01_mnist_light_ryazhskikh.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "01_mnist_light_ryazhskikh.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [
        "6ho0LqzRy0ND",
        "xWgamMlWvjna",
        "xHzUpI2AQIIH",
        "3djmzlwKQWYg",
        "jYl__UuYQo15",
        "GFfFilTHvs-o",
        "o2ALLlVvtCZW",
        "-xSaT3b6tOj7"
      ],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/darja/NeuralUni/blob/master/01_mnist_light_ryazhskikh.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "metadata": {
        "id": "qiYEYNhreXuj",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Распознавание рукописных цифр из набора данных MNIST – Light"
      ]
    },
    {
      "metadata": {
        "id": "uZOV5o91eFvc",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.datasets import mnist\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Dense\n",
        "from tensorflow.keras import utils\n",
        "from tensorflow.keras.preprocessing import image\n",
        "from google.colab import files\n",
        "import math\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from scipy.misc import toimage\n",
        "%matplotlib inline "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "Ff_DgIyAejTE",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Подготовка данных для обучения сети"
      ]
    },
    {
      "metadata": {
        "id": "OdrY5RsOekut",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "0c326a2c-d99e-4aeb-e505-7d903c2eff93"
      },
      "cell_type": "code",
      "source": [
        "# В Keras встроены средства работы с популярными наборами данных\n",
        "(x_train_org, y_train_org), (x_test_org, y_test_org) = mnist.load_data()\n",
        "\n",
        "x_train = x_train_org.reshape(60000, 784)\n",
        "x_test = x_test_org.reshape(10000, 784)\n",
        "\n",
        "# Векторизованные операции\n",
        "# Применяются к каждому элементу массива отдельно\n",
        "x_train = x_train.astype('float32')\n",
        "x_train = x_train / 255 \n",
        "\n",
        "# Преобразуем метки в формат one hot encoding\n",
        "y_train = utils.to_categorical(y_train_org, 10)"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n",
            "11493376/11490434 [==============================] - 0s 0us/step\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "TrR8BAzAV1mM",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# print(x_train[5])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "ZCO5TovheqfC",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Подготовка нейронной сети"
      ]
    },
    {
      "metadata": {
        "id": "SrL0zD3YiRVa",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Утилиты"
      ]
    },
    {
      "metadata": {
        "id": "8cIV_GMjfBCr",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Структура для хранения параметров нейронной сети\n",
        "class FitConfig:\n",
        "  def __init__(self, dense, validation_split = 0.1, epochs = 80, batch_size = 100):\n",
        "    self.validation_split = validation_split\n",
        "    self.dense = dense\n",
        "    self.epochs = epochs\n",
        "    self.batch_size = batch_size\n",
        "    \n",
        "  def __str__(self):\n",
        "    return \"Dense {}, validation split {}, epochs {}, batch_size {}\".format(self.dense, self.validation_split, self.epochs, self.batch_size)\n",
        "  \n",
        "  def csv(self):\n",
        "    return \"{};{};{};{}\".format(self.dense, self.epochs, self.validation_split, self.batch_size)\n",
        "  \n",
        "  def plot_title(self):\n",
        "    return \"Dense {}\".format(self.dense)\n",
        "  \n",
        "# Структура для хранения обученной сети и результатов обучения\n",
        "class FitResult:\n",
        "  def __init__(self, _config, _model, _history):\n",
        "    self.config = _config\n",
        "    self.model = _model\n",
        "    self.history = _history\n",
        "        \n",
        "  \n",
        "# Функция для создания и тренировки нейронной сети с заданными параметрами\n",
        "def fit(config):\n",
        "  model = Sequential()\n",
        "\n",
        "  # Уровни сети\n",
        "  input_dimen = 784\n",
        "  for dense in config.dense:\n",
        "    model.add(Dense(dense, input_dim=input_dimen, activation=\"relu\"))\n",
        "    input_dimen = dense\n",
        "\n",
        "  model.add(Dense(10, activation=\"softmax\"))\n",
        "\n",
        "  # Компиляция\n",
        "  model.compile(loss=\"categorical_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n",
        "  print(model.summary())\n",
        "\n",
        "  # Обучение\n",
        "  history = model.fit(x_train, y_train, batch_size=config.batch_size, epochs=config.epochs, validation_split=config.validation_split, verbose=1)\n",
        "  \n",
        "  return FitResult(config, model, history)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "T48XnP7Bh6xo",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Обучение нейронных сетей с различными параметрами"
      ]
    },
    {
      "metadata": {
        "id": "Mp5nIKE4h5zx",
        "colab_type": "code",
        "outputId": "9e3cf056-9752-4e6d-b164-79bdfe6faf72",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 30229
        }
      },
      "cell_type": "code",
      "source": [
        "configs = [\n",
        "#     FitConfig([1]),\n",
        "#     FitConfig([1, 1, 1]),\n",
        "#     FitConfig([10]),\n",
        "#     FitConfig([10, 10]),\n",
        "#     FitConfig([10, 10, 10]),\n",
        "    FitConfig([100]),\n",
        "    FitConfig([100, 200]),\n",
        "    FitConfig([200, 100]),\n",
        "    FitConfig([400]),\n",
        "    FitConfig([400, 100]),\n",
        "    FitConfig([400, 400]),\n",
        "    FitConfig([800, 800]),\n",
        "    FitConfig([800]),\n",
        "    FitConfig([1000]),\n",
        "    FitConfig([2000]),\n",
        "]\n",
        "\n",
        "i = 1\n",
        "n = len(configs)\n",
        "\n",
        "results = []\n",
        "\n",
        "for config in configs:\n",
        "  print(\"++++++++++++++++++++++++++++ {} of {} ++++++++++++++++++++++++++++++\".format(i, n))\n",
        "  print(config)\n",
        "  print(\"++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++\")\n",
        "\n",
        "  i += 1\n",
        "  \n",
        "  res = fit(config)\n",
        "  results.append(res)\n",
        "  "
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "++++++++++++++++++++++++++++ 1 of 10 ++++++++++++++++++++++++++++++\n",
            "Dense [100], validation split 0.1, epochs 80, batch_size 100\n",
            "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/resource_variable_ops.py:435: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Colocations handled automatically by placer.\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense (Dense)                (None, 100)               78500     \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 10)                1010      \n",
            "=================================================================\n",
            "Total params: 79,510\n",
            "Trainable params: 79,510\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n",
            "Train on 54000 samples, validate on 6000 samples\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Use tf.cast instead.\n",
            "Epoch 1/80\n",
            "54000/54000 [==============================] - 3s 59us/sample - loss: 0.3722 - acc: 0.8965 - val_loss: 0.1654 - val_acc: 0.9572\n",
            "Epoch 2/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.1727 - acc: 0.9506 - val_loss: 0.1241 - val_acc: 0.9685\n",
            "Epoch 3/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.1246 - acc: 0.9644 - val_loss: 0.1042 - val_acc: 0.9705\n",
            "Epoch 4/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0980 - acc: 0.9716 - val_loss: 0.0961 - val_acc: 0.9730\n",
            "Epoch 5/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0802 - acc: 0.9762 - val_loss: 0.0913 - val_acc: 0.9733\n",
            "Epoch 6/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 0.0668 - acc: 0.9806 - val_loss: 0.0843 - val_acc: 0.9752\n",
            "Epoch 7/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0565 - acc: 0.9838 - val_loss: 0.0811 - val_acc: 0.9758\n",
            "Epoch 8/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0490 - acc: 0.9861 - val_loss: 0.0777 - val_acc: 0.9778\n",
            "Epoch 9/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0413 - acc: 0.9884 - val_loss: 0.0781 - val_acc: 0.9768\n",
            "Epoch 10/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 0.0365 - acc: 0.9893 - val_loss: 0.0752 - val_acc: 0.9793\n",
            "Epoch 11/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0313 - acc: 0.9913 - val_loss: 0.0776 - val_acc: 0.9763\n",
            "Epoch 12/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 0.0270 - acc: 0.9929 - val_loss: 0.0744 - val_acc: 0.9792\n",
            "Epoch 13/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 0.0227 - acc: 0.9942 - val_loss: 0.0777 - val_acc: 0.9792\n",
            "Epoch 14/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 0.0203 - acc: 0.9948 - val_loss: 0.0778 - val_acc: 0.9788\n",
            "Epoch 15/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0172 - acc: 0.9959 - val_loss: 0.0825 - val_acc: 0.9775\n",
            "Epoch 16/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0154 - acc: 0.9963 - val_loss: 0.0837 - val_acc: 0.9783\n",
            "Epoch 17/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0127 - acc: 0.9971 - val_loss: 0.0848 - val_acc: 0.9792\n",
            "Epoch 18/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 0.0118 - acc: 0.9972 - val_loss: 0.0852 - val_acc: 0.9773\n",
            "Epoch 19/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0099 - acc: 0.9978 - val_loss: 0.0816 - val_acc: 0.9792\n",
            "Epoch 20/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0082 - acc: 0.9983 - val_loss: 0.0886 - val_acc: 0.9797\n",
            "Epoch 21/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0076 - acc: 0.9985 - val_loss: 0.0916 - val_acc: 0.9770\n",
            "Epoch 22/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0057 - acc: 0.9993 - val_loss: 0.0888 - val_acc: 0.9788\n",
            "Epoch 23/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 0.0071 - acc: 0.9984 - val_loss: 0.0927 - val_acc: 0.9785\n",
            "Epoch 24/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0053 - acc: 0.9992 - val_loss: 0.1057 - val_acc: 0.9750\n",
            "Epoch 25/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0044 - acc: 0.9993 - val_loss: 0.0884 - val_acc: 0.9807\n",
            "Epoch 26/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 0.0033 - acc: 0.9996 - val_loss: 0.0937 - val_acc: 0.9783\n",
            "Epoch 27/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0042 - acc: 0.9992 - val_loss: 0.1056 - val_acc: 0.9765\n",
            "Epoch 28/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0064 - acc: 0.9985 - val_loss: 0.0932 - val_acc: 0.9802\n",
            "Epoch 29/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 0.0038 - acc: 0.9991 - val_loss: 0.1043 - val_acc: 0.9782\n",
            "Epoch 30/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0051 - acc: 0.9986 - val_loss: 0.1037 - val_acc: 0.9788\n",
            "Epoch 31/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0016 - acc: 0.9999 - val_loss: 0.0969 - val_acc: 0.9797\n",
            "Epoch 32/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 9.8188e-04 - acc: 1.0000 - val_loss: 0.0980 - val_acc: 0.9805\n",
            "Epoch 33/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 0.0077 - acc: 0.9976 - val_loss: 0.1122 - val_acc: 0.9768\n",
            "Epoch 34/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0045 - acc: 0.9987 - val_loss: 0.1110 - val_acc: 0.9777\n",
            "Epoch 35/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0016 - acc: 0.9998 - val_loss: 0.0986 - val_acc: 0.9798\n",
            "Epoch 36/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 6.9037e-04 - acc: 1.0000 - val_loss: 0.0983 - val_acc: 0.9810\n",
            "Epoch 37/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 5.3118e-04 - acc: 1.0000 - val_loss: 0.0983 - val_acc: 0.9812\n",
            "Epoch 38/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 8.3615e-04 - acc: 0.9999 - val_loss: 0.1093 - val_acc: 0.9775\n",
