 0.3396 - acc: 0.8506 - val_loss: 0.3485 - val_acc: 0.8477\n",
      "\n",
      "Epoch 00006: val_loss improved from 0.35142 to 0.34848, saving model to ./model/dog_cat_classify.model\n",
      "Epoch 7/100\n",
      "18747/18747 [==============================] - 5s 253us/step - loss: 0.3146 - acc: 0.8647 - val_loss: 0.3215 - val_acc: 0.8534\n",
      "\n",
      "Epoch 00007: val_loss improved from 0.34848 to 0.32153, saving model to ./model/dog_cat_classify.model\n",
      "Epoch 8/100\n",
      "18747/18747 [==============================] - 5s 256us/step - loss: 0.2959 - acc: 0.8730 - val_loss: 0.3429 - val_acc: 0.8489\n",
      "\n",
      "Epoch 00008: val_loss did not improve from 0.32153\n",
      "Epoch 9/100\n",
      "18747/18747 [==============================] - 5s 253us/step - loss: 0.2719 - acc: 0.8860 - val_loss: 0.3596 - val_acc: 0.8344\n",
      "\n",
      "Epoch 00009: val_loss did not improve from 0.32153\n",
      "Epoch 10/100\n",
      "18747/18747 [==============================] - 5s 252us/step - loss: 0.2641 - acc: 0.8865 - val_loss: 0.3158 - val_acc: 0.8604\n",
      "\n",
      "Epoch 00010: val_loss improved from 0.32153 to 0.31585, saving model to ./model/dog_cat_classify.model\n",
      "Epoch 11/100\n",
      "18747/18747 [==============================] - 5s 252us/step - loss: 0.2454 - acc: 0.8964 - val_loss: 0.3352 - val_acc: 0.8501\n",
      "\n",
      "Epoch 00011: val_loss did not improve from 0.31585\n",
      "Epoch 12/100\n",
      "18747/18747 [==============================] - 5s 251us/step - loss: 0.2316 - acc: 0.9034 - val_loss: 0.2955 - val_acc: 0.8707\n",
      "\n",
      "Epoch 00012: val_loss improved from 0.31585 to 0.29551, saving model to ./model/dog_cat_classify.model\n",
      "Epoch 13/100\n",
      "18747/18747 [==============================] - 5s 251us/step - loss: 0.2199 - acc: 0.9087 - val_loss: 0.2893 - val_acc: 0.8710\n",
      "\n",
      "Epoch 00013: val_loss improved from 0.29551 to 0.28931, saving model to ./model/dog_cat_classify.model\n",
      "Epoch 14/100\n",
      "18747/18747 [==============================] - 5s 256us/step - loss: 0.2107 - acc: 0.9129 - val_loss: 0.3070 - val_acc: 0.8731\n",
      "\n",
      "Epoch 00014: val_loss did not improve from 0.28931\n",
      "Epoch 15/100\n",
      "18747/18747 [==============================] - 5s 251us/step - loss: 0.1966 - acc: 0.9178 - val_loss: 0.3023 - val_acc: 0.8731\n",
      "\n",
      "Epoch 00015: val_loss did not improve from 0.28931\n",
      "Epoch 16/100\n",
      "18747/18747 [==============================] - 5s 251us/step - loss: 0.1844 - acc: 0.9231 - val_loss: 0.2875 - val_acc: 0.8749\n",
      "\n",
      "Epoch 00016: val_loss improved from 0.28931 to 0.28753, saving model to ./model/dog_cat_classify.model\n",
      "Epoch 17/100\n",
      "18747/18747 [==============================] - 5s 256us/step - loss: 0.1709 - acc: 0.9301 - val_loss: 0.3019 - val_acc: 0.8728\n",
      "\n",
      "Epoch 00017: val_loss did not improve from 0.28753\n",
      "Epoch 18/100\n",
      "18747/18747 [==============================] - 5s 251us/step - loss: 0.1689 - acc: 0.9308 - val_loss: 0.3354 - val_acc: 0.8758\n",
      "\n",
      "Epoch 00018: val_loss did not improve from 0.28753\n",
      "Epoch 19/100\n",
      "18747/18747 [==============================] - 5s 251us/step - loss: 0.1600 - acc: 0.9357 - val_loss: 0.3402 - val_acc: 0.8728\n",
      "\n",
      "Epoch 00019: val_loss did not improve from 0.28753\n",
      "Epoch 20/100\n",
      "18747/18747 [==============================] - 5s 251us/step - loss: 0.1536 - acc: 0.9372 - val_loss: 0.3399 - val_acc: 0.8761 loss\n",
      "\n",
      "Epoch 00020: val_loss did not improve from 0.28753\n",
      "Epoch 21/100\n",
      "18747/18747 [==============================] - 5s 251us/step - loss: 0.1521 - acc: 0.9403 - val_loss: 0.3176 - val_acc: 0.8707\n",
      "\n",
      "Epoch 00021: val_loss did not improve from 0.28753\n",
      "Epoch 22/100\n",
      "18747/18747 [==============================] - 5s 252us/step - loss: 0.1417 - acc: 0.9406 - val_loss: 0.3122 - val_acc: 0.8755\n",
