second.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Working on Fashion MNIST dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import  tensorflow as  tf\n",
    "import  tensorflow_datasets as tfds\n",
    "\n",
    "import math\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import tqdm \n",
    "import tqdm.auto\n",
    "tqdm.tqdm  = tqdm.auto.tqdm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "dataset,metadata = tfds.load('fashion_mnist',as_supervised=True,with_info=True)\n",
    "train_dataset,test_dataset = dataset['train'],dataset['test']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tfds.core.DatasetInfo(\n",
       "    name='fashion_mnist',\n",
       "    version=3.0.0,\n",
       "    description='Fashion-MNIST is a dataset of Zalando's article images consisting of a training set of 60,000 examples and a test set of 10,000 examples. Each example is a 28x28 grayscale image, associated with a label from 10 classes.',\n",
       "    homepage='https://github.com/zalandoresearch/fashion-mnist',\n",
       "    features=FeaturesDict({\n",
       "        'image': Image(shape=(28, 28, 1), dtype=tf.uint8),\n",
       "        'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=10),\n",
       "    }),\n",
       "    total_num_examples=70000,\n",
       "    splits={\n",
       "        'test': 10000,\n",
       "        'train': 60000,\n",
       "    },\n",
       "    supervised_keys=('image', 'label'),\n",
       "    citation=\"\"\"@article{DBLP:journals/corr/abs-1708-07747,\n",
       "      author    = {Han Xiao and\n",
       "                   Kashif Rasul and\n",
       "                   Roland Vollgraf},\n",
       "      title     = {Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning\n",
       "                   Algorithms},\n",
       "      journal   = {CoRR},\n",
       "      volume    = {abs/1708.07747},\n",
       "      year      = {2017},\n",
       "      url       = {http://arxiv.org/abs/1708.07747},\n",
       "      archivePrefix = {arXiv},\n",
       "      eprint    = {1708.07747},\n",
       "      timestamp = {Mon, 13 Aug 2018 16:47:27 +0200},\n",
       "      biburl    = {https://dblp.org/rec/bib/journals/corr/abs-1708-07747},\n",
       "      bibsource = {dblp computer science bibliography, https://dblp.org}\n",
       "    }\"\"\",\n",
       "    redistribution_info=,\n",
       ")"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metadata"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "60000\n",
      "10000\n"
     ]
    }
   ],
   "source": [
    "#  getting number of test and train examples\n",
    "print(metadata.splits['train'].num_examples)\n",
    "print(metadata.splits['test'].num_examples)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The value of each pixel in the image is integer from 0 to 255.We have to convert it into normalize value i.e b/w 0 and 1."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def normalize(images,labels):\n",
    "    images = tf.cast(images,tf.float32)\n",
    "    images /= 255\n",
    "    return images,labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_dataset =  train_dataset.map(normalize)\n",
    "test_dataset =  test_dataset.map(normalize)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we  just take 1 image  and reshape it to 28*28 and plot  it"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x27839de0828>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for image,label in train_dataset.take(1):\n",
    "    break\n",
    "image = image.numpy().reshape((28,28))\n",
    "plt.figure()\n",
    "plt.imshow(image)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlAAAAIzCAYAAADPgiexAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAgAElEQVR4nOy9d7gkV3X2u3bnk/PMmZw0QTOjnIWwQEKAsbAAAwYDljEgDDjwYGzJ/vyZy73YF/va5jP+bEA2+WKQLbAlsIyQBkkglPNoNJImaHI8OfTpVL2/P6p7v2/PqZ7TfVL3Ga3f88wz61RXV1i1a3fVevda21hrRVEURVEURamcUK0PQFEURVEUZaGhD1CKoiiKoihVog9QiqIoiqIoVaIPUIqiKIqiKFWiD1CKoiiKoihVog9QiqIoiqIoVTKjByhjzJuNMS8ZY3YbY26ZrYNSKkP9X1vU/7VDfV9b1P+1Q31fP5jp1oEyxoRF5GURuU5EDonI4yLyXmvtC7N3eEo51P+1Rf1fO9T3tUX9XzvU9/VFZAbfvVREdltr94qIGGO+JyI3iEjZCxkzcZuQphnssjJMFKeVa42LiMjSJX2B6ybzcXzP4GEyYbLOztqws08ebxcRkehYzi2zqfQMj7gyRmWwz1rbU/izbv3P2LZGERHJR4xbFsrRQzs/vxsJXG5D+CAf9f+PnByfxaOcmpSMS8am+Qir8n8tfH8msRDbfpH0cuw3hG5D8jE08lCG2niCGn8eZvxAck6OrxIWov/TK/39ReJwei5FP3mmTPDAmsDFxe3kPAg38VcmZniUU3NK37MgfH8mcUrbL2EmD1DLROQg/X1IRC473RcS0iSXmWtnsMvKiPT0OrvvujUiIvLZP/9a4LpPJVc7O0q928b4UWcfy7U7+ytfuEFERBb/YsAt83a8NLMDrpB77e376c+69T+Tfu0lIiIy0YWm1tAHPxv6gbAsKFPf5jXgg7Fe/2F20ZcfxQp5b3YO9jQ8areduqgq/9fC92cSC7HtS8hvq7v/8BIcRx/acnIF7oOmfbg/kptTzrZJLN/wscfm5DArYSH6f9ct/iEtXouX5+O7u51to8EPUKEUrhE/S3Wt8/v8oZFGt2ztbzwzK8d6Ok7pexaE788kTmn7JczkASroMX1SizTG3CQiN4mIJKRx0heUaaP+ry1T+l99P2do268t6v/aob6vI2byAHVIRFbQ38tF5MipK1lrbxWRW0VEWk3nrE68l3rrpc4e2IRTMRSQiA/6u/z87/2mW3b0Kqz797+OyNSKyJCz3/HtT2H5Nkh04VX+//tv6HLL7NuvdHb7boRUWr73SGUnMj1q7v9K+O6XviAiIrePbnHL9qQQDR3ONjh7wos6O0R9QjIXc/Ytyx4QEZE/M7/tli36x4dm8YgrZkr/19r3ZzB12/aTb0cw4P7//SUREQmbJ90yz6J/CJupc3j6PEjVr1t8k4iIpHe1umVrb354+gc7ferW/8w91/+tiIhE6ZFj+3pEoLIWvwM8ZIPtznCSlvs/LO102T4gr5m1462QBeH7VwszycJ7XETWG2PWGGNiIvIeEblzdg5LqQD1f21R/9cO9X1tUf/XDvV9HTHtCJS1NmeM+V0RuVtEwiLyNWvtjlk7MuW0qP9ri/q/dqjva4v6v3ao7+uLmUh4Yq29S0TumqVjqYiB377C2cnFiM02HkOUMjYOO19IoPMSCLat+i+EZf/hy9c52yaxfNVG2OPLITOFPH/bLQewD49ixAObYWc/iGPt/Prsh9pr4f9KOPLHkDSXRPxBltc34x5vaYWPusNTZ4cczY05uzPsZ00++1EMrHzgHxsmfWc+qFf/vxqoJ98f/RTa+3Of/idn78n6fcgxD2NQNkcxQHwoz3IettdC0t5ekq8fu/TrIiLSeDmWXbT3Y87u/sr8yXn15H9m4m0Y1hE1D4qIyN+ceL1bFqdEoXQ++OcvRNl5zWEM3xjI+n3VLYvvdcvsFec52zz87HQPuyrq1fevRrQSuaIoiqIoSpXoA5SiKIqiKEqVzEjCmy/CZ61xdrYJse72PVxEiCS1+ORMT142sgaST24zEhoiKWyDizdyLZCSIpDFz1FnU9p2wx5dSVLVuZucnX/uxUnbOJPofRPKlKStn9Fy3IPPB6hu0+4s7IzAkWGqIBg1KHaatH5I/YPtyGx6QK6ajcOeP0LUYKaoYRU6f7OzX3k76pGlV2ScbcLwVShC8nXOb3+WCv+JhzZpovieicC2OaqDk6PGXzA7F424RWPJhLMzw1SUdgLnuPFm1MrJpyBhLWR2f+FyZ9/6q19x9pNpXBdPfH8M5SHhpSyGBrSHgt9fs5SB+mJ6ibMPZv1td4Uhad/8h/8K+8J3O3vDRx+v4CzOPI5cBZ+OFsZvsAzXHYXvuEByiArS5akgXdRMvj+jZI8vR/tvnt4hKwsYjUApiqIoiqJUyYKIQE2s65pyHS9Gb8oUJAoFvOBbGrEZH8HKxoOdo3HJXPG/+N08BRH4JYWjVdFR2ONrUbul4bmAEziD+M6G7zl7uPBiFzV4b2sJ8TQ5cJ5HNeLCdBF5+WhhLpd1cUQ7qono1BRTOI8yx2guOUdERPo/gzfm83tQBLfXIrI3nMGbLw96zXHDdJ/nA9fNU2Pl5SmqxzWexYBlW1g/m8d7Vy5O9XMWwe5tQeO//qnDzv7O//0rzg6sk2b4Pq6v8jXjP17r7D3nftnZOzKYzuOYhzhEsc5Qfw7LnqLtvbEBNZ5C1Ma/M7rI2a1hROzaQ370iiNa58dRAuiVt/6zs69Y+WvYxi/vKXdKZxxeK+6t8YL/eeC4F1iHUiSdjwYuZ8KF+yhKbXRsGe43jUC9+tAIlKIoiqIoSpXoA5SiKIqiKEqVLAgJjyeTtSXSGUL8YZ7APDJ5ADhLeaEs13CiHUWDZUCO+pq8/wGH3EvkPPpejkocBYxFPGNZRLWddmZ82SFKMpLH0tHkaZz8dcrIeSmZLFGNvguTtbbcNqfT50wfI2LC/rHbHCSFE7+LGkI3fswv7XLvybPdsu39GEQ8kUFj5RnhEzFIZw1Rmnm+ILXly8wub2l5OITrk6VtZ3LoIvJ5f/047SOVxjHFSc47MNDh7K8lUQ9t1cchJx29cZ2IiHT+DgZe5/bTPKkszdbo/sm+8WJn/3DLF539X8lOZ3dR5/JLCZxLSHx/TNhht2wvXCcHqB2cpCSLZdFBZ6+P4rttBX8k8xjEf9CDlP1fNKD/rq3fdvY7r/09Z0e2IfnijCSPNp2yfttM5iFDt0UwiD9Ng8U9shtp0DmTKsh8SU5YSgSuqlRDDWX7cAf6KW8Q913pMZX/vkagFEVRFEVRqkQfoBRFURRFUapkQUh4Ex0I5Wcp1SFPkhspRCWYqSKCwepGWcmtKCGyKsL1ozItFELuodoikTI7OkOILOkNXJ4NeEbPk9OzFDoPlblYCboYQ/n4pM9HVmMbLVMfam2wItab3KgiSZzz/X0bRURk1/EetywWg8yTJ3kiFML3sh7uj5FRZGiFwmVuigKeF+x7lvbyXDcqVDwOkvjGII9k09Sd0LFmUlg+0IcrdNFZfobhRLhM/pI9/fHPB/vex/XIqD5QGU3x6QzWbzL+teskiS9K90MxS0xEZHEYmXxRA5s9cKRQg+6I10brYtsJAwm1MQRpdd8HsZWzMAPSGUkoCf+2mslSHNd+agxBbh2mtGv2aT7gB8Lj4SKTRxQolVCBbMdTJOWu9GVr7vd42EEkTPUEadjB+EH0N4mTuFgTy/x78+z/+Urw4YU5sztwFf94yn+kKIqiKIqiBKEPUIqiKIqiKFWyICQ8znTINVMhQD76gGKXIsiaY0muRPrzgsOHvA7LdUH11mJj2EZyCckiiKhLuhUbgchy5pBf1BG4fLSQAdMeQjidM+w6w3ASz0J/xONsmtPHydMd9VVwsSwBoequp5FRlftN//yzE2hklqSwME23kqRpU1injsSpMGlBoisnHkejWDdE4fB0CrKcofTWopyYocy7aCOuXy5LU2PEgqvLhmj6mLFsvLDvcjp67d/vfuM8TImSJUmxqUx79qitDhXskx46qqiBJNtCEhJPdVTMHjuV4n44Y4xlO77QyTyWv/+cx5z9iExdMHIhEx2DE9oKHfB4DvcKy3NRqnS8qQEFSZM0TGCAiqAWs1rD3FzP7JEZcwff21RM+eUvX+rsX7kIGaOrGvpEpPw0O81UcLY3MuTs2DnUx5EWV7zHvvwnl019fKeh9j2UoiiKoijKAkMfoBRFURRFUapkQUh4LL/lWmhIPEkD2SbYiSEqdFZQI8oV4CwJwfLivA1epRC/DWcoG4Ds7AZk0IT3QnvMUnZe2eJdZyDFIpglsh3NTfVsptvZx7Ltzn5/Kwoq7sxMDr1yZcXcYkghC4FDf4LskjVvQhZIMZuuoQXhaE5WyWSCJTLOTGGKXzWh4DQSQ9JfmmS5EGW08P6zGb+7iFJmYHHZqdtLJCAhcfbgRB8E7OG0f3+M/+pit2zJ31FWTB3Ma/jG1u3OTpGEx/JbknT9oAKwYZbZiHLfY1kuQ4VjQwGfs4zRE0LfkyK5+Iqm3c5+RFCk9UwkOkr+L049ydfEcFYlfPfE2BpnX936krOPUp80VihaeoyKl76aCiTPGO5Mytzbrz3vRWcfmcD8sYeS/nU4noSkenIQGXa2TIbyki4Uol3Zgt/alQ0D/veomO100AiUoiiKoihKlegDlKIoiqIoSpXUrYQXbm0NXt6B7JdcAzJXOOONs51sqJCJlJ9+ppblLKHCZsLp4O01NOL4UjFIeNFROqaJCTnTMIeOBy4vSh2cSdceQrN7bHyds//70GZnf+iCf3d2RnBxkYFEMtNo3TZjh21plOzlF4mISPvVx9zyLW1HnX3XPv/8w1QAk6U1j7LcOGTNAl0+i3ciRMyDsxiznOFH2Xt5KsxZsr1CFiDLdvx5hGS7iSRl8nHxuxasc6Lfv8c/9eEfumXfHLje2R3feDjwuOeT82K4V/tIdWihbLr2COy9Weq3CuvwXI6hchV/SwrKYp0YdTNFOY+z9NZHg2W7URrecFUCMsYXgvd+xhAfhA+c2E++GMxCQj47gcy7Hz5xgbOvfD0kz0MTGG5xYNy397Vj2EFiYIFkANcDZTLvmBcHIOef1d7n7EUJP2OS5dhjBzEfZUkWfpwyfdPoh8YSkF6L8+sOXb/FLWv9LuZRtdnKhoVM+ctjjPmaiFwvIiestVsLyzpF5DYRWS0i+0Tk3dbaM3swT43YYZ+QPjkqMSHdXf0/bxT979FDnPp/ftC2X1vU/7VF+576pxIJ7xsi8uZTlt0iItustetFZFvhb2UOWCqr5AK56tTF6v95Qv1fO9T3tUX9X1vU//XPlBEoa+3PjDGrT1l8g4i8rmB/U0TuF5GbZ/G4RJYhlBel+cI404ei3iXZEFxIszhXUYQkvtI59IIz8li240KaxWmtvDgWNpzExi/oPeTsBw8hSyCcoX220IxtKWRcBdFhemTCjp+6eO79XyVe/0Dg8lBAFh4XJLyoEVlXzzYvD9wGF09LhCZnNDUdnLsJqWbL/5lOkVfeXSjENwCZ5z5vvbNTE3642SNZjAtP5knCEyo0KvHgrDlXYJOWldw/XrDfeP47G3B7lMh64eCsWEvbCEcoY5LmywsXzu3vfvIrbtkn/+guZ//70K9IanxA7MPPi4y6gqPz2vbbQhgmcNxDO2ihvmec+pBEQMYdS3Is53llqjBye2fCBbE2XE4GJFi2ag4lTrNmeRZK38O0HuAMRd8H/emmwHXb25POXnIffL7+TRiOcI9A4pnI+qIgF1HteGnyfHuzxUL0/+kw/JtKTdjEEeG8egnk012ji5y9JObL0E+lVrhlrYvGnN3WgN/R5c0opLm++YSz+zPI4GspFN7sOw/H1PrdCk+EmO4g8sXW2qMiIoX/F5Vb0RhzkzHmCWPME1mZu8b2KkP9X1sq8j/73hud1BEq00Pbfm1R/9eWqvse9f3cMeejb621t4rIrSIiraaz4hF3yTWovxGZwNd6OzH1xfEVGBDY8wweabNNp38u5IhSSY2nMlGnIPiYmOu7nnX2gxY1VzjqlV9Obf7kydPvaIZM1/8zIW1pepZCDaKTOQx8bQ0h8rEztczZz768EhtBYEbGLQYCNpni4D5sY+nP8SZST7Dv2xqX2k1f9AdCvvf2e906/3kCg1dPHvHbfLgpuG5QCeHgS8mRn+LrUcnE51wTJ4HIEL8RhmjKGEM9RHGPXIOKazx5OY5M8ahOuq9onbNX+4N4X96NGjzfegVTK/Rdn5dcX16yz4sIZt2omNlu+xzVCVEkKUkJEly3qRiN4qhTuQhUUP0okdLoVbSwnO8Hjn61h+DbbAVRqrmmFn1Pww4MDI8WBi2HKOqap2vI0cLWHzzl7Pa/xgBi/m5zzH8Q6QmjMUYfQ92i2nsczJvvTZkfymLo2vC9H1xzadf/iz7wlyL3Obs5SglZhZppcUrYuHzpPmePZhFlHcogasy1vk6kEYHa2uQrRS2bg5WTSpluBOq4MWaJiEjh/xNTrK/MLur/2qL+rx3q+9qi/q8t6v86YroPUHeKyI0F+0YRuWN2DkepEPV/bVH/1w71fW1R/9cW9X8dUUkZg++KP2it2xhzSEQ+IyKfF5F/M8Z8SEQOiMi7ZvvAss0IhceGEbaL0WjwzCIsj44hSjm2lMLoA35gtcy4zNL6ESztlZnWxStIcRwh9+LY+CXxw86O9+M4UlSyIt9A01+UOawi2+2jMignpaBjn1vw+Zz7fyb8Zd/5zv6jLn9G7RjVbQob+OVops3ZzbuCZ4pnSSPh6u/QAOhHnpvJ4Z6Wov/zkpeZtH87kZL88364/xcj0CfXNaPWyTOptSIiYmjMq5cpM3CcaivZMstd9LzMVC8s21lq/Ib8zcuLg8d5YDvXeCp3A+XzdPPR+u0xfxBviBIswt/DjdK389uSenmv5MfGReax7Ye3bKS/nnFWmqS6qOFpXdC2D5NUnQj7fVXK4n5PUP2ocMncUVPPCVLc57Es7plWA5ljTRSDcQ96wckpfG7ejpcC1ymyEPue3GFIeKOF6UJaIvDRMEk9LOFx3Z89WdR+aqLvLmrwpbtdmV63LJ/EQPTZZrb6njmFs0xoaIaJ+HaoAf72RjD8Ri49x5nvveYXzv7unoucvawN9cuWNfiy9et7XnbLOKGoOYz2fiDd5exnhpGYdGAE1/XJ6GoREVndDglvnI6/0mmkKsnCe2+Zj66taA/KjDjHYEzIvfb256y1Xy38qf6fB4r+f9RukxE7wGmC6v85pvtD73P2gY/9sbb9eUb7ntqifU/9o1O5KIqiKIqiVEndzoHhkZrD2XEHB5Gdt3E9wrVpWRK4nWINqVyCMmhyZWavLyPb8f6LkftIEiE+lhvXRDHSP0qZQ1kslnwM689dBaPa8e0HUfzts2/fISIiXpkclWVxFNENlamez1NXtIX8MPCOTD3lvFTHvQ+d5+zeTTQGtMWXd0JUWynUQOeJ5JKSyLmXC556paRBFxeVkf4ilFmXI9mQs+kSzb6cwVO5MNZUoIdT1l6qcJPHUbZFuh7CNDd950MqmU9GNrUHLueMrFEK8fNbaH4O30mLGWQs/bE8OJznKaKCBwcMb4GM0bxjdo+v3niokOHbEIbUwxIe9yvMgIfOui0Cn+YKRQV/2Hcerz0bh3pmkOesXt/2SBr1Xn+hs8f/GPLcd5+/2NlLurE8mUW26WN9q0RE5KxWDHlopWszkkPneDKF63d4FHL3+g5kvHfG/LIy5zajbuN/XfRLOP7Ht596doFoBEpRFEVRFKVK9AFKURRFURSlSupWwmNyjXjOmxhFpsmeTI+zuztxKlzkMpz2JZB0K2W+lZHqSqZ1IUJZLC9KgQ0HEGrc//HOSd8REVl6H9bZ8x5M35FtpuksAr+5sFnzA2Qajd3gZ0eEBeHyZB5h3RVRhMCjo8H+b6SpE4rFAj9/9BpaY0IWEpdcgkyS/ZQZIuN+uwg1wT9cqLJEtqNpXfK5Mu9BTkZjaY0+J5UnEqXw+0EUqCXXy0SPfw2jg9g3JaaJjQVfv3wTTTVD087sGfSzZZquh2wnt+NatuytzfvdwKbguzJBmXItlLFziOoDhqYop1iuYCZP38LF/1iiCxc6rvYwqtq38QWibTeZ4KKFQ2dhP82Ba5w53DO4VUREemIYSzEQQtvO2ODr7FFcgaf/iBSyLfcOIcur49Uo4VWQrRZZ4svvL3wGxZHP3gi57MAjq5yd2EDZecTSZvx+Fn1+3+4NbtniLnweC1P/VaYK9kAa194V2yS1fmALrnXH44GbmIRGoBRFURRFUapEH6AURVEURVGqpG4lPC9GoW6eGowyh9b1YlT9S9ciC2/V92mOqcTkZ0TjBUwxXyGFKXnEjEFqOP+8vc7+cTJ+6ldERCTXTUXbImf2c2tk25POLs4E3xKCv/hycjG0UAW1y4rzW92zE/MMrpenyq1el5zVhHb7/PHJ2aOpfkq3o3npOIOtnBTH2EIGnYly5l2IbHwxk4bEmu+FLJSjrL7iHHj51uC5wsJlsvAM2d449tPd6Bch/MTKn7plX06j9tCSB/qdPXelUiczsSxY/uK58JoN7vNxi4bL7TmIWBmJr9y3WM7LFiQnXjacx3GcTdLKcS948uqJpZUVCDwT2DXsD/FYvAgSUWkmZcOk74iI7E4tdnZDGG19wvOzwhKR4PaxIOC560zA71C5ApL8vTLrHPuDK5392vf7vwEvPLPCLUv9Bfq65rPwvcEOVA72qH9a3oL03KaYfx2iJNs1R3Ft+pKQ5waOY7hM4iAy+TLD3c4e2eRfw+HVyMo8eSWua8c3pCLO7F9yRVEURVGUOUAfoBRFURRFUaqkbiW8kuwefswjSWFFE0J8u4cRKoyOYl6cbIsfogt5nG0XvJ+yx0L7DxfUDW8RCnQ9vRPhw1vpexPLEJoMJ0iqylSpGy5gvjXih03f0rTfLRvluQXJueXmK+QMpVDhmX/RPbHglRcAdx+C/Li8HW34SMhvmGNDaE+Gimo2NpO0RsUzubAlF9K0KX8dM0KftyFMHWkIliKaWrCfkrnwChKiIRmEI/vhEI61oxGS7fo2SJZrGlAIL1y4Ee8d3oLtNeHcc1PM1TZXRDvSgcuTJRlx8HOK2ie31aK0ylldLCHly2QLcaZeljqopsLcbV6Z7zFD+eD7I9QZfG5nImnP912YOnwuqnnf8Nm8trNu24WCjx/c9LCz90/4fRlnh6EMcJ1TlHdZfrNVyLm2TIbtay+A/Tr0Zff9wJ/Tbsle+H7oLNw/Y0jCE5NFe47F0CednEBW3JJGX4aNh/H5i4OLnD3Y34LtJSlDOUKS7Vk4lnjn5MzttWuPT1o2FRqBUhRFURRFqRJ9gFIURVEURamSupPwTNQPPbPMlmeZLYGw495RFDTrfI4KDlKWW3FOPZ7PrjSFqTrCGf+7yRWQ57ofwQHuaKesqitoTjGDjIGSLMAznC+9crWIiLzv3H93y17MIlzO2XnhdLBfolQUcMz6IfjWV1KB6y4EYpTFk6CQdCLq2y29EAaGk8GZQteve97Z6xKYT+8jbQedfSDnZ7mNkvT0wPhGZx/NQIbuiCJra2kUoXi+Pu3h5KTjGPIgubFU0hXC9njOsQ8/+pvOvnKNn716fgsK7L20dBM2fujwpP3NB+2tOE+eX66FJEqPfJqkbK6EyZLt91Xl5lxjqY7tkhELtJ/GsH8sHmVj9nuUwWRxX3EhTb6tutvHAo/lTCQoW46z6o6lWukT+C51DD5NnE1+zPvXYmMzpJ5HJPja1h0BmXOh8zc7+5V3+H1BaPPopPVERFLjkISLQwNERFashiR/dSfmpn26wZ+H8PhW9DFXrt3j7MPjqGC5shn9XZZ+7FluLc5Z10ZFZDc3Y3/ZZbhPnhtd5uyhNO7N4TQy7nIFebcxin2sbMJxoEc6PRqBUhRFURRFqZK6i0CFGgpPiSXTrcCO0zQXJ8fwptAxgCdsLz55IGc5ODJVbioX3kY+6v+Ro8GijSex7/RjGPiWvhRve94JPAnb8Omne1iImAiaks3hre1EX+ukdUN0cVtDiCR1PDs0aV2R0kG4RXJN2N8CeQd0pLI49hzVKTt5zH9bix3BGXVchAHYJ2mg5PcfvNTZix9BA/3rK+Db5hX+wMv3rEWdrCubdjn76iYM0t6bRY2UJ8bXOPt4GtevOCVG8U1cRGQRTZPBA547InhT3BTHm2L7fbgPdnX4dXouaUWCQXgX3v1qVbFoaTPqBo2WvLlPPei7MTR5kHa+XHZEyTawTrjMgN3iwPD2EEVRPBowS/dVGxVVO+nB5to6wbGGM4fFjZPPsFjLSUTkHT24L74uGNVsG4Jb3kRBzlgW46HjiwLXrVdO/C5qNQ2di+hLcUat1h+hj+FcBbuSasZ14Per/34oLk8cQP2slv1+v27fgTplj0ZWO7uzFf3DVW3ok1hteCW9aNLyAxn0U5yA1J/Fs8Bbu5919jjVSTucwbRZxUhXWwQR5oOp4CnZTodGoBRFURRFUapEH6AURVEURVGqpO4kPIkWZqTnchVlHvNG+xG26+1D6DzdjbAdZj2g6SzKyXZshoPD9UEyHy+LD8DOUM2XyDhOIkfjgoMnfll42ArkzyK9lBTwh6+81dn5518M3EQ7yXzRwgbzZa7PQmCgD2Hyy3oPOPuV3b50lmuEL3ubIEOwhNe2CrVoTnZjcCTXSfMe9UPW337mWrfsnxdf7exLzsGgzs8u/6Gz39iNGeYfo4GX94769ZoiNJj6cBqDQYeywQPeX7fkZWcPbqY6SAf8kPnAUtzH+eTkgerzzepmTCGTJB0jSh3EvhyOczwPaaA3DPlvKO/7zqtA+mOZz6MxCyGyBzx/eEBXCMcXpmPaTckZ7dRnJunW7E2gPZ3pEl6mIG82h9F/8CDyqxuQcMESHtdHY0kpHvLtPamFJduZcEjCzb4U33T9Mbd8DUmcHTFfyspejYbz4OOok9V0AJ124iTWyaJLkmwz2vPu3/J/xz9z1Q/cskMZSGTdEew7STLbqIdB5zwkoHuudzUAACAASURBVLgOX0u+Z5bH0GdxYgtvY2UM901xyqUjWfRfcZqGKbxhHU7sNOXopoxAGWNWGGPuM8bsNMbsMMb8QWF5pzHmHmPMrsL/HVNtS6melE3Kk/YBecjeLSKyRf0/f7Dvx2VY1Pfzi7b92qL+rx3a9ywMKpHwciLyh9bas0XkchH5hDFms4jcIiLbrLXrRWRb4W9lljFiZL2cK1eaN4mI7BT1/7zBvm+UVhH1/byibb+2qP9rh/Y9C4MpJTxr7VEROVqwR40xO0VkmYjcICKvK6z2TRG5X0RunukBmZifJUGJPhJmOY9C6vHDyFYKJzGqP9MMKaGoorHAxNkFLBUWazyJiKRbWeajLxfkPy4Rz1l/DYPYYC6OkOA4y4NVjDyLmwaJizufvPgd2Zz5f9rY4MzCrg4/E3HCInQeN7i4e36B0PlqQbbWz6jM0+YofNoY8ttH/zm49kv/e5rHPAXse+PLLbPj+wmqdUIZVd3P+xLBhj/b4Za9NAS5IBKDH0IkozU0YButnXBcqtP3M89OHh6E357+xQZnv6X7953d3o3s0b/Zivpdn1u0XUREfpHCvh8YQ5h/7xgyZCJ00xzOYf+ta5EFNnTYX75rHOdo0/i8Vm0/R7VoeNqUNVHImZ89eb6zz25Au2VJrSgfhKusOxeUdcrEqf9iuWLbCOr6fHbRo85Oemgf+anSkkv2s0D6njJsP+TXA3pP72NuGcs+Q/lgP/PV6ozgXljT6Nc8emKQ5iGR6qf/qITZ7HsyXQ1y+H1bRUTkwyv+yy1PUY2xaGGsSyP1R7/xFrQhrm82bpHJmKIpg5ooA7UofR7Molbj2Qmq20TSGteXY8n0WsqG3JP1r8NXB5BFyBl0BzLYD8t8fD68T69w7ftIg+yiWnip1RTYm4mExxhjVovIBSLyqIgsLjxcFR+yFpYwvDCJifq/JuT9pHr1fe3Qtl9b1P81Qvue+qXiQeTGmGYR+b6IfNJaO2JMZW8yxpibROQmEZGENE6xtlKOnM2JiKwTkQ+o/+eXnM3JhIyLaNuvCdr2a4v6v3bMRt8TbdFhUnNFRQ9Qxpio+A9P37HWFofVHzfGLLHWHjXGLBGRE0HftdbeKiK3ioi0ms6pY9nxyRIep3It70KI33sYo/cz3bhBi8UuRSDLcdZWOQkt1zD5eyLlM/KCaNyPMGA+TqHz45RxE6kugyxv8/KcPCwiMjDn/p8uZYr//eoKX/YZzSM02xZC2LfMpPHSRFPfRGnme6+wn7GNmUnfmQuKvo9KTFI2OSu+X30H5K0fdm51dmJ9pLBPtI+TIyjMmk3hpmhqR0g9FkGou28Y67c2+aHsFSuQfZJdRoU7BxG+7mxBGJ1nNv/wTz7k7I5lfubfX27+T7fs5q6dzt7ThsKE3xlCoc8nkmudfc4ihPG3F/5f1oB7GmVDfWrR9jnzLUyZtFEDCeDBE8jSuXQNshkzs1wZJiw0fUwxA5U+740gG/PhEyiA2tj7tLNDJOGFqpQTF0TfU4bGx/3fhNZLIemwvPPQxNpJ3xERkaOQalm6WhnzJbzbBi50y5bOkYQnMrt9T+//ekhERO54Fhm5e96N/uSNF/l343gOv6mtUfhqcQzZpd1RZNBFkeYux3LIoCu2W86wW9GAfuiHQxc4+2eHcS/l7ocUt+QhmnbokedERGT/ZyHhvfiRf3L2M2maioekSZ5GqeS+Lhzfg0NnuWVXdO929vdXV1aeuZIsPCMiXxWRndbav6OP7hSRGwv2jSJyR0V7VKrCWisvyBPSJC0ipYK7+n+OYd/HJMEfqe/nAW37tUX9Xzu071kYVPK69BoR+YCIXGOMeabw7y0i8nkRuc4Ys0tEriv8rcwyw9Ivx+SADPrv5pvV//MH+35cRkR9P79o268t6v/aoX3PwqCSLLwHpfyMcteWWT5tbMQPk4doIu08RdOODUNeWL0dckD/61YEbs+L+YfOsh1n1eUSODWW8KLjLOHR8RW2YzxyCcn74ZOURRTBF7NUYLOaLLx20y1vkHeKiMi99vYXrLUX08ez7v/Z5orC3Gs8uxRLIbn2yTOm++sEZ/XlC6FXE537+QTZ94/abTJiB86nj6ft+9jdTzg7etMWZ296h5/uEQ/DJ73tCJ3vG+lx9qFjGNfQ1IpQe1MDQtmxwmz0E1ncQJkcfB8ieSqZQqg9koBskafMv9ExPyvoE4/8hlu2cjEK2H1mLYpxXtoEWWttBOtsT2DurJ6YH6L/2V9f7pa1yiPOrlXbZ1miHINJZPqyjMYU5QOWo5mS4pnUxXL3kOVMqcJddIQq8a6I0Lx9qeCyvDSioaRY4FQs9L6n92G/ffX/DmRtluQeG11DayOjKzYYClw/VmgX3lMovjhXzFXfE74PMvuG+7B8X/HzDkS7Tm5Elu7ONoy1yDbTDyJntOfQn8SG/L4nvguFO+8+zPOioo/rOV2a2yms+Xf0JZfs+xi28SjkQTOIeyJ3FPuXgPFjJob+8ktpyHndvmw9JTqVi6IoiqIoSpXoA5SiKIqiKEqV1N1ceJnlvjTBEl6uEaG3N69G1s/zhyDj5BIrnc1z8kQm/LBiOdnMUljPUCZZtil4eTECznIfZ5LlDh129rImhCxfNMtofTq+NX5Rttwr+4MPcIGzPeVLqxujkDmyliSSSHByzp/ue7uzv7/+TmeHis/8Q5VlSdQVoULoO4/zX/5XaJhP/r7fhovFR0VENnVi7O6nr77b2bvSvc7uDGP9dTEk5bSEJktHnXRjsQh6f3K1s0fzkIjGPIT0ixlMT4xg3Yf2Qwb54CvI2Iu2YN/LuiFr/8Hqbc6+7/B6ERHp+S5ku5Iwe5nMzrmGC+6VY3ycM7VwPU96TbS8crmskmKbRQmpP499tFHxwrGxxKTviIiM5lkq5I5w7mXwWhI56GfNcbFRviZcOHEf+SUOlagki6s4f1r0DJ5E0BscxB+PwOZk6TKJ04EED9CYPjxfaufzWD616C6B/Yml7L3poBEoRVEURVGUKtEHKEVRFEVRlCqpOwkvH/Of6UrmpWtHWP+nh5AZsEgQzlu8DRl545tQ3T6c8cPUXgzPipUUxrRlPJNuDRe2h2Xtu4OzbM5vOeTsZ7pQMLGhD+c2tnWxiIgkFrqEV0Z6uf2gX3Tupq0vu2VJizB684vBAeGXH1nt7NB6XLui/Jc4MbXMUnfkAwLNj2135lnvn/xx35aNzr7lLb/t7PGVwUFrkzt92+YEM27jNAVVSRZSnCL6ySX+dW08RgVnKSHJLMbGswbXdeAxyNd/u+99zu46OTl8HmpGxlR+tDZaCWfHJamir0fzPXoZztTCeQ+RhLcs4jsvI9W1VS6eGQ7IRvXywVJdPssFZ/G9JF1oLtJ6ppM77P8m8HxnGxNHnb2hARlaT8pSfJHCCuz/bKEtJPprIy0r9YdGoBRFURRFUaqk7iJQ6Tb/bW2im96mGvDEP3QUA7N5FkUehB0PiObM1ok2TPF5ZAkG9zaHMaCX3/wnuvEW2HDyzHibMWG8ZdscwhlHXyhcJQTgJEkRqt6HUX+Fad0duNhFrzp3VjRscMHj7UCNlKU7angg80Stok5MK83yPmJRW+mEh8H6n77sJ85+IbPY2aMUHTrm+VNb8AzzPLUFL+d6TyO0DR7EXBxozoPcX8igv7n5sh/TsWJqnpTFIOrmMEf9Xh3vz7vG8UuxJo4++fz4QWd/lyJQ48vRP3kUjez3/Oho59MIy57Zw/CVqXh13EGKoiiKoiiziD5AKYqiKIqiVEndSXjxYV+a6TsXz3aRMUheW177irNRxeMUalhLhkvH/39Pv9HZHUM4jvFlOL7EMT/UfmYIeZPpedL/v/E9GFS8mK5P9DjK7rMo13wE8gZP/dId9gfpNh1MiqLMBTy4eBPVCloUxgD3T7QflGDGyiw/PR4lVoTN6d9rD+Swj5WRZvpkhGwsX0K9/I50n7OfLBkEceZyYBTTHl2+DMM7NkSbglaX+AD6pwtJ8nu+MA7D1Kg+mVJ/aARKURRFURSlSvQBSlEURVEUpUrqTsJr2O7XTuqNY2qWxHFkxbxs1zt7mZwM3kidhFhjzyP7pW0vziE6gUyc8KCfdTTbJe/nG+sFZ8W1fcefpuOst3zQLfOSaHYbXn488Hvxu7B8489/E/spXNo1jz037WNVlNNx2zuvcfa3eiHzRO990tmhFtQW2vsvmM4mM4gMuvYlvqTGtZfYbooH149LZpB5l89Pfscd70O/wlPmrPsgMjbzqZSzvddd6OzIMGfhvQrSOkVk9I4lzn7z0d93ts3jWmyQJ5y94quoL/jatZ9ytvH89TceeGFOjlNZeGgESlEURVEUpUr0AUpRFEVRFKVKjJ1HucsYc1L85Lm+qdZd4HTL3JzjKmttz3S/rP6fEbPh+/0yd22jnqhX/2vbnz7q/8qo17avfc/0Kev/eX2AEhExxjxhrb14Xnc6z9TzOdbzsc0W9XyO9Xxss0W9nmO9HtdsUs/nWM/HNlvU8znW87HNFvN9jirhKYqiKIqiVIk+QCmKoiiKolRJLR6gbq3BPuebej7Hej622aKez7Gej222qNdzrNfjmk3q+Rzr+dhmi3o+x3o+ttliXs9x3sdAKYqiKIqiLHRUwlMURVEURamSeX2AMsa82RjzkjFmtzHmlvnc91xhjFlhjLnPGLPTGLPDGPMHheWdxph7jDG7Cv93TLWtOT5O9X1tj1X9X9tjVf/X7jjV97U9VvX/XGGtnZd/IhIWkT0islZEYiLyrIhsnq/9z+F5LRGRCwt2i4i8LCKbReSvReSWwvJbROSvaniM6vvaHqf6X/3/qvS/+l7b/pns//mMQF0qIruttXuttRkR+Z6I3DCP+58TrLVHrbVPFexREdkpIsvEP7dvFlb7poi8rTZHKCLq+1r6XkT9r/6fAxaI/9X3tUX9P4fM5wPUMhE5SH8fKiw7YzDGrBaRC0TkURFZbK09KuJfbBFZVLsjU9/X7shERP2v/p9j6tj/6vvaov6fQ+bzAcoELDtjUgCNMc0i8n0R+aS1dqTWx3MK6vvaov6vLer/2qG+ry3q/zlkPh+gDonICvp7uYgcmcf9zxnGmKj4F/E71tofFBYfN8YsKXy+RERO1Or4RH1fS9+LqP/V/3PEAvC/+r62qP/nkPl8gHpcRNYbY9YYY2Ii8h4RuXMe9z8nGGOMiHxVRHZaa/+OPrpTRG4s2DeKyB3zfWyE+r62qP9ri/q/dqjva4v6fy6Z55HzbxF/tPweEfkf87nvOTynq8QPiT4nIs8U/r1FRLpEZJuI7Cr831nj41Tfq//V/+p/9f2ryPfq/7n1v1YiVxRFURRFqRKtRK4oiqIoilIl+gClKIqiKIpSJfoApSiKoiiKUiX6AKUoiqIoilIl+gClKIqiKIpSJfoApSiKoiiKUiX6AKUoiqIoilIl+gClKIqiKIpSJfoApSiKoiiKUiX6AKUoiqIoilIl+gClKIqiKIpSJfoApSiKoiiKUiX6AKUoiqIoilIl+gClKIqiKIpSJfoApSiKoiiKUiUzeoAyxrzZGPOSMWa3MeaW2ToopTLU/7VF/V871Pe1Rf1fO9T39YOx1k7vi8aEReRlEblORA6JyOMi8l5r7Quzd3hKOdT/tUX9XzvU97VF/V871Pf1RWQG371URHZba/eKiBhjviciN4hI2QsZM3GbkKYZ7PLVzagM9llrewp/Lgj/Z5b4+4s0Z92y/FDU2aEcrWxg5ik26jXADsU8ERGJHcKDv02nZ+dgT0NKxiVj03SE1fl/1n1v6FCm+RJU0W7iMewmnanii7N7fAux7QeRXo5jaGhCuw0Z+Chs8s5OnmjE8v7xOT668tSz/004TH+g3WW64iIikkd3I4nj1Ibz8LP1vAp25G8704trYuO4bolD1A8Z7sCwbUv7rJRT+p668v10Mdw/sB2haxmCD23Et70YLSMXW/oa/47YMhpbbNi/Dvlo8Aqh4aSzR+0At/0SZvIAtUxEDtLfh0TkslNXMsbcJCI3iYgkpFEuM9fOYJdS6mxmDn9Epk2Irqotc+NUcdz32tv305+18X+VHPzIlSIi0vGaY25Z6j8WOzsxQJ1YmDq/JtiDW+Gj5jXDIiKy7I/REXov75nFIw7mUbvt1EVT+n8ufW/icWdX9QBZ5YNNZPlqZ+f27qt8N1F68MpW8eBVhoXY9oPY+wdXOPucy3c7OxHGm0R7bMLZT37xAiz/9sO+MU8Pz8y8+L94XlWeU7i1DX/QfXHwA2eJiMjEYmxv498fcLYdG3O2NzSMbXC/ncfDT7FN7//oxW5Zem3K2Zs+tQ/rJhLYBG07P179Q/ApfU/9tv0q2mWI/MMPSqGeLmwigWuZ7WkWEZGxlfhethH7y7bAztGmc404DpPHOit/7F+HiV5amWj60dPOvifzr/sDV5KZPUAFPclM8pq19lYRuVVEpNV0zvxur/LmGnv35SIicvQtiICs+h4uWHzbs84OtbXgixmsn3ztRmcffI/f0S3+ES5uy22PBO88X+atpuTBqoI3n2Bq4/8yvPyVS5z96df+2Nmb4l8REZFrG+g8z53+fp4sREEe/881btmLE0uc/aOf4jjW/dHD09/R1Ezp/1n3PXVQlTw09X3U/7H+iz/6mlsWEjywPpta6ezhHN6qP9gBv6Us7J8n/R+kL/zgV92y1X8W7GN+aDIRdDM2lwtavVrqqu2X3M9MwP3/uqufc/Y/r/iFsx9Lo7+5NI6QyZqrtzq7/dsFg/tAahMmgu/ZHLY3Bw9Zc+P/qX50W9A/T/zSJnyNfBAfRLtbcVefiIjs/HSrW7b3C53OTg2j35AsfhNaX0J7jVzb5+wnL/o3ERFZey+uyca/RqTCLlvk7Ewn7qds83JnN+4bcXb++RdlGtRX2y9D8Z7n+33XP+I5b+tWPJOMZvBb+v7ljzq7JwJfxYx/Lx3OdrhlR7Ptzh7OQaZoDKMNHJrA+gfGYd/wrmdEROQbe/FCk/8hHt4aK3zpm8kg8kMisoL+Xi4iR2awPaU61P+1Rf1fO9T3tUX9XzvU93XETB6gHheR9caYNcaYmIi8R0TunJ3DUipA/V9b1P+1Q31fW9T/tUN9X0dMW8Kz1uaMMb8rIneLSFhEvmat3TFrR1aBntr/YYTfrv0EpIQ3tm539usST4mIyEgeWvXoNZAxaHyhjFrss50eLY95P3H2irD/3eZrEHb82V9ivMe2kS3OvucfXuPszq+R1FFO2quCOfd/GSLLlzn7Sw/d5uyvU+j8qZFVzv7P8fNFROTPMvCRRyPEExGEeJtjkKVG0tCmc7R+OOT7v6cBYwl6EhjL8L43/szZbb+MsSR3b0UYfzaoif/pPghthYSR+gLO8829GEt6WaMvn0YNfLwqAsnhnNigsxM06PUkjRVIGLTV1TFfzrjtff/LLXvs1yCl/vXTb3T2uvdBGi+R7WZh/E6t2j6PO+OBy/lkMmh1R7gdY3Q+suheZz+XgW+zFvfHzgy2d+918PXHzWt9g/1GdrmxZrM9Hm1e/U/tJXvRemdHxuG76CDaf4k/CgOPN/3eTnx+FmTrgfOand3/Jvw+LHoa8ue+JZB1LrzrYyIisv5L6MtNB2Qhu7LX2bHDQ7Cj+JnNLMI+473+WNDcseNSKbVq+xXB7bJ4f9C937IL98zoBtxLY2nY3z54ubOP9OO+yRd+AyJRbC8UCu4/IhG0Dc9Dv7a0HZLgl771VhERaduDdZtvLzMU5zTMZAyUWGvvEpG7ZrINZfqo/2uL+r92qO9ri/q/dqjv6wetRK4oiqIoilIlM4pAzSllwvu5exGCfWLzl5w9RhLdwRwkuiczvkiXtwgTfqsf0tqGRqTXN4YQ3s6TnNeXQ/bHG5onR0t7w5CQbu5BFsFffg4ZNxvf9JvOXv3rWL7Q2PN3yGL57jBSrB/sW+fsjjgkiJaof126EpDcWJLL5Cm87SHE2xrH9WyOQtpLhLOTvpejYh+PD0A+3NJ21Nn9H0JouOurc5qdN2PKlSgY+CAk63/4n//b2RmB315ML3X2g+MbRETkgsZ9btl51MaH8giHp6jMBmUEl0h7xQyYZydwD3ZHRp39/1/+VWfveREZSd9+/1twPo9DXi+e53zU8Zo2lGHHx1lOfBy8Eddoy8eeFxGRr698wC07kKPr6WEAQXsIy5MWbTthcL3uPuynVv9VP6SsrzxwjbPX3g7pKXz/UzhWlu3KpOjXjCkk3chKZLBlI5RtdxJ9TL6B6spNZCctz1x1tlvWuAN9Qtd/I8Ou+6ckc1J5g7U/h12UbQ3J52Yc8qEZoYw8ku0spenHDkE2z60s3CPHTwhWrsNyPNMhP/k8Ol7Ctel4B3zV0wAf9yYgs129eJezvUIfzzXS+Pd61KPyBlQUan8Sv1ctEdxjRwvj8BtPYBvhNgzzKClrcRo0AqUoiqIoilIl9RuBItK/gto+287+irN/gSBFyZtylF5qwoV3xZ4w3hRu6sYb4T+ewBvc4SQGrV3VheKMv9n+hLNThchUlqIebO/MwKUe1dx5/qqvO/v17/y4s5tuR8RqIfDGtahd4tF5L25AJKKJnvQbChGjqAl+2y2pwEz+GqK6RIxX8P9YFlGaZA5vj8UIlYhImqJU/Rdh/10IlNQl5SIy534U0ZtxGnT87ASibm1hvNk1h/0bZJRKuTcYfJ6ka+LRm2+UogLNIfh5NO+/5XVEEE306B3s58kNzn5t48vYz+fwhtnwJpxPXUeeilCUhgeDH/4GIn3/z5Y7nH15ArWdMgWfPpKiUvoCm++JkTyWJwza8EEPg46PFypav7sVRf5+722IiPf9Kt6mP33gBmePvQ179/r6pa6YIuKSW4xaP+Es+odcC9pldABtmgdpewk/EtGwB+dsW6hfaaXq3BPUFqmYY3gpiv4WI0mGagTmehC1CI/iB8mkqNBvO/ZpEuiTQmk/AmzacY7eICJUC5rQ5HJV8UFWEhD95v57KIv7YDyHa5wvlL/ySBkKm+C200C/AWy3R9FOirk0XEPMrsI9LRqBUhRFURRFmRv0AUpRFEVRFKVKFoSEd+QDCP2FSarjAZbtNKBsvGQQph/2HqYZJeMUOv/CMswzNECDark+1EAeg9KK2+b6OB5V1+8MI4w7lEdoMkQDffvejVBi0+2yoNjadNjZgzmEwDtjkHUmPJx3MQw7TmFahgeAs5wX4QlVc1Fa3/djiIbxLm9EzZVjExjwvygGWbFjeWUh2bqDQsxv7XrG2SdzkA6aqc0xRYloyIOEkJe+wHVZYE1Qe+7zIH0Xp3sJR+D7eAgh8h4aUL4ni0Hkv7HiMWf/hwTOyVm3cA2l33oM/r++8R5nP5+Fv7ZnaCBqoW23hnB9UjYaaHMfwlJ2q0HfN1SQ+fbT/bC/ZGYcDKT9p1WorXj7zyCt/sfmheX/Yi0nERGPJn5NHMU9n11EU7z00MD8n+/zjQb4hSWbksmEadB3iaxIcl3Iy0/6XiiN9pHroPsshmPiOT7He3F8HdsL5xAgd52JhPsh5Q9lINXFQuh9cvRby+Gd4vKmCH7no/S9Mfp9GcjgOvDvyHWtzzv7qR0XiohIpg3XPR/D7xn//p8OjUApiqIoiqJUiT5AKYqiKIqiVMmCkPAuXnEwcHmUwnOlIXA7aTl/PpBHSHc4g3UTFN71Aie9xrY58y5cpioML2fp8cJlh5xdZzkxUxKjaUEOp5E9ciwF6YIz4Yr1msrVfgqV8V2Irm2efB0pyFLjlIXHs2zzvl8cw9QKjTGanX4BYS8/19nnxH7u7Geo3lMT1RAaz1PmSsFvw15wRmPGBrf9jjDW/94ofFvMhOGpYbg9cFZmJ9VGWx9DrbUf9mIm+2qmsKgVh25DzaW3N6F+2H+MQwrrjUAebiG5rliPhut0se+YUJl32XGS+YrZeYspo3iAhgmkaOjCQykc30fa0H/+7V/9qrPX3lzf9dBERCzJduEUJJv0CvQ9Q+vgg46XJ2d22jC2YbLwv42QXFRmepyg7fAvgzkCSTxCGZvZrciMjfbheoVo2pKiUmsaOEvzDMELyLruQ4bhZZ2493dSPx0PB98fxaU85CMuWDdOch4/F4zTcBKW0uMH/GPp/2Xa9xCuO1rX6dEIlKIoiqIoSpXoA5SiKIqiKEqVLAgJ79L2V5zt0ZQTUcpW4QJbTFFGywsXCkS4L0bbCPreqQRJexwyTFEZec7Uy1rYr+nY7ew7BUXbFgLnxCE/bh/HNAvvXIRio/cPY+qEvrQvQXTEEMbmaXIiFHrlwmixEMKzXFBtOOvLr/EIPr+0Y1/gsT4ysMbZF3ZDxngpcO365OCbkBkSp6bHsp2Xxwd5C3m6KBeNefDfziykzLMiCG+nLfx5KAf57bkkJMRFsZHCdoOLoraQtJSnd7MeOtaBa3BNWv+1/iW8D29EYczjlJG4LBJc8JCnkfAC3k+Dlp1uHS6qWdz2VwcxXcx725HhyBJeOxVU7c/juK95PTIJ9015JLWBC5aacZy/SaGNHn09puho34114vsoy7SQfWe8Mtl2+eC+vyxFmZu+Z5op865vwNknLoAsN7IRMuyGr+Na2EJRzZLinmcINkDC4yKhByZw/ThTjpmgqY46Y8nCuh59ztnewVnevO0XM0vwQeFaLv7iQ1h37WpnBwuJk9EIlKIoiqIoSpXoA5SiKIqiKEqVLAgJ78KGfc4eyWMkfVOIZ5DHs+AxD1JQPkByKyfPlSMowy9UZh6eTpKesrTKWB4yxkWJfc5eCBJeuBvHuDiMQmZdURTPvCiOApvfGHuNs9c2+3mGx1IoLMfF0jjbjn1aUlCNKF7Pnjhkpq89c6WzP3oBMtU4I6+c7FTvRM6jgoHUnnrCKFr55ka0rTtJ8igW22yLQDZ4JgXZ9dxWyB1xkn9+MIIMv6JsJ4LMuhTNw7clhut+QRzX8sEUpETmxC/jkUGW4wAAIABJREFUWFv/NXCVuuKcBKTfNPmf+xBuwyzhzQacgVqkk+YiLC0ajL6H5bwUZZVd2oLhEPsEGUj1hF1JUgv15X2XIiN0+AL0Q3kqdtr4GPoFu8wv5momsG7Vsl0VmBgkpyxNsxcZRpuIHMf9PHyJf5+1Pgu59YxhijkOD40jz21pE7JYWYrj34NicUyeN48zu3kuVpbtWAbM0D3x0sf9LNWzPok5b3N79+EAueDqaU5FI1CKoiiKoihVog9QiqIoiqIoVbIgJLx9mW5nb4ruD1xngEKz7SFe7odPOVMuNEXRzVPhYo/FApq8rJMyyU5SkcgmCqmP2zLZAAuBToRbD+WQXdJCc7D9ycEbnL37URSR+613+FlMDxrMxxUnmZPDtNGSrEWEvTlUezjlH8s1HTvdsqceQabYwDmInfMceZc0Q7p4XnB89c4bV77o7BTJOS00B92auz/m7E9f9hNnv7nJ99GODOal43nxdmYgHaTLSE/r4yiCmSoUbbw8jnvwW4OXO/t/9CHD7l/WQ5/jeSjftvlZZ2NmqvrldQkqzJrlgpjBc2EyRfktVCbLiAmXyQYO2vb1LdudfdKjIoyGiwnCHqf756IErt2/1ZmEF+7wJTovQbJLJ6TgjpcgRXtxnHckiT7EtDQ7Ox8uXK84ZB9Dc9sJFdgUytSbSn6yseCZ0kwUy1fcA0np5Y/gHAZes8zZqQ5//83NZ2AhzaIEVsaXDRFch3IZ9JGS4R2+3dlI88iGKbs3g37/6ASKOjNf2X2Vs9vX+RmT5pJz3DL7+PZJ35mKKR+gjDFfE5HrReSEtXZrYVmniNwmIqvFz4Z9t7U2OK9XmRE77BPSJ0clJkhFV//PH0X/e0Lp1Or/eUHbfm1R/9cW7Xvqn0okvG+IyJtPWXaLiGyz1q4XkW2Fv5U5YKmskgvkqlMXq//nCfV/7VDf1xb1f21R/9c/U0agrLU/M8asPmXxDSLyuoL9TRG5X0RunsXjKmFd7ISz+YlvgEbYf/TPP+nsR//qS84+kvYzMLioJRfgDMrSOx1FKZALZi6JIET79j/6HWd/+XN/7+xGygjjucFElsrp6DA9MmHHT108r/6fWIuiZyP54OyqFQ14CXqGouHXNBwREZGHRjGnGGd2cZYRz6XG12XYUri+4P9NsaNuWZjS0zjrIiiDqVpq7X+WOEdpXrRLSUY4+y9QwC97B9pld0HC4KKKPB/UdpKS2VfLoriWXMixqZCBuSGKcPnPjp/l7JaP4VjXPAgpZdsEjqk5wnOVBUshRWrte5HSOSx5/sseykY9RoVKs5TpUyxkWknxzIqOpTBsgGU73naTcNZpcB/3YqZy2W6+/Z85z5eAx5dCcgvzXKW/QMZU6vWQXjJXQdrr+gEFY7p9Kceky0ioOcrM5XnxvDLrFyU/nk+vkaJzE2jb2Vacw+IHSPodpwyxbn974SFkDuZCWLcjX/v2XxWUuWYi/r1ts7hPIr2Lnd0Vx3lxn82yXQNnUReGyYxRFt6uEcz3yHPondVy0tk7h9DeBw5jKIpJFLZ3A7a3+nE6lylk3CLTvbMXW2uP+vuxR0Vk0RTrK7OL+r+2qP9rh/q+tqj/a4v6v46Y80HkxpibROQmEZGETK9kfUsIT7Fc0+T8OJ7+O14ckyAaC2+B/BZWbdSJKUaeukLBb9Jte/C2z8d3IIenaT6fuWY2/D++GM0kaXFOlzViSppheivu2o5rdNvoJhEReapvhVt2ohk1oTJl6j2VYzDtn8OeLN4+Gvq47gfeKHjKmGJNJBGRcCtsbwTRsNlmNny/pRF1lspF1LzdGCDfE8Hg1Ubj+4IHjsd4GiOyQxRpylgeLM21hYrtHFGs5hjug9zeI8HHR8d9USOO9RHZELT6rDFd//ObMjNOEcDFFO3hviBF6xTrMo1T3axyg8WroaSmGb0od9Ibe4ra/igN4u+NoObOXFOt//MR/5ijSarv9wxF67tQByq1CD5oojpoXOfJhk4/kLmEclGnIEJ0H/LXaD9eAvfQ2DKs33gM1yVT6IYyy3BeMTr+3H7UIKuW2eh7KtxRxatamvYmHkY7HMzgtyNBkSSOTKULvxM83VdLFP3QmqZ+Z+8awzPlwUGKOjVi27FCckj3xYhY5q8639mhBzHl0emYbgTquDFmiYhI4f8T5Va01t5qrb3YWntxlAYjKjNC/V9bKvK/+n5O0LZfW9T/tUX7njpiug9Qd4rIjQX7RhG5Y3YOR6kQ9X9tUf/XDvV9bVH/1xb1fx1RSRmD74o/aK3bGHNIRD4jIp8XkX8zxnxIRA6IyLvm8iATJfWBgtfhGg7PpBHS3VCYfXsnlf9IzMK0Hp0Uxt2RwSBGeeS5wPU5pF7N/rfbR2VQTkpW0iIi5xZ8Pq/+H12NYz+WxUzpqyMImz7Uv9bZXPamOG3NPYnNblm5KVuao2laBxtJBQwy5MHQ0XH4c2kCtZ+4VtfmxCFn/8fl1+G7P3lCTkfR/3nJSy3af08EEmPpdDTB0udtxy5x9vvW/1hERJJ5vIHGaEA5DxAvV8toqmmPdu5GXZsNEizhsVTIA9ojS/wBnrmjxyZ9R6R2bb//DahndSiHoQFhgdTQFoIs10I6zlAe8nyxlhnLdpXUnZsK3kYX+XMX3ZtLaaof3k8L3TdTMR/+z11zkbNfeafvr/hxqqX3YwRYzEq0tYZjaP/xZ3HeoXbYNl/wE8tM4RkM6C9uh7ZhPLRt24D7rOklTJM0cDYGMufRbKRYxmhiMb4X7UeSTq37noqocLC1iEhyPeo5ehb3fI5r/lmux8Xf9tfh345emr5lxzASYvb3I+nJy+FaRaK4VukR3+dH02hr5gb4ft2Dpz0VbHOqFay17y3z0bWV7UKZCeeYy5x9r739OWvtVwt/qv/ngaL/H7XbZMQOLKeP1P9zjLb92qL+ry3a99Q/OpWLoiiKoihKldTtVC6R5QjXNlLYLlmm7DtzIIeshvPjfog7a3mag+CMi3wF2y6G5TvCkJUeSgfXRmI47B6n3USW+XWgcoeD5Y96IN0D3704gVDpr7fswvKDCFMvo9otxzw/pH5yArWDOPOuMQLJYyQLPyYoo4ivy0jGD70ezuIamxz29+DJdc4+pwM+/dnYJmcPrcW1Qy5f/RBeDzm0JYRskPaS7E3SAoidh6nOT6H0FtcKCpXUQKvu/YlrHBVp3Hv6Wk4iIu0hSNx8741f4GdmxstIeLWCs6ZaqC5PhmTTRlp++TPvdPZ/bP2Gs/cXpj0qJ4NWIufxtXN3IW1udQSZTdd//aPOfu4j/wA7g/u3kbIqTSFL2KY5o3h+iT+FTN4VLRv9ZUNoL6aNpuXgDLUmOKH7WUg5thX9TFFeshGS3LIkg1eRQVYC149iCYvrIFGtKL5eLQexvOdR/7iHNyNTLLMYGcphzFa1MKDzt97kYSrHrkD/sY5+A9pjyUnripRm3BWHbvBvQZa2cXIc170xwUNBsL2xCRpM3+z3pTmS8Fach9qCkVXIGpd9gYfnb7/8R4qiKIqiKEoQ+gClKIqiKIpSJXUr4fVfjRBaewiHmfVyQauX0BXmopr+M2L5GdOnX1SzSHsoOATJ8JMqn0/ftav8Zd+qXwlPmuHzEymEmDvCkA8SuyC/GYv1N0b9LJqeBpTu74jDX5wpN05FMHl5PILttRYKN3J2mqEw+pEhhPw/vvJ+Zz8/gTGY2ZaZX/O5ZGJdl7NbDcLRjXTYaZuVIHKpybd0U2hqiYaLXYbLSNyhgCKQ5WbLWXP3h5z96Bu+6OyTHr4wuMGX/3rvmvLw5pV0J9pTntpWkwmWULN3QAhecj6msHkh60sMnPk2k+mFivsfpemU+FqtvAtS1tiHcM2jhq8tzie8yD/u3EFkqM433hAKKjbc8ZiIlA7fsC0kyUUpYypHctlOFGfNXLIRqw8WpECW6lhmCk3vWpT0HnRMPMULM7YF16LlICTvhif9LOZ2khVNCm1s6l+6+iJEhaPzKb/NR9audsuWXIHfuF3DuGfWt2HqlXR+8lRIIvid5s/HsthfQxTrNkXhQy6a3RBFnxkO+ffNUBKZtUub0BaPno0pqlTCUxRFURRFmUX0AUpRFEVRFKVK6lbCGziHQm8G4fKEmTqw2RNGFkfSzQOFU50N2Y7prKA4XVMo+HwGtvr/t5/6hTrCRCATDKSD51Vq2Q9pIETVTocLRRxzJF3wHEcchmUmcliHZ9ouZuGxLBVKUTG75yHhRc/D93aOITvNq/OZDUZWBd+WXLzxXbvfSp8gi+2i9fuc/XLWl01DNBdWiTxH2UFhkjaiElzotfjd4nZFRJqvPBm47rIf4Rzi13EWIHaabZK6JNeNUP84FfbrCUMa2JlB22rbEzy3ZVDW4mxQbjhC5DiKyPJxc55kgiS8gdf6snbrv9ZOwgsisxryTvQ4pHobozaFKczEJHBDmxzPhRcQHwhVXvhx8ncL/0eQ/WWjsEMkv1mS9iIncN/moBiJdPu9Pst25WTAhUA+M3lYwUufQL97bnyPswdSuPlLCivT70QuYF5TnvNuZxbbbo3jN3iEsuJZtmtP4Lkgl/f3MxHF3bEkAQlvf2tlc7RqBEpRFEVRFKVK9AFKURRFURSlSupWwsv2IPQ2YRHibKEMtp2Z4Oy3DVGEB4vz1CUqUO0qmZuqOB/ZjgwkpC2xYFmLj28FhX1H8gg35trrP2QbT+BacOFLJjKB0HkkCQloR9rPqDk6iuy97kaax4oy7GIh2MUQq0ipnOcVlvd7yHYKkZzS9Tz2zfPlDaZwjVI9wVlm9cL4UrTDuOECsJAqDn0P87X1kIT3wSWYxGm0IF+zlORR9h4X0uQsPL4NOGtsvCDHjuZxPW5ef7ezbxUUAOXNcSZbmua9ynTU53WINsJHPPcm9yGjFj5I7BsI3E6mcK7sw5I5Hmkb4YAMx1OXY269MjIUFXjcm8P9sYKykhMk1Q6v9Y+LSlXWBeEU9YnUb3LRynwVv1yGCnCWDRlMtymyIp6A5GQPIOPMLsf8fMlBaHgm6fdPtomWLTAJz0RwIWwOxx7e6GexXXr5S27Zc8eXOnt5O+Tm45TZnc7RPIg0N2pRujuaQmvlYssjHmQ7lgR5yMKJcdwT2UI2cIKy97qiGJrQeKSyOSM1AqUoiqIoilIl+gClKIqiKIpSJXUr4TFZyihpDSFU92/DW51tosFzgx3O+SG/RSXFNWfOwRzy5rbEKCMsgeP7zhBmM/9MD+Y0S1oKD0ZmkBUyT2QyVPhzojlwncYTkPYmenAtrmvy57q6o/V8rEsyIEt1kRCuczOFb9vjyJ54ecLP0EmRjJRtg89bdyI0vDmGwoIsFdqW+g6Tx85DitEAFU0czsMPvdtOOJtz5n6lEW3rx0k/NJ6gOfSyEpxdEiOpkKVslpmKkuhhavu/1owsqVtpe+1PHqftoY03kky77oL6yv4qclYvMgtTJL8tDuM+uH0Uhfa8PfsCt1PtXINTUbwW5YYa2EaaS9IEF1rNkAw2sTZYjq81licw43nnKAsvTLVhDWVShXIBWhzPVzcT1Tjou3SJbThYzpIjuC40fZvk2/yhJmaCrkN0QfwkO0rOk3jpY90iIrJkFE6LUR/cScWUOfOOM+SXNqBvKc4JyXNpHqWM8IkstQGS8I6MQPLj35fN3X7/FKf+KBGiQpsTwffPqWgESlEURVEUpUrq93GXXkL4DTZM0xJ87+WLnL16Mz8JP+as4hs3v7WVe4MrOziTaCrUoTpmg9/kzbpVzr79ZURiPrdou7P5qZVrLNUr7K1yNbSiLyCaMPaGdc7em/XfAEYzGADdGqWpLWhQa4oGiycieANoovo7i5r9SOJFiYNu2Ve6qObKTzFoMW6C32xMuL59flEvfLkigje17jCif6mViAJ5Gy+lbyPSOVKIXvFgem77XO+JB1uG6I0wSpGp4r2UzHMhLRzfkU9f6eyV3z9Mx42kjgcm8EY4nvXvj3orB3VWCyJQSRqAz7WynhvD1EAmNj8RzWJUqT/PHqMaaMdw3LsyqJGztvGAs096OIfuRXjDrwtMQN9C0aOSmks56qsbaBqpXMCA8eCyZqVUEUqwZaK44lEtvBU0YPqnaB9956GPSy3x7+eGPahtZBvqvEidiEiIw2hwbv4qqAwXXbxLREQOjHS4ZV1NwVGnZA6/k5xIlKV1joz7/QbXj0rRgHNDUaeN7YjOC8bvyzkt6Fd3JxeLiMhLI4vcss4YBpFXikagFEVRFEVRqkQfoBRFURRFUaqkfiU8CjVnbbC01vpDSBqjZwXLMt4MZj8PImVPL2OMrUfMsP1O2vdVMPl8bKqykvG1JESS16LG0cB1vJOQD2wYEl6xXlPWw3ken0DdD56mJUIy0kiG6nrEJl//+5Mb8EdQ6F9E0pR8MEYSos3Ut89PvAtt6LfP+n1nj6zEOXTc+7Czk3dhQDNPs5Iw/iDLEI1+TZTUlYLvS2oS0XVgWbtYkyhLtdgeS0Nq/Z3f/qGz7/ybLme/eRUkRpuFHNske6Ue4XowKZLwWBLe3r/E2Z2tmAKCcYkOrHhQf1Su9lM5ivJruSliMpfgnnhmHEkz72xGnbAjHq5/T9NYYbv1i/Hgo3wsTMuxTr6J50eZYntUEypwqpfTftkUtkGyFQ1gN3Tf8HEnTuD3wYvh3nYyJPdfAdOh1AOcHJVPBddIGv1TtLlcIdkonUVbHRzFoO++BKS4TpL2WhuCtz1emPKLp2wZSEIe/NyWO5z96Sff6ewPbUE/+dO+TZO2WzK1lOX2VVlyl0agFEVRFEVRqmTKByhjzApjzH3GmJ3GmB3GmD8oLO80xtxjjNlV+L9jqm0p1ZOySXnSPiAP2btFRLao/+cP9v24DIv6fn7Rtl9b1P+1YyI7on3PAqASCS8nIn9orX3KGNMiIk8aY+4Rkd8SkW3W2s8bY24RkVtE5OZZOzKPs+aC6XwG9XKOXNMZuA5LE0UqybYrR7EOBU/JwIwug0uX3N8fuE7J+eRPP32MESPr5VxpNR1yr719p4h8Yl78T2SS8GFjZOoQc7odz+XrY35GRGcDwrRdcUgkEx5lpVAWXiwEL3Fdj6IUd9vBi92yMomBJTVvmqlWl4Qru/7s+0fsvTImQ/Pi+9xBZIuEyS7XU96y/sfOHspTRktB5+AaT0w5KYjhrL3ifcPS3yjVqfpEOzIj7xRIeCzbVUOt2j5LmF6Zd8yTL3U7u311sITUGVB7rpLposoeV8H/mTIZwIeuxbUfObIe+1z0qLOzJdPK2MIxBTPv/i9IpDaCY7RUF8nkcd9yk851wf+RftRKswn/uzYMf3EmX8mus5TNGC4j8busvjL9Bw/NoGy68DH8ToVykPCK/ZZtwHUr1oQKhcKz2/cUZcIyw2HcanEct4nhuPKjwUM3Xvn8Fc7e0LjP2Tte8bMQwzHcS90d2EZvU/D2Gmh6lgzN17O8ya/v9/ixFW7Zc5d+19lbv/hxZ6/9G2ThX/wShgn89MRGZxeniRmloThcE0rylcnrU0agrLVHrbVPFexREdkpIstE5AYR+WZhtW+KyNsq2qNSFXHTIK14yciL+n/eYN8b/4dPfT+PaNuvLer/2hGPNGvfswCoagyUMWa1iFwgIo+KyGJr7VER/yFLRBaV+c5NxpgnjDFPZCUdtIpSOTFR/9eEvP+err6vHdr2a4v6v0Zo31O/VJyFZ4xpFpHvi8gnrbUjpkzm06lYa2+VwiwPraazcu0syplAweSfe9HZqV+/InCdReHgUOF0KWbhdZWZGmZ8BU7R2/FS4DolxCoLFeZsTkRknYh8YF78T/Cu1jfRFCI2+Nh5pvSnUitFpHQm7HLwFC9cUC1Ocl5T1F/n91fd65b9efsHnY38PpEXMhC9OAuP5eFKyNmcTMi4yDy1/WrD6EMesltaQpAwilJbObmZSVEomyW6IJoMrtMJDx6/fwLHF16Mft07jjYTakL2TX6ikFGTP30e2Hy3fc7GyZdk8cKPvQ9hc/1bg9t2uAK/V0M18t/I05BQGy9EG/IEEglnxp6OefN/sT/h7ZeRUlLdWCfdR9L/gQEcd6NffNFMIVvNFiXSH59DcmLyyvw9mqKGpUSROeh76PsmAr/ZQnamTeNhi23+3pFP47e281zc2y/sQ/HQaIPfzlqakDXXEsf2WmPwSTP1+0MZyLFcbPOlPr8/ue38r7plZ9/6aWev/PxDONTzN+O45WlnpT34ORqe3OeUKxJ9OiqKQBljouI/PH3HWvuDwuLjxpglhc+XiMiJct9XZkbe5uU5eVhEZED9P78UfR+VmKjv5x9t+7VF/V87tO+pfyrJwjMi8lUR2Wmt/Tv66E4RubFg3ygid5z6XWXmWGvlBXlCmvzYynH6SP0/x7DvY5Lgj9T384C2/dqi/q8d2vcsDCqR8F4jIh8Qke3GmOJEW38qIp8XkX8zxnxIRA6IyLtm9cgomhauIGzprUaoMJnnrB+/8c0k844pZuElaXtjeZrbbeXpw7UiIjEOo4ZOf1zD0i/H5IA0+5P6bC5cg7n3P5HYA3knez7C1PdMBGcfeaSWDeR8eYOz8IZS+N5oCitnciSd5IOf7WNRX166ve8SLBsJ9iHPB3ZRF+YD23+wO2j1SbDvkzIm8+X7kjB6ZuoMth+dPNfZ/7z6Tmc/nvYzfjgTtSXE9wl83xiaepzEQKEoKhfmvKbhiLP/cQDXxI4FzyvlZDuRKaW7WrX9pBebcp22p/Hiv+fG3sB1XBHf6SfelVCcl7A4J96pZLrgz6XbgnfKkmRDIaO2XI9Vs76HJDeTo6KVcfQP6Q6sk2mhviJIrqtESaV1TBnp1QbFG8JlLi5vrxH9HW861+yfT3SEtlH4bRhKHpq9vseImIK8aHOQ54OyY8NbkKk2cD6GQJy4Dm2uvRPZ5cePYU7Opna0pIaYv35TDPuI0lAMbofZPK7rikZkLP7kAApf/vN53xIRkbd+/1Nu2br/C7Idk6GsTM40DirazNneTKWFNKd8gLLWPijlu4BrK9qLMm3aTbe8QfzKqvfa21+w1l5MH6v/5xD2/aN2m4zYgfPpY/X9HKNtv7ao/2tHR9MKeYPRvqfe0UrkiqIoiqIoVVK3c+GxtFXJU96mZZjvqY8kvIz1s36G8pXPl1QJI1RAcDQ/4uw1i4OLZzIl5zOFhFcP5KM4xrYIzetUJjCZOheh3Pe3bRcRka0JFFlsDQXPd5ShfEueJyxKVfOKGWfrowj1XnfRFmwbtdXkgQHMDdZEmR7xQ1NLNHUDSxIsZdPy0df2OfvCf8HceXe94YuTNrc3h4KzLOfly8wZyUUbr0z4w2C2Z1rdsiu+g0yYtbdg3ikRkvBKsqrqeeY1n+cGlzn7be1P0ifwkbf7FWdnunuczcMHwqYg0VD7Dc2Cntcexj14woOfV6/DMKXY7+zHF/4eJt+zG1v99Z+RBQLJKuE0ziOSIl0sRVJ0KMDX3My5XVKzLDtHXlGRpczAPO2jpNAnZdN5J3B/tu5b6exMk7/BREnh0DmYp9OWSndF+j+MbLr+y33JrbWbpXea43EEv59DQ8ik7V2CfriLhmkUs6hjlO3GRTJZtuMClizbnb/4sLN/60k/03rdHz4y6TxOxYvBnz8aRODOmMm/tXm6H9KcPl5h5qZGoBRFURRFUapEH6AURVEURVGqpG4lvGIhLpHKkig+uwrZnCsjKGzXaPw5dBJm6tBo3CBbKU97DdFzZtb6Ickxi6KBi8LY35fPgob0cbkqcD98PpFo/Usa4QmEOVmO2JfpCVpdojGEZF/M+uHefVmsG6NCjby9bJk5vpjU/2HvzePkKqv8/89Te/W+JN3p7AsJEMK+KaAiqKCC4ILCoOJvdBiccVxHZZwZncVxXL46Oo46ojjIuIsLiNsIsqhASGRPAtn3Tncn3Z1ea39+f9yq53M6fSvd1UtVNZ7365VXTt+6dZdzn/vUvefznPPkvGskZYxsrX8LaQxTojqtjnPKbRw6dcL9VCUyOynEW1eG59e8Y6OzP7DmLQCAj/7qB27ZyhALDe7JMMumKSCyJHMszHlWlOv/0yFv7OqOc+nXlZCynaCI3DgX2NXNIpTzT5A5arXjVwYQqGdf1SskvMGc//ycU6WQ1ZcWUvdwjgV97SQKAcr7jTJ5lbxH59uJFbKYnFMuNMB2Fx5kG0008/gbZTHLgqTmXwsVJs37xmTpF5MpUiA47vU9uej4+VXl8QNAcgELzA6edzaPbxfPofckbxiI9ZMaZxATCiHY4vW/Wz5NCTFez9+wgkA3PMyhKcEQfRKJsY3XximThgNCzhTtrzAkICQyRuW8p/UhbuOZvg5nN8bpn5BIWVzyhmfGn1hAXms5l6E4DiHRhQLjr2tGZHuHi8wZejyq5M5RFEVRFEWZO1RtBCoknn7D4inyaM6/ask/vPZtzu4+nzNeRwYLbzViZfFCLJ/+U/ViQKCcWVu8TYeHveU58RKSruPnCx6S02087SxZK0oeSiA4s9M9zAYtz/FaHM3wze+uveucPQ9bnb30Gp73v0Fm384cPwEjWmvwqO86D+1f7uzhBRw4HvQfwz6nGDMotEi0J7t1BwDgYyv5BvyhHbw2K8OMLm1I8M30kjgHIL/p2eudHXm5GJjsR5E3wrlG8DlGmpr8g8hjuHDVDmcvFtHvmPEGadeYIhGLKZIWI54bA9zf/et+6uzLxH33hKgr1iQG7O4dLUTI+mf0+KaNrK0k6kDZGO/hxt0ikpYQHbqIJBUGZOcior6cqCUlA3ZyfHEgyW2EBsdPZ2LD/GIgkR73OQCk60WUcCGX1+9j75+Jj4885aIz/5OcaI9i27tPAACcuWq7W775EOuXhULe79D8FiZEyYjmaIptOBZmGwqKqI6cKqVQX2koLer8iYHjRxK8x2R9phuXPOjsr5/B3xc/jIxUip9RWRcsVCSqlMgfa1ZEoPrT/G2oqgQxAAAgAElEQVRD+vjTWRXQCJSiKIqiKEqJ6AOUoiiKoihKiVSthLe4iTUoGgIc2Par0Rq/1ZF7YrOz51VhYZP1SYYsL40zrLio5ajf6lXFAVH39mPz6eeRHEPqVehy3HTS7539N82Un1acvsrZ7WU9olmiyCBtE/auj5yy4dOrOID+r7dRdr2iloPsP3/kPGf7ynbFpLo5LNtJQhyXjRqRfPLxwyf5rA3s+vTJzl559QnODu+Ljlu3yCwhCKTE8AHhxoBUavPfDQjVSMpQWVHqbhE4zcXH97/a2XesusfZh52MUiUSXr5dZWL0eaiGPhxdxD608fe7nZ05xPpXUngx+eWBKLcRjot55eSAcyn9CfnGpsZPmyOFt5zPdCgAEH+S9mIxW93gm17g7MT8/HAQISsGigxgnw7RvixWfd+T5navZWLDeUvG39vnNnCZ7N97M/T9vhEmnwyk6c+Mz/RbGTF2ZijF63B5B39HVkY5LdKtr+KPTW6Etdb8+rJiiTShJPvDjEhMCgidtiZfFzAm5MM3zVvv7M/jynHn4odGoBRFURRFUUpEH6AURVEURVFKpGolvK3bFjr73mUM/b3rt29x9hps8P2uESFbF5o1s/isKLMBkv6z2r/jt3/u7P+55FZn79zuZUKswd5ZOrjp07qRvvvLc1n+/757mOmzQtYDMrNb12QcRSSs//rpq5z9xKXMPmu/r2qb/YziN9u65LPvf7Oz7/jIs87uuWmRWGsLzcJ1fZ5IdcWo308ZpSfLEP/W4TaxFrOVan7C0P/qn8zqoU2JbT/ilEYveEWTs2Nf9KSYKA6M+05FyLer2O/Z5qSEFg/zPBCd/HRMsk/OFumfy0X0qMjwG/Iy26JPUKrKDYi6XjO105EE7OObAADzX8PFRxbzPk+c6P0ObTqB02IdXS2OoIN+a6hnzThZt6k+wnWao946svZTROjRX3voJc5ec5PMot4FP2x2fJ9js/5yZ8MGtmeZhSftwbz0GBSa+rs2XufsFdue8t32sWgESlEURVEUpUT0AUpRFEVRFKVEjC3jNAvGmB5407QfnmjdOc48zM45LrPW+s+fMgnU/9NiJny/B7PXNqqJavW/tv2po/6fHNXa9rXvmTpF/V/WBygAMMZstNaeU9adlplqPsdqPraZoprPsZqPbaao1nOs1uOaSar5HKv52GaKaj7Haj62maLc56gSnqIoiqIoSonoA5SiKIqiKEqJVOIB6pYK7LPcVPM5VvOxzRTVfI7VfGwzRbWeY7Ue10xSzedYzcc2U1TzOVbzsc0UZT3Hso+BUhRFURRFmeuohKcoiqIoilIi+gClKIqiKIpSImV9gDLGXG6Mec4Ys90Yc3M59z1bGGOWGGPuM8ZsMcZsMsa8J7+8xRjzG2PMtvz/zRNta5aPU31f2WNV/1f2WNX/lTtO9X1lj1X9P1tYa8vyD0AQwA4AKwFEADwJYG259j+L59UB4Ky8XQ9gK4C1AD4N4Ob88psBfKqCx6i+r+xxqv/V/3+S/lffa9t/Pvu/nBGo8wBst9butNamAHwPwFVl3P+sYK3ttNY+lrcH4c2+ugjeuX0zv9o3AVxdmSMEoL6vpO8B9b/6fxaYI/5X31cW9f8sUs4HqEUA9om/9+eXPW8wxiwHcCaA9QDarbWdgHexAbQV/+aso76vLOr/yqL+rxzq+8qi/p9FyvkAZXyWPW9qKBhj6gD8CMB7rbUDlT6eY1DfVxb1f2VR/1cO9X1lUf/PIuV8gNoPYIn4ezGAg2Xc/6xhjAnDu4jfttb+OL+4yxjTkf+8A0B3pY4P6vtK+h5Q/6v/Z4k54H/1fWVR/88i5XyA2gBgtTFmhTEmAuBaAHeVcf+zgjHGALgVwBZr7efER3cBuCFv3wDgznIfm0B9X1nU/5VF/V851PeVRf0/m5R55Pyr4I2W3wHg78u571k8p4vghUSfAvBE/t+rALQCuBfAtvz/LRU+TvW9+l/9r/5X3/8J+V79P7v+16lcFEVRFEVRSkQrkSuKoiiKopSIPkApiqIoiqKUiD5AKYqiKIqilIg+QCmKoiiKopSIPkApiqIoiqKUiD5AKYqiKIqilIg+QCmKoiiKopSIPkApiqIoiqKUiD5AKYqiKIqilIg+QCmKoiiKopSIPkApiqIoiqKUiD5AKYqiKIqilIg+QCmKoiiKopSIPkApiqIoiqKUyLQeoIwxlxtjnjPGbDfG3DxTB6VMDvV/ZVH/Vw71fWVR/1cO9X31YKy1U/uiMUEAWwG8HMB+ABsAXGet3Txzh6cUQ/1fWdT/lUN9X1nU/5VDfV9dhKbx3fMAbLfW7gQAY8z3AFwFoOiFjJiojaF2Grssjm2ocXawI+Xs5NEoACB8aHhG95dt5XkEj8zstosxiL7D1tr5+T+ryv/PdxIYRsomjVhUkv/V99NjLrb9dJu3v0hvwi2zmeyUt2eM1/ym+tI7HarZ/yYScXa2NuzsTNz7PzTKdQOpnLOTzRRgWhsGnd2f4m+JHQg6Oxfz/B6K8hrmBvgTGulP83tJ/gaNPVhhT/IyHtP3VJXvi5Fc4fkwHMq4ZemU/+NGQw0v0PBh+j5bQwcF8qcfGuUyc3RkZg52Ao5p+2OYzgPUIgD7xN/7AZx/vC/EUIvzzaWT27oRLW0SHUb6hec4u+kf9jj7uV+tBgAs/sRDk9vvJOm96oXObvmfh2d028W4x96xR/w5u/5XxrDe3nvsopL8r76fHnOx7R+6/gIAwKL/fdYtyx7pnfL2ArEYACCXSEyw5sxTzf4PLV7m7P5zFji79xTv4adlMx+aavfxx3rn6+POfsvLH3T2T3ef5uzcb1ucPXCS9zDQvozXcPi+Nmcv+ekhZ2e37fQ9VhPiT67NZHzXOZZj+p6q8n0xtv3LWQCAhW39btmB/fQlsvx9f+VZTzv74dvOcnb/aXwgrdvhPRi3bKbPYnc/OnMHfByOaftjmM4DlPFZNu5JxxhzI4AbASCGmnFfUKaM+r+yTOh/9f2soW2/sqj/K4f6voqYzgPUfgBLxN+LARw8diVr7S0AbgGABtMy7dhzaAXfNl58F6OWzw4/5+xIgE+pL77+FwCAS27kW+ChTL2zP/Dff+HsBY8wJNj1t0ln53IM9Z67cC8A4LK6X7tlP37TGdz20+3OXvXBWY1MVcT/imNC/6vvZ42Kt30Tpmxk00KuCVDy+cy7vgYA2PjnK92ye7tPdHZrjNL/kQQllkW1fGv/xKJfOPvFP/0AAGD1u9cXOSjx22pEflBu6rJhESri/+xLGZ04sjbm7Ix4Pqg5xN1c9MonAQCPnM7fjO7nGnl87Yzk3bV3HbcR5fU8dA4jVis7jgAAXjx/u1t220oRofoijyMUWO7s2rdyG5lDXc4uRKMmG4nKU/G2X4xAPX9X1yz2znNNQ7dbduAgfWUibJP/2H6Ps1+85ExnN7QPOXvAetuOHeb9xRZQOaaThbcBwGpjzApjTATAtQDumpnDUiaB+r+yqP8rh/q+sqj/K4f6voqYcgTKWpsxxrwLwK8BBAF8w1q7acaOTDku6v/Kov6vHOr7yqL+rxzq++piOhIerLW/APCLCVec2sZ9F4/cQntPotXZW/s5SD5o+N3dg946vw2d5JblLEPdl/8ZZbaOtx119oNHVjv7aIrBwp5EHQDgZ4OnumXxMAe7Xf8KDkb87pKznb3i2qd8z2c6zKr/lQlR/1eOSvt+jGwn6Hwvx/N+4Kue/aI3PuaW/eOKu52dsMwYS1lKE6+p5VCCy7Zc6+yWJz3BYPSq89yy+J1iIK3sM62Q7UpMyJkM5fJ/sJ2DtA+dxn64cQ9lr2Q9fdewm7Lcves9We7S859xy+7pXst1N3IQeS5Du5c/K6BQC+wc8gao79rc4ZbNe4IiTnIDh290nko/20/y92H12yjhlSjd8XtV2u9kT1vl7K+s+i8AwE3br+MKKSF4DfLR4693v9bZ4SG21aFdlFsL3zx8AX3Z9n+LnJ3Zf2DKxz0dtBK5oiiKoihKiegDlKIoiqIoSolMS8IrF8HVzGLpqOlz9nNHGd6NBBmyTmcZ0k1mvVNM5bhsOMXA7I4eZl9Ewgypjo5ynUiUy6NN/eP2YYUk+NBhHuu6hZ3OTjQxHJntp1SoKEqVUMigm0TWWnDtGmcfvHSes+2LmEGXTHgS3S//yLpC9edTYnpbM2vTpcW77AcPscbcjicWOzt3micL9QpJrmHpBc5u2kF5I3YvhwzYJDOKZ0POm01yizk0o66T9ZyCo7Qj4pyGlkSdvfo7nhS67SRu40WnMVv7oUHKeeEh/1hCsoVtoW6H91sy2sF9Dy3mvlPNXB49IqS9eVz+fObwaZRBP3noMgDAji0L3bJgM2XvXIz+6R5h9l42KgplZujb8LBnJwN8ZBlZx21HVMJTFEVRFEWZG8yJCNRz72Sk6crYXmfL2imjWQ7IjIry8YUB4ykRMaoRg77rm5Pj1gWAUD1rUIykue1CREsOHJfRKLmNpbWMlt1zEwd+LvrkzFZFVxRlBrDjIwWhJYwAbf4I33gRE1GqJPubsKgz5N6mG/j53T9gxOilb9/i7Pc/cY2zAxsanJ1bxP2ERvLvuzkRcVnOYz56unh7fzVr061+V5GB5nOAXIQ/USZrfe3woOjvRaXvZEs+GvV5RqC2/BWVhdPO2+Hsg0O8bsNJrpM8yOjIgg1ePaed7+DxhQ4x6pI+kb8l6SQHvNfVl79yfCUYoPiCRzuXAgACo4zRZMP+jxttNZxGpyfAgfiFqJMk2sftJZq5PDJuzfKgEShFURRFUZQS0QcoRVEURVGUEpkTEt7VL2EIejTHYF19hKHRgrQGAMkM7YKcVxcWA9jEdEJSfpPIbcgB6qFAbtz3pGzXFGYYV8qK667kVDJ9n/TdpaIolcRH3tr2Ts6aEW5ifSa7R8wfImdNCY/fhhkS3axQJd69gTWeAtu4vZzcdFJIIHlJMDQihhoImSPWw75xeDllrb4bXuDs5tvKM/H5TJFu9BdnjFBbjbhu4WF+kIt4vslG6MP6z1OS23QxCz5l49xG41b6tP4ol4+0e5JgboS/B6kGfh6vYd8/EuNg9mXNHMqRXUCJSk7r8nwgsIRTEw0PexJmeFA0eDHNUVC04dYzeF+Z1Rw6Y57itbL5W8iI0lnBdOXlaI1AKYqiKIqilIg+QCmKoiiKopRI1Up4iSuYtTaQ4VQ/SVHPqV7IZQkhlx0ZZQw8m/OeEdPWX6qT8lsx0kLOi+UlwYRYJjPyakKUClM5rhMSMefQCs4Ontm1Z8L9K4pSPuT0IZmFvJ9NFzOujOhObJBSQjDF/qTQtdiAXJd2XS2HIAzEhW4HISdt59JovnxcN2eIGiPnZWr4vWg3dzS0iOuIxKU5QSZO50k/StkuOEpdJ5CgvJaLeT5INbIfHlpMSXDh7/k9k5HSn9ynqPPU4C0PHxGZgSdScoqJOoJDcR7H9h7WCYu+nhmWbV96fkl4Nke/RWOp/DLeM+FBfp4T98wjB8XvYVr4tk5kXebNoMhAzYYn/u2ebTQCpSiKoiiKUiL6AKUoiqIoilIiVSvhZd512Nmn1LFMe3eKIdChLDMd6usYDu9PMHMmkw8rZoVUlxWhxoDxH8lfyLYDgFSG4fBoPiMvKD6/cP5OZ/ek6vg9IeEtrzni7Ic+xlmrV79NJTxFqSbSazjLeygmCuZG2A9EhsVUTkEh24lX0kBq/DKZRTQ0zGKLMqssKGZeSbRw24XthUbH7wMAAlkhNzWxX0s1cuOBGPeZS1RngcegnPYqynOSxTMzcfo/kBLnl6J0VpCUZBZeQGRupeq53OSkDd/1I4PeB9Fe9uutTSwCOSgKcAbr2W4yGW47+DJO9YMv4XlFVpxnQ7PXtvpDvJZjMujENC1DXfzNjLawcafFtC4FqVpmueZCKuEpiqIoiqLMOfQBSlEURVEUpUSqVsJrfAtDo1/8u8ud/Y+v+rGzU5aH/50PvNrZr/v0vc7+3x1eNl9jnOHqtJDtUiKbzhSR82ojDMcOp7ww7YUdlO3u/M6LnP2iax5z9v/X9jtnf2I3j6/uKUqPfzLM0CzwQ9ecDwCo++H60vZpZCqUiNHnj2UuSBtFMf6hbBPx2qpNJn0/P/DjU5zdcjvnlYz/9FG/1f9kSDUxozeXE21VyAfpJfRpbBvbjpxN3iXTyaYvsvCk5GFEVlIgzeXpRi7vz6shsd7xsh4ApGuF3UApy4jhC6ZOrFSl7dwu7XC2GT89obeOkE3TdfJnjHYhm06MpBiTVTc2MduKdbhUykSFfWZ5uREWRZaPDjKTsl5kWB7to88bxVATk5+3z2aEtjWXyR1fUgtICU/I0EvXcIjOtQs3OPufe692tk14FyveLYrIJos0jjKiEShFURRFUZQS0QcoRVEURVGUEqlaCS/b0+PsE95P+4efOYvrLORcRjW7dzj7ivqnuX7kTABj5bkaUfhSBh1lRl48xHVkmDaVnwPv2mZKSFs+zdjkPU0vdPaDAzzWRZ98yNkdYMjyeU9BXpqEbLf3oxc4e8GL6KOhJCXPdfM2AwC6Hl/hlmW37+LuQmzSVsgvRobuMyK7Y7GXcbX5owvdspM+wHkLc4OUkquWIr4tJt0VSG9iRus1//YLZ39jIeXmti8/hKmw7+95Le0Z9KGc3yqdL5RXc5DXpuPLG/m9tNCnysjAMqn5sB8Iieyr0y9gm/vj0Epnx7q5TqrBkxjkyAAzWETmEDKblJxk5lLhVgqIyxoa5cYHV8n0MZrBfvFHE685DjMzuJpItlHyCqb8ZZpsRBQQjQWFzXONH/GcJzPpJiMJ2iJhhcJ2AmwSCIoNRqOirQS53I7y+FprOF/crn86FwCw/B/m1vyExQh1Mgsx2eI14rq9ouBpknaiRTj5wy3ObPkOC5MuX8VCo4d+5/XTNd30a2SAv8uVYsIHKGPMNwBcAaDbWrsuv6wFwPcBLAewG8AbrbV9xbahTJ1NdiMOoxMR8CFC/V8+Cv7PQjx0q//Lgrb9yqL+ryza91Q/k5HwbgNw+THLbgZwr7V2NYB7838rs8BCLMOZuOjYxer/MqH+rxzq+8qi/q8s6v/qZ8IIlLX2QWPM8mMWXwXg4rz9TQD3A/jwDB5X0aytTOchLhd28hJOEBUUGRXNMW+4fzLLUw2Iz6U8J5Hr54TQt6DWkyPqRfpLQQYCgBUfmdlwbLOZj1E7fOzi2fd/iZgww7djpBcfeSl44gnO3vKhJmdvvfyLzv5I1znOlvMV3n/Q++6C4QHf4yiW0SIT72TbOni1Nw/TvIUs3Jo7hfJg8yODVel/KVUGZOFBIcvs/ztPRqs7wGvQdDvb5/J/pP3L/1zj7I89dLuzP9L8Vmcv+bfxct72z7/A2Re/4BlnR5PMUh1Oi9D+hfTl8oZeAMAfD7DwbfBnlFKbd6Yq4vsEpy5DTmTEiUQ5DKaYihVupqYW2O/Tpeb8M7+saNeyG7IhWSxQbCe/WGaBxUTsIdI+4uzkEc5BJl+Tbc3kM4Ar1ffE9h119vAayjujrXTe6Hz6rnGnKJ4p/FsovCmzv0RTRKCIAiTXz4hCnoXlUsLrHKAkGgjwujXEmG13pIYb7BqkhN201X//BeZK318gLOTphQ1e/3w4ydkXo0fZCY+0i/kGN3DIzWtq2YbvaT7o7CND3m+slG6rgakOIm+31nYCQP7/tmIrGmNuNMZsNMZsTOP4YzKUSaP+ryyT8r/6flbQtl9Z1P+VRfueKmLWB5Fba28BcAsANJiWyRcAKjIw1kT5BiUHyaaaeCq39DLsWYheFOo3AUBLnE+5yPGVRU7fMiTW96sPlRYjDbMdfEvCfg5+DtRyMGRuROxzGnWQSmXK/i91P0UG/KYu9wZKDv8NpzB4x8o/OPuiNN/Izt7wZmef0U4/nlLX6ey/WOV995m7FrtlG7/AKEjjt0V9qCJ+Dpx2krOjfd41P7KZCQmRpfxe3SO+m5gUM+L7gJiyIsa2L9tTtshg4NGTvLfg/3j7/7pln739FN91ZdLG1y5+sbP//rffd/ZPLvcSMv51yV1u2cIQo1hXb7mWx/eFdmfXPLDF2SExKL+wx6Vg25DxQxONwuTSQNrIMj2TZqr+l7WcbErUahLryKmcMmnZh3CdQrchIxryPGS/IgeLB1JiTyIRojBVixGRk2BKJL6IQcypHGsSyYHTNlK+vKEp+/+57c6ObRXRu6vOdXbP+Vw/2i/uEREdyoW9ixHIyF2LmljCLzLSJwedy8HlhXXGfE9EEVe09Dr7zR28L75540udndm5G+WgXP2+JHZYRODCXt9zWHyeqeHNkRORQMn9o1ynP80oauzI+FOYy1O5dBljOgAg/3/3zB2SMgnU/5VF/V851PeVRf1fWdT/VcRUH6DuAnBD3r4BwJ0zczjKJFH/Vxb1f+VQ31cW9X9lUf9XEZMpY/BdeIPW5hlj9gP4GIBPAviBMebtAPYCuGY2D1Ji0/6DhIUSh7CIcY/kRw3WRigxRcTowWxAhBVFeLdOrB8W649mvGkevnGEMmGgnwP95LjE3KiYKmGKst3Tdj360IO8jn1a3ufT9n+xeklutHWJx2svON3Z297OqTA+dqF3fx9McTDhjw6yPta+Pg4ib2tgDZDOEQ6M7h6lzPeKNk8Oem0L6wW9+V84uPnZf+A0EPtTlFabQ7xGX3mW13y4z2tPV57+pFsWfzl1gP/9vuf/HHKYdvsvyHE+U8mMWa2e5yvrUEnZLv0yJk0MLWY8vPk2Sgerb/CmFQru4P62/wflzhPeR30y2MDBsJkDHLz50V/y1LZe82Vvuz96n1t24s0cABoZ2SPOgvZUJ1x4KvHgrLT9iZAzvkvJLRsTg/EjYi6KMd+lbTI+EoO/mgSTFX+I19qAWF6o/yQHSstBzosaOfj66G7eV0JlQrqRMvBEnf9s9T0lIe4POcVQayvr7R05j78Ji38lBv0nvN54tE1clCLblhdjzFQu4qsFeVBOnxMQUu7WrvnOvvUlTETxqg2UTsH/M9L3lIF5T7GPnRf1+vJton1mRf5CoEiJt/Ujq5zdm6z1X6mKmEwW3nVFPrp0ho9F8eFUQ7H/HnvHU9baW/N/qv/LQMH/6+29GLC9i8VH6v9ZRtt+ZVH/Vxbte6ofncpFURRFURSlRKp2KpdijJmSQ2gDaTHCPyZTMfLIqVkKMhwAtMYYduxPcdT/UIrxxmZRm+JIwgsrNoe5zMb9UwqKHWs1MNUZwIfeSAnozA8+7uzrWv/H2d8+wmk8vnfQy5zpqGHdpq17Fjh7zTLW8lrXROlo1zCz4pqFXLJpyKsTtDvBz5fHmIXWEqIMKGW7RWEWzPnx2bc4e3PKyxZLiFj99iQzyABRdGe65PICb0BOAz9ezpOyncw63f33lD5Tzfzeg1d9xtmXdXzI2Yv/3ZM23/kosxt/8trPO/uD7+O1zA7419Vac/MTzv7NFd79sfrdzHSckWYt/JG8nOcY6xSZq4/dMRN7mhxC8wqIGj5W1ISqDwt5XtZ5EjWc3EgCKQnJDD8hz41J9BW2lBML2XlhMbtQZFjIs92igFUzNRLbxTbUexLttntR9YwZaiD6rNZbKVWPdFwgvjG+uFM6Tj9HhkWNrfDEWVyy7lBk0PturJ8+D0Tp5+SjIhtbUqSm4fONwCZOb5TJ68xStguKagqmSA2u+3rW+C4P56csSjTRl6GRyk/lohEoRVEURVGUEtEHKEVRFEVRlBKZcxLemIwxQZ+oD9iXZhG5wYQXQ4yGGP6VM2iPZCi/1YQYju1LUM47NCxmMM+TFakaR09mxkv9U2IlU73Pp1IaMkFKKGaRJ6/tfw1ltuhLWQ7tL1axiOJTQ5yC4193X+nsfuG7C9u9KT0e6+W6Jy5nYcxXt3P6j/VHmbmytJaS2/ZBZrec3OBJfjVBXqu9SYbOO1PM3lse43HvS1Hy687weo7kK7rJzM3B7AzKdhPhE9LPiGmJRtopLd765i85+6tdFzv7tn5Oe/P+t/7Y2Z9qeC0AYMV1lDtOPMBrvfUr5zl7zTuZ4STJJShVrR9eddzvycKxPdee5uyBl1NKfcWq55y9MOoV0FwbZ9HUK2s2cNs//CseyGO+hzcr5OLsH2yG9/C85WyTv73/DGfHxBQWmRoh4RUSWsW2jZD76hopTY/GRXbcqJjiJSFt7//hxdxizWF+Hr+f2Zutr6E03n2IkrRQvqsXIXnZ7MQyTf3eIgVz81O5jMmMzHLdbITXNpTw30bOJ4FPFtJMpPkTmmopImjLe9z4ZWY+P2Q9OfTg0KjXZyebeb6y0GYxtneysPpZy/Y6e3TE821uPv0d6eX9U6kRMtX7C68oiqIoilKl6AOUoiiKoihKicw5Cc9lMh3DvFNZ0b4zQRmnqcYL88k5i6JBBvyao8z06RIFGwMiLaYhShmjkJ03kKHMM7BCZOfIg6qy1DsTjSC0eDkA4C2/fMAtv+0As1jiIS8M+575LE75s24Wyfz8pkucXRdnWsW8Gso0KxuZFdeb9mSdU8XM2m0Rhnof6F3t7IVxFgKU1+uNCyjrjOQ8/3eleY2DIoCbFJNayeV9WcpLfULHaA17x10jUkSub2aBySfATLVpUwjfyzYsM/Lyy//t68wSfOv33uXst9z9TmfbKM/tv199j7PP+N/3OPv/veGbAID//DXnqDv5x6zt88BVn3X2X4CFYYvxg+9fDAD49xu+65bd/RDbxheX/MLZUXOfs0csM2DlHdGT9e6bfRlK4E+nuO6qHzBEv3vCo5s+JuzJuYEWkS50kPf5+etYHHTjdykb93FqRTSeyrY/tNHLikvHeNZGJAjXigyuwUa2iWCCbVgqPoUmGj2Z98mRNH238luU7Y6+ht/L1nD/kT6ZATp3kdl5cq47OY9dQaLLimKOQfF5SnTWsnimzBAbszzvRjlvXirDP3Kt47O/x/E8kYFb0T8AACAASURBVOsmonvEc+7IYg6diRwV7brYT+MhStmxleP9KX1vDvSM+7zcaARKURRFURSlRPQBSlEURVEUpUTmhoTnI3Mcy0sXbHP2s4PMOinMezecZrZdbZih89Gs/zxJdWIdKScF83MfDQoJL3OWqGwnKFqsskKF1VJNYey/yitEKYtFrqyn7DCQ9s7rzm5mGbVGKc+9cPFuZ0eDPL8hkc0YFvHZ4WxeFhG5SHuyXPdlrVucvXlkobNHxTqPDS939up4FwCgRRTJlIVT02KisJOizPaLxZjRUS8mYtqR9uQ8mZn5uxHKijOGMTAhr62ZcGjM8gK5Ye+crrvnJresYW2/s09t4/n0XMDl+/fw/Lfe8BVnX7bQu4Zf2/OfbtkN734/t3EFfZy4kpl1sZ8xsy5zz1Jn33eiV7DzUJY+vhuU8F71IW47Mkh/fuaLzB4czB0/w1FK54FUeQvlmWD+fVLc73LeuWHRJms76fPuy3jMoSCP2fgoOgExP15Pn9SQ5IGIYpsBmeGXP44D/F4oIiSrbTudHTTMaJXFPSP+XVV1MYk+MVBHST4XlJmKYp7TeP7iyTkHRZcsr62cSzWY87/+BTlPyk/JBH8/bOL5IY/OBIXs93ALh7+YHK9ZJu5fxDQ4IobaiDloTT77PiCuX7ZHJTxFURRFUZQ5hz5AKYqiKIqilMickPCk5GGTIlvlBIapnx3kHGgJIcuZfDg8LELrUk6StpTqciLuWxtkLL6QhTeYZrbAOYv3ObtoULEK5kMKpoD6vZ4fZCHQQkFDADijzpsTTUphch65GsMMpVMjXL4zw+KlBzLN4/a9NNTrbCmhSbnvpiYWVNyR5vU8kK1z9sG0t+19aWZByTnvutMskvm5/a9w9p4+HlNBhgWAvi5v/ebH2MbaNkid4xnMCNbCpr3zLvxfjDU3Muvw1r2/d/alD7OwZOgja539N8v43V8f5Nx1o7/27o+bxOdnbOT8hf+279XOvv7Tdzv7RV/c7uyeHLd34Xf/FgCw6iM8Ppth22kAsxf73/JCZ68N895bLxLcgvl7Lyi0rhFRuTB42H9+vtkil/QOLiDaR7qOx945wrYV+u0fnd3xoZXOPrif7TJcly/kKHtZ0Q3kUqKAbYr3o1/mF0AJL5AU/VTEvy/pH6VUasSce6l6v7XnIPNFRVAp0Ul/xQuSrPxcyqNiuVhHLveT8KRkmBHXMFA7cRZesbn9nm8MD3vtLxTmOcp5BXP+08eOCemcUMPM+uciXqVsW2Uhnyo7HEVRFEVRlOpnTkSgUKScf+8LOBD6lAhjP52j4weGR8SAtIx4jJV2Mkt3ZHNcHg5we00xrzZNSrxWDqb4thds5feyRxh1kdOlVOrNI9A3jNofrQcA/DL+Yre86yX0TajXO6+xb2F8PZNTS2Tq+bo3byXPdU0zr8XyGm+A+oNpFst5qpeDxbs2iCljesVM26PcZ+0h7qf2Z96b/1gf8rVaTlFjU13OXmAPwY+2/Pomwlcis5THh3yU0+wt9so0OUw4hNB871w7r2bkdOzgYe+/+Y8zovb2bYudPb+RUbmGl3OamuwnuInVt7NWVHSNF8FZJHZxOMlo3vdX/crZDyfZPq96mNtY8Wecm2ilfXjcIRfjyOUcPJqwWWHzXqoPJHAsw5Z+tiKqXA6CrV70KBD0L1Kzr481lxaD0dLWuKgl18c+yQ06Ft1XTgzojtfz/FNd/u3LBsTUMPkB6GPuwbj/1bByIHyY55OLzrGBzkUi95l5/gPwZb+VqvO+GygSGLL+45jHIAeXF2pMpWvlVDNiOpgId1SoKQaMjThPZmqa5wPZo959Hm3nuafFJYv2HfsNj3lP8mJGr6Y/+1d6v0uiXF9VoBEoRVEURVGUEtEHKEVRFEVRlBKZExJeMcmrdy1DqbJ+zJjv5uO0RnxeF2YcMC4GiB8VMl9ODCStCTEMeTQV974X4vdaY5Rc+joYwoeQ8KqNxm9xwG/T90SJ/XVrAAAHX0q5IjGP34t30efzfynCs3Vc/8hBbm//Iq+mUqyLMkfdEP1fF2QNqlQb5aVAmv4fWUBZrvf6c71jauVxRAalzMFjleH8dAPXz3JzSLTmB/qK6TassFsf8c4l/VP/emGTJd0UxaHX5Ad1v/tOt3xBiIOwF+XteqE5PJaghHdqlNPhDFv6+OwD1Bne18njPL3WS254+0HKl1/qX+LsM9e/1dnt/0UZeoUYIC0JxLx1cglKT4FaUY9nmPfBBSt3OLs5EHf2oiCnIWnMn6cUzFaF2QY+i/JiamvGLQvXs60mE/4yW0QUpxkj18Xy7VKcoBzkHBOSTyIs618JiUgqWIUB46Kry9X5S0Ijw2zkJiLa9iRkq4oziYSbdININhDTs0j/FqS20IgYOC4GgMsaW3JweTEKg6DlYGh5baX0GziBmRvZLaxRWA1DOcpBIJmfRkdInFbIzdFd/v5u2Mz+MCkSShLzvfVjPdXVgCeMQBljlhhj7jPGbDHGbDLGvCe/vMUY8xtjzLb8/+NTr5Rpk7Aj+KN9AA/ZXwPAKer/8pHp7ce2n38Zm3/4KST6DkF9X15Gs0N49OjP8Pu+HwDa9suO9j2VQ/p+GEe176lSJiPhZQB8wFp7MoAXAPhrY8xaADcDuNdauxrAvfm/lRnGwGA1TsMF5jIA2AL1f9kwgQAWnf8arL3mw4g2tgHq+7ISMAGcVPtCXNT8RkDbftnRvqdySN/XoAFQ31clE0p41tpOAJ15e9AYswVeYs9VAC7Or/ZNAPcD+PCsHGURYmKai2ExlYjMuEPeDkl5TsSx0zn/rJSkmGUbInJfqBsVErF6ue0jZ/OFoFmUEJpq9kXUxBGFk0By8DqyGfW/DCXbJzYDABaw/A9CHcyUS53Q4excmM/fo/Ppr2CKUsjQwlB+XUo9wSQ/NxmGvTN1oqZKUoTlRbS3pts71vp9UhcRZlbKeVxnVMiAuRC/MO9p77rU7GRbkiF3j30I21EkpuH7UM8w5n3Vy2L70VfbxCe0A+u8TMVsg5As1wpfLhHtVmRA5ppFmlGa1+T/Dp0DALj9Uba9+K8ec/aizKbjHTKAsdlEudQkZpvPs/W/WKfq5CuXO9vuFZJfXpKSWZ7hAR7/8r6HEUIAWfQCs9T2JbbWu88CIvOtqW7U2f0b5/t+b0GMdcOelNvLy3JGSnKiu5lfS8mzt4YSuJwSZMys9fnNFKsrJckNUP4INPC6Zeq4wULGqk36pzaVo+/xZRJ18rJxtpOAkPCyMS4vKECRYVHrT9z7shbRmKxj4VP5U5KpyQ8HEX0MRI0t2W4GT2I9sBrOVjXp3wHpe+Nd5Kr53Z0MNuxzDUVbllMhjVll87F9r0e62fNb9Eh1jToqaRC5MWY5gDMBrAfQnn+4KjxktRX/pjJDRKD+rwg5ZAH1fSXRtl9Z1P8VQvue6mXSj3PGmDoAPwLwXmvtgDGTG8xljLkRwI0AEMP4QZrK5MjYDACsAvAW9X95ydgMRjEMaNuvCNr2K4v6v3Jo31PdTOoByhgThvfw9G1r7Y/zi7uMMR3W2k5jTAeAbr/vWmtvAXALADSYlsnPYRKQU2X7hz3fuJLTUjwmsoukpFaY1iUEMZWLnPFdxMhlpl40RFlrSEzbYnyy/XIi/tvzAu6n+ZtipWlM35KzOTyFhwGgt2z+F2Q6mcUVkLZYpwn+tBZZ7sf08tyOT+0EnxcLrBd8H0YECTsyq77PPfMsgLGqTOtDwp5oA5Og1AYw0bQzMvNO0vjtR4Rd4k7l9svY9rMNXpahEfJ8Wy2LeYYf978CrRFR8NMnpi/lITliQBa7RERK0iJTS/TQoUFv/cx8sW7W/wc11s0vphspl+Rktl+DNzXN8Wa1r3TfU4xULZ0a6+P1ykbHT4kTHhJS/jx5AYQp3JiNGt91ChJUuklk72VEIWZRwLl/Ffcz5vGlhN+BcvY9M01onid9RyP8HR2uEW0v7V+sVv7W1wWZ7Rtu8mwbqhv3lUoymSw8A+BWAFustZ8TH90F4Ia8fQOAO4/9rjJ9rLXYjI2o9aptd4mP1P+zjPR9BDH5kfq+DGjbryzq/8qhfc/cYDJjoC4E8BYAlxhjnsj/exWATwJ4uTFmG4CX5/9WZpijOIJD2Is+b5riter/8iF9P4wBqO/Li7b9yqL+rxza98wNJpOF93sUzfXApTN7OHLHRUJ8gnsOcX61pfUsWtkvssBGM15YNSYKZtaHmHUis/diQYYbj4r57aJiRunCfHkZkU4T9pnTaxyTkCT9aDLz8DK8AQBwj71js7X2HPHx7PlfGeP79fZeDNjeM8TH6vtZptxtv1BYMZ3mvVon+orMVmZqop1jdx/vF7KRT09pY7zfzTC7XDm3HlL+28iKbMvQYDB/nJRCYgf9u/DafSIj7GTu3+aE/BSL4nhUqu8xIZ6TzBAuFHIFmBEHAOYwzzUtlvvNmyZqMyIgi+6KAptyuSyMWpD20lJFkkVNEzzu4XXTm7Rtrvc9TfWehBcUBakHo3RWYj5/d4uN0OpONzi7rsbz52hkjkl4iqIoiqIoylj0AUpRFEVRFKVEqqsqlWQS2Qp79nOStjNO3+/swwmG+Qrz1MnMu8GMf+i6IPcBY+e6k3PhDae90KMsxtkYZrG9+H5/lxox79Ik1ElFUcpNPkU8FuP9Xivu/SNbtjs7uIpznfXy9h9T+FIWCC1QKB4KAA01lP5HhZBhQ3LuNlkk0rPDg3zvTTf6dybN27ntoctEIUk552P8+BJepShWbDIg5xkVjCnoO492/PB434ypm1zkJ2ZMUU3RnWfyqbwBkZhq0kJvrSuxYy+UJJhGhna1MpyfN7K5ljeHlJ5TtUWyFAWyyHVNft7IYb8CnRVEI1CKoiiKoiglog9QiqIoiqIoJVK9Ep6gWFbGwl9xefPZI84eScvMOi/0Vxf2z4rIiDChlO0GRfFMKf8V1pFy3/LYYWcvu7vP2WMCukEROxbnoChKdVCQgoaHmO0VlpOhiezZobUcPrC4ntLegQYuR9Rb3wz6d7PJNJdLKSiYFBlmon/K5g9LZollG/zlrvDjO5ydyazk9pLinTnkPw9otZJewKxFKc/lwiKDTkyxViiwmWjmeabrxbpCissKNVPKfHKdQlaf3IfcXnZI/E6FeXzBk1dznXHzbD4/SeYlvHDDkO/ndhKhmyFxUeoinvN7k5OrxF4uNAKlKIqiKIpSInMiAmVz/gPH6n7A6SJ2vI9vfufP3+3sTUc7AACJLN8I+pIctpbO8nVDDgwfTTPCVFvP15CmiDco7ro27vtrB17CbTwppt6W51BkxnNFUaoDkx/MK6cb29y3wNlR7Ha2FUkhJ9axSPf6CKM9za3e23e8gyGLhIg6LWk46uxdou+RvV0qxfVTI/kEFvF5qCsCP6wYmDzmnV0MUE+3eP3gXHmLTs5jRCLZyKNu2MO+tV7U6spF8jX74kwqmvc0r0Vo1F8JMBkx2Llx/ARTsmZUJs7rkxSRsGUnsU0cOXeRs5tEBMqEvG1PNF3SXCS40wuXRjvoYyNqkMWOTlwLsT8dd3ZbfBAAsCuqg8gVRVEURVHmNPoApSiKoiiKUiJzQsKbzNQnXe9hXZa/+8FXnf1A1BvA96IaDvT8Y2IJN13kGXJ15JCznxbr39R0AABwysPXu2WLX7/J/6CmOH2Loijlp2+1JzvMb+EE93KKKDmbbs1P1jv7kZ9Q5lnbwXp02YWtAIBMQ61bFmxkl9sLDopuSVCaCCbYV2Sjoq7Rfk/GsHsPumW5wUHfczHtHNJwzcmPOfu761/g7OFF3vnW+26h+qh9lsk6hz9JeWcgK3wUpUz0jtUPAgBWRXg9Hxw80dk9KZ55V4L2ifW80p9Z8LizfzrsSYErwzyOtBgN/buRNc7+z42XOHv17Q/7no/NpH2XPx+o2+f93xJlcpesb1a7fcDZxapn1YkabKH8nDpBHUSuKIqiKIoyt9EHKEVRFEVRlBKZGxLeZHj0aWe+d/kFzj7y9hcCAD7HUhyIrGH4sKmGpeZHRYZM72GGdBsfY/bHnbd5+1k8WES2k6hspyhzhnmPe/3CgWZOGbK5r83ZrfCXYiSZTkr/yNuy2lIcpSG/W0pvkt2+y9l/PJPvyWvwaIlHUD1kt+109uLXc3lo0UJn91+41NlfXXwlAGBwNTPBWhYzS++UebxW7TFKoU/3c3snPHu6swO7vatXe4Ay0sK79zk7s4f2alA2LcrzcAqXAk3bPfntsQOL3bJwH1uzfW7nuO8cy2+2neTsExd6smrtwerymUagFEVRFEVRSkQfoBRFURRFUUrE2DKGEY0xPQCGARyeaN05zjzMzjkus9bOn+qX1f/TYiZ8vwez1zaqiWr1v7b9qaP+nxzV2va175k6Rf1f1gcoADDGbLTWnlPWnZaZaj7Haj62maKaz7Gaj22mqNZzrNbjmkmq+Ryr+dhmimo+x2o+tpmi3OeoEp6iKIqiKEqJ6AOUoiiKoihKiVTiAeqWCuyz3FTzOVbzsc0U1XyO1XxsM0W1nmO1HtdMUs3nWM3HNlNU8zlW87HNFGU9x7KPgVIURVEURZnrqISnKIqiKIpSImV9gDLGXG6Mec4Ys90Yc3M59z1bGGOWGGPuM8ZsMcZsMsa8J7+8xRjzG2PMtvz/zRU+TvV9ZY9V/V/ZY1X/V+441feVPVb1/2xhrS3LP3izEuwAsBJABMCTANaWa/+zeF4dAM7K2/UAtgJYC+DTAG7OL78ZwKcqeIzq+8oep/pf/f8n6X/1vbb957P/yxmBOg/AdmvtTmttCsD3AFxVxv3PCtbaTmvtY3l7EMAWAIvgnds386t9E8DVlTlCAOr7SvoeUP+r/2eBOeJ/9X1lUf/PIuV8gFoEYJ/4e39+2fMGY8xyAGcCWA+g3VrbCXgXG0Bb8W/OOur7yqL+ryzq/8qhvq8s6v9ZpJwPUMZn2fMmBdAYUwfgRwDea60dqPTxHIP6vrKo/yuL+r9yqO8ri/p/FinnA9R+AEvE34sBHCzj/mcNY0wY3kX8trX2x/nFXcaYjvznHQC6K3V8UN9X0veA+l/9P0vMAf+r7yuL+n8WKecD1AYAq40xK4wxEQDXArirjPufFYwxBsCtALZYaz8nProLwA15+wYAd5b72ATq+8qi/q8s6v/Kob6vLOr/2aTMI+dfBW+0/A4Af1/Ofc/iOV0ELyT6FIAn8v9eBaAVwL0AtuX/b6nwcarv1f/qf/W/+v5PyPfq/9n1v1YiVxRFURRFKRGtRK4oiqIoilIi+gClKIqiKIpSIvoApSiKoiiKUiL6AKUoiqIoilIi+gClKIqiKIpSIvoApSiKoiiKUiL6AKUoiqIoilIi+gClKIqiKIpSIvoApSiKoiiKUiL6AKUoiqIoilIi+gClKIqiKIpSIvoApSiKoiiKUiL6AKUoiqIoilIi+gClKIqiKIpSIvoApSiKoiiKUiLTeoAyxlxujHnOGLPdGHPzTB2UMjnU/5VF/V851PeVRf1fOdT31YOx1k7ti8YEAWwF8HIA+wFsAHCdtXbzzB2eUgz1f2VR/1cO9X1lUf9XDvV9dRGaxnfPA7DdWrsTAIwx3wNwFYCiFzJiojaG2mns8k+bQfQdttbOz/85N/xfEwcAmEzWLbKpVEmbMEERKA14tk1npn9sJZDAMFI2acSikvxfLW0/11Tj7EwLX57GvkfxNKO7RspwVBNTFW3fiMs/mRfPOq/tp9rEslG25aC8DeTmxG6yjTkuHvK+G+4ensSxFtn2FKmY//P9B4Cx5ySwAX5gcjNwshNR5DiKIo9pJFHy7o7pe+ZGv/884pi2P4bpPEAtArBP/L0fwPnHrmSMuRHAjQAQQw3ON5dOY5d/2txj79gj/pwd/xuf3mGKUUoAMOtOBQAEDw+4ZZlde4qt7kuwroHbq/UeADKdh6Z8TCX/EAJYb+89dtGE/i9b2y/hfEZfep6ze64fdXYqya7A5ri91W99bPxGAkGxv5ywZ+/HqyxtfwJMOOJsm574JcCecQYAYO97+PIQeLre2fV76K9AmnY2Qv/3v5IPS/GH6wAAC77w0MTHGhLXMzP9l40Z9b9fHwMAJv9wmaO/Cv0HANggvycfmrJxnmto0Oe6BEp94iFjHs4yXlu3YbZ/KzZtRPOXy4MJ+t9ufGb8Tia4f4/pe/R3t8wc0/bHMJ0HKL9WOe7qW2tvAXALADSYljK8HowlUO91WLnBQbcstGKZs3dft8jZubD4Xpp2407+SNR/7xGfnYgfFHHzzzKz4/8JfgSD7XydHj2Tfuw9mc5L14mDzPcdl79hk1sWDdCf9xw8kd/L0I9ntB9w9jvb2YH81cffDQAY6VjpltUe5DE3b2HExDz8pP9JzMwP/YT+n9W2X8JDU2DdSc6++uP3OHtltNvZ3Rk+pN7YeNDZt2xZ6OwfnZy/9kXa+Ez/aB+H2e17ivzAF3to6n/LC2mv4fLLXrkRAPB/Cze4ZX84h20/bOjHWkN/nRIRURfBB9ecCQC44/Rz3LKWR3jftX37KWfnhicRpZo60/O/KTL0Nt+ukq8+1y26+Qu3O/u2rgudHTL0YzzIzjqZYx8SFuv4ERBPPDnrf81DAV6jvpT38hYRy+S+5Tbk91rDvBa//PpFzm77kvcgbILigWzi+2ZO/O7+qTCdQeT7ASwRfy8GcLDIusrMo/6vLOr/yqG+ryzq/8qhvq8ipvMAtQHAamPMCmNMBMC1AO6amcNSJoH6v7Ko/yuH+r6yqP8rh/q+ipiyhGetzRhj3gXg1wCCAL5hrd00wddmDxF2D0Sjzi5Id/v+4QK37L3X/9TZ3/5gu7Nr1+/i9hqpQ2395yZnp9/sxehbrtjKdcsn2zlmzf8FOVKcU+oySgZd53EcSOQovxbpZ5S4tpN2ww4vfP30/RzLkP3XPmef0sqxTKfWUbb7+aF1zv7oO9/m7MYWT0YJJShdJFr4HrD/EjFY8lJe8xXf4UtaZuduTJeKtP9JyHaJKzjGad81nhzQ2jrklrWEaD88dIKz94y0OPtwmuN0do3Oc/bWr64FANTtoO8XfprjcWZZtuN+Ztv3UmKS43HOYZtM/juHBHTvp7RnArwudz98FgBg56n0YVOE486k5COlIMmOAX5375YF3j6auL++C4R9EeXw1vvZB7Z842EeX4njuPyYtv/t8aW1PVfSL/+1/xJndw2xXc6vZTve09vs7GSSbbMg0VnhZyNlu5wY0B/idc5muVzqZdm0tzwQ8r/3cml+L1or2oTYZ9tV7IcCt48fXjLRPV51v7t/4kxnDBSstb8A8IsZOhalRNT/lUX9XznU95VF/V851PfVg1YiVxRFURRFKZFpRaCqiTGyXYK1Nvre5mXIXHglM7JcNhGAGB519hghrqfHmauu5+KtX/fkrGEhCS75OGWMKYfIp5BaPyvkJQuZbTewnGHx9kfFOYljTtf5P4sn5scAALFOkR33sr3OPhSLOftggplgIXAdnHmKM7NRbz+RIcoAoYRI8KEKiEQTj6n7JR3ObpkBCa8iFGkX277ILOaFq9luV4e9a7WvjxL0wwOrnL2ulnLC4SQl6x0jLHnyeDezVMONSQBA3Ut73bK951KaXfRltpPgfT7lD+YKRST5597Jthp8iv4yohe1USFPRTx72x+Wc12x6Sy7rDF5VOFB3le5qPigJb/tASFTidJkRpSfOHwed9TyDbGbKcp2s4Zs0/nhA687d6Nb9IcukW0b4bEvquH4gX39bN8Sv8w6KeflMlLaE/2XWCcgpD3jIwnmhM8jNTy+lnpm3nV28/ia5lPC7X+xlx0b/TmzNCva9ysloxEoRVEURVGUEpnbESgRAZFRJ8kpN3mFyzZ863S3rB0iYiQiV0WrW4s30jXv8N6O9v6Qb96hjgXOlgUex0SjMmKQqN9bRpW9eQyft9zZUTFAPBsXAyzF23T8UNLZ4S4Wzcy25t/UQ/xe7qIznB0Y4ltbrp7XIiiWQxTQq33Wi7CklnDgaKpB1KCq4X5qeniAA0tEU3/BabQfYe2cuUTuJWc6u/2Ew84+eIh+aWvz3tLlINZHDi539smrO509mqUPU6KWTv9evj13nOD5vn+YdYoCoqZX3/s5sHfefZM8kSontHK5s2MNbOOpAVHzKi5uBBHVCI56fkyLauJWDDI3Wa4bHmC7Tc7n+rmYqESe37aMctmsqCEk+hAjjiP5StZViv5SRDsqxZhoD32Xe5F3X55Vx6Syu57jvbq8/Yizz6hnhPq+0dXcXM6/npPbR4r+amhhlGhZMxNbNu1jtDotCswWEgSsuG5WDDi3YZ5Lc4yRpoNZ3pP9Sd473Wd699ySnx/3kJUqRiNQiqIoiqIoJaIPUIqiKIqiKCVSfgnPp87QlCkie+3/Ow7wflHNbwEAXT/Y7pbJPUvZzgRkaFas5TNVS6KHodhnP7Tc2Se8jxLemAGbcpB4tQwYPw7JRj5b13ZSfhxpp9RTt4+SRtd5nKT23//qe87+xEfe5m1jH0Pa536RA4y/9xCnwVi2RsifH6N01PNh7udbp38XAPDut7/LLYv08/jStZQB5Zxi0aOUPZKtXEeO451LHD6V7S+ToH9MkOeZyk+Ns2oepY9ndnFQuKz3tLKWMuCduyhPy8HNqYzXXSRGKU3nkrw3knG2DVYvmtsMnM46cYle+lYocQgO0gdy0Hc27q0vpTqJDXLddJMYfC7McN/4edcyTaKWkZDzwv1iGpOlrC2UauR1ror2XqQO1IEXeW16T5KtJyDac3uc57Q8wmQJKdvJOlyF5WNkPSFtDg8zKeCKEynly3tETnBnM4WJzEXcQVzbdJA/p3KaGMmRYfaTiROSvusocweNQCmKoiiKopSIPkApiqIoiqKUSPklvIJ0V0TGKszoLiU0E6FkILFJhkCH3vgCZ19wNWs+NQY96Wj43OVuWexuzkIvpOrXdwAAIABJREFUpcSiMwzY8XLjmtsoSd34rTud/eHsnzl71d8+wi+IzBM5+7YJestzKZGlV4GpYQpkLjkbAJCq5/Wp3yMygcShhbsZUh/pYDh8fpDLj7zBy3Tp3s26OcNdnEIkJLKP9nVzOpG2pRQbhllCCu/e/iYAwM5r+b21H2Xxp2TLUmdbkb0XHeA5jMxjs68KSWMKHF3N8xGT1yAoMoH6+71P0iJTS9aquWMrsyE/eOpvnD2yh/W4Am3Mbu3t87Znj4r7UUxrEawT7WQGpgypBvpX0XfRZjbE9DC9LjPebIPot0a874YHRF2hNO1sTEh4rSID2PqvU6j5JGW9dLOYgiQursVTjc5ONnI5xbwKUiQLr/HCLgBjpw+S8lttiP19UGjLtXVso5kMfZNJe7YJCz+L7WWF/PzMMGU7ExJSrZAEC9O9JIf8ew0p7WXENDFSVkynuc/29n5vHzWU9XIjorNTqh6NQCmKoiiKopSIPkApiqIoiqKUSPklvIJ0J8O48NHOhKwnpbpiHHwpt3FNHSWduw95GUWxux8d9x2AkiFwnNnkfeTGoeUM4T81usTZgUWU9sYwRioUts8k7JM6plki2eLtu34/jzFTw7DzwAox47iYQqFpK7fxpvtvcvb5a3YBACJLeB6PPMCpWdILKO/EIlyn69Xc3okLmHGzZedCAMDCJcwsy7WxUN3AMvqueSudm42JTJ3qTHwsieB8yhayUGZdDe+VgbyMMNRHiSBWL2QQkeH0P3uYuZqLiPtxVHQRKW974UG2gTHSkzy+JQudnZmrU+cAGD6NfrYDlG7CYgqVTIPwlyykmfDsTL3oy4TPAykxDUiPKNgoM7vEtjPNnq9l1l9gRBRyFBmAubUsatryYhaJxFdReYoMUbhhmTfk4dadF7plctqUjCjwek6U0wmd0c7+/ulutrtsvshlNCqydEeZKSoz6PrTzGqVkpu8twpyXiAiin+O+v+EtkTF1FUi2zKT5vqtcW+d9Gkc0jBXC/v+qaIRKEVRFEVRlBLRByhFURRFUZQSKb+EV5DmfDLbACFZSdnsfBb2O3oC5YihxXz+i3Vx9f/ccCm3lw/TBv/fYrcstJRzIBWjJkZpqbWW4dhAPvtjcJCzgX9nE+eayogMpe2fZ2bgfDEFVU0XQ8qxP+4EAGT7GGYvt2wnCaS98zNZhp1TjWwmIZEkko3R/7JoZePjlDoeHfHmqYp2i+yYNl77+W2cN0+G6z9w2t3O/rsH3+Ds2AEvBH8wQNkudwrD8n/zzh87+4d/xnaQWMh2E8jw3IZffz4AoPZH6zGXiAupribC9jRGcsjbcZE9NrKLGXZnnsvisn/csoLbqGX7swlet1BeOpKZX8E6IY8k2E4ybdwPdk50NtXL2qWcL3BrZ5uzMyOiuK6Q7SAyvrK1eflNSMayeGaOiavIRcRKIlMsNCwL8Hr7zIpsx+CweAcWGX7tTcyE3b6Vc7utwT5UFaJI8QU1OwAAnx14ue+q8aDoNw2/Nz9CuXI0yb6gIOHVRtmXD+ZEzqq4V3JWSKHC/8Eo23o65bXvkMh0TUkJT0iC0QDvoUiUduIo+8Zcvr8bWUj5kL2UUg5CK5Y5O72QvynmD09M6vsagVIURVEURSkRfYBSFEVRFEUpkcpl4RWZA67vBm9utMNn8fPwAlHALslwbF0DM9466hnG3fd7ZsUlF3hh35woSJfbzqKO4UGGXXMiQWOohuv31zLbzOYLBzYvooSXGaBsF5/PY/3rtQ84+wsLXurspjZmkHUOenNtDYtMvmVfEVk2DzyOchLr9qShvpMZTO49VWSltNHnzVvZfJp2MLw+9C765gMrHwYA/PBvL3fLUg08v4de+0Nnn/Dtdzr7o09d6+zgMmZCrfymVwT1ml8+7JZ9/wsvc/YPO8929v5/5HkFf8eLW9Mlsqb8pymreqRslxVF+05sYZHY7ufmAwCaT6ZMOtTE67qjlwULzaiQpBLcXlBkihUKQoZWsA3EhTzSv5f3SU5IWXP5LW3nvZQ2g6dRFkvL8xMZeVYm7OazGU1KFFUU8pCVE+rV+mfyZUXmWWE/RhRjzNWI74nNdT65gPtcIOZKrLICp+bsteKvPwAYK+XnhCzWEGK7y4qTTVvhj+x4Ka6Q7QYAXQG2Udkwh9KU1gpz3gFApI7yW6EIZkRkC6eCoqisKKQpi37WxWknjlCuK9y3fWt4/HNOwvOZJ7ZUirXJnZ/iPKkdZ3jzpMYv2zWlfQDwPdb9V7GA6mgb29SKP0xukxM+QBljvgHgCgDd1tp1+WUtAL4PYDmA3QDeaK3tK7YNZepsshtxGJ2IiJrZ6v/yUfB/FnKckfq/HGjbryzq/8qifU/1M5mXw9sAXH7MspsB3GutXQ3g3vzfyiywEMtwJi46drH6v0yo/yuH+r6yqP8ri/q/+pkwAmWtfdAYs/yYxVcBuDhvfxPA/QA+POHejIEJeVKKDNWZM1lYccGfeyG6nh3MmjPPMXOiVswrFe/mm9H+5RxBHziNkkVwlzf7kwx1yyyitIjoGhGChciWseHxkk9fD2eVal/Kom4ntzAd8LO/fRU3N8RtbxPzNRXmcbru5D+6Zd+97MXOXvHAfIzacVmDU/P/JDAPe/MItlAhQ1sH5YDcAGWMQL2Y3+5szkHXEKPk9vm7r/C2UUd/dl5Mf/5oiNlaDSfSj4kNrc4+fxWzxZ65eB0A4EvbL3bLmhrZDjKfYsZRo8gebPodU8EynYcwWZpNef1/PIKtnCswGKAPB0aYzvWKlmec/VDgRADAgW428kgt7zs5R15oVMhGYqqvnJhfDXllVsp2zTWUVfpzUuqevjZaDb4fXSgyYof85+SUBTFztfRX6IjX/gJFlA05L56U36T8mZonijbmL7OcC29MCeKsHI7AbTQ2iqKOEUrZE0l45fB/38nsQ3495P0OhMJCNrO8h9fEed8Ghfbem+LvQ0AUKs2kPD81hEXR2ZBP0WYAoxnhFyGzRsWxjBrv+sv59mTRTZHIh4YQ9xkJ+WdVp/Ly7Gib/zFVQ/ufkCnKdsETWTx025/Pd/aYDFOOysG6Fi8b9pe3MON9zU2PlXYcPutk4j7rlcBUhye0W2s7ASD/f9sE6yszi/q/sqj/K4f6vrKo/yuL+r+KmPVB5MaYGwHcCAAx1Pi+9djHNzl78wavdtK8k0Q0YvO8cd8BgJF2/7fcxIgYlNbi6cdG1m0Rb2cy6mSjYjb5JJcHxBtHrj7/NiHeUo48zSfoBzpE/ZsGateZCLcRDvFJOJZ/w9nQy3oUq77fz/1hehzr/6lQLGJjIvRzeIhvWYeHuZ+mtd6A+cR2RpT++WLWavr0x6939uv+9j5nfzfNweD3b1nDnZ7r+W6ReKtLNvPtMX4XI3kR8cZRicpaM+H7MdsTEb+gqGGTEYNeD2fY/tpXHwYAHDrI6GxKrJsJs00GZUMrUreogIw61YfFNEviFrNm+hGo6TIj/o/Kwd3ifVNOdyMGMcvaQoWEk2zQ3xe5kHS02IRYJ3JY1E/LR7fSHexDA/1s+7mY/0D0lIyYLOVUJ9iyzfe4ZorJ+L//RNqHko0AgLiceiXFn6gFISanJEQdwZ5R3hdjD2D8orAYAJ4VfklmxU+hcGNAXM9CtCkQkIqE+FxEIhdH+PtVjKOjXkgx1+wzn9c0mem+ZzKEFnNAtq31QjuDa9nvD7fz/jl6okjqamJ7XrKQfhv4OdWEDV8+EwAQfzUVkF3foXJl9zAK2f4or0/dToaxAjtEDbTFnqoSFN1X5CivX2DdSfzgaRRlqhGoLmNMBwDk/+8utqK19hZr7TnW2nPCYjCiMi3U/5VlUv5X388K2vYri/q/smjfU0VM9QHqLgA35O0bANw5M4ejTBL1f2VR/1cO9X1lUf9XFvV/FTGZMgbfhTdobZ4xZj+AjwH4JIAfGGPeDmAvgGsmvUefOlAy9Bfr8Z7pXvxSDhy+s4ODZyN9okYNx+khLAacpXopLa05bfzUBb2jDGn2HGp0duComBFdRGmzTQz7FmbilrVKsjVCBkzIWhMihiwGG67t4EDzQH5HkSBD0ke3M1T9tF2PPvQgjSQAnJb3+dT9P1mkBGPEc7aQxXIjHJw6uJhvOQM7aEePeN/975u/7JYdTFNSWnbjVmcHhI7xNyff7+zb/+VKZ3de6u1/XpyDK4eOcvBysIV29jDrbU21XknB/znkMCPtfxqkO+i3rGUo24o6ULc+x9oprXXe9QnGhWwhphoKxoSwKZuqkH+skLIz+QHlO/dw2MWZq/fwi0KGGjP9j//pTEil2n6glnJAuEYMuu/nYH0TLiKuiyEBhbpyVvjfCOnPikHfcgB4+IjwmFRQ83bwiJDtokJCyvoPSh/p4flk6yf/zlwO/2dEjbeB/IjeGpGkMDRMn9cHKB1LUqJWlh9yCpig0KqlhCeR1zbmMwA8FByfVASM9X+9+HGSfbskk79vIzX+El5Z+56Ajy+K9JPBZvZDe//yZGcnWsfL/dlG+q+pnb9rYTHlTngjE7L6n6ZsN3i6GO5TmJJtD+XabL04vjaue/AS8Tt+EbcdHqTkl1ySXz/H752y6oCz975CZJddhaJMJgvvuiIfXVpkuTKDnGrOd/Y99o6nrLW35v9U/5eBgv/X23sxYHsXi4/U/7OMtv3Kov6vLNr3VD9zuUiwoiiKoihKRSj/VC4+U7iklzGLbeHvPGmm97WitsciSkWjtQzpNm7yP3wp823d7D242xqG+1Yu5bi72qU9zt7fIySSHu4nMMj92GHPDiZEzRURRpc1ROraqSsuamT4MhZiyLYx7IWlF0RZu2pDRGSmDo+rAzK7+GVPFQnl2jTDs6EkffDwNZ9z9iVf+iAA4L3PvInfE/Jnyxd4nTf+E6/bE7s4tU1zA9ePHfBCv0/mWHdquZzJPltEWplivZJqIt0Q9l0eFFmdZgMl6Z6zPX/KTKKckOdkRtJoA/0WEDWhxkjZ+emQQj08jpEVIuNVZJXlwnP33czEJh50K6VNI7s0KdHl3RjuZf8h6zONkdwE6RYxZKCWts3LryHRv8kpqmS9OiMzBsXxpRtE5qzv3svLKUs6nT2a9Y4oLLLcIiIjrz5AuUXe5YkM/SslusIQipAoxFUvplXpHmQfP5AQ1dZ9Mk8B9u2yNtSwqCuVE9/Lij6uNnz8elux6Mxn4ZXMBP1j4orznN15EeW+bNS/vw12eL/ZuR4WWgr8nL+vrX383sGXC4lVDKNZLn6bCz6MCTm2e4Ty3KC4fvULeY2X1rNI+4Fh9o17nvWy8BqX8nd5foy/15v6RLbqcZi7vZyiKIqiKEqF0AcoRVEURVGUEim/hJfn8F8yW0iWwDf5SOKKAMOkL1mxw9kf6/i1s9+/jsPj//joamfLaSnqd3jhxkCSYcd9h5j1l24TEzUKCcLWi9C5nE095h2gaef32hops71uyRPOjhlu4zeHma2wpaed287//5GTf+WW/exaTuUy/ytiTpVy4iO1jkOEfftW078HRUh9yZe8KmTPrmJhsq9f8g1n3/z+1zv7dfOedfboKynbNv8fZYeHt6wCAKz9Zxb3PHo+r2c8WyQULaXJyZxbFZJqEAURxXIpW8iEJJPXlqRkKmWjpMiECQhJWtaFzIkQfXjA+yArJNORNK+NEdlm1vhPezInyE3cPoyULsTqgahofwOhYz9GTvS4gTHzt9D/wQGx0qAo4pufjiq1gH2PHF4QEBnA2WaR+SeW5yLV9c58dtNeZx/MF9KsEZJXNExpJmboW3kWybT/z1jhNj+apoy0spGZuV0HmWmVSPFekPeTMePbQm2Ex9cHDkEwQh7vFJnGUkKX2diF+zIW4fWUGemZ/cwKm3XOOxUAsPP1IstNTH+2au1BZweO0G+ZLma0y6Ez4U5vOwExhVf9/oxYlz5cvoKFoxfVUlLrT/G6vbb9cQDA3hQLc751wUPOlkVWbzv8Imf/+ln+7sY2c3vNF3hFhs9fwPb3+wMreI5Fpvw5luq6mxRFURRFUeYA+gClKIqiKIpSIhWT8AYYLcOar+x39r43eNlX/7HwAbfsrmFKXncMruM2UsyiqFnBLLa2elFVM0/XAEfsZwYZFq6t54h9mSnXHKOEdGIdC1+6z0OU7U6NsVjnJ3a92tn2k8wu/NdbvubsZ9s4wv/XR7ziXtuTPMfIQBVITJMopBmI0f85kbj0rudYOqzzn72Mwq9fwvP/v4FTnX24i3O33RE9y9ld72co+9knGfptWOgVkEwtp29lpphNHT/jZS6TrhUhcrlczBeWa6EzavPZeclR//nSZOm8SD+3nWoQ7U/MVp7Jh/RDw2JOq1FRXFJkIeUq1rPMACEefCZVRB6Sc2iK2yMgJJrggPdBqllkSbawfWaHhIQ6wo3IIQgQ8msukrdFAcgxmY9xmW4njk/IEZk4l1fDBB/LooedvWPEm/M0JLLwZMZbvZDCjop7Pp0VczoGx0svcq6881t3O/vRGs4/msv5Z6fKYynMhSeXSXk8kKZ9NMMbJySkR6m9F/ZZE6aEJ/u1wCxLeDs+w2E0oXyN0pZn+HmYP4EYeoj98byUKII9KgqTigbVc6bXzoQbEEhx3aGl/CB7O2XAh85hIc26Zfw9/o9er/RVYpRDA36w9SXOrt8t5gYV+4yI+XJXX86hQM1R7+Tu33OCW7aomfvrzE0utqQRKEVRFEVRlBLRByhFURRFUZQSqVigvW4fQ2uZPZTAQi/xwnlv2Ha1W7atk4Uls72ME8rCcfWiaOXOXZTDHEUG1c/roBS3romZBjuGGEr91lMsIlZX781xlBSh/VesZPbYvg0Mda64hxl079vCQpItccZGk1lvO/sizNq47EO/c/Yj3y5zuTvfDLWJMxJEJBsH9jFTYuVpnk//7tnXuWX1UcqmIZG5NZTktT3j8i3OfuSxNc4uZLR0ncs4bcdDvPa2WCHN5wEyRB4QIeZYnLLQYC1XyubXsVnxniTumfnNnE9PJIkCB2Pwwxay70RoPyuOQ6q+uZB/kcg5QTNlZZuUKYk0Y62c6yzbQ4koIwpYZuZ7bTs4KgpfirkIfRK8AACpViH5ifsjcDj/XSE35USmlBESko0LqX1IdvOVHx4QbKB/Y4ZzKSbyhTQzok01RNhXNAbou8cS3EY6zWtUE+O9UGiPg6JfaYtwqEeLyJ4eEHPuSRkwM6Z9e76Li0LIMktPyql7Rzl/6+IaZpk9Lut85redFBKkqWN/Pxt5rHv+hbKdHPrQ8QfvnIYWcv+pOp57IMNzG+7g8SabRYanKBJbkLXlXHgDy3hGdZ3j5xgEgGW/ENsI8r4qyH+JVh5fP38WMPBq/gac1M5C2bKI6pEEMyZ39nnX5+JlnHM3IjL/d3fxN+x4aARKURRFURSlRPQBSlEURVEUpUTKKuHZ+hpkzjsbAHD0JIbWxMxvuPkkr6Dkv295pVtWV8twefsCzo9z0TyOqu9LcyT/yjVc53Day75rDI2MWwYAsQDDsSfHKOHtGmYI7+2ns2DXs/mMwAFR5OuQCCe/5JKnnL33Xk6gfUotCz8+2c0svNZa77gyogLiT3ed5uwFoJRVbeQSvC7pVaPODgkJaNE6L3z99I+ZPfnB99/q7NvjFzi7I8YsiLU1vBYbh1iEszbqheg7z+L+wMtTHClNztGimlIWi4vMnVFRBFDqQk01no9GR8R8dQkRos/IPDxBkcKbBQlLFtKsj7MNyP1k/VXAOYGtEfOiRdhPhQ5zebpRzL8mpJDGRvYzgz1ewcFAUmbSibnwhOQWTMpCpuK9Ns3r5Ypjynn4RPFGG5P6kNhnLc9hdB6Pm4JGebFL2f8N5CihLIh58tqmBDOxWmOU2cKG7fVAmhLZRKRFO0/m6E+ZQSftaOj488I1hNnmA+Lip8W1fbyTQzn+6uQHnf0zn5BFRkh4CSGJzZiEVxODOcXrf5PzhbSb4MHsvyQvnzb+/+29eZQc13Xf/329L7PPYBkAQ2wESIAkxF0UtW+WRMmmF0mRraMwsWzmF0sn0k9KfqZ/OraTWPGiIymO48Q2HTmiE8YSLcoSLWoxCYmSKFIkIS4gQYAEQGyDdQAMZuvptV7+qOq638ZUYaYHM1094P2cgzMP1d1Vr269el19v+/eGywf24KMm9QZGrfTwVGI9bkieVo+N/kukdkKFK3KkatXr5DaiK/vke/393XsAgCsT4qsN+nIdThGCZSPVeX7/UxN3r+7KONux6gbgcnJM4t7JEHo2u+KdPwKwlEPlKIoiqIoSpO01ANV6TA49kb3V5wJWez7n168DQBQKMivvURSni75l8KTo+v8dn9afqnsnVw+Y3tfUn4SnyNvVYkS1ixPygLD3pR4OE6UxcOU9apBn3bkyZY5Upan2BU52d/hSVkkvrJTFu/WS2Hc1HXA3/b4t8QD1c6YNJVZ2CUeua5Dcm1373HzXA3sFXv+u7/8qN/Ojsivj109cm0fmZbtlz8q5Rcq33XLPSxbI8cu0aLbXIf8rq5RuQDEyNtil+hCc/qBN0hj6/jjVDl8QGxRqHum2MlGOYEmC3JPJF6Se6LcQ3luKLdTzPN8xMpUDobKaFhaQO3El+4iclMS714qS+VRYuSZIk9fNS9t9urVf51Xc7QwlrxBlvbh0GL1OOXZ4gXoZc/DFCMPFAdvgLxRSNEYp1JUlXz012Xiym6/zR6h5Sl3Xny6OuRvW5s7G7iPF6bEw8OLvhvKFnljPazUS5xzPNG1SCfE88I5puqO6zQtNq5xTi4q71PeL98Zo5tkTornyKtjZx672EsL6AN73Tw2HkOlyxu7NC6cOC+Ad4+b7BGvTiYj90G8R95bG6ScWXEZz3kqSVO3LefMKlGJr7eufNlvj5TFY/T0iKg2T+5b57f/8/i7AQAdB8TeiQLno/KbSE1SbqqU9JVtmz3jvifZLa+nuArT2rllSVMPlKIoiqIoSpPoA5SiKIqiKEqTtHYReQIoeaUm0meDn90evN4t+cElW350ZlPge1dlxwO3c42ChJfsoishrsmTJXGOVm1wPxzSS3aPrfTbW7rdxeC5hMhDsZCELoWqLAOsp44HAIfczMfH3b48Mb7B3/axX/tHv/3AZ+eWjyIKYiSX9b5Mrtyj4k91kq7LtUZSyOofTtDrXBKDF9vK+89tkwWjca+MQOewLPIrrCBpJRPsejVxcrVXLrxItF0pifKB3acl11meKj6UJH0Z4t7izBi76smF39tJgRWdIuFxPhtT5sXN3l/K9xLn6vIFkqFKS2dx/vmwhGeMyJzV5XTP19guJMFQqQl40l2iQDnvaMrlHE7xSbkPanmSIOSyIJ53+2UnxM4xXjfeI/LQiuUSkHHimCwfYNkjKlhKGauJ9J/2Ano499KmrJTR4kXkHLjD5V6snSlR1igPWjo2U2Y6H5alyiThJbzF5WHzfYzkueQhkSZ3jovcmKC+lr2Ajjjdh9VFCGoxEwUkvv8zAMDWA1K+xqakj2bKnbPtuMzNpkuktdpKGUPF5XLNSt20NKIg9oxNeLaalHsmV5HXn3RkmYqNkWyZkz7lsvx94H52cpDuk7S8Xqa5cWoVye5U3qhK7anL3M9S7BaSk7K//PDcroN6oBRFURRFUZpk1gcoY8yQMeYHxpjdxphdxphPeNv7jDEPGWP2en97Z9uX0jxFW8DP7A/xmP0eAFyl9m8dbPspjEFt31p07EeL2j86yhOjOvcsAeYi4VUBfNpa+7QxphPAz4wxDwH4FwC2W2v/2BhzF4C7APz2hY9mgV7XpVepBGe5WJNwo9v2FUSiqKf4B4AtXZJPaZJqW4xVxK2Yj4vbMJ9wpZ7h4uzjjPNDrc5I+v0K+fkOevmhelLidi1TJF+M8uNn4pTGviJSAMt/Qz3ucYanJHrvRgq/MDDYhG3oMr142H5tN4CPzdv+C4xJUk4ViroqrKJK20l3e6Ikdin2ifRnQgLi6JIjLl53P+qo1CtvqGbIbZ8OHle21rxsx7b/qX0YkzgXqe2rHeJW/sOt3/Lbf/DgR/x2rEOM5Xhv50rzHPHDJXXOcD6hTrIVfTbmRZU5GdlHV0ak8dFuOfbUkNwT8xWhoxr7Tofcq8VzlNCK5cpxGWfJCYqUy1KepZPu9ulrRNLeskbmr10HJHoycZLmEL4WDf1yt9s0Re+RfIiyzFMcDZgYkXuFpsZZWSz7Tw/QXFETO65JuRF3HLl1VXrYb+8sS+cTNHFwlF0uTfKrd72qVOqFSdJxGqLpSAYsU+RYxpPfSvR9EKN9cE4oCtTDsUnRl1b3ibR69CzpTvV+eLKUU0ksytxTPSClc0xSbB/LuuPcrhe5kcde7JxEuee4TTIgKLLe8XKpORmK0qX8ZtV8gtpUDoZy3fF3Q6zi9oZLwNTLuwBAnL5fTJWuQ1Heb2jphvWWlsTPSW4qts1cmdUDZa09bq192mtPANgNYDWA2wHc473tHgC/GLwH5WJImyy65EeGA7V/y2DbG3dNnNq+hejYjxa1f3QkO7p07lkCNLUGyhizDsB1AJ4AsMJaexxwH7LQmFCcP3OnMWaHMWZHbXwq6C3K3EnhIuxfQSnoLcoccFAD1PZRomM/WtT+EaFzT/sy5yg8Y0wHgPsBfNJaO264HMYFsNbeDeBuAMgODtnci66rMPfGkQt9DN99eavffuvlknTrUCE4hX9nUgYJR0mcLrmS4NmShLNc3inH5kSa+wsSwtTJ6frJl1iPynAoeo/dyVVD7nxy9XJkyUBKHiRPeNtXdEj0w9ePXee3M2tq3ufLwFFsBPCR+dq/y/TNPcRjDvu33SJ5Gocis8idaj17cGLFBtcsVflm921uRFyvTpKSofW4Nk1NUYJH2gdic/hN0GQpl6qtYhpTwEWM/aZsH8L6zzzpt//qs6/x2wNTj/vt2C9IifKiJ23EqTRFtSTjfU1eZOpDkAR2CIhkAgDCS/RdAAAgAElEQVQ74EooiaMinadulySnm9eJdIsjUobhYmMeq7YKoIVjn8cQRdtxWYpP3f6A3+6Ji5y/LCGRwY+/1o0e/vILt/jbdu0VO7/lmj1+e2qTyCmcPJJLStWlrfGyyIodNO9NVuS67H5ZpBhD0UeVjubjhhba/hzE1k0ZEEe8EhxX9Jzyt+0vy/PB107e6Lc3dch7nnHEpg2RdbYuh8k5Hy3JUo4Clclh+S2XFKnwbIESBGdcW6djMqI5opotW+kkmTst3yVD+VHZ95T7nZShxJ3T590sizn32IqcZ63efi4ksn2ehPU2GdJeaGxIu041YFszzOluMsYk4T483Wut/bq3+aQxZtB7fRDAqbDPKxeHY2t45syDAHBW7d9aHOtgJx5HEimo7VtP3f7QsR8Jav/o0Lmn/ZlLFJ4B8CUAu621X6SXHgBwh9e+A8A3F757irUWL5x9GB2JPgA4SS+p/RcZay1exA7k0YkUGqrjqu1bANsfOvZbjto/OnTuWRrMRcJ7PYCPAHjeGPOst+3/B/DHAO4zxnwUwGEAH5htR6lT0xj6b88DANLvEUnt6K+/zm//wl43aeXWNVKVmaMeuklaq0fYAY31iU6WRFqqR7+tzkn0AzNVFbd3sSbm4Kg4TrZZrw7O21J0bJYPy7S/gYys9t87LlJh3nPBc608liP3XAsc+9YeZHoHAWCrdw3mZf+mYGkrxG1ssyI7cM0hcHLSgI+y3MdqUWJa3OhcNZ6j7DLn3PfE6b0s/VVWSGSL4TLaXP9ujsnqxnAGJ3AYHehGAZNome3DcEg6mApeT5hPiVu+5EUW8eVLUMEnjlZtsEiSbFWbeQGrXWR77seul8J6Pi/Y/mjh2K92UHRSnhIvnpAx+Rd33+63l+8QCe/UDTKvrXrElWu6bhUZqPxOkUieu0eSBXcfkDnkSK8kDm6Qr/vcdvqcXK1TNINzhFLqFnmh0kV1zIKDVANZLPuzhD+YFEnruanLAACv7ZYbd7gsEubzL0mNvDfeutdvcyQcS2p1rdBQ9CR/T3DCzFRKtnel5DvGQOaTgaw71pMk4YVJVA7Z+aWjElF+/dVH/HbN62tnSub7c96uC8cOtNfcowQy6wOUtfZRhI+Tty9sd5Tz6Vi5Adf9xhcAAM/8j0+/aK29kV5W+y8iPWYA78D7AQBP2O0Yt2evpZfV9osM2/9h+zUd+y1G7R8d+VUb8A6jc0+7o5nIFUVRFEVRmqS1tfAcB86EG2028t+u8re/7zM/9Ntv7XwRAFB0ZG3+joLUiRutiou8Iy6uz6QRt+qusUG/XXfHjpbFjf7kmcv89upukfY4ooVr4U2VyaXvuYXjIRkgU+QWjpEwcrrY4bc5mWE9mm+KEov+m7UP++1PrZNzbzfsXCL1vPp27Fm3lNcuUeRaRVTniKSL7Bmxabzo2Z2OzRF+ob7SS4EYGc4Jjm1LU/LWuDfOYiRhVKguWJYkvBjvLkThrI9bJxk89pvta7tiamKAdYMSEXf4mES23fbzT/jtN39MoukOlUWeX/UxV57ampLkmc+VZB+QKRArEzIPTVmZC05UZCnBUNKNeCxamRsrVqbwX+kQeXDDw7/ut21JrkWpL/obxFDoU1dc5LLjRVcuu63nOX/bzyakdlt8XM6D53tOfMlReHEvsq5G5z9NGXob5uqGpRfxwO3174cEHTuTpUg2urdoCCFxQNYwbbnpmN/+XmKLu42SQx9B+873ykzUA6UoiqIoitIk+gClKIqiKIrSJC2V8JiO+37qtx//ushyj37v5wEA377yG/625fGdfvtUrYO2S2TbrrLUlSr0igt8MOUmC3ypsNLfNpwSt/hokZL/EfVoO6Ax4WBdlmPX7lzg6JAE6SUHJ9wok7FpcfP+zgu/5LfX/a4kSdzX1BFbQEKevxui6YqUSNOT8DhKz4YlaqTH+eQ0yZwk59l4fMZ7Y+wvv+h0lUubg6cl0WxX3pVHEpRIs0J1wXoTEj3m8ExA9e/AtdaaUX+WmGzHxCdEyt+7T+aNfgn8wvZhSY75k/cEyy4jZ9xoYKdI0biDItVNTlME8IjMQ4Zr4SXI/mnvvqLX+SfwHz4lx+miyNWp1ZTU8UBgV1tKxzGZH16cFkmznryYZcnRsnw3OJQQdPeULNPgWo9TJZn7bX07SXicmJQplkTa49qmxfLMNI/DBfn+KBWD00Dy18PKJ+Re2Pm+oRnv5UjwriMXm9pRaSXqgVIURVEURWmS1nugAhYe2yqlsv8L15N027+WGon7h2VhJibkif/abZIv5DJKkb86Le3//r9dj1b6Fik58c4hyVfzk5Py63G6Kvs+MSG5pCYm5ddh0ssXsrpXfkmensz77bFRaWOCKnmfpF/+L8kvko6/dxej5j8sv2h7vrrDb7ezQ8VJBj9/s8eonqOJF3qzt4MXi4d5OFITlDcqPvNNvG9+PfrlsguLidF5cjmcnPxKT1M1+vqCWvaWZjLy+t6ClMlwUjTS2MNBC8brt258bA7TRpPlctqJiSsl988v3/SU335s7Xq/vbVHykFd03nUbz9+VuaT39rwCADgQEnsfO933uy3a6tlAfVvv+VBv82lYb526ga/fargzkkdlDeIPdtvf50sZv/bva/12zctl5x6Tw7Jouz+v0YkjFwnfb46K3mRJmuuJ2YVLai/PC92rl4j880/H/iJ3z4+LdeLg3G6+04DAA7mxSvLJbjyVLJlWbeoGT0pKS9zxXJJ9N3nleAazMj31eiA3HvZhNxb+yDfWZlvyPjvS4iykfAWsR+b7pL+r5DviYb0mUpboh4oRVEURVGUJtEHKEVRFEVRlCZpvYTnufNNkhb7UVXozGnXDcqLyHdvFNfo96eu9NslyhX1wa5n/PZv7P01v73mDx+b0YVnqZ3HKzNeB0AO2MZ2EMtD2s3Qfa8sqm8QPBqkkHnufJGodIr9a2npZ60y87mcF3rH5XI2lHXhkiyxavAi8rpcl5oUGbTYK27vSrf0SZZmvnrgBfrL864sUUrLbT5eFqsMpkUqiQ+KbNRFuW3Gx0W+znuL0qeH5rBAfInJdszZK2Q8fWHwab/9SI/MHH+0cZvfPgmR+wGRnO7Fmhn7Xo/HZ2wDgH8InWVO+62s1w6z/sPUj+l75Tr/n/U/8Nt/2rPOb38Hshi6lWz6W8mt9amhD/rt37zGleXOOSJe/cM/vMFvr3hKJo6Pvu5jfruap7miX95TX0RupuV6jtNyjFxOpNASLSKvOjJ/8aL0cyX3szmS/g4ekBmf81RlTss+EsOH/PZ12YN+++GU+132zJ51/rat3xNJU5eTtz/qgVIURVEURWkSfYBSFEVRFEVpksjyQNlqJXB7/BHXZX771e/wt429Y7PfPne5uEmpEgUe3P9Wv935xGG/veBu0KDyJXORK8LKnhjvGdaGlMdoYykkOzwh/zESSRKrUHV0L1Kvmgsu02JIcuLcTjYWLOclql55hjS5yKkcTPaQ9GkOBUeWFBytyjgFkd+G7hj228O/5tYKqVFOoCqlPfvqVrlmXT+VF7KnxXJ5kl5jFVfmWP0jSYjUICct4cg7Zv29YsMNV0hJlNgpkXM24qcIZLYSNmwjwwOeRivbjvc3G3S8VfeJJHVzzwf89uiuAb+9IUROXGxquyQKeqOstsD3kff+St3cyzBzCQYArP3O4vTtfLoDtvFV3YxjAe9ohO/aL1wu9XtiOOLtQ2W7pYp6oBRFURRFUZpEH6AURVEURVGaxNgWutqNMSMApsChJZcmA1icc1xrrZ0tKDAUtf9FsRC2P4TFGxvtRLvaX8f+/FH7z412Hfs698yfUPu39AEKAIwxO6y1N7b0oC2mnc+xnfu2ULTzObZz3xaKdj3Hdu3XQtLO59jOfVso2vkc27lvC0Wrz1ElPEVRFEVRlCbRByhFURRFUZQmieIB6u4Ijtlq2vkc27lvC0U7n2M7922haNdzbNd+LSTtfI7t3LeFop3PsZ37tlC09BxbvgZKURRFURRlqaMSnqIoiqIoSpPoA5SiKIqiKEqTtPQByhjzbmPMS8aYfcaYu1p57MXCGDNkjPmBMWa3MWaXMeYT3vY+Y8xDxpi93t/eiPupto+2r2r/aPuq9o+un2r7aPuq9l8srLUt+QcgDmA/gA0AUgCeA7C1VcdfxPMaBHC91+4E8DKArQA+B+Aub/tdAP4kwj6q7aPtp9pf7f+qtL/aXsf+pWz/Vnqgbgawz1r7irW2DOArAG5v4fEXBWvtcWvt0157AsBuAKvhnts93tvuAfCL0fQQgNo+StsDan+1/yKwROyvto8Wtf8i0soHqNUAlZ0Ghr1tlwzGmHUArgPwBIAV1trjgHuxASyPrmdq++h6BkDtr/ZfZNrY/mr7aFH7LyKtfIAyAdsumRwKxpgOAPcD+KS1djzq/pyH2j5a1P7RovaPDrV9tKj9F5FWPkANAxii/68BcKyFx180jDFJuBfxXmvt173NJ40xg97rgwBORdU/qO2jtD2g9lf7LxJLwP5q+2hR+y8irXyAegrAJmPMemNMCsCHADzQwuMvCsYYA+BLAHZba79ILz0A4A6vfQeAb7a6b4TaPlrU/tGi9o8OtX20qP0XkxavnL8N7mr5/QA+08pjL+I5vQGuS3QngGe9f7cB6AewHcBe729fxP1U26v91f5qf7X9q8j2av/Ftb+WclEURVEURWkSzUSuKIqiKIrSJPoApSiKoiiK0iT6AKUoiqIoitIk+gClKIqiKIrSJPoApSiKoiiK0iT6AKUoiqIoitIk+gClKIqiKIrSJPoApSiKoiiK0iT6AKUoiqIoitIk+gClKIqiKIrSJPoApSiKoiiK0iT6AKUoiqIoitIk+gClKIqiKIrSJPoApSiKoiiK0iQX9QBljHm3MeYlY8w+Y8xdC9UpZW6o/aNF7R8davtoUftHh9q+fTDW2vl90Jg4gJcBvBPAMICnAPyqtfbFheueEobaP1rU/tGhto8WtX90qO3bi8RFfPZmAPusta8AgDHmKwBuBxB6IVMmbTPIX8Qh54ZJJv12tSvl/s3Lg2KsZPx2okAPkOSPcxLynmqHvCdecLcnz0z726zjhHQkpIPze2bFBEZPW2uXef9tW/szJuYatbQiK9vCzEXbDdmolpJ26thUwAepPU/bzkYRUyjbEh+pKfsvqu1nOX8Tj8vLtdr8D2O8A8XkRrmY/TXDUhz7dUpr5Li5fMlvlx25LjEa8OlY1W9PFDKy/XDhwgcyNBDm+cM4jHa2v0klg1+om6BhHg65WdheNZqI6N5BzMx8b8gkb8vl4D7Ng/PmnrayPVNZQcfocOcFpxQPfG/Yd4Dlt8dmjuF0uiL/ebky4/XF4Lyx38DFPECtBnCE/j8M4LUX+kAGebzWvP3CezUBA7LJySCxYpXfPv22tQCAU7fKRN+5T0574DmZ0JyUfDEUlsl7Rt4kF6rvSfdmXf6/npPPFYInNpMINq916HycuX8BPWy/doj+u7j2b3YCDpm8Yzn3pjp452v8bamJ4F3Ep+lBleaf8fXSXve7j888NNm54Qt9Ab9EnrDbz9/UlP3nZPt50nD+1eqM1+Nd3X67dm5s3seJZdwvc5NJz76/oPsYmPc1acnYDyLsPJhZzmn/J1/nt6+9Za/fPjop1yWTkOu2rvOs337kmS1+e/O/fnLmzmPyjdPwoFwJ+QKf50NWZPafA4mVa4JfqP+wpQd+JOgbukpzRUXmeGdK5vNYV6fftll33JsSfXEngh8QqgcOBW6fD+fNPW1le+bEh2/12/bNowCAwj4Z47Ay9uJF+iANSXZ0OB10fbwfGJvXn5Btbx++uA7PkfPGfgMX8wAVNLPMuCONMXcCuBMAMshdxOGU81D7R8us9lfbLxo69qNF7R8davs24mIeoIYBDNH/1wA4dv6brLV3A7gbALpM3+w/eUzM+yPjhH9Vx3IyGPb9vng1Vl933G/ftHy/394z4f4SuyEz6W/7j++b4U0AABTpF9maRIffvn+yy2//16G3AQC2/T/yK+Xh4c3S1wf7/fayvyRvCf3yMwlxOVvr1BuBfboAi2P/IML6zr9yQ/q//39cDgD4j9f/nb/tpeKgdDolv7aTRq7z8wU5tV/v+4nfvmPfpwEAvfeIbYO8LgBgkil6D/1qXBjP1Kz2XxDbh9BwbnQdSu+5yW/33HUYAPDuZS/423Ix8biO1cTl/pqs/MiKQ/zrP5jc6re74+6vypcKK/1tL39C7kHzmHhlG/rKnpGQa9Ukizv2F8h7llh3GQDg8hsO+9vuHPyh3/69l2/32wf3r/Dbt79R7Lj1jXJaP9h8LQCg9rLMb+zBtmHe7IWX9lo391wMsYAYqbLMA7Yk94JdI/avdouCkXpJTstMe+8PkwxbQ9va/qr37/bbd63+DgDgx5s3+dt64vKduTJxzm9vTYoX+6nScr/t0JqaKcf1/n2w45S/7XW/+XG/3f/XM5WJVnAxUXhPAdhkjFlvjEkB+BCABxamW8ocUPtHi9o/OtT20aL2jw61fRsxbw+UtbZqjPk4gO8BiAP4G2vtrgXrmXJB1P7RovaPDrV9tKj9o0Nt315cjIQHa+23AXx7gfri4rmhLa3Sj3XKQr4DX1rnt1f3iGw3Ni3RKvftvl525y1cq02J2/U7tW1+2+REUrDTtCCwRrJVXt4TT7ode2DsGn9bLiOu4Mt+9RW/vedt8p61H3xejrNAC50Xxf6z9CdscSpfo/2fudpv/9KVrmv1T/cHL2K8eblIR4+fkNXi+ZQcZztJpF0fdq/52bgszF32j/v8dm1kJLivc1kM3CSLYv8LQefA51Z8381++/1/8j2/nY65csXhksjKuZh8bkNa3OEFRxaGf2v0Wr/dkxS3+7FSDwBgTXrU39b/Fzv89qOfusVvJ7b/TPq6MLJdAy2zfcj9EF8mQTnjb9rgt89tkjkkdesZd9uwSKVnV8vSgD/bIrK2c6WIAbdkZB+vefJX5Zifd+ee4hOyWHdgl9i2c6dcz+orB4PPYYHkvJaP/YuBo6Q5grQgkdTntsgyjVK32GjFUxLxEuvvcxu8+JwXkZdbExXWrra/ueeA3/7s8HsBAFVH7J2Jy1hdnRUJ7wcUdfr0qKiTA7TspupFrPbHH/O3nX2NXFeZ4VqLZiJXFEVRFEVpEn2AUhRFURRFaZKLkvAWhXpeE4ooYUloWddJv33kVK/fTiSDI1Dqzthsr7hrqxVxuzrkYrQs1cXFPRijdlCapIlJSRL5wri46zeuFDnp5L8Rt/uKPxM35LzzLrUS6ltipUSr7LlLJLe33CKRXr21PX57/+QAAOAdq17yt9235zq/XV0m16Jclfa2gdN++0enL/fba7rdiI1f+JRIVYc/1ue3v/2yRIWt/Uu5trEfPhNwYu1PWI4ne6ucZ/e/O4wgUl5UY3dCZLhD0wOB7x2ryRhOxORemq6J9N2XdJOYZmIiVWxIimw08GcP++2/+Mp7/fbQH9B4b2cC5p74CokK2vfJjX672ilzgo3L/ZE9Krub3OWOy84tIld8dtdtfvvrN9zttx+blnvpz09I5GMnLQ84M+HOLYZ+9h55N/X/3dLXxJhEum7+c0kbVD1CuXOWwtyzwNhOieJ2yBYdx8TOfY/KvF0tyvZYSM4nn2ij8yIhtu1Kv70s8ZDfrkt3/WmZe6pWBm5HXOxacCSiuCMp25OUbbMn7X5/n6mJBL5xy4zgw5ajHihFURRFUZQmaT8PVADVVfQrjDw87CWq1S78LFgu06lSRtRYjLxLlDreOrRg13J75vESCfnFyvs4Ni4LE6eul3MQHw6WxK+/Ux8X75n5uTN+e3VavIHPnFrtt9kjN9DrLsLcOSyvv3G95LFJGLHd+Ij8unglL8sCDV2jE5PuYvW/PvV6fxsv4u/vkYWH8d+XdLfjd8sC586v/FROLsDr0E6w1yk+IDYZ+IIs2EyTx2j7aflF2JVyf7UNZsb9bVNUF2dvgTwW9GuPfylOVGRx+aTX3tApY4B/PcYon9+//GfiIfzHnRJAkP1mQDbtdiFgDBz4V+L95PpCXS+TN4IrgtA0lJxw541p0+Nv63tB3vznl73Fb39n+41+Oz1KZaTkVvLLXyQpk398X/AUPr1Srue+O2Vh7rrfJQ/UEph75kTIIvH69upxmadwVLwWHPgyfKsYOnbTZX576MukXHjZxTkHW3yl3ENhWckvZU68Sbz/FarD4ng3wkhR5vRcgvLVOTJuObCFSxpl4+Lprpc3OlaRe2lDp6gUB+fV+4tHPVCKoiiKoihNog9QiqIoiqIoTdJ+Ep7nRufFyt3dshBtuiQL9XhxN0hSc+yFc/4YchM2yHO14M85nBPqvL/n7yNGEl6FFquvWim5cxLr1/rthSw6uZDUy1AAgPMO6XutKkNmkt5fKIpbO50R1+uEl5+LbcQLkx87IotnTVoklMmSSEfJgAABlu3K1CeqAISRKZF7p355ym93foV21KbSXRB7/63kw1od2+m3jxakYGeK5LzTnvv8OL3OcI6WzlQpcPvqnJRZeHOfFMKt8/SEjJPRkizQ7aXFo2v+P/ncmW8GdqWtcN4gebCKg5Rnaa+Ms5oMzwbZzgQMp9wxmj+okPh398pi8dQ5eU9SFFck6SbzUuGgSqXNDKd4oukwe4LKYKyVTtXeIjny4o88PbOzS5zaCQlqqOdKi18hMuzo9RJEkT8h0tGKp2T8Z5496Lerm6RQ8cgHrwAADD4i0lH1xZcXoNdLl4l1MuiGy30zXl9GuZw256UQcMmR74Aa3UCdCVq0TwO6LvnFSS/nReZRoR4oRVEURVGUJtEHKEVRFEVRlCZpPwnPY+IWkbm6s+KWLVWCu2wpooTdfPVouTBRby5xKPzZuvzHMhXDsl06Je7/BEeSvVOqfQ/c3Z4S3vDt4rrOUbTduUnRDypGznXDMonMOnRW8nNVvOhHjqQ7PCGvO07wlanRdo54HDnlRjbGkrK/5f2ieYwXMoGfu361RB+dWfjq9C2hY4tIqaNliRqaroo7vBKbGQnEkhxLqSxl8z4GSbZ7ZL9UU39ih5t7qrBKbN99xVm/vb5H2hy999peiRj8fk6keacgMl87cfpaGeOxopxrRYK2kJHhjqoMOVianrw0XA2RdMVeuRY8PjmDEFXVASkdqAcuUeULDigGpedq+Fx8iu6f66SzKx/BJYEzKnm24v0yt7z8SbfEjrNaonHvvHa73/7h+yWXWu0lKQd14oErZPt2WQYwfYu7DKDwiyL9Hd77Wr99xV/JfePslFx4lzKdm8T2R4pi+3o+p66ERDE+MyZyf09Ktncm5Ppk42LbaYoYrtOXEEnwnya20CtHZ7y3FagHSlEURVEUpUn0AUpRFEVRFKVJ2lbCO/Z6kSIGKMoqSUkraw67w+WzHAlngsJiiIYovBA5KZ4QN35x3PWvV0+LKzzWL5EDA5TIkftXpHM4+zpxUw5INYe2YvxKSuBIsumm5VLm4Pn9IvPtK1LJj26JeJv0TDo9Je7YcwXRNMIUtKlp0THK4+kZr2e7xe3Lsi7LIsu75VrsHV3mtwe2rfTbznO7gzvQhmzsk+ifq7qO++3tUyI5VGpy35S8MRcn+ZTbPD4TDRKrRNNksjJWJ9e6Y37j1eIu78/ItS6SDMhReCsSIm3Ell/lt52DwSVooqawUgZlZkRsVFhLif1Gg8t2UE5A1PMKcrAQR+zZMzKuYyHTFEt0/lTGwUes2NK9NLmelg+MyZuml9Pc6JUJ4mStSxGWgod/SyIoExvdjKPVA5LM8d5OSVha/rAkZVz7eyLhdd4tUavHbyV77Xel3eP7ReLt2Say9cFfkkS3l0mQ7CUNR7xXHRln4xV3ruhKyjx9dackMf1n3T/z2z+ZXue3n5yQcklp0qor3o2zLCHLNY6MyvVboxKeoiiKoijK0kAfoBRFURRFUZqkbSW8Vdsk6da5aZHLMklx641RzTWW7ZqBI5FMPHgfXHMv3+tGD5RzYrrKFEVBdVJSsLT488+SbLVlncgv7ZrGMd0vURLTJZHfbu496Lf35KQOVGVUrtFpkoZicfcMnTJpDRLYAueI/IcKcSO/TaQhlvBMyt0f1zDkKLOuvLiMmSLJfONXSI3CjucC3942cN2tSYps647L9blxmUhhPzgiUXOdXrLRaTr3MAkvmxR5imXA/rzII1fd6N6TwxPiOj9YErlveV4kU65jlSIZvbROEhnG20zCi69wx3MtK/NAnpSBzOtEiiwck/PIiLLaEHHnS2okw3F0XOaU2D9Ow5aTdHKiTH9/9LOXgpwajv3zNz3jt3/8P2/y22NXyLWIr3Gjgattdh2ahWtErvr8Y377+KfcGp6VG8RIk+dEfvv6R/7Ubz/yKyKD/5fHxOgJCXxF7ohX2/ANtExjO8l2/0WO/WqE69v1p935e1v+iL9tU1q+089QqOnJqkimy1JS6PFkWebpem29vKEovQkKf40I9UApiqIoiqI0iT5AKYqiKIqiNEnbSXjxLa4EcW5aulYuS7snK77uiZC6czGW5eYp7YVRP05nB0lcVKvNIVkkkxAZo1yWLHxbusSVubvflUBqZySaox3IZ0VPGxsXmW1NSvoZJ8mzkiJpqCjXolYPRQqpT+iskOPUKpTwcYrcs1WKlPQiyyaN6BXZvOxjtCB95WSnqQRJv+vlOBKf057EV0vEYH9Wktbtnhr02+/ufd5vf3taotz68hdOVMlyXplku2SIPHrEk+5yIXIfv5frHTJTq0SS7Ap8R3TYFe69aKj2ZZxk5UKZzulqkRqy98uYO7eRZDmKyPOPQUo2J8SscTLOsNp63u3GtfByx+UeLNwqsvfOs6upH/KeWIUS1C7zpJODM/u5lDDpmVG6ANC3xx2nldeJocskM33++Lv89pUdMievHpIsqaOvyP1XXDbzu2T5z9ozGWyrKH/6HvEAAB7dSURBVNPygL7U1IzXHz0nSwr+5dqDfvumL3zCb/PYf+Bffc5vf/b4e2bsrycm37ux08FzTCuZ9QHKGPM3AN4H4JS19mpvWx+ArwJYB/f2+6C1djRsH8r82WV34DSOIwVaB6T2bxl1+9cgDw1q/9agYz9a1P7RonNP+zMXCe/LAN593ra7AGy31m4CsN37v7IIrMJaXIc3nL9Z7d8i1P7RobaPFrV/tKj9259ZPVDW2h8ZY9adt/l2AG/x2vcAeATAby9Eh4ZvcxMeJuMS2lKy4qpj2eGaQUnM9dSLG/x2okdkPhsiHdUxJljiMyEfq0d5jU2IhFQl+fDqDRJhd3BMIpS4Hz8+IcnCOm5w3ejJfwqW8HrNMkzbGa7RRbN//cTjJH3WxsX+uwoiDRQnOFyI90FRLGlXg6iWSOqpcjZB+hhJfzYrbneTo7bXL6cgQ7cwKRJKrEt+rXVkRH9h+WV6kDMRXpiW2/88qsslQiUdE2lhpCjiY4ayNyaSM+M6OUlmxZn9NxNLcdWA98cpMyQnmd3cJTUrp6oyNs7VRHMq9lDE5Cz9aLXtpza4PXKSMiit4dqB8t4PX/GU334w9hZ5P0l0dbWIE2nyfdIwNdng7XzMutTRkIyTjvf+K57123+/+3q/3ZEKPlBhtTuHZZ9CIFGP/YslO+zKrKUpmR8MRVSXKfHjiZLcZyfOSJuCYJEZce04uVE+l5iUOcbSm20lQL9tkqVg/xIlSO5PSl+Hvbp4l2Xley1tqN7mFyViMbF2yG8P/JbYtkLXp55UM0e6d3Ii+iXc8+3BCmvtcQDw/i4Pe6Mx5k5jzA5jzI4KSmFvU5pD7R8tc7K/2n5R0LEfLWr/aNG5p41Y9EXk1tq7AdwNAF2mb9YV3ZVbvMWZtDh13YA8xW7qklIi7++Tn04fPyWL/aYL8lScTFW9fgS7lHh7YwkY8sBQeZC0txi5UpCn6fdse8Fvf6j/p377dyZ/2W9306LzVR2Sjn7ne908Ipv+KbB7F02z9o9fvh4AkE/JL6gRev0DvU/67a/hhsB9xJK0oLy+MJy8TgnyWpT5FzktSo/FA/YBIN/pegBLKbF/mX4FZXPSbyfkmseWB+eKWmiatX0QxeXi6SzRLzL+9cxwqaP62GYvEluEvblhNCw09xbwn6GF+j1ZGdfX5iWf0A/PSV6dTEy8gjXOk7TINGt/J+5ax+mWX7njG2WcbeiRRfw7x8UT27CPlBwmXvKsHeLNDnF+N2znIVz3jDUuciePkiP3wQ1r5Vo8ObJZdtJL90dicRfhLsT4n9NxssGLyEvL3XGazss5s9f8fQOSBO5bp1/jtznHXKVL2sa7j7i8kY1TgEutfbL6tcz2NDfnKOJizCvl8vaeF/1tT5ZkHmjYR0HmY4dqe3EZmFLNnXuS5EFNyddoZMzXA3XSGDMIAN7fU7O8X1lY1P7RovaPDrV9tKj9o0Xt30bM9wHqAQB3eO07AHxzYbqjzBG1f7So/aNDbR8tav9oUfu3EXNJY/B3cBetDRhjhgH8PoA/BnCfMeajAA4D+MBCdeiyD7g5bbiERfHnxL36nXes8dvPbBE3OueB4oW0IsuRa53koWpVPhe2oJwXifsV7mmR8ysTks7/6dw6OQ7t7+SYyB7Tjyzz21u+4rraw+qhP2+fwChG4OnY2zybL5r9S0Pu4r+aQ4vaE3IeT1Pl7IYF4CTbObQYHJ6L1+TlDNOUR2j6rOTHSo/K83xyJeXQmqByJmOuBpTKUmgvXU8eB5YWz7JUu3pApJjZqNvfgYNWjP/zme5n2S74dp0gXayxOrprz4vx37P6VJevp6i0T6EiMtBQUha5J2jldAx0v1EOo9lo9djP3/8EAGDz/cGvx7aKFDa+V4TtiU+TROpwziX3by1YYQqV8BreQyqr8fKhcamZYp8ce+8vS26wyWukfeWPREapjc9d92i1/eeLmQ5e4xOruXbaskJyPL1QXeW3D5elHM+pgsxDLNFVKAeev43mmPgJCXZyMnKhncLF54eKeu6ZC/Hx4KUE9bJTMRrAv/ncP/fbK7Hbb9dG5F56viITRGeCJTw3aCZJE1K8uGjK5JyZSxTer4a89PYF7osSwDXmtX77Yfu1ndbaL3n/Vfu3gLr9n7DbMW7PrqGX1P6LjI79aFH7R4vOPe1P9HGAiqIoiqIoS4y2K+VSh/NopB+UaLvLHwx+/4n/fZ3fzlFpj7q7lSPsWM5h2Y7f4zgUnce5QzzJb6BHKnK/tFfcwvZtUr49iwN+WzI/NRIm3UVFuccdEnWpEgBAdklybQkuscLP4hRxF+t0dQzOZ1OmfTudYoGSke3VAukecdYK3T8cyVeJBV/DmhMs4a3pEAmPIwzbES5zwNF0JyclD1RnXCLhuGRN3RI1+hxH1YVtT4SUcqm3OUrp9Jj048Wi/EhOU76WFI2Zai56t/t8qb34cuD2EgU5JcdozAXMriYs8DEkUo/zPCU8VahKqtLUkBy7euiI385Qu31iw1pL6oiboPtsUQxWmQ6OPOSyW4BI4ny9iitdS/ZQmSu0UeRdFOSOyvywIjHmt+tzy/tyMsP+wWO9s+7vIMmqHL2bjrvzScpwBOo8OrzAqAdKURRFURSlSfQBSlEURVEUpUnaTsIziQt3yVKUCxxxn3Z1ioxRogrRQSVZWJ5rOHaDnNfwCb9VLLku4HRSJIquFZOYlRhF+8VIQvTO1ym2JrnjbFTTXqQP2SjXI7btT9C5UkI/fhTn0itOPdEaSWhOhs6fovcstRuuRfLCCR/jFHWZo/ItFUrGypF/XDX8tBftuRClFxYDJ8FRdXI+U9Micb43J2Pn31+4clEDYYk02fYs5yXjrp2zpIL05WRsnKxIcZb+lIwTdsXPUlmpPQir42SD5ceGSDlq1y9XaLTdHGzRYC9vf4mCbKx0UrRjRvRenk94Tm1I9hhyPkuZWKdE09X2uUso3rFSbPFdZ6vf3jMpyZeZqQmxYyJNSz9Srq15CYItUVLNSrstyFh8Jq6S8z9QkqTol3XMrG+85rsS2R02o/9kfJPfXpuR99e8G4G/lntenkbUqAdKURRFURSlSfQBSlEURVEUpUnaTsKbtZ5QiNuZI4OC6t7NJfIuDH5Puew+c7LEFVZzLQw+R1ttL7dvqds9l1JZdJoERSH2xCRBXKwkz99OnM+Jnss9mS+Rp8SYJdl3jCLsag3RdlyXkDZ728OuYZL6WqRzWNYhst10TRJBxgdcl3/1uCTbaycqokgglwiWGY9XRS4rknxd9Wo4stwcBo/gsLqRdTnv1IRE3g12SWLGkyWR8G7qkgjUZwpr/XYtP3v9vcjhOSZMzmNYwgupYzdfGvbn3VYJytFY6qPXQ+bGhmUPl6Bsx+stTFrubXhlVV+X3+tvetBe5bd/rk9qmH7jlERxO0W6hzooAbOXNHJwg4z52qhIVfHeXtrenksCFpobNh/024enZTB2enXsClbmfWfnnln39+iRDX77ui1Sz3HYuLaN00yV3HXIb0cVC6keKEVRFEVRlCbRByhFURRFUZQmaTsJz8eEPNtZctbNwb1el3o48i5IEgIapQuWBB1KJlj3y3OEV5jkEd4p3l97JWJzvPpxfM4TUxLxVebMfmzHLEl4ZUrQ6EXk1UjWS6REUqqOSsRLckz2naHIxqlJrrfkRWOESHznJiUJXpqOw1Ehx6dFaqqt9uoYtqmEV5XTgUMG53qOGRpPbJe6vFah6L1kLHi8hcnQsYAQsl6KvNt3XCJvrrryuN/uj8v1e7Rwud+26SUg4TUJ55Z1+PZY6FP1LlGM1CHKV4pYVu6lWkmiUTnq11565gccOqnazBM8WhFpbahTkuhyxNfOI5IENtMjUXvx/pn7+9AqSex8H4Ij+V4tjJVlgtrYQXUBvfkkY6hG7ZDYuHpkOHB/9uluv53cOnOuqlBlT2dyasbrrUY9UIqiKIqiKE2iD1CKoiiKoihN0r4SHiXJbEgEh3jQu5siLPLOhGS8Mw1ykdtm2S4sMWfDPhrc6O0bCVNX6DghZfGsSANvzkoIkGU78imxQultD4uaixeCn+ErFE0WT8hYqHm1DesRZnwMoFHaeuuQRN/sPieu9jLpLAkvK2S7/pKoZchuZOROqseVphqCW5ad9NvFqntuYyW5fql4iL1Jkk6EJNis18JLUsQlX9caWTFO+lWaZcN4+479piDduCFSjpKMxusKA00PzhymL95fgzxYv8y0v/g09SNLeu85qUv2aiIoqnnCEbusz5/x25szItuPDElk6aZOqd/2/aObZT9Pu3L/R2+Rz93fu8Vvt4Ok1GqOPzTkt9/0Gw/57R9NXAkA+HZhhb9t6ppBv50OkfASlE+a55CgRJqILUCY60XSrt8biqIoiqIobUv7eqDCFog7zS0iD9yFw4txZ3qXZnZl5ntqDR6Q2fuxVEoo1E8ly7mDqGTLmEOlC5JkF9qHScgvh7r3qCMnHpPxiZzfdgZkf5xbi3+oW1rQXi/bUpymXE7koVrZI78C39kteV6+tecav33tZVKp/kzeXQRN2WPaCodKSSTIk5Ol0jR8TfpS4iF8ccr1unWmxPalWvAtz14l9kBlqUr9mencjNfDuCJ5ym9/pSpBCGaWsjxtR1jAB93DlFasIUdT/aYImx7mlDOKAzU80znk5Wrw/NIi8lcVlAfKmZILkFjpej9+rfMxf9unxt/lt3+lRxaDP5qSBeX3PXOj346Pyv2SP+NejGdpgX5tkyyMxpPPz6v7S5nsSRmAy+ITfrvkzJxnMidJvQjZX+pc8Ctx72Yp0H1n6TpEhXqgFEVRFEVRmkQfoBRFURRFUZqkfSU8InTRNbnzuIQHS251uY4XLje0+b2cB4rzQ9EhY95xWMK7KOoyZJvIejFPsWmQaWKBzQbDpHMi9fBC7ri3n6lpkXGYVFY+V6KcUN15yTU0QZ+t77swFbwat1ARfaNiyf2el9WJkxXZn5OMfiHiBaHL0JUQl/VkKdiexwqSR6W+6LtmKQcXXdcqSaPpuEi2E2XZd3dKrkM9h87wRI+/LZWW67cxI4tvj9WkBg0f55KBlg/UcpST68zMPGkNUl3Y7viFsDd521kdiVUpOCUbPCYueRLBc8H01a681hvPBb6+jkoj7T23zG+n8rK9TMsXpi7zttEsePK1Ms5XPNlEny8ROo7LvNEfk/mp/l2aoiiIcr8s5mcVmuk6QiW/aP6Ohd4U0TLrzGaMGTLG/MAYs9sYs8sY8wlve58x5iFjzF7vb+9s+1Kap2gL+Jn9IR6z3wOAq9T+rYNtP4UxqO1bi479aFH7R4fOPUuDufw0rAL4tLV2C4BbAHzMGLMVwF0AtltrNwHY7v1fWWAMDDZhG2417wKA3VD7twy2fQ5dgNq+pejYjxa1f3To3LM0mFXCs9YeB3Dca08YY3YDWA3gdgBv8d52D4BHAPz2gvWsoSI6befIOxsiv8UuLOFxFF6SovBqJDVYExxhVt9PjVy7dg55oBpL05AuM4t0lzZZpOG7Ph24E9mi2T9ecvvTEJGYkv4OxPOynU6pp0MiLBrs6F2XiXFx33Z0kTw3IvlXYtPiis+nODpP9lf08kNlsvJ6YVKkixRFk71SkjIjcZKuuLSJvcBPCLa9cUdBa8Y+YUPyBrF9inSpzpXEzvVIvXhITZGGcjAk4dWSYhSW6zb3upF1fZQLbGxaZNeXCpJraygl+XZSIeVjZqPVYz+QkNonsVywLJSQoY2yp+405HUKW41A45AvV1B0Xug+khefI49pC/vPhbLIPhylffJmd1740piMy76kROk+Ver3269f8Yrf/sbYNtqfNOv34t+efoO/bXyzHE8yHl087TD3zIXcbsmJVaNvyrRXY6hIIaOlHnncCJPwnITsI0+SYD0Cudhs2bRFpqnFCcaYdQCuA/AEgBXew1X9IWt5yGfuNMbsMMbsqCD6sMMlTgpq/0hwUAPU9lGiYz9a1P4RoXNP+zLnByhjTAeA+wF80lo7PtfPWWvvttbeaK29MYlX6SLHBaBqqwCwEWr/llO1VUxjClDbR4KO/WhR+0eHzj3tzZyi8IwxSbgPT/daa7/ubT5pjBm01h43xgwCOBW+hwVkgaPVYiTtOBU2hzxbcpLBwH3EFzc5oGMd7MTjAHB2se2fGXXPhaMQOTHmdwt0M5K0F28o6TETS/Lc4GpJuFYsijO3NiWyyGhBpKjurETQTZfd93dQqZnChPSJ+50jF3CcJNyE4Ui0gM4SddsnkULRFlo/9rsrgZt70iSDUrRKheTOrnQRc6VICTaTJLktz0/67VPTribVkQz+RbshK1F4W1NSUqY8l/olIbRy7AcSkkjTkISXmKREjqxN1ANseVOY/MbBe7NJeJzPk7sXX3h5I3L7zwFnfCJwe2G9e+8cr4gMvWtMyom8setlv31gSuQ8JjUi90V5mStLnavI3LRq08iMzywUkc89c6B6SJISHyQ75+LuEoMkDdaJIRmssnCjkXiJllrQTVHfTzFsTUNEzCUKzwD4EoDd1tov0ksPALjDa98B4JsL3z3FWosXsQN5dALASXpJ7b/IsO1TaMjyrLZvATr2o0XtHx069ywN5iLhvR7ARwC8zRjzrPfvNgB/DOCdxpi9AN7p/V9ZYMZwBidwGKMYAYCtav/WwbafwjjU9q1Fx360qP2jQ+eepcFcovAeRWMgGvP2he0ORKJrss5djmqDTYYkbWwGjtSLUf2uehSe5Zo8i5jjq8cM4B14PwDgYfu1F621N9LLC27/ZME91/6MRKuc7pLIu4wRO3PCuWmSP1kuK1ddl6upUj07kk03LJdorX01cc+WWE6lIvP1BKaZhESNJTPSLlblcyzhdWVEzuJzO4pw2PZP2O0Yt2evpZcXfuwHwGMrn5gZlQI0urX5ril79myQNSl5YJkyMnIUHsPSXl/ajb6bpNp2nJiT7T1Sk4vGUXiZnBx/Nlo99pvBUPLGxLRYvcrOgmbmhZBEmkGzIKsYlu41G1/YhKXtbH/GmSAJ72apeQkvwpplpK6UzANnqiIk5fm+GKclAR1i33hH1XuvjPM71/3Ib/9d55bgPs2Ddph7muUbozf47TXp0Rmvl3pnvyESExR9Tfp0fQ5bchKeoiiKoiiK0og+QCmKoiiKojTJkqiFFwrJfMMjktGe63PV39IgyZHbm2volWh7lSQnhyKbOGpPXm+v5F4Xg6m6Ntg/KlEpk5OiS/z342/126kUyWhkx4Ykj54UZ0PqD56cDI7H4CSPnEgzmXSPeYYi9q4aPO639zy0yW9/rvQuv81ReztGLvPbq2rtWWPJh8ZWZ1zkhypFtmVIouAkl/1pV6qsWU5wJ++tXiiLaAB5L7Lmuu7D/rYHSyKZFByRPk5UJSKnh+rpVcpLe8rxSch5OEmSeag2XX2YN0TSzWW4ca7g2aaWhiy/l848NF+OvEtq06Hifg+crcoShG1dItpPODKv/fTIOr8dz1FSWUq0nN3jytIjQzJnPWk2+u3DH5d7Yc0fPTaf7i9pvn9A5t6PXeVKm5xc00nNvo9YITjquC7D8jVrB9QDpSiKoiiK0iT6AKUoiqIoitIkS8Kfbsg1zaWpTFxkjHhCpAlLfu94QJJLrvPGEWGN8tzcny2DjrFUqWXc8/7w+h3+tu+d3Oq3nx1e7bcrk+KTvflqkXV2nZXaU34S0hAJjyP2TEiWwYbkkJ4Ud/Jsl7/tvcue99vF+7r99p7lIkNufo3km+PjVDqWAUD75upl6Zn0n1W5Mb/daURyqMt2ADBdc7M6pmLyeomkP47CK1YlA2QmEexGHy25sunZssinnKyTo53O1ETmyMZE0q1OL4kpxyds7rGdYoMaDZ7kJL0nYAoJVU1D1LcYl3nzLhHfJnQJ4SRoqUHDvrnQXnAt0UuF4kZKnutFT7+pc4+/7YXpIb/92KjIbyu6g6Pmjrwoc1ntavfifnCFzI2HygN+O/f60/Pt9iVBbKfIp7lr3OsQpzmrlpn9e9KUKAqPk1l7cwvPK+2AeqAURVEURVGapH1/Djb568jSL3V+zjXeL656/qDzKdJi0ArlHqpSmxeX8wL0OpMj+Rnblir57+8GAPzdF2QBNq07xsbnzvlt57mdfvs/HH7Ub385J+li7nvlOgBAYkzsOV6ShYD10iwAUKsEX6PRCSqb4ZWV2TQoHqX/c/Rmef2lfX77su9IYMHOkvzaTI7LcTY8dggAEJwFKXqSHfKL7EwleJx94uD7/fbn193vt+8959qFPUNzgT1dDrlG6qUV0uTxuiIjC/h3F1f57TVxWTiejZMbpXZpLHS2SfHkkYMNlAoLtfowDznlhpItNK3wwnF2HtUdiXw5eddOmvrU0NmQA10iJFbLuLvy3x7022NvvRwA8Mw16/xtV2eH/fbalHiMTlTFc90XFzfit3Kv8dvjZfeCfv5zH/K39X/pcfkcpDTMq5H+F2fOM7zou2fduRmvz2BcbD9GueQynit2uBxccicq1AOlKIqiKIrSJPoApSiKoiiK0iTtJ+HVFznOwdUc6xWJprNDJAPOSVRfJL65R6pm86LaoYyknD9VlkVwp0uyWI3z2Dw9sgYA4JCfvdjBJdjnAJ9bE+fbCuolCPr+5+PBr4d87qOXvcFv/85+kfYuu9Jzk18p7+V8Qd84IRUKRqg8y6Y+ca9fnpdrtyV7DADwxMQGf9s+yv3CTuT0t5/y2xu/HdzvdpXu6lQrIsv0J2WBeDImi8gf//L1fvv//cYH/PbEje5YZZ2nmpbfTLWUvJCaCr6yTkDlBEsLqx/dKaV49v4LWVD7X3/lb/z2SEKuT7Jz7qVc2plaLkVtyh+XD8h7Q6alqadhQbmTomAKyiXl0NTiS3j8uQSVTeqS6bz9JvZFJEVGKshc3fH3TwAAfvz3IiP95Npf8NvH3iy5ylY8KfdW4qy0TVnkZ3vIlf/6HcklFcvJ8oJYr+yvevRYc+dwCdC5S+aCvKdlDyVlW0N5rhBMTmS7VcmZkh/v73u4dsbrrUY9UIqiKIqiKE2iD1CKoiiKoihNsiQ8vdYJkLwA1M6c9dvFxySNfOwEvb/stndjOe1Dmk+nOC9K8PE5Bb2TdN9Pah+Sc0hR30A752Lx+sY5tkBtWyYJJqTvf7Rxm+zuxqsBAIffJXmbpjfIPvqWj/vtnqzkFHr+xKDffmbPFX576CEvv8gjT9MRz2I2DF8kikqyNU/0a7fr4LHiu9LvvZtlDNfLqgDA8qckhw1LB9kWyAgsmW7+K5E7/vKWt/htziuV//GlEbGaOC02zx+Wc+K8TfXC8SzV8ev889UpUr4viuQzNPc53lwVL/JYlZ2kR4sIomH+vASxLOH1i4wW99qmJvd7dedLfnvls8HRqbUYRTNmZLlBYq0ridssJf6ifds4XVBR+V411CgC+t9/1Y1U7NslY2/NV3866z6qBw757d/7tixHuOb6AwCA4f8lSzf6EbzMpJWoB0pRFEVRFKVJ9AFKURRFURSlSYxtoXRhjBkBMAXgUs95P4DFOce11tpl8/2w2v+iWAjbH8LijY12ol3tr2N//qj950a7jn2de+ZPqP1b+gAFAMaYHdbaG2d/59Klnc+xnfu2ULTzObZz3xaKdj3Hdu3XQtLO59jOfVso2vkc27lvC0Wrz1ElPEVRFEVRlCbRByhFURRFUZQmieIB6u4Ijtlq2vkc27lvC0U7n2M7922haNdzbNd+LSTtfI7t3LeFop3PsZ37tlC09BxbvgZKURRFURRlqaMSnqIoiqIoSpO09AHKGPNuY8xLxph9xpi7WnnsxcIYM2SM+YExZrcxZpcx5hPe9j5jzEPGmL3e397Z9rXI/VTbR9tXtX+0fVX7R9dPtX20fVX7LxbW2pb8AxAHsB/ABgApAM8B2Nqq4y/ieQ0CuN5rdwJ4GcBWAJ8DcJe3/S4AfxJhH9X20fZT7a/2f1XaX22vY/9Stn8rPVA3A9hnrX3FWlsG8BUAt7fw+IuCtfa4tfZprz0BYDeA1XDP7R7vbfcA+MVoeghAbR+l7QG1v9p/EVgi9lfbR4vafxFp5QPUagBH6P/D3rZLBmPMOgDXAXgCwApr7XHAvdgAVzNuOWr7aFH7R4vaPzrU9tGi9l9EWvkAZQK2XTIhgMaYDgD3A/iktXY86v6ch9o+WtT+0aL2jw61fbSo/ReRVj5ADQMYov+vAXCshcdfNIwxSbgX8V5r7de9zSeNMYPe64MATkXVP6jto7Q9oPZX+y8SS8D+avtoUfsvIq18gHoKwCZjzHpjTArAhwA80MLjLwrGGAPgSwB2W2u/SC89AOAOr30HgG+2um+E2j5a1P7RovaPDrV9tKj9F5MWr5y/De5q+f0APtPKYy/iOb0Brkt0J4BnvX+3AegHsB3AXu9vX8T9VNur/dX+an+1/avI9mr/xbW/ZiJXFEVRFEVpEs1EriiKoiiK0iT6AKUoiqIoitIk+gClKIqiKIrSJPoApSiKoiiK0iT6AKUoiqIoitIk+gClKIqiKIrSJPoApSiKoiiK0iT6AKUoiqIoitIk/xdjXhSiKK9XwQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 720x720 with 30 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10,10))\n",
    "i = 0\n",
    "for image,label in train_dataset.take(30):\n",
    "    image = image.numpy().reshape((28,28))\n",