            "Epoch 39/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 0.0083 - acc: 0.9973 - val_loss: 0.1029 - val_acc: 0.9792\n",
            "Epoch 40/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 0.0021 - acc: 0.9996 - val_loss: 0.1080 - val_acc: 0.9795\n",
            "Epoch 41/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 6.5896e-04 - acc: 1.0000 - val_loss: 0.1039 - val_acc: 0.9805\n",
            "Epoch 42/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 3.8497e-04 - acc: 1.0000 - val_loss: 0.1040 - val_acc: 0.9800\n",
            "Epoch 43/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 3.3067e-04 - acc: 1.0000 - val_loss: 0.1038 - val_acc: 0.9807\n",
            "Epoch 44/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 2.9269e-04 - acc: 1.0000 - val_loss: 0.1064 - val_acc: 0.9798\n",
            "Epoch 45/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 2.6970e-04 - acc: 1.0000 - val_loss: 0.1043 - val_acc: 0.9793\n",
            "Epoch 46/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 0.0091 - acc: 0.9972 - val_loss: 0.1165 - val_acc: 0.9783\n",
            "Epoch 47/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0066 - acc: 0.9977 - val_loss: 0.1197 - val_acc: 0.9773\n",
            "Epoch 48/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 0.0012 - acc: 0.9998 - val_loss: 0.1179 - val_acc: 0.9797\n",
            "Epoch 49/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 3.4774e-04 - acc: 1.0000 - val_loss: 0.1135 - val_acc: 0.9793\n",
            "Epoch 50/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 2.4562e-04 - acc: 1.0000 - val_loss: 0.1154 - val_acc: 0.9793\n",
            "Epoch 51/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 2.1263e-04 - acc: 1.0000 - val_loss: 0.1145 - val_acc: 0.9792\n",
            "Epoch 52/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 1.9340e-04 - acc: 1.0000 - val_loss: 0.1141 - val_acc: 0.9795\n",
            "Epoch 53/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 1.6923e-04 - acc: 1.0000 - val_loss: 0.1149 - val_acc: 0.9793\n",
            "Epoch 54/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 1.5282e-04 - acc: 1.0000 - val_loss: 0.1158 - val_acc: 0.9790\n",
            "Epoch 55/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 1.4135e-04 - acc: 1.0000 - val_loss: 0.1169 - val_acc: 0.9795\n",
            "Epoch 56/80\n",
            "54000/54000 [==============================] - 3s 47us/sample - loss: 0.0077 - acc: 0.9978 - val_loss: 0.1540 - val_acc: 0.9730\n",
            "Epoch 57/80\n",
            "54000/54000 [==============================] - 3s 47us/sample - loss: 0.0043 - acc: 0.9986 - val_loss: 0.1248 - val_acc: 0.9780\n",
            "Epoch 58/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 5.0389e-04 - acc: 1.0000 - val_loss: 0.1195 - val_acc: 0.9802\n",
            "Epoch 59/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 2.1099e-04 - acc: 1.0000 - val_loss: 0.1180 - val_acc: 0.9798\n",
            "Epoch 60/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 1.6576e-04 - acc: 1.0000 - val_loss: 0.1183 - val_acc: 0.9800\n",
            "Epoch 61/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 1.4556e-04 - acc: 1.0000 - val_loss: 0.1178 - val_acc: 0.9798\n",
            "Epoch 62/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 1.2515e-04 - acc: 1.0000 - val_loss: 0.1192 - val_acc: 0.9802\n",
            "Epoch 63/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 1.1519e-04 - acc: 1.0000 - val_loss: 0.1175 - val_acc: 0.9800\n",
            "Epoch 64/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 1.0206e-04 - acc: 1.0000 - val_loss: 0.1185 - val_acc: 0.9795\n",
            "Epoch 65/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 9.4104e-05 - acc: 1.0000 - val_loss: 0.1186 - val_acc: 0.9800\n",
            "Epoch 66/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 8.2910e-05 - acc: 1.0000 - val_loss: 0.1200 - val_acc: 0.9800\n",
            "Epoch 67/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0115 - acc: 0.9970 - val_loss: 0.1348 - val_acc: 0.9765\n",
            "Epoch 68/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 0.0022 - acc: 0.9994 - val_loss: 0.1255 - val_acc: 0.9797\n",
            "Epoch 69/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 2.5807e-04 - acc: 1.0000 - val_loss: 0.1222 - val_acc: 0.9805\n",
            "Epoch 70/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 1.4062e-04 - acc: 1.0000 - val_loss: 0.1227 - val_acc: 0.9798\n",
            "Epoch 71/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 1.1221e-04 - acc: 1.0000 - val_loss: 0.1233 - val_acc: 0.9807\n",
            "Epoch 72/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 9.7515e-05 - acc: 1.0000 - val_loss: 0.1233 - val_acc: 0.9798\n",
            "Epoch 73/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 8.5898e-05 - acc: 1.0000 - val_loss: 0.1223 - val_acc: 0.9803\n",
            "Epoch 74/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 7.6841e-05 - acc: 1.0000 - val_loss: 0.1227 - val_acc: 0.9803\n",
            "Epoch 75/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 6.8197e-05 - acc: 1.0000 - val_loss: 0.1219 - val_acc: 0.9808\n",
            "Epoch 76/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 6.0663e-05 - acc: 1.0000 - val_loss: 0.1221 - val_acc: 0.9807\n",
            "Epoch 77/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 5.5312e-05 - acc: 1.0000 - val_loss: 0.1238 - val_acc: 0.9807\n",
            "Epoch 78/80\n",
            "54000/54000 [==============================] - 2s 43us/sample - loss: 5.0612e-05 - acc: 1.0000 - val_loss: 0.1229 - val_acc: 0.9812\n",
            "Epoch 79/80\n",
            "54000/54000 [==============================] - 2s 44us/sample - loss: 4.7117e-05 - acc: 1.0000 - val_loss: 0.1226 - val_acc: 0.9813\n",
            "Epoch 80/80\n",
            "54000/54000 [==============================] - 2s 46us/sample - loss: 0.0040 - acc: 0.9991 - val_loss: 0.2010 - val_acc: 0.9700\n",
            "++++++++++++++++++++++++++++ 2 of 10 ++++++++++++++++++++++++++++++\n",
            "Dense [100, 200], validation split 0.1, epochs 80, batch_size 100\n",
            "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_2 (Dense)              (None, 100)               78500     \n",
            "_________________________________________________________________\n",
            "dense_3 (Dense)              (None, 200)               20200     \n",
            "_________________________________________________________________\n",
            "dense_4 (Dense)              (None, 10)                2010      \n",
            "=================================================================\n",
            "Total params: 100,710\n",
            "Trainable params: 100,710\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n",
            "Train on 54000 samples, validate on 6000 samples\n",
            "Epoch 1/80\n",
            "54000/54000 [==============================] - 3s 54us/sample - loss: 0.3226 - acc: 0.9080 - val_loss: 0.1257 - val_acc: 0.9657\n",
            "Epoch 2/80\n",
            "54000/54000 [==============================] - 3s 51us/sample - loss: 0.1327 - acc: 0.9612 - val_loss: 0.1045 - val_acc: 0.9700\n",
            "Epoch 3/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0946 - acc: 0.9715 - val_loss: 0.1005 - val_acc: 0.9715\n",
            "Epoch 4/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0735 - acc: 0.9779 - val_loss: 0.0815 - val_acc: 0.9765\n",
            "Epoch 5/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0562 - acc: 0.9830 - val_loss: 0.0744 - val_acc: 0.9783\n",
            "Epoch 6/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0445 - acc: 0.9862 - val_loss: 0.0814 - val_acc: 0.9778\n",
            "Epoch 7/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0372 - acc: 0.9884 - val_loss: 0.0776 - val_acc: 0.9805\n",
            "Epoch 8/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0304 - acc: 0.9908 - val_loss: 0.0854 - val_acc: 0.9770\n",
            "Epoch 9/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0264 - acc: 0.9916 - val_loss: 0.0878 - val_acc: 0.9770\n",
            "Epoch 10/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0207 - acc: 0.9936 - val_loss: 0.0806 - val_acc: 0.9792\n",
            "Epoch 11/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0199 - acc: 0.9935 - val_loss: 0.0859 - val_acc: 0.9788\n",
            "Epoch 12/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0158 - acc: 0.9945 - val_loss: 0.0934 - val_acc: 0.9787\n",
            "Epoch 13/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0142 - acc: 0.9954 - val_loss: 0.0963 - val_acc: 0.9785\n",
            "Epoch 14/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0115 - acc: 0.9964 - val_loss: 0.0802 - val_acc: 0.9822\n",
            "Epoch 15/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0115 - acc: 0.9961 - val_loss: 0.0897 - val_acc: 0.9812\n",
            "Epoch 16/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0130 - acc: 0.9955 - val_loss: 0.1060 - val_acc: 0.9775\n",
            "Epoch 17/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0095 - acc: 0.9969 - val_loss: 0.0991 - val_acc: 0.9812\n",
            "Epoch 18/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0112 - acc: 0.9962 - val_loss: 0.0865 - val_acc: 0.9812\n",
            "Epoch 19/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0111 - acc: 0.9964 - val_loss: 0.1034 - val_acc: 0.9788\n",
            "Epoch 20/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0076 - acc: 0.9976 - val_loss: 0.0997 - val_acc: 0.9803\n",
            "Epoch 21/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0093 - acc: 0.9969 - val_loss: 0.1074 - val_acc: 0.9802\n",
            "Epoch 22/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0036 - acc: 0.9990 - val_loss: 0.1071 - val_acc: 0.9790\n",
            "Epoch 23/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0046 - acc: 0.9986 - val_loss: 0.1109 - val_acc: 0.9797\n",
            "Epoch 24/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0127 - acc: 0.9954 - val_loss: 0.1070 - val_acc: 0.9793\n",
            "Epoch 25/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0090 - acc: 0.9969 - val_loss: 0.1169 - val_acc: 0.9790\n",
            "Epoch 26/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0049 - acc: 0.9984 - val_loss: 0.1069 - val_acc: 0.9800\n",
            "Epoch 27/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0061 - acc: 0.9979 - val_loss: 0.1201 - val_acc: 0.9780\n",
            "Epoch 28/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0072 - acc: 0.9975 - val_loss: 0.1146 - val_acc: 0.9802\n",
            "Epoch 29/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0049 - acc: 0.9984 - val_loss: 0.1174 - val_acc: 0.9800\n",
            "Epoch 30/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0064 - acc: 0.9978 - val_loss: 0.1148 - val_acc: 0.9793\n",
            "Epoch 31/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0078 - acc: 0.9977 - val_loss: 0.1259 - val_acc: 0.9795\n",
            "Epoch 32/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0065 - acc: 0.9978 - val_loss: 0.1139 - val_acc: 0.9822\n",
            "Epoch 33/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0056 - acc: 0.9983 - val_loss: 0.1195 - val_acc: 0.9797\n",
            "Epoch 34/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0044 - acc: 0.9984 - val_loss: 0.1156 - val_acc: 0.9822\n",
            "Epoch 35/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0060 - acc: 0.9980 - val_loss: 0.1129 - val_acc: 0.9802\n",
            "Epoch 36/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0087 - acc: 0.9971 - val_loss: 0.1159 - val_acc: 0.9812\n",
            "Epoch 37/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0032 - acc: 0.9990 - val_loss: 0.1390 - val_acc: 0.9803\n",
            "Epoch 38/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0043 - acc: 0.9987 - val_loss: 0.1285 - val_acc: 0.9818\n",
            "Epoch 39/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0027 - acc: 0.9992 - val_loss: 0.1396 - val_acc: 0.9798\n",
            "Epoch 40/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0066 - acc: 0.9979 - val_loss: 0.1241 - val_acc: 0.9815\n",
            "Epoch 41/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0046 - acc: 0.9985 - val_loss: 0.1655 - val_acc: 0.9767\n",
