DeepLearningHelloWorldMNISTMLP.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Assignment 1-d: Deep Learning Hello World! (3-layer MLP+Dropout+Optimizer)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Objective: To be able to improve upon the 3-layer MLP + Dropout in part 3 by adding different optimization for MNIST Classification"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Step 1: Taking care of the imports which includes numpy, datasets, models, layers, optimizers, and utils. <br />\n",
    "You will also be able to tell if your set-up is correct/complete."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "ename": "ImportError",
     "evalue": "cannot import name 'np_utils'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m-----------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mImportError\u001b[0m                           Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-9-efb294208580>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0mtensorflow\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlayers\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mDense\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mDropout\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mActivation\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0mtensorflow\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0moptimizers\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mAdam\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m \u001b[0;32mfrom\u001b[0m \u001b[0mtensorflow\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mutils\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mnp_utils\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     10\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0mmatplotlib\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mpyplot\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mImportError\u001b[0m: cannot import name 'np_utils'"
     ]
    }
   ],
   "source": [
    "from __future__ import print_function\n",
    "import numpy as np\n",
    "from tensorflow.keras.datasets import mnist\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.models import model_from_json\n",
    "from tensorflow.keras.layers import Dense, Dropout, Activation\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "from tensorflow.keras.utils import np_utils\n",
    "\n",
    "from matplotlib import pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.rcParams['figure.figsize'] = [10, 6]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Step 2: Set-up some constants to be utilized in the training/testing of the model <br\\>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "NB_EPOCH = 20\n",
    "BATCH_SIZE = 128\n",
    "VERBOSE = 1\n",
    "NB_CLASSES = 10   # number of outputs = number of digits, i.e. 0,1,2,3,4,5,6,7,8,9\n",
    "OPTIMIZER = Adam()  \n",
    "N_HIDDEN = 128\n",
    "VALIDATION_SPLIT=0.2 # how much TRAIN dataset is reserved for VALIDATION\n",
    "\n",
    "DROPOUT = 0.3\n",
    "\n",
    "np.random.seed(1983)  # for reproducibility"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Step 3: Load the MNIST Dataset which are shuffled and split between train and test sets <br\\>\n",
    "- X_train is 60000 rows of 28x28 values\n",
    "- X_test is 10000 rows of 28x28 values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "First 100 train images:\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiwAAAFYCAYAAABwLKejAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAgAElEQVR4nOydeXxM1/vH33dmkkkmkYhsluySCLHviZ0WbS1VFC1FF1VEqaLtt/11U13su6Iobanat5ZSVCV2UUsiIRIEIZGQPbPc3x9DIgQVM5nccd995dXcLedzzNxznvOc5zxHEEURGRkZGRkZGZnyjMLSAmRkZGRkZGRkHoVssMjIyMjIyMiUe2SDRUZGRkZGRqbcIxssMjIyMjIyMuUe2WCRkZGRkZGRKffIBouMjIyMjIxMuUf1ODfbCmrRDgdzaSlGHtkUiPlCmRR2m7KsH0Am6amiKLqXWYHIdTQH1l5H+V00D3IdTY+11/Fpfxcfy2Cxw4FmQgfTqHoEB8SdZVLO3ZRl/QB2iKuTyqyw28h1ND3WXkf5XTQPch1Nj7XX8Wl/F+UpIRkZGRkZGZlyj2ywyMjIyMjIyJR7ZIPFROjaN+Li6tpsTj7CxdW10bdraGlJMjIyVkTckkZsuxxNp5O3UNYKtrQcGZknwnWfC+6RFR/rGbMaLIJKhapK5cKfc1OaEzevKfnb/VAG+nN9Yw22XY7m54v7ODu1uTmlmBVDmwbMXDybY2FLMGDgWNgSpi6Za2lZZUJ2r2b8X8JRhMa1LS3F5JybFMbm5CNsuxxN7otNLS1H5j+idK2EKsCPy2PDSfg2DIVGY2lJT4wytAYb2s1BK+oZ7nKGjLqVLC3J5AiNQhHD6nHpw3C2XY5mc/KR+35yt/mjsLOztNQnQlCryX2xKUGH1JaWYjHifmjMTJ/NHEj0e6znHivo9lEoawYhqm243KYiuc2zqeSczd56v9533+85Ffh2dmcO1PmF89pcvkl5lqp7pbkJo7ZjY8bNXU6wjS0GDCRotdw0qGmghvznmmC/6wSGvDxLy3wgud2bkuuqpNLiqFI9f62xgi8Tu5pYleW5Ojqc3X2+QyvaGk9I8+v5VKGoHUL8h/a8XieSMa7bCs/X9BxK0KAjFlRmApKvMjKuL3+GrrG0EpMjhtUjfpAt09qvwEbQ8Yx9JlpRgQHDfff+WXsV9Ze/jv87l9GnpllA7ZOjdHdj15z57M1TMcm/K7rzZR4nbFHi5jXlUMdpZBpEnPbYP9azJjNY9G0bMnXpHIJtbB96n1bU83+zBqHKFgn7bQQVknWoU3PRHD5gKillgtLJiezWIYye9gvt7LO446xamh7Ozrlh7PtsJn8umk+tn0YQML50xkBZcLm1Ak31DFhciocVSkSfXDp4xLJTCDe5NkuS5W2gkuLh32UpUdCpMUmvGnin4R5GucQBUGdRBJorIhnh+fj+rMB222ELqywdQpM6nB2tZHfL2bgr1ShQsCXHhYR8D4a7nGF564V82WQg4qETlpZaavQZN0m6FAShllZiesQJN4gNWfuf748OX0ynZsNQb5GmwXKHVnY6vvKphOIpM1jaNoihgsKWYUmdcfv+8fpGkxks6jOXOZLnTbBNyn3XxlxpTkKWG0urr+amQcRzZmSx61IcvF5aVo1DTebcd/4Lj0P84RjO4MSO/Oi3A6da5ful+rzLb3wb07FUzyqr+xLbZjH1D/anqoQ7g3vJ6t2MNT1mAALzM0LY8XJjHJJOlTDekwbXh4Yxa9wcGqv1KFAwMPEZGjhf4PibMwBQoCC8Uj8qbXvEHypnKN3diZtRjU3hcwmwsQGMLvYlt7xZ37MlBrUNwzefobFaT66nPVKeSFB6etCqZpylZZiF5N3eEGL8PSpPzetb3wKBwo6hecM4lvhtt5g+c6EUrDOENLd7U9zGnCe/jxLdlavFrl0bFs63ntP46ZYv6R/6oODx+keTGSy6K1eZ9W1vvuqcjfJfR44PmwXAhNS6nH1Ggz7jCq+EDSNxJPhz3FTFWgRd+0asqD8bBcYR+OCkDhzeUZMTb8xmV64dHodzOZsegs3EXSjKNMXP42Mj6Er9rGpRDgC555xMJcfi5HVpyqdfLybYxvjB/biwM5VPRz7iqfKJYGNL3jP1WPPhJKqq1LyR9CxJk2vgsCWaXRof9qwLZk3QRgBuRbsitaiI5P5BnGozA7ApPPfTLW/WvxiO/kwcQgMrckdUcOD5SocKD681Eqj4bzD609I3Yny+OUyPVf0AEAq0BJ0v7m3PcHNlx/4KPGOfCUD7E31w2iXdAcQd9KIBrUaFtUWy9P9mM4OdLvJMo3ew21zcYBk4fCv11Wre+rIHlfY+/syDSU28SkuiqD48Ge+JUYT+/ToAGxe0QZ9xEwAh6jj+/aRtrNwJsA20UWHAQJfY7tx4QaBinEit5SOY0a4Tir3HcFkahVbUs6bu4nK7YsjQsj6t7JJL/byfg9E69t6hN5Uki3Olfx7t7POwEZS8mdSJyjOkaawAXBnRmG2L5lJVpab32a7c6KZEs/YAYn4+l/vXLDRWfs+pQOD3Fy2s9vGp1i2x8PfVWZUJP/oqa15qhf7MWQDS61iPIa0/e56PN/UpPD71ykwSX3KzoCLTIWoL0J85i/7M2RLjOVJeCqaObWrh8eXLlTDk5JSlRLNxrZHNo2+SGFcKKmLAgM6++Gjd0KYB3R1PohX16OxKN5I3adAtUBgIpb1l9D6Evnqa6/OUYJB+pyY0CiX1vVyCbWw5kg9/ZdUibaU3rulROP+0H2fgXn+Fp1JN2qgcPHZZQvHDSepij4eydCsoVH4+9Kpk7PDsz6cj/U8XVF7VONVqCVpRT4wWLkwNxgFpxVbdIX5WM868NAsDUPPPoYS8n1gsSHHoOxsKf5/w1UBcLpbfOKsH8paaWsMj8P5Tj8Opq7glxRX7HuZ4lnP35mNS/f390M/SKsqW6++EEdI/Fk9lkR+i5rjzkm5vRK2WOG0ewTZ25PoXWFqOSYmf2Yx1rrOYlxFMxf3Jhf2hsqIzqe9nU1WlZvTlcDx/OFKqUBCzTaLVHB/H4KQOLPHdSWbvJuYqpsxQaDTov7vF/sbLOa/L472PhrPvuep47Ln2yGebVimfQVWqQKOLNS/28dbCA1yc7kALtYEfbnlBxi1TSytzlKE1aLwlsfC4z9qROKyRprFybkpzzrw0l5uGPDqf7kWNiLhCY0Xh4ED6wDC6O55BgYJav0XgslSCxgpGr0Pg6P2otx5Cl3S/h0jbJNMCqsyLjaBEcfs/a+baiHCeO5XB9o8mF4tf+fJ6Q8R8aXfy+pRrjDzX59E3SgxljUCWd5lHjqhl7f86ort4qfBa/Fx/Ihv+zI7cCsQ3yUfMzy9VGSb3sNxBn3GTtHdqcmFjLh9MWMaHL/dAPOaM91dRIEovzDa3TSjbQoy5Vd58dzQV1u+/z5siVTwO/7fZYKWbKyk9g6n08iX2BP8A2DFvzot4pEh32uQOSd1cWe16DFDyyrmuBH9zTpKjOKWnBz/2mIsBA71jX8H22aTCuX5F/VrUXhzDBM+ZgJoW0X2p8VmMJOtZEhf+LxydRiwM2HwpyGiIjbjUFvs/jkoyuP9etKK+xOW+UkYZWoO4wS60aXmy8Nxm71m362n01J/V6ugzbww+61IwZJ6zkFKZByG2qE/fHzbTWK0n5I93CV5/sPBa4oQwDreeCqgYv+h1qlH6/sJsBguA4XgMfT8fy8+fTia6+TJoDqEOIwhaeAVdQqI5izY5db+MRoGCwUkdsL/rw3gYNoISrQhKoXw3lbmVFMX24jS0aoCoFLj4jJqCqloUtnq2t5qFjQBX9Wo+SejBDYMBjUKP54FMyXcENwaHsW7oJMCGoRfboB2oRn/9gqVllQrBTk1jtdEEsR9pi+DrTfxQLzo+c5TRHgvwUdljAPSiiPCrG/qMeMsKfkKUTk7kNQ3C5sMU/g0xBvob3zvjv8GuXA2Xhvgg6mIsKVPmAYgt6jNoyTq6O6Tec6W4B2nk2T5U+zbSaozrOzhWknYsjmBjy5URjTn8/qzb752Cl+ofZeO3YQR+fhxFZQ+6Pb8fJQL1I1/H55snG9ya1WABqLQ4ihFnhuP0zSVWBGzj1GuzCfF+kxqfK9DHJ5i7eJOQMSCMjz0nY8CWI9tr4fMfLcQ7o6E/YmoRxFEzq3x88vNsMCCy5KNpbBxRv/D8eNdFKBDIFQu4rNcz+3pbntkxiorHbKmyPQUh6RLXY+zxVGolndsCjKO7yAmz4fai16hLfngnnnz4Q+UYMS+fA/k2NFNr2bBjZbHR+I5cN+K1Iu3sszhcYEvFZdKcCgJjttCCNnUYPXc57ex3kqLPZ1euC/8X150VoUupqjLGPNgptCS8XJGAM3blOoHj04wS8b4prjuDvTv8UXMdrV4djvPP+8tYnXlZ03AhEbSwtIxSc3VoYw6+PwMDoBVh2a1qTKx8gIn9D/DRM8141vl32tlncSDfDp/eT95XmN1gARD2RZPTy4MmfSI4MH4Gse0W8apfR262LIvSnxydPTgrbInKUxOw7PIjp4IUGg2xk2sDR3g14TlC3i2fQWKB/Y8R+vUIvJsUXym061ow13/3wvWUFts/DgFagjEmFdMDyePDaaKOYmVWtbIXbWLiPtIUjsYBfL6RZl6gO+hTrvHpO28yef5c6toal/lO2NON4KV5qFJu4rHiBu28/2LgrjcLP1OpobCzI61PA/ZOnAlA6IoIvHbpUW85hGuVLFZsa8QYV6PR2Uyt5d9BMwm7OBLPZcclv7rk7o7cKfzR8XPlHWFfND+82JkPBrnis60AZW7x1jX+DRtiO8+zkDrzcfGfotwzUuX60DAix08n06DltNaB/73/NnZpBeycmMgSv+1MrHwABQoMQGPbAkafjWFGz5cwHC+9t7NMDBYwNqSeM6+RN06HRrBlod9muvQYhWaddAIb0/SOj5zKUmg0nPmmDrHdZ/N7jjOX5wRSIb38jgr8Pyx5lF2FB0+JaFpfB+DjXT0J5r9Nj5VHDG0aMKHx+sLjZ0/2xfGwdL0rd7DddpiP/Iv2PrrzGWV2b8oWnw1oRQX2idLM4iuo1cROrUtsd6Ox0v3MiwRPSkCfcg2Vtxf1Nl5grOtpbhoKaLZmDFVCrrGzzq9EfTKTPv26kDqzDnZpWgCUu8uf1/NR3B3DsqfeCro1fwP2/2thVU+G/nQcAeNKvlYz3h06l62essDxotHqrCCIKGtJM59Orddi2JjtycQF/agyJRLN7RWVaWPqMnpWK6ZV3Vt4r1IQGHuiJ1WPn36iMsvEYDG0rM+53nbUrp+IRjA2lLNuNECzQVojvPf39SaYB+9JYmjTgGvv5RLTeDYdTvTBoXMCFSi/xsqT4rtByr4I+GrpAmrbGOvw/pXWOPezjuXZD0Jnryjs8PyXXpBc0LigUnFmej1iu83hki6fbt+Pw2/xOXQp19A+04ja3x7jU48jLLnly/L/dSVw7X6Ubq60fTaC7D43WddgIV4zjVNFm7NdWRAcYOEaPT4hf73J6fYLCo/jhtgSbL1NDCkvBVpagllQ3H75lIKAwV6auViObKvFjZVuVDlTPEQi19OOCPe/ABuafzECt+PZAHifTX7i9tWsBovQuDZxI21Z2OJHWtsVLUXLF7Xsv+EPhivmLN50CMb05TNarmAOJW/rnvRFGGtem0qwjS0NDw6kao8nsyRlzE8DW0XhdFDUkoZ4pEt/tdPDqLByP0yxtIrSc3FsU2K7zeCyLp/e34zFb30CN9r7I/avwOraM3BXqgldGUHwglQ0Z4yjPX1qGk4r0nBaAb2GjcOz1+0UA2MqAqcsV5lSoo6zh/aWVvHkCGo1Gb0b4LLhFIbMkpefXxkTzoaR34HV5YIFl6VRzB/ny1DnJOJH2xLY39KKHh+fz+8Pgla6u3Opp45AGzU/Z1YptleQKQaDZlnMr/L3JemLMAb9vIUzHRYWM1Y+SmlM+/+NRttWIsYKgAgGDLSxTyN+aSMMbRqgrBFI6pAwDDu9GR4fx4k3ZhNb4Enzo/1w+97h0X9T4igFBenB0hwZAFxcXRsbQVn4U2X3vasUrI/Mvs0tLeGJmPeWMa2AnQBdh/5N97/+ZfekWeypt8JorPwyksDxhwsz3d6Lx9xIxPbJxp9j0jNWALy/jGRFZrXCXCznOy9CUa+mpWU9Fnldm2K73YU9385CqOp533VVlcqkRITza8TkwuDpFH0+NrnS9ujey+T9nUjU5RD8tvSmgx5E/JhAYjp8T1S+klXdWpn875vUw6Ly8+Fmoyr0+eIPhlYsvvvmmCvNiZrbmEpLD+JikObqBDtBRcyz8/mnlR3x+ZUZ7JxYeO3dy634I7I+Qe9asX/2LvSiwYxpB82LoU0Dptf/Ca2o56Yhjya/jyIkyfo9YjcDJPqB3ebvrBCaqU9QSanmI7doALrEvsSFKC8CVt8k8NQRRJ3UJroen6UXwukX+htAsZU0UqHTV3sKg6JjP3KCrGbFrvcNj2K9xxYMt/eIGpjYibNLauC6Vpr9xsPQI2DItY7Va8pawXzZYyV6UWTwxqEExpm+LzSJwaKqUpkbix14x38P/SoU3615RHJLjs6rj9vqk1TKlOYXznP3Nca/Hca3lY36W9sV0NIuEYBj+Qr67RlC8OAjBFlxvEpJ5DSR5oqLvEq2tLTLBpRsy/EheMghK0vFVTLV9uRgM0IpyU4OILJdVZq92p6b9QpQXbcheH4yqqvX8Mu7+FR8fnfIX1oZJllahWmIeeb7B1xREJWn5q0DrxH4Vjyu2dLsOx5FdZU9aYOb4vqD9Ov38trd9HC8RsP9gwkcZZ6+8IkMloJOjSkYfYOPArfS0T672LUUfS6tN44h5ONYKmVESbpB0cedI763H7UiIjj98qzC8yFbh1Fjbg7Bxx4ciGutWOvW6NaMsC+apbc86FchmZzQKtjelTpbCujTbuA5M5I7kwjW70spGZfoG8xJr8FwlzOWllIq/hrZgmXDmnK8xeL7rv10y5sr2oosPtqCwIV6AvZFS7rveBhL2iwm3ZCL279Zkk6lcIevNvSkX/+Z2G8136ajT2SwJL6oIK7Ob4XHczKqM2NPRwS9QMiE8wSlHLCaVRe6hEQCRyfSbXTRvkjBHLKKL9rjkr/DHX196TYjTtFXibjUnvneeywtpcyZ9n0v+r0/gyqfnCUto67kl8Q+jehPx7GtthPbuNMWSSuLr3L3UfwPamg08l1+fHs6tW0F2p/ow83dlfH9NRnd+SSCHrIa01oYG9OLXr7HUGTnW0U/GTA+im7jm+CK+bxFT2SwBL9zkC7vNCp+7nbOB2v4AGRKpvK0SJ6f1pAAoi0tpVTozidxqTl0odGjb7Yyqi0/Q58Xu/Br4Gba/F8/Kr3ijD7jpqVlyTxlGHJyqPZNJB99Y8wX5EgCjiQ8VV6zSl3i+AsHwHqCbs2N7NeXkXmK0KemUdBTR80db7On3gr0Ib6WliQjIyPznyizTLcyMjLlA31qGkED0+hGE0CeEpKRkZEGgij+9ygMQRCuA0nmk1MMX1EU3cuoLKDM6wdyHc2CXEeTY+31A7mOZkGuo8mx9vrBQ+r4WAaLjIyMjIyMjIwlkGNYZGRkZGRkZMo9ssEiIyMjIyMjU+6RDRYZGRkZGRmZcs9jrRKyFdSiHWWzsV8e2RSI+UKZFHabsqwfQCbpqWUdQCXX0fRYex3ld9E8yHU0PdZex6f9XXwsg8UOB5oJHUyj6hEcEHeWSTl3U5b1A9ghri7LyGtArqM5sPY6yu+ieZDraHqsvY5P+7soTwnJyNyD0CiUN+LOMzxezkApIyMjU16QDRYZmbuI/7EhX6xeSneHVKYlPmtpOTIyMlaOom4I6VuCWHNpP0KTOpaWU64xX6bb5nU5382BT3uuYmpcBzJPuAJQ/YtjGPLyzFasjExpUPn54P9bCpurLsQATEmrjWaQ9qna20RGRqZsyX++CUvmTeNAnjd1to2g1sWLcpvzEMxisCR/EM7WYd/ho3IE4NVGq7izz1zLI2/jsOaAOYqVeUyULi5cfKMmqjzIqF+AjWMB/7SYx+vnehF3tXjMk+6aPf4bdKh2Wt8uqkKjUAq+u8WUqv8ACuouHYnHEQOaZCv5ngoCNzYFsarOYoY/OxB93DlLK5J5TFJGhiO2T2dg4AGGVDwNwOjkDiR30aC/ft3C6soGZUVn6u9Kp4PTKaZ064X+1BlLS3oibvZvzuavp9Bk+yhqjo4j+NZh2Vh5BGYxWHx/TODyEHt8SvjrC6dM4w3Ve1T4db85ipZ5DGK+DuJs19n3nLVnQ9AWCLr/fl1PPTPTQ1iwpSOBy9MxnIwtE53mJs9Dw7aQpYXHmmQBzVorMVYAZYUKfBWyDh+VhovdPak6STZYpILKqxo5S1QcqjWLGK2WcQk92X6tJgsCf2W+117m7/FlYy1XS8s0C8pAf7RVKgJgk5pFcid3NnnMZuFNb7gqbSNNaBDKuomTeTOhJ8FvHUNv0FtakiQwi8Giu3KVNxZGsOOd76iicmRjtoZuDjkA1LTVcOVZHRV+NUfJ5QtlrWAMDmriX3VgRfdZAAw6MhjvXictrMzIhHZrih1HF+iYcrlTsXMHzvvRzD+RIMdr/J/bCd5ziee9/vG0ODEM5/JRjSdCaBTKsBmrUNwO52rxvxF4LI20sCrTor91i2XXWtDB9y/y3Kx/K47EL8Mw2IBdjZscbbocgPkZAWwOdbGwssenweYL9HI+TPCGCGp9eQHxSjIAQ5q8w7zV83jTOYFpU16g+hjpDwDFFvVJHCFSu9plAF6tvJNuDukA1Fg3DJ/gyygFBbvTayDY2VlS6hOhcHCg8rwLTL7emoKeOrBCY0VVpTKJgwKY8sYPdLA39v21fhpBwPioJ/u7phBXEl5fR7KkXyM+cjvD2fzK4JBQeC1kZhYGcxVcDsjq3Yyr3QvY3HIOwTZ2GBC5E988stYu1lGmaQIeyE8vd2RWbWdcTt4EQJGZiy4hsdg9gaSTBmS4erJpfxJdNbcASHs+D+efyliwGYgb6Eh3h1S6xPZAOdQWl/gne6HKK7GLa8Lnf2EXfNPSUsxC7otNSQ1V4dTiGsfqzkApGFNV3GlnBjufQXE6SFLeiKyXm/Op+xyaHx1A8LCDxaYLxEMn6LD+feJ7zeXbrr+wYEyAxXSaiosdNJxqPavwON2QR4MDQ5hadxVneswFQC8KxK4MwTNZuoOK2Km1WF5tOq91eA19qvV5O/Ofa0LNCcdYX3kLBgyF7+DJ/jNpUP3JBuzmC7oF1s5qjyFC4GO34lMHBjsbcxZrMRJ/rUu3oBN84znv9hk7EnU5dNwbgcMxe6rNP44hO9uiGu/GcDwG5+NFjfrDjMgrfUPoqtkBQLohF+/FSrPrMzc1Dtuw3HMqq7N8EN53Rh9/ytKSzIbHXqML/e/Gi+gf8Mp9hqnUUAX44fFLGr3cDgEQYvMPXio1ChR8fK0pX3gcKna/jaDE2yYNkI7BoreBZbeqoVxTsubqv+VBL3BX3ULp5oo+Na2MFZqOs9Oa80/P7wB76kYOIi/NnlpfXabaxVNMatMfpyWLaKSGQ/kiVRYfl+yAV6HR8OOzC3k9oadVxpKpvKrRadJOhruc4WC+koH738J9sx0F/W4Q2fBnRtbaxQY771IvvDGrweK6MIqoHTWYtEnL2EpFH07WF9k4djZnyWWLqlpV4ie7E9NyCScKtHxyrQnb57TALToTRXY+gTHHgIcbBOUVhZ0d8YtDiGw1CbAHoO+ACGx2Szv4Nn1QGFOqzMaALR/v7EnN7DSszzFbHKWgwElhR9LLVan2TaKl5ZSarN7NeO+rFXRxuLuDVgPQtftAlFdu0L3KYLJ9HBjz7c88pzFOKyy63Bq4WvaCS4nL+hOs2RSMy62SvX7KPKPPpYXaQNJbNfD6WrpeB4NGj4dSw/rsigSMz0SXcAIdoKhfi5vjM2miFriiz+GNRePwypZuPeO/qIufajs5n1VFKaHv4n/Ff10qw13O0DO+G9q2V6hONABCTCgn14o873iGNU06oth7rFR/36wGy7UR4WTU1rHRZR13p3y5sb8yjiQ8+EGJcfrLqsS1/p7A7UOo+V4C+vR0XIlCBEl3gtk9m5HWN4cz4YsBe7LEfFrMHoP3IemOcACUnh5cDy9ysNtkKO8b7Vz4NJy8aloAgocUH61LFb1o/NQMEndwVhl5rpixckOfT4cF46i8Px+bw0eMUyfJl7k8qkGhsbI6qzL6V6XlFXykN/ZEPLMyAoiomEBOgLZsRJkJv7Uis1oGMLziOT6d7IDP687gVgntlEz2hqznRIGOvsvG4TtRusYKQIfWxxl5vifKXUctLcUspBYYU/gn/uFPNa7cdz1e64zN1Zul7hfNYrAITerw4o9/8ZrTdDQKW+7NT+e39oakOzwApZMTZ76oxdfPr2DyV2G0+HsEIb/9i74cTfk8CdqOjdk+YxZqoegrYhBFHC8aEHUSX3yn09GqzhlsBCVaEar9XVSfpC/CQBT4ot/P9HC4AYDNZSXPt3kJfbz1GNlSJbNPc+b7TOaOR2VDthtzR7+M95b7O7KAKqmFv3/890sEX7IOw/MOYn4+WXrpBp/ejd3e08w90Zrhrc4xte4q/tfjTT748OfCoNtXFo/G90tpGyt5XZsyrepserz0BpBSeD6nRzPsr+YhRB23nDgToUBEgQJ1uojCwQF93UDiB9nSok48e7JD+OPt1gjx0aX++2YxWNLqONKnQjwahabE62fGaAgaaI6Sy47Yr2ty5sU5ND/aD4/V/2LIzpa8EXY353sJxYwVACeFHfu+m8tH7zdkzc7mBKzLQ9hX+i+fpUh7vgbrfGaiFRVszHZBnZKDCBjaNMCj2VX+rL0KgEu6fLZm12SIcyLBKy8QNyAY/Wk5Xb8lqTz8HFVVRmNlxKW2xEyqjcOW4kvQVZU9udyzOr8GTgJsGXGpLdW2WV9Sb4VGg5vKGJukyJKW9+heDDk5aDONn2s7+zwiJ8xGgYABCP37dQJXXZO0txoguV8BczJqojh5DgNwdXQ4y9+dSqjNUdINubSePxbvr6RtlA2vshMDBtoOOWRc2GsAACAASURBVIDibZEJnosAaPHZSK4vzEDgyfoLs7zFlRZHEb7wfVL1JXsbqnhmmKPYMiWhx/coEFCudi1XgbSmwnc9dI3rwpGC+5uJiR5HOdNvLptWLSRuQRNUvt4WUFg6lK6VyPQzriDZlWvH2N9fQTx2CqFRKKnv5bKz9mqO5Stoeug13ho4kl8/eg6AIPuUh/1ZSaAUpN9pX1gWyMTU+ryR9CyX+7rhsPr+fDlxowLY/9EMfFW2TL9R64H3SR2xVgBvOV8EwGdb0Xuq8qpGdq9mnJsUhtC4tqXkPTZ2F++fq+wS252AyTr0Z85aQJFpiW49n19ndMSQk4PCzo7/e+cnXv9qNF3rPkPHr95n99BJ3OrX3NIyn4ijuf7c0OczsfIBJnoeBuCyLh/XE6bpI80Ww+LzRSRdz44hr6KxkRRVsGbMd1S3cTRXkWXK2KsNmOh5mE8/WcLE3EE4rpJ+HoS7UW89hH4rfFbzFQoqVyC7ii1p3XI41WoJCowdvgolZ1/4nsF12pLSQimJfALpnYI5NnQGAMM2vEHQmP2o/Hwo+O4W+0PWcl5XwCv/RFBjaCz6+kG8MnEb53V5TDn8LEGnpT3vfCeGRcq4Lopi/yIbIOP2T3EyXgsj6tXJgC05Bi3LVjyL13lpj1rvRaHRQJAvyW2cCs89+93fLHu9Kf1DDlHXfhcvaLJI1OXQLeBtvHpaUOx/RFCpqBiWUti2ADwf2w06XELkkgWVmQYxrB72wlGE27PPeW3rMP5QTaovjEIPuM+PonlYBO3ePcmlFRaV+kRsDnVhbefRZHqriPrMmJS0y+G38dr/r0n+vlmDbp1+2U/hKyUIdAx4n3Mvz2eY/x5+rtVBcu71gk6NsdtzEkNeHqdf8CRk3HBiX55DyKTJDEt8Bw6esLREk6OPiUcZA06A0y/QdEQE7Qfv57vKhwvvWeKzm5oThuP/UfnPYZJWp6hBvJNsy/+3lNtp+eHNd0cTtP4guc81YdsiY+6HkC2jrCbwFsDthMRjkB7CP18bV34BtJr9Pl7fStdYUVSogOBdhWthrqQ10dOvmfH76mF7mYiK/xS7d1SlE9SobwxyHL3pNWZs0WKbko3XSWks1b+5yZe/664qNq1uQLCa3XlzqhljjSqdygLA/tA5apywK5ZbJ2iejvlr9tDlzj42EsX2j0NUalG/8Nh7IpgqXaVZDZa7Udjbc+7l+QBk6u1AV/5H43dQBfjReF083Zzm8sbUUXjOikR35SohU5TwMvio7Ml3s7sdBmjdeMyO5NT3try5tw2LvPcUXfDPsZyox0DrrEeBgg4ne2HPeQxtGtCj0jIUKKi7MAKf9ZHFst/WXRhB8GfS7fRKwvG0dS7hjp/TDBshGu3t1tFr5y2TNZRliaJCBWIn1eT91r8z1HlPsWvndLkkaiuSKxZgLxgNs5p/D8ZngbJw5UkgRsNGCv40ZaA/8UMqE1t3Dgbg02sNWBNXn5Mtl1C3YjJWkEy7GMpzyegBfdqN+6+dTS57QWbiej1N4aIGU1JmBkvstFDA2PBPW9sNv7jyPxq/w/gd6wlSZdFhwTi8ZxV1XjHjvQDoc64zmoMJVtkJlISoLWD3iXpwl8EinCs5wLo8YsCAQSzytGhFFQbyIDSTkWdjcVce4rf0pix9oQP+qTFPzecqZRR2dtSunYRW1GPAQIO57+J9VJpxK/Zb7DhbfT7phlyej+1D/EVPqm5WocwTcYhLQx93jsSYa7zhdImVWe4EDruAPj3d0rJLxeXnq3D61VmAQK3lIwiadAb7Vx2hJWxNCMUH6/Bai8KjY8jSng/mvG5jGSkyL6pcEa2oZ2DiMyjydCZrQ03icVNVq4phpzfXhoeXfD3Ajx2dpxUeB6yS1sv1xqp3ADjxzmy2Jh9la/JRXo65SvxL89iY7cKtT70lnWXyblQBfpyb0pzUt8MeeI+gUtGsVlHeklyxgMoHpNGt+24ymvw76/xK/nNNuBxuT4CNcbQTHb6YjvbZGEQFe6c0Qx+fINmO4GGI9raWlmBSlE5OXHy3Ib8FbgKgyaHX8Ft0VhIxVSWxNvBP1mS70HdABHS4RNCgIzisPoDd5oMYzl8kbn5TnneIY38+LBnaXbLf0ZwezVjx3mQAwj8YTuCX/0Jld94bZlylp73gYEl5JkUQHx5DJqjVuA5OovPuiDJUZSaa1qFS/4sczBeIWxpi0l21TeJhuTzXiWM1V7JgRFV+Su6CQ2IWhujT6No34kaImp5D/yoMtvXf/BYh506botgyI+CDKNrqxqKpk868Oj8DUMfuIi+ceRHGuaCK/leSrud7Ufn70nrDKTZWWkvX+p1KtIpVfj6c/qAyZ/3mF56bk14Hu00Hy07oE6DMN3BZl09VlZo/F82/vdNFUQd+XpfHK/9EEPSzdQVR301SNxe8pZ/yATCu+rr+oxtHGhgDqet+H4HPhAOS3v1WLxqIya2Gat/JYu2KwsEBw0YXzobMJ90g8MGYd9DskqYXCSC5AwTb2DH4QlsqLo9CVKtJ6uFKiPoKCpSo06wlggWcj6awO8+GpCEh92UkFtRqzi+rwQiPPQgDBKQcYaZ0d2foz2vopLlJ3cUj8Vto2pkUkxgszvMrMLJaE2ZWPcSQuQtYk+XED8ktmR8wA//bhopeNDD/pi81x8VJMrma38fGf/hPiwVEJd/+sQ6uzVLzfiWjNayt5YXqaB6GzEzAOK8e93ko23tOxk9lnP5RCgrOa7PY8kk77JGGwaL66wj9/vc+Ae+c4Ue/HYXn6+17HeF0BdyjdQStl0ZdHgcx6VJhVlRrQl+9Gv80WAzAT7e88flC+vFGP9zy4mO3k9T+ZSBVXW5y/lRVKiQqePPNLQypuJsxV8M4OaYumt3SNVYAEMGAiEFUIKjVXB/YkOPvzOJUgYGaf7+Jv8Sz2t6NLiGRiMVvs3vYJNoxFt+FZyio7cv1+naMfHstqbob/P5iI/SXpP1+xs+oRifNTZoeeg2/T0wf9mESg0X9+yE2vdSEnWuacCpiLj0db9GzxlagaAnzKW3B7Z1SrXO3WGsg7283aGD8/Y9ffuCL1DqcyzbuLF3d4Tqb3eYCRbEq57VZDBgzBof10mo4nX/aT9pPFIvG97WSufIHYcjL41qBcc1etbYX4UsLCzIBQpM6JLxnjEVadDOALb3CAGmtPCyJNTU9+GZmVw72mIqNoIAQ4/nOJ/rzy4TncFqxHyXSXmIPoHTNByAu3Z3QPVfZ5G1cBvv2p6PwXyadGMf/ivdXkbRlLLuHTcJlhHHV0HldHp3Xj6HG+GgMedI2VrJ6N+PvllNI0YN6fUWzlGGyoNvgtw6h0Gio4WiM93Coc4OjjX8FIE6bzXuDI6ziJbNmvLbeoEnLfhxqZEwE8H9uJ8Dt/vtyxQLqbB6J3zoDDtukZaw8zaw+U58vPaLxtM/kuqXFmICUj7WcaLwcgLk/dcXrtPWMyINGHuDVkS2KnXPiHGA9O/yq4jTQBvbVX4kCgRMFOnpueJfgdSclscKpNHh/FcmAr4p/rkHsl3x9laE1mDfJOC370udjcV1qHoPTpKuEDDk5+P2vSGgnitZiy8ZK+cdwMhbPvhqaDBxOVuschHMaWj9rTPizJyEQAMe/NVSKySd4t/VNm1g7gV/mM2ZZU45tqoUX0u7cxbB6eDgaA91r7RpC0HZpLmF+mgmYf45Q2xHs6D+Jj5Of5+D22gR+Gin5zvtpQ1GhArHjHahho6RV9ABcfzCfd6zMljXLSANDTg7u86Jwn2c8vvCx8f/+WEmU5lOM/tQZYhoheWMFIH6gmtiQdazL8iBouhbxsLVl7LB+dFdT8P8ohbc+agncwtcKvpdPIxci6nCmwyzevdwK197JZjU4rScMW0ZG5qmh2g5j7MrUr/vKxoqMjIXQdmzMtNcXEvLbcBJfrYYhx7wJRGUPi4yMjORwWH2Abqub4IL1BWfKyEgFm+2HmRIYSiD7yyTBpiCK/33mVxCE60CS+eQUw1cURfcyKgso8/qBXEezINfR5Fh7/UCuo1mQ62hyrL1+8JA6PpbBIiMjIyMjIyNjCeQYFhkZGRkZGZlyj2ywyMjIyMjIyJR7ZINFRkZGRkZGptzzWKuEbAW1aEfZ7KCZRzYFYr5QJoXdpizrB5BJempZB1DJdTQ91l5H+V00D3IdTY+11/Fpfxcfy2Cxw4FmQgfTqHoEB8SdZVLO3ZRl/QB2iKvLMvIakOtoDqy9jvK7aB7kOpoea6/j0/4uylNCMjIyMjIyMuUe2WCRkXnKSN0UzJC4BAo6Nba0FBkZGZn/TJkbLEKjUNwjK6IK8CvrosuEnB7N2Jx8hG2Xo8l9saml5cjI3EeLKufp6XiLwTPWI6jVlpZjEtLeDOPaiHDifmjMtsvRuEdW5NqIcPTtGlpamoyMjIkodWp+pWslBGcnxNQb6G/d+s/PJXV1ZoPvUkKHRRD44WVEbUFpJZQ7ro4O58VBe9CKxiTF7T7fx6p6bfCdeNiq6vm0kPBNGLED5hC4cSjB71jH7tTKGoFMrPwTYMdrTqmssK2JmJ9vaVmlQlGhAtpGQaj+L4WtQZNxUdgBoBVhie9O+HAnq7I8+PxoF4JGX0Wfmoao01lYtUxJKCpUILVXbbqP3sXHbrHoxaIt9BpMGYE63Zjg1OVMDkJk+d+I9cbgMILejGW5306UgqJYfdqd7En6zip4b06lwNMR2+jz6NPTLahWOpTaYIn9LIiYl2ZTb9G7+H7633fZ9Diig7fgVL9ZvPjTQMTo06WVUK64Ojqc0F4xjHc9VnhuvOsxxr91jF6/9Ud/Os6C6kyDslYwMe87caLjbOwFW5oc6Yd7tzOWlmUWlK6VWNj7ewyIHO0yndbn3qfqZOnvJqtzc8Txdsfe9uSLqLMvWljR42NoWZ/kNhqqtb/I7yELb5+1K/Helx2v8XLrxXAEaq4cTsDaPIR90WUn1swoawYxZvNavh3QXxId+R2ExrVJGl+02MXRPp/IBrMBo8F5N0fGzCr8fXp6MDsHNkc8cqpMdJaGG4PDWPXZJLxU9hgAg1h8l52dtVdDbVgy2JtqNjc4nB3ALZ3x+3vkk0aotx6ygGpp8MSbH24cNIn+59/HZel/24Qsu7LySYssNyg9PcgM92fUtytoZvcPlRS2AMzPCMFG0POGc7yFFZqOK2PCWTfyO5akhxE+9T1yqhmI6TuHrk0GIR46YWl5JkWwseXsmBq0sPsTgL15blT9J8vCqp4cwcaWa+PyCo8zV1dBbUi0nKBSktxGw/Fhs4qd25jtQp5oU+xce/sk3JT2hccxfedQPy0Cr31lItNkCI1CUeQUoI8poT2Zm02G3gFVRm6ZbD5nKiJWrqajffZjPzfKJY6X1h7n2V3vEjToiBmUPTl57gJeKvtH3jfYyThY6Ghf1H7unHmK7wr6Y7OjfNbN0jxxDIuvypYln01F16HRI+9VurgQ/vbhwuOz/ZyftHiLcWNwGKxUsW3WLLo5pBcaKwA/LuzMokUvWFCd6VDY2ZHwTRh/jZrEs5vGcLRTFapMjcRvs3GKS5Hx+I1OeSe3c31ODZxdeDzxs9dg/78WVGQa4hbV5liTlZaWYRLanegNwN95toT8NpwfmjViWQ3vYj/tF42zsMon58brYcxb+z0Jfe9PS5H/XBNWBK1h3NZXJOfBHbFzQLHjOG0BdRZF8MNNn0c+66Oy553GuxHD6plL3hNxYOT0Uj/bwT6HdUtno9xVFVVlTxOqshyCSoXQILTwR6HRlPpvldrD4pBU5CkJtFGh/vgKwilPdFdTHvhMQT1/JlX5vrRFlhtSIsI59MGdEZ4SG+Eer9FtT+d95yXIxVENOT1gFvXmjyXoy8jCUdyFTmqi8pXo4xMsqs/UKENr0H7CP4XH/RI64bLxFIaHPCMFLo8LZ3/7SVCGCaDMhf8P5xC/19HVfxCKnAICT5W8tb3nIS28XXR8SZeLXaq0Nnu1fTmFbkeHlDjtntRNQCPYErA6r4Qnyzc1FubSbc6rhceCVo9vTCQbf23Bdx90osbkHBotO0UP5yPUtb2/HX3W4TQbKz9D6bs+81Hv53cZ1XUzALNPtcVjmdHbktJYRUTvTQAMdIpHLdiU+LxGsGVD8CZeqP4GwkP607IkfWAYLj/+t1kUpZsr13oEU/nVRBxt8nFQFrDAe3nh9Rq/DKf62P/2t+6l1B6WatMPEroiovB4XfAGLgyo/tBnbC+lsyKzWuFx4IqbpS3eYqREhLP0vWloRT1X9LmsyvJgf54eragvPGeTaWwU75yTKqoqlVn89gxaRPfF55siz5iyVjCf9lhlQWXm45lfD/GRm9FFe1abz5VZgRgyMy2s6sm48l44/4ycgofSgU+vh1pazhOju5piDKA9dAL9qftjqAS1mrjvmzBq5opi57vMH4fbgtI1lJbi6+C1ZKU4lnjNsXIWCso06anJSGnuhOHf2MKfO9Nd+tNxBL12FMO/sRwZWJvlN8Lvf1afy1tfjkKz7kBZy/5PBIyPYmMtVzbWcsWn9wnsNh3EbtNBfD+NLDzfbM571F48gtqLRxCj1Zb4d871LjkuyxK4rj350OsqX2+ujQgnfkZzqmzO58Bnc9gQtIVAh+vs+ac2H6QUzcBE9p1cah2l9rCIOh3B357jxxd8GehkTEzXZ8BfRP0SgC75conPFHi50K9CcmmLtDg3Bodx6INZaEVjIzHuYjfSW9zg6ug+RI0xugE7LxyH9w+RXB19/4smJQQbW7ruPEFkThDub2aiu2uVk/fSi/R1vE7g5rcJxroCxN51OVvoTXlh82iCVu23qJ7HReniQk54IDdCbMhplMOG8HmE2kYD9tSaNwzfjenwe/kNWHxScl5qht3wy8SFzC92fkWmJ34rk5HSGiFloD/uigcHem9vtIArelDmapGW3wg8Zt9fL2VFZ0SvKtRdHgtAdbs9hXEed7igy6XXt+PwWCztAHivr4v0fzD9eQSnClybpWZlncX43I5/6drqMDGWEngPjxq0Pf97NEOdNxSuiFIKClr++xLOb2upnrSfbWPC+ea9IygFBeF7h1Od0gW+P1HQrf76daae6MDAFosBGOt6gi6BYSjuMlgUdnYkjTXmQmjRRTpR7PdydXQ4v42aBNiRaigwGitjvIAbAExMbcTaNa3w+eYwIlBtaQzjXmnFd1X2WlJ2qbn2ZiPecI6iw/B3sL9StKT32vBw1lWdzgVdAbW+TpFUB/Ao4hY1RilEYxD19EvoRI3xJyU3FaQP8mLA5E284XwVgDitgeA9Awn4TofPycMIwf4WVmg+rrwXTuR7U+9ztYf8OpyAdfkozh97wJPlE/3Z85woqIKg0aHQaDDk5BS77qHUMCG1Nhw/U+J1KaGs6EzKck/2N/z5gffEaQsY+OVYPH6QtrFyL/q0G5B+k9y9zbCpW3T+74VNcEcaHkGDqOCaPoevr7Xj4PRGVDp6A4fYc+gMehK/CmP3gO8ADSszXajxcUap+40nXiVkc6gCtCg6vhxuj9ceyH++CVfCVegcRE6/POO+5+ZlBKG4niGZDuHDoSvwUhobwg6/jCXggyjuGCtVd2VwbJ0/3omRhSMdfXo6GdqKlhFrArz6nmfBTT80vx8vrJPK35d5Y2ZhIyjpsGkMQYnl0yVbGs4ub8Ci8CXoRQPfpoWS+4odhuxUS8t6fA6eYHWjAH6t3wkAVXoO/jHHSxyB6xykOZ1wL4aW9Tn/oh3RfYsbK6n6XF46NZDgxRkYTsZaUGHpGf9XH+K6zqPFb31x/VCJ4d/i9ehX8RDn9j7HuWm1cJSYN7AY7q7sb7jiobe8MmMMla3MWLmD0t2VYyNnAUbvyujL4VT5/ZJkBoS/THiOrTEtMESfxpmimLL0gWHsfW0yrgoNa7Jd+LFnJwwJpX8Xn9hgqTopkqYtXuNwk58AiB4xC0YAGJdl2QhKtOL9oTLDK57jh37PU3VyydNH5QlDmwYE2B7CRlDSpVojAu6xeg3Rp0s0vBSCQbKBt+sCtxqDo/KNdVW6ueK/6iqN1BDy15uEfHZWUssoH4a2Y2MWhC+htV0BILBiRXu8Lkq3YTTkFCXXuvczEjIy+TsPWtuBy3OXYVrZ6zMVglqNWC+Yvgv/4DWnZMCGfFFLpsHYzLf9cSx+n0RJZlBUErU+S6KDfy+i6v+K8g8FTY/1JifflhZeCSgFBdVV9hzcXhvfVdL9vgIIWTl8lNKYiZ6HH3jPjBHzmXBsEMrdR8tQmflR+Xoj/FhkmmhFPXt+a0TVJOl8pk4r9t/3niV+FcafAybhqtCwJceRxX26IJ58sunoJzZYANyn22P4ueRmQSuC4QFNRn7j8p/bQgyvxxsL1lPbRnysAFqliwtOqhxJBt1m92oGHCVgfS4AmX2bM+7Ln+iqucXyzCrUeO8S+tQ0y4o0IRcG6WhrpwUE5mb44/vbFasxxu5Fl3yZ1Tea0LrqIVL+qYoPiZaWVGrOTmhIzCtFy8+HXGzLwQ11CuMD/CTiTn8YuqspaF5Q0fjtETh0vcoLXsYGv7b9JfSigfAPh+O34rDkYljuRXflKidfr0Wz5s0Lz+383xQcFUVbR7Sw03Kur4rg3WWvzxwIKhXJo5rSru8hplQxesduGvJotmYMgd9Jx1gpiYzXwjg4cCqOgoaVWe7M/rw3Tsee3ANoEoPlYSy7VQ09Cr7+5wWUt5Sc6jvr0Q+VIzwmJ9HN4fGXliUPqsnqKqVfj29JNFfySTPk0n3hTqrbXqOZOpIbBgNKwZFvV/TC57q0X6a7ufBZOP+2no4Boyfs9+4N0Z89b2FVZYPmqjS7OWVFZ7R1Avi8W9FKtQGJz5I5wAmvBOv5bt5B1OnwmBMJcyASY76ndSP78+L42bj9nVwsIF7KGKJP435XLOYry54hflEwMW1+KDx3ostMurwQgXqL9IP9k0c15ejo4v3he5eeI3CUhKf2gJv9m7Pyy0k4Cho25TixvG9nkxgrYCaD5XgBbLzZkC0LWuEx19iABHMIfduG0NccJZYNDX8YhS+PbhCFJnVYONIYt7M6qypCjrTyJAj7onl28lga9/uXxdfDUK525dNPlvDJpTD8Jh+XtIv9bpQ1Anm791ZsBCX5opZ6u94h8Ky0AjNLi1404HBNmn6kxOGhxTLdDkx8hlvdQJ+WaDlRZUymn7W8hQ/GkJND1V9tiQsvINjGaKipBRtEQdqxV4lfhfHty8tprP6HOzErZ7X5jHluIEJuPvDf9+Yrb9zs35y/vpmJWtDwc6YHK3o988TTQHdjEoPF9mQS9SNfp7l3InvPBRIwV0TYF43HIzr3yY1W833lVg9NNleeUN949D1Ckzr0XvYnDWwVROUrWf5WVxSJ0usEK8+I5NIMcBPiiVtcgZZ26Uz5NACbbOtIGa2oX4sXftnLEOdEAOpsHknwUOvY4PBuhAahFLgbG0VVtg5Vyk0iv6/MmrFx5LgrqeDthVhBw6WJSuzWVfzPyaEsRtM6fDNoaeFhv4RO5Pa3R592/55IylrB6J3syP08k3aVizLBXsp14dwXNSW7Z4tCo6Ff+31MTQ/CkGbdm+Y5nr1JpqEoi3jIrjcJ3vGvJAdN+c83IWd4BiP8tvKC5iZ3B9genNMQl5hy/u49gozXwlj55STUgjGd3y8DnkM8adptW0xisOhT0/DpncZloDr/vXPupLnJ93ble3v7uwNnD4+bRZcZJW9BoHRxIWtlRXbWXgoY8z780r8zisPSM1buJufFpsR1nEuN1aMJ2iFtV+XdpNd2KjRWAHw3SXN6pCQUdnYY6gYRN0RNZMfpVFEZE4/dNOSSpBOoa2tMSPXy5/M4+JGWiooCgm0c8E95C5cfLan84RjaNODV77fwnKYoJ8SxJG+8a6tQJxkNlrPTmiMqjZ/lN51X0sPh/lFG4+8iqLxVulNHCk93PnXfS431wwjKtJ6VevciNAil/uITNLqrizBk2WDIk5bH+g6Z3ir21y/aGuOSLpf2v79HzU8TcUmRtrGirBHIS2N34KPSsD8f3p04HNfDpu8vzB7Dcjc2qTnsydXQxr4oX8C575zxe9UWsZzOwyZMqcmpSX8SbGN0Q3pGOWEQFRzZVouKcQaqR8SiEAw4qXL4rso2rugL6PDLWKqvzkQ8/PDsgFLgy8kL2ZmrIeSzM1YViJpXqWjl2sDEZ3A4ckEySwgfhsrPh2uz7TjU8E4qbKOxsjtXgR5H5l1ux5HTJediCfqxfL6Dd7D5N5EJR5/n1btiGmLbLeJUCx1x0zwA6O5wCMUjEnjnVJW2cXqhpzFbuP8a6X1js3s246XP/yw83ji+A/bJxv3IFGm3ENU2iI72xI9VM7HJ2hINTqny9qgNxY7bbx+N9+8gZkl7PzaVVzWGbdlMZ/scNuU4MXXsK7iuN48BVqYGi+FkLF8PH4hy7hJa2hmt5OMtFtPD4Rn0GeWzsXRYfYCRughmTptFsI3ADz67jCt/3tpeeI9x6bae15M6E7MqhIAZkZKP2gfQdWhEY/V+ms15D6906Y5IS+KziGWFv8ctDcH1qrRHOHc4N9iL2IZzC487x75A9iwvHLZGI+bnA6kEI8H8MhhzGwW9mU/zX/sVy9kRaqsi1PZOx/ZgY6XmLyNQpwsETo+W5JTCHXLqGVfvqf6S3vRsvpOCCJeiXacjFhT9PiK5JQ0cL/CG84USn62zbxAekdJMEyGo1bSyPwsUuYvinp8Pz0PbE725lVuUhr/isgo4bj8JoiiJZICZjavR2T6HXLGA6e/2w36r+abWy9RgAbDddpiv3h7ElwsW0lhtHLNnta2B/fryGz9gv/4gH53qx5nhHsT2nnPf9b/zbPkgpifug9OpbCUraBQODrSeGsW2HA98552yKu+KGFYPb9VBQEnNlcOpvsh6proCVqUxqWd19qUFkrChOlWmRqIh2SoMaDAGYqpWuTKg0rMsBNsZjgAAIABJREFU9/vzgfdtyXFm+sh+aA4nFp6rnnYQDHpJGyuGVg2Ia/8Dw5JbArmWlmNSZlf7p8TzMVotWlGB3yQQD0lzYJH2SkPclX+VeG13nd+Kn2hq/N/Yq804PLEZDmvK97SfOk3LFX0Or8QMwN7McWFlbrAA2Ow4wogpI3hxyG5+3NOKkN2x5b5D1McnEDgqgcaJEQx+aytDnOOYnxHCjws743JWS6Uth8p9HR6HCyPrsdFtFi+8/AZCRun2fSivXGnhQIBKByhRFAggWkt3DvpTZ9hRuwKQQhWkEcz+uFRcHsWtzS407RdBlo/I6QHGXCw1d79JhX3GQMaK5wpQb7OudxIgz80WAyLbT4QSzIOTrJVXnM/nU/ufwewLn4ez4uGb+9WLGoher6D6yKvoU64Bpg3gLEsqLYmig/NY2g04SGfnE3Swf7DnZFWWB9VU6dgIemZMnsVHa5qWodLHR7H3GEM6Dcbx+g2zv28WMVgAPOZEEjnHliAOSKpRqTw9kt+nV+T322Zw5f+wzFmKDBuwiW/TQlEcPG01o/M7VJ0cyd633aimysBnW76l5ciUAn16Oh5zI/EAunxgDIR/nIB/qXKpk8ixAgM1xyVIqt28g3L3Ufx2Q4sJ73Ny8Oz7rtdcOZxK/xrjBX1+PoSo00myniVReXokMdPhRIf+jG1gnBpyefYKO2uvLryn1k8jqP5bJnnu9jicvsqg3mFUlUAfc2e3bXNjMYNFpnxT3y6JIYtG4KUt/y9LaZgTFAyAEutK8y1j/SxNbWXcME/C+H0cRZeP719xWZ2i6VlrGyjdQbXzCFV33j6YDF0o+ncIIAoRY6SLDqg6+f7l+k8zssEiUyJfBDTESwKWvYzM00Tw0IOcs7QIGRkL8fD1fzIyMjIyMjIy5QBBfIyAQ0EQrgNJ5pNTDF9RFN3LqCygzOsHch3NglxHk2Pt9YP/Z+++w6I41wYO/7ZRFURQeu8aCyIIGDVq7Mau0ZijMcbYazTty4lH060x2Hs0xpgYS2wxsSeCvSsKCKg0BSwgfcv3xyqIHdllmXXuXFxhZ2d3nld2Z555q1hGvRDLqHPGXj54ShnLlbCIRCKRSCQSGYLYJCQSiUQikajKExMWkUgkEolEVZ6YsIhEIpFIJKryxIRFJBKJRCJRlVeueVhMJKYaMyz1FUsZBeRSpCmUVMrB7qnM8gHkcCuzsnt8i2XUPWMvo/hd1A+xjLpn7GV82b+L5UpYzLCkiaS1bqJ6hsOa3c/eSccqs3wAuzTrK3OoGFCxMsZFNiGh5yL+yLVgYWgTVLduPfM1QivjizD2MorfRf0Qy6h7xl7Gl/27KDYJiZ5JIpcTPzuMQ91m0SWuPeO3DqC4rruhwxKJXko5fcOwj7Yi4btwct4MM3Q4IlGlERMW0TNdHxZKbJ/5hP0xgcIW6fiMP4T0X+NawVkkEor0DkUsc9vLxbfn4TPuArIa1oYOSS+y3wpja8pxxsfHkP1WGHJXF0OHJHpeUhnxs8PYmXqKxG/Ddfe2Onunl1VYfRLXNkCyx5mdqaco+tudxK/DSfw6HJl9bUNHV2HFbRuz9IPvmZzRAP9Jwl3eXaRV/HowO1NPsTXleMnP5em6O6GI9OtO/zB2v/ZDyeNlbnvBrqYBI9KvYo2K5mY57J72AzlLFYYOR/QcpK8EkPq7Pxf7zKNYo8IzRHcLOIoJSwXFvmfC+ebL2eL/B8UaFX/V2cD5gXM5P3AuXx/aQsrHEYYOsULShhTyiomEX/98FXVenqHDEVWQyafpFGtUZX7O9JtD7Ipg8ruGGjo8vZH5+xC7JAT/Ywp2pp7C/5jwLn7SBoGs+GoWLnLzkm2NjvwH9ZUUA0alezL72mT84c+UqcvKbE875WCgiETlkTuzkBOhqwG4qy4keZebzt67wqs1y2pYI7G0JLm3B9n1ip66b8Dsu6jPXazoIauUBa+tLvl9SkZDojM9Sx6v8PuZRl3Pcf1bQ0RWcfndQjkePpewYwPw/Dja0OHonaZpQ1KbWXB69FwAZBIp3eLacX2RJ1Y/H3rGq6s2WV1/eq7fz2sW/wLmjzwf23YxDeJH47q58mPTl8yh4dxsrOSNoFP84LS+zHNHbrhjTbyBInsxl8Zb4KMwLXncL7ENzlMlaIqfft4VmrthHnzg/zstzQsofmDlmNP95/BazFhsVlb9c5HU0hKJiyMA6a1qoZFIyPFWP7KfpFiC74oM1JeT0CiVlR2mzkmC6zLU48+Sx+HLJ+L+TZTO3v+FE5YbIyPIDstncIMoJtleeK7XrIpwZkP7EJRXdFdFVFWsyHblZFcP5FeuAiD3cOPCLls+ctzJW+M/wGX9VZTXkg0cZflk1pVjKpFzO80K4TduPVnm++HIumayud5c7GTm3D+tqDUqfvPZzsEvFIxxGIbjLN198Sqb2kROv+pXARNDh6JXMn8fYibaMLflajpZlO1nNSY1hINLGmO3KFpwyUrsohDWNl9YZtudV7OALMMEpEcFNjJcFY8v1+zP5zFBPZIaq6p20pK5zplDQb88387/gdcvdMfinWKUKan6DUyP5B5uvPXzn/SpdqNkm/0x3SZhL5ywnPy/+SVVyhtztUOmp57rTG6mBdViy1a33g0sIrb9IgZYpTD93R64Tza+hCWlyKYkEZM72HNppDOtzfMAE45NjKRZmzex7mjYGMvLq00i+woU+C/Lx1iXyLwxKoIFEyIJNoWGh4ZistsatQnMHzOXUFNtqZuaFZPn9OjdkZDcmPLoiSN85jhyggo412rhY14hTDdnQ2LDJQCEnepF8WbtucluUTRQjB1V+0L3JCF1Ewi+V7lSrFFR988R+HHMsEHpWOyKYGLbLgaOA6CQyB7Zp6mZlBw3CTUqObbysjQpYnJGA24XWzx1P1NpMdMcjrGrzkaathyB9U/CTVhUttXLJCujUl6l2tl0dJmyvHDCsuyOAzdVlvwS2fbeyQCcOf/IfjJ/H65GlFZjVrtqnJe+Tlan2Nt1JNWPXKXOtutsst9e8lyxRkXhttogoLs6uYM9n7ht4z/bh+N79LChw9Eb376XCDaFV1aOwuur06jz8pCamfG2/zD2dNT2FyjWqFDcFWZ3r+y3wjgwfR7ai4D2AuC/fiS+Yw/hQBRFn0egaC3TXhwqdToq3YtdEsJcv9WEneqFWaQN1juOIqTv3JNowhsw1PGnkscN/30Pv/eMK1kBQCOhWKMqsyniZD9yo+2oFpHBvgZrAGjYIYY7K51RJlfdvjsWvW5yrNACTWHhU/eTKEx4/5/mLHY9wPVmKqx/euruVVpCb6uS30elvEpyn1ook67q9BgvnLD8GqjtAPW0OxbpKwE0XH2BzbVPAtD6XC9qb7qE6omvEK4gEym75y94ZPvvd+34ckU/nOcJqzkh5jMPwkzBPOXRuxxj4mqunfxOopSgLijk9n/CkWg0vN7obEnnxsaH3sXtf8L6+z3o4YuA79jS/jgSzQPPC/heInNoOImdFjAmNYSa40F16aihQ9KZuPflNDcr7aciP1vNgNHonszXi7Q2DmxsORuQEFMMR/M9WRDbHOchmdjlXCW2VkNooN1/qftOetoOhCqcsKiys59rP2kNaxa77gTA8zdh1+Ka+GnLvDbHnuQ3a6NMKp3/TWZfm7jZTqivm+G/5CaqC7EvdIwKd7p9HJmVFZnd6zJ78rySanUAs/9ZocpK1Mchq5zpWXU42DUAVWo6zoXCu9h93HoLAO7r040ywbzv1C0XcDjKxoEzuNbfmqTiNAZZlTZZtjzbG493rwr236D5pLKdhXfll1amS83MKLQV9knyvqZDtDUOW042pOZrcqp72WC6wziSll4NThg6BL1RNwui3YL9jLaJo1ijreLrtXsEfu8dw4EYVIAkqC5ne/3w9DcSGZTcw40Z9bUd279a3xuPxNKKjOK2jQn+7giba+8AYFtna5Y0a4oy/Xq5j6OXeu6YaQEc/HpumWQFwHl2ApI9zkj2OBM/OwxZXX99HL7SSBq/goOsbCZdrFERPHM0/7ZxR5l45ZlVglWR3NEBN0UWw5KboU58cpWeLNC3EqPSD9Phcj5OD8FHYUpL84IyyQrA7b0Oz323VBWtPx9U5vG3/x1Q8nvC/wUZxYXgznYffnA6ypjUEEzT5EydtIKC0beQ+fsYOrQKkzR+hdZWF5AiKfmxfy2FhG/DSfguHKmZmaFDrJCU18wZVuPpI0elicm8smdYmW2pU/QZVeVJ6S/8cyhA7Ahtn800VT4Oh0tv7+QO9ig+SmfKvVYWgE4Wd8D8xT63eqlhcfXMeOz2xa77Sh/4Q1O/vtTsKhfccC6JwoSc7o2Y9NVP1DXR/hMmKgt4+9w7WH9jieO/UYK9IwfIDXKljXk+Y/+qj4eybJOf3MuDCx/VQmFdyM7w+dxUm9Bn0xh8xgtz2K8qLoELTc3oWq0tSKRcfdeXk2MiAWgQPRC3748LuaWEiy2XlgwNbXuuLzb/XivpBPdhr40Gi0sXCjuEUP8L7ZDlgKXD8VqTgfulaEY5/ofETkvoSC9Dh1hhWfWr09o8jwfrwf6qswHqgBQJV/vmMTm1I1ndzV/ojtWQZDWsKfTPRyHR9qGKvOXL7i718Ut4tH+ORKops5/LhHydduY0lKJ7FZ5b8qwwO3sNFXDn7TCUZmU7lNkurdqdxSVuuQCsy26A2ZYjJduvLrDlRIB26o+xqU3ZeTGQi62WvvBx9JKwmE+uzuu1hpfZlmsvo/qbqeSucSLHTcI/Q6ZzsOEvjIp6lStNpYKZS0D6SgCXhltzqdu8Mtu7L5yEiw7Hm1cFNpfKXqqz3gvns49W09b8Jg3/fY82B0dxqcVyRrb5i51YPeFdqj51QQEUFEBoPca8swmAQVda4zn2JkoB1pDdp9jnWDLSYkW2K5btE8qc5GUSdcnz3eM64/qVsD6/6eFy9jkdJWDpcNw/jxb0TcKLcpGbs8xtL3XHjsJ3nrxKd0R9WMw0P86/Nq8koV66tj2uCY9+BtWeLpx9bRHFGoi85cuOYc2RJgh4aZDQeqgsFBTYKZj7n0UAdLC4Rb0jGwBwkx/XJqPKPL5Ma0/UjvrYGjLeZ5C7OLMidCUAm5MbYElCyXONHbU11gcKTEga4Eq1tubQqgLHqkigT3ToDA9W+BR2CqHee+e4PsgBk5hobIC+/4wmdNYx5jr/S+fgQXDojF5C0SVlq2DGLlpLB4ucR56rEW98p8scVyn3VynJGhzO7v/NYn2OJx2GvYXn1iNIgupCC1i4tR2eAh0uep8s0Je4cTIGWV3jrrqQtM+8kaccN3RYLyy/WyiDHH8rmXpArSl7x5Y1JJyGpt+X9BvIWO5BDdIMEeoLc/88mo5reuF+qexnb27L1U94hfBUSykmpriYQEXpVBFfZ9bj77QA9tcrnQzv/IC59G/Wltx+LoKZ76npK3FlHhfaqZGYmj61GX3bmJbI/xXO91JWw5qbnQK54yvFvfkVXrW7zDCbhdhIH528UYWEnXfrMOtQG5x2yrG6dAf16RjcqOI3EiaKku4f+RvtSxIWZatgPnecw4ECa74Y9S6mMUdpvKJi5xi9j9Usah+C/+RzJE4JQBVT+gGV7T3Bxs2vAvDhmjX6DqPC1C2CmLT4p5JkpVBTTOtzwq9yfhqrJG1FtNzRgT//N4PwQ0P4vVldzLYeQWphQb1l2gkDa50QcqOJVkrbWlxosQw1apqs+QCzC8I46T9JZj05XSwf30Qgq2FNfvts/BTaZCV85rgqPxHXk6gulR22fGVqOJ0sCvDcNuSR54TIZOcxVt8su9bTyhPhWH2goM3AIUzPqlOyfY3nX1x5S3fToOvb0d2BZR6HhMYiq2VX8lhW159r/40gbmB1QNukeT3UFKGQuzjz+Yk9HJw2n3ND5rLNfwuf2F4oSVa25ZWO9nr1szGMdm/K1ro2+A0+RrVfD6E+HWOo0F+Y/wBtfyS5izO1vkjERW7Oe38NxnTHUZK+DOc7552szbFHk5f/Qu+v/4RlXBa1TO5iefLRzpv2R4q5qy4sM2Svquoyf8+9ieCg88WudBw1howoRwNHpR/mKXeJLS7gevsikEhQ3bxFxMHhWP9eDVVmFhKFCU57pHxtf4yAn0Zi9YeAq2fRjlQ4OGkWatQE7h6K58fRgusP8CzfHehU8vuluV4cbbKi5LFlmjBGCmUOffoijVemhjPzrRVsyzMjcMatSopKvwo6hzLE9t8y22LbLqbP73tBKkEmKfu3e73PEYTiwT5Un15vTPZ7tmWatG41sOHY0O9LOoarV9bG+dsqXtvwAE1BAReLtNcIn83DqLN8JF0iutIltDNdQjuzsEsn9hVoa85q/3XlaW9VpWly81iV7QzAD25budsnjLxXnFjt8TcA0gIpcZFNODhwBttz3fm5XztU12887S2fSD9NQg/wq5HB5Fqn6L3OiZxZoZhv1n6hMt8PR9nxNtWkVT9jvj46gsHW3/Pp9Sacb2OD5G4G5oXJmL8r7FFOT6I+HcM/+T7Evb6UoPGjcJwVhWffM8hsa3LrrTA+mbKKThZ3CVg9Cq+PoxHG5e7x5F4eSL5Ix1SiYPKNIHwHGOcQ0oAFOaiBuDlhXGo574kdcauy45MXMGZI6fT698UuCSGx0xLgFGGnet2bh0X4tSsAlhcz+P5Ga+Y4HSyzvX/1NN5ZubTMHDtpqnz2/ByKY1VvQrhnXlwLBgT/DMB0h5N4f9qQ2jvCuNGxkIstl6KQnKJYo+1j5ffnUPx+EVbHflVmFgu/6MEcGykBK06jzs0t+z1LgehcX14ze76lbaoq1fUbfHe6LQOarcBaasbMb+dRoCltwrzY535/TzO++aUPbidf/POp94Rl3wV/cN3Hbz7bSYvM5+h0JwDaWRzCVKItVOMZo3Gool8ySeNX+HT0GhQSGYVqOaqsmyXP7Q5aAZgSXSjDIk24nTMfZ+l3Xek4dTr7x88gaoR2+frq0gKamu4iRZVHneWT8PnqpKCTFYDQjbF8aneWX+/W5siExsgwjoRFI6FkVAVAyus25E4K4lLLeSXbGh99G4duMYJIVgC25Znxg9NRtk06yxddO/Nfv610sigATjEmNYT4/u5YX4o3qs63qvhErrzpwfh1Ecx2evo5st+FATjOrJrn0ce5lW5VJuE623IRtNT+fj+hLtaoOFJohtsmYc40bb1Gm2QJ/Tz5LD5T8ln2uxuDra/eW0Ki+JF96qwZhfdXxyo06lLvCUvAqAsELh1MTItlOMrM6WJ5v6pWm6wE7h+MT+SRKjt0tKimGWFmKYA52/c2xptoJMF1iR1UDQupdmKqQVGD8Pn35NPfSGBsVkbzutcknCNS+CtQO3LmVJES//3v4xWpwSNa2DUrErmc9OGhfGw3h+uqQr7+8U1c9grnZP8sD85gW6xRcWj89/d+p2Sb02RhnUg/nz6InUOO8YPTUTo1XE/YqV6MSrHBfQP3JokzjlqVhykTkkjs7UbA2JH81WMGbvJHO2xOSm9CjaHFgkk+4dFOtw9LVuazM9efdeM7YLZTOE1d5ZGYZ4fMTpjJ2INUMXH80TOC7yZ0ZHKzzfSvru1cG3GyH3lRdnisSMA742iFpzDRe8KizsvD992LdLPpROw4T5Q1SwO2OS7He9EhNJqqmq5A8msKHGXaE8SG3rOZ1KgXQ1y33Eu8pAy60hr/ccKdCfVp3D/XVrt3pFHJNm+E3V/lPmWz+nw1djnJykLeWPohrl8bT7ICYHIH0lRFj724rch2Zfa6brhfFM5oC9AuYHhpEbSjIQDWxJeMYjN2yqSr+Iy/yugvuxA/yZ/Q5jFkF5vhYnGbXbuD8I28gjJFWIvK3nqzGu/90o6l7jsfea7tub7c3ueA87dRmBjZIo8PivqzPqohBwwdhk6oYuLwGwLrXmmFdP0u+lW/jsWiGtTcEqWzRFrvCQto57lQp6Xj9VF6ZRxOpzy25nHwTQVNzbRDC7cGbC557roqn7gFgdTIFOYIi5eVJKQe3y1fSH0TGX5bJlDnpxRB3Zk+D/sfouhq8SEnR0c+8tzvgbVxI6rK1mqKnkyVdRPPj6O5PzXnZcCTaEF+fpXXkrk9xI//rg7jWwdtbXX4zHFYpqmx+fcalsnGdRPxslCfu8jPqU3Ya5GN2Tbd3hRVSsIiZJKo03w9YACfrV5JuGlpPUr3uM7cjnSjxu9isiIkslq1eHXZEeqbyBh0pTWBs2+iTBRuD/2ncf42ikDbUZzpN4dd+TVKpuWvjrA6L4qMl+pCLOeCoTPBACV9GYWYgL0Ir0UJtD46FLPrxtWlQNMqhVQ9vK+YsDwHycFTfOXV8KGtaVgKbKKtl50s0Jc+G/cTYZ6I/99j8B92HnWBcfZ7uM97UjTdJ4UCYqIiElU1yrR0TLeli7Wdz0lMWEQvBYnChCtfmhAZ9xqLf+qB77pDgupwKhKJRC87MWERvRQ0xUW49Dxv6DBEIpFI9IIk5RmhI5FIMoDKavB312g0tSrpWECllw/EMuqFWEadM/bygVhGvRDLqHPGXj54ShnLlbCIRCKRSCQSGYLwZ6wRiUQikUhk9MSERSQSiUQiUZUnJiwikUgkEomqPDFhEYlEIpFIVOWVa1izicRUY4alvmIpo4BcijSFkko52D2VWT6AHG5lVnaPb7GMumfsZRS/i/ohllH3jL2ML/t3sVwJixmWNJG01k1Uz3BYs7tSjvOgyiwfwC7N+kqfE14so+4ZexnF76J+iGXUPWMv48v+XRSbhEQikUgkElV5YsKiJ0lfhrPs6r9cmRpu6FBEIpFIJBK8SklY1Ltd2Zl6iuRPIirjcJVOE94AdbMg1M2CkNYPKNluLzOnX9f9BoxMJBIZDamMhJ8bEnJKhWKfI7FLG1PYMQS5g72hIxM9B1XLRrQ6m8vO1FPc/dMLWa1K7WpTafK7hXJlajizkqLZmnK85Gdn6ilkvl4Vem+9ryUkNTPDxiyXYo2KsQM2sWlJIKrMLH0ftlLInZ1InmfN38HzsZGaoUbN6SKY/MZ/SvbZne6POYkGjFL0PGR1/HBcnspi1wOo0SBFwvzbnvzZvh7Ka8mGDu+5pW0K5O41Kzw3Kp+4j1ncdUGVSXSPWoX3bBX7vvBlS92fGKnoxJi5u1BIVHzWbSDq0zGGjlBv4meHAXD5zYWPPNds5FAsNh6u7JDKRd0iiBnLF1DPREGmKp899dbRou1orNdkGDo0nUn7IIIvh60kwOQgnnIz1MhRP7DEbLEGxu3YQoFGwbjtA/BblYPmePnWd9N7wiKpZom7hTZBGWR1jc3VGoMRJCwFnUNx/CSeIx6bAZOS7Q1MQGVtxnvd/uJkkZrqA/N48qVDICQSZN4eJPZ3pHmnk/SzPcy0zj1RxcQZOjLdCK1H9tRcNrruRY303pdMyvs14vn7p0CULQwd4PPr7H6eKSEnUfdQI71Xlof/vzXXluO5Hhwf2gCOnDV0yDohq2GN2R8mbPD5m2KNCoCDBQre3T4EgIAFt1BdiDVkiC9M7mDP9U5eFNpKuBtQRGy9ldxSa7g6248pcQ6M+n0jXdceYGMd47tjj58ddi9JOfXEff6Zt4h2GxtWXlAvIKW5OfVMFJwvLuLjjoO49IklO76eybiTgwT1uZT5eKJJTkNdUFBme26vJqwfMx13uQkPXg8f1sI8D4CYXpE0vDUWt+PlO77+V2tWqSlUaw8T8OtI/NKf/METijv9w1j/zQxqyUwBCNw1FCQwvNF+9nWqQ9z/KdhQ4zzHCi1Qpl83cLQvTubnTVJve5p1Pcl8599Ltqep8pDk5BkwMt25PjqCheMjcZHnM+VGBH/88irua65ws5krB2bM42ycC36kGzrM53Y8SErQx6PJ9SmiX/CRx+7jbHqLKbVPMmWxhqMNZZUcoW5Jq1fnTqe6rPluBk5yUwo1sD/fAjNpMeFmhcT0mAtAhEd/7N4wcLDlJDE1JXZmQ7Z0/p66JuYPPCOjtsySX2bNpPvkSVwqcGJwjfNsbtDcqGpZ8ro3eWyNyn0DrjTn4KE6APhwqLLCqpCuf47B78IRJDfC8JabEzfAFq+PDR3V87vWwxG339SoE8sO5LE6k0mntRM59vYsUpUafBSmxBSpWX+nMb9vakZQmxh+9NhVsv9P2a64b75DeVcy1HvCovJ3ZbrDCgBMXHIfycyE6ND0hRRrzPkysz7HOnvie+0Eed2bML+gFbXm32Zt4EJMJQqGn+iPG8K7g1W/2pCbH+Wxq+FKrKRm/J5rg+/fQ5DI1cS2XMZbMf/BPNk4mrkWjo/kUL43a2Z2oObyaNxdr3KlvztNup+h+Zk+1PlfmuBqyJy/jQLg+BO6qB1u1Zb3Vi9kc2I9nLhQmaHplMzPm/QZcqKDIwFTzhVpGDxrHPaRUcjsbIn51pOLHRYAcPuOJXaGDbfc0tZ5ERMylwW3A+m9pCOea5JR16jG5b7W7Og3HWuphMKaEuzk2TT8azR+p48ZOmSd+mfeIqBsYuJ0QPNA80+2YBIVY+E0Leqx50NV7GX8Fqt4NfMDbOKVZDSU47y/gNSRRZwaMufeXtrz0fq7Dvz2Vms0J8vXHASVUcNihIo1KtSoWXuhMd63EwCw2HgYJ3kT9nz/MwAdL/bArbewkpW8Hk0Y9+1ampkfxFZqTsD+ETitNcFy/0V8s4+jbhEELSElxh4fI+iXkzAtnBDTE7wd3RLv5dEA5DRyorCmhsWuB/C/PBir5MsGjlL3dv+0jGKNFNOt1oYO5YXJbGxo9Gss/611gtNF8Fb0ELzmarCP0iZrxXXdmNNC+13seqkb/h+koDJkwOWkbhbEkZDFRJzoj90bsbhSeqHwPAWjZ/dA6eVI/ZkXuKOypM6XmYJLrB+W170Jqc0l+IzXJiGlfVOMNzGReuYaOgSdUSYk4TQjifxuoWikGuasmIefwoT7iYpCIsNn32AUFy1wOxlRTxLJAAAgAElEQVT1QsfQe8KSOE7fR6h8fluGc/GNeZxtvpTxe5oRNzEIjUTC5G+WA7A3vxpMtQOuGTbQcsqzkxKZ1IqpeeaYbK6B149HQK165EQvK6jUiRb1pnebg6gfqpQ033yEwAkOqNGgiDV/wiuFK2FaOMWa4/ivH4nvsmhDh/NCZDY2pKxwYGOtvzhWKOPTscPw3lLa/CU1M4PPM2hncQeA/FnOmF1/fPNYVSUrUKLSaBjifZA/7AIeHaiQX8DlEVI2uf1Fk+/H4ZTwYheAqsI+2opV7toalXbjtf1RqnpH2vKQ3WtYCK8Xx6Wh4Qxt9xcAs4J/ZV6tZqgyhN/5Vu7sxNX+Hvw+cvojnW6/yGjExl+b4TPtCBrli6fWeh/WrFCUXu4K8xT6Plyl8B97mo4xPQGY7fQP/1uxjC+XL6GFeR5786sxY0h/pPtPGjjK8rNbHI1p2yQcusVQc0U0qMumKqZTtH05fGYbR63D8aENkCJhbMM9XP65ITl9w8j904vffbaxLc8ar2WVPmmmXsVFNuFC/7lMvhFE4HRhJdMPKmrgyZGQVQBMmDwSsweSFU1EA/wPKvkjYKOhwtMJzdGz1N06imE1Uvji6A5SPo7QJmL3xC/24IfwtTSbOhanGcJOVvK6N2GV+wEAvNcNM3A0+uESeYJ60QNY7bGbI5/PIzG/FuGn3qS9eR7XBvoaOjydSFtQnWNj59zreKvV/kIv2r87nJM9vHH5OqpCyQpUcpOQ77ziyjyc3mgKCzHtl0/AFyO5+MY8GpuqkCIlulDBzMFvIdt/wtAh6kVYTeE3A5Vx5CzNz/ZiT711DGuRgLpF6UiaI3e9UCanGDpCnUmYFs6lHnO1Q+/fcBV02bLqaC/c+/MtsD1xs6QGMHViBPNGzMdLfpf1d93pVS2dKTeCsTwYJ6jmoPv8R52g7aqB+My5xLkx8+nRsQ2FA+2JGefIhiZz+KTbO9idFmYt2YMe7KtyvznI2KgLCnB/O5524e8BYHrlJsUdakFDCOp5jhuRpmgKCw0cZcU4VM95ZFtKljU+l7NQJiTp5Bh6rWGRO9jzTX3tnc6ZIhWyu8L+gzxIlZGB3REZ0nv/KSQyDtwNEGTNSnl8eqMR6qybhg5DZ6q1T6Dp52MIPvo2U24Es/iOB1Kk/PHLq4YOTWeuTIngQv+5bMuzpu2wkYJOVgAUnbTV50fyvEluZ0vahAgiThexbOQcPOR3eb/DYL443RGAtadCUN26ZchwX5hGqUQSdZrLIQXU+34E0903su3gZhJ6L2TUxDFGNSIIYJX7AeyjrYifHUZe9yaGDkfn1AUFyPaeQLb3BMqEJBxWnaVrXCdWuO0ju1uQocOrMMlgExouGUvDJWNpfa4XAGebLWPmrjXkdwvVyTH02yRkbkYHC23WNTG+t6DGmz+LslUwX326FDVqogtlHCxQ07n6aTQRDQwdml7I/LwZaXOC/Wk+Fa7Wq2psl0Xj2C2G40FS3rdOKjPZkTFQ+uahRo1KIyWrjpyUjyIoeCMUQusZOrQXYr7ABoBJtmc5+sEcjn4wh4/tTrPuZhOGN+sHKem09brI4UIFfguKDBytbjhNi2JgzICSxzcDhT0c/UEPNgOtcj/A5TcX8s+8RSWTxRkrdU4O6T95kK0u4LoRFFWZkITb/6Jw+18U5u0SqfvLaP4tMMNHYcrueQu4MjUcmY1NhY4hriX0AlSvNeKjxatK+qx8MWgQb/81jEATKQmjjKND6sOSettjJTXDdEFNQ4eiN1mDw5Ei4WShFOfvhN0v4EFO60wI2DGc47kebB85jdNj5rJr4QK2b/yRrSnHyRosrPWuLC9mEPj7KHbklZ78AnYMJ6Z1dZRXrnGjZx2mO0bxV3Y945gYTyoj8Zf67K+3nvozRzDgSnMujJhvNEud+Iw/RDunhnivG1Ymebn85kKjrGl5kO3SaL6/GcKg1vuQmJoaOhyd8v7gEF8NfYf2Md1Ro+bs4LlgJyYsle5pHWzffSXaKNf2MGuShRIVlvHCrF5/ptB6bJg8HTUaJnwy0tDR6JT55iP4vXdMW4PUczjNPhhB/cWjaX62F2rUtB110NAhlosqPhHfMYdZ4OtDF+cQujiH4PfeMVS37yCzrUn01LlIkbLmqBHctgI3RjQhtvkqfDcMx3FmFJmd5UTecmfa4OVGdUH3GX8In/GHaDZyKAOuNAdK+7cYs1XRTfnE9gIyZ0dDh6Jzil3HMRus0Y6cBWImVGw2JDFhKaerv9Uj+F4i/PmX7yLbW7aDrZfpDbAwvuGwr9RK49vMBsYzHf9Dsr0tcZSZM/lGENXXGWfHP9COPqn+yyHcpkSh/rE2UiM7Bdzo5o8aNb/erU3APOHPcSF3cebniTMACJirHdqsyrrJhgltaGSSSb1PTyN3cTZkiDpnsfFwyURx9xOXl8GtUAdDh6AfShVJRdpEpcbZio3zqbSzVeZdy8o6lN5cHxPBmYiVKCQyAnYOx2ZlaQ99qWUxUqTMTWypsx7RokoSWo/ZX81DjZq/I5saOppKk/FGgdGVucfYPQBMW/gm6lPCncX3vtSu7gSaWLAmxxbNldIFK012HqNt5IfMdz5EwmB3A0aoe3ndm9A0TPh/u/K67WNcNw/3Jb7jwSDrJACkHTIr9F56/Rcqci5tr7JdKvyERfp6FmrU3FHnY7+nNFPUhDfgYqul/JFrg/UI4+qwCSCzs2WGyw7OZjsZOhS9SOhVjRBTbd+VmsuFP0z0eaR8FMGlFstZfNvHaMos93Rnku0FctRFmN4s7yolVVO2j/Z8MvW3Po8sa+L2WzJ31PkUWwnznBM/Owz7aKsyzVp53Zvwz7xFJfOy3K9pEVU910dHsDXlODtTTyHz9SrZnt8tlCtTw7kyJYLBsYmcHhGJFClTMhpi/27FuhTodR6W+HeMa+b/Hh6nAbiilCAthpw3w7gVKGX/4OmAGXMSW2OeYGRzlQBpff2xlZpzbYkvNahYhlzVyF1d+LbbGtRoGPTjaNwwns62TyJ3daFL339Ro2H2rg74Gsm057FDtQn1lzdaYPOjcSRh99kfe3QmmUIPOywkT14ZtyorXYUZmHcA5t1/pnRxXGObl+VunzBGfPEbh3O82PNbiKA79itbBfPHpGmoMaVYA1d7OmB6056abyaz1Gf2vZlutYn0/XR6y4pmOGRUrMzGlVHo2e+JDZlke5ZAhYK9syIfeMaEgL3vETAxVfDreTyOdZdUAKyuCH/hyofdWWJCF8tb+O0Yit8U4Z5AyuPOEhOm1D5J8NG38R1rHBcEWV1/JnXdzHVVPrFvuoIRrHUFIFFKuKPOp1rMzTKT30mC63Jn4l0UEhnm14XXlPC0VZhBO9TZmJIVgNRWalQaCf4W6UwfM4c6jqPxW5lNNfu7hg6t3ExPJ/J61AjONlsGwPHRcx54tjSJTiguZnxCbwq/c8RhZ8XPr5WSsHx2IxjzfRcEP7tFzTmWBPQZSewb2hVgP01vgoPpHfa+2Rif8yeNMlm577IyH0XqHUHOGPokWYPDia43l3m3vQW5KvOLSNsUyPF6P9H8TB8cuxnPxGMRa08z0OoKyUrtKCJj4bc0g0u95LTdcIwFGzoAUGSnIrHrYtKUd/Fa/wG+04SXaHuvG1aStHivG0bTsAscPFSnJEkxxsUOA+Zn8/NYb9TBAcRF2hPbez75vYowlSg4VaTCc+UVwZyDVFk38fnMmoRdxfgoSodjF2iU7MqzRyFR8n+L3qFGggqLDYcxQTeTVeo1YbE5poAO8OuREPxyj+rzUJVCvuc4fnug87Dge1vUnKM6cMmQYeldX5ejnCp0QhWXYOhQdEZWw5qBE7ajkMj4YXc7fJONZ6G1h12ZEoFtk3TS4mpxKWQ+fjuGGk2CJn0lgDo/xjK65klATs8ZH2JvRM16qkvxjPt4NAu/+55xg5NKtjc/2x2LTy3wPS7Mz63P+EMlixz6cIjrGGeS8iD1mYsASKJOExMio84XI/nz7em4yU3ov3os7snCasZUxScydNw4GJ7B36/8SsD2EVS7pChZ28pJD99DvSYstRZG03lhMH4IP1l5WSV9Ec6wGgvw2fcO3g+0LwtdYSMf3q+xi0FXWxGw4KZR1Rw9zHl/IVP/s56BWYNpMWEkfusOGUWyAnAjwoZNDocBOQE7huMXaTzJyn3V1x1i0rqyc8qYk4hxdCt+SalVePxfNMP+T7sEiDvCSlbuM990BDZBF0Iq5Tov9mERPVVxTW1Dnv1G45qFUb7nOF2cQ4Ccez/GS77nOFO9GuGN8a5z9Wl6EwInCnORQ5FI9HzEhEX0VL4jD9NuZEOqIcyqZ5Fxs1scTZfFIWjHItwxdDgikUiPJBrN81csSiSSDOCK/sIpw12j0dSqpGMBlV4+EMuoF2IZdc7YywdiGfVCLKPOGXv54CllLFfCIhKJRCKRSGQIwhvALxKJRCKR6KUjJiwikUgkEomqPDFhEYlEIpFIVOWJCYtIJBKJRKIqr1zDmk0kphozKmfV5QJyKdIUSirlYPdUZvkAcriVWdk9vsUy6p6xl1H8LuqHWEbdM/YyvuzfxXIlLGZY0kTSWjdRPcNhze5KOc6DKrN8ALs06ytzqBggllEfjL2M4ndRP8Qy6p6xl/Fl/y6KTUIikUgkEomqPDFhEYlEIpFIVOXpJWGR2dhQ1D6E2MUhXJ4ZRuFfHrxyXEr6+AiQyvRxSJFIJBKJRFWQ1MKCor/d2Z5yguFx8S/8PjpdS0jaIJDrX6hZXX8lAQpTstT55Ko1uMjNuaUuYNrEY7S4NhzL9cazLo361YakNrPgzKi5ZbbLJFJiivIY12coHDlroOhEIpExKGofwo0gBZ++s46+1TJoOXKYdqVckeBdnhFG/FsLmXXTi9Xz21N7vvGtOF4UFshfdRajBlSaF68n0WnCYvrDTZw0Ut7YNwrJTQX2h6HGmSyUNS2R5RbR9ef9yN+/Dut1eVTDkATVJaG3FV/2+pmelrdQP7TYu1qjwkdhSrvlB9k6oRWKv44ZKFKR6OUg8/fh4gg7AL7vuIoulnkA1FkwAveZp0j4tAF/D5jOnIzmXGhqgrqgwJDhPhe5owNpi6yZ/8pCgu8tmK4Gfpgdyc4vXinZb8P017H7OxFlWrphAhU9UdzcJnivK0L6z6OrpUstLNjXZwYqjQVjbeI5O+AsqfMNEKQeSRQm5E4sXZg0XWn9wu+l04SlaJgVqpg4fEkr2aYCJGi/ZJnK6vwauIZ37LqgyszS5aErlczOlsClMWxxOHpvy5NHmY22iWNR83Z4/FU5sVWIRMK1z8LLbNJIQKKBQls1F3vPo/6i0Ug04PF7JqoLsQYKVM/C6pP+cTFHQlbRZuRIwd7Jql5rRFJnE9b1/IFgUxNUGjUBP43E66PoR/aV2dbk4lRf/AKTKZjuhOmOo495x6pJamlJ6pAGjBm6gQFWKSXbi+/dQ5weFgnDAP4BzPms9r/0N+sAAkhYrnf25HDw3Ee21zNRULfmxZLHE765yGtv9caqpyXq3NzKDFE/QutxeZyMn8KWEWIqYf5tT36c1RHbZY9+dqsquYM9yX292dJ5JjkdTfhfYASawsJH9nOUWRggusoTOzOISw20WdiqbGe2DGgBnHuh99JpwqKKiXvic0XtGjOh5nxeO/0ONlkv3oZlSDL72lyZX4s1jZZT1+Tx/3S31AWcLbKiuVlRJUf34uSuLiT3cGPMsA28bTWnzHNSpKhRA9qk88zQSNSo2TuwGiN2voPv6nw4dMYAUeuW3MMNtZUF6jMXWfbrfBxlFryyZCxum8pWz2a+H47d4qp/0kzbFEhU4/mYShSAjGKNCoBzb/9Am+iySZjMvjZ+22+y2UF7Unnl1VF47DBE1OUn8/Wiw6bjDKtxAIBEZQHtdo3F7KoJxf55xLRYVmb/EcnNufR1XcxvV/0ktLBDCD/+3yzA5Ln231fvN+p+NgrPT6r+5/NpEqaFs7vvdBxl5qhRo0bK+zXi6Tp5Oj01k6i5XBjly2voxolJcwFt1ZhEIuFlXGp4aMs9Jb9/tacrvsdevEuIThOWJ5HVsObN2Ts4WSSn1tA8lAJdIfraf3w4FRbJ0/7Z1mXXYfGKTpwYH1l5gVVQ3nIZR+rMefaOD2hpfpeYbnPZ3MaOj/b2IfCjWFS37zz7hVWQRGGCzc/Z9Km1i3m+frjIq9Ejvg2ekRdRPbSv0qJS52wqN2XrYLLH53A0aBVSFCy87cWOPmGQeoOiBp6kNjPDbeuRMifOhEh7NjtoM5Se8Z3w+uLkvRS1aitNVhJYdseNyNVdcd+cid+FY0gtLIhf6ldm/+uqfM7NqYfV5kMGivj5yexsmbfgB/wUj09WOl7sgrdVJnOcDpbZvrHfLCZN64jq1q3KCFNn5K4uxHzozKUe85Fygvm36zIrug1um6VYXr7N51vXEmJqQVZDNTUNHazoud15O4zuVjMBM+KLCwmYfeORc2p56D1hkbs4Y/NrLm9bXabT8DGYJVf9O5vHkXt5ENb79CPb/X8fQbUkGWatMzjY8BeWLer0vDdEVcbbLmUz3jq/jsY068kdo5a8O5fGptqPXXfLm3TtPJ+We0dT/ZeqfyF4mNTMjFsbXNjm/itjU8PJ694QleYEN7/zwDRLOM0i9+VNvE1U/V8JihyL++okNHn5qG5dAkC27xau+yiTrNwYFcG/4dMBM84UqSjumCOIvh0Alz63ZnONBA4WKNjctxkuZ6JKTobZnetzocW8Mvv3+HwSNj8L4+5c41DrscnKXXUhwX+Mx3fkYa44OhDeaRTLP5tNoEIBoH2NTFizVchdXWjwx1U21d7MvNvebB3VCpPj8fhlH4PQelwcbkOQqZrmZ98k4LOYCl3wRJXLYUginnIzMlX5DB03AfP4il3/9ZawyD3diRviRP9O+/nM7hzZajVXu6oxbxCB57IEwXUOC90Yy6d22tE+xRoVZ4pk9N80Ev/J51Hn5CBf68AbTgNwOH0EqY0NzVv34UD9Xw0c9fP5dVBbZk5UczJ8BQDD2vzN3j7BT+yjMvXLRkiC6rJ568pKjFL3pNWrc+nrOsQ1WMDkjAYk9rYn8x3tsHuLuJuPPTGa3azatYPbX1kNmOK+OgllSuoT95O7umD9Sy4rXadhLTUnTZXPiMmTqJErjAv6g9KV1khSbpQ8lvl5k9kzr8w+bS/0wG7TBcFc7FK/fHxNXuP1E/Adr70xUKalY7s0nTEpo0ltLmds1628b51ESn9/HOYIZ6RJ7Hd2bKq9meCjb+PYLQYZJ0r+TiktqxPXYy5qpNzZ60C17ASDxloRN/sEUWOV8L5fLyp2aWMSfZdSrIGRSd110hdQL6l4frdQ3t25l5iB8/jMTtu5xkpqRnyHxZwdPpfmf13Wx2H1pqhdY/rVKL3bPlMkY7JXMD4TDqHOyQG0Jw/N8fNolEqoZYNzNQE1jxw6g1v/OOqtGsMPtwIYY3OR2EG2yOr4Pfu1wOFCBeY3ivUcpG5Jzcy48bMjcT0WsCzbhX8/DQOJhO2DprEtrxrcyXns66r6CefrjKYAJA3wQFq9+iPP5/QNI+G7cMbu3cmPHruwk5kD0HbFh1W+bA9z/l3B+SIlPatlkrfWCpmdLQAWy+5w7tUVJfuNTW2Ked8cwTRZFnYIYV9w2b43gb+MpGuDtvh+eOKR/U13HMXzk2j+SGsAwL6JMyhqH1IpsepCc6941Khx7BbzyHO5vkWo0aBGjfuaSp91v0JSXy1bH1BnxDlkNjYA3Hw3nKuTI7DdJbDq+OeU3zWUpS1XoNKoSVbmk7bQWyfvq5caFkWOijmJrfkwxp5qSVKc15Z2sr0y2IeoETNZMmM83hMF0oQwMQNPuVnJw/6bRuLDk2O/1tGOE15rKyMyndEUFuL5STR7VwcTv7I2rVuc4toK1yfuf2l4ac/2ocf/g9ue45URps7c7B3EkUbaJoPIi69h4iLnShd73OQWfJIchup66R27ukUQN/3NcNiVhjIhyUARP58LneyZtLUJp0ZFsnagPXlq0zLPd602oyRJua97XGe85sYJpvbhPvNNRxhhNpb9M+fxd93fab2mFybfuuNiUdoJPKa4mBPfN8Q6SyDnGoDxGVSTav9uicoCpqZ0wm9RxnOPrKwmNaV4XBYmf+ozSN1Z7HoA/33v403ZYb9xkU2I7TAfKRKOF0pRJqc84R2qpldfLzv/1lLX/Rw8LqVAraCR6UFspOaPvCZ2Zh2qIex5yorbNmbyrGX3Bp5I6LzwQ1x+1k2Nn14SFvnu48h3gw+JAGVOhC5f36BP695s6TWLcRMj9HF4vfL/fYS2GcjQgeiJ6kIsSaH3Hz1h2HJYfb5poZ1MZ/DVlniOSBfcxc5m7VHqdR3A2fBVnAr9CUJLn3vX/h9GrHkbgPFBuxhotQBziQlN80ZgXcUTFmVaOnFdnQmd158uHmf5zK704j34akt+SQ7B2yqTha77Afj9rh3qN1WoMjIMFXKFWG87T0it0SyeMIfdr6yHn0qfiyku5v3/G4f1GgElK0AHx/Mlv4+M74u09TXg9jNfd32TG3ys/b2j03n28+gFsSpSo+HH8GV8t78jCdu9APDqmMAln/n3RihKGfTjaNwQTjNXftdQJjvNAsoOWW5qqgYK4Ql/G9M7QjuTPurqO0peMysGtMPR3X9L09n1oVJGCT3s5k+uMNkQRy6/xG/CiakzD5BwqBBqH5GUNAM9idOMKOo3eYdzET9qN1TtgSXl1m/ln/SslglIubC8LrYZwmpKANAolbgPSCC8z0gyG6kxdcwr+XsVI2NW6LqSfYddbU9UvBd+F+8KYliiMjmF2l3hmIsH7f3CSrabnojH9HYSlzuFwGJtwjJlTT/crgvnQvAwdU4O9pFRjL4zhn+/KZ2v5HyRkmH/FV6yAqCQqJDeO2lIy/GJs4+Mwi9wOIndFiOTCOeWavKNIIbZRrHRZzvqMRqk9y50gfsHE9NiGdvyrPFadgWloQMth+onUhmV2IuNPtuf+zUXiwspsJGj0GNc+iar68+eV+eivpeQzd3QEY943V0fDJKwAFSXqpG7OFf9aj4NJbPYDjr8Lp4/Pd8JUK2WlM5+K4Sr3HNKHxtBZ8vpqDHhh1sB2C4RXrJynzovD5uV0