    "    plt.subplot(5,6,i+1)\n",
    "    plt.imshow(image)\n",
    "    plt\n",
    "    i+=1\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Building up our model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = tf.keras.Sequential([\n",
    "    tf.keras.layers.Flatten(input_shape=(28,28,1)),\n",
    "    tf.keras.layers.Dense(128,activation=tf.nn.relu),\n",
    "    tf.keras.layers.Dense(10,activation=tf.nn.softmax)\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(optimizer='adam',\n",
    "             loss = 'sparse_categorical_crossentropy',\n",
    "             metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "BATCH_SIZE = 32\n",
    "train_dataset = train_dataset.repeat().shuffle(metadata.splits['train'].num_examples).batch(BATCH_SIZE)\n",
    "test_dataset = test_dataset.batch(BATCH_SIZE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train for 1875 steps\n",
      "Epoch 1/5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1707/1875 [==========================>...] - ETA: 2:54:25 - loss: 2.4817 - accuracy: 0.187 - ETA: 35:08 - loss: 2.0620 - accuracy: 0.3688  - ETA: 19:40 - loss: 1.8203 - accuracy: 0.447 - ETA: 13:43 - loss: 1.6808 - accuracy: 0.468 - ETA: 10:34 - loss: 1.5770 - accuracy: 0.492 - ETA: 8:36 - loss: 1.4921 - accuracy: 0.510 - ETA: 7:17 - loss: 1.4117 - accuracy: 0.52 - ETA: 6:19 - loss: 1.3401 - accuracy: 0.55 - ETA: 5:36 - loss: 1.2786 - accuracy: 0.57 - ETA: 4:46 - loss: 1.2066 - accuracy: 0.59 - ETA: 4:03 - loss: 1.1348 - accuracy: 0.61 - ETA: 3:36 - loss: 1.1035 - accuracy: 0.62 - ETA: 3:05 - loss: 1.0630 - accuracy: 0.62 - ETA: 2:42 - loss: 1.0287 - accuracy: 0.64 - ETA: 2:24 - loss: 1.0032 - accuracy: 0.65 - ETA: 2:09 - loss: 0.9806 - accuracy: 0.66 - ETA: 1:58 - loss: 0.9477 - accuracy: 0.67 - ETA: 1:48 - loss: 0.9283 - accuracy: 0.67 - ETA: 1:40 - loss: 0.9162 - accuracy: 0.68 - ETA: 1:34 - loss: 0.9001 - accuracy: 0.68 - ETA: 1:28 - loss: 0.8817 - accuracy: 0.69 - ETA: 1:23 - loss: 0.8687 - accuracy: 0.70 - ETA: 1:19 - loss: 0.8522 - accuracy: 0.70 - ETA: 1:15 - loss: 0.8421 - accuracy: 0.70 - ETA: 1:11 - loss: 0.8274 - accuracy: 0.71 - ETA: 1:08 - loss: 0.8129 - accuracy: 0.72 - ETA: 1:05 - loss: 0.7991 - accuracy: 0.72 - ETA: 1:02 - loss: 0.7906 - accuracy: 0.72 - ETA: 59s - loss: 0.7830 - accuracy: 0.7326 - ETA: 57s - loss: 0.7747 - accuracy: 0.734 - ETA: 55s - loss: 0.7699 - accuracy: 0.736 - ETA: 54s - loss: 0.7614 - accuracy: 0.738 - ETA: 53s - loss: 0.7577 - accuracy: 0.741 - ETA: 51s - loss: 0.7516 - accuracy: 0.744 - ETA: 49s - loss: 0.7460 - accuracy: 0.746 - ETA: 48s - loss: 0.7413 - accuracy: 0.748 - ETA: 46s - loss: 0.7334 - accuracy: 0.750 - ETA: 45s - loss: 0.7270 - accuracy: 0.751 - ETA: 44s - loss: 0.7206 - accuracy: 0.754 - ETA: 42s - loss: 0.7154 - accuracy: 0.756 - ETA: 41s - loss: 0.7111 - accuracy: 0.757 - ETA: 40s - loss: 0.7062 - accuracy: 0.759 - ETA: 39s - loss: 0.7013 - accuracy: 0.761 - ETA: 38s - loss: 0.6975 - accuracy: 0.762 - ETA: 37s - loss: 0.6925 - accuracy: 0.764 - ETA: 36s - loss: 0.6873 - accuracy: 0.766 - ETA: 35s - loss: 0.6827 - accuracy: 0.767 - ETA: 34s - loss: 0.6777 - accuracy: 0.768 - ETA: 33s - loss: 0.6701 - accuracy: 0.770 - ETA: 32s - loss: 0.6660 - accuracy: 0.773 - ETA: 31s - loss: 0.6618 - accuracy: 0.774 - ETA: 30s - loss: 0.6604 - accuracy: 0.774 - ETA: 29s - loss: 0.6560 - accuracy: 0.776 - ETA: 29s - loss: 0.6512 - accuracy: 0.777 - ETA: 28s - loss: 0.6456 - accuracy: 0.779 - ETA: 27s - loss: 0.6418 - accuracy: 0.780 - ETA: 26s - loss: 0.6379 - accuracy: 0.782 - ETA: 25s - loss: 0.6341 - accuracy: 0.783 - ETA: 24s - loss: 0.6301 - accuracy: 0.785 - ETA: 24s - loss: 0.6254 - accuracy: 0.786 - ETA: 23s - loss: 0.6212 - accuracy: 0.787 - ETA: 22s - loss: 0.6184 - accuracy: 0.788 - ETA: 22s - loss: 0.6167 - accuracy: 0.789 - ETA: 21s - loss: 0.6141 - accuracy: 0.790 - ETA: 21s - loss: 0.6114 - accuracy: 0.791 - ETA: 20s - loss: 0.6091 - accuracy: 0.791 - ETA: 20s - loss: 0.6066 - accuracy: 0.792 - ETA: 19s - loss: 0.6059 - accuracy: 0.792 - ETA: 19s - loss: 0.6029 - accuracy: 0.793 - ETA: 18s - loss: 0.6013 - accuracy: 0.793 - ETA: 18s - loss: 0.5982 - accuracy: 0.794 - ETA: 17s - loss: 0.5968 - accuracy: 0.795 - ETA: 17s - loss: 0.5964 - accuracy: 0.795 - ETA: 17s - loss: 0.5957 - accuracy: 0.795 - ETA: 16s - loss: 0.5955 - accuracy: 0.795 - ETA: 16s - loss: 0.5946 - accuracy: 0.796 - ETA: 15s - loss: 0.5930 - accuracy: 0.796 - ETA: 15s - loss: 0.5920 - accuracy: 0.797 - ETA: 15s - loss: 0.5880 - accuracy: 0.798 - ETA: 14s - loss: 0.5866 - accuracy: 0.799 - ETA: 14s - loss: 0.5848 - accuracy: 0.799 - ETA: 14s - loss: 0.5823 - accuracy: 0.800 - ETA: 14s - loss: 0.5801 - accuracy: 0.801 - ETA: 13s - loss: 0.5792 - accuracy: 0.801 - ETA: 13s - loss: 0.5783 - accuracy: 0.801 - ETA: 13s - loss: 0.5771 - accuracy: 0.801 - ETA: 12s - loss: 0.5767 - accuracy: 0.802 - ETA: 12s - loss: 0.5755 - accuracy: 0.802 - ETA: 12s - loss: 0.5734 - accuracy: 0.802 - ETA: 11s - loss: 0.5720 - accuracy: 0.803 - ETA: 11s - loss: 0.5713 - accuracy: 0.803 - ETA: 11s - loss: 0.5702 - accuracy: 0.803 - ETA: 11s - loss: 0.5676 - accuracy: 0.804 - ETA: 11s - loss: 0.5650 - accuracy: 0.805 - ETA: 10s - loss: 0.5629 - accuracy: 0.806 - ETA: 10s - loss: 0.5620 - accuracy: 0.806 - ETA: 10s - loss: 0.5612 - accuracy: 0.806 - ETA: 10s - loss: 0.5595 - accuracy: 0.807 - ETA: 9s - loss: 0.5583 - accuracy: 0.807 - ETA: 9s - loss: 0.5575 - accuracy: 0.80 - ETA: 9s - loss: 0.5571 - accuracy: 0.80 - ETA: 9s - loss: 0.5554 - accuracy: 0.80 - ETA: 9s - loss: 0.5552 - accuracy: 0.80 - ETA: 8s - loss: 0.5537 - accuracy: 0.80 - ETA: 8s - loss: 0.5520 - accuracy: 0.80 - ETA: 8s - loss: 0.5505 - accuracy: 0.80 - ETA: 8s - loss: 0.5496 - accuracy: 0.81 - ETA: 8s - loss: 0.5484 - accuracy: 0.81 - ETA: 8s - loss: 0.5478 - accuracy: 0.81 - ETA: 7s - loss: 0.5468 - accuracy: 0.81 - ETA: 7s - loss: 0.5457 - accuracy: 0.81 - ETA: 7s - loss: 0.5442 - accuracy: 0.81 - ETA: 7s - loss: 0.5430 - accuracy: 0.81 - ETA: 7s - loss: 0.5415 - accuracy: 0.81 - ETA: 7s - loss: 0.5405 - accuracy: 0.81 - ETA: 7s - loss: 0.5392 - accuracy: 0.81 - ETA: 6s - loss: 0.5380 - accuracy: 0.81 - ETA: 6s - loss: 0.5370 - accuracy: 0.81 - ETA: 6s - loss: 0.5357 - accuracy: 0.81 - ETA: 6s - loss: 0.5346 - accuracy: 0.81 - ETA: 6s - loss: 0.5339 - accuracy: 0.81 - ETA: 6s - loss: 0.5327 - accuracy: 0.81 - ETA: 6s - loss: 0.5320 - accuracy: 0.81 - ETA: 6s - loss: 0.5316 - accuracy: 0.81 - ETA: 5s - loss: 0.5320 - accuracy: 0.81 - ETA: 5s - loss: 0.5311 - accuracy: 0.81 - ETA: 5s - loss: 0.5302 - accuracy: 0.81 - ETA: 5s - loss: 0.5297 - accuracy: 0.81 - ETA: 5s - loss: 0.5293 - accuracy: 0.81 - ETA: 5s - loss: 0.5290 - accuracy: 0.81 - ETA: 5s - loss: 0.5290 - accuracy: 0.81 - ETA: 5s - loss: 0.5285 - accuracy: 0.81 - ETA: 5s - loss: 0.5285 - accuracy: 0.81 - ETA: 5s - loss: 0.5276 - accuracy: 0.81 - ETA: 4s - loss: 0.5272 - accuracy: 0.81 - ETA: 4s - loss: 0.5270 - accuracy: 0.81 - ETA: 4s - loss: 0.5267 - accuracy: 0.81 - ETA: 4s - loss: 0.5259 - accuracy: 0.81 - ETA: 4s - loss: 0.5256 - accuracy: 0.81 - ETA: 4s - loss: 0.5245 - accuracy: 0.81 - ETA: 4s - loss: 0.5239 - accuracy: 0.81 - ETA: 4s - loss: 0.5235 - accuracy: 0.81 - ETA: 4s - loss: 0.5229 - accuracy: 0.81 - ETA: 4s - loss: 0.5226 - accuracy: 0.81 - ETA: 4s - loss: 0.5219 - accuracy: 0.81 - ETA: 4s - loss: 0.5210 - accuracy: 0.81 - ETA: 4s - loss: 0.5203 - accuracy: 0.81 - ETA: 4s - loss: 0.5195 - accuracy: 0.81 - ETA: 4s - loss: 0.5189 - accuracy: 0.81 - ETA: 3s - loss: 0.5192 - accuracy: 0.81 - ETA: 3s - loss: 0.5188 - accuracy: 0.81 - ETA: 3s - loss: 0.5182 - accuracy: 0.81 - ETA: 3s - loss: 0.5181 - accuracy: 0.81 - ETA: 3s - loss: 0.5178 - accuracy: 0.81 - ETA: 3s - loss: 0.5172 - accuracy: 0.81 - ETA: 3s - loss: 0.5168 - accuracy: 0.81 - ETA: 3s - loss: 0.5164 - accuracy: 0.81 - ETA: 3s - loss: 0.5158 - accuracy: 0.82 - ETA: 3s - loss: 0.5154 - accuracy: 0.82 - ETA: 3s - loss: 0.5148 - accuracy: 0.82 - ETA: 3s - loss: 0.5146 - accuracy: 0.82 - ETA: 3s - loss: 0.5147 - accuracy: 0.82 - ETA: 3s - loss: 0.5142 - accuracy: 0.82 - ETA: 3s - loss: 0.5137 - accuracy: 0.82 - ETA: 3s - loss: 0.5132 - accuracy: 0.82 - ETA: 2s - loss: 0.5127 - accuracy: 0.82 - ETA: 2s - loss: 0.5120 - accuracy: 0.82 - ETA: 2s - loss: 0.5112 - accuracy: 0.82 - ETA: 2s - loss: 0.5108 - accuracy: 0.82 - ETA: 2s - loss: 0.5101 - accuracy: 0.82 - ETA: 2s - loss: 0.5101 - accuracy: 0.82 - ETA: 2s - loss: 0.5095 - accuracy: 0.82 - ETA: 2s - loss: 0.5090 - accuracy: 0.82 - ETA: 2s - loss: 0.5089 - accuracy: 0.82 - ETA: 2s - loss: 0.5085 - accuracy: 0.82 - ETA: 2s - loss: 0.5079 - accuracy: 0.82 - ETA: 2s - loss: 0.5073 - accuracy: 0.82 - ETA: 2s - loss: 0.5064 - accuracy: 0.82 - ETA: 2s - loss: 0.5065 - accuracy: 0.82 - ETA: 2s - loss: 0.5058 - accuracy: 0.82 - ETA: 2s - loss: 0.5050 - accuracy: 0.82 - ETA: 1s - loss: 0.5043 - accuracy: 0.82 - ETA: 1s - loss: 0.5042 - accuracy: 0.82 - ETA: 1s - loss: 0.5041 - accuracy: 0.82 - ETA: 1s - loss: 0.5038 - accuracy: 0.82 - ETA: 1s - loss: 0.5033 - accuracy: 0.82 - ETA: 1s - loss: 0.5030 - accuracy: 0.82 - ETA: 1s - loss: 0.5027 - accuracy: 0.82 - ETA: 1s - loss: 0.5024 - accuracy: 0.8240\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1875/1875 [==============================] - ETA: 1s - loss: 0.5019 - accuracy: 0.82 - ETA: 1s - loss: 0.5016 - accuracy: 0.82 - ETA: 1s - loss: 0.5011 - accuracy: 0.82 - ETA: 1s - loss: 0.5005 - accuracy: 0.82 - ETA: 1s - loss: 0.5001 - accuracy: 0.82 - ETA: 1s - loss: 0.4998 - accuracy: 0.82 - ETA: 1s - loss: 0.4996 - accuracy: 0.82 - ETA: 1s - loss: 0.4990 - accuracy: 0.82 - ETA: 0s - loss: 0.4986 - accuracy: 0.82 - ETA: 0s - loss: 0.4981 - accuracy: 0.82 - ETA: 0s - loss: 0.4980 - accuracy: 0.82 - ETA: 0s - loss: 0.4975 - accuracy: 0.82 - ETA: 0s - loss: 0.4971 - accuracy: 0.82 - ETA: 0s - loss: 0.4968 - accuracy: 0.82 - ETA: 0s - loss: 0.4961 - accuracy: 0.82 - ETA: 0s - loss: 0.4956 - accuracy: 0.82 - ETA: 0s - loss: 0.4951 - accuracy: 0.82 - ETA: 0s - loss: 0.4954 - accuracy: 0.82 - ETA: 0s - loss: 0.4948 - accuracy: 0.82 - ETA: 0s - loss: 0.4943 - accuracy: 0.82 - ETA: 0s - loss: 0.4941 - accuracy: 0.82 - ETA: 0s - loss: 0.4934 - accuracy: 0.82 - ETA: 0s - loss: 0.4930 - accuracy: 0.82 - 17s 9ms/step - loss: 0.4929 - accuracy: 0.8273\n",
      "Epoch 2/5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1435/1875 [=====================>........] - ETA: 52s - loss: 0.7753 - accuracy: 0.750 - ETA: 18s - loss: 0.4779 - accuracy: 0.812 - ETA: 15s - loss: 0.4177 - accuracy: 0.835 - ETA: 13s - loss: 0.4108 - accuracy: 0.837 - ETA: 13s - loss: 0.4137 - accuracy: 0.841 - ETA: 12s - loss: 0.4075 - accuracy: 0.847 - ETA: 12s - loss: 0.4167 - accuracy: 0.848 - ETA: 12s - loss: 0.4064 - accuracy: 0.850 - ETA: 12s - loss: 0.4037 - accuracy: 0.851 - ETA: 12s - loss: 0.3967 - accuracy: 0.854 - ETA: 11s - loss: 0.3907 - accuracy: 0.855 - ETA: 11s - loss: 0.3857 - accuracy: 0.857 - ETA: 11s - loss: 0.3917 - accuracy: 0.855 - ETA: 11s - loss: 0.3890 - accuracy: 0.855 - ETA: 11s - loss: 0.3949 - accuracy: 0.853 - ETA: 11s - loss: 0.3918 - accuracy: 0.855 - ETA: 11s - loss: 0.3914 - accuracy: 0.855 - ETA: 11s - loss: 0.3931 - accuracy: 0.855 - ETA: 11s - loss: 0.3905 - accuracy: 0.856 - ETA: 11s - loss: 0.3907 - accuracy: 0.855 - ETA: 11s - loss: 0.3901 - accuracy: 0.855 - ETA: 11s - loss: 0.3876 - accuracy: 0.856 - ETA: 11s - loss: 0.3868 - accuracy: 0.855 - ETA: 10s - loss: 0.3873 - accuracy: 0.856 - ETA: 10s - loss: 0.3876 - accuracy: 0.856 - ETA: 10s - loss: 0.3855 - accuracy: 0.856 - ETA: 10s - loss: 0.3906 - accuracy: 0.855 - ETA: 10s - loss: 0.3886 - accuracy: 0.856 - ETA: 10s - loss: 0.3917 - accuracy: 0.856 - ETA: 10s - loss: 0.3903 - accuracy: 0.857 - ETA: 10s - loss: 0.3902 - accuracy: 0.856 - ETA: 10s - loss: 0.3906 - accuracy: 0.856 - ETA: 10s - loss: 0.3890 - accuracy: 0.856 - ETA: 10s - loss: 0.3889 - accuracy: 0.857 - ETA: 10s - loss: 0.3897 - accuracy: 0.856 - ETA: 10s - loss: 0.3867 - accuracy: 0.857 - ETA: 10s - loss: 0.3876 - accuracy: 0.857 - ETA: 10s - loss: 0.3865 - accuracy: 0.857 - ETA: 10s - loss: 0.3854 - accuracy: 0.858 - ETA: 9s - loss: 0.3881 - accuracy: 0.857 - ETA: 9s - loss: 0.3861 - accuracy: 0.85 - ETA: 9s - loss: 0.3862 - accuracy: 0.85 - ETA: 9s - loss: 0.3890 - accuracy: 0.85 - ETA: 9s - loss: 0.3902 - accuracy: 0.85 - ETA: 9s - loss: 0.3876 - accuracy: 0.85 - ETA: 9s - loss: 0.3864 - accuracy: 0.85 - ETA: 9s - loss: 0.3880 - accuracy: 0.85 - ETA: 9s - loss: 0.3871 - accuracy: 0.85 - ETA: 9s - loss: 0.3876 - accuracy: 0.85 - ETA: 9s - loss: 0.3872 - accuracy: 0.85 - ETA: 9s - loss: 0.3869 - accuracy: 0.85 - ETA: 9s - loss: 0.3866 - accuracy: 0.85 - ETA: 9s - loss: 0.3872 - accuracy: 0.85 - ETA: 9s - loss: 0.3883 - accuracy: 0.85 - ETA: 9s - loss: 0.3874 - accuracy: 0.85 - ETA: 9s - loss: 0.3868 - accuracy: 0.85 - ETA: 9s - loss: 0.3874 - accuracy: 0.85 - ETA: 9s - loss: 0.3887 - accuracy: 0.85 - ETA: 8s - loss: 0.3902 - accuracy: 0.85 - ETA: 8s - loss: 0.3905 - accuracy: 0.85 - ETA: 8s - loss: 0.3921 - accuracy: 0.85 - ETA: 8s - loss: 0.3922 - accuracy: 0.85 - ETA: 8s - loss: 0.3909 - accuracy: 0.85 - ETA: 8s - loss: 0.3915 - accuracy: 0.85 - ETA: 8s - loss: 0.3922 - accuracy: 0.85 - ETA: 8s - loss: 0.3923 - accuracy: 0.85 - ETA: 8s - loss: 0.3917 - accuracy: 0.85 - ETA: 8s - loss: 0.3915 - accuracy: 0.85 - ETA: 8s - loss: 0.3914 - accuracy: 0.85 - ETA: 8s - loss: 0.3907 - accuracy: 0.85 - ETA: 8s - loss: 0.3909 - accuracy: 0.85 - ETA: 8s - loss: 0.3905 - accuracy: 0.85 - ETA: 8s - loss: 0.3908 - accuracy: 0.85 - ETA: 8s - loss: 0.3903 - accuracy: 0.85 - ETA: 8s - loss: 0.3891 - accuracy: 0.85 - ETA: 8s - loss: 0.3884 - accuracy: 0.85 - ETA: 8s - loss: 0.3874 - accuracy: 0.85 - ETA: 8s - loss: 0.3869 - accuracy: 0.85 - ETA: 7s - loss: 0.3866 - accuracy: 0.85 - ETA: 7s - loss: 0.3866 - accuracy: 0.85 - ETA: 7s - loss: 0.3870 - accuracy: 0.85 - ETA: 7s - loss: 0.3867 - accuracy: 0.85 - ETA: 7s - loss: 0.3867 - accuracy: 0.85 - ETA: 7s - loss: 0.3868 - accuracy: 0.85 - ETA: 7s - loss: 0.3858 - accuracy: 0.85 - ETA: 7s - loss: 0.3857 - accuracy: 0.85 - ETA: 7s - loss: 0.3856 - accuracy: 0.85 - ETA: 7s - loss: 0.3847 - accuracy: 0.85 - ETA: 7s - loss: 0.3841 - accuracy: 0.86 - ETA: 7s - loss: 0.3834 - accuracy: 0.86 - ETA: 7s - loss: 0.3828 - accuracy: 0.86 - ETA: 7s - loss: 0.3832 - accuracy: 0.86 - ETA: 7s - loss: 0.3829 - accuracy: 0.86 - ETA: 7s - loss: 0.3818 - accuracy: 0.86 - ETA: 7s - loss: 0.3819 - accuracy: 0.86 - ETA: 7s - loss: 0.3819 - accuracy: 0.86 - ETA: 7s - loss: 0.3820 - accuracy: 0.86 - ETA: 7s - loss: 0.3817 - accuracy: 0.86 - ETA: 7s - loss: 0.3811 - accuracy: 0.86 - ETA: 7s - loss: 0.3812 - accuracy: 0.86 - ETA: 7s - loss: 0.3816 - accuracy: 0.86 - ETA: 7s - loss: 0.3817 - accuracy: 0.86 - ETA: 7s - loss: 0.3815 - accuracy: 0.86 - ETA: 7s - loss: 0.3807 - accuracy: 0.86 - ETA: 7s - loss: 0.3812 - accuracy: 0.86 - ETA: 7s - loss: 0.3807 - accuracy: 0.86 - ETA: 6s - loss: 0.3801 - accuracy: 0.86 - ETA: 6s - loss: 0.3799 - accuracy: 0.86 - ETA: 6s - loss: 0.3799 - accuracy: 0.86 - ETA: 6s - loss: 0.3794 - accuracy: 0.86 - ETA: 6s - loss: 0.3790 - accuracy: 0.86 - ETA: 6s - loss: 0.3792 - accuracy: 0.86 - ETA: 6s - loss: 0.3794 - accuracy: 0.86 - ETA: 6s - loss: 0.3788 - accuracy: 0.86 - ETA: 6s - loss: 0.3786 - accuracy: 0.86 - ETA: 6s - loss: 0.3788 - accuracy: 0.86 - ETA: 6s - loss: 0.3786 - accuracy: 0.86 - ETA: 6s - loss: 0.3784 - accuracy: 0.86 - ETA: 6s - loss: 0.3781 - accuracy: 0.86 - ETA: 6s - loss: 0.3777 - accuracy: 0.86 - ETA: 6s - loss: 0.3783 - accuracy: 0.86 - ETA: 6s - loss: 0.3788 - accuracy: 0.86 - ETA: 6s - loss: 0.3785 - accuracy: 0.86 - ETA: 6s - loss: 0.3791 - accuracy: 0.86 - ETA: 6s - loss: 0.3788 - accuracy: 0.86 - ETA: 6s - loss: 0.3799 - accuracy: 0.86 - ETA: 6s - loss: 0.3795 - accuracy: 0.86 - ETA: 6s - loss: 0.3790 - accuracy: 0.86 - ETA: 6s - loss: 0.3789 - accuracy: 0.86 - ETA: 6s - loss: 0.3786 - accuracy: 0.86 - ETA: 6s - loss: 0.3786 - accuracy: 0.86 - ETA: 6s - loss: 0.3786 - accuracy: 0.86 - ETA: 5s - loss: 0.3789 - accuracy: 0.86 - ETA: 5s - loss: 0.3796 - accuracy: 0.86 - ETA: 5s - loss: 0.3791 - accuracy: 0.86 - ETA: 5s - loss: 0.3789 - accuracy: 0.86 - ETA: 5s - loss: 0.3786 - accuracy: 0.86 - ETA: 5s - loss: 0.3786 - accuracy: 0.86 - ETA: 5s - loss: 0.3783 - accuracy: 0.86 - ETA: 5s - loss: 0.3777 - accuracy: 0.86 - ETA: 5s - loss: 0.3775 - accuracy: 0.86 - ETA: 5s - loss: 0.3776 - accuracy: 0.86 - ETA: 5s - loss: 0.3772 - accuracy: 0.86 - ETA: 5s - loss: 0.3780 - accuracy: 0.86 - ETA: 5s - loss: 0.3781 - accuracy: 0.86 - ETA: 5s - loss: 0.3778 - accuracy: 0.86 - ETA: 5s - loss: 0.3778 - accuracy: 0.86 - ETA: 5s - loss: 0.3773 - accuracy: 0.86 - ETA: 5s - loss: 0.3767 - accuracy: 0.86 - ETA: 5s - loss: 0.3765 - accuracy: 0.86 - ETA: 5s - loss: 0.3759 - accuracy: 0.86 - ETA: 5s - loss: 0.3756 - accuracy: 0.86 - ETA: 5s - loss: 0.3760 - accuracy: 0.86 - ETA: 4s - loss: 0.3759 - accuracy: 0.86 - ETA: 4s - loss: 0.3767 - accuracy: 0.86 - ETA: 4s - loss: 0.3770 - accuracy: 0.86 - ETA: 4s - loss: 0.3774 - accuracy: 0.86 - ETA: 4s - loss: 0.3769 - accuracy: 0.86 - ETA: 4s - loss: 0.3769 - accuracy: 0.86 - ETA: 4s - loss: 0.3767 - accuracy: 0.86 - ETA: 4s - loss: 0.3765 - accuracy: 0.86 - ETA: 4s - loss: 0.3770 - accuracy: 0.86 - ETA: 4s - loss: 0.3772 - accuracy: 0.86 - ETA: 4s - loss: 0.3774 - accuracy: 0.86 - ETA: 4s - loss: 0.3778 - accuracy: 0.86 - ETA: 4s - loss: 0.3774 - accuracy: 0.86 - ETA: 4s - loss: 0.3773 - accuracy: 0.86 - ETA: 4s - loss: 0.3775 - accuracy: 0.86 - ETA: 4s - loss: 0.3773 - accuracy: 0.86 - ETA: 4s - loss: 0.3775 - accuracy: 0.86 - ETA: 4s - loss: 0.3776 - accuracy: 0.86 - ETA: 4s - loss: 0.3780 - accuracy: 0.86 - ETA: 4s - loss: 0.3782 - accuracy: 0.86 - ETA: 4s - loss: 0.3778 - accuracy: 0.86 - ETA: 4s - loss: 0.3773 - accuracy: 0.86 - ETA: 3s - loss: 0.3768 - accuracy: 0.86 - ETA: 3s - loss: 0.3772 - accuracy: 0.86 - ETA: 3s - loss: 0.3770 - accuracy: 0.86 - ETA: 3s - loss: 0.3769 - accuracy: 0.86 - ETA: 3s - loss: 0.3769 - accuracy: 0.86 - ETA: 3s - loss: 0.3771 - accuracy: 0.86 - ETA: 3s - loss: 0.3773 - accuracy: 0.86 - ETA: 3s - loss: 0.3768 - accuracy: 0.86 - ETA: 3s - loss: 0.3764 - accuracy: 0.86 - ETA: 3s - loss: 0.3765 - accuracy: 0.86 - ETA: 3s - loss: 0.3762 - accuracy: 0.86 - ETA: 3s - loss: 0.3757 - accuracy: 0.86 - ETA: 3s - loss: 0.3756 - accuracy: 0.86 - ETA: 3s - loss: 0.3755 - accuracy: 0.86 - ETA: 3s - loss: 0.3757 - accuracy: 0.86 - ETA: 3s - loss: 0.3760 - accuracy: 0.86 - ETA: 3s - loss: 0.3756 - accuracy: 0.86 - ETA: 3s - loss: 0.3751 - accuracy: 0.8632"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1875/1875 [==============================] - ETA: 3s - loss: 0.3748 - accuracy: 0.86 - ETA: 3s - loss: 0.3749 - accuracy: 0.86 - ETA: 2s - loss: 0.3745 - accuracy: 0.86 - ETA: 2s - loss: 0.3746 - accuracy: 0.86 - ETA: 2s - loss: 0.3751 - accuracy: 0.86 - ETA: 2s - loss: 0.3747 - accuracy: 0.86 - ETA: 2s - loss: 0.3745 - accuracy: 0.86 - ETA: 2s - loss: 0.3746 - accuracy: 0.86 - ETA: 2s - loss: 0.3746 - accuracy: 0.86 - ETA: 2s - loss: 0.3746 - accuracy: 0.86 - ETA: 2s - loss: 0.3745 - accuracy: 0.86 - ETA: 2s - loss: 0.3741 - accuracy: 0.86 - ETA: 2s - loss: 0.3739 - accuracy: 0.86 - ETA: 2s - loss: 0.3738 - accuracy: 0.86 - ETA: 2s - loss: 0.3733 - accuracy: 0.86 - ETA: 2s - loss: 0.3732 - accuracy: 0.86 - ETA: 2s - loss: 0.3729 - accuracy: 0.86 - ETA: 2s - loss: 0.3730 - accuracy: 0.86 - ETA: 2s - loss: 0.3728 - accuracy: 0.86 - ETA: 2s - loss: 0.3726 - accuracy: 0.86 - ETA: 2s - loss: 0.3725 - accuracy: 0.86 - ETA: 1s - loss: 0.3722 - accuracy: 0.86 - ETA: 1s - loss: 0.3720 - accuracy: 0.86 - ETA: 1s - loss: 0.3717 - accuracy: 0.86 - ETA: 1s - loss: 0.3719 - accuracy: 0.86 - ETA: 1s - loss: 0.3717 - accuracy: 0.86 - ETA: 1s - loss: 0.3718 - accuracy: 0.86 - ETA: 1s - loss: 0.3724 - accuracy: 0.86 - ETA: 1s - loss: 0.3728 - accuracy: 0.86 - ETA: 1s - loss: 0.3725 - accuracy: 0.86 - ETA: 1s - loss: 0.3721 - accuracy: 0.86 - ETA: 1s - loss: 0.3719 - accuracy: 0.86 - ETA: 1s - loss: 0.3719 - accuracy: 0.86 - ETA: 1s - loss: 0.3715 - accuracy: 0.86 - ETA: 1s - loss: 0.3713 - accuracy: 0.86 - ETA: 1s - loss: 0.3711 - accuracy: 0.86 - ETA: 1s - loss: 0.3715 - accuracy: 0.86 - ETA: 1s - loss: 0.3713 - accuracy: 0.86 - ETA: 1s - loss: 0.3711 - accuracy: 0.86 - ETA: 1s - loss: 0.3712 - accuracy: 0.86 - ETA: 0s - loss: 0.3714 - accuracy: 0.86 - ETA: 0s - loss: 0.3713 - accuracy: 0.86 - ETA: 0s - loss: 0.3713 - accuracy: 0.86 - ETA: 0s - loss: 0.3716 - accuracy: 0.86 - ETA: 0s - loss: 0.3718 - accuracy: 0.86 - ETA: 0s - loss: 0.3717 - accuracy: 0.86 - ETA: 0s - loss: 0.3716 - accuracy: 0.86 - ETA: 0s - loss: 0.3716 - accuracy: 0.86 - ETA: 0s - loss: 0.3717 - accuracy: 0.86 - ETA: 0s - loss: 0.3715 - accuracy: 0.86 - ETA: 0s - loss: 0.3713 - accuracy: 0.86 - ETA: 0s - loss: 0.3710 - accuracy: 0.86 - ETA: 0s - loss: 0.3708 - accuracy: 0.86 - ETA: 0s - loss: 0.3715 - accuracy: 0.86 - ETA: 0s - loss: 0.3714 - accuracy: 0.86 - ETA: 0s - loss: 0.3716 - accuracy: 0.86 - ETA: 0s - loss: 0.3714 - accuracy: 0.86 - ETA: 0s - loss: 0.3711 - accuracy: 0.86 - ETA: 0s - loss: 0.3711 - accuracy: 0.86 - 13s 7ms/step - loss: 0.3710 - accuracy: 0.8650\n",
      "Epoch 3/5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1556/1875 [=======================>......] - ETA: 41s - loss: 0.4654 - accuracy: 0.875 - ETA: 14s - loss: 0.3710 - accuracy: 0.893 - ETA: 12s - loss: 0.3575 - accuracy: 0.884 - ETA: 12s - loss: 0.3655 - accuracy: 0.877 - ETA: 11s - loss: 0.3479 - accuracy: 0.883 - ETA: 11s - loss: 0.3556 - accuracy: 0.874 - ETA: 11s - loss: 0.3551 - accuracy: 0.876 - ETA: 11s - loss: 0.3424 - accuracy: 0.881 - ETA: 11s - loss: 0.3535 - accuracy: 0.878 - ETA: 11s - loss: 0.3518 - accuracy: 0.878 - ETA: 11s - loss: 0.3484 - accuracy: 0.877 - ETA: 11s - loss: 0.3414 - accuracy: 0.880 - ETA: 11s - loss: 0.3400 - accuracy: 0.879 - ETA: 11s - loss: 0.3359 - accuracy: 0.880 - ETA: 11s - loss: 0.3368 - accuracy: 0.879 - ETA: 10s - loss: 0.3352 - accuracy: 0.881 - ETA: 10s - loss: 0.3316 - accuracy: 0.883 - ETA: 10s - loss: 0.3308 - accuracy: 0.882 - ETA: 10s - loss: 0.3309 - accuracy: 0.881 - ETA: 10s - loss: 0.3316 - accuracy: 0.880 - ETA: 10s - loss: 0.3331 - accuracy: 0.879 - ETA: 10s - loss: 0.3322 - accuracy: 0.878 - ETA: 10s - loss: 0.3367 - accuracy: 0.877 - ETA: 10s - loss: 0.3369 - accuracy: 0.876 - ETA: 10s - loss: 0.3400 - accuracy: 0.875 - ETA: 10s - loss: 0.3390 - accuracy: 0.875 - ETA: 10s - loss: 0.3409 - accuracy: 0.875 - ETA: 10s - loss: 0.3402 - accuracy: 0.876 - ETA: 10s - loss: 0.3414 - accuracy: 0.875 - ETA: 10s - loss: 0.3406 - accuracy: 0.875 - ETA: 10s - loss: 0.3409 - accuracy: 0.876 - ETA: 10s - loss: 0.3431 - accuracy: 0.875 - ETA: 10s - loss: 0.3460 - accuracy: 0.874 - ETA: 9s - loss: 0.3463 - accuracy: 0.872 - ETA: 9s - loss: 0.3457 - accuracy: 0.87 - ETA: 9s - loss: 0.3458 - accuracy: 0.87 - ETA: 9s - loss: 0.3464 - accuracy: 0.87 - ETA: 9s - loss: 0.3462 - accuracy: 0.87 - ETA: 9s - loss: 0.3463 - accuracy: 0.87 - ETA: 9s - loss: 0.3457 - accuracy: 0.87 - ETA: 9s - loss: 0.3462 - accuracy: 0.87 - ETA: 9s - loss: 0.3447 - accuracy: 0.87 - ETA: 9s - loss: 0.3438 - accuracy: 0.87 - ETA: 9s - loss: 0.3419 - accuracy: 0.87 - ETA: 9s - loss: 0.3419 - accuracy: 0.87 - ETA: 9s - loss: 0.3404 - accuracy: 0.87 - ETA: 9s - loss: 0.3411 - accuracy: 0.87 - ETA: 9s - loss: 0.3407 - accuracy: 0.87 - ETA: 9s - loss: 0.3425 - accuracy: 0.87 - ETA: 9s - loss: 0.3423 - accuracy: 0.87 - ETA: 9s - loss: 0.3427 - accuracy: 0.87 - ETA: 9s - loss: 0.3420 - accuracy: 0.87 - ETA: 9s - loss: 0.3414 - accuracy: 0.87 - ETA: 8s - loss: 0.3417 - accuracy: 0.87 - ETA: 8s - loss: 0.3412 - accuracy: 0.87 - ETA: 8s - loss: 0.3409 - accuracy: 0.87 - ETA: 8s - loss: 0.3415 - accuracy: 0.87 - ETA: 8s - loss: 0.3407 - accuracy: 0.87 - ETA: 8s - loss: 0.3395 - accuracy: 0.87 - ETA: 8s - loss: 0.3394 - accuracy: 0.87 - ETA: 8s - loss: 0.3410 - accuracy: 0.87 - ETA: 8s - loss: 0.3416 - accuracy: 0.87 - ETA: 8s - loss: 0.3413 - accuracy: 0.87 - ETA: 8s - loss: 0.3422 - accuracy: 0.87 - ETA: 8s - loss: 0.3424 - accuracy: 0.87 - ETA: 8s - loss: 0.3427 - accuracy: 0.87 - ETA: 8s - loss: 0.3415 - accuracy: 0.87 - ETA: 8s - loss: 0.3411 - accuracy: 0.87 - ETA: 8s - loss: 0.3419 - accuracy: 0.87 - ETA: 8s - loss: 0.3415 - accuracy: 0.87 - ETA: 8s - loss: 0.3405 - accuracy: 0.87 - ETA: 8s - loss: 0.3408 - accuracy: 0.87 - ETA: 8s - loss: 0.3410 - accuracy: 0.87 - ETA: 8s - loss: 0.3412 - accuracy: 0.87 - ETA: 8s - loss: 0.3420 - accuracy: 0.87 - ETA: 8s - loss: 0.3411 - accuracy: 0.87 - ETA: 8s - loss: 0.3405 - accuracy: 0.87 - ETA: 8s - loss: 0.3406 - accuracy: 0.87 - ETA: 8s - loss: 0.3408 - accuracy: 0.87 - ETA: 7s - loss: 0.3417 - accuracy: 0.87 - ETA: 7s - loss: 0.3420 - accuracy: 0.87 - ETA: 7s - loss: 0.3412 - accuracy: 0.87 - ETA: 7s - loss: 0.3411 - accuracy: 0.87 - ETA: 7s - loss: 0.3411 - accuracy: 0.87 - ETA: 7s - loss: 0.3416 - accuracy: 0.87 - ETA: 7s - loss: 0.3410 - accuracy: 0.87 - ETA: 7s - loss: 0.3418 - accuracy: 0.87 - ETA: 7s - loss: 0.3421 - accuracy: 0.87 - ETA: 7s - loss: 0.3425 - accuracy: 0.87 - ETA: 7s - loss: 0.3419 - accuracy: 0.87 - ETA: 7s - loss: 0.3415 - accuracy: 0.87 - ETA: 7s - loss: 0.3417 - accuracy: 0.87 - ETA: 7s - loss: 0.3420 - accuracy: 0.87 - ETA: 7s - loss: 0.3424 - accuracy: 0.87 - ETA: 7s - loss: 0.3417 - accuracy: 0.87 - ETA: 7s - loss: 0.3415 - accuracy: 0.87 - ETA: 7s - loss: 0.3424 - accuracy: 0.87 - ETA: 7s - loss: 0.3420 - accuracy: 0.87 - ETA: 6s - loss: 0.3417 - accuracy: 0.87 - ETA: 6s - loss: 0.3411 - accuracy: 0.87 - ETA: 6s - loss: 0.3409 - accuracy: 0.87 - ETA: 6s - loss: 0.3403 - accuracy: 0.87 - ETA: 6s - loss: 0.3400 - accuracy: 0.87 - ETA: 6s - loss: 0.3400 - accuracy: 0.87 - ETA: 6s - loss: 0.3390 - accuracy: 0.87 - ETA: 6s - loss: 0.3398 - accuracy: 0.87 - ETA: 6s - loss: 0.3395 - accuracy: 0.87 - ETA: 6s - loss: 0.3396 - accuracy: 0.87 - ETA: 6s - loss: 0.3397 - accuracy: 0.87 - ETA: 6s - loss: 0.3402 - accuracy: 0.87 - ETA: 6s - loss: 0.3391 - accuracy: 0.87 - ETA: 6s - loss: 0.3390 - accuracy: 0.87 - ETA: 6s - loss: 0.3395 - accuracy: 0.87 - ETA: 6s - loss: 0.3396 - accuracy: 0.87 - ETA: 6s - loss: 0.3387 - accuracy: 0.87 - ETA: 6s - loss: 0.3389 - accuracy: 0.87 - ETA: 6s - loss: 0.3387 - accuracy: 0.87 - ETA: 5s - loss: 0.3383 - accuracy: 0.87 - ETA: 5s - loss: 0.3375 - accuracy: 0.87 - ETA: 5s - loss: 0.3376 - accuracy: 0.87 - ETA: 5s - loss: 0.3372 - accuracy: 0.87 - ETA: 5s - loss: 0.3368 - accuracy: 0.87 - ETA: 5s - loss: 0.3372 - accuracy: 0.87 - ETA: 5s - loss: 0.3370 - accuracy: 0.87 - ETA: 5s - loss: 0.3367 - accuracy: 0.87 - ETA: 5s - loss: 0.3364 - accuracy: 0.87 - ETA: 5s - loss: 0.3359 - accuracy: 0.87 - ETA: 5s - loss: 0.3360 - accuracy: 0.87 - ETA: 5s - loss: 0.3363 - accuracy: 0.87 - ETA: 5s - loss: 0.3365 - accuracy: 0.87 - ETA: 5s - loss: 0.3361 - accuracy: 0.87 - ETA: 5s - loss: 0.3358 - accuracy: 0.87 - ETA: 5s - loss: 0.3356 - accuracy: 0.87 - ETA: 5s - loss: 0.3350 - accuracy: 0.87 - ETA: 5s - loss: 0.3345 - accuracy: 0.87 - ETA: 5s - loss: 0.3344 - accuracy: 0.87 - ETA: 5s - loss: 0.3341 - accuracy: 0.87 - ETA: 4s - loss: 0.3341 - accuracy: 0.87 - ETA: 4s - loss: 0.3339 - accuracy: 0.87 - ETA: 4s - loss: 0.3349 - accuracy: 0.87 - ETA: 4s - loss: 0.3355 - accuracy: 0.87 - ETA: 4s - loss: 0.3355 - accuracy: 0.87 - ETA: 4s - loss: 0.3356 - accuracy: 0.87 - ETA: 4s - loss: 0.3355 - accuracy: 0.87 - ETA: 4s - loss: 0.3353 - accuracy: 0.87 - ETA: 4s - loss: 0.3354 - accuracy: 0.87 - ETA: 4s - loss: 0.3357 - accuracy: 0.87 - ETA: 4s - loss: 0.3358 - accuracy: 0.87 - ETA: 4s - loss: 0.3359 - accuracy: 0.87 - ETA: 4s - loss: 0.3359 - accuracy: 0.87 - ETA: 4s - loss: 0.3358 - accuracy: 0.87 - ETA: 4s - loss: 0.3358 - accuracy: 0.87 - ETA: 4s - loss: 0.3359 - accuracy: 0.87 - ETA: 4s - loss: 0.3360 - accuracy: 0.87 - ETA: 4s - loss: 0.3357 - accuracy: 0.87 - ETA: 4s - loss: 0.3354 - accuracy: 0.87 - ETA: 3s - loss: 0.3355 - accuracy: 0.87 - ETA: 3s - loss: 0.3356 - accuracy: 0.87 - ETA: 3s - loss: 0.3358 - accuracy: 0.87 - ETA: 3s - loss: 0.3356 - accuracy: 0.87 - ETA: 3s - loss: 0.3353 - accuracy: 0.87 - ETA: 3s - loss: 0.3356 - accuracy: 0.87 - ETA: 3s - loss: 0.3365 - accuracy: 0.87 - ETA: 3s - loss: 0.3367 - accuracy: 0.87 - ETA: 3s - loss: 0.3366 - accuracy: 0.87 - ETA: 3s - loss: 0.3366 - accuracy: 0.87 - ETA: 3s - loss: 0.3369 - accuracy: 0.87 - ETA: 3s - loss: 0.3367 - accuracy: 0.87 - ETA: 3s - loss: 0.3365 - accuracy: 0.87 - ETA: 3s - loss: 0.3360 - accuracy: 0.87 - ETA: 3s - loss: 0.3360 - accuracy: 0.87 - ETA: 3s - loss: 0.3357 - accuracy: 0.87 - ETA: 3s - loss: 0.3354 - accuracy: 0.87 - ETA: 3s - loss: 0.3358 - accuracy: 0.87 - ETA: 3s - loss: 0.3354 - accuracy: 0.87 - ETA: 2s - loss: 0.3354 - accuracy: 0.87 - ETA: 2s - loss: 0.3354 - accuracy: 0.87 - ETA: 2s - loss: 0.3357 - accuracy: 0.87 - ETA: 2s - loss: 0.3356 - accuracy: 0.87 - ETA: 2s - loss: 0.3360 - accuracy: 0.87 - ETA: 2s - loss: 0.3360 - accuracy: 0.87 - ETA: 2s - loss: 0.3364 - accuracy: 0.87 - ETA: 2s - loss: 0.3368 - accuracy: 0.87 - ETA: 2s - loss: 0.3366 - accuracy: 0.87 - ETA: 2s - loss: 0.3366 - accuracy: 0.87 - ETA: 2s - loss: 0.3366 - accuracy: 0.87 - ETA: 2s - loss: 0.3369 - accuracy: 0.87 - ETA: 2s - loss: 0.3368 - accuracy: 0.87 - ETA: 2s - loss: 0.3368 - accuracy: 0.87 - ETA: 2s - loss: 0.3366 - accuracy: 0.87 - ETA: 2s - loss: 0.3366 - accuracy: 0.87 - ETA: 2s - loss: 0.3369 - accuracy: 0.87 - ETA: 2s - loss: 0.3365 - accuracy: 0.8754"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1875/1875 [==============================] - ETA: 2s - loss: 0.3366 - accuracy: 0.87 - ETA: 2s - loss: 0.3362 - accuracy: 0.87 - ETA: 1s - loss: 0.3361 - accuracy: 0.87 - ETA: 1s - loss: 0.3361 - accuracy: 0.87 - ETA: 1s - loss: 0.3355 - accuracy: 0.87 - ETA: 1s - loss: 0.3356 - accuracy: 0.87 - ETA: 1s - loss: 0.3353 - accuracy: 0.87 - ETA: 1s - loss: 0.3352 - accuracy: 0.87 - ETA: 1s - loss: 0.3352 - accuracy: 0.87 - ETA: 1s - loss: 0.3349 - accuracy: 0.87 - ETA: 1s - loss: 0.3353 - accuracy: 0.87 - ETA: 1s - loss: 0.3349 - accuracy: 0.87 - ETA: 1s - loss: 0.3348 - accuracy: 0.87 - ETA: 1s - loss: 0.3346 - accuracy: 0.87 - ETA: 1s - loss: 0.3343 - accuracy: 0.87 - ETA: 1s - loss: 0.3342 - accuracy: 0.87 - ETA: 1s - loss: 0.3341 - accuracy: 0.87 - ETA: 1s - loss: 0.3341 - accuracy: 0.87 - ETA: 1s - loss: 0.3342 - accuracy: 0.87 - ETA: 0s - loss: 0.3341 - accuracy: 0.87 - ETA: 0s - loss: 0.3340 - accuracy: 0.87 - ETA: 0s - loss: 0.3340 - accuracy: 0.87 - ETA: 0s - loss: 0.3340 - accuracy: 0.87 - ETA: 0s - loss: 0.3339 - accuracy: 0.87 - ETA: 0s - loss: 0.3337 - accuracy: 0.87 - ETA: 0s - loss: 0.3337 - accuracy: 0.87 - ETA: 0s - loss: 0.3334 - accuracy: 0.87 - ETA: 0s - loss: 0.3334 - accuracy: 0.87 - ETA: 0s - loss: 0.3332 - accuracy: 0.87 - ETA: 0s - loss: 0.3334 - accuracy: 0.87 - ETA: 0s - loss: 0.3332 - accuracy: 0.87 - ETA: 0s - loss: 0.3332 - accuracy: 0.87 - ETA: 0s - loss: 0.3331 - accuracy: 0.87 - ETA: 0s - loss: 0.3337 - accuracy: 0.87 - ETA: 0s - loss: 0.3337 - accuracy: 0.87 - ETA: 0s - loss: 0.3341 - accuracy: 0.87 - ETA: 0s - loss: 0.3338 - accuracy: 0.87 - ETA: 0s - loss: 0.3338 - accuracy: 0.87 - ETA: 0s - loss: 0.3334 - accuracy: 0.87 - ETA: 0s - loss: 0.3332 - accuracy: 0.87 - 12s 7ms/step - loss: 0.3332 - accuracy: 0.8765\n",
      "Epoch 4/5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1536/1875 [=======================>......] - ETA: 42s - loss: 0.2239 - accuracy: 0.906 - ETA: 15s - loss: 0.3370 - accuracy: 0.868 - ETA: 13s - loss: 0.3553 - accuracy: 0.876 - ETA: 12s - loss: 0.3629 - accuracy: 0.872 - ETA: 11s - loss: 0.3554 - accuracy: 0.869 - ETA: 11s - loss: 0.3516 - accuracy: 0.870 - ETA: 11s - loss: 0.3676 - accuracy: 0.869 - ETA: 11s - loss: 0.3677 - accuracy: 0.871 - ETA: 11s - loss: 0.3566 - accuracy: 0.872 - ETA: 11s - loss: 0.3496 - accuracy: 0.870 - ETA: 10s - loss: 0.3467 - accuracy: 0.870 - ETA: 10s - loss: 0.3476 - accuracy: 0.871 - ETA: 10s - loss: 0.3451 - accuracy: 0.872 - ETA: 10s - loss: 0.3405 - accuracy: 0.871 - ETA: 10s - loss: 0.3419 - accuracy: 0.870 - ETA: 10s - loss: 0.3357 - accuracy: 0.873 - ETA: 10s - loss: 0.3353 - accuracy: 0.874 - ETA: 10s - loss: 0.3398 - accuracy: 0.872 - ETA: 10s - loss: 0.3386 - accuracy: 0.873 - ETA: 10s - loss: 0.3369 - accuracy: 0.873 - ETA: 10s - loss: 0.3322 - accuracy: 0.875 - ETA: 10s - loss: 0.3297 - accuracy: 0.875 - ETA: 10s - loss: 0.3297 - accuracy: 0.876 - ETA: 10s - loss: 0.3307 - accuracy: 0.875 - ETA: 10s - loss: 0.3316 - accuracy: 0.874 - ETA: 10s - loss: 0.3299 - accuracy: 0.876 - ETA: 10s - loss: 0.3284 - accuracy: 0.876 - ETA: 10s - loss: 0.3307 - accuracy: 0.876 - ETA: 10s - loss: 0.3322 - accuracy: 0.875 - ETA: 10s - loss: 0.3353 - accuracy: 0.875 - ETA: 10s - loss: 0.3348 - accuracy: 0.875 - ETA: 10s - loss: 0.3328 - accuracy: 0.875 - ETA: 9s - loss: 0.3316 - accuracy: 0.876 - ETA: 9s - loss: 0.3296 - accuracy: 0.87 - ETA: 9s - loss: 0.3279 - accuracy: 0.87 - ETA: 9s - loss: 0.3280 - accuracy: 0.87 - ETA: 9s - loss: 0.3276 - accuracy: 0.87 - ETA: 9s - loss: 0.3267 - accuracy: 0.87 - ETA: 9s - loss: 0.3248 - accuracy: 0.87 - ETA: 9s - loss: 0.3262 - accuracy: 0.87 - ETA: 9s - loss: 0.3262 - accuracy: 0.87 - ETA: 9s - loss: 0.3281 - accuracy: 0.87 - ETA: 9s - loss: 0.3276 - accuracy: 0.87 - ETA: 9s - loss: 0.3256 - accuracy: 0.87 - ETA: 9s - loss: 0.3254 - accuracy: 0.87 - ETA: 9s - loss: 0.3233 - accuracy: 0.87 - ETA: 9s - loss: 0.3216 - accuracy: 0.87 - ETA: 9s - loss: 0.3215 - accuracy: 0.88 - ETA: 9s - loss: 0.3207 - accuracy: 0.88 - ETA: 9s - loss: 0.3199 - accuracy: 0.88 - ETA: 9s - loss: 0.3210 - accuracy: 0.88 - ETA: 9s - loss: 0.3210 - accuracy: 0.88 - ETA: 9s - loss: 0.3206 - accuracy: 0.88 - ETA: 9s - loss: 0.3216 - accuracy: 0.88 - ETA: 8s - loss: 0.3211 - accuracy: 0.88 - ETA: 8s - loss: 0.3205 - accuracy: 0.88 - ETA: 8s - loss: 0.3203 - accuracy: 0.88 - ETA: 8s - loss: 0.3197 - accuracy: 0.88 - ETA: 8s - loss: 0.3201 - accuracy: 0.88 - ETA: 8s - loss: 0.3204 - accuracy: 0.88 - ETA: 8s - loss: 0.3201 - accuracy: 0.88 - ETA: 8s - loss: 0.3203 - accuracy: 0.88 - ETA: 8s - loss: 0.3208 - accuracy: 0.88 - ETA: 8s - loss: 0.3205 - accuracy: 0.88 - ETA: 8s - loss: 0.3194 - accuracy: 0.88 - ETA: 8s - loss: 0.3194 - accuracy: 0.88 - ETA: 8s - loss: 0.3193 - accuracy: 0.88 - ETA: 8s - loss: 0.3202 - accuracy: 0.88 - ETA: 8s - loss: 0.3199 - accuracy: 0.88 - ETA: 8s - loss: 0.3198 - accuracy: 0.88 - ETA: 8s - loss: 0.3195 - accuracy: 0.88 - ETA: 8s - loss: 0.3189 - accuracy: 0.88 - ETA: 8s - loss: 0.3190 - accuracy: 0.88 - ETA: 8s - loss: 0.3191 - accuracy: 0.88 - ETA: 8s - loss: 0.3187 - accuracy: 0.88 - ETA: 8s - loss: 0.3183 - accuracy: 0.88 - ETA: 8s - loss: 0.3175 - accuracy: 0.88 - ETA: 7s - loss: 0.3178 - accuracy: 0.88 - ETA: 7s - loss: 0.3188 - accuracy: 0.88 - ETA: 7s - loss: 0.3184 - accuracy: 0.88 - ETA: 7s - loss: 0.3178 - accuracy: 0.88 - ETA: 7s - loss: 0.3174 - accuracy: 0.88 - ETA: 7s - loss: 0.3166 - accuracy: 0.88 - ETA: 7s - loss: 0.3170 - accuracy: 0.88 - ETA: 7s - loss: 0.3168 - accuracy: 0.88 - ETA: 7s - loss: 0.3168 - accuracy: 0.88 - ETA: 7s - loss: 0.3162 - accuracy: 0.88 - ETA: 7s - loss: 0.3158 - accuracy: 0.88 - ETA: 7s - loss: 0.3150 - accuracy: 0.88 - ETA: 7s - loss: 0.3152 - accuracy: 0.88 - ETA: 7s - loss: 0.3152 - accuracy: 0.88 - ETA: 7s - loss: 0.3155 - accuracy: 0.88 - ETA: 7s - loss: 0.3151 - accuracy: 0.88 - ETA: 7s - loss: 0.3148 - accuracy: 0.88 - ETA: 7s - loss: 0.3144 - accuracy: 0.88 - ETA: 7s - loss: 0.3141 - accuracy: 0.88 - ETA: 7s - loss: 0.3137 - accuracy: 0.88 - ETA: 7s - loss: 0.3141 - accuracy: 0.88 - ETA: 7s - loss: 0.3140 - accuracy: 0.88 - ETA: 6s - loss: 0.3133 - accuracy: 0.88 - ETA: 6s - loss: 0.3136 - accuracy: 0.88 - ETA: 6s - loss: 0.3144 - accuracy: 0.88 - ETA: 6s - loss: 0.3141 - accuracy: 0.88 - ETA: 6s - loss: 0.3144 - accuracy: 0.88 - ETA: 6s - loss: 0.3144 - accuracy: 0.88 - ETA: 6s - loss: 0.3144 - accuracy: 0.88 - ETA: 6s - loss: 0.3148 - accuracy: 0.88 - ETA: 6s - loss: 0.3144 - accuracy: 0.88 - ETA: 6s - loss: 0.3139 - accuracy: 0.88 - ETA: 6s - loss: 0.3138 - accuracy: 0.88 - ETA: 6s - loss: 0.3140 - accuracy: 0.88 - ETA: 6s - loss: 0.3141 - accuracy: 0.88 - ETA: 6s - loss: 0.3142 - accuracy: 0.88 - ETA: 6s - loss: 0.3136 - accuracy: 0.88 - ETA: 6s - loss: 0.3134 - accuracy: 0.88 - ETA: 6s - loss: 0.3135 - accuracy: 0.88 - ETA: 6s - loss: 0.3129 - accuracy: 0.88 - ETA: 6s - loss: 0.3123 - accuracy: 