            "Epoch 42/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0077 - acc: 0.9975 - val_loss: 0.1362 - val_acc: 0.9788\n",
            "Epoch 43/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0025 - acc: 0.9992 - val_loss: 0.1336 - val_acc: 0.9803\n",
            "Epoch 44/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0044 - acc: 0.9985 - val_loss: 0.1397 - val_acc: 0.9795\n",
            "Epoch 45/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0031 - acc: 0.9990 - val_loss: 0.1309 - val_acc: 0.9808\n",
            "Epoch 46/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0039 - acc: 0.9988 - val_loss: 0.1369 - val_acc: 0.9780\n",
            "Epoch 47/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0081 - acc: 0.9974 - val_loss: 0.1379 - val_acc: 0.9802\n",
            "Epoch 48/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0026 - acc: 0.9992 - val_loss: 0.1091 - val_acc: 0.9830\n",
            "Epoch 49/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0041 - acc: 0.9988 - val_loss: 0.1266 - val_acc: 0.9800\n",
            "Epoch 50/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0066 - acc: 0.9981 - val_loss: 0.1304 - val_acc: 0.9803\n",
            "Epoch 51/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0047 - acc: 0.9986 - val_loss: 0.1467 - val_acc: 0.9788\n",
            "Epoch 52/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0058 - acc: 0.9983 - val_loss: 0.1397 - val_acc: 0.9808\n",
            "Epoch 53/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0031 - acc: 0.9990 - val_loss: 0.1269 - val_acc: 0.9818\n",
            "Epoch 54/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0031 - acc: 0.9990 - val_loss: 0.1396 - val_acc: 0.9822\n",
            "Epoch 55/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0046 - acc: 0.9985 - val_loss: 0.1376 - val_acc: 0.9822\n",
            "Epoch 56/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0067 - acc: 0.9978 - val_loss: 0.1306 - val_acc: 0.9825\n",
            "Epoch 57/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0028 - acc: 0.9991 - val_loss: 0.1260 - val_acc: 0.9825\n",
            "Epoch 58/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 6.8321e-04 - acc: 0.9998 - val_loss: 0.1272 - val_acc: 0.9835\n",
            "Epoch 59/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0065 - acc: 0.9980 - val_loss: 0.1493 - val_acc: 0.9805\n",
            "Epoch 60/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0056 - acc: 0.9982 - val_loss: 0.1386 - val_acc: 0.9803\n",
            "Epoch 61/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0027 - acc: 0.9992 - val_loss: 0.1206 - val_acc: 0.9827\n",
            "Epoch 62/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 1.7618e-04 - acc: 1.0000 - val_loss: 0.1270 - val_acc: 0.9838\n",
            "Epoch 63/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 3.7390e-05 - acc: 1.0000 - val_loss: 0.1233 - val_acc: 0.9843\n",
            "Epoch 64/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 1.6325e-05 - acc: 1.0000 - val_loss: 0.1229 - val_acc: 0.9843\n",
            "Epoch 65/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 1.2431e-05 - acc: 1.0000 - val_loss: 0.1230 - val_acc: 0.9843\n",
            "Epoch 66/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 1.0090e-05 - acc: 1.0000 - val_loss: 0.1229 - val_acc: 0.9843\n",
            "Epoch 67/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 8.3095e-06 - acc: 1.0000 - val_loss: 0.1231 - val_acc: 0.9843\n",
            "Epoch 68/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 6.8628e-06 - acc: 1.0000 - val_loss: 0.1233 - val_acc: 0.9845\n",
            "Epoch 69/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 5.6535e-06 - acc: 1.0000 - val_loss: 0.1235 - val_acc: 0.9848\n",
            "Epoch 70/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 4.6950e-06 - acc: 1.0000 - val_loss: 0.1241 - val_acc: 0.9853\n",
            "Epoch 71/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 3.8472e-06 - acc: 1.0000 - val_loss: 0.1244 - val_acc: 0.9852\n",
            "Epoch 72/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 3.1571e-06 - acc: 1.0000 - val_loss: 0.1245 - val_acc: 0.9852\n",
            "Epoch 73/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 2.5858e-06 - acc: 1.0000 - val_loss: 0.1256 - val_acc: 0.9855\n",
            "Epoch 74/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 2.1307e-06 - acc: 1.0000 - val_loss: 0.1257 - val_acc: 0.9855\n",
            "Epoch 75/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 1.7444e-06 - acc: 1.0000 - val_loss: 0.1262 - val_acc: 0.9855\n",
            "Epoch 76/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 1.4200e-06 - acc: 1.0000 - val_loss: 0.1271 - val_acc: 0.9855\n",
            "Epoch 77/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 1.1609e-06 - acc: 1.0000 - val_loss: 0.1277 - val_acc: 0.9857\n",
            "Epoch 78/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 9.5813e-07 - acc: 1.0000 - val_loss: 0.1288 - val_acc: 0.9857\n",
            "Epoch 79/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 7.8421e-07 - acc: 1.0000 - val_loss: 0.1292 - val_acc: 0.9855\n",
            "Epoch 80/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 6.4235e-07 - acc: 1.0000 - val_loss: 0.1292 - val_acc: 0.9853\n",
            "++++++++++++++++++++++++++++ 3 of 10 ++++++++++++++++++++++++++++++\n",
            "Dense [200, 100], validation split 0.1, epochs 80, batch_size 100\n",
            "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_5 (Dense)              (None, 200)               157000    \n",
            "_________________________________________________________________\n",
            "dense_6 (Dense)              (None, 100)               20100     \n",
            "_________________________________________________________________\n",
            "dense_7 (Dense)              (None, 10)                1010      \n",
            "=================================================================\n",
            "Total params: 178,110\n",
            "Trainable params: 178,110\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n",
            "Train on 54000 samples, validate on 6000 samples\n",
            "Epoch 1/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 0.2929 - acc: 0.9158 - val_loss: 0.1177 - val_acc: 0.9688\n",
            "Epoch 2/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.1127 - acc: 0.9657 - val_loss: 0.0988 - val_acc: 0.9725\n",
            "Epoch 3/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0736 - acc: 0.9777 - val_loss: 0.0800 - val_acc: 0.9757\n",
            "Epoch 4/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0533 - acc: 0.9841 - val_loss: 0.0794 - val_acc: 0.9765\n",
            "Epoch 5/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0404 - acc: 0.9875 - val_loss: 0.0804 - val_acc: 0.9780\n",
            "Epoch 6/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0311 - acc: 0.9900 - val_loss: 0.0779 - val_acc: 0.9788\n",
            "Epoch 7/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0253 - acc: 0.9916 - val_loss: 0.0853 - val_acc: 0.9760\n",
            "Epoch 8/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0196 - acc: 0.9940 - val_loss: 0.0748 - val_acc: 0.9790\n",
            "Epoch 9/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0160 - acc: 0.9947 - val_loss: 0.0788 - val_acc: 0.9798\n",
            "Epoch 10/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0146 - acc: 0.9950 - val_loss: 0.0863 - val_acc: 0.9797\n",
            "Epoch 11/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0130 - acc: 0.9957 - val_loss: 0.0820 - val_acc: 0.9788\n",
            "Epoch 12/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0107 - acc: 0.9966 - val_loss: 0.0827 - val_acc: 0.9803\n",
            "Epoch 13/80\n",
            "54000/54000 [==============================] - 3s 50us/sample - loss: 0.0094 - acc: 0.9971 - val_loss: 0.0973 - val_acc: 0.9760\n",
            "Epoch 14/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0132 - acc: 0.9951 - val_loss: 0.0800 - val_acc: 0.9828\n",
            "Epoch 15/80\n",
            "54000/54000 [==============================] - 3s 51us/sample - loss: 0.0082 - acc: 0.9973 - val_loss: 0.0780 - val_acc: 0.9825\n",
            "Epoch 16/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0083 - acc: 0.9974 - val_loss: 0.1046 - val_acc: 0.9787\n",
            "Epoch 17/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0103 - acc: 0.9965 - val_loss: 0.0945 - val_acc: 0.9808\n",
            "Epoch 18/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0076 - acc: 0.9974 - val_loss: 0.0916 - val_acc: 0.9808\n",
            "Epoch 19/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0046 - acc: 0.9986 - val_loss: 0.0850 - val_acc: 0.9850\n",
            "Epoch 20/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0088 - acc: 0.9970 - val_loss: 0.1015 - val_acc: 0.9790\n",
            "Epoch 21/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0071 - acc: 0.9979 - val_loss: 0.1124 - val_acc: 0.9797\n",
            "Epoch 22/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0071 - acc: 0.9977 - val_loss: 0.0930 - val_acc: 0.9823\n",
            "Epoch 23/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0052 - acc: 0.9984 - val_loss: 0.0947 - val_acc: 0.9827\n",
            "Epoch 24/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0064 - acc: 0.9979 - val_loss: 0.1118 - val_acc: 0.9813\n",
            "Epoch 25/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0078 - acc: 0.9976 - val_loss: 0.1071 - val_acc: 0.9810\n",
            "Epoch 26/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0068 - acc: 0.9978 - val_loss: 0.1016 - val_acc: 0.9807\n",
            "Epoch 27/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0042 - acc: 0.9986 - val_loss: 0.1029 - val_acc: 0.9823\n",
            "Epoch 28/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0061 - acc: 0.9978 - val_loss: 0.1153 - val_acc: 0.9792\n",
            "Epoch 29/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0049 - acc: 0.9984 - val_loss: 0.0945 - val_acc: 0.9812\n",
            "Epoch 30/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0079 - acc: 0.9977 - val_loss: 0.1075 - val_acc: 0.9810\n",
            "Epoch 31/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0018 - acc: 0.9994 - val_loss: 0.0952 - val_acc: 0.9832\n",
            "Epoch 32/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0056 - acc: 0.9983 - val_loss: 0.1377 - val_acc: 0.9772\n",
            "Epoch 33/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0067 - acc: 0.9978 - val_loss: 0.1126 - val_acc: 0.9803\n",
            "Epoch 34/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0035 - acc: 0.9989 - val_loss: 0.1068 - val_acc: 0.9813\n",
            "Epoch 35/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0038 - acc: 0.9988 - val_loss: 0.1330 - val_acc: 0.9798\n",
            "Epoch 36/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0039 - acc: 0.9987 - val_loss: 0.1204 - val_acc: 0.9822\n",
            "Epoch 37/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0066 - acc: 0.9981 - val_loss: 0.1198 - val_acc: 0.9815\n",
            "Epoch 38/80\n",
            "54000/54000 [==============================] - 3s 54us/sample - loss: 0.0054 - acc: 0.9982 - val_loss: 0.1094 - val_acc: 0.9832\n",
            "Epoch 39/80\n",
            "54000/54000 [==============================] - 3s 51us/sample - loss: 0.0044 - acc: 0.9985 - val_loss: 0.1131 - val_acc: 0.9827\n",
            "Epoch 40/80\n",
            "54000/54000 [==============================] - 3s 50us/sample - loss: 0.0023 - acc: 0.9992 - val_loss: 0.1044 - val_acc: 0.9842\n",
            "Epoch 41/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 2.1433e-04 - acc: 1.0000 - val_loss: 0.1074 - val_acc: 0.9835\n",
            "Epoch 42/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0096 - acc: 0.9971 - val_loss: 0.1264 - val_acc: 0.9785\n",
            "Epoch 43/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0046 - acc: 0.9985 - val_loss: 0.1155 - val_acc: 0.9808\n",
            "Epoch 44/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0031 - acc: 0.9989 - val_loss: 0.1163 - val_acc: 0.9837\n",
            "Epoch 45/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 6.9478e-04 - acc: 0.9998 - val_loss: 0.1145 - val_acc: 0.9822\n",
            "Epoch 46/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0066 - acc: 0.9980 - val_loss: 0.1284 - val_acc: 0.9798\n",