      "\n",
      "Epoch 00022: val_loss did not improve from 0.28753\n",
      "Epoch 23/100\n",
      "18747/18747 [==============================] - 5s 251us/step - loss: 0.1424 - acc: 0.9423 - val_loss: 0.3405 - val_acc: 0.8713\n",
      "\n",
      "Epoch 00023: val_loss did not improve from 0.28753\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(X_train, y_train, batch_size=64, epochs=100, validation_split=0.15, callbacks=[checkpoint, early_stopping])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2451/2451 [==============================] - 0s 153us/step\n",
      "정확도 : 0.88 \n"
     ]
    }
   ],
   "source": [
    "print(\"정확도 : %.2f \" %(model.evaluate(X_test, y_test)[1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3Xd4XMW9//H3bNGqrHqxVSzJvTdwB+xgEgOmmIADpveSBEggoaTA5XJJhZtcQvglAdO7MZCY2IYETGyaiww27gVXSVavK2m1bX5/zEqWhWRJtlb1+3qefXb3nNmzs8Kczzlz5sworTVCCCEEgKW7KyCEEKLnkFAQQgjRSEJBCCFEIwkFIYQQjSQUhBBCNJJQEEII0UhCQYh2Uko9r5R6pJ1lDyilvn2y2xGiq0koCCGEaCShIIQQopGEguhTgs029yilvlJK1SilnlFKDVBKrVRKVSulPlBKxTcpf6FSaptSqkIp9R+l1Ogm6yYrpb4Ifu4NILzZd52vlNoU/OxnSqkJJ1jnm5VSe5VSZUqpZUqptOBypZT6o1KqSClVGfxN44Lr5iultgfrlqeU+ukJ/cGEaEZCQfRFlwDfAUYAFwArgZ8DSZh/83cCKKVGAK8BPwaSgRXAu0qpMKVUGPB34CUgAXgzuF2Cnz0FeBa4FUgE/gYsU0o5OlJRpdRc4DfApUAqcBB4Pbh6HjA7+DvigMuA0uC6Z4BbtdbRwDhgVUe+V4jWSCiIvugJrXWh1joP+BhYp7X+UmtdD7wDTA6WuwxYrrX+t9baCzwGRACzgBmAHfg/rbVXa70U2NDkO24G/qa1Xqe19mutXwDqg5/riCuBZ7XWXwTr9zNgplIqG/AC0cAoQGmtd2itjwQ/5wXGKKVitNblWusvOvi9QrRIQkH0RYVNXte18N4ZfJ2GOTIHQGsdAA4D6cF1efrYESMPNnmdBfwk2HRUoZSqAAYFP9cRzevgwpwNpGutVwF/Bp4ECpVSTymlYoJFLwHmAweVUquVUjM7+L1CtEhCQfRn+ZidO2Da8DE79jzgCJAeXNYgs8nrw8CvtNZxTR6RWuvXTrIOUZjmqDwArfWftNanAmMxzUj3BJdv0FovAFIwzVxLOvi9QrRIQkH0Z0uA85RSZyml7MBPME1AnwGfAz7gTqWUTSl1MTCtyWefBm5TSk0PXhCOUkqdp5SK7mAdXgWuV0pNCl6P+DWmueuAUmpqcPt2oAZwA/7gNY8rlVKxwWavKsB/En8HIRpJKIh+S2u9C7gKeAIowVyUvkBr7dFae4CLgeuAcsz1h7ebfDYHc13hz8H1e4NlO1qHD4EHgLcwZydDgUXB1TGY8CnHNDGVYq57AFwNHFBKVQG3BX+HECdNySQ7QgghGsiZghBCiEYSCkIIIRpJKAghhGgkoSCEEKKRrbsr0FFJSUk6Ozu7u6shhBC9ysaNG0u01sltlet1oZCdnU1OTk53V0MIIXoVpdTBtktJ85EQQogmJBSEEEI0klAQQgjRqNddU2iJ1+slNzcXt9vd3VXp0cLDw8nIyMBut3d3VYQQPVSfCIXc3Fyio6PJzs7m2EEtRQOtNaWlpeTm5jJ48ODuro4QoofqE81HbrebxMRECYTjUEqRmJgoZ1NCiOPqE6EASCC0g/yNhBBt6RPNR0II0dUCOoAv4EOj0Vq3/Nzwusn7hs95/B48fg/1gXrz7K8/uszfbFnAvJ6TMYdxSeNC+rskFDqJ0+nE5XJ1dzWEEM007IC9AS/egLdxx+sJePD6vXgCHtw+Ny6vC5fHRY23hmpvNTWeGrOs4RFc1/S1pmunHkiOSJZQEEKIen895e5ySutKKXWXUuYuM4+6sqOv3WVU1leao+smO/yADpzQd0bYIoiyR+G0O3HanUSFRZEckUyUPYrosGii7FGEWcOwKNMKr1BYlAWFamyqbf6+4XWYJYwwq3k4rI6jz5ajy45Zbg3DbrF3SROwhEIn01pz7733snLlSpRS/PKXv+Syyy7jyJEjXHbZZVRVVeHz+fjLX/7CrFmzuPHGG8nJyUEpxQ033MBdd93V3T9BiBPmD/gprism35VPfk0+BTUF1Pvr8Qf8+LW/sekkoAP4tVnWsM6v/QQCZrkn4KHCXdG4s3d5Wz4LD7eGkxiRSEJ4AimRKQyPH37MTtRusTfufMMsYY3v7VZ74w7YbrHjsDqIDovGGWYCINIeid3SP7tu97lQ+O93t7E9v6pTtzkmLYb/umBsu8q+/fbbbNq0ic2bN1NSUsLUqVOZPXs2r776KmeffTa/+MUv8Pv91NbWsmnTJvLy8ti6dSsAFRUVnVpvITqbL+CjuLaYPFce+TX55LnyOOI6Qr7LvC6oLcAX8H3jc1ZlNQ+LebYoS+N7i7JgUzazLLjeZrER74hnbOLYxp1+4yPCPCeGJxJpj+yGv0Lf1udCobt98sknXH755VitVgYMGMCcOXPYsGEDU6dO5YYbbsDr9XLRRRcxadIkhgwZwr59+7jjjjs477zzmDdvXndXX/RTAR2g3F1OSV0JJXUlFNcVU1JXQmldKcV1xRTXFlNYW0hBTQF+7T/ms8kRyaQ50xifNJ6znWeT5kwj3ZlOqjOV1KhUImwR3fSrxInoc6HQ3iP6UGltzuvZs2ezZs0ali9fztVXX80999zDNddcw+bNm3n//fd58sknWbJkCc8++2wX11j0VQEdoKq+ijJ3GaVu0xZfWlfa2C5fXFvcGAJl7rJv7OwBouxRJEUkkRieyKSUSaRFpZHmTGvc8Q+MGojD6uiGXydCpc+FQnebPXs2f/vb37j22mspKytjzZo1PProoxw8eJD09HRuvvlmampq+OKLL5g/fz5hYWFccsklDB06lOuuu667qy96MK01Nd4aKuorqKivMBdeGy661pU1vi6tM8/l7nJ8uuWmnMTwRBIjEkmKSGJ04mgSw83r5MhkkiKSSApPIjFCmmf6IwmFTvbd736Xzz//nIkTJ6KU4ve//z0DBw7khRde4NFHH8Vut+N0OnnxxRfJy8vj+uuvJxAwvSN+85vfdHPtRVeq9daaJhp3KRVus6OvrK+kvL6cyvrKxp1/4zpPZYvt9XDsBdfUqFTGJo0lMdy8b1ieGJ5IQkQCsWGxWC3WLv61ordQrTV3dMrGlToHeBywAou11r9ttj4LeBZIBsqAq7TWucfb5pQpU3TzSXZ27NjB6NGjO7PqfZb8rULLF/BR5i5rbJYprSv9Rht9w7paX22L27BZbMQ54ohzxBHriG183XRZfHg8cY64xiP+CFuE3LEujksptVFrPaWtciE7U1BKWYEnge8AucAGpdQyrfX2JsUeA17UWr+glJoL/Aa4OlR1EuJk1fnqOOI6YnrfuPLJq8kjrzqvsQtmubu8xRuaosOiTbNMRFJjj5qG94kRicQ74hsDIMoeJTt40W1C2Xw0Ddirtd4HoJR6HVgANA2FMUBDx/yPgL+HsD5CtMkX8JFbnUuuK7exm2XDc54