0.88 - ETA: 6s - loss: 0.3123 - accuracy: 0.88 - ETA: 6s - loss: 0.3122 - accuracy: 0.88 - ETA: 5s - loss: 0.3122 - accuracy: 0.88 - ETA: 5s - loss: 0.3121 - accuracy: 0.88 - ETA: 5s - loss: 0.3119 - accuracy: 0.88 - ETA: 5s - loss: 0.3116 - accuracy: 0.88 - ETA: 5s - loss: 0.3119 - accuracy: 0.88 - ETA: 5s - loss: 0.3115 - accuracy: 0.88 - ETA: 5s - loss: 0.3114 - accuracy: 0.88 - ETA: 5s - loss: 0.3109 - accuracy: 0.88 - ETA: 5s - loss: 0.3113 - accuracy: 0.88 - ETA: 5s - loss: 0.3114 - accuracy: 0.88 - ETA: 5s - loss: 0.3111 - accuracy: 0.88 - ETA: 5s - loss: 0.3109 - accuracy: 0.88 - ETA: 5s - loss: 0.3109 - accuracy: 0.88 - ETA: 5s - loss: 0.3107 - accuracy: 0.88 - ETA: 5s - loss: 0.3102 - accuracy: 0.88 - ETA: 5s - loss: 0.3102 - accuracy: 0.88 - ETA: 5s - loss: 0.3105 - accuracy: 0.88 - ETA: 5s - loss: 0.3104 - accuracy: 0.88 - ETA: 5s - loss: 0.3106 - accuracy: 0.88 - ETA: 4s - loss: 0.3110 - accuracy: 0.88 - ETA: 4s - loss: 0.3112 - accuracy: 0.88 - ETA: 4s - loss: 0.3114 - accuracy: 0.88 - ETA: 4s - loss: 0.3117 - accuracy: 0.88 - ETA: 4s - loss: 0.3117 - accuracy: 0.88 - ETA: 4s - loss: 0.3115 - accuracy: 0.88 - ETA: 4s - loss: 0.3114 - accuracy: 0.88 - ETA: 4s - loss: 0.3119 - accuracy: 0.88 - ETA: 4s - loss: 0.3124 - accuracy: 0.88 - ETA: 4s - loss: 0.3126 - accuracy: 0.88 - ETA: 4s - loss: 0.3128 - accuracy: 0.88 - ETA: 4s - loss: 0.3135 - accuracy: 0.88 - ETA: 4s - loss: 0.3143 - accuracy: 0.88 - ETA: 4s - loss: 0.3143 - accuracy: 0.88 - ETA: 4s - loss: 0.3145 - accuracy: 0.88 - ETA: 4s - loss: 0.3148 - accuracy: 0.88 - ETA: 4s - loss: 0.3146 - accuracy: 0.88 - ETA: 4s - loss: 0.3148 - accuracy: 0.88 - ETA: 4s - loss: 0.3145 - accuracy: 0.88 - ETA: 4s - loss: 0.3144 - accuracy: 0.88 - ETA: 3s - loss: 0.3145 - accuracy: 0.88 - ETA: 3s - loss: 0.3150 - accuracy: 0.88 - ETA: 3s - loss: 0.3144 - accuracy: 0.88 - ETA: 3s - loss: 0.3148 - accuracy: 0.88 - ETA: 3s - loss: 0.3144 - accuracy: 0.88 - ETA: 3s - loss: 0.3141 - accuracy: 0.88 - ETA: 3s - loss: 0.3147 - accuracy: 0.88 - ETA: 3s - loss: 0.3145 - accuracy: 0.88 - ETA: 3s - loss: 0.3146 - accuracy: 0.88 - ETA: 3s - loss: 0.3149 - accuracy: 0.88 - ETA: 3s - loss: 0.3147 - accuracy: 0.88 - ETA: 3s - loss: 0.3145 - accuracy: 0.88 - ETA: 3s - loss: 0.3141 - accuracy: 0.88 - ETA: 3s - loss: 0.3140 - accuracy: 0.88 - ETA: 3s - loss: 0.3134 - accuracy: 0.88 - ETA: 3s - loss: 0.3135 - accuracy: 0.88 - ETA: 3s - loss: 0.3135 - accuracy: 0.88 - ETA: 3s - loss: 0.3136 - accuracy: 0.88 - ETA: 3s - loss: 0.3134 - accuracy: 0.88 - ETA: 3s - loss: 0.3134 - accuracy: 0.88 - ETA: 2s - loss: 0.3133 - accuracy: 0.88 - ETA: 2s - loss: 0.3130 - accuracy: 0.88 - ETA: 2s - loss: 0.3129 - accuracy: 0.88 - ETA: 2s - loss: 0.3126 - accuracy: 0.88 - ETA: 2s - loss: 0.3126 - accuracy: 0.88 - ETA: 2s - loss: 0.3126 - accuracy: 0.88 - ETA: 2s - loss: 0.3127 - accuracy: 0.88 - ETA: 2s - loss: 0.3127 - accuracy: 0.88 - ETA: 2s - loss: 0.3124 - accuracy: 0.88 - ETA: 2s - loss: 0.3126 - accuracy: 0.88 - ETA: 2s - loss: 0.3123 - accuracy: 0.88 - ETA: 2s - loss: 0.3125 - accuracy: 0.88 - ETA: 2s - loss: 0.3129 - accuracy: 0.88 - ETA: 2s - loss: 0.3130 - accuracy: 0.8848"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1875/1875 [==============================] - ETA: 2s - loss: 0.3131 - accuracy: 0.88 - ETA: 2s - loss: 0.3132 - accuracy: 0.88 - ETA: 2s - loss: 0.3129 - accuracy: 0.88 - ETA: 2s - loss: 0.3128 - accuracy: 0.88 - ETA: 2s - loss: 0.3129 - accuracy: 0.88 - ETA: 1s - loss: 0.3129 - accuracy: 0.88 - ETA: 1s - loss: 0.3129 - accuracy: 0.88 - ETA: 1s - loss: 0.3131 - accuracy: 0.88 - ETA: 1s - loss: 0.3128 - accuracy: 0.88 - ETA: 1s - loss: 0.3128 - accuracy: 0.88 - ETA: 1s - loss: 0.3130 - accuracy: 0.88 - ETA: 1s - loss: 0.3127 - accuracy: 0.88 - ETA: 1s - loss: 0.3126 - accuracy: 0.88 - ETA: 1s - loss: 0.3128 - accuracy: 0.88 - ETA: 1s - loss: 0.3125 - accuracy: 0.88 - ETA: 1s - loss: 0.3126 - accuracy: 0.88 - ETA: 1s - loss: 0.3126 - accuracy: 0.88 - ETA: 1s - loss: 0.3124 - accuracy: 0.88 - ETA: 1s - loss: 0.3123 - accuracy: 0.88 - ETA: 1s - loss: 0.3124 - accuracy: 0.88 - ETA: 1s - loss: 0.3124 - accuracy: 0.88 - ETA: 1s - loss: 0.3124 - accuracy: 0.88 - ETA: 1s - loss: 0.3121 - accuracy: 0.88 - ETA: 1s - loss: 0.3122 - accuracy: 0.88 - ETA: 0s - loss: 0.3123 - accuracy: 0.88 - ETA: 0s - loss: 0.3126 - accuracy: 0.88 - ETA: 0s - loss: 0.3127 - accuracy: 0.88 - ETA: 0s - loss: 0.3126 - accuracy: 0.88 - ETA: 0s - loss: 0.3124 - accuracy: 0.88 - ETA: 0s - loss: 0.3122 - accuracy: 0.88 - ETA: 0s - loss: 0.3122 - accuracy: 0.88 - ETA: 0s - loss: 0.3118 - accuracy: 0.88 - ETA: 0s - loss: 0.3117 - accuracy: 0.88 - ETA: 0s - loss: 0.3119 - accuracy: 0.88 - ETA: 0s - loss: 0.3120 - accuracy: 0.88 - ETA: 0s - loss: 0.3118 - accuracy: 0.88 - ETA: 0s - loss: 0.3121 - accuracy: 0.88 - ETA: 0s - loss: 0.3120 - accuracy: 0.88 - ETA: 0s - loss: 0.3122 - accuracy: 0.88 - ETA: 0s - loss: 0.3119 - accuracy: 0.88 - ETA: 0s - loss: 0.3118 - accuracy: 0.88 - ETA: 0s - loss: 0.3116 - accuracy: 0.88 - 13s 7ms/step - loss: 0.3115 - accuracy: 0.8853\n",
      "Epoch 5/5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1480/1875 [======================>.......] - ETA: 41s - loss: 0.2938 - accuracy: 0.906 - ETA: 15s - loss: 0.2570 - accuracy: 0.923 - ETA: 14s - loss: 0.2810 - accuracy: 0.906 - ETA: 13s - loss: 0.2834 - accuracy: 0.901 - ETA: 13s - loss: 0.2750 - accuracy: 0.907 - ETA: 13s - loss: 0.2726 - accuracy: 0.903 - ETA: 13s - loss: 0.2670 - accuracy: 0.901 - ETA: 13s - loss: 0.2689 - accuracy: 0.903 - ETA: 12s - loss: 0.2759 - accuracy: 0.899 - ETA: 12s - loss: 0.2785 - accuracy: 0.898 - ETA: 12s - loss: 0.2887 - accuracy: 0.894 - ETA: 12s - loss: 0.2910 - accuracy: 0.894 - ETA: 12s - loss: 0.2939 - accuracy: 0.894 - ETA: 11s - loss: 0.2889 - accuracy: 0.895 - ETA: 11s - loss: 0.2896 - accuracy: 0.894 - ETA: 11s - loss: 0.2927 - accuracy: 0.894 - ETA: 11s - loss: 0.2991 - accuracy: 0.890 - ETA: 11s - loss: 0.3054 - accuracy: 0.888 - ETA: 11s - loss: 0.3073 - accuracy: 0.888 - ETA: 11s - loss: 0.3091 - accuracy: 0.887 - ETA: 11s - loss: 0.3090 - accuracy: 0.887 - ETA: 11s - loss: 0.3078 - accuracy: 0.888 - ETA: 11s - loss: 0.3090 - accuracy: 0.887 - ETA: 11s - loss: 0.3074 - accuracy: 0.888 - ETA: 11s - loss: 0.3064 - accuracy: 0.888 - ETA: 11s - loss: 0.3095 - accuracy: 0.887 - ETA: 11s - loss: 0.3077 - accuracy: 0.888 - ETA: 11s - loss: 0.3055 - accuracy: 0.889 - ETA: 11s - loss: 0.3050 - accuracy: 0.889 - ETA: 11s - loss: 0.3041 - accuracy: 0.889 - ETA: 11s - loss: 0.3054 - accuracy: 0.889 - ETA: 11s - loss: 0.3104 - accuracy: 0.887 - ETA: 11s - loss: 0.3125 - accuracy: 0.887 - ETA: 11s - loss: 0.3110 - accuracy: 0.888 - ETA: 11s - loss: 0.3100 - accuracy: 0.888 - ETA: 10s - loss: 0.3089 - accuracy: 0.889 - ETA: 10s - loss: 0.3081 - accuracy: 0.889 - ETA: 10s - loss: 0.3082 - accuracy: 0.888 - ETA: 10s - loss: 0.3067 - accuracy: 0.889 - ETA: 10s - loss: 0.3073 - accuracy: 0.888 - ETA: 10s - loss: 0.3073 - accuracy: 0.888 - ETA: 10s - loss: 0.3091 - accuracy: 0.887 - ETA: 10s - loss: 0.3085 - accuracy: 0.887 - ETA: 10s - loss: 0.3101 - accuracy: 0.887 - ETA: 10s - loss: 0.3108 - accuracy: 0.887 - ETA: 10s - loss: 0.3124 - accuracy: 0.887 - ETA: 10s - loss: 0.3140 - accuracy: 0.887 - ETA: 10s - loss: 0.3125 - accuracy: 0.887 - ETA: 10s - loss: 0.3110 - accuracy: 0.888 - ETA: 10s - loss: 0.3102 - accuracy: 0.888 - ETA: 10s - loss: 0.3092 - accuracy: 0.888 - ETA: 10s - loss: 0.3091 - accuracy: 0.888 - ETA: 10s - loss: 0.3081 - accuracy: 0.888 - ETA: 10s - loss: 0.3073 - accuracy: 0.889 - ETA: 9s - loss: 0.3076 - accuracy: 0.889 - ETA: 9s - loss: 0.3078 - accuracy: 0.88 - ETA: 9s - loss: 0.3080 - accuracy: 0.88 - ETA: 9s - loss: 0.3080 - accuracy: 0.88 - ETA: 9s - loss: 0.3071 - accuracy: 0.88 - ETA: 9s - loss: 0.3060 - accuracy: 0.89 - ETA: 9s - loss: 0.3049 - accuracy: 0.89 - ETA: 9s - loss: 0.3045 - accuracy: 0.89 - ETA: 9s - loss: 0.3045 - accuracy: 0.88 - ETA: 9s - loss: 0.3037 - accuracy: 0.88 - ETA: 9s - loss: 0.3035 - accuracy: 0.89 - ETA: 9s - loss: 0.3034 - accuracy: 0.89 - ETA: 9s - loss: 0.3029 - accuracy: 0.89 - ETA: 9s - loss: 0.3020 - accuracy: 0.89 - ETA: 9s - loss: 0.3018 - accuracy: 0.89 - ETA: 9s - loss: 0.3016 - accuracy: 0.89 - ETA: 9s - loss: 0.3009 - accuracy: 0.89 - ETA: 9s - loss: 0.3004 - accuracy: 0.89 - ETA: 9s - loss: 0.3002 - accuracy: 0.89 - ETA: 9s - loss: 0.2999 - accuracy: 0.89 - ETA: 9s - loss: 0.2997 - accuracy: 0.89 - ETA: 9s - loss: 0.2993 - accuracy: 0.89 - ETA: 9s - loss: 0.2989 - accuracy: 0.89 - ETA: 8s - loss: 0.2991 - accuracy: 0.89 - ETA: 8s - loss: 0.2993 - accuracy: 0.89 - ETA: 8s - loss: 0.2996 - accuracy: 0.89 - ETA: 8s - loss: 0.2990 - accuracy: 0.89 - ETA: 8s - loss: 0.3001 - accuracy: 0.89 - ETA: 8s - loss: 0.3015 - accuracy: 0.89 - ETA: 8s - loss: 0.3027 - accuracy: 0.89 - ETA: 8s - loss: 0.3037 - accuracy: 0.88 - ETA: 8s - loss: 0.3033 - accuracy: 0.88 - ETA: 8s - loss: 0.3027 - accuracy: 0.88 - ETA: 8s - loss: 0.3020 - accuracy: 0.89 - ETA: 8s - loss: 0.3015 - accuracy: 0.89 - ETA: 8s - loss: 0.3007 - accuracy: 0.89 - ETA: 8s - loss: 0.3008 - accuracy: 0.89 - ETA: 8s - loss: 0.3011 - accuracy: 0.89 - ETA: 8s - loss: 0.3013 - accuracy: 0.89 - ETA: 8s - loss: 0.3017 - accuracy: 0.89 - ETA: 8s - loss: 0.3010 - accuracy: 0.89 - ETA: 8s - loss: 0.3009 - accuracy: 0.89 - ETA: 7s - loss: 0.3007 - accuracy: 0.89 - ETA: 7s - loss: 0.2999 - accuracy: 0.89 - ETA: 7s - loss: 0.2994 - accuracy: 0.89 - ETA: 7s - loss: 0.2993 - accuracy: 0.89 - ETA: 7s - loss: 0.2993 - accuracy: 0.89 - ETA: 7s - loss: 0.2991 - accuracy: 0.89 - ETA: 7s - loss: 0.2995 - accuracy: 0.89 - ETA: 7s - loss: 0.3001 - accuracy: 0.89 - ETA: 7s - loss: 0.3000 - accuracy: 0.89 - ETA: 7s - loss: 0.3001 - accuracy: 0.89 - ETA: 7s - loss: 0.3002 - accuracy: 0.89 - ETA: 7s - loss: 0.3003 - accuracy: 0.89 - ETA: 7s - loss: 0.3006 - accuracy: 0.89 - ETA: 7s - loss: 0.3008 - accuracy: 0.89 - ETA: 7s - loss: 0.3008 - accuracy: 0.89 - ETA: 7s - loss: 0.3004 - accuracy: 0.89 - ETA: 7s - loss: 0.3006 - accuracy: 0.89 - ETA: 7s - loss: 0.3000 - accuracy: 0.89 - ETA: 6s - loss: 0.2994 - accuracy: 0.89 - ETA: 6s - loss: 0.2994 - accuracy: 0.89 - ETA: 6s - loss: 0.2991 - accuracy: 0.89 - ETA: 6s - loss: 0.2989 - accuracy: 0.89 - ETA: 6s - loss: 0.2991 - accuracy: 0.89 - ETA: 6s - loss: 0.2993 - accuracy: 0.89 - ETA: 6s - loss: 0.2990 - accuracy: 0.89 - ETA: 6s - loss: 0.2993 - accuracy: 0.89 - ETA: 6s - loss: 0.2995 - accuracy: 0.89 - ETA: 6s - loss: 0.2991 - accuracy: 0.89 - ETA: 6s - loss: 0.2995 - accuracy: 0.89 - ETA: 6s - loss: 0.2988 - accuracy: 0.89 - ETA: 6s - loss: 0.2984 - accuracy: 0.89 - ETA: 6s - loss: 0.2987 - accuracy: 0.89 - ETA: 6s - loss: 0.2989 - accuracy: 0.89 - ETA: 6s - loss: 0.2982 - accuracy: 0.89 - ETA: 6s - loss: 0.2981 - accuracy: 0.89 - ETA: 6s - loss: 0.2978 - accuracy: 0.89 - ETA: 5s - loss: 0.2975 - accuracy: 0.89 - ETA: 5s - loss: 0.2980 - accuracy: 0.89 - ETA: 5s - loss: 0.2981 - accuracy: 0.89 - ETA: 5s - loss: 0.2977 - accuracy: 0.89 - ETA: 5s - loss: 0.2977 - accuracy: 0.89 - ETA: 5s - loss: 0.2973 - accuracy: 0.89 - ETA: 5s - loss: 0.2969 - accuracy: 0.89 - ETA: 5s - loss: 0.2971 - accuracy: 0.89 - ETA: 5s - loss: 0.2969 - accuracy: 0.89 - ETA: 5s - loss: 0.2963 - accuracy: 0.89 - ETA: 5s - loss: 0.2960 - accuracy: 0.89 - ETA: 5s - loss: 0.2954 - accuracy: 0.89 - ETA: 5s - loss: 0.2953 - accuracy: 0.89 - ETA: 5s - loss: 0.2952 - accuracy: 0.89 - ETA: 5s - loss: 0.2955 - accuracy: 0.89 - ETA: 5s - loss: 0.2955 - accuracy: 0.89 - ETA: 5s - loss: 0.2955 - accuracy: 0.89 - ETA: 5s - loss: 0.2963 - accuracy: 0.89 - ETA: 4s - loss: 0.2963 - accuracy: 0.89 - ETA: 4s - loss: 0.2961 - accuracy: 0.89 - ETA: 4s - loss: 0.2959 - accuracy: 0.89 - ETA: 4s - loss: 0.2961 - accuracy: 0.89 - ETA: 4s - loss: 0.2959 - accuracy: 0.89 - ETA: 4s - loss: 0.2960 - accuracy: 0.89 - ETA: 4s - loss: 0.2956 - accuracy: 0.89 - ETA: 4s - loss: 0.2956 - accuracy: 0.89 - ETA: 4s - loss: 0.2954 - accuracy: 0.89 - ETA: 4s - loss: 0.2951 - accuracy: 0.89 - ETA: 4s - loss: 0.2952 - accuracy: 0.89 - ETA: 4s - loss: 0.2951 - accuracy: 0.89 - ETA: 4s - loss: 0.2950 - accuracy: 0.89 - ETA: 4s - loss: 0.2955 - accuracy: 0.89 - ETA: 4s - loss: 0.2958 - accuracy: 0.89 - ETA: 4s - loss: 0.2957 - accuracy: 0.89 - ETA: 4s - loss: 0.2961 - accuracy: 0.89 - ETA: 4s - loss: 0.2963 - accuracy: 0.89 - ETA: 3s - loss: 0.2963 - accuracy: 0.89 - ETA: 3s - loss: 0.2958 - accuracy: 0.89 - ETA: 3s - loss: 0.2958 - accuracy: 0.89 - ETA: 3s - loss: 0.2958 - accuracy: 0.89 - ETA: 3s - loss: 0.2959 - accuracy: 0.89 - ETA: 3s - loss: 0.2960 - accuracy: 0.89 - ETA: 3s - loss: 0.2967 - accuracy: 0.89 - ETA: 3s - loss: 0.2969 - accuracy: 0.89 - ETA: 3s - loss: 0.2970 - accuracy: 0.89 - ETA: 3s - loss: 0.2968 - accuracy: 0.89 - ETA: 3s - loss: 0.2964 - accuracy: 0.89 - ETA: 3s - loss: 0.2961 - accuracy: 0.89 - ETA: 3s - loss: 0.2959 - accuracy: 0.89 - ETA: 3s - loss: 0.2962 - accuracy: 0.89 - ETA: 3s - loss: 0.2963 - accuracy: 0.89 - ETA: 3s - loss: 0.2963 - accuracy: 0.89 - ETA: 3s - loss: 0.2961 - accuracy: 0.89 - ETA: 3s - loss: 0.2959 - accuracy: 0.89 - ETA: 3s - loss: 0.2958 - accuracy: 0.89 - ETA: 2s - loss: 0.2957 - accuracy: 0.89 - ETA: 2s - loss: 0.2958 - accuracy: 0.89 - ETA: 2s - loss: 0.2961 - accuracy: 0.89 - ETA: 2s - loss: 0.2960 - accuracy: 0.89 - ETA: 2s - loss: 0.2962 - accuracy: 0.8923"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1875/1875 [==============================] - ETA: 2s - loss: 0.2960 - accuracy: 0.89 - ETA: 2s - loss: 0.2960 - accuracy: 0.89 - ETA: 2s - loss: 0.2957 - accuracy: 0.89 - ETA: 2s - loss: 0.2959 - accuracy: 0.89 - ETA: 2s - loss: 0.2956 - accuracy: 0.89 - ETA: 2s - loss: 0.2960 - accuracy: 0.89 - ETA: 2s - loss: 0.2959 - accuracy: 0.89 - ETA: 2s - loss: 0.2962 - accuracy: 0.89 - ETA: 2s - loss: 0.2965 - accuracy: 0.89 - ETA: 2s - loss: 0.2966 - accuracy: 0.89 - ETA: 2s - loss: 0.2966 - accuracy: 0.89 - ETA: 2s - loss: 0.2969 - accuracy: 0.89 - ETA: 2s - loss: 0.2969 - accuracy: 0.89 - ETA: 1s - loss: 0.2970 - accuracy: 0.89 - ETA: 1s - loss: 0.2970 - accuracy: 0.89 - ETA: 1s - loss: 0.2970 - accuracy: 0.89 - ETA: 1s - loss: 0.2972 - accuracy: 0.89 - ETA: 1s - loss: 0.2971 - accuracy: 0.89 - ETA: 1s - loss: 0.2971 - accuracy: 0.89 - ETA: 1s - loss: 0.2968 - accuracy: 0.89 - ETA: 1s - loss: 0.2966 - accuracy: 0.89 - ETA: 1s - loss: 0.2965 - accuracy: 0.89 - ETA: 1s - loss: 0.2965 - accuracy: 0.89 - ETA: 1s - loss: 0.2965 - accuracy: 0.89 - ETA: 1s - loss: 0.2966 - accuracy: 0.89 - ETA: 1s - loss: 0.2964 - accuracy: 0.89 - ETA: 1s - loss: 0.2964 - accuracy: 0.89 - ETA: 1s - loss: 0.2966 - accuracy: 0.89 - ETA: 1s - loss: 0.2965 - accuracy: 0.89 - ETA: 1s - loss: 0.2966 - accuracy: 0.89 - ETA: 1s - loss: 0.2967 - accuracy: 0.89 - ETA: 1s - loss: 0.2963 - accuracy: 0.89 - ETA: 0s - loss: 0.2962 - accuracy: 0.89 - ETA: 0s - loss: 0.2960 - accuracy: 0.89 - ETA: 0s - loss: 0.2960 - accuracy: 0.89 - ETA: 0s - loss: 0.2962 - accuracy: 0.89 - ETA: 0s - loss: 0.2962 - accuracy: 0.89 - ETA: 0s - loss: 0.2962 - accuracy: 0.89 - ETA: 0s - loss: 0.2959 - accuracy: 0.89 - ETA: 0s - loss: 0.2956 - accuracy: 0.89 - ETA: 0s - loss: 0.2956 - accuracy: 0.89 - ETA: 0s - loss: 0.2958 - accuracy: 0.89 - ETA: 0s - loss: 0.2957 - accuracy: 0.89 - ETA: 0s - loss: 0.2956 - accuracy: 0.89 - ETA: 0s - loss: 0.2957 - accuracy: 0.89 - ETA: 0s - loss: 0.2957 - accuracy: 0.89 - ETA: 0s - loss: 0.2957 - accuracy: 0.89 - ETA: 0s - loss: 0.2956 - accuracy: 0.89 - ETA: 0s - loss: 0.2957 - accuracy: 0.89 - ETA: 0s - loss: 0.2957 - accuracy: 0.89 - ETA: 0s - loss: 0.2961 - accuracy: 0.89 - 13s 7ms/step - loss: 0.2961 - accuracy: 0.8921\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x2783a1ed6d8>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "training_examples = metadata.splits['train'].num_examples\n",
    "model.fit(train_dataset,epochs=5,steps_per_epoch=math.ceil(training_examples/BATCH_SIZE))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Evaluate the accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - ETA: 35s - loss: 0.2553 - accuracy: 0.875 - ETA: 3s - loss: 0.3464 - accuracy: 0.867 - ETA: 2s - loss: 0.3530 - accuracy: 0.86 - ETA: 2s - loss: 0.3559 - accuracy: 0.86 - ETA: 2s - loss: 0.3629 - accuracy: 0.86 - ETA: 1s - loss: 0.3657 - accuracy: 0.86 - ETA: 1s - loss: 0.3593 - accuracy: 0.86 - ETA: 1s - loss: 0.3650 - accuracy: 0.86 - ETA: 1s - loss: 0.3635 - accuracy: 0.86 - ETA: 1s - loss: 0.3563 - accuracy: 0.86 - ETA: 0s - loss: 0.3596 - accuracy: 0.86 - ETA: 0s - loss: 0.3677 - accuracy: 0.86 - ETA: 0s - loss: 0.3701 - accuracy: 0.86 - ETA: 0s - loss: 0.3663 - accuracy: 0.86 - ETA: 0s - loss: 0.3709 - accuracy: 0.86 - ETA: 0s - loss: 0.3719 - accuracy: 0.86 - ETA: 0s - loss: 0.3737 - accuracy: 0.86 - ETA: 0s - loss: 0.3729 - accuracy: 0.86 - ETA: 0s - loss: 0.3703 - accuracy: 0.86 - ETA: 0s - loss: 0.3699 - accuracy: 0.86 - 1s 4ms/step - loss: 0.3760 - accuracy: 0.8645\n"
     ]
    }
   ],
   "source": [
    "test_loss, test_accuracy = model.evaluate(test_dataset,steps=math.ceil(metadata.splits['test'].num_examples/BATCH_SIZE))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy on test data : 0.8645\n"
     ]
    }
   ],
   "source": [
    "print('accuracy on test data :',test_accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "for image,label in test_dataset.take(1):\n",
    "    image = image.numpy()\n",
    "    label = label.numpy()\n",
    "    prediction = model.predict(image)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32, 10)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prediction.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([4.3992968e-05, 1.0571389e-05, 1.5527538e-02, 8.3221693e-08,\n",
       "       9.4727588e-01, 7.6189152e-11, 3.7136629e-02, 7.0900885e-12,\n",
       "       5.3553781e-06, 2.7345719e-09], dtype=float32)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prediction[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "label[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.argmax(prediction[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As above 2 values are same, hence we can say that the prediction was correct "
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}