            "Epoch 47/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0064 - acc: 0.9980 - val_loss: 0.1094 - val_acc: 0.9830\n",
            "Epoch 48/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0016 - acc: 0.9995 - val_loss: 0.1133 - val_acc: 0.9827\n",
            "Epoch 49/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0011 - acc: 0.9997 - val_loss: 0.1136 - val_acc: 0.9837\n",
            "Epoch 50/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0028 - acc: 0.9991 - val_loss: 0.1132 - val_acc: 0.9822\n",
            "Epoch 51/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0089 - acc: 0.9973 - val_loss: 0.1068 - val_acc: 0.9825\n",
            "Epoch 52/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0054 - acc: 0.9983 - val_loss: 0.1196 - val_acc: 0.9828\n",
            "Epoch 53/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0024 - acc: 0.9994 - val_loss: 0.1258 - val_acc: 0.9810\n",
            "Epoch 54/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0032 - acc: 0.9991 - val_loss: 0.1180 - val_acc: 0.9832\n",
            "Epoch 55/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0022 - acc: 0.9993 - val_loss: 0.1307 - val_acc: 0.9820\n",
            "Epoch 56/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0043 - acc: 0.9986 - val_loss: 0.1492 - val_acc: 0.9793\n",
            "Epoch 57/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0043 - acc: 0.9988 - val_loss: 0.1349 - val_acc: 0.9820\n",
            "Epoch 58/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0026 - acc: 0.9992 - val_loss: 0.1235 - val_acc: 0.9825\n",
            "Epoch 59/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0022 - acc: 0.9993 - val_loss: 0.1127 - val_acc: 0.9825\n",
            "Epoch 60/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0011 - acc: 0.9996 - val_loss: 0.1232 - val_acc: 0.9832\n",
            "Epoch 61/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0066 - acc: 0.9980 - val_loss: 0.1244 - val_acc: 0.9827\n",
            "Epoch 62/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0025 - acc: 0.9993 - val_loss: 0.1110 - val_acc: 0.9845\n",
            "Epoch 63/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0053 - acc: 0.9985 - val_loss: 0.1243 - val_acc: 0.9827\n",
            "Epoch 64/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0038 - acc: 0.9988 - val_loss: 0.1174 - val_acc: 0.9828\n",
            "Epoch 65/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0020 - acc: 0.9994 - val_loss: 0.1182 - val_acc: 0.9832\n",
            "Epoch 66/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 8.0850e-04 - acc: 0.9997 - val_loss: 0.1111 - val_acc: 0.9847\n",
            "Epoch 67/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 8.0560e-04 - acc: 0.9998 - val_loss: 0.1421 - val_acc: 0.9823\n",
            "Epoch 68/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0098 - acc: 0.9971 - val_loss: 0.1525 - val_acc: 0.9792\n",
            "Epoch 69/80\n",
            "54000/54000 [==============================] - 3s 50us/sample - loss: 0.0059 - acc: 0.9984 - val_loss: 0.1197 - val_acc: 0.9842\n",
            "Epoch 70/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0012 - acc: 0.9996 - val_loss: 0.1089 - val_acc: 0.9853\n",
            "Epoch 71/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0024 - acc: 0.9993 - val_loss: 0.1038 - val_acc: 0.9857\n",
            "Epoch 72/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0030 - acc: 0.9990 - val_loss: 0.1200 - val_acc: 0.9823\n",
            "Epoch 73/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0036 - acc: 0.9992 - val_loss: 0.1214 - val_acc: 0.9832\n",
            "Epoch 74/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0027 - acc: 0.9992 - val_loss: 0.1255 - val_acc: 0.9822\n",
            "Epoch 75/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0032 - acc: 0.9990 - val_loss: 0.1349 - val_acc: 0.9818\n",
            "Epoch 76/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0029 - acc: 0.9990 - val_loss: 0.1090 - val_acc: 0.9853\n",
            "Epoch 77/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0012 - acc: 0.9996 - val_loss: 0.1089 - val_acc: 0.9862\n",
            "Epoch 78/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0050 - acc: 0.9988 - val_loss: 0.1283 - val_acc: 0.9832\n",
            "Epoch 79/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0061 - acc: 0.9984 - val_loss: 0.1229 - val_acc: 0.9852\n",
            "Epoch 80/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0023 - acc: 0.9994 - val_loss: 0.1111 - val_acc: 0.9852\n",
            "++++++++++++++++++++++++++++ 4 of 10 ++++++++++++++++++++++++++++++\n",
            "Dense [400], validation split 0.1, epochs 80, batch_size 100\n",
            "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_8 (Dense)              (None, 400)               314000    \n",
            "_________________________________________________________________\n",
            "dense_9 (Dense)              (None, 10)                4010      \n",
            "=================================================================\n",
            "Total params: 318,010\n",
            "Trainable params: 318,010\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n",
            "Train on 54000 samples, validate on 6000 samples\n",
            "Epoch 1/80\n",
            "54000/54000 [==============================] - 3s 50us/sample - loss: 0.2778 - acc: 0.9202 - val_loss: 0.1172 - val_acc: 0.9692\n",
            "Epoch 2/80\n",
            "54000/54000 [==============================] - 2s 46us/sample - loss: 0.1147 - acc: 0.9661 - val_loss: 0.0872 - val_acc: 0.9745\n",
            "Epoch 3/80\n",
            "54000/54000 [==============================] - 2s 46us/sample - loss: 0.0751 - acc: 0.9777 - val_loss: 0.0891 - val_acc: 0.9740\n",
            "Epoch 4/80\n",
            "54000/54000 [==============================] - 3s 46us/sample - loss: 0.0542 - acc: 0.9844 - val_loss: 0.0755 - val_acc: 0.9782\n",
            "Epoch 5/80\n",
            "54000/54000 [==============================] - 3s 46us/sample - loss: 0.0391 - acc: 0.9888 - val_loss: 0.0734 - val_acc: 0.9793\n",
            "Epoch 6/80\n",
            "54000/54000 [==============================] - 2s 46us/sample - loss: 0.0294 - acc: 0.9914 - val_loss: 0.0684 - val_acc: 0.9812\n",
            "Epoch 7/80\n",
            "54000/54000 [==============================] - 3s 46us/sample - loss: 0.0224 - acc: 0.9939 - val_loss: 0.0658 - val_acc: 0.9825\n",
            "Epoch 8/80\n",
            "54000/54000 [==============================] - 2s 46us/sample - loss: 0.0181 - acc: 0.9949 - val_loss: 0.0701 - val_acc: 0.9793\n",
            "Epoch 9/80\n",
            "54000/54000 [==============================] - 2s 46us/sample - loss: 0.0138 - acc: 0.9965 - val_loss: 0.0660 - val_acc: 0.9815\n",
            "Epoch 10/80\n",
            "54000/54000 [==============================] - 3s 47us/sample - loss: 0.0111 - acc: 0.9971 - val_loss: 0.0724 - val_acc: 0.9800\n",
            "Epoch 11/80\n",
            "54000/54000 [==============================] - 2s 46us/sample - loss: 0.0090 - acc: 0.9979 - val_loss: 0.0659 - val_acc: 0.9845\n",
            "Epoch 12/80\n",
            "54000/54000 [==============================] - 3s 46us/sample - loss: 0.0069 - acc: 0.9984 - val_loss: 0.0692 - val_acc: 0.9848\n",
            "Epoch 13/80\n",
            "54000/54000 [==============================] - 3s 47us/sample - loss: 0.0063 - acc: 0.9983 - val_loss: 0.0771 - val_acc: 0.9820\n",
            "Epoch 14/80\n",
            "54000/54000 [==============================] - 3s 46us/sample - loss: 0.0072 - acc: 0.9981 - val_loss: 0.0775 - val_acc: 0.9813\n",
            "Epoch 15/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0047 - acc: 0.9990 - val_loss: 0.0775 - val_acc: 0.9813\n",
            "Epoch 16/80\n",
            "54000/54000 [==============================] - 5s 88us/sample - loss: 0.0060 - acc: 0.9983 - val_loss: 0.0822 - val_acc: 0.9798\n",
            "Epoch 17/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0042 - acc: 0.9991 - val_loss: 0.0724 - val_acc: 0.9832\n",
            "Epoch 18/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0032 - acc: 0.9992 - val_loss: 0.0745 - val_acc: 0.9845\n",
            "Epoch 19/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0035 - acc: 0.9988 - val_loss: 0.0809 - val_acc: 0.9825\n",
            "Epoch 20/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0068 - acc: 0.9978 - val_loss: 0.0863 - val_acc: 0.9818\n",
            "Epoch 21/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0042 - acc: 0.9989 - val_loss: 0.0731 - val_acc: 0.9848\n",
            "Epoch 22/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0033 - acc: 0.9989 - val_loss: 0.0817 - val_acc: 0.9825\n",
            "Epoch 23/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0020 - acc: 0.9995 - val_loss: 0.0777 - val_acc: 0.9847\n",
            "Epoch 24/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0037 - acc: 0.9988 - val_loss: 0.1230 - val_acc: 0.9755\n",
            "Epoch 25/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0056 - acc: 0.9983 - val_loss: 0.0879 - val_acc: 0.9827\n",
            "Epoch 26/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0022 - acc: 0.9995 - val_loss: 0.0763 - val_acc: 0.9850\n",
            "Epoch 27/80\n",
            "54000/54000 [==============================] - 3s 50us/sample - loss: 0.0022 - acc: 0.9994 - val_loss: 0.0859 - val_acc: 0.9837\n",
            "Epoch 28/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0016 - acc: 0.9996 - val_loss: 0.0829 - val_acc: 0.9842\n",
            "Epoch 29/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 2.0841e-04 - acc: 1.0000 - val_loss: 0.0783 - val_acc: 0.9857\n",
            "Epoch 30/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 9.0302e-05 - acc: 1.0000 - val_loss: 0.0790 - val_acc: 0.9863\n",
            "Epoch 31/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 6.7473e-05 - acc: 1.0000 - val_loss: 0.0781 - val_acc: 0.9868\n",
            "Epoch 32/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 5.4730e-05 - acc: 1.0000 - val_loss: 0.0779 - val_acc: 0.9873\n",
            "Epoch 33/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 4.7082e-05 - acc: 1.0000 - val_loss: 0.0782 - val_acc: 0.9877\n",
            "Epoch 34/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 3.9982e-05 - acc: 1.0000 - val_loss: 0.0789 - val_acc: 0.9870\n",
            "Epoch 35/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 3.4195e-05 - acc: 1.0000 - val_loss: 0.0785 - val_acc: 0.9873\n",
            "Epoch 36/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 2.8873e-05 - acc: 1.0000 - val_loss: 0.0781 - val_acc: 0.9875\n",
            "Epoch 37/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 2.6386e-05 - acc: 1.0000 - val_loss: 0.0784 - val_acc: 0.9875\n",
            "Epoch 38/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 2.2053e-05 - acc: 1.0000 - val_loss: 0.0797 - val_acc: 0.9872\n",
            "Epoch 39/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0193 - acc: 0.9944 - val_loss: 0.0898 - val_acc: 0.9835\n",
            "Epoch 40/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0026 - acc: 0.9991 - val_loss: 0.0998 - val_acc: 0.9823\n",
            "Epoch 41/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 4.2795e-04 - acc: 1.0000 - val_loss: 0.0851 - val_acc: 0.9850\n",
            "Epoch 42/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 1.0740e-04 - acc: 1.0000 - val_loss: 0.0860 - val_acc: 0.9850\n",
            "Epoch 43/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 7.5971e-05 - acc: 1.0000 - val_loss: 0.0853 - val_acc: 0.9853\n",
            "Epoch 44/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 6.0648e-05 - acc: 1.0000 - val_loss: 0.0855 - val_acc: 0.9853\n",
            "Epoch 45/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 4.9703e-05 - acc: 1.0000 - val_loss: 0.0855 - val_acc: 0.9858\n",
            "Epoch 46/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 4.1161e-05 - acc: 1.0000 - val_loss: 0.0854 - val_acc: 0.9858\n",
            "Epoch 47/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 3.4780e-05 - acc: 1.0000 - val_loss: 0.0858 - val_acc: 0.9860\n",
            "Epoch 48/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 2.9156e-05 - acc: 1.0000 - val_loss: 0.0855 - val_acc: 0.9860\n",