dishPhZ2gu77+738B3wcD+B2/hyAg3aPiBCGQ2lTE5B/sD36/5dzm1v7SlxTY4jXsuSBHUheBxZDWtyu2WX2dYrahjez/ldrWqKNbKSc8eNDW7UJvm5XyvRSCjWqCo0/XllOx4k5f2Q4cSP016Oqkeb47jqHCwENWr+e64LTskXDBxl+SivJUN/Z+oNHMWsQcseeX7UpkGYZknZOmwabnJtLczlYlukg27Aukd2FwapjJjx1XG81+Rc959BeP5Xt9cHgyQsBXYSctTCa5N82cldXRj0/naqS03YlV+dvX2CeWKzkYDIPdw41nsWH19v+phkpSzri9mCbg68OSicVRNmAQq+W9sLtxTh1q4AyKysuPZ+XU420c57FFNcTJ5agcJE6GmYll2PazDv2fsBIJWhMRVmk4Lm6Fm8+5c+Tt4UyKWQ5bx3rRVO3YWVrNynTE7B9asUZn8V+Mhz3vf6QLbyG0t82yUAdLK4y8YaGTx5bF/VJvP3IrbDIgCOFEpwXqH7DsUGScOXjSjfJGWGon61IV/2+rlcr5FaWJD0ZTj/RCwA4I1LXfD67sWmIa5qLkxxYGQN7Qiv0TsGGs3U/OltnbGSmrH5z7Bn7qs+JcyT530hI04SqFCw7I4bXiuvGTqcCrv4ZSAnxmqTlZZne/Nhz8HMSGlHoL2wpk14km6Op5A2ePSC9zgPXjCErrP7edRouDS7rqFD0ata+4xnlNDNYNuS30ec6Y/JTt2vn6ezhEWiMCF2cQgSU9Mn7yOXE/9TEMEmMjpvmqCrQ+uN9N9TfLb+rXK9JuHTBpwbNBcbqXZUUV6xyTP7vAhB6sQIYtstQiGRsTnXDt/Vec9+kQDIbGzwH3iRv/PN8V30/FXvQiSzrYm9STZpqnzmL+uK8oqwE5bLM8LY1EV781Nv2Sis3r6D5vj5Z7yq6vslIbjk9/etk+i69gDShnWe+brkb0orzNclNNJLbJUhLrIJU2qfxG/HUKqvE2az3vPKCBdmjdjDsvuFMXOqdmmPgUmv4zLspl6Oo7OEJbdzEPGdFiH90xZJ8KNZsbR+ADX2WxHbchkzbvoTMEPYJ8vHufp5BP8MnFHyOFFZgHpJbQNGpBtyVxfavxWNGjXFGhULRveGI2ef/UIBSBgfwGqP3UxcOljwF/BnSe0fwKd2Z/kirR2OM4XdFJTfNZTNvbQzvLa90AOvOZdQZWYhd3Uhwka4k4sBOA27zWtne5c8Hmx9lc4//0P8rDCkFo9fKC+7XxgL6q8BoPmZPnq7YOhdaD0u9dCODnLbLJx+OC9MoE14D/to6k+Emmr7XZ3LcECZfl0vx9HZJ6L6vlj+zLdgu/92hqzdQsbwcO72bkL87DCu/FqP+VuWssZjF99k1eHAG4GC7L+yMXwh3S9koHqtEXIX55Kf3J5N6H4hg62Dp2Ej1U5DfLwQBo+bgOV6YY+pB0ju4caX9tpZCtuc74l5tHE0BWkiGjDzrRW0Pt8D18hHl10wJspWwSwb/z0A/25vYOBoKkZWw5qlP8zGT2HCX/mWmPe+jSpLe4FOnVuN0TZxnDzvaeAoX5wyLR2rntfpHte5ZNv71klcfHMeyWs8HpkULq97E378dmbJBcPmP7f1dsHQp+x+YWzduBIpEsK+GYvZFmEu4/K85O6u/Nji0Q65QqMJb4CrvDRBDndKQu7qopdj6azTrerWLeYM7MuFRf8wwSaObp+V9hSTSaSoNBb0S2zDzc/ckSU9vWNjVWJ1GQ4UmNDcrAg/hQl+1lcZvGZJmX2k3B8RZE6isoBuiz/E9asoLAQ+AdB9Yz8xLwwAACAASURBVIZtKPldPa82quwkwwWjQ/HDZBzM8YM5tVDnJhk6HL3KcTOhvomMW+oC7I8Ka1bih8X+NxBP+R7SVPl8/elIqt3WNhukTYhgV6Pp7M6vScCiu4LuHK3OzYWOGlpu6M3eer+VbD8RuprM4HzyNFB0bySQrewgNlIzduRVZ2F4eEnyJjR1xp5DjZrQb8biuOqc4OZ2Ki+Nqcm9eVmETZ51l1yNCffHIV7Pt0KTn6+XY+m0zk1y8BS73o3A77cRxBYXsK9AQceLXQj6egRvNGpP9mt3kO0TTrICYLssmi8TOj97x3u6LZmE61fCrm5/UPYObwZYpSBFSovT/choIOfWwHBDh1VhMh9Pdrf4gbVHm2C6TXiLHJbX7fa5APSP7YvpdmGXV2WhPckPjuvHzTpSrv4vArfDluwfPwMbqRmTp7wr+M7RoB12bz0ojzbvD8Nvy3C25VkDYCczx0NugY/CFB+FKTZSMxKVBfx3/jva2XA1wptHIWFaOMvc/mXYtVbUnhuFKjv72S8SOrWaG6rSvoCf3mjExTnC62Ssir3MbVVpU2X6Es/nnpW5vHQ/rPnIWXyOwLhx92exTaY2yYLOlk0/tSL+t0J8FE/uUBx0eAB2Syxw23XMmKZd4UD9X+/NNgn7G6xF2kBK23eHGjiqirs4pjZvnhtEna/SBT8PyfMovnmvE/hCJ6orrqMpLjJwRBW3PWATBDy4xRS/HUPxX3vUaL6DyrR0TLel47cNFgZ15f/aWxPa9Swr3P5Brbm3YnpcZ7IWu+Ogo+nPK1vW4HB2953OoQJzrn3iiwxh3dS+KFV8Ij0nTeSfWdrOqjtWR+C4Vph/w3m+fiWj762fsmxNRRls4jgh0Rw9yziPpy8j4Ix2dIKxnCifpO6Bd/GJjhV0Aip3duLnzvOZNHEEyivG0Wz3LLWOyKAr7Pt+Pv7NRuA7RrjlDvw8EbqUPj5TpKLP5jF4/1aAX/RJNGohfzqfTHPyPC4nIfUbaEfDB55Jw4o0g8VVUXda5+MoM2dUYidke1+OZOW+6usO0XGddkSXo4CWHjAUMWERPbdXT/bHs99pQScrADnLTXl3xWhcN7w8J4ia53JoED2QoiI5AYtuC/pvqMrIoLNzcJltT1uMVFS1aYB5t71R9ZcZOhRRFScmLKKnevDCUNMIZrUFMG+XiOu9hTlfFppj53Dtpf1dyMmKyPh4v3WKHdQAhDdyVFS5JJpydNCSSCQZwBX9hVOGu0ajqVVJxwIqvXwgllEvxDLqnLGXD8Qy6oVYRp0z9vLBU8pYroRFJBKJRCKRyBBegqkERSKRSCQSCZ2YsIhEIpFIJKryxIRFJBKJRCJRlVeuUUImElONGZb6iqWMAnIp0hRKKuVg91Rm+QByuJVZ2R2oxDLqnrGXUfwu6odYRt0z9jK+7N/FciUsZljSRNJaN1E9w2HN7ko5zoMqs3wAuzTrK7PnNSCWUR+MvYzid1E/xDLqnrGX8WX/LopNQiKRSCRg9tFWxM8OM3QYIpHeiRPHiUQikUDtTD0FwADgumFDEYn0TqxhET03mb8P/S6msjP1FF8kHkXTtOGzXyQSifTCPtqq5Pfr4S/B6sZGQubvw53tPuxMPcXO1FPELgkxdEiCofOEpfj14GfvBGQNCUfu6qLrw4v0RFbHj26bouhfPY1ijYogEylvLdtu6LB0Su7qQtHf7mxNOc7WlONcnh5u6JB0JnVSBNOTDpHw7ZPLJLOtibpFEHJHh0qMTPQi7KOtWOV+AADvdcMMHE3F3e3dBE3ThsT90KTkJ/XDCO72bsL10RFoIhoYOsQKk/n74H9Mwfa96znUcD0A2/LMSOy0hMyhxnOuCT6pZnvKCf5IOVry0+VCFspWwchdnCv03jpNWHL6hjFryXy6XMhCYmr6xP0yhoeze/IsshaZ6fLwVUrSF+FMTzrErW2+hg6lwu72CWPxn8sZZHUNgLGpTQFQSIxnVZrkTyLIWy5jW+BvFGtUFGtUnOk3h+RPnr5KtxDERTZh5YjvqaswoUe76MfuI7W0RPWrBX/+vIzkPl6VHKGoPB5MVgZcaY7PeOEv/Fjtt8NIDp7Cd8zhkh+naVFU++0w9pFRKK5k8P/snXdcVfUbx993sDcIKhsFBHGmImhqprlnzjI1915Z1s/KsmHDLe6Ro6xcmbnSHJUpKqI4kCEiKjhRVGRe7j2/P65eJHGA93I5t/Puxat7zvnC+Xy9Zzzf53m+z/fuDn9jyyw1icsasGP/Rua5RxEW0512zbvT2r0OU6YPACD6k0Um4WnJ6RxKLevLFKDWPUdVgprBDslsWbOAlLmOL2S06M1gufdGGBFfzaOmuRnDHS4ikz155pUiD1SChpXV13Blc3V9SSgfyBVc+zWYUwPncVttjcNXZTcdzBDIawTx5qfbqaiwAuCL9FpceNWMVmdfN7Iy/ZDXrgE9467x2/Bv2V39l8eO7xzxLRenNnqqAV7emdVmLXXMlcSq8jk+vm6xbWQKBa+6JZSxMuMgt7REVjcEWd0QY0spMdldG+qMFTDdUJCimj+XPm3EjZHaAYPa3YV3/P8wsqqSo6jmz9ikeC60X8bYKw14ZdAQHNoloU5IAqDCkki2Z2sH7hfaL2NsUjwXPxOntyWncyhdvtxDB5urT2xzpOEK8vwrlvocejFYFI4ONJ10mDrmSgpQE7B3MJp81RPbO38XyZ5sTwLNLJkYvAeFi7M+ZJQLCprX4XiDtShRMHrZcOT/xBhb0gvhtzKFoQ4pABzMNeOvDxqhyczEum8uP7YSt/chp0son81fxht2l6isMC+2TWWFFccHzSFl8vOFOsszvVa9g/yvE8UftLKkmU08eUIBDhcLylaYgZHb2FDwaj0uTWlE6qYQGh+9w9Zta9i6bY2xpZWI7K4NObBgiW67yahhRlRjWJIGuHJmyHwOTp5D0ixxzoBSVPNnx/6NtLfOJWj5CBLqq7DYGfVYu9H7+xIW052g5SNob51L/GAReltCazLy2w0MdTz71Ga1N4zH/HhSqU+jF4MlZbkX09yOA1Dn0EAC+h0HzfOFC/raXSOtX5A+ZJQ5ae834vLGGjoXl8LJidcjdgNQM7If3itL/8WUB5K/CefLSvsArbHy6ZhBuhuu4Np1Ci5eNqa8F+ZyJw31zfOfq+34HlsMrMYwqF95ierm19mU5USV5U8uUZE0rioNLGQMv9wSq1+PlqFCw6FqWY/ERaG0PprG798v5cyw+cSEreF9l1gA+qa8ZmSFJcNvUpzuc5NRw7DefMSIagxLqxba94mFTIl9sjjnhozavg2AsJju+EwpPhQLEDwjA8sIJ3ymRNLavQ5+24fovC1iYcfm1XS1vYGZTIGZTIEShe7zo9tne0ZwtW+NUp/nhac1q5u/xPf1FgNKTuer8Fho9qJ/UhRomtVlx6hvydQoeM+mLwCqjbYMd7hIzch++I64jvrmTSOrLD0Z/cM5+uZMbOXaUMiAvwYSuOPx0YFYOT8jjMTWCwCFbp+ZTPFYu4f7fM3TIawFHD5VVhJfGKWHOx9+9x2BZjZ0+OlN/FKLf2gqfb15o702zBD1ew28OVSWMvVGTpdQcpwUvDQihiWekagF7UtvU5YTnRM7IpcJaAQZcXGeVNmkRrk32siKn59H81ZM3VjJ7tqQTyvNAqw4na/CJTYXgE9X98FLJNdmXtsGtLeOISymOw7tnj5wVSckYfFINDZwSBRBn40gfvAidh0zI6H+k6MV5YXA/YMIr3KBxd6/a3fIQCVonRavnuyLQq7hj5prATjwv1m8kjsBlxVPNuKexAsZLAWv1mPRynlUVWrzGwZ9NZ4K+59PxBex7ejR8HsA5o5ezFeza72IlDJFUdGN1LH5eCis6XqhHeqEJGR1Q5hTdTmJKgHfUTdEbawA5NvLdMbKyNSmBL9zDtNJsYW4NxbobqhHedK+ppaZjBhsTqBI8hvllpbcWGJLU0s4r7qPz/bcJ7a91sqDKRV+BcB3yx00ZSVSTygqujHowGFaWx/GQqYdMKkFGZ3PtUd4Q0DIykZ974ru+g3kivHElgJt3oo2FNTvYlOTNlYAXMan4CTXvlMGzJyA2/5DyCwt8b1gI5pn0HsR2nfbs4yVJ+EzJZKw+t05XGcjr7QdUmwoqTxRbVoW8hVFnxxDL7bl9K/BeP90kTvLi+YAykr5kHkhg+VyS3OdsfJZek3cfjrz3A87r/6X+OKfGnxU4QyWsvJvQT5K3FfeJIUtJSa/gPufe2DGNcZv2EiQmQXBa0bhd73klmN5QhFYlQ6DDui298QGE3jnmBEV6RdFQBWgcHQdpwJHeT7eD67l+4KKhtsnELQok7SWThyeMAeAzS0WMK7jGCy3lv+Qyc2+dTlSdwFqAfpPnIjNgeJfcnIbG14Zpj3WLaktssvXylLmCyFTKkmcU4/Fbb6juVUudzVqfsyswrfHW1NxiwUOfyWjvn7D2DJfmEdDQQcPV4fZxbdz/1swCWOmun1h0qZVuvaNosnNhdwnG93ljfbWubpk2tLi0C4JrkDumAwsdupJmIFou+EIAxwK3USL7gRwZ6AL7gmHKAD21YxCJRS2d15ZunfkCxksS3tprf7ofDV/TGuCXebzDz81mZncK9B+oQ7yPBQh1VDHlv9ZCmnvNyK+9Vxi8gXe+HkcfnsiyegfzmtWx2kR+zp+H5b/l9mzuNypIltc1wMQnQc+G2Rcea8RTXoe17U5m1EJZruWe8v/32ia1KX1or9023EqGPbFOHJdZZwYEwFAw+0TCBx+FA1gE1KY8BdoJkNlI6e8T8aXKZVYdNPWPd1w3wWHgxd5Uhrt+Y9rsa3SAm5pcsia4o781hOScsshua3qktB1IXJkaIDGq9/Fb2o0VVXaPohlNP40kmaHsctnsW77fK/FT27cC5og7nCR0tebQCuRuDGfweeJHXDgxfIYg5aPoFW7Y5T3N+OKxe0Z+kFhX3eGOMIjfW/Xdzg7vi+8duud0BBdt+S5SS9ksLxipUEtwIfJr2O37vkuMqWfD3k+LgB4WPwJQKCZJd02/sX64PJdsOrW4HAWDluIEgUfJHcGGZyb35C/O00HrFFOcwZNCppmdUmvYYXbAnHEW/9No56FL61lN5phPSmNPwK3Fm3kDvsjLJnbsi0FKZfKWGHpUARUofWivxjuGM/D3JWhX4+jwoNYatOLo5AJELzjbLEvu1ZneuP0z+UnvvzLC2kbAomp+QPxqjzWdGmJ+uq5J7aVV7kPwI/3Qp48g6icYnk9m7051rxmlQPA3N7fMTl1IK6Lxe3hfJTiDJR/F4prHHZWl99ypakM/81lIs0gnB/gSV+7ayhkcuLys7FMF5f3/VE+DtzGPF5sQolNGsxzj6I15beq+K1B4Rz5YG6RnBUnij5zLnRVFgm3b9jVmCqUcQ7LQ97z+Z1hiwYQPPtWkf23wtxQdb9dZF8vv+O86/y4vdjIKpn1lG+DRd41ncYWWhfl70FbKLwWrVHI5KxaM493L3diTOUVDPp+tNF0vijdnAvDP4u9Cr0RPZLakfxbVfoO2MV4p0SaW+Xyw093uC6SsgFXX6vEGKdz8CBjHcDtcGHOhv1PWqP738bKw7b7a26gaZNR2P+UVjaCS0HKF+GcCZ0PyOi5ZCKecU82mi992oiTjecCCiL+aIM/4hrZCtGxzO3Slfnz77LW/xdes5Lx97CDRC8W56ySf6Nd0LCwLELVdcNx/1vAf3PR7+k6ILK0nCcyqNsuANSChjk3WmC2RzyJ0Y8y9koD5rlHMc/YQgzM9TGNmDdhoW479PBg/MZnPDaoi3t9fpGQUJUPjBASanq6K/trbKKFVR5JnRZDp9L9nUsF2YwZMBYFx5/d2IhUeEfgpbZaQ+S+t4ZFnVbQwiqPA7lKll5rhgZtsbyhJ/ris/O+MaXqlQsFuQyOfwu7QflUTj3E7iNNGL8uEYA3XY8wm2AjK3w2t4aEs/K92aiEwoKGKkFN/Ag7Ap9R2fzRkYFMeErDckDrNseQI6Pp6e54fl30oSCvU52L7R3JDcjjRMv5WMuikaNgVkYA1T5PFGUIRXMmHs0r0L77ePbPXYi3xS2icTW2LL3wqHeltXudJxqU/zZsxEpO51C6280CrEkpyCZqZR1cHxmFKwKrok48bzyBJWD3jvowOIq8tg1EFzYvCe57bnJsWBUaPpjmNKXmdlZV7gCpRQd1YdPGceB/s3TbV38NpnKXOErKCw1FrFpfIPR47xL9ztDLTQnYOJKAjSPZm6PNHO52chCK/eXbWAHt9LNKcw5Rac4hqm7KxUt5l0RVLl/1eYtbjTPIaHybjMa38ewWK6rpr89ifHIPrFpfoODBRZhTsTDje8TefsaS9dzkdAll+eQ5BJoVGit7c6wBONJ+NomLQ1E4OhT5HUVFN2R1Q5j62QpAm4gbsm00jjueXhipvHD1vCsKOztujGzEtQmN8Dxsy3e/LeX0yPmce20ZtjIL5A8M7IWHXkV96/Yz/mL55qEdOufMq8YVUsZkd21YxLARc5n+rEoKvJXa+/Lj1I64LipqrFzsVvoKqWVNlbXaWaIPZwuVltv1y3kAWibDTKZGiYKT+ebM/Lo3QtTpx5oJcnQ1WZbfLX2Y7IVDQq7dL9LRtiVJE6uh8SnM4raLtMLucgE3XlJSZV5hARwhJ5eAbO1NtevlmrSwOoZ6v8uLyihbZDLOd7PER6nkpRXj8TkszlyVZzHxqjbh9OomX9we+JzltYL43zerAcgW8nGO1ktU0aCorOVFjJVWZ3pj86ktIwY/qG6r1KAO9IajhTdaylB/jg2bg5lMgUooTMQVixfi744zyesAvso/H9lrza9Zjrz7Zy/i22tzscJjelFtzCnKueNIR+LCUBzilFSMKHrPuY5OAaDgkriXwnhesrs25EpTmc5Y6XexqejL9Oe6FN6j5zJccSZDt61xsMY80xiqSoc6IYmwGO205M93+Jd6evP85t8z9koDoHzm8iR+aMN6hwQKUNDv4FD8nzD759Ecl22jXqXyn6VzULzw20bIy0Odl4ff5OKF+mwpPltfaFyHjo7iKo39kFuDwkjstYBxV5rg84lpGisAjeySWNPxVdwStX2UW1sTP9qOVlZZAETcrkuFpeJKcGx1pjcOg/MoSE1+Yk0VRUAVBvf6XbddY+9wgifFi8JY2XqsLtM7HqKywlq37/vMSqy+HM7dX9xxXRRJIFGoUtUoZQrunK6AU96Tk3LLG0mdF/P9q5VYe6KdbtkLRUAVZvuuAazw3l3OR6QloN/FprpkWm3YR4vWSIkp0k7sxgrAuL7aWkA31NnYz7IrckyWX4AySyxmtRbnCbB9uyWH62wkbMezC8j9G0U1f9pbxzDxx/r4lCJBtSzw+MmMbfUr09X2BsPr/s22Di2w3Ga4mbJGGx7LDsaw9U4dmlQSV30PuZ0d1j2uEa/KI/bjWphjevHJkUfeIu6V5XSzTWfmLAWOX9dBba2kYOItEmtoR3Tr77vxT/96QKxxxT4nDxNn09KcsUkt/ppTenmS/Z2C3dU3PNijoINHPQI4LgpjBSBw+FFe/7YH96u7cbWxAv8V1+DOPczTL+KKtjS/3M4O+YNocMDSq+V+1tOjND7Zk4O119N33Xcsu+tFFfMbtLKOQS1YE7B5BAG7xTul999cD7+nS6Z90nRmU6x6+9qCSXjsKzoQ1JyMw/mkkQSVEnVCEgvad4Dt2zhcZyN+y4YQOOT53hd5bRvwXsT3bM+2fGpZf2Njue0oJz73oavtDUY4xuI18zarrnfQhYWSvw3n9df0p9+o/vxTdzyYLNfgsf26aF4I5z4JIaHGQuocHYT776ZnrABUma/h7zBzmlrmc6juT7Cu6HGVoGbO1z1xPlF+b6R/89Ad2a9+JGu+Cydw4ONGS/Z3CrYHb0AlwMCLbTi1I0g0pcAfpSA5BcvkFPy2Fe/dTB1eEwvZXwxPbYI67ckrq5ZHnN+VU+vzvvT0P8HkCqdZeteXYfvaYHXJjKBlSaJ5jjwvTUYNK7LgIRROa/afcBhrTMNYyW9dnyZW8wBLKh7LM7YcvaFOSGKefxCjlzXgQvtlcEW7tpBqiysVlhR9fiqq+XP9FVc+e28l7a1j2J5tqTV4XrCWiyFJ/jacZS7TsZDZAtDV9gZdf/0OJQoKUMODAb0SBcig/pxxuP9Z+meqUQ0W88FyTpuJJ/MbYGCr/WRocqg4r7yXDys9ssiTfNunD6of1tHc6r5uNA6gQUPIb6MJLGWlQmNglqUhUSUQaCbjfZcTvN/qBOvjPVELRXPO+9lHczDXkqXXmpHZywqvVPEZK89DzQfZ+dHXPXHNK+8lqYqiPpuIZzc4bGlPp6C+yFLSCLyjfSiamrECYL35CK03F63BIbbp58/D3VGZBJpZMisjAMvzN0Xl9XseAodEUW/YCBoPOcbhOhuhDvDJv1sVhvn8tg8heEYG6oTya6wAVJkUSQvNe1SofUO3VhBQZC2hYrdLiVENloILT149trzSzu4UDXaPI3C/uEJZJebwKWb7B/Pu+43QNLhHfr4SWaoVVtdlBM4U14vcastRBruOJ6v1faLDtbN++thdLeYGUjDyu+F4fXkIHkn4MzV8rW89u1E5R5ObCzHimLEl8Ww+DNpJAWpWrW2NZ4q4ni/PS4UlkSQsgXbVupPcxxXH+o+vN/fQ8xJIlGgM8CofRKL09KDh3EEcabjiie2GXmyLx5+ZL5TgX/6neJQj5HZ2mMk0KCzFcim9OB7fmMbDw2V5JBWiQ2Br8cdrrx1HhRgB35ibonlQlJY9cxozddoJ7iU4m0jFEglToEnMm3h+ZRrPm6ehTkjCZ8qTPCfl26PyJApS0/AdBy2XvsWeWj/o9tfeMB6/LfkAmF+9h5Dw+JTnkiAZLCVA7mDP+rv1sbERzyJcEoUIJ2Lp6hla7LGHZaJN3VgBcFodSYfV9ahaTmceSPz3WBpY5bFy7hLioiA1DYd20I3CGW2Phi/18WyVDJYSUJCaxuHaZlSi5BX6JCQkJCQkJEqPaSy6ISEhISEhIWHSyATh+VNgZDLZTaCsMmV9BEEo0xB7GfcPpD4aBKmPesfU+wdSHw2C1Ee9Y+r9g6f0sUQGi4SEhISEhISEMZBCQhISEhISEhLlHslgkZCQkJCQkCj3SAaLhISEhISERLmnRNOazWUWgiVls3x7LlnkC3myZ7fUH2XZP4BMMtLLOoFK6qP+MfU+SveiYZD6qH9MvY//9XuxRAaLJTY0lLXQj6pncETYWybneZSy7B/AHmFjma9NIPVR/5h6H6V70TBIfdQ/pt7H//q9KIWEJCQeILOw4NyChuy6EsPFqY2MLUdCQkJC4hEMbrBkd21IxUh7MnuFPbuxhIQRufn2S0xsvgOVoEYmzfaXkJCQKFcY1GC536Mha+fOZIX3fnbOnE1Ol+LXcZEovyj8/Yi4eJB7b5i+wVlhSSQz/m4LgFkd012tWUKcyGsFcXtAuO4nr10DY0uSeEFyOoeyI+04HWKl583zYDCDRREcwDtf/kRlhRUAKkGD1ZUcQ51OwkDINAIqQU