rjzJ32THlwyxhjRdeRyaMJCUypXFHnxSRRHJEMokRiXIxVvQaoQyFdOBwk/e5wPRmZTYDl2CamL4LRCulErXWpU0LKaVuAW4ByMzMDFmFRf+htaaotog9FXvYW76XPRV72FO+h68rvsYT8DSWs1vsZqcflcaZg84kIzqjsQdOujOdxIjExjtahegLQhkKLZ3/Nj+v/inwZ6XUdcAaIA/4xpU0rfVTwFNgril0bjVFX1ftqWZvxV72lJsdf0MAVHmO3uSYEmHuhp0+ejrD4oaRFZNFmjONpIgk2emLfiWUoZALDGryPgPIb1pAa50PXAyglHICl2itK0NYJ9GLVNZXsrl4M5uKNlFcV0xAB47/wAyRoLVuHD4hz5XHkZojjdt02p0MixvG2dlnMzx+OMPjhjM8fjixjthu/KVC9ByhDIUNwHCl1GDMGcAi4IqmBZRSSUCZ1joA/AzTE0n0Q1prcqtz+bL4S74o/IJNRZv4uvJrAGzKRlJkElZlRaGwWqyNg481PKzKilIKCxYsFot5VhYmpUzi0vhLG3f+qVGpchFXiOMIWShorX1KqduB9zFdUp/VWm9TSj0M5GitlwHfAn6jlNKY5qMfhqo+omfxBrzsLN3JF0UmAL4s+pJSt7mUFG2PZmLKRM4bch6TUiYxLmmcDJXQT2it0W43AbcbS0QEyuEIWYhrv59AdTV+lwu0xpaUhCUi9P/OGm4D6KkHJyG9eU1rvQJY0WzZg01eLwWWhrIOPdHx5l44cOAA559/fuMgeX1Bna+OQ1WHOFh1kJ1lO/my6Eu2lmzF7TdTg6Y705mZNpPJKZOZnDKZoXFD+207vvZ6CdTXo+vqCLjdjTvIQF1d42vtdhOoc6N93hP4AsDvI1DvQdfXo+vdTV7Xm+9ueO2pR7uPvlZWG5bwcFR4+LHPEeFYHM2ewyNQ4Q4sDgeB+noCrhoCLhcBlwt/jYtATc0xywI1NfhrzHv8TYbbsNmwREVhjYrC4nQGH1FmmdOJJSq4LCrKLA8Lw+9yEah2EXBV4692mR1/dXUwAKrNuupqArXfvE/E4nRiS0rClpx89JHS5HVwnSU29pideqC2Fl9ZGf7SUnylZfjLmjyXlOIrK8VfWmbKlJejrFas8fFY4+LMIz4ea1ws1rg4bN9Ybl5boqNl6GzRe/gCPo64jnCg6gAHqg5wsOpg43NBTUFjOauyMiphFAtHLGRSyiQmp0wmJTKlG2t+8ur376dy2TJq1nyM9tSjAxp08BEImPsWNBAIgNZoHTDvG9YHAo07fHwt37EcKiosDBUejnKEYQlzmCPzcEfja0tCpNnRh4WhA350XTCkamsJlJUdDaq6OhMobcwBbomMbLITNzv4sKREs3NvssziCDffU/PNMPGXV+A9nBtcVoNuYefe8Nss0dEmPKKjsUQ7caSkYIl2YnVGm3XRTizOaAB8JSX4iosbn+u2bsVXXIyuq2tx27akJFAKX1lZi2UAE16JidgSErAPGkTExIlYExLQPi/+igr85RX4Kyqo370bf3k5/spK8++kJTYbAx94gPjLLj3u3/hk9b1QWHk/FGzp3G0OHA/n/rbV1ffddx9ZWVn84Ac/AOChhx5CKcWaNWsoLy/H6/XyyCOPsGDBgg59rdvt5vvf/z45OTnYbDb+8Ic/cOaZZ7Jt2zauv/56PB4PgUCAt956i7S0NC699FJyc3Px+/088MADXHbZZSf1s1sS0AH2lO9hS8mWY3b8h6sPHzMEQ7Q9muzYbKYMmEJWTBbZMdlkxWSRFZPVJ8bT8ZWXU7VyJZX/+AfuzV+BxULkqadiTUsDiwWUQlkUoECp4DJQyqwzy1SwnBVLRDjKEW6ewyOwhDvMc8TRI3JzVB6BxeFARUSgbDaznQ5SNpvZ+dvtKEvnnpHpQMCcWTQ5y7FERJgdfmQkytr5w2tov9+ElMuFrq83wRIdjcVx8veGaK0J1NTgKyrGV1JsQqPJAw22xERsSYlYExKxJSY0eU7AEh7ese8LBMwZTXk5/ooKfE2Cw19RQfiokSf9m9rS90KhGyxatIgf//jHjaGwZMkS3nvvPe666y5iYmIoKSlhxowZXHjhhR06/XvyyScB2LJlCzt37mTevHns3r2bv/71r/zoRz/iyiuvxOPx4Pf7WbFiBWlpaSxfvhyAysrO6cSltWZvxV42FGxgQ8EGcgpzqKg38z6EWcLIjMlkaOxQ5g6aa3b+sWbnH++I7/Y2U/eu3aDAMXgwyn7yE6ZojwfXmjVU/uMfVP9nNXi9OEaMIOXee4k5/zzsKb37jKczKIvFBFcXtM03fqfVijU6Gmt0dOdvWymsTidWpxPHkMGdvv1vfJ/FgjU2Fmts9/WG63uhcJwj+lCZPHkyRUVF5OfnU1xcTHx8PKmpqdx1112sWbMGi8VCXl4ehYWFDBw4sN3b/eSTT7jjjjsAGDVqFFlZWezevZuZM2fyq1/9itzcXC6++GKGDx/O+PHj+elPf8p9993H+eefzxlnnHFCv0Vrzf7K/awvWN8YAg138aZFpTEnYw5TB07llJRTSHOm9biB1bTWuFavpnTxYupyNgKg7HbChg8jfOQowkeNxBF8tsbFtWt77i1bqPz7P6havhx/ZSXWpCQSrryS2IsWED5qVKh/khBdqu+FQjdZuHAhS5cupaCggEWLFvHKK69QXFzMxo0bsdvtZGdn426jvbW51gYrvOKKK5g+fTrLly/n7LPPZvHixcydO5eNGzeyYsUKfvaznzFv3jwefPDBFj/f/DsOVB1oPBPYULChsRfQgMgBnJY6ixmRY5ioMkmsUfgKivBuziXg2kV5XBzW+DhsCeZU2RoXjy0hHktMTKc3S7T5O7xeKpcvp+yZZ6nfswdbaiop99+HLSEB985d1O/caY7y33mn8TO2gQMJHzkSx6ijYRGWlYmyWvHm5VH57rtU/v0feA4cQDkcRJ91FrEXLSBq1izTdCNEHyT/sjvJokWLuPnmmykpKWH16tUsWbKElJQU7HY7H330EQcPHuzwNmfPns0rr7zC3Llz2b17N4cOHWLkyJHs27ePIUOGcOedd7Jv3z6++uorRo4cSUJCAldecQVRkZG88MILaL/fXMwEc4ET0/5a+PVWvjiSw6bCL9h65EsCpWXEV8Og+iju9CeR7k4irlpjKa3AV7wcvH+nFmh6OU85HOj6+pYrbrWaXhQJ8VjjE0wPioR4bPEJ2DMyiJo+DXt6eof/Hi0J1NRQsXQppc+/gO/IERzDh5P2u98SM39+Y5NR7IVHy/uKi01I7N51NCw++aSxx4sKD8eeno7na3OPROTUqSTefBPR8+aFpHlCiJ5GQqGTjB07lurqatLT00lNTeXKK6/kggsuYMqUKUyaNIlRJ9DM8IMf/IDbbruN8ePHY7PZeP7553E4HLzxxhu8/PLL2O12BiQlcf8NN7B++XJ+8dhjKIsFu83G4w88gHvHjm9s01dYSNkPbycbyAYuOmZtNRanxpaSgm1ACvYhw8zr5BRsAwZgS0nGnpKCNTkZS1gYAbcbf3m56WZXVo6/vCz4vhx/WRn+CvO6fvdu877JdQ57ZiZR06cTOWM6UdOnm54cHeArLaXs5Zcpf/U1ApWVRE6ZwsD/ehDnnDnHvZZhS07GmZyM84zTG5cFPB48e/eakNi1k/oDB4g9/zxiLriQsIzOCS8heouQzqcQCv19PgXt85k+11VVjTfdKJsNS3QMyh7MeKXwBXx4Ah7q/R48AU9wOGdFfl4JRevfYVjCCAZGp2Gx27ElJWMfkGL6X0dFhbTu9V/vo3bdOmrWraN2/XoC1dUAOIYPI3LGTKJmTCdy6lSsMTEtbsNz+DClzz5L5dvvoD0enGfNJemmm4iYNClk9RaiL2jvfAoSCr2A9nrxV1fjr6oi4KoBNMpuxxoTgyUmBktkJL6AD5fXRY23BpfXhT9gmkMcNgdOuxOn3UmkPZJdO3f1mL+V9vlw79hBzdq11K5dR+3Gjaafu8VC+JgxRM2cQeT0GUSeMpn6/fspe+YZqt57H6xWYhdcSOINN+AYMqS7f4YQvUK3T7Ijjm/Lli1cffWx8wk5HA7WrVsHmCaNQFWVCYLgzTkNN8xYY2NQ4eH4tZ/K+ioqq4qo9ZoyNoutMQSi7FHYrSffFTNUlM1GxPjxRIwfDzffTMDjwb15MzVr11Gzbi2lz79A6dOLwWYDnw9LVBQJ119HwjXXYh8g3T+FCAUJhW4yfvx4Nm3adMwy7ffjKynBX1lJIHiHpCU8HFtKCtaYGJTDQUAHqPJUU1ldhMtjhspw2BykRKYQHRaNwxq6sWJCzRIWRuTUqUROnUryHbcTqK2lduMX1K5fhzUunrhLvycXe4UIMQmFHkAHAmZclJIStN+HJSIC24ABpnkoGAQuj4tKVzHVnmq01tgtdpIikoh1xBJu69hdk72FJTIS5xmnH3NRWAgRWhIK3UgHAvjLgmHg82FxOglLScESGYnWmhpvDZWuEqrqqwjoAFaLtXEuX5moXQgRChIK3UAHAqbrZnGxCYOoKOyDBmGNiqLOV0eF6whVnip8AR8WZSEmLIZYR6xM6C6ECDkJhS50NAxK0D4vlsjIxjDwB/zku/Ipd5ejlCI6LJrYsFicYc5+O4y0EKLrhTQUlFLnAI9jJtlZrLX+bbP1mcALQFywzP3BORj6FB0ImBEPi4vR3mAYZKQ33hNQWV9JQU0BvoCPxIhEkiOSe9yYQkKI/iFkh6BKKSvwJHAuMAa4XCk1plmxXwJLtNaTMdN1/r9Q1acrXHTRRZx66qmMHTuWp556Ch0IsPzNNzll/HhOmTmTc6+/nrCsbDzJydx0xx2MnzCeMePH8Nyrz2Gz2BgSN4SBUQMlEIQQ3SaUZwrTgL1a630ASqnXgQXA9iZlNNBw62oskH+yX/q79b9jZ9nOk93MMUYljOK+afe1We7ZZ58lPj6e2vJypp92GmePG8etd9zBB6+/zvBTT6XC48Ea7eR/7r0XR5SDN//zJgD2ejtDYofI9QIhRLcLZSikA4ebvM8Fpjcr8xDwL6XUHUAU8O2WNqSUugW4BSAzM7PTK3oytN9PoK6OQG0tf/j1r1m2ciUAh/PzeXbJEmbPns3I4Hg8iYDb52bFv1bw+7/9nkh7JKlRqYRZw7r3RwghRFAoQ6Glw97mY2pcDjyvtf5fpdRM4CWl1Dit9THz0WmtnwKeAjPMxfG+tD1H9Ccj4PWia2vNTE81NQTc9YBmzYYNrPrkEz5esYKoxES+vWABp86dy9dLlqCUIqADFNcWU1pXSkAHGBg9kMzoTDk7EEL0KKEMhVxgUJP3GXyzeehG4BwArfXnSqlwIAkoCmG9OkT7fGaoiZpaArU1aG9wsnSLxdxklpyEJTKSuj17SExNJXboUHbu3Mnadeuor69n9erVbN21FUeKg+KSYrIGZnH+Oefz4lMvcsr/nQJAeXk58fHx3fgrhRDCCGVfxw3AcKXUYKVUGOZC8rJmZQ4BZwEopUYD4UBxCOvUIVprPIcO4c3PJ1DjwhIRgX3gQBxDhhI+ahSOwYOxDxiANTqac+fPx+fzMWHCBB544AFmzJhBQmICv37813xv4fe44IwL+OVtvyQ9Op0HH3iQ8vJyxo0bx8SJE/noo4+6+6cKIQQQwjMFrbVPKXU78D6mu+mzWuttSqmHgRyt9TLgJ8DTSqm7ME1L1+keNGxrIDgYnT01FWtCwnGbehwOByuD1xMAquqryK/JJzWQyn/W/ofkyOTG+w2cTicvvPBCyOsvhBAdFdL7FIL3HKxotuzBJq+3A6eFsg4nSgcCeAsLsTgcbQZCc+XucvJd+YTbwkmPSe+zYxMJIfoeuaO5Ff6yMrTHgz0r+4QCIcoeRWZMptyNLIToVSQUWqB9PnxFxVicTqzRznZ/rsxdxhHXEZxhTgZFD5JAEEL0OrLXaoGvqBgdCGAfOLDdn5FAEEL0BXKm0Eygvh5fWRnWhHgs4e27FlBWV8aRGgkEIUTvJ6HQjK+gAGVR2FPaN91jaV0pBTUFRIdFkxGdIYEghOjVZA/WhN/lwl9djTU5GWVrOy8lEIQQfY3sxYK01uYswW7HlpjYZvmSupITDgSns/0Xr4UQoitJKAT5KyoIuN3YBw5EWY7/ZympK6GwppAYR4ycIQgh+pQ+d02h4Ne/pn5Hx4bO1mh0bR1YFCo84hsj+TlGj2Lgz38OHBsI6c50LMrCfffdR1ZWFj/4wQ8AeOihh1BKsWbNGsrLy/F6vTzyyCMsWLCgzbq4XC4WLFjQ4udefPFFHnvsMZRSTJgwgZdeeonCwkJuu+029u3bB8Bf/vIXZs2a1aHfL4QQDfpcKJwI7fWitcYSFt7i0K4NimuLKaotMmcIzozGm9oWLVrEj3/848ZQWLJkCe+99x533XUXMTExlJSUMGPGDC688MI2b4QLDw/nnXfe+cbntm/fzq9+9Ss+/fRTkpKSKCsrA+DOO+9kzpw5vPPOO/j9flwuV6f8TYQQ/VOfC4WGI/r2Cni91O/egzUmmrBBg1ot1xAIsY5Y0p3px+zcJ0+eTFFREfn5+RQXFxMfH09qaip33XUXa9aswWKxkJeXR2FhIQPbuPdBa83Pf/7zb3xu1apVLFy4kKSkJAASEhIAWLVqFS+++CIAVquV2NjYDv1+IYRoqs+FQkf5CgsBjW3AgFbLHC8QGixcuJClS5dSUFDAokWLeOWVVyguLmbjxo3Y7Xays7Nxu91t1qe1z2mtZe4FIUTI9esrpIG6OvwVFdgSE7GEtTz7WWldKUW1RcQ54loNBDBNSK+//jpLly5l4cKFVFZWkpKSgt1u56OPPuLgwYPtqlNrnzvrrLNYsmQJpaWlAI3NR2eddRZ/+ctfAPD7/VRVVXXobyCEEE3121DQWuM9UoCy2rAlJ7dYJqADFNcVE2WPIs2Zdtwj9bFjx1JdXU16ejqpqalceeWV5OTkMGXKFF555RVGjRrVrnq19rmxY8fyi1/8gjlz5jBx4kTuvvtuAB5//HE++ugjxo8fz6mnnsq2bds6+JcQQoijVA+avqBdpkyZonNyco5ZtmPHDkaPHt2h7firqvAcOoQ9LQ1bsH2+uYYRT7NisnCG9Y17C07kbyWE6P2UUhu11lPaKtcvzxR0IIC3oADlcGBtZRpMrTWl7lIcNgdR9qgurqEQQnSPkF5oVkqdAzyOmXltsdb6t83W/xE4M/g2EkjRWseFsk5wdK6EsKysVpuEarw11Pvq22w2OlFbtmzh6quvPmaZw+Fg3bp1nf5dQgjRXiELBaWUFXgS+A6QC2xQSi0LzrYGgNb6ribl7wAmn+j3tbd3jvb58BWbuRIsxxluotRditViJdYRmi6e48ePZ9OmTSHZdmt6W1OhEKLrhbL5aBqwV2u9T2vtAV4HjndL7+XAayfyReHh4ZSWlrZrp+crLkb7/WY4i1ZCpN5Xj8vjIiE8oc8MYaG1prS0lPB2DgcuhOifQtl8lA4cbvI+F5jeUkGlVBYwGFjVyvpbgFsAMjMzv7E+IyOD3NxciouLj1shM6NaEZbISKz797darrK+klpvLTpKU6pKj7vN3iQ8PJyMjIzuroYQogfV12qrAAAgAElEQVQLZSi0dBje2qH8ImCp1trf0kqt9VPAU2B6HzVfb7fbGTx4cJsVOnz77dR+9jlD//U+tuCdwc1V1ldy7dJrmZc1j0emPtLmNoUQoi8JZdtILtB03IgMIL+Vsos4waaj9qpZvx7XBx+SeOutrQYCwNLdS6nz1XH1mKtbLSOEEH1VKENhAzBcKTVYKRWG2fEva15IKTUSiAc+D2Fd8ObmETZ0KAnXXtN6mYCX13a+xvSB0xmZMDKU1RFCiB4pZKGgtfYBtwPvAzuAJVrrbUqph5VSFzYpejnwug5x15i4i7/LkGX/OO68yx8c/IDC2kI5SxBC9FshvU9Ba70CWNFs2YPN3j8Uyjo0pazWVtdprXlp+0tkxWRxRsYZXVUlIYToUfpGf8tOsLl4M1tKtnDl6Cv7TDdUIYToKNn7Bb20/SWiw6JZMLTt2dGEEKKvklAA8l35fHDoAxaOWEikPbK7qyOEEN1GQgF4dcerKBRXjLqiu6sihBDdqt+EgscXYO2+b96dXOOt4a09b/GdrO8wMOr4U2UKIURf129C4U8f7uGqxes4WFpzzPK/7/07Lq9LuqEKIQT9KBSumZmFzar4vw/2NC7zB/y8suMVJiZPZELyhG6snRBC9Az9JhRSYsK5dmY2f9+Ux+7CagBW567mcPVhOUsQQoigfhMKALfNGUpUmI0//Gs3AC/veJnUqFTOyjyrm2smhBA9Q78KhfioMG48fTDvbSvg3R05bCjYwBWjrsBmCemN3UII0Wv0q1AAuOmMwcRF2nls7dNE2CK4eMTF3V0lIYToMfpdKESH27nmtHhK1TpmpZxLTFhMd1dJCCF6jH4XCgCW2LUoAhzYd4rMWyyEEE30u1Bw+9y8vfdNhjmnsWm/jTV7Srq7SkII0WP0u1BYvm855fXl3DPzZtLjIvjff+2SswUhhAjqV6GgteblHS8zKmEUM9Om8aNvD+er3Ere31bY3VUTQogeIaShoJQ6Rym1Sym1Vyl1fytlLlVKbVdKbVNKvRrK+nx+5HP2VuzlqtFXoZTi4snpDEmO4g//3oU/IGcLQggRslBQSlmBJ4FzgTHA5UqpMc3KDAd+BpymtR4L/DhU9QEzZ0JieCLnDj4XAJvVwt3fGcHuQhfLNueF8quFEKJXCOWZwjRgr9Z6n9baA7wONJ/B5mbgSa11OYDWuihUldlXsY9P8j5h0ahFhFnDGpfPH5fK6NQY/vjvPXj9gVB9vRBC9AqhDIV04HCT97nBZU2NAEYopT5VSq1VSp3T0oaUUrcopXKUUjnFxcUnVJn3DrxHmCWMS0deesxyi0Xx03kjOFRWy5s5uSe0bSGE6CtCGQqqhWXNG+5twHDgW8DlwGKlVNw3PqT1U1rrKVrrKcnJySdUme9P/D5vXvAmCeEJ31g3d1QKkzPjeGLVHtxe/wltXwgh+oJQhkIuMKjJ+wwgv4Uy/9Bae7XW+4FdmJDodEophsQNaXXdPWeP5Eilm1fWHQrF1wshRK8QylDYAAxXSg1WSoUBi4Blzcr8HTgTQCmVhGlO2heS2mxZCs/MA7+vxdWzhiZx2rBE/t9He6mpb7mMEEL0dSELBa21D7gdeB/YASzRWm9TSj2slLowWOx9oFQptR34CLhHa/3NOTM7gy0cDq+DXStaLfLTeSMprfHw3Kf7Q1IFIYTo6VRvu5t3ypQpOicnp+MfDPjh8UkQlwnXL2+12E0vbGDd/jI+uXcusZH2k6ipEEL0HEqpjVrrKW2V6z93NFusMO0mOPgJFGxttdjd3xlJtdvHUx9/3YWVE0KInqH/hALA5KvBFgHr/9ZqkTFpMVwwMY3nPj1Aiau+CysnhBDdr12hoJT6kVIqRhnPKKW+UErNC3XlOl1kAky8DL5aArVlrRa769vDqfcF+H8fydmCEKJ/ae+Zwg1a6ypgHpAMXA/8NmS1CqVpt4DPDV+82GqRIclOLjklnZfXHiS/oq4LKyeEEN2rvaHQcCPafOA5rfVmWr45recbMBayz4ANi1vtngpw51nD0WieWLWnCysnhBDdq72hsFEp9S9MKLyvlIoGeu9AQdNvhcrDsHtlq0Uy4iO5YlomS3JyOVBS04WVE0KI7tPeULgRuB+YqrWuBeyYJqTeacS5EJsJ61q/4Azww7nDsFsVf/j37i6qmBBCdK/2hsJMYJfWukIpdRXwS6AydNUKMasNpt4IBz4+bvfUlOhwbjljCMs25/Pe1iNdWEEhhOge7Q2FvwC1SqmJwL3AQaD1K7W9wSnXBLunPnXcYrfPHc7EjFjuXfoVeXLRWQjRx7U3FHza3Pq8AHhca/04EB26anWByASYcGmb3VPDbBb+dPlkAhp+/PqX+GTOBSFEH9beUKhWSv0MuBpYHpxVrfePATH9VvDVwZcvHbdYVmIUj1w0jg0Hynli1d4uqpwQQnS99obCZUA95n6FAsxkOY+GrFZdpaF76vrFZmyk47hocjoXn5LOE6v2sG5faMbsE0KI7tauUAgGwStArFLqfMCtte7d1xQaTLsFKg/Brta7pzZ4eME4MhMi+fEbm6io9XRB5YQQomu1d5iLS4H1wPeAS4F1SqmFoaxYlxk5H2IHwbq/tlnU6bDxxOWnUOKq5763vqK3jTArhBBtaW/z0S8w9yhcq7W+BpgGPBC6anWhpt1TC7e1WXx8Riz3nj2K97cVyixtQog+p72hYNFaFzV5X9qezyqlzlFK7VJK7VVK3d/C+uuUUsVKqU3Bx03trE/nOuVaMwlPG91TG9x4+mDmjEjmf/65nV0F1SGunBBCdJ32hsJ7Sqn3gzvx64DlQOtTmAHBHkpPAucCY4DLlVJjWij6htZ6UvCxuAN17zwN3VM3v3Hc7qkNLBbFY9+bSHS4nTte+wK39/gXqYUQordo74Xme4CngAnAROAprfV9bXxsGrBXa71Pa+0BXsfc59AzTWvonvpyu4onRzv4w6UT2V3o4pHl20NcOSGE6BrtnmRHa/2W1vpurfVdWut32vGRdOBwk/e5wWXNXaKU+koptVQpNailDSmlblFK5SilcoqLi9tb5Y4ZOA6yTof1T7fZPbXB7BHJ3Dp7CC+vPSTDYAgh+oTjhoJSqlopVdXCo1opVdXGtlsaWrt5d513gWyt9QTgA+CFljaktX5Kaz1Faz0lOTm5ja89CdPb3z21wU/mjWSCDIMhhOgjjhsKWutorXVMC49orXVMG9vOBZoe+WcA+c22X6q1bpjz8mng1I7+gE418jyIyTjudJ3Nhdks/GnRZPwBzV2vb5JhMIQQvVoo52jeAAxXSg1WSoUBi4BlTQsopVKbvL0Q2BHC+rStoXvq/jVQ2P7rBNlJUTzy3XGsP1DGnz+SYTCEEL1XyEJBa+0Dbgfex+zsl2ittymlHlZKXRgsdqdSaptSajNwJ3BdqOrTbh3sntrgu5MzuHhyOn/6cA/r97fdg0kIIXoi1dvuyp0yZYrOyckJ7Zf844ew9W24eztExLf7Y656H+f/6WM8vgArfnQGcZFhIaykEEK0n1Jqo9Z6SlvlQtl81HtNuxW8te3untqgYRiMYlc997+1RYbBEEL0OhIKLUmdAFmnmSakdnZPbdAwDMZ72wp4ae3BEFVQCCFCQ0KhNdNugYpDsPu9Dn/0xtMHM3dUCg8t28byr+T+BSFE7yGh0JpR50NMOqxrf/fUBhaL4s9XTObUrHh+9PqXrNpZGIIKCiFE55NQaE1j99TVUNTxnrKRYTaeuW4qo1NjuO3lL/js65IQVFIIITqXhMLxnHIdWB2w+vfg6vjwGjHhdl68YRrZiZHc9EIOGw+Wd34dhRCiE0koHE9UIky5Aba9DY8Ng7+eAR88BPs/Bl/7Zl6Ljwrj5RunkxLt4Lrn1rM1rzK0dRZCiJMg9ym0JRCAI5vg6w9h7yrIXQ8BH9ijYPAZMPQsGHYWJAwB1dJwT0ZueS2X/vVz3L4AS26dwbCU6K77DUKIfq+99ylIKHSUu8rM0rb3QxMU5QfM8rjMowExeDaEx37jo/tLavjeXz/HaoE3b51FZmJk19ZdCNFvSSh0lbJ9wYBYZcZM8rhAWSFjKky/FcZdfEzxXQXVXPbU5zgdNt68bSapsRHdVHEhRH8iodAd/F44vN6cQexcDsU7Yf5jMO3mY4p9lVvBFU+vIyXGwZJbZ5LkdHRThYUQ/YUMc9EdrHbIPg3OehBuXQMj58OKn8JnTxxTbEJGHM9dP5X8ijquWryOitr2XbQWQohQk1AIFZsDLn0RxlwE//olrH70mNVTsxN4+pop7Cuu4drnNuCq93VTRYUQ4igJhVCy2uGSZ2DCIvjoEfjwYWjSXHfG8GT+fMVktuZVcuPzG6jzdGycJSGE6GwSCqFmtcFFfzHzNHz8v/D+L44JhnljB/KHSyey/kAZt728kXqfBIMQovtIKHQFiwUueNwMyb32SVj+E3P/Q9CCSen85rvjWb27mB+9JlN6CiG6T0hDQSl1jlJql1Jqr1Lq/uOUW6iU0kqpNq+M91pKwbm/g9N+BDnPwLI7jhmWe9G0TB44fwzvbSvgxhdyKK6uP87GhBAiNEIWCkopK/AkcC4wBrhcKTWmhXLRmKk414WqLj2GUvDt/4Y598Oml+GdW8F/9ALzjacP5pGLxvH5vlLOfXwN/9lV1I2VFUL0R6E8U5gG7NVa79Nae4DXgQUtlPsf4PeAO4R16TmUgjN/Bmf9F2x5E5Zed8w4SlfNyOLd