            "Epoch 49/80\n",
            "54000/54000 [==============================] - 3s 51us/sample - loss: 2.4636e-05 - acc: 1.0000 - val_loss: 0.0856 - val_acc: 0.9863\n",
            "Epoch 50/80\n",
            "54000/54000 [==============================] - 3s 51us/sample - loss: 2.0982e-05 - acc: 1.0000 - val_loss: 0.0851 - val_acc: 0.9863\n",
            "Epoch 51/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 1.7349e-05 - acc: 1.0000 - val_loss: 0.0857 - val_acc: 0.9865\n",
            "Epoch 52/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 1.4541e-05 - acc: 1.0000 - val_loss: 0.0851 - val_acc: 0.9870\n",
            "Epoch 53/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 1.2136e-05 - acc: 1.0000 - val_loss: 0.0864 - val_acc: 0.9867\n",
            "Epoch 54/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 1.0484e-05 - acc: 1.0000 - val_loss: 0.0851 - val_acc: 0.9870\n",
            "Epoch 55/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0213 - acc: 0.9943 - val_loss: 0.0970 - val_acc: 0.9828\n",
            "Epoch 56/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 0.0014 - acc: 0.9996 - val_loss: 0.0924 - val_acc: 0.9840\n",
            "Epoch 57/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 1.4323e-04 - acc: 1.0000 - val_loss: 0.0926 - val_acc: 0.9845\n",
            "Epoch 58/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 6.2741e-05 - acc: 1.0000 - val_loss: 0.0916 - val_acc: 0.9845\n",
            "Epoch 59/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 4.7539e-05 - acc: 1.0000 - val_loss: 0.0917 - val_acc: 0.9842\n",
            "Epoch 60/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 3.8924e-05 - acc: 1.0000 - val_loss: 0.0921 - val_acc: 0.9838\n",
            "Epoch 61/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 3.2280e-05 - acc: 1.0000 - val_loss: 0.0921 - val_acc: 0.9843\n",
            "Epoch 62/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 2.6880e-05 - acc: 1.0000 - val_loss: 0.0917 - val_acc: 0.9845\n",
            "Epoch 63/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 2.2642e-05 - acc: 1.0000 - val_loss: 0.0920 - val_acc: 0.9850\n",
            "Epoch 64/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 1.9118e-05 - acc: 1.0000 - val_loss: 0.0924 - val_acc: 0.9848\n",
            "Epoch 65/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 1.6032e-05 - acc: 1.0000 - val_loss: 0.0919 - val_acc: 0.9850\n",
            "Epoch 66/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 1.3484e-05 - acc: 1.0000 - val_loss: 0.0921 - val_acc: 0.9852\n",
            "Epoch 67/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 1.1161e-05 - acc: 1.0000 - val_loss: 0.0925 - val_acc: 0.9850\n",
            "Epoch 68/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 9.4122e-06 - acc: 1.0000 - val_loss: 0.0919 - val_acc: 0.9858\n",
            "Epoch 69/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 7.8799e-06 - acc: 1.0000 - val_loss: 0.0922 - val_acc: 0.9850\n",
            "Epoch 70/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 6.4558e-06 - acc: 1.0000 - val_loss: 0.0931 - val_acc: 0.9855\n",
            "Epoch 71/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 5.3129e-06 - acc: 1.0000 - val_loss: 0.0935 - val_acc: 0.9852\n",
            "Epoch 72/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 4.4332e-06 - acc: 1.0000 - val_loss: 0.0938 - val_acc: 0.9860\n",
            "Epoch 73/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 3.7879e-06 - acc: 1.0000 - val_loss: 0.0931 - val_acc: 0.9857\n",
            "Epoch 74/80\n",
            "54000/54000 [==============================] - 3s 50us/sample - loss: 3.1286e-06 - acc: 1.0000 - val_loss: 0.0937 - val_acc: 0.9857\n",
            "Epoch 75/80\n",
            "54000/54000 [==============================] - 3s 50us/sample - loss: 2.5086e-06 - acc: 1.0000 - val_loss: 0.0936 - val_acc: 0.9855\n",
            "Epoch 76/80\n",
            "54000/54000 [==============================] - 3s 50us/sample - loss: 2.5963e-06 - acc: 1.0000 - val_loss: 0.1123 - val_acc: 0.9827\n",
            "Epoch 77/80\n",
            "54000/54000 [==============================] - 3s 50us/sample - loss: 0.0181 - acc: 0.9951 - val_loss: 0.1161 - val_acc: 0.9815\n",
            "Epoch 78/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0012 - acc: 0.9997 - val_loss: 0.1088 - val_acc: 0.9835\n",
            "Epoch 79/80\n",
            "54000/54000 [==============================] - 3s 49us/sample - loss: 0.0013 - acc: 0.9996 - val_loss: 0.1092 - val_acc: 0.9832\n",
            "Epoch 80/80\n",
            "54000/54000 [==============================] - 3s 48us/sample - loss: 2.8253e-04 - acc: 0.9999 - val_loss: 0.1003 - val_acc: 0.9847\n",
            "++++++++++++++++++++++++++++ 5 of 10 ++++++++++++++++++++++++++++++\n",
            "Dense [400, 100], validation split 0.1, epochs 80, batch_size 100\n",
            "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_10 (Dense)             (None, 400)               314000    \n",
            "_________________________________________________________________\n",
            "dense_11 (Dense)             (None, 100)               40100     \n",
            "_________________________________________________________________\n",
            "dense_12 (Dense)             (None, 10)                1010      \n",
            "=================================================================\n",
            "Total params: 355,110\n",
            "Trainable params: 355,110\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n",
            "Train on 54000 samples, validate on 6000 samples\n",
            "Epoch 1/80\n",
            "54000/54000 [==============================] - 3s 60us/sample - loss: 0.2548 - acc: 0.9260 - val_loss: 0.1059 - val_acc: 0.9728\n",
            "Epoch 2/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0972 - acc: 0.9707 - val_loss: 0.0863 - val_acc: 0.9763\n",
            "Epoch 3/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0619 - acc: 0.9811 - val_loss: 0.0775 - val_acc: 0.9773\n",
            "Epoch 4/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0431 - acc: 0.9865 - val_loss: 0.0690 - val_acc: 0.9810\n",
            "Epoch 5/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0317 - acc: 0.9899 - val_loss: 0.0734 - val_acc: 0.9783\n",
            "Epoch 6/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0248 - acc: 0.9917 - val_loss: 0.0733 - val_acc: 0.9808\n",
            "Epoch 7/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0195 - acc: 0.9936 - val_loss: 0.0645 - val_acc: 0.9832\n",
            "Epoch 8/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0166 - acc: 0.9946 - val_loss: 0.0721 - val_acc: 0.9803\n",
            "Epoch 9/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0171 - acc: 0.9940 - val_loss: 0.0778 - val_acc: 0.9803\n",
            "Epoch 10/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0093 - acc: 0.9970 - val_loss: 0.0714 - val_acc: 0.9840\n",
            "Epoch 11/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0120 - acc: 0.9959 - val_loss: 0.0852 - val_acc: 0.9820\n",
            "Epoch 12/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0138 - acc: 0.9954 - val_loss: 0.0824 - val_acc: 0.9812\n",
            "Epoch 13/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0097 - acc: 0.9970 - val_loss: 0.0884 - val_acc: 0.9805\n",
            "Epoch 14/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0082 - acc: 0.9976 - val_loss: 0.0965 - val_acc: 0.9782\n",
            "Epoch 15/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0098 - acc: 0.9968 - val_loss: 0.0927 - val_acc: 0.9810\n",
            "Epoch 16/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0078 - acc: 0.9974 - val_loss: 0.0968 - val_acc: 0.9802\n",
            "Epoch 17/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0089 - acc: 0.9970 - val_loss: 0.0904 - val_acc: 0.9842\n",
            "Epoch 18/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0059 - acc: 0.9979 - val_loss: 0.0985 - val_acc: 0.9825\n",
            "Epoch 19/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0075 - acc: 0.9976 - val_loss: 0.0997 - val_acc: 0.9807\n",
            "Epoch 20/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0102 - acc: 0.9968 - val_loss: 0.0922 - val_acc: 0.9828\n",
            "Epoch 21/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0056 - acc: 0.9983 - val_loss: 0.0923 - val_acc: 0.9812\n",
            "Epoch 22/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0039 - acc: 0.9988 - val_loss: 0.1185 - val_acc: 0.9802\n",
            "Epoch 23/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0058 - acc: 0.9980 - val_loss: 0.1207 - val_acc: 0.9800\n",
            "Epoch 24/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0112 - acc: 0.9964 - val_loss: 0.0885 - val_acc: 0.9833\n",
            "Epoch 25/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0038 - acc: 0.9986 - val_loss: 0.0920 - val_acc: 0.9848\n",
            "Epoch 26/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0013 - acc: 0.9996 - val_loss: 0.0919 - val_acc: 0.9838\n",
            "Epoch 27/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0017 - acc: 0.9993 - val_loss: 0.1136 - val_acc: 0.9820\n",
            "Epoch 28/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0107 - acc: 0.9967 - val_loss: 0.1046 - val_acc: 0.9828\n",
            "Epoch 29/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0074 - acc: 0.9977 - val_loss: 0.0971 - val_acc: 0.9830\n",
            "Epoch 30/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0025 - acc: 0.9992 - val_loss: 0.1023 - val_acc: 0.9833\n",
            "Epoch 31/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0029 - acc: 0.9990 - val_loss: 0.1115 - val_acc: 0.9832\n",
            "Epoch 32/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0062 - acc: 0.9979 - val_loss: 0.1016 - val_acc: 0.9835\n",
            "Epoch 33/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0071 - acc: 0.9979 - val_loss: 0.1012 - val_acc: 0.9832\n",
            "Epoch 34/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0028 - acc: 0.9991 - val_loss: 0.0998 - val_acc: 0.9838\n",
            "Epoch 35/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0010 - acc: 0.9997 - val_loss: 0.1100 - val_acc: 0.9828\n",
            "Epoch 36/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0041 - acc: 0.9987 - val_loss: 0.1283 - val_acc: 0.9798\n",
            "Epoch 37/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0087 - acc: 0.9975 - val_loss: 0.1210 - val_acc: 0.9813\n",
            "Epoch 38/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0032 - acc: 0.9989 - val_loss: 0.1091 - val_acc: 0.9830\n",
            "Epoch 39/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0050 - acc: 0.9986 - val_loss: 0.1132 - val_acc: 0.9823\n",
            "Epoch 40/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0045 - acc: 0.9987 - val_loss: 0.1141 - val_acc: 0.9802\n",
            "Epoch 41/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0060 - acc: 0.9981 - val_loss: 0.1117 - val_acc: 0.9825\n",
            "Epoch 42/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0068 - acc: 0.9982 - val_loss: 0.0946 - val_acc: 0.9840\n",
            "Epoch 43/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0046 - acc: 0.9988 - val_loss: 0.1004 - val_acc: 0.9850\n",
            "Epoch 44/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0025 - acc: 0.9992 - val_loss: 0.0977 - val_acc: 0.9848\n",
            "Epoch 45/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 5.2399e-04 - acc: 0.9998 - val_loss: 0.0950 - val_acc: 0.9863\n",
            "Epoch 46/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 2.8585e-04 - acc: 0.9999 - val_loss: 0.1103 - val_acc: 0.9850\n",
            "Epoch 47/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0068 - acc: 0.9980 - val_loss: 0.1329 - val_acc: 0.9807\n",
            "Epoch 48/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0056 - acc: 0.9983 - val_loss: 0.1132 - val_acc: 0.9822\n",
            "Epoch 49/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0024 - acc: 0.9992 - val_loss: 0.1328 - val_acc: 0.9802\n",
            "Epoch 50/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0047 - acc: 0.9985 - val_loss: 0.1321 - val_acc: 0.9800\n",
            "Epoch 51/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0047 - acc: 0.9986 - val_loss: 0.1164 - val_acc: 0.9830\n",