7H/+03thSDo/Rwp2P9EwC4zbEyspoX5+bwcC5vrMG2tGgub6zBskv/UDHSntyOpjtwyOkcyo3RjegQm8GuKzFMTIrlxuhGKJycjC2tVCQuaUDqphBSN4VQfWUiUV8u0v18HPEdWd0aGluiYQmrxcSkWLalRbMtLdrYavTO7SAlGgSGOiaR9n4jlJ4expZUrjHI4of33gxj+bTZBJqZ6/ZlagQ4+mJLS0uUPQXJKWzLrEVt64v8U6Ex6vRbxpZkMKx+zmdm5cNMv1Udi3PXKDC2oBfgbp8wjnw8Hw0aNMCJ8JWABSu89xM77w/GmI3F+pcjxpapVzJ7hTFt2lIaW6oAUAnQ3ErN0f9FENxoEAGD89BkZxtZ5bOR1a9Btoc1t/tnkdhwMWYyRZHjR/O0/WthBTXnzOJ12TvYbDSt7/IhiUPNedkyC5UJh2jlyAA5J8fOp/XRwShS04wtqdyid4Pl3pthLPpybhFjRUL8tLHKZm6AJzITNVhUreozwn0lV9U5rF/Wgopph4wtqdQo/P1oPenAE48Hm8tp9PERTv/ljPrW7TJUZlh+nTETJ7llscfimq2gzvgxeE4r39+r0LgOHZbuZ4zTw4kSCnomt+DUvkBdG8cEAWTw1zcRuClsaPzhEWI2GkevIbkxshGbm88GZMSpICrHz9iSDIIG4cGwwnTnwChCqpEwxIn4HgsYevkVrne0RH3zZon/jt4NFpehFwkx1/7ZQZeaM8tzJw5yS1r+PQZ/Tuj7dEZHbmmJ3MWZuMleAJi7ZhP/8vcALL7jwfJZnXBZHmlMiS/MzqshvOucwPmeVviLuyvFIq8RxAcLV9PUMp/6UUOoFFG+X2rPIrtaBTYmWaPSKNiSXFO3XzjhwMkREQB85hZFx6AhyA6ahsGi9PJ86uN++q3q+P6UVu69Zl/8sIxQCzOuFtznp3u1+G1SC2xOXMbnatEbTxEcgEpQYyEzM5JSw5LxdjhHP4xAJWhLjnTfO5LAwceMrMowPPSwaP9vGii9PEl93Zuxw39BIdNQSRlHc6v7aIClXn9Se1F/vLqXA4Ml9pwnBEDg78NwPmqG2ce7AJBfKX7kI1YU9vbEfR1Ex9ATzHPfgVrQ6I6pH7gvhzhcZk6r+7gsN5JIPaFZ6AYL4I3mB4lC8exfEBnxI+1pbpXLG8mt8Rx9r9y/1J6FxfYoPLdDjKMTnndidftVLevBCCMKMyBJM5xxeIJ3BeBirjMFF8q8REeJ6b1tNAA2lxS4Tz+EBVHFXo/pM8D2Kf0VO879LhlbQpmQFZBvMh4WuaUlQnBVlLNvE2yfxi8VN2v3I3/Qv0LsttiV6hx6N1iC3omn04K+BCWeRWZtRdz75tSz0PdZjIcisCrpH/pezgAAIABJREFU4W7UHRXDVo/FD/bKyRMKWH/fE4BKyru8ZmU6CcbpIQZJdSoX5LVvwMrWWovy6ryq2KYW5gIoPT3IruGO5dX7aE7GGUtiqVHfuVtk+9OlKzCTKbS5Had7YHMwxkjK9MvNEeH80nAWUBiG3nS/At1s040nqpQEjHl2Lkpm7zBWhswGtInhh6eEYslRAysrOxQh1ajplKDL3Vl5z8skvSsFr9Yjse0SnYdle7YDFueui3PAJJNx/tO6nO47r1gD5d84n7rzjBbFo/c3kSYzE2LOajeys8kVzACVvk9T5ijs7Yn7JoiNbeZTx7zoP1u6Oov2k9/F8ftIFPb2pA6pwYl35gOQf9PaGHL1iuVt08x4U9jbU2nyeRpbquh6rgOO/1yEShXJDPMho/993vKP4h3n39ibY83MQW8i/0t8IU25nR0FdfxJ6qukivIfVIIVGjTc2V8JO7ub2vtV5Mx6d0mRnLmJV8NIHBHEsq/u8nvwZiMq0z8yM3MyumURYq41VgZdehmbwxdQG1mXPkl53YWNrkdRCQpUgprZ67rgjbjDtMVxoatS51nRoOHjM51wTz1rbFklIqdLKOk1lSwbMJ/6FkWN5tfO9ORyqgv+q9R8unoF9S20V2ngrmEEniydAWq6Q2c9Irez4/wyX5JeXszDf7Lt2bbsu1edLUdewv1POY7rtTHm9K4hOmNl6OWmBP0vTvQPE4dk8RucxXGnXXW2+C5gxV1vhDcEEAQK1pqxN2hRkXYtrLIZOUhGwF9GEvoCJHwRQlz3iAdbha7O6LFzaXp1LI5rxJuUlN+mAVcG5BFqcRRQ0D+lJfdUlmh6FCDcPM2dDeFcmJyLn9ISf+sbXA6sgzrxvLFllxqZmTkJC2tzofEyAD66UZOr/SuhvplkZGX6Q25pSZ5L4di75vqxBHwVjSkOmdwDbiLX5a7IEQ47GltSibj3ZhhLps2hmllhmsC8jCD23ggiIdGD4EnxBN67wNTkaOpaaHTHq0+5WmovksENFrlMo3XtiTyfKP+eBR0TO3BtvQ92l9XYXLiLcP4igbmFVmVOl1AWfToXUPJSVB8qf6lEuCf+qdyXW5meXaus4susaQsAmLe2M15XD3F5Yw1OBq0GYPjlZuw7UR3bC0p2jv7WmFJLTU7nUH7pNJeHsfGWo0fT9fM/GOWUAECtUadI3WwnWi/L6zN3M9wxGVBQ458B+PU5i1BQWICrwpJI1o4M5aMKpxjvlMiq7q3xnCZeg+XSpPpcaL9Qt71u98tUSRCvwVkcQnBVTnefp9u2uCVHyMszoiLDcGtQOJE15xfxsLicFVkwSACNIGPlXV+WLOyMVboGx1O3Ec4mEkgaarTJt/UsotEAdzX5rF3SmoqppfeWGfxNpBHkqAQV3rvyDX0qg6HJzCRwSBRqwJUrAI95TeR2dnhPSqSOudZY8RhwDXWGaVUvTM6qAJhGnzJCK1HPAtrFd8F3eRLXh4bza4MZRObZ0H/XUILfjyfw3lHu9A3HRibOZLg8ewX7soKJz8/gu7c7YX3kGGvdWjNqitZgme/5J9U/G4P/hMNGVlpykr8OZ6jjPEBO4JYRVBt7HKHg8Qf+oVENYN2psheoZ5SVK9Gjp9bFNzqtIYdWvETV5VEm53m4MbXwO5x8vT6+v9wUvYe6OFqNPlhkdlB0nhzLreLKQ7L/6TBjs8dgteUobg9Cdo9+V0ovT0K3JQNaz8raJa2p+IIzMA36JM7tGEo1s3sAKPeaXpXChygcHRgQfYrVPvuIuFMFj35pJmWsOAdpa69EHQgyshL9ILezw3eM9qV9e70nar9KrP9wOn5KS+akvkbgiKOo72mv21xnGVuzvAmcLb5RnuP3keyqYc/Kaj7IIk+CRk2FJZHUnzPugStajuAivoGEIrAqozruRI6c+5o8qn97rVhjBUBeUJrUvvKFvHYwbfbEMdU1FpWgZtfel3BdHPnEPouWsFp8Xf0XzGQKzGQKTr0koI47Z2xV+ie0JsNdDulmB2kQGLB6jLFVlQqrLU82smTfq5lc4TTTb1XnzzZBL2ysgIENloxAJRUUxZc4l9cOFm257H+T+WoQ3WwySFTl8uuk13QvO1Ohm7d2NknTV8Qf3gK42z6E1b57uFCQi8VdAZtvr+KttCLk74HkDiqcbievEcTgodv5em1PhBOxT/mL4qLS4ZwHD0oNcS2WcLePuNaJqrI2lVGO5zmRr6H9u+9QkGLaU2DPveXIGKeLDLr0Mi/NHUOVD0wrDPSQwqq2aqrtGWJsOQbjXF9rKiuskCNDjpxPbtTFe6ppJRWn/q8REX4b0KDhYDt/CvRUvbfMfd1yOzsSl9fnk80/ED+7SlmfXu/ktW/AznlzSVNnM3L4OCy2RxlbksE4tLOWsSXohevttd6SP7MDyHaTs67q70TmKfCbrUF9TuvCvD0wnHbrIhnumIyF6TjLAFDezeHkI46VsHfEM2VU4eiAjVL7/c1Ia4PduqeHs/KctYnGczIC8dlwzeD69M2dfuHs6zWd+5pcDuyrifu3pvVie0hOl1A2N1+g266000QrpYfWJOH1hTrPigYNx4fWNrYqvZLxdjjbhn+Lu9KCI3lmejNWwMAGS61uhVO0kmaHIbe2JnVYTRLbLqGeBQR6i+8B8ijZXRuybYl2BsaAAeMw/930jBWljxctbbXeBZ/f7j6jtfjIqajNAriiciK1pR2pkxuRuLw+ez6bxXDHZJbe9cV9U7KRVeoXzal4hs0Yp9uOvVPZiGpKRsqoEL5w04aX701wf2pbVct6fDBvDQBJ2W46Y1RMTPzoR7yVthzJs8Fvsml6VgDSuqsINNPOzJiWXg+75CwjK9I/CkcHOqz8W+dZkSMj5M+hCFGm4bl+SNCwWJ2x8nm/t/X6tw23WnM1fwZX/Fu3fbZnBK8evc7xCRG6fQlJT3/glGcU9vZMn7kQK5k54VEDTTZH5+z/KlPHXMkfOVbIr5S8lHJ5JHC6tqjfW3YpyALus/SuL91s04kZFUHMqAgS2y7BWmZOyOrRbG/sT8FVcRvWxXHfuzBd80KM6a0Qq2lSl0mLf6CFVTaReQoSPq1hbEml4u+72ryxcIsczk8PN7IawxHfvLAc+NpjDU1yodwrfUMY6pik86wMvtwM3+Uinz77CIqKblSMtGeNz99syarAl737Iv9Hv8UpDWKwKAKr0mjDGd2qqQBKFIx3SgTgVL6aZu+MImhivCFOb3BULevhtUdFAwsZ7eI74d5VXMV+SkK/8IMoZHKWXmmK+voNY8vRC5oz56gd2R8zmYLTjVcx1CGlyPGfMivSqd1b+H10VNTJ0zeHh3Php8fdzfI61fmy64/IkXNdnUfVieKbJfQkZBYWnJvXkCErftEZK1PGDMFihzi9n4njgwGwlpvT5pXjpHwRzp2+WsPl0qeN0DSrq2urqOhGyhfhRX7EwsNE2/Bj/U2yqq3Sy5MhI7YiR46ZTIEcOVfG+qLYf9zY0vRGepuqLPPei0pQ8/6+XgYxOg0yrTlxmCtbXIpPUmx04g3cht3HLu1wqUrzlgfujstkocdBovIE1J+7oSDV2JIMilrQcOpoVapiGh4WNGp8PiuAnUV3/5blxGeL3qLy/GMIKvGV4v83GXXUbA5bytD+43H55QxyZ0eyqldi4OzNdLZJ51S+wDtj3hFVWXffJQlM7VGHT1xjGPPTBrI0Fsz+9A0s7mgnVOaOziChtrZWyd4ca74e20+0xgqAMvYCgX/3I7HpGuZ7HIGBR4jOy2fpmGbs8JjPz2+48vdd7UrOruZX2eG2mwuq+2QL2mJe73xU/o2W1P81QiVEoxLUuH+CaN8LT+PsJ5X41WELGjSoBGh6qif2JuRFSllXi+8aaHOQXovtRvCkeINMRzeIweJ8Wga9tZ8nX6/P1VwHDkVXw3OvgMvOkxSIuBDQrUHhrK81nYN5tkwdMgjln6YZCjJ1NKfi6eBR77H9lThkUrUtgs3lHJg2jwmjm1DL5jiDHC7p1vno+etY/LeJy7uiTr/FHZV2uYtWVllAFl2nL3isXaIqn6/HDhe1sQLa9aCcdtgQrOhL6ypxzKl8jHoW5izxjATk9LG7RR87bW5L09NdCT5XB5+vNQjR4pnV5vVa+V+U8kW4OLURiW0jdOt4fXKjLg5TrE3mOZPTJZQNDefpKt5qFrihvpdikHMZxGBxWhVJh1UPXwYCcIcAjui2xIqyciXeemcnlwvs+WzEQMz2mp7r8t+s/6UZTj2zqLLJdBZz/K9gl1h4e892P/DgkzYKXH/uOAIXnRblaPb8AD86zHVlof/PeCuLlk3IE1TUWzWBqnMSsUgXt7HyEMc1kTiugYQaQYTWa/DEdhW2JWB164Kon7GmSJWfb1LNaSQ/dljIgNVjHkxhNhHvikzG3gWLADPW3PNgfXAlrAzosTW9musGQlmpIgnv+vKb4w4CN4zEf7e4RqalxXvqIXZOdUTGSWNLkSghnqvjWTCwmq4U/+RrDfklqj6eu2R47jyOJjfXyApLh+ZMPLSAPm+9i8OAVLYFbSH4z8HYHbRCrhbwXRJpktVRNWficTrz5OOm2GdTQB13joCx8MnYeia3iOP56Q3RcJQjeWasmNoFOwz7XpQMluckL8iDhN4L8d89hKDJp0Q5MpX4b6G+dZtdNezZxcNRuYbAB6MfU7h+HX44DD9AB+pRFfGtpC2hRXg1jQ489MiLP3fsv4IioApfdFwHwHufjMDxZ8NPuxfnIilljKKiG4HTz7I3x4Kg2VlosrONLUlCQkJCQsJoaGyt6Gp7gwEprXD5vWxWDJc8LM/BudnufOCymy/69Ed2SgqNSEhISEj8txFOxNLJowFluSCuTBCeP0VLJpPdBMoqpdtHEATXMjoXUOb9A6mPBkHqo94x9f6B1EeDIPVR75h6/+ApfSyRwSIhISEhISEhYQykHBYJCQkJCQmJco9ksEhISEhISEiUeySDRUJCQkJCQqLcU6JZQuYyC8ESG0NpKUIuWeQLeWW6lGVZ9g8gk4z0sk6gkvqof0y9j9K9aBikPuofU+/jf/1eLJHBYokNDWUt9KPqGRwR9pbJeR6lLPsHsEfYWOaLaEh91D+m3kfpXjQMUh/1j6n38b9+L0ohIQmJf7HrSgxJs8OMLUNCQkJC4hEkg0WiWOQ2NmT0DydlXS1ePZ3FjS1BCOG1UTg6GFuaQcnu2tDYEiQkJCQkisFglW4Vrq7Efe5H99AoplU8hplMwcFcDUNWjMZ7RjRCXp6hTi3xgqR8EY4s8D4nG8/T7RvvfBb5RjlfpNfinwlhKPdFG1GhYcju2pADC5YYW4aEhIQJIrex4fbrtXjlnUimuR1HIZNT6+gbuE+VI5yMB420fOWzMIjBonB1pfeBE/Sy2wHAnIwg8jRmOCmzODFyLjVtx+L3P8MvlFQWKAKr4vXDFRZ6HHz8mExOs2FDsdxquOW2DcGwLrsY43Su2AXyJleIYdeSCyyqXcfk1lS60rQwl81/wn9jNW5T4U6/cHpN2sU4J+2aJgqZHLWgIWDzCKq9fwZNVhbyOtVJa+GI+9yjCAUFRlYs8V9C4eKMfJMF/wTMB7SLj2oENdENfoBt0KVxVwpSLhlXpAgwSEgo7nM/etld5USenLAvx7GvvisHalmyvUkAU240QG1pGtV15ZaWVFh9k/ke/6BBeOxHJajBNLoKwN4cawBaW9/F7y8NisCqRlakP7K7NqRx2FkAmowaZmQ1EiVCJuNuh/sPjGyBNHU2L596nTpzRnOqy1ySPqmF3NqahAmWBHZJRF4Ow5pya2vk1tbc6RdO4or67LoS89jPtV+DuTqxETILC2PLlSgBCldX1Out2RywDYAFd6oS+tUY1mZW1rUpWKHh8seNjCWxzJmSfJxtadFsS4vm2vjn77fePSxCeG0SOy5CA0y73B63hYd0I3X1rdtsPV+DaovTEbvzS+Hvh/OaDFZ4F83avlCQyxdX2jLVfQeeSisjqdMve3LsmD6uLxa38pj/9U02BfzGbPcDVB85Bv/x540tTy88DAX1u9gU681HjKymZCi9PGm36yRDHVII/mvQU9smNPsODQJyZDQ93Z3rt+1x+80Su5/F61GSW1hwuvEqdmbb8dnX/bG9UoDNzihsSEY9TmB/7+m806gzv3tFMKbXCEg/ZWzJj7Ez6RBqQQNoBz8XVDkAmMmgskI7UDjeYC00gJ+HuvJDt9fQnIk3omLDIKtfg6TetszqvIb21veBQm9Z45M9sfvGDvlfJ4yssmRcHBxATFAEAHUO98P7Mw3Zk+/Tx+6qrs22oC3kVVNRX/kOPp8cMpbUF0amVHJzUAOcY3OQ/xNTbJsboxsRYHYQDeYcy1PglKB67r+vd4PlYntr3edrK/1w4lqR4xV+sKbAWUH/hMtMW90Lz2ni+3IUrq7YrrrHSp/Hp5gtSW/C9fB7bD5TizFO54yg7sWRyzTIkbE5qwJTV/XBc9ohLIgCQD2+Ojs3ONHe+i7xPRbQYv8IrLaIK+T1bypG2gNaY+V6+D0jqyk59+u4oxFOo0FDXLMVaNAgR17s/xfe8WeoYxIgZ1/Nddr9zTQ0thmLywpxhmnNdjkSm1/AnFFv4rL78T5UVFix1m83fVM6wuHyZ6wAnMrP5bbamhFrhqPMBvfp2uei0sOd5MG+AOR65XO6zXx6295k0Yx8bNoYUbCeEcJrc7WJDZFjZ2EhMwPQDXQ1gnZ4e6D2OvLWqgg7OgiP12ONpLRkKD3cGfLWDt22xR57kt6Asy/PJzoP+mwZhf15OTsmfUsFhRWRA2cQLpuIzxRx3oup74YSPWYufS+04X5nZ9S3bhc5rvTxYsLo9TjIzQH4MOl1LHZGPfffN0gOizbBVobz6XuPRURsdp+hddQVetrewGvwQr6cVscQEgyG0seLu/Xd2eK3sMj+SwU5bLxXl9NjamLmm4GzUrwj1t/GteAX21bYRqbgeb2oQamJOcuCwT1o++NSAC53EAjcYgSReiJpdhi7fBYDcOHbYKwRl3cFwHLrUbZtdWJb6EDSmtsVPZYuPGaIbKOB7nNBi3r8vmYpuRXKtBaVXpnvt5FMjQKLK/d1Lzl5rSCqrriApazwEXc0yZcAbhf/R4zMe77aafTeFL3fCtKu4D31CnJLS3JfqcmR5ja8YqniTrZVGZbyMizXxjXij3enIwcsZJYAjLvSmE5OJxj380Dc/ylg9wrtPWohM+O7uqv4+JFruDyT0s+XUY5bAai1ZAzeS46QsbQuACNmjsF/gfb7HvR7XwLXXWZ6pSNEDphJ5+Pjsfq1/A4EFfb2qO8VHdzJbWzwbp0CwHyf33jb7W34l8Fyr547veyukqBSs/RmM9SLKgIpz31evRssvtuyUA1QU99CQ/K7Svw/C0Adp/U0yO3sOL/MjxGOfz1IOjLT9+kNTsJoT+LenP/Y/nEXeqB65SoyThI3r2ERd5/YMNsTjRk8MWxndjKZeRlBjHWK5/3GO9hi6YUmN7csJeqF7K4NOd9L+yBsMmpYkVBQdteGuiRc978FcYSJjp7Go4TPuAtdlGiKTa8WD1OvtGGx119UW51EXD3tvuorE/m6UhQP0/QiMgKoNjJBlD2VW1sTP686SW21YcsZt6vhMyYDU0kbzmqYjZNca6h0TOiEoq+A5l4m8yyb4nszEpmZOa0GDmP3d9r+Dz/9Fq4kGFPy8yGT4dSsMMLgt+4Gao0alyNm7H/FkopHCgf06qQLHJkRhmr6IWzlFrw5bTu/xYWjTkgyjvZncO3NEFwXFw6EFC7O3F3ryL7AdQA0X/gennFFjW+ZhQUu41IAGPnuOO74K/DYXLIIi/6Tbg+fYnOWMwCnmy5n7NbfsPqrIuciGlJ1Xz6nm6zQNf31zkt6P72h8f/5Hmvueei2r6tzqLV0DAUTnI2oqmxR37nLmlWtARjkcImsNrWMrKh0+E2KAx7PW0maHcaBBUs432sx53st5sCCJey6Unw8Vuwkd1uCXOTlmK62UaAS1HxTKZLUTSF0PXuTaZW032e6OoeWw0fwR7gHmqwsIystHQlf1SKp7VLddm2rS1ya74jmZXF5p4sltCZbGi3SbSYd96Ig7QqazEzUN2+i9PLE+U8bvlhU2P+s0+J41sprB7O/5gYAPrlRF9K13gaXZZHM9A9BOHamSHv7nw7T9JNxgPa5ent22eotCY8aK3IbG62xUnMd2RoVQbuH47PscYPy0qR6rPPfRo+kjqS1FHBKKHkmq0GeVKvbNKfmmrG8dPhtKinusc5/GwmvL2Sm+z/sz7HlpcNvc1Odx86d4nDrPUpSb3v62afptjueGIz31EMIJ7QxVaWPF9Vrmf70NK/lsfyU6fHshuWYNT5/A9pQ0EOSZofpvC4AVdcNL3LMpAitiUpQo0GDz9oyr2iuN9R37lLjlzHIkRMTtoZBDpeQI+eqOoc3hk3AcutRNJmZxpZZagR7FYfzYG6GP1fV2bxmlUNM6A+MWbme5G/DkVtaGltiqemzZieBZuasv+9G6LE+BKy5w/UxjVB6eVLQoh7j9u9ite8eQi0ELhXkED5lNFVnicC78i+2Xwx5LJ+jONw2nWXiVe1zppJNZrn/bmUWFgyLOakzVur9NoHAAdGo028VaadqWY8Ng2YCsMRvI84nFKXyWhskhyX5Gzv8emgtsHEdx3C/sgKzbAGXA2kIuXl4Xo9lwN43mNJjPWs+8jKEhDLhQkEurt8WvaByAtzYFaAdDWQL+cg0JjSv+RHUd+5yR22NmUxhbCmlQmt8xFB13XD8Nx/W7XtorDxMwPXnMFUZXsSIMRWuTi5Ajoxqm0YTkCqCkNdT8P85F7oVbq/NrMx3H3TBamf5zQN4XgIHRPN5g/4IUafZ0vEdroUpmNJ9Pb1tb9K+zwK6NWyLqpcDBdeuG1tqiZDVC6Gx1T80Ofk2LqMLcM/PRMjKxjLDEXVFR9SWcppbFYaaO6yYhPeKQ6KcYWq21fG52qnv3GXrsVBmdjzMBv8ddKr2FpyMM7C60qFwcuL+z460tf6HbVkufDm7DwGLik8Wbj37b/zNtObG2+d64/bjmVKFZ/VqsMjt7Ej8LIRqQ+J0F5Xl1qM8fKU/GnP92G/rg0/iMVgUFd1wCtZajivvebFxcCtkB58cKqizbRyBO8T5wFS4unK9q/9j++0uF+iyujWCHJWgJj1EidevZa2w9Dyau+L+d6FB+bAOy79nC7n/LUCvstVocEJrsu2lRWiwosrm559WWB5RBAdwc3IO19U5pKvNCDFX8sXWblT9VZwzLYpDiDoNaJ+nvlvhx+Uv89lgdw70m8Em/534fzKc4PmOqGPF43242NEBb6UVi4PXMvqlsdhs1BrNDj8cRgBuvVdYn2PT/Qr4LUkSlbGS2N9e99kp8flz/Px/UkFHQyjSHwpHBzJ+dObPGj9zV5PPvIm9cd1a/P12a3A43e2nAxYsyKiGolcO6lJ6PPVqsFwdUJOzPedSK2ssvh89/WEx9FhfFtdbq8/TG5yUBW6crLsagK+PtiHg4HEjK9IzcgXXRzUk009D00axbPaaV/QwcjI0uZxTaevL1DDPA8ypECuu9L+HybSP5q5kd23IGh9tUt+TpjY/atyInYvt7aissGJ7tgMW566LOoGz06ZDxOdU5u1+Y7kdZMGRj+fzUngimdbWJleN+SEFFy7i++FFmt9/j79GTiep02Ia+/TE6YMgNKfEUZ+lyqpUkt7OI8TcgrnTI1j/YSjrI0OxTVbi0PIam4Knw4Ph7tS1b+B9XVwlMBSVcowtwWAkfBrM2Vra2jIWMjlDZ2yCGZCptmTuus6EtT1NS6ezKNDQyOofKiq0xQ4XnGiGf3rp6+gYbC2hZyEk2uIcmo2yii8FySnGkvHcJC4KJT58ASBnb441gUNjn1rE9qo6B9tko/3zlorroxpy5IO5AA/qdjyOg9yc+hYPxzna/qW2lBEgIg/LQ+/Ko9OYHzVioKjB8jA5VxQzhZ6H0Jqs