208nMcrBdc9t4OF3t8t1BiFElwllKKQDh5u8zw0ua6SUmgwM0lr/83gbUkrdopTKUUrlFBd3fLTSHumMu+Gc38KOd+GNq8B7NBNHDozmH7efxrUzs3j20/1c9ORn7C2q7sbKCiH6i1CGQkujwzV2u1FKWYA/Aj9pa0Na66e01lO01lOSk5M7sYrdbMb34fw/wp734bVF4KltXBVut/LfC8bxzLVTKKxyc/4Tn/DqukMy77MQIqRCGQq5wKAm7zOA/Cbvo4FxwH+UUgeAGcCyPn2xuSVTbjBdVvevhlcWQv2xZwRnjR7Aez86g6nZCfz8nS3c9vJGymvkDmghRGiEMhQ2AMOVUoOVUmHAImBZw0qtdaXWOklrna21zgbWAhdqrXvowEYhNOkKuGQxHFoLL30X6o6djCclJpwXrp/GL+aPZtXOIs59/GOZyU0IERIhCwWttQ+4HXgf2AEs0VpvU0o9rJS6MFTf22uNuwQufQHyN8ETU2D902aAvSCLRXHz7CG884PTiAyzcuXidfz+vZ145Z4GIUQnklFSe5r8L+H9X8LBTyBxmOnCOuq8Yybwqan38fC723kj5zATB8Xxp0WTyEqM6sZKCyF6OhkltbdKmwzX/RMufx2UBd64Ep6bD7kbG4tEOWz8buEEnrziFPYXu5j/+Me8/UWuXIQWQpw0CYWeSCkYeS58/3M47w9QugcWz4WlNxyd6Q04b0IqK388m7Fpsdy9ZDNXP7OeLbkyB7QQ4sRJ81FvUF8Nn/7JzMug/TDtFjjjJxCZAIA/oHnhswM8sWoP5bVezpuQyk/njWRwkjQpCSEMmXmtL6rKh49+BV++YuaAnn2PmdXNZmZuq3J7eXrNPhZ/vB+vP8BlUwfxo7OGkxIT3s0VP0F+H+Suhz3/NtdXJl1xzLUVIUT7SSj0ZQVb4d8Pmmk/47Lg2/8FYy9u3GEWVbt54sO9vLb+EHarhRtPH8wtc4YQE27v5oq3Q10F7P0Adr8Pe/8d7J6rAA3jv2dGmw2TMyAhOkpCoT/4ehX860Eo3GIuUA+aDo5oCHOCw0mxx8Hb2yv5aH8dFoeTi2aM4sJpIwmPigN7RM856i7ZC7vfM4+Dn5kmsshEGD4PRpwNQ86EDU/Dql9Bymi47GVIHNrdtRaiV5FQ6C8CfvjqDfjk/6C6ADzVoNu+d0ErC8oRDZFJMPlKmHozhMd0QYUx918c+tycDexaCWVfm+UpY2DEOeaRMQUs1mM/t/dDeOtG85u/+1fTVVcI0S4SCv2V1uCthXqXuUDtqTbP9S52H8rnw81fU1FRTpbTz+mZ4QzyHUTtXw2OWJh+qxmPKXgBu1P5PLB7JWz7u9m511eCNQyyzzA9rYbPg/istrdTcQjeuBqObDIX28/8xTfDQ4i+yOs2/89YTqzTqISCaJHWmhVbCnjsX7vYX1LDlKx47ptYy5RDz6F2/hPsUTDleph1B0QPPPkvLNgKm14xZzO1pRCVbJqERpxjmoUczo5v0+uGlffAFy/CkG/BJc9CVOLJ11X0XH6v+fdTUxx8lAQfTd8Xg7vSNIsqi3nQ8Fq1sjy4LjrVnDEPmXvCO92Q8dWbf+sf/6+ZqGtMSzMQtE1CQRyX1x9gSc5h/rxqL0cq3QxJjuLuiX7OKX8N2/a3wGKHyVeZmeLacwTfVF05bFkKX75sjugtdtPUM/lqGHpm5x3Zb3wBVtwDzhQzREj6qSe3vYAf9q+BAx/D6AvMdRrR9XI3wtr/B9VHju70m40H1shiM02gUckQlWR65QGgTTOq1sFHINis2uR10+VF203oxGbCqdfApKsgJrWrfnHLfPXw5Uvw8R+gKg8yZ8F3HoZBU09ocxIKol28/gArthxh8cf72ZJXSXyknR9OsnKl920itr0BaJhwGZx+FyQNb31DgQDs/48Jgh3/BH89DBhvgmX890J3JJ/3BSy5BlyFMP9ROPW6jn0+EIDcDbB1qWnaqgnOdmexwdxfwqw7pXmqq9SUwof/bY6KI+LNNaaoxOAOP7jTb3ydbDojhMd1zpG9rx52/hM2Pm8ODJTVNGueeh0Mndu1/wZ8HnN2/fH/QuVh04HkWz8zZ8Un0TlEQkF0iNaaDQfKWfzxPv69oxC7xcLVY638MGwlCTtfA58bxl5k2vEHjj/6wfIDsOlV86g8bP4nnXCpCYPUiV1T+ZpSePsm0xtr8lUw/zHTu6o1WkPBV7D1Ldj6tqm3Ldw0a427BDKmwnv3w/Z/QNbpcPHfIDaja35LfxTwm53xqv8Bd5W5rjXnvq7r+NBc6dcmmDa9Ys5SYgeZs9zJV0FsetufP1F+L2x+DdY8aq6dpU8xszQOPatTegpKKIgTdqCkhuc+3c+SnFzqvH7OHWzl/rhVZH79KspTba4HDJ8H294xTS0oczQ1+SoYOR/s3XCzXMAP//mN+R8qdSJc+tI3m71K9pgg2LLUDB1isZl6j1tojgqb7oS0NkG38l5zlHj+H01giM6VmwPLf2KaGbPPMGd7KaO7u1aGzwO7VpjA2veRuf4w/Gxz9jD8O5139uD3mWtua35vDrLSJpsOFMO+3andxiUUxEmrqPXw2vrDPP/Zfgqr6pmYpPmfgZ8yPu81VF05xGebttdJl/ecI+ldK+HtW02TwiWLIWlk8IzgLXN2gILs080OfvSFbTdrle2Dt26GvByYeDmc+/vuO4LtS2pK4IOHTJt5dCrMe8T8N+kp9840V7bfnD18+bJpYoxJN2cPw+eZDhnOFLB28OZQv880W67+nfl3ljoRvvVzc8Yagr+DhILoNB6fue7w9Mf72JZfRXqkn5vHwWmz5jB8YGzbG+hqpV+bbqtF22mcATb9VHNGMPYiiEnr2Pb8XnMGsuZR05Rw8dOQOb3Tq90vBPyQ8yysegQ8rqNNRY7o7q5Z+/i95sDjixdM12qa7D8jE8E5EKIHgDP4aAgM58Cjr+2R5iBl9e+gdK+59nbmz8xZdghDUUJBdDqtNev2l7H44318uLMIrWHEACfzx6dy/oRUhqX0oP+xPbXw6f+Zft3jLoGEwSe/zUNr4e2boTIXZt9rxp6y2k5+u/3F4fWw4qdwZDMMng3nPgopo7q7Vieu4jAUbjU3jboKg89F4CqA6kKzLOD95ucsdrM8ZSx8634YdX6XdIPtEaGglDoHeBywAou11r9ttv424IeAH3ABt2ittx9vmxIKPUNRlZv3thXwz6+OsOFAGVrDyAHRnDchlfnjUxmWcgL3H/QG7ipznWHza+aC9MVPQcKQrq1D8W7Y+a5pf45MbNIlM9hTJzLJ9NQJDpTY7VzFpqlo08sQnQZn/wrGfrfnNhV1Fq1NV9rqAhMUriLzuqbY/NsZfWGX3hPR7aGglLICu4HvALmYOZsvb7rTV0rFaK2rgq8vBH6gtT7neNuVUOh5CqvcrNxyhBVbCthw0ATEqIHRnDc+lfMmpDIkuQ8GxNa34N27zDhN5/4+tCO4BgJmRr6d75ruvqV7zPKoZLPTCfha/pwjxoRD89BIGmGO1DvajNZRlbmwfRms/i14amDmD80Z1oncsChOWk8IhZnAQ1rrs4Pvfwagtf5NK+UvB67RWp97vO1KKPRsBZVuVm49wootR9hwwNxwNDo1hvPGD+S8CWl9a46HisPwzm1m6tQxF5keSp01RIjfCwc+MX3nd66A6nzTdz77dHNj3cj5pnuk1uCuMN1ya4qhtuEu39bel5ogAxMOQ74Fg+eY7UbEnVydXUVHb/7bv8ZcPAXzHec+CskjTm774qT0hFBYCJyjtb4p+P5qYLrW+vZm5X4I3A2EAXO11nta2NYtwC0AmZmZpx48eDAkdRad60hlHSu3FLB8yxE2HjQBMSEjlmtmZnP+hFTC7X3gprCAHz77k7lwGpFgLmjHpJlHbEbwdbp5Pt69E2COpvd+aIJg93tmyAZbBAw7ywTB8HknHzoBv2kH37ca9q82o9J6a013y7TJJiCGfMvcMNVW1+LaMjj4qQmA/R9D8Q6z3BEDWaeZs5HBZ8CAcX2/qagX6Amh8D3g7GahME1rfUcr5a8Ilr/2eNuVM4XeKb+ijhVbjvDGhsPsKXKREBXGFdMyuWpGFgNje+kkQE3lfwlrHoPyg1CV2/KwDBEJJiBi048GR0yGaf7ZtdLcfOerM3fzjjgXRp9vxocKiwxdvX0ec0f3vv+YkMjNMWcStnDInHE0JFInmvA4+Lkpt38NFGwBtOlNkznTBMDg2TBwolyA74F6Qih0tPnIApRrrY/bx1FCoXfTWvPZ16U8/9kBPthRiEUpzhk3kOtmZTMlKx7VV44oPbVm7J7KXDNjXlVe8BF8XZkHdWVHy8dkmPGhRp9vxrjprp2qu8qcPexfbYKiKHgJ0BFjzmS0H6wOGDTNBED2GebsyBbWPfUV7dYTQsGGudB8FpCHudB8hdZ6W5Mywxuai5RSFwD/1ValJRT6jsNltby09iCvrz9EldvH2LQYrp2VzYUT0/pG01JbvHUmJPweSB7VM5tYGq8TfGJ6Og2ebQKhraYw0eN0eygEKzEf+D9Ml9Rntda/Uko9DORorZcppR4Hvg14gXLg9qah0RIJhb6n1uPj71/m8/xn+9ldaJqWFk0dxFUzskiLk52PEJ2hR4RCKEgo9F1aaz7fV8rzn5qmJaUUZ48dwHWzBjM1uw81LQnRDdobCnI1SPQYSilmDU1i1tAkDpfV8vLag7y+4TArthSQ5AwjOzGKzMRIshOjyEqMJCsxiuzESOIipT1biM4iZwqiR6vz+Fm2OY8vDlZwoLSGg6W1FFS5jykTG2E/JiQyEyLJTjLBkex0yBmGEEjzkejD3F4/h8pqOVBSY56DYXGwtJbc8loCTf5JZydGNg69MSY1RgJC9FsSCqJf8vgC5FXUcbC0hq+La/hoZxGf7yvFH9BkJ0YyPzj0hgSE6G8kFIQIKnXV86/thSz/6ogEhOi3JBSEaEFZjYf3txWwYssRPvv62ICYPz6VsWkSEKJvklAQog2tBcS541OZPTyZyZlx/eMmOtEvSCgI0QFlNR7+tc0M3tcQEGE2C5MGxTFjcALThyRySmY8EWESEqJ3klAQ4gRV1nnJOVDG2n2lrNtfxta8SgIa7FbFhIw4pg9OYMaQRE7NiifKIbf6iN5BQkGITlLt9pJzsJx1+0xQbMmrxB/QWC2K8emxTB+SwIzBiUzJjic6vIOTtwvRRSQUhAiRmnofGw+Ws25/Kev2lbE5twKvX2NRMCzFyYSMOCYOimNiRiyjBsYQZuu6KReFaI2EghBdpM7j54tD5azfX8ZXuRVszq2krMYDQJjVwui0GCZmxDIhI45Jg2IZkuTEYpEeTqJrSSgI0U201uSW1/FVbiWbcyvYfLiCrXmV1HjMNJhOh41x6TFMDJ5RjE+PJT0uQoJChJQMiCdEN1FKMSghkkEJZogNAH9A83Wxi82HKxrD4tlP9+P1m4OyyDArw1KcDE+JZvgAJ8NTnIwYEC1hIbqcnCkI0U3qfX52Hqlma34lewpd7C1ysbuwmqLq+sYyEfaGsHAybICTEcHQGBQfKWEhOqRHnCkopc4BHsdMsrNYa/3bZuvvBm4CfEAxcIPW+mAo6yRET+GwWc0F6UFxxyyvrPWyt7ia3YUu9hS62FNUzWdfl/L2l3mNZcLtFkYOjGHGENM9dkqW9HwSnSOU03FaMdNxfgfIxUzHebnWenuTMmcC67TWtUqp7wPf0lpfdrztypmC6K+q3F72FrnYU2gCY0tuJV8eLsfrN91jx6XHNobE1OwEnHIPhWiiJ5wpTAP2aq33BSv0OrAAaAwFrfVHTcqvBa4KYX2E6NViwu2ckhnPKZnxjcvqPH6+PFTO2n2lrN1XxrOf7Odvq/dJSIgTFsp/JenA4Sbvc4Hpxyl/I7AyhPURos+JCLMya1gSs4YlAUe7x5qQKG0xJKZlJzBpUByJTkc31170RKEMhZaugrXYVqWUugqYAsxpZf0twC0AmZmZnVU/IfqciDArpw1L4rQ2QgIgKzGSSYPimDwojsmZ8YxOlRvtRGhDIRcY1OR9BpDfvJBS6v+3d68xctV1GMe/z8xe2+1u2+2FbQu9Emwt2AJSEMUajKm84JJAVJCgb/AFJJL4QrxFQ8Irg/qGKDUSagQVkJvGCNoYlBdQesPeIFRoS9ulte122y3dy8z8fHHOTre7W7osnZ3tzPNJNufMOS5n5DcAAAkaSURBVGfP/uefs/vsOf85v/NF4AfA5yOiZ/B6gIhYDayGZEzh3DfVrDINFxJb9nWyaU8Hm/Yc5dV3DvP85uTXsq4mw9JZzSy/aArLL0qCYlZLg0uJV5lSDjTXkAw0Xw/sIxlovj0itg3YZjnwNLAqIt4eyX490Gx2brV3nmTTnqPFoNiyr5OeXAGA6ZPqi2cSV86bwmVzWqivcaXY81HZB5ojIifpXuBFko+kPhoR2yQ9AKyPiBeAnwJNwFPpfyN7IuLGUrXJzIZqa2mk7dJGbrg0udGuL1/gzfbjbHqvoxgWL20/ACRnE8svTCrFXjW/lcvnTmZCnQewK4lvXjOzszrc1cP63Ul9p3XvHmHb/qSceE06gL1iwVRWzJ/KFXOn0tLo+yXGI9c+MrOSOd7dx4YBIdFfKVaCxRc0c9X8JCSunDeVaU11HpcYBxwKZjZmuvvybNpzNAmJXYfZsLuD7r5CcX19TYaG2iwNtcm0+LomS31thvqa09dNrK9hbusEFk1vYtGMJn989hwo+5iCmVWPhtos1yxs5ZqFrcDF9OYKbN3fycbdHRw72UdPrkB3X57uvgI9uWTancvT01egqyfHoa5eevryxe2O9+TozZ0KlSkTalk0IwmIhWlQLJrRxKwWFww81xwKZnbO1dVkhtx9/VEUCsH+zpPsPJgUCvzv/5Lp37a+T8cHfcXtGmuzLJwxsXhGsWB6E7MnN9I2uYFpE+sdGKPgUDCzcSeTEXOmTGDOlAmsvGTGaesOd/UkYZEGxc6DXbz27hGe23z6bVB12QwzW+ppa2lMgqKlgbbJjcxqaaCtpZFZkxtoaaz1eMcgDgUzO6+0NtXT2lTPigWtpy3v6smx69AJ2ju7ae88yb6jJ2k/msyve/cIB451kyucPoY6oS5LW0sDrU311NdkqMtmqM1mqKsZ8JU9fTpw/fSmepa0NXPh1MaKCReHgplVhOSJdi0snd0y7Pp8ITjU1cP+oydp7+wuTts7T3K4q5cTPTk68gV6c8lXXz7oyRXozeXpywe9+QL5wvAfzJlUX8Mn2iaxpK2ZxenXJRdMoqH2/LvRz6FgZlUhmxEzmxuY2dzA8lHuI1+IJDTS8Nh39CQ72o+xff8xdrQf4+kNe4uPXc0IFkxvYnFbcxoWk1gyq5kZkxpG9bP7Pyla6jMSh4KZ2QhlM6KxLksjyRnA9En1LBvwkKRCIXiv44NiUGxvP87G3R38+Y1T4x3TmupobqglH0G+EBQKkc5DYciyOLUs4MFblnLHirklfY8OBTOzcySTEXNbJzK3dSKrlrYVl3d+0MeO95OgePP9Y5zozZOVyGZERiKbgWwmk0wlMhmdWp/OZzLi0jNcGjuXHApmZiXWMqGWqxe0cvWgwfHxyMXTzcysyKFgZmZFDgUzMytyKJiZWZFDwczMikoaCpJWSXpL0k5J9w+z/jpJGyXlJN1ayraYmdnZlSwUJGWBh4EvA0uAr0laMmizPcA3gCdK1Q4zMxu5Ut6ncBWwMyLeAZD0B+AmYHv/BhGxK11XGG4HZmY2tkoZCrOB9wa83gusGM2OJN0N3J2+7JL01ijbNA04NMrvrVTuk+G5X4Zynwx1PvXJiOpjlDIUhqvaNKpnf0bEamD1x2sOSFo/ksfRVRP3yfDcL0O5T4aqxD4p5UDzXuDCAa/nAPvPsK2ZmY0DpQyF14GLJc2XVAd8FXihhD/PzMw+ppKFQkTkgHuBF4EdwJMRsU3SA5JuBJD0aUl7gduARyRtK1V7Uh/7ElQFcp8Mz/0ylPtkqIrrE/U/uMHMzMx3NJuZWZFDwczMiqomFM5WcqMaSdolaYukzZLWl7s95SDpUUkHJW0dsGyqpL9LejudTilnG8vhDP3yE0n70uNls6QbytnGsSTpQkn/lLRD0jZJ306XV9yxUhWhMMKSG9XqCxGxrNI+a/0RPAasGrTsfmBtRFwMrE1fV5vHGNovAD9Pj5dlEfHXMW5TOeWA70TEYuBq4J70b0jFHStVEQoMKLkREb1Af8kNq3IR8S/gyKDFNwFr0vk1wM1j2qhx4Az9UrUioj0iNqbzx0k+UTmbCjxWqiUUhiu5MbtMbRlPAnhJ0oa0lIglZkZEOyR/DIAZZW7PeHKvpP+kl5fO+0sloyFpHrAceI0KPFaqJRTOWcmNCnNtRFxOclntHknXlbtBNq79ElgILAPagYfK25yxJ6kJ+BNwX0QcK3d7SqFaQsElN4YREfvT6UHgWZLLbAYHJLUBpNODZW7PuBARByIiHxEF4NdU2fEiqZYkEB6PiGfSxRV3rFRLKLjkxiCSJkqa1D8PfAnY+uHfVTVeAO5K5+8Cni9jW8aN/j9+qVuoouNFkoDfADsi4mcDVlXcsVI1dzSnH5/7BZAFHo2IB8vcpLKStIDk7ACSarlPVGOfSPo9sJKkBPIB4MfAc8CTwEUkD4K6LSKqatD1DP2ykuTSUQC7gG/1X0+vdJI+C/wb2AL0P//l+yTjChV1rFRNKJiZ2dlVy+UjMzMbAYeCmZkVORTMzKzIoWBmZkUOBTMzK3IomI0hSSsl/aXc7TA7E4eCmZkVORTMhiHp65LWpc8NeERSVlKXpIckbZS0VtL0dNtlkl5NC8U9218oTtIiSf+Q9Eb6PQvT3TdJelrSm5IeT++WNRsXHApmg0haDHyFpGDgMiAP3AFMBDamRQRfJrnLF+C3wHcj4jKSO177lz8OPBwRnwI+Q1JEDpIKm/eRPNtjAXBtyd+U2QjVlLsBZuPQ9cAVwOvpP/GNJIXOCsAf021+BzwjqQWYHBEvp8vXAE+ldaVmR8SzABHRDZDub11E7E1fbwbmAa+U/m2ZnZ1DwWwoAWsi4nunLZR+NGi7D6sR82GXhHoGzOfx76GNI758ZDbUWuBWSTOg+BzeuSS/L7em29wOvBIRnUCHpM+ly+8EXk5r7e+VdHO6j3pJE8b0XZiNgv9DMRskIrZL+iHJU+kyQB9wD3AC+KSkDUAnybgDJCWTf5X+0X8H+Ga6/E7gEUkPpPu4bQzfhtmouEqq2QhJ6oqIpnK3w6yUfPnIzMyKfKZgZmZFPlMwM7Mih4KZmRU5FMzMrMihYGZmRQ4FMzMr+j9SRV+arzpuRQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.plot(history.history['acc'])\n",
    "plt.plot(history.history['val_acc'])\n",
    "plt.title('model loss')\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['loss', 'val_loss', 'acc', 'val_acc'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "해당 catcat103.jpg  이미지는 고양이 으로 추정됩니다.\n",
      "해당 catcat11 - 복사본.jpg  이미지는 고양이 으로 추정됩니다.\n",
      "해당 catcat15 - 복사본.jpg  이미지는 고양이 으로 추정됩니다.\n",
      "해당 catcat22 - 복사본.jpg  이미지는 고양이 으로 추정됩니다.\n",
      "해당 catcat27 - 복사본.jpg  이미지는 고양이 으로 추정됩니다.\n",
      "해당 catcat30 - 복사본.jpg  이미지는 고양이 으로 추정됩니다.\n",
      "해당 catcat31 - 복사본.jpg  이미지는 고양이 으로 추정됩니다.\n",
      "해당 catcat37.jpg  이미지는 고양이 으로 추정됩니다.\n",
      "해당 catcat4.jpg  이미지는 고양이 으로 추정됩니다.\n",
      "해당 catcat40.jpg  이미지는 고양이 으로 추정됩니다.\n",
      "해당 catcat48 - 복사본.jpg  이미지는 개 로 추정됩니다.\n",
      "해당 catcat53 - 복사본.jpg  이미지는 고양이 으로 추정됩니다.\n",
      "해당 catcat56 - 복사본.jpg  이미지는 개 로 추정됩니다.\n",
      "해당 catcat67 - 복사본.jpg  이미지는 고양이 으로 추정됩니다.\n",
      "해당 catcat69 - 복사본.jpg  이미지는 고양이 으로 추정됩니다.\n",
      "해당 catcat70 - 복사본.jpg  이미지는 고양이 으로 추정됩니다.\n",
      "해당 catcat70.jpg  이미지는 고양이 으로 추정됩니다.\n",
      "해당 catcat72 - 복사본.jpg  이미지는 개 로 추정됩니다.\n",
      "해당 catcat9.jpg  이미지는 개 로 추정됩니다.\n",
      "해당 catcat99 - 복사본.jpg  이미지는 고양이 으로 추정됩니다.\n",
      "해당 dog개1.PNG  이미지는 개 로 추정됩니다.\n",
      "해당 dog개2.PNG  이미지는 고양이 으로 추정됩니다.\n",
      "해당 dog개3.PNG  이미지는 개 로 추정됩니다.\n",
      "해당 dog개4.PNG  이미지는 개 로 추정됩니다.\n",
      "해당 dog개5.PNG  이미지는 개 로 추정됩니다.\n"
     ]
    }
   ],
   "source": [
    "from PIL import Image\n",
    "import os, glob, numpy as np\n",
    "from keras.models import load_model\n",
    "\n",
    "import tensorflow as tf\n",
    "\n",
    "seed = 5\n",
    "tf.set_random_seed(seed)\n",
    "np.random.seed(seed)\n",
    "\n",
    "caltech_dir = './binary_img_data/img_test'\n",
    "\n",
    "\n",
    "image_w = 64\n",
    "image_h = 64\n",
    "\n",
    "pixels = image_h * image_w * 3\n",
    "\n",
    "X = []\n",
    "filenames = []\n",
    "files = glob.glob(caltech_dir+\"/*/*.*\")\n",
    "for i, f in enumerate(files):\n",
    "    img = Image.open(f)\n",
    "    img = img.convert(\"RGB\")\n",
    "    img = img.resize((image_w, image_h))\n",
    "    data = np.asarray(img)\n",
    "\n",
    "    filenames.append(f)\n",
    "    X.append(data)\n",
    "\n",
    "\n",
    "X = np.array(X)\n",
    "X = X.astype(float) / 255\n",
    "model = load_model('./model/dog_cat_classify.model')\n",
    "\n",
    "prediction = model.predict(X)\n",
    "np.set_printoptions(formatter={'float': lambda x: \"{0:0.3f}\".format(x)})\n",
    "cnt = 0\n",
    "for i in prediction:\n",
    "    if i >= 0.5: print(\"해당 \" + filenames[cnt].split(\"\\\\\")[1] + filenames[cnt].split(\"\\\\\")[2] + \"  이미지는 개 로 추정됩니다.\")\n",
    "    else : print(\"해당 \" + filenames[cnt].split(\"\\\\\")[1] + filenames[cnt].split(\"\\\\\")[2] + \"  이미지는 고양이 으로 추정됩니다.\")\n",
    "    cnt += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}