            "Epoch 52/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0014 - acc: 0.9995 - val_loss: 0.1199 - val_acc: 0.9833\n",
            "Epoch 53/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0016 - acc: 0.9994 - val_loss: 0.1264 - val_acc: 0.9833\n",
            "Epoch 54/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0075 - acc: 0.9979 - val_loss: 0.1410 - val_acc: 0.9825\n",
            "Epoch 55/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0045 - acc: 0.9987 - val_loss: 0.1358 - val_acc: 0.9812\n",
            "Epoch 56/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0038 - acc: 0.9990 - val_loss: 0.1491 - val_acc: 0.9817\n",
            "Epoch 57/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0016 - acc: 0.9996 - val_loss: 0.1247 - val_acc: 0.9838\n",
            "Epoch 58/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0015 - acc: 0.9996 - val_loss: 0.1348 - val_acc: 0.9828\n",
            "Epoch 59/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0061 - acc: 0.9983 - val_loss: 0.1212 - val_acc: 0.9827\n",
            "Epoch 60/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0051 - acc: 0.9986 - val_loss: 0.1307 - val_acc: 0.9808\n",
            "Epoch 61/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0017 - acc: 0.9993 - val_loss: 0.1174 - val_acc: 0.9830\n",
            "Epoch 62/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 7.7439e-04 - acc: 0.9998 - val_loss: 0.1096 - val_acc: 0.9833\n",
            "Epoch 63/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 4.7191e-05 - acc: 1.0000 - val_loss: 0.1071 - val_acc: 0.9843\n",
            "Epoch 64/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 9.9032e-06 - acc: 1.0000 - val_loss: 0.1076 - val_acc: 0.9845\n",
            "Epoch 65/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 6.1019e-06 - acc: 1.0000 - val_loss: 0.1077 - val_acc: 0.9845\n",
            "Epoch 66/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 4.4892e-06 - acc: 1.0000 - val_loss: 0.1086 - val_acc: 0.9848\n",
            "Epoch 67/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 2.8983e-06 - acc: 1.0000 - val_loss: 0.1114 - val_acc: 0.9850\n",
            "Epoch 68/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 1.8748e-06 - acc: 1.0000 - val_loss: 0.1131 - val_acc: 0.9850\n",
            "Epoch 69/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 1.3186e-06 - acc: 1.0000 - val_loss: 0.1161 - val_acc: 0.9847\n",
            "Epoch 70/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 9.5157e-07 - acc: 1.0000 - val_loss: 0.1184 - val_acc: 0.9848\n",
            "Epoch 71/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 7.6642e-07 - acc: 1.0000 - val_loss: 0.1194 - val_acc: 0.9843\n",
            "Epoch 72/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 6.4156e-07 - acc: 1.0000 - val_loss: 0.1194 - val_acc: 0.9850\n",
            "Epoch 73/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 6.0040e-07 - acc: 1.0000 - val_loss: 0.1202 - val_acc: 0.9848\n",
            "Epoch 74/80\n",
            "54000/54000 [==============================] - 3s 57us/sample - loss: 4.4490e-07 - acc: 1.0000 - val_loss: 0.1209 - val_acc: 0.9850\n",
            "Epoch 75/80\n",
            "54000/54000 [==============================] - 3s 57us/sample - loss: 3.7538e-07 - acc: 1.0000 - val_loss: 0.1221 - val_acc: 0.9848\n",
            "Epoch 76/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 3.2184e-07 - acc: 1.0000 - val_loss: 0.1232 - val_acc: 0.9848\n",
            "Epoch 77/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 2.8280e-07 - acc: 1.0000 - val_loss: 0.1236 - val_acc: 0.9848\n",
            "Epoch 78/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 2.4865e-07 - acc: 1.0000 - val_loss: 0.1239 - val_acc: 0.9850\n",
            "Epoch 79/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 2.2462e-07 - acc: 1.0000 - val_loss: 0.1248 - val_acc: 0.9852\n",
            "Epoch 80/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 2.0051e-07 - acc: 1.0000 - val_loss: 0.1249 - val_acc: 0.9852\n",
            "++++++++++++++++++++++++++++ 6 of 10 ++++++++++++++++++++++++++++++\n",
            "Dense [400, 400], validation split 0.1, epochs 80, batch_size 100\n",
            "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_13 (Dense)             (None, 400)               314000    \n",
            "_________________________________________________________________\n",
            "dense_14 (Dense)             (None, 400)               160400    \n",
            "_________________________________________________________________\n",
            "dense_15 (Dense)             (None, 10)                4010      \n",
            "=================================================================\n",
            "Total params: 478,410\n",
            "Trainable params: 478,410\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n",
            "Train on 54000 samples, validate on 6000 samples\n",
            "Epoch 1/80\n",
            "54000/54000 [==============================] - 3s 64us/sample - loss: 0.2303 - acc: 0.9305 - val_loss: 0.1027 - val_acc: 0.9683\n",
            "Epoch 2/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0867 - acc: 0.9731 - val_loss: 0.0854 - val_acc: 0.9743\n",
            "Epoch 3/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0545 - acc: 0.9823 - val_loss: 0.0692 - val_acc: 0.9785\n",
            "Epoch 4/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0390 - acc: 0.9876 - val_loss: 0.0717 - val_acc: 0.9787\n",
            "Epoch 5/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0291 - acc: 0.9903 - val_loss: 0.0811 - val_acc: 0.9768\n",
            "Epoch 6/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0245 - acc: 0.9919 - val_loss: 0.0866 - val_acc: 0.9798\n",
            "Epoch 7/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0204 - acc: 0.9934 - val_loss: 0.0766 - val_acc: 0.9803\n",
            "Epoch 8/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0180 - acc: 0.9938 - val_loss: 0.0812 - val_acc: 0.9820\n",
            "Epoch 9/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0154 - acc: 0.9949 - val_loss: 0.0854 - val_acc: 0.9793\n",
            "Epoch 10/80\n",
            "54000/54000 [==============================] - 3s 57us/sample - loss: 0.0155 - acc: 0.9950 - val_loss: 0.0819 - val_acc: 0.9835\n",
            "Epoch 11/80\n",
            "54000/54000 [==============================] - 3s 57us/sample - loss: 0.0113 - acc: 0.9962 - val_loss: 0.0866 - val_acc: 0.9817\n",
            "Epoch 12/80\n",
            "54000/54000 [==============================] - 3s 57us/sample - loss: 0.0126 - acc: 0.9958 - val_loss: 0.0898 - val_acc: 0.9803\n",
            "Epoch 13/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0096 - acc: 0.9969 - val_loss: 0.0990 - val_acc: 0.9780\n",
            "Epoch 14/80\n",
            "54000/54000 [==============================] - 3s 57us/sample - loss: 0.0137 - acc: 0.9956 - val_loss: 0.0940 - val_acc: 0.9827\n",
            "Epoch 15/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0111 - acc: 0.9962 - val_loss: 0.0910 - val_acc: 0.9822\n",
            "Epoch 16/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0112 - acc: 0.9966 - val_loss: 0.0868 - val_acc: 0.9813\n",
            "Epoch 17/80\n",
            "54000/54000 [==============================] - 3s 60us/sample - loss: 0.0066 - acc: 0.9978 - val_loss: 0.0918 - val_acc: 0.9827\n",
            "Epoch 18/80\n",
            "54000/54000 [==============================] - 3s 60us/sample - loss: 0.0052 - acc: 0.9984 - val_loss: 0.1148 - val_acc: 0.9792\n",
            "Epoch 19/80\n",
            "54000/54000 [==============================] - 3s 59us/sample - loss: 0.0081 - acc: 0.9974 - val_loss: 0.1081 - val_acc: 0.9802\n",
            "Epoch 20/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0121 - acc: 0.9962 - val_loss: 0.1086 - val_acc: 0.9793\n",
            "Epoch 21/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0074 - acc: 0.9977 - val_loss: 0.0987 - val_acc: 0.9825\n",
            "Epoch 22/80\n",
            "54000/54000 [==============================] - 3s 57us/sample - loss: 0.0086 - acc: 0.9972 - val_loss: 0.0833 - val_acc: 0.9845\n",
            "Epoch 23/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0045 - acc: 0.9986 - val_loss: 0.1000 - val_acc: 0.9840\n",
            "Epoch 24/80\n",
            "54000/54000 [==============================] - 3s 57us/sample - loss: 0.0058 - acc: 0.9981 - val_loss: 0.0994 - val_acc: 0.9820\n",
            "Epoch 25/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0058 - acc: 0.9984 - val_loss: 0.1205 - val_acc: 0.9795\n",
            "Epoch 26/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0100 - acc: 0.9971 - val_loss: 0.1005 - val_acc: 0.9825\n",
            "Epoch 27/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0075 - acc: 0.9978 - val_loss: 0.1044 - val_acc: 0.9802\n",
            "Epoch 28/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0061 - acc: 0.9982 - val_loss: 0.1020 - val_acc: 0.9832\n",
            "Epoch 29/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0049 - acc: 0.9984 - val_loss: 0.1077 - val_acc: 0.9818\n",
            "Epoch 30/80\n",
            "54000/54000 [==============================] - 3s 59us/sample - loss: 0.0078 - acc: 0.9979 - val_loss: 0.0922 - val_acc: 0.9835\n",
            "Epoch 31/80\n",
            "54000/54000 [==============================] - 3s 59us/sample - loss: 0.0044 - acc: 0.9988 - val_loss: 0.0992 - val_acc: 0.9842\n",
            "Epoch 32/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0063 - acc: 0.9981 - val_loss: 0.1133 - val_acc: 0.9828\n",
            "Epoch 33/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0067 - acc: 0.9980 - val_loss: 0.1241 - val_acc: 0.9780\n",
            "Epoch 34/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0055 - acc: 0.9983 - val_loss: 0.0956 - val_acc: 0.9837\n",
            "Epoch 35/80\n",
            "54000/54000 [==============================] - 3s 59us/sample - loss: 0.0027 - acc: 0.9991 - val_loss: 0.0991 - val_acc: 0.9843\n",
            "Epoch 36/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0039 - acc: 0.9989 - val_loss: 0.1003 - val_acc: 0.9830\n",
            "Epoch 37/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0055 - acc: 0.9987 - val_loss: 0.1102 - val_acc: 0.9817\n",
            "Epoch 38/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0066 - acc: 0.9981 - val_loss: 0.0994 - val_acc: 0.9828\n",
            "Epoch 39/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0037 - acc: 0.9989 - val_loss: 0.1223 - val_acc: 0.9813\n",
            "Epoch 40/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0073 - acc: 0.9978 - val_loss: 0.1202 - val_acc: 0.9793\n",
            "Epoch 41/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0057 - acc: 0.9985 - val_loss: 0.1284 - val_acc: 0.9797\n",
            "Epoch 42/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0024 - acc: 0.9994 - val_loss: 0.1525 - val_acc: 0.9793\n",
            "Epoch 43/80\n",
            "54000/54000 [==============================] - 3s 57us/sample - loss: 0.0095 - acc: 0.9977 - val_loss: 0.1156 - val_acc: 0.9823\n",
            "Epoch 44/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0032 - acc: 0.9991 - val_loss: 0.1249 - val_acc: 0.9825\n",
            "Epoch 45/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0057 - acc: 0.9988 - val_loss: 0.1270 - val_acc: 0.9822\n",
            "Epoch 46/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0049 - acc: 0.9985 - val_loss: 0.1276 - val_acc: 0.9835\n",
            "Epoch 47/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0056 - acc: 0.9986 - val_loss: 0.1372 - val_acc: 0.9783\n",
            "Epoch 48/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0049 - acc: 0.9986 - val_loss: 0.1438 - val_acc: 0.9783\n",
            "Epoch 49/80\n",
            "54000/54000 [==============================] - 3s 59us/sample - loss: 0.0058 - acc: 0.9983 - val_loss: 0.1246 - val_acc: 0.9832\n",
            "Epoch 50/80\n",
            "54000/54000 [==============================] - 3s 59us/sample - loss: 0.0076 - acc: 0.9984 - val_loss: 0.1125 - val_acc: 0.9837\n",
            "Epoch 51/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0014 - acc: 0.9995 - val_loss: 0.1011 - val_acc: 0.9845\n",
            "Epoch 52/80\n",
            "54000/54000 [==============================] - 3s 59us/sample - loss: 0.0010 - acc: 0.9998 - val_loss: 0.1352 - val_acc: 0.9798\n",