6huBBg1fftUX51RxeiLktYKIH2/LVVUkR7+tj92fh6l03J6pw+uw1m83zduN1I3aTRXPrw7RM3ocs5Ys4GDt9YRN641zJ4Uo1qUpSLnEqKFjqfvVcb6uFEUtt2i+6PzoGmVaY+VSQQ6up8RnUpvH2EBT7edrDa1wP2BcPfpEY1+gS9a3lpnT0/aG7tigoRG6z2YyBSrBit05Now/2hv/vi9W9E+vb9R71bQ3ScVjz3ez5AoKyBaJFaoQdF/Q6M0DqZpX/EM+bYjWvb4vuwru34prRDB8ROkKqvRuHMkJOwdRJDY+zF2B4g2QR42YRxdDrLpuOP6It7bOoyR3t6WBhYyoPDnO34nTWAG49rIzia0jCNowCv912u9Gfe8eO5a8zCcfxZDWWkPgRiOLLAPMdh/jrcUTODhmJofr/kx7s3CEvPJvsIBWe3ySL7X6hfHHwG+pqLAiTqWi857RJD6Yxt3hu0l4/yquZymA1647MFb7Wf5yBsx49u8oKrrRLELb14iMAOTXb5fLqfgBy1W8/Ofop7a53SaHs81WcCpfzYLOnakS++IzLfVqsFhf1iZgXq+vwPcZI+4pPdbjIs+jwLcilPNkMYWLM/Zu2jLRgTuHUe2j48V6VxQh1VjdYGXZitMT6UPDGeQwn4cTx0amNuXQltpUPJqH2R7tqCencyjXwhXI/LI487K2n2YyBVPdTtD399fIaGws9S+O/4TD0Etbor9q0+E0DjurCxFVXTfcZBZDVHp58nWXtWgQGLB6zGOFysSCEF6bGROXcF+TR7XPE4vkNihNMwr0VDy+OcSH3V9hrnsklyfWE1UF8YLkFLw/TaFP9DuobORY3VBhNVKb89E35TWqrLosypClTKXmujqHigornKxzkJmZI6jyn/o7BVUr857LTgCWxL6M77XyWZlZdugkjk+5xO72CWN8nb2WA2bGAAAeG0lEQVRcKMilz/fv4hOrn4GRXqc1K7O1oYS9/aaT2zG02DayeiFMTY6mp+0NonK9ym2p7Ee5/7I/xxr8AIAsS4mQl/dYG1ndEOr8EEe9B+uSfb7j9bKUqBc0j/y3L7ImridU5FbQFvfL7B1Galc1wzvs4tTLK3TtQiL7EJmnYGTlfUZW/3w8moeSNDusyM9DzvdarJvy3GTUMJMxVgASvqlAJ5sMtmc7UGWFeKcyV5hxiaaW+dTbMuGx6aIqOxmpBTko7opzBltp2btdWzUvp2J5HJM/G8utR3Hce56UgRp+rLeCBXeqkvmmDQUXLxtbWqlQxybQI7Y/AH+EbEJexfup7RXBAfRasQvQGmr+76QbXKMhUPr5MH7KOjraxjF0+Hi9LjOgVw9L5T9vs2aEB/3s05g7L4JuLcYR+HGsLlSgCA4g6T0FdS00ROaZsbJ/R6D8GyxFcMxHXqd6kV3n3rNgWoNf6Gpzm3R1Di9vfJfAj0+XS1fe8xLfYwEZ3XKxkMk5M82CGuaRWMvMi1RFvanOo+IiS6ZqBpFd0Qx7EYRMrDcfgQXaz8VNVe53sSkHD1fXVbcVY6n+p/FDmNbY/PhMJ9xTzxpbTqlQ2NvjYXWH0WkvEzQ57rGZI6H9TtDu6Aiqvlv+r0dD4FNdvFW2L/cPIL55BFfVKr77rh2VL4rHU/Qssv2dsHjCJK7cDqH0m/6brsbXpXmB2KWJ7/qV29igXlFAV9sb1FzzHn479Rty1qvBojkVz6JZXen36XyCzeWc7RHBhMZNiFoYjteAJIa7/0ozq2z259gycsfbBBwWx8vA5sI9PrpRjy/coklouQxaPrntjJtN8X/nsOiMFeubGjI1+dg9WJQK0C1QpU2yLXqpxOVrGP3uRGz2aL9De8RDk1HDdLkpj/Jo7RVTJO39RjSwOI4GOe5fidf7cKNXCFsqzqfaryMJuFf4DFE4OhA/ryrT3Ray52BtIyo0LimpFQhEfN6ze2+EsXb0LMCMtosm4TlT/MaK+mc3eFAI/LOIZXx+f0CR1aYV/n4kjKrI8R6zsZaZc12dQ/uZk/DYd05UU7gfEj8jhPhqC6jx1xD8Pz+p9/eg3qexuCyLpNXl4QyYs5ledleZ7X4AvjigWy0WYOyGgQQvFE9cUnMqnvVRDfiiffQT29zX5JFcoOTolw1EOSq33nyEHjnj+X3Fwme2PZMv8HH7vticFV8/QdvX1puLK2suvpWaS0KVdsloEGh6qif2R08bW06pUDg50WjYMQCcTj2IaIfW5FZtW7LcZSS0iKD6X8MJnBIrukHDs8joH06Fg9dQJ10o9rh5nQwAXP80L/Z4eUZhb8+ITzYSbGbG6LSX8f1RPO+Hp+Gy8RTBrQYT98pywi3UfLlqKYMWjSPXWeDjLhvwMDtNU8t8wJzJ1+vzZ0QYlVaKrzieTKkkIeIlYjvOZ0+OPf5f5xqkpIBB5t2a/x7Fz03qsqpuF8zev8a5VDcaBSQTtT8Y/6WpVLkSTcEzko/KG07HlTT37qFbG+Lf9D3fDdUrV0VprDzE/PcoOnk873IJiQbVIqFfFI4OvFYhDjkyrp+rgD3njS2pVMjsbWlkp51tsH7ydA5O8KWWxRFCzJXEqVQE/zyegCmnRbtu0JNQBFZl3OT1XMhzZe0vrz52/Js+q2hrHQ3IcP65+EkB5RW5jQ0Z61x5w+4639wKIbWbCwWXxZm38m80WVn4z8rnaLiMUAuBuuZyjo+LeKzd5Ov1ifwyFOdN4py1d35aA+I7zeOuRkVE7+4Ip848+5dKgcEKhahv3sRs903YDQGkchPwJVK0VrPr4kgUG11o3H40ec4yjr1beNEFbh2B768C5og3dixh2uS95M9Qxz1E5ckJWnQbdWhNEKGXpeDiZVb1bofZuq10sgFvu+uMTnuFMzNq47DvHFXTxReOfR7Uief5oX51/PapODPk8dXiASbfqMf+OeE45YsrpHlpbG1iakUw8WoYR+bWx/GyOF/aT0KIjuWzKs9a6FfARsSDXdtqWu9e+IaJ+B8z3PUnrspmRkadfgun1dqbqcPserr9gYiz/L7Efws5chpYyGi78Sh/pAeT18zYikqHcCKWpYFVKFy/NwdbDovOjV5SNJmZJDcyp708XLdP1bgGFzqbEfzVBe3zqUBcL/vcDqFsG/4tYEXisEAco8WlX6KQ1rE9CJh0zKDePclgkZD4D2BxLZOjeVq3tBoZ6j7iTbr9L/PvOh7KfdEE7EO0nutZEfPxVFoxLb0mimsZou3Hfx23ztrlIAwdipQMFgmJ/wDqs4n/ckunGU2LhMRDJvs9Wq/ritF0SIgDmSA8v00kk8luQpnNl/MRBMG1jM4FlHn/QOqjQZD6qHdMvX8g9dEgSH3UO6beP3hKH0tksEhISEhISEhIGAO9luaXkJCQkJCQkDAEksEiISEhISEhUe6RDBYJCQkJCQmJco9ksEhISEhISEiUe0o0rdlcZiFYYmMoLUXIJYt8IU9WJid7QFn2DyCTjPSyzviW+qh/TL2P0r1oGKQ+6h9T7+N//V4skcFiiQ0NZS1eSIyyii8dt0XhZX6Lj2YPxG1B8StyHhH2vtB5SoM++lcS9ggby3xJVamP+sfU+yjdi4ZB6qP+MfU+/tfvxTILCcmtrTm3qh5ttp5giMNl2lhlY39RqmsoUb5QN3+Jby4cYdeVGM5PD3/2L0hISEhIlAllVun20vg6nHtNu2hXhiaHsZc6YHvmmlSKWaLcoGpZj0+XriDEXIlKUGN+p0w9rxJ64H6Phvw9ZxEKmRy1oKHmwtF4fVm8F1dCQkJcGNzDIjSuw40tQcSOXgjA1mx7mi55j1uNMyhIuWTo00sYiKpRllx5t5GxZegNuZ0dfl8kEG6hXULvYK4Z3t8eM7Iq/aNwdCBxSQPaxt5hW1o0lzfWIH2Y+D1JCkcH8v/w4cuvl6FBQCWo0SAw9M0dxpYmUQIULs4oqvmjqOZPVreGKAKrFh6zt4e9niy++A+aJnWNqLJ0KFxdufxRI6YmR7MtLZpdV2KQ7fPgyiTTeY4WR/rQcHZdiSHtg0Yge7FBoEE9LEo/H75au5gQM3NUgkDwj6MJXHYDr0RpxFOekVlYUBAeguLP48Uel9cKYnLF5bQTnrVkujiQW1uTvNyP37xWAhC0bzCBM/MQVGeNrEx/yGsFce1lZ9oO+YctbovRoEEDnAhfCeHQ7P4YHNYabll4Q6IIrEqVtanMdv/lsWM1LS+z9MN2+C5JQJ1+ywjqJJ4XpYc7Zz93J7H1kgdXJ6y868uqlHAKNrniduAGr7qdwlNpRUaQJS4HjCy4BMjrVKf3z7vpZac1oDWASoAt1X5lv7ctM6L6oNhf/PNW7HQY9TcqQc2JMRF03tIbddy5Uv8tgxksV95rxMRBGwkxM2dtphtf/r+98w6L4lr/+Gd3lqrSBSLSpItGIzYQNbEmlkRjv5bEGo39WhITf3pNMYnh2k0sMbFGjRq70RivJjFgxwYCIiiKijRRQdgy8/tjVSR2XVyGzOd5eB5mzuzu9zy7M/Od97znPeu64DcxptwvAV8eyOgfxrUaegL3PrhdsrLATbB5kZJKFbVbZU5FGs1K3/PNCRychJifb2ZVpsHwah0qTknnI6/l1LI07uuS3I7E3X5obkGP3rsZ43yKhv8+zOmV5tX6rHgsv8KMKg9+CGpireX4+3NofWAQFr/J37CobW250aYm6a1F+tSPZkrlOAAC9r5LtX8dM7O65yNppDcJrWdzb+C/r/05+tdKIzW0kHc+HINaJZpP4DOSNSicD8b+yFsVsogtUtNj92Cq7BLQ2ar4ZOL3vGZzkzHDiqi6x9xKyz6lYlgurKuBm306lTXXAZg+rys+c8pvVCXlS2NIfVuPKHw11qy88RKfb+wEgO/mAlTRx80p76mxbp+BOtvO3DJeCCqNhvwFxWHKs3OCscuXZ6Th71wZFcGCEXN4xcp4kX89vjP81xXLHYfwIgOAv5b70CLmFFPc/qR1r9HYr5BX3292bciH7lFASQOdpr/FB2kdWFVtJwABU+M595sZBJoAqVFttHYWpDfVENY4kc0+xuF1NSp0t9eCi2+6mHaEmVPmc3F25StEN4kCLGmT0AGLIZZ32zbt+Yk2+4ZRyVnNCMcE5FY+7KYXdKyQwwmtxLgx7xO44cDdtrFvduZIgyXmE1fKSBG1eKPS9wBkGG6B/vlCFiY3LBpfb8bX+JWmNik4CQKhi/+N97wDj3+hjNC4u5Ha34/AVmeZ5vMzvpojt0OYloiI9KiUTo/eswFY0+ElVgZXNa/gp2Tfy2s5VCQx+SEXwNyQii9YUelxdmo94kONyeCBW4YQtPYw5WU50PFD1hBmBW0S3sbwlRtWe04g6UrOGNT5uuOg1pInSlRM15pJ6bMh2NmR1/0GXppis5IrFtLn7cEUuljzybxFZlT3/KRNjqBl+0OMrDwXL40NalSIt3+di/O82HD5FbYGbwKg/uGeuJJgTrnPjODmip97Jo5qa9okdEDd/EKJSHyDI/+iUowNUaMXoEbNRxl1cV4UYza9T4tfVDzBFYYSPO0ctpeL74Uab0921F0AWCHF2ptPYClyppcVYVbG/9+YM54qZ54vcGFSw6LxrEqNn8/Tu9IVwJbQxUPx+eQQklg+BoLyejUku6aKRZ0XEG699fZey0e+pp51Gj+83hHLHYdKX6CJEFRqeuwcRCAHH3qMGhWSzCfRCHZ27On+NWDDX4UWBH9zHVGvR12pEmoXJ5AkWSeGC4hYqAQElYheBahLfmFqW1vSxurx1VgTEdsDJ5mNoZ+OCiKpwfwS+yL3vY/voeNUqOZzN4FarmzvN42qGhvuRI/6pr3Kgd2h+C27Cnk3yH3DEz43Huv2mYVsjbYu0IMAu9OIiFze5YkHF0q0uw/M5c09+4i0LkQETtSRV08N1/LwH73/vhmxFzp54ibY0CW5DZ6flc8RCI2D8SEoVV+I7ZXn/95MGltL7ePFVFfjRW9+njd+C84j6eU9cflan3CSZzakXVwu66ZGcarXbMKtix56/Myc6iW2z+hcZGVWAAySyKOufjk1QERCJa/rxv1sqoibYEOBpGXg+vcQTyRws2tDvHbr2bjvZ37ct4YzS8IQKr/QwpkmY/La7ugkA5uCNrLj+2+52b52ifYr/WpztOESYooEnD6zNpPKZ+ffjX4tsR24aQh+A5IBMKRdJPinoeaQZTL6jBpDzcXDqLl4GG+93JKM8Ov4TIzBkHQWQ8ZVMiN1qFHxVXYo6pR0c8t9ZtR/xrIjphYA0wcsQhUWWqK9zs7L9LU/B0D1NcNftDyTowoL5dLYCI6MmUOG4RYpm/0e/yIZIjg60reGMRLW/Xg/HJY/f1TMpBEW+8gMBJXRAy0/1wD79OT7D1ILCI72XOgXTL6PAfs4Addvyq67XP5pFN6aO1EUqxJtC6/5syc7kGufeN3dlxtoxaiPi2eXjF3/Dr7IJ3z5JOgryfvJFYzRwA+8jVGyqZnhVBsfw41uDfnmy1mEWhpPC1uVJYktF9J4aTfs22SaU+4z4bcqh77NmrPYexcAriNSKEgNRToSR+oX4azoPguA/wwegMV+eU3hvtYnnGYVpnNvhDN47CnEggIAJL0eTb68Q4C2Gw7gvcH4/9/POMHfl4mNtiIisfhQJIHZ8vr+/k7wgmv83saWpjYF6NasZ043Yw5g22X7GOyQgojRrASMl++QreDsxJk5XmxrNO/2PUVNgaRCEoxthuwcc0s0KXktgxjnbEwcy71aCVM89pnMsGhecqeLZ6zx6RzI3+OKPcWGReNRhaxm3lwLgvi+8wBjR1oFdUC1yBJJV3bHzw8XCazJacAfy+ohNctFd9gRp9MG7E5mYUg6iwXFNzONV3FNi035LgQsuCSr4niCm+sTHXdap8Nj/nHkl7NvHDuO/8idcCsD869V42QXXwTnPMLHH7xrVjolt2Wt/xbUqBnut4cfvSPQn7/wmHcuWxjiEsl5qzJtVnVie8h6Vvlt581pHUnNfJmfGszkisGOmsv6UW3PEVndBAQHe7Sdcgm0MJqVmCKBDycMptKtkrly9meMv9MQCwtzyCxVkqY48K7dJY5oRYLm3pLV9/cgDHGJTJg2gP2T59LWtpC/vjvJFNdYABbm+bCldxP8j+yXdT8v9A3mVNNZ3GuyvTWWHB45ix/e9WFLy1ro0y+ZT6AJEVycyQk2Bi9OaA0Ez7xpknuFyYaE9J6V8bbMAiAqJwivFSl32zJGRPDqzkRmTZnLhM7r2VZQnLT5XcCPqAN8TCXD5PSeOJbJ/QeQWFeH2+xo3DucxvOzaCqsP4Ah6ex9x2u6XAUgw1DEf5b1RJ/6wpe2eC4u9vK/b9+dEOblMca/Ka/9TOf9g2Q79TcnwoOkdsbch5m73sCQnEpG5yC+dDcO3QWvHYq2RQ515o4EoGvFq+Q29DCb3ufBkJmJ1WCBDTeNRnRz8AbiGv9AiKWa97f1xXdCTJl+WHgQUtWXOFi3eA72tLQ3qLRmP0glb2eOS2NYml0+i3LNarAKgGVZkUixcWZWYxrc9l6lS3IbdJKBya7GiQwTr4axuboz0hH599Fz42U23HRldm4wwVuG0qFRR+pHjWTDTVf62p8jYazX499EJtyM9OOzXis4oTXQZfMIxFOmSQg3WYQlbRx0qHANgAa2Z/ndqQ5cvkLq6pc5FjmTAklH7/YDEY/FY3i1Dm1XfgdAt1N9cYxPMpUMk/M0425J39dla8gcQEPLpePwkWEiVWT3owgqNQMjf+fon56sqfYrcPRuqfM79Gy8hDbIv3Ccy1EVGm9Ptn5sTL7tlNyWgHFHkXRaqn51gNoM59iwOeaW+VwYklNZ9nYrKm9eRxNrLRYqAd8d7xE4Ul5TmO+Q3tqpxLbU9+GJ76eG1oD18sohexxi01eobrEPPVYcnvUK9sjze7yPB1RBPdnND7j/wVCOGJJT+SHIG4BADqIH3GekMftSNzpNn0dC13k0++N9bDfIf1at65gU3qyQywktBC65YbLImMkMiza9ePnpxtZ6flpynl9/b8ihRtNpHPsu1sudqHjMeGJp8oqI02kJtXj0DBu5oLa1Ja/9y/zWPAontZqQDcMIXn5VlkXyfv+5DoZh0YxzjifZLpa2iV1ITPAg6PuCu8dMW/cdg+J7lRjykxMZEcWnj/PmeHB0wOV2IbzEP33x0V0xNkoihaG3zCHR5FwPdsBZXYCIGp0E9rHyPfeadD/yxMde/lBXikrMw6URWqpqbPgoo67s6uY8DMHfl3OfWxPrv4p7A/+6b7WoX9xCwWbB8dAVNuc70q5CNlfD1PhsMLei50Pj7saR5KpQDfIlS1SiWPYMS8DSG9CleHtOlWjoEY3/5lGEzM3DELcfta0tiV/WxG2/itW59fnU9RhZ55xwNJUIM5E8qRanes8GrGh2shsBww7I0qwAVP0iGn/vwdjHafDYmo6Uep5A0kv84AolgZzrtsi1ckDDOsaIXpr+FoglT6WXYoozjrL7NSThtbnMu+aHw4ls2X6n+Z0aMD1qLiGW8iq49SS8c64F0vUbD2zTtQhjWo0VAPzvf7WpVg6S36VGtdlcdy5gw5atDfEuB30CSB7gTlz4XEBNzcXDaNz6BN9U/YPtwRtp3mEINhsfXmJB7uhTznG6sArtKsi/EjPAtSa+fBaxlpgigf7rhlDtuOl+oyYzLKrEVLqltLo9hFBM1V9VGOISEezsaBF9ka2O+8CYAE7IH30J+jBelombdyjo2IAzfb5FJ6mZnhOMw0CtrJJsH0TgYOPF4UH9EJydsFbJ9dZdkpZ7RxBw/Sga52LLfKGHnqCUAPSOtrw98n/Mu+bHr13qYzhddoctH4XGowpzo2aXS7MCcHR3MN7Z918QNZ5VGT1/Bc1tCogpEghYnCFbw3kvWjuLEoXyygN52/2JCljKkSJ4b85wvGdEs9uhAWLVvQDYXsiXdbLtPwnNS+60+PhPula8Sqv4twlYZNrzzmSGRSwo4MrMmjC75P6FM2Zw6AsvrNU6OlXIBeCkVkf3QwOoNk2PeOPBT0dyQHB2Yv3s6egka77KDiWmrT/6CxfNLatUyWofRIjFLnPLMAlV3XNRaTSIGZl0O/s6a/x2kPDad+xuaEtdqxwGpLyNdrCdbM2KrkUYwseXCbFUExnbE8fPbZi88gfqW5Xfy7/Kyoqkr2sjVTDcNSuThg/E6oz881hUFpZYj7uEGhVp+gK8J8k/unL1/QgO1pqDiEjkxBG4/xCN4OzEO02NKxtOulqvXCTcPoqCjg0Y6jgDkPdsNo27G4ljfNjoso1uZ1/HakIlDGdOmfYzTPlmFbceo0aNYczts4BXrY1jx4EW1gRaXL2dtAnhx7pReVA+3uknZe2aBTs78lY6YK+2ZHpOMOsWNcPtgvySbJ8Wl4PZXNTLO6/jQKoP+MDuGusIXDQIu+NWLPKcxZ2x8+Y2BYSsHoPf2AMgZZhT6nNhMSGD7UHbWXLdg8ofqLjUvAL1rSRERDSFcj77ilndeyZDEkZidc1A7/9uwUEo4M0KxefhxDMdsdkmf7MCoH3tZXYGLUAEXl8+Dh+ZDwepwkKZMnoJAGGzR+K56uh90fZtqyPwoPxeV1VWVlxsLVFRbazx5bFXXjP27iX3VV9Gt9lGTJFA/lh3OHzS5J9hUsMiFRXhNSWaqOkRDBtdE6+maWwP3gzATbGQNsNG4PTLcfRFD68UKwcEZycyl7qwr6ZxauWqhS1xK8eLO95Lvp8DDmp5Dy+4brOC14z/J7VaCK3g3kS/Rse6E/jFGQySfG/qOf3C2RDwNae1atZ0boYYl0Bht3BERGbmVJfVWiyPItRSwx9fz3tou9VUhxeopnRJ7VI8i8ZvmTyT+u8ls64db9jeQCcZqJguIhYWAnB+UDCbXHYxNasWXt8lyL6fYDQmhvrV0eTdQjyRcHdf4sxaJLSfhwjU+qsf3r89eUJ5WaOoZw6D7M/xS0El1AW6Ukn1KJXVmsUbN/D6xHgDv3fqqw0HZR1VAZDCaxG5YD9jnHdyUV/EwP4jcfvtn2FWAKy3HOTMbHmHLu1/jqVmyDA2v2NcXRsgVivS7ZehBKwowulgHAaZLynx1qg9VBasyBEN5IU6kPlRbbY1Mq6G+3v3OkCiuSU+M78kVGdGlUefcx3PtEPsokfIlXck90G8frojFinyqu/0IArcVIiI5IhaMprpueUSgVXLTI7VNg4RLTkUIfsKvncQqrizafVCjmvhvaiR2GSJ+I+KZ5OX0WzvuVWRasPlm2d1vUdD2nn9yVfZoUQ380DMKp2FOEvFsJRXhMqVuTKhkDHOp/g0sw5/TA7H5rfym71eXpGKivCeFMPwSY1K7L+z2GN5usGFWFiwZ/qdOjKWzM4NxhAnX7MCEDTiHBHLexD9yqr72m6KRTT7YixVtl/EkHnZDOpKjz71o40rNke5Ienluyjn33ETbEho/S20Nm5nGIpoFzWekJXlI7pyL7UsYf9HxiUx1Kg5oZXounEEQfOzMGTIs0yEEBLAjq9ncFprycR+AxGySm8RVcWwPCGCsxNXFjuxsfZiBqa1IetdV2wS/5lmZUlWY3NLUHgMuz5uQs+5h2+v9guRsT3R7nbBY8lpINe84p4TQ24uTu1yaUfYA9tdiZb9TL2/I/j7MsllPSISN6tocHr8S8o8vjNPUd1xOAldjVGGSVfrsW11BN5rL+GWEl2uzIp0/QbzcoMY6lj8sBC4ZQheW8F/635591WtxlZlyZCo4bjuLd3RBsWwPAGCnR2ZS12YWX0NfZP+haZFGnDd3LLMxtl6hfhywtwyFB6B9ZaDDN4SeXfbiSQgSd4Xxn8wqnvqBc3/v1lM2t1Jdmtb/R3D9ev4j95Pu9HFxtOjHJpNAEN2Djtr2LGTenf33Ynoyh1DXCLtPMJwfQHJ0fLOnnwBqK2tyfvJhdU1v+eT/v1umxUFBQWFF4c+5RyTM2ux55Y17ywYhSG9fA13KSg8CUqE5TFot7oS6ZzCiIiuCOmlNzanoKCg8CgO1RY4RCgeRJerPCsFhSdFJT3F1E2VSpUJvKj0dG9Jkiq/oM8CXnj/QOljqaD00eSU9/6B0sdSQemjySnv/YNH9PGpDIuCgoKCgoKCgjlQclgUFBQUFBQUyjyKYVFQUFBQUFAo8yiGRUFBQUFBQaHMoxgWBQUFBQUFhTKPYlgUFBQUFBQUyjyKYVFQUFBQUFAo8yiGRUFBQUFBQaHMoxgWBQUFBQUFhTKPYlgUFBQUFBQUyjz/D9royyhUAm8fAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x432 with 100 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "(X_train, y_train), (X_test, y_test) = mnist.load_data()\n",
    "print(\"First 100 train images:\")\n",
    "for k in range(100):\n",
    "    plt.subplot(10, 10, k+1)      \n",
    "    plt.gca().axes.get_yaxis().set_visible(False)\n",
    "    plt.gca().axes.get_xaxis().set_visible(False)\n",
    "    plt.imshow(X_train[k])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Step 4: Preprocess the input data by reshaping it, converting it to float, and normalizing it [0-1]."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(60000, 784) train samples\n",
      "(10000, 784) test samples\n"
     ]
    }
   ],
   "source": [
    "# reshape\n",
    "X_train = X_train.reshape(60000, 784)\n",
    "X_test = X_test.reshape(10000, 784)\n",
    "X_train = X_train.astype('float32')\n",
    "X_test = X_test.astype('float32')\n",
    "\n",
    "# normalize \n",
    "X_train /= 255\n",
    "X_test /= 255\n",
    "\n",
    "print(X_train.shape, 'train samples')\n",
    "print(X_test.shape, 'test samples')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Step 5: Convert class vectors to binary class matrices; One-Hot-Encoding (OHE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_train = np_utils.to_categorical(y_train, NB_CLASSES)\n",
    "Y_test = np_utils.to_categorical(y_test, NB_CLASSES)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Step 6: Create the model with 3 layers: Input:784 ==> Hidden:128 w/ dropout ==> Hidden:128 w/ dropout ==> Output:10 (with Softmax activation)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "module 'tensorflow' has no attribute 'get_default_graph'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m-----------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                        Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-7-a8d4ebd6b6e3>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mmodel\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mSequential\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mDense\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mN_HIDDEN\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput_shape\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m784\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mActivation\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'relu'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mDropout\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mDROPOUT\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mDense\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mN_HIDDEN\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/.conda/envs/keras-gpu/lib/python3.6/site-packages/keras/engine/sequential.py\u001b[0m in \u001b[0;36m__init__\u001b[0;34m(self, layers, name)\u001b[0m\n\u001b[1;32m     85\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     86\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m__init__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlayers\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 87\u001b[0;31m         \u001b[0msuper\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mSequential\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__init__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     88\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_build_input_shape\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     89\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/.conda/envs/keras-gpu/lib/python3.6/site-packages/keras/legacy/interfaces.py\u001b[0m in \u001b[0;36mwrapper\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m     89\u001b[0m                 warnings.warn('Update your `' + object_name + '` call to the ' +\n\u001b[1;32m     90\u001b[0m                               'Keras 2 API: ' + signature, stacklevel=2)\n\u001b[0;32m---> 91\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mfunc\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     92\u001b[0m         \u001b[0mwrapper\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_original_function\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfunc\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     93\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mwrapper\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/.conda/envs/keras-gpu/lib/python3.6/site-packages/keras/engine/network.py\u001b[0m in \u001b[0;36m__init__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m     94\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     95\u001b[0m             \u001b[0;31m# Subclassed network\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 96\u001b[0;31m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_init_subclassed_network\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     97\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     98\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_base_init\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/.conda/envs/keras-gpu/lib/python3.6/site-packages/keras/engine/network.py\u001b[0m in \u001b[0;36m_init_subclassed_network\u001b[0;34m(self, name)\u001b[0m\n\u001b[1;32m    292\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    293\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_init_subclassed_network\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 294\u001b[0;31m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_base_init\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    295\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_is_graph_network\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    296\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_expects_training_arg\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mhas_arg\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcall\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'training'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/.conda/envs/keras-gpu/lib/python3.6/site-packages/keras/engine/network.py\u001b[0m in \u001b[0;36m_base_init\u001b[0;34m(self, name)\u001b[0m\n\u001b[1;32m    107\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    108\u001b[0m             \u001b[0mprefix\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__class__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__name__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlower\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 109\u001b[0;31m             \u001b[0mname\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mprefix\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;34m'_'\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mK\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_uid\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mprefix\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    110\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mname\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    111\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/.conda/envs/keras-gpu/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py\u001b[0m in \u001b[0;36mget_uid\u001b[0;34m(prefix)\u001b[0m\n\u001b[1;32m     72\u001b[0m     \"\"\"\n\u001b[1;32m     73\u001b[0m     \u001b[0;32mglobal\u001b[0m \u001b[0m_GRAPH_UID_DICTS\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 74\u001b[0;31m     \u001b[0mgraph\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_default_graph\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     75\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mgraph\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32min\u001b[0m \u001b[0m_GRAPH_UID_DICTS\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     76\u001b[0m         \u001b[0m_GRAPH_UID_DICTS\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mgraph\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdefaultdict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mint\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mAttributeError\u001b[0m: module 'tensorflow' has no attribute 'get_default_graph'"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(N_HIDDEN, input_shape=(784,)))\n",
    "model.add(Activation('relu'))\n",
    "model.add(Dropout(DROPOUT))\n",
    "model.add(Dense(N_HIDDEN))\n",
    "model.add(Activation('relu'))\n",
    "model.add(Dropout(DROPOUT))\n",
    "model.add(Dense(NB_CLASSES))\n",
    "model.add(Activation('softmax'))\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Step 7: Compile the model with categorical_crossentropy loss function, Adam optimizer, and accuracy metric"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer=OPTIMIZER,\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Step 8: Perform the training with 128 batch size, 250 epochs, and 20 % of the train data used for validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "history = model.fit(X_train, Y_train,\n",
    "                    batch_size=BATCH_SIZE, epochs=NB_EPOCH,\n",
    "                    verbose=VERBOSE, validation_split=VALIDATION_SPLIT)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Step 9: Evaluate the model on the test dataset (10,000 images)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "score = model.evaluate(X_test, Y_test, verbose=VERBOSE)\n",
    "print(\"\\nTest score:\", score[0])\n",
    "print('Test accuracy:', score[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Step 10: Plot the accuracy from history"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(history.history.keys())\n",
    "plt.plot(history.history['accuracy'])\n",
    "plt.plot(history.history['val_accuracy'])\n",
    "plt.title('model accuracy')\n",
    "plt.ylabel('accuracy')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'test'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Step 11: Plot the loss from history"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('model loss')\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'test'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[Optional] Step 12: Save the model (serialized) to JSON"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_json = model.to_json()\n",
    "with open(\"model.json\", \"w\") as json_file:\n",
    "    json_file.write(model_json)\n",
    "%ls"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[Optional] Step 13: Save the model weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save_weights(\"model.h5\")\n",
    "%ls"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[Optional] Step 14: Load the saved model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "json_file = open('model.json', 'r')\n",
    "loaded_model_json = json_file.read()\n",
    "json_file.close()\n",
    "loaded_model = model_from_json(loaded_model_json)\n",
    "loaded_model.load_weights(\"model.h5\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[Optional] Step 15: Compile and evaluate loaded model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "loaded_model.compile(loss='categorical_crossentropy',\n",
    "              optimizer=OPTIMIZER,\n",
    "              metrics=['accuracy'])\n",
    "score = loaded_model.evaluate(X_test, Y_test, verbose=VERBOSE)\n",
    "print(\"\\nTest score:\", score[0])\n",
    "print('Test accuracy:', score[1])"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {},
   "source": [
    "- mkc"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}