            "Epoch 53/80\n",
            "54000/54000 [==============================] - 3s 59us/sample - loss: 0.0068 - acc: 0.9982 - val_loss: 0.1318 - val_acc: 0.9822\n",
            "Epoch 54/80\n",
            "54000/54000 [==============================] - 3s 59us/sample - loss: 0.0070 - acc: 0.9984 - val_loss: 0.1232 - val_acc: 0.9815\n",
            "Epoch 55/80\n",
            "54000/54000 [==============================] - 3s 59us/sample - loss: 0.0024 - acc: 0.9992 - val_loss: 0.1293 - val_acc: 0.9825\n",
            "Epoch 56/80\n",
            "54000/54000 [==============================] - 3s 59us/sample - loss: 0.0048 - acc: 0.9987 - val_loss: 0.1665 - val_acc: 0.9797\n",
            "Epoch 57/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0059 - acc: 0.9981 - val_loss: 0.1360 - val_acc: 0.9840\n",
            "Epoch 58/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0029 - acc: 0.9991 - val_loss: 0.1213 - val_acc: 0.9853\n",
            "Epoch 59/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0024 - acc: 0.9995 - val_loss: 0.1176 - val_acc: 0.9852\n",
            "Epoch 60/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0063 - acc: 0.9986 - val_loss: 0.1413 - val_acc: 0.9812\n",
            "Epoch 61/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0043 - acc: 0.9990 - val_loss: 0.1198 - val_acc: 0.9837\n",
            "Epoch 62/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0034 - acc: 0.9991 - val_loss: 0.1298 - val_acc: 0.9827\n",
            "Epoch 63/80\n",
            "54000/54000 [==============================] - 3s 57us/sample - loss: 0.0036 - acc: 0.9990 - val_loss: 0.1283 - val_acc: 0.9825\n",
            "Epoch 64/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0051 - acc: 0.9989 - val_loss: 0.1233 - val_acc: 0.9835\n",
            "Epoch 65/80\n",
            "54000/54000 [==============================] - 3s 57us/sample - loss: 0.0059 - acc: 0.9986 - val_loss: 0.1424 - val_acc: 0.9822\n",
            "Epoch 66/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0050 - acc: 0.9989 - val_loss: 0.1225 - val_acc: 0.9847\n",
            "Epoch 67/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0047 - acc: 0.9987 - val_loss: 0.1540 - val_acc: 0.9808\n",
            "Epoch 68/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0046 - acc: 0.9991 - val_loss: 0.1413 - val_acc: 0.9825\n",
            "Epoch 69/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0033 - acc: 0.9991 - val_loss: 0.1227 - val_acc: 0.9840\n",
            "Epoch 70/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0020 - acc: 0.9994 - val_loss: 0.1269 - val_acc: 0.9838\n",
            "Epoch 71/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0030 - acc: 0.9993 - val_loss: 0.1308 - val_acc: 0.9837\n",
            "Epoch 72/80\n",
            "54000/54000 [==============================] - 3s 57us/sample - loss: 9.2198e-04 - acc: 0.9996 - val_loss: 0.1342 - val_acc: 0.9837\n",
            "Epoch 73/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0088 - acc: 0.9984 - val_loss: 0.1387 - val_acc: 0.9803\n",
            "Epoch 74/80\n",
            "54000/54000 [==============================] - 3s 59us/sample - loss: 0.0045 - acc: 0.9989 - val_loss: 0.1197 - val_acc: 0.9837\n",
            "Epoch 75/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0019 - acc: 0.9996 - val_loss: 0.1137 - val_acc: 0.9848\n",
            "Epoch 76/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0042 - acc: 0.9991 - val_loss: 0.1224 - val_acc: 0.9850\n",
            "Epoch 77/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0014 - acc: 0.9995 - val_loss: 0.1299 - val_acc: 0.9845\n",
            "Epoch 78/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0014 - acc: 0.9997 - val_loss: 0.1342 - val_acc: 0.9822\n",
            "Epoch 79/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.0073 - acc: 0.9982 - val_loss: 0.1543 - val_acc: 0.9827\n",
            "Epoch 80/80\n",
            "54000/54000 [==============================] - 3s 57us/sample - loss: 0.0061 - acc: 0.9986 - val_loss: 0.1236 - val_acc: 0.9860\n",
            "++++++++++++++++++++++++++++ 7 of 10 ++++++++++++++++++++++++++++++\n",
            "Dense [800, 800], validation split 0.1, epochs 80, batch_size 100\n",
            "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_16 (Dense)             (None, 800)               628000    \n",
            "_________________________________________________________________\n",
            "dense_17 (Dense)             (None, 800)               640800    \n",
            "_________________________________________________________________\n",
            "dense_18 (Dense)             (None, 10)                8010      \n",
            "=================================================================\n",
            "Total params: 1,276,810\n",
            "Trainable params: 1,276,810\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n",
            "Train on 54000 samples, validate on 6000 samples\n",
            "Epoch 1/80\n",
            "54000/54000 [==============================] - 4s 74us/sample - loss: 0.2010 - acc: 0.9384 - val_loss: 0.1183 - val_acc: 0.9658\n",
            "Epoch 2/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0801 - acc: 0.9749 - val_loss: 0.0807 - val_acc: 0.9770\n",
            "Epoch 3/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0493 - acc: 0.9845 - val_loss: 0.0783 - val_acc: 0.9755\n",
            "Epoch 4/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0356 - acc: 0.9890 - val_loss: 0.0717 - val_acc: 0.9803\n",
            "Epoch 5/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0278 - acc: 0.9908 - val_loss: 0.0739 - val_acc: 0.9800\n",
            "Epoch 6/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0272 - acc: 0.9908 - val_loss: 0.0856 - val_acc: 0.9783\n",
            "Epoch 7/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0189 - acc: 0.9941 - val_loss: 0.0820 - val_acc: 0.9798\n",
            "Epoch 8/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0196 - acc: 0.9937 - val_loss: 0.0756 - val_acc: 0.9787\n",
            "Epoch 9/80\n",
            "54000/54000 [==============================] - 4s 71us/sample - loss: 0.0159 - acc: 0.9949 - val_loss: 0.0747 - val_acc: 0.9825\n",
            "Epoch 10/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0135 - acc: 0.9956 - val_loss: 0.0779 - val_acc: 0.9787\n",
            "Epoch 11/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0157 - acc: 0.9948 - val_loss: 0.0651 - val_acc: 0.9855\n",
            "Epoch 12/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0133 - acc: 0.9962 - val_loss: 0.0793 - val_acc: 0.9823\n",
            "Epoch 13/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0129 - acc: 0.9958 - val_loss: 0.0909 - val_acc: 0.9810\n",
            "Epoch 14/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0087 - acc: 0.9969 - val_loss: 0.0936 - val_acc: 0.9808\n",
            "Epoch 15/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0131 - acc: 0.9959 - val_loss: 0.0955 - val_acc: 0.9813\n",
            "Epoch 16/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0083 - acc: 0.9973 - val_loss: 0.0831 - val_acc: 0.9837\n",
            "Epoch 17/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0073 - acc: 0.9980 - val_loss: 0.1034 - val_acc: 0.9800\n",
            "Epoch 18/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0139 - acc: 0.9962 - val_loss: 0.0867 - val_acc: 0.9813\n",
            "Epoch 19/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0085 - acc: 0.9972 - val_loss: 0.0873 - val_acc: 0.9853\n",
            "Epoch 20/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0045 - acc: 0.9987 - val_loss: 0.1271 - val_acc: 0.9773\n",
            "Epoch 21/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0141 - acc: 0.9957 - val_loss: 0.0959 - val_acc: 0.9815\n",
            "Epoch 22/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0067 - acc: 0.9981 - val_loss: 0.0906 - val_acc: 0.9843\n",
            "Epoch 23/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0056 - acc: 0.9981 - val_loss: 0.1057 - val_acc: 0.9820\n",
            "Epoch 24/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0102 - acc: 0.9973 - val_loss: 0.1005 - val_acc: 0.9822\n",
            "Epoch 25/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0083 - acc: 0.9976 - val_loss: 0.1123 - val_acc: 0.9823\n",
            "Epoch 26/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0070 - acc: 0.9981 - val_loss: 0.1096 - val_acc: 0.9842\n",
            "Epoch 27/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0107 - acc: 0.9971 - val_loss: 0.0911 - val_acc: 0.9827\n",
            "Epoch 28/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0061 - acc: 0.9985 - val_loss: 0.0968 - val_acc: 0.9827\n",
            "Epoch 29/80\n",
            "54000/54000 [==============================] - 4s 71us/sample - loss: 0.0029 - acc: 0.9991 - val_loss: 0.0943 - val_acc: 0.9843\n",
            "Epoch 30/80\n",
            "54000/54000 [==============================] - 4s 71us/sample - loss: 0.0054 - acc: 0.9985 - val_loss: 0.1088 - val_acc: 0.9838\n",
            "Epoch 31/80\n",
            "54000/54000 [==============================] - 4s 71us/sample - loss: 0.0084 - acc: 0.9980 - val_loss: 0.1190 - val_acc: 0.9818\n",
            "Epoch 32/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0137 - acc: 0.9965 - val_loss: 0.1078 - val_acc: 0.9828\n",
            "Epoch 33/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0051 - acc: 0.9986 - val_loss: 0.1015 - val_acc: 0.9833\n",
            "Epoch 34/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0061 - acc: 0.9982 - val_loss: 0.1109 - val_acc: 0.9843\n",
            "Epoch 35/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0054 - acc: 0.9986 - val_loss: 0.0986 - val_acc: 0.9857\n",
            "Epoch 36/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0063 - acc: 0.9984 - val_loss: 0.1062 - val_acc: 0.9842\n",
            "Epoch 37/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0064 - acc: 0.9983 - val_loss: 0.1035 - val_acc: 0.9817\n",
            "Epoch 38/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0052 - acc: 0.9985 - val_loss: 0.1262 - val_acc: 0.9800\n",
            "Epoch 39/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0065 - acc: 0.9982 - val_loss: 0.0913 - val_acc: 0.9850\n",
            "Epoch 40/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0047 - acc: 0.9988 - val_loss: 0.0980 - val_acc: 0.9838\n",
            "Epoch 41/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0061 - acc: 0.9983 - val_loss: 0.1114 - val_acc: 0.9835\n",
            "Epoch 42/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0042 - acc: 0.9989 - val_loss: 0.1234 - val_acc: 0.9832\n",
            "Epoch 43/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0094 - acc: 0.9978 - val_loss: 0.1120 - val_acc: 0.9842\n",
            "Epoch 44/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0055 - acc: 0.9987 - val_loss: 0.1011 - val_acc: 0.9850\n",
            "Epoch 45/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0017 - acc: 0.9996 - val_loss: 0.0993 - val_acc: 0.9858\n",
            "Epoch 46/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0048 - acc: 0.9989 - val_loss: 0.1234 - val_acc: 0.9810\n",
            "Epoch 47/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0085 - acc: 0.9983 - val_loss: 0.1281 - val_acc: 0.9838\n",
            "Epoch 48/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0094 - acc: 0.9979 - val_loss: 0.1235 - val_acc: 0.9820\n",
            "Epoch 49/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0019 - acc: 0.9995 - val_loss: 0.1076 - val_acc: 0.9867\n",
            "Epoch 50/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0025 - acc: 0.9994 - val_loss: 0.1156 - val_acc: 0.9848\n",
            "Epoch 51/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0023 - acc: 0.9993 - val_loss: 0.1025 - val_acc: 0.9870\n",
            "Epoch 52/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0096 - acc: 0.9976 - val_loss: 0.1244 - val_acc: 0.9835\n",
            "Epoch 53/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0037 - acc: 0.9991 - val_loss: 0.1227 - val_acc: 0.9827\n",
            "Epoch 54/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0050 - acc: 0.9987 - val_loss: 0.1134 - val_acc: 0.9850\n",
            "Epoch 55/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0069 - acc: 0.9986 - val_loss: 0.1325 - val_acc: 0.9830\n",
            "Epoch 56/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0089 - acc: 0.9980 - val_loss: 0.1379 - val_acc: 0.9837\n",
            "Epoch 57/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0050 - acc: 0.9988 - val_loss: 0.1413 - val_acc: 0.9827\n",
            "Epoch 58/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0041 - acc: 0.9991 - val_loss: 0.1143 - val_acc: 0.9838\n",
            "Epoch 59/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0042 - acc: 0.9992 - val_loss: 0.1403 - val_acc: 0.9815\n",
            "Epoch 60/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0123 - acc: 0.9978 - val_loss: 0.1252 - val_acc: 0.9835\n",
            "Epoch 61/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0022 - acc: 0.9994 - val_loss: 0.1167 - val_acc: 0.9848\n",
            "Epoch 62/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0027 - acc: 0.9994 - val_loss: 0.1232 - val_acc: 0.9852\n",
            "Epoch 63/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0087 - acc: 0.9984 - val_loss: 0.1359 - val_acc: 0.9843\n",
            "Epoch 64/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0033 - acc: 0.9991 - val_loss: 0.1311 - val_acc: 0.9835\n",
            "Epoch 65/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0035 - acc: 0.9993 - val_loss: 0.1244 - val_acc: 0.9847\n",
            "Epoch 66/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0082 - acc: 0.9983 - val_loss: 0.1516 - val_acc: 0.9828\n",
            "Epoch 67/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0062 - acc: 0.9986 - val_loss: 0.1347 - val_acc: 0.9842\n",
            "Epoch 68/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0057 - acc: 0.9989 - val_loss: 0.1241 - val_acc: 0.9853\n",
            "Epoch 69/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0036 - acc: 0.9994 - val_loss: 0.1384 - val_acc: 0.9855\n",
            "Epoch 70/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0044 - acc: 0.9994 - val_loss: 0.1387 - val_acc: 0.9852\n",
            "Epoch 71/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0066 - acc: 0.9986 - val_loss: 0.1523 - val_acc: 0.9823\n",
            "Epoch 72/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0060 - acc: 0.9988 - val_loss: 0.1484 - val_acc: 0.9847\n",
            "Epoch 73/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0036 - acc: 0.9995 - val_loss: 0.1429 - val_acc: 0.9852\n",
            "Epoch 74/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0082 - acc: 0.9984 - val_loss: 0.1449 - val_acc: 0.9842\n",
            "Epoch 75/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0057 - acc: 0.9988 - val_loss: 0.1319 - val_acc: 0.9853\n",
            "Epoch 76/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0040 - acc: 0.9993 - val_loss: 0.1461 - val_acc: 0.9848\n",
            "Epoch 77/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0048 - acc: 0.9991 - val_loss: 0.1576 - val_acc: 0.9830\n",
            "Epoch 78/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 0.0067 - acc: 0.9990 - val_loss: 0.1542 - val_acc: 0.9827\n",
            "Epoch 79/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0015 - acc: 0.9997 - val_loss: 0.1473 - val_acc: 0.9848\n",
            "Epoch 80/80\n",
            "54000/54000 [==============================] - 4s 70us/sample - loss: 6.0744e-04 - acc: 0.9998 - val_loss: 0.1609 - val_acc: 0.9830\n",
            "++++++++++++++++++++++++++++ 8 of 10 ++++++++++++++++++++++++++++++\n",
            "Dense [800], validation split 0.1, epochs 80, batch_size 100\n",
            "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_19 (Dense)             (None, 800)               628000    \n",
            "_________________________________________________________________\n",
            "dense_20 (Dense)             (None, 10)                8010      \n",
            "=================================================================\n",
            "Total params: 636,010\n",
            "Trainable params: 636,010\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n",
            "Train on 54000 samples, validate on 6000 samples\n",
            "Epoch 1/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 0.2457 - acc: 0.9292 - val_loss: 0.1050 - val_acc: 0.9718\n",
            "Epoch 2/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0958 - acc: 0.9714 - val_loss: 0.0929 - val_acc: 0.9733\n",
            "Epoch 3/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0604 - acc: 0.9817 - val_loss: 0.0725 - val_acc: 0.9787\n",
            "Epoch 4/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 0.0415 - acc: 0.9877 - val_loss: 0.0591 - val_acc: 0.9835\n",
            "Epoch 5/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 0.0282 - acc: 0.9921 - val_loss: 0.0649 - val_acc: 0.9822\n",
            "Epoch 6/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 0.0221 - acc: 0.9939 - val_loss: 0.0668 - val_acc: 0.9818\n",
            "Epoch 7/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 0.0151 - acc: 0.9956 - val_loss: 0.0656 - val_acc: 0.9813\n",
            "Epoch 8/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 0.0124 - acc: 0.9967 - val_loss: 0.0615 - val_acc: 0.9817\n",
            "Epoch 9/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0111 - acc: 0.9969 - val_loss: 0.0633 - val_acc: 0.9827\n",
            "Epoch 10/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 0.0087 - acc: 0.9976 - val_loss: 0.0709 - val_acc: 0.9827\n",
            "Epoch 11/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 0.0081 - acc: 0.9974 - val_loss: 0.0837 - val_acc: 0.9780\n",
            "Epoch 12/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0074 - acc: 0.9977 - val_loss: 0.0675 - val_acc: 0.9828\n",
            "Epoch 13/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0040 - acc: 0.9989 - val_loss: 0.0704 - val_acc: 0.9838\n",
            "Epoch 14/80\n",
            "54000/54000 [==============================] - 3s 54us/sample - loss: 0.0061 - acc: 0.9983 - val_loss: 0.0806 - val_acc: 0.9807\n",
            "Epoch 15/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0083 - acc: 0.9972 - val_loss: 0.0807 - val_acc: 0.9812\n",
            "Epoch 16/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0063 - acc: 0.9980 - val_loss: 0.0770 - val_acc: 0.9827\n",
            "Epoch 17/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0031 - acc: 0.9989 - val_loss: 0.0712 - val_acc: 0.9830\n",
            "Epoch 18/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0015 - acc: 0.9997 - val_loss: 0.0705 - val_acc: 0.9855\n",
            "Epoch 19/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 3.4729e-04 - acc: 1.0000 - val_loss: 0.0669 - val_acc: 0.9860\n",
            "Epoch 20/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 1.3866e-04 - acc: 1.0000 - val_loss: 0.0646 - val_acc: 0.9867\n",
            "Epoch 21/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 9.2897e-05 - acc: 1.0000 - val_loss: 0.0655 - val_acc: 0.9867\n",
            "Epoch 22/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 7.6335e-05 - acc: 1.0000 - val_loss: 0.0667 - val_acc: 0.9872\n",
            "Epoch 23/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 6.4256e-05 - acc: 1.0000 - val_loss: 0.0671 - val_acc: 0.9867\n",
            "Epoch 24/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 5.4177e-05 - acc: 1.0000 - val_loss: 0.0681 - val_acc: 0.9870\n",
            "Epoch 25/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 4.3935e-05 - acc: 1.0000 - val_loss: 0.0691 - val_acc: 0.9868\n",
            "Epoch 26/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 3.6913e-05 - acc: 1.0000 - val_loss: 0.0694 - val_acc: 0.9865\n",
            "Epoch 27/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0248 - acc: 0.9928 - val_loss: 0.1004 - val_acc: 0.9803\n",
            "Epoch 28/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0073 - acc: 0.9974 - val_loss: 0.0816 - val_acc: 0.9840\n",
            "Epoch 29/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0021 - acc: 0.9994 - val_loss: 0.0743 - val_acc: 0.9847\n",
            "Epoch 30/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 3.1099e-04 - acc: 1.0000 - val_loss: 0.0678 - val_acc: 0.9862\n",
            "Epoch 31/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 1.6829e-04 - acc: 1.0000 - val_loss: 0.0683 - val_acc: 0.9867\n",
            "Epoch 32/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 9.0146e-05 - acc: 1.0000 - val_loss: 0.0690 - val_acc: 0.9867\n",
            "Epoch 33/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 6.9449e-05 - acc: 1.0000 - val_loss: 0.0690 - val_acc: 0.9873\n",
            "Epoch 34/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 5.6792e-05 - acc: 1.0000 - val_loss: 0.0701 - val_acc: 0.9870\n",
            "Epoch 35/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 4.6061e-05 - acc: 1.0000 - val_loss: 0.0703 - val_acc: 0.9872\n",
            "Epoch 36/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 3.9589e-05 - acc: 1.0000 - val_loss: 0.0708 - val_acc: 0.9877\n",
            "Epoch 37/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 3.2315e-05 - acc: 1.0000 - val_loss: 0.0708 - val_acc: 0.9872\n",
            "Epoch 38/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 2.6423e-05 - acc: 1.0000 - val_loss: 0.0707 - val_acc: 0.9870\n",
            "Epoch 39/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 2.1673e-05 - acc: 1.0000 - val_loss: 0.0718 - val_acc: 0.9863\n",
            "Epoch 40/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 1.7605e-05 - acc: 1.0000 - val_loss: 0.0717 - val_acc: 0.9877\n",
            "Epoch 41/80\n",
            "54000/54000 [==============================] - 3s 54us/sample - loss: 0.0063 - acc: 0.9988 - val_loss: 0.2227 - val_acc: 0.9615\n",
            "Epoch 42/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0264 - acc: 0.9932 - val_loss: 0.0799 - val_acc: 0.9830\n",
            "Epoch 43/80\n",
            "54000/54000 [==============================] - 3s 54us/sample - loss: 0.0017 - acc: 0.9995 - val_loss: 0.0830 - val_acc: 0.9845\n",
            "Epoch 44/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 4.8176e-04 - acc: 0.9999 - val_loss: 0.0769 - val_acc: 0.9852\n",
            "Epoch 45/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 1.9824e-04 - acc: 1.0000 - val_loss: 0.0762 - val_acc: 0.9865\n",
            "Epoch 46/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 6.9279e-05 - acc: 1.0000 - val_loss: 0.0766 - val_acc: 0.9860\n",
            "Epoch 47/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 5.4225e-05 - acc: 1.0000 - val_loss: 0.0770 - val_acc: 0.9858\n",
            "Epoch 48/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 4.2279e-05 - acc: 1.0000 - val_loss: 0.0774 - val_acc: 0.9858\n",
            "Epoch 49/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 3.3472e-05 - acc: 1.0000 - val_loss: 0.0771 - val_acc: 0.9860\n",
            "Epoch 50/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 2.7906e-05 - acc: 1.0000 - val_loss: 0.0778 - val_acc: 0.9860\n",
            "Epoch 51/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 2.2541e-05 - acc: 1.0000 - val_loss: 0.0783 - val_acc: 0.9860\n",
            "Epoch 52/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 1.8623e-05 - acc: 1.0000 - val_loss: 0.0796 - val_acc: 0.9858\n",
            "Epoch 53/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 1.5086e-05 - acc: 1.0000 - val_loss: 0.0799 - val_acc: 0.9863\n",
            "Epoch 54/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 1.2018e-05 - acc: 1.0000 - val_loss: 0.0795 - val_acc: 0.9860\n",
            "Epoch 55/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 9.9622e-06 - acc: 1.0000 - val_loss: 0.0792 - val_acc: 0.9865\n",
            "Epoch 56/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 7.8905e-06 - acc: 1.0000 - val_loss: 0.0803 - val_acc: 0.9860\n",
            "Epoch 57/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 6.3744e-06 - acc: 1.0000 - val_loss: 0.0815 - val_acc: 0.9865\n",
            "Epoch 58/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 5.1296e-06 - acc: 1.0000 - val_loss: 0.0823 - val_acc: 0.9865\n",
            "Epoch 59/80\n",
            "54000/54000 [==============================] - 3s 54us/sample - loss: 4.1156e-06 - acc: 1.0000 - val_loss: 0.0811 - val_acc: 0.9862\n",
            "Epoch 60/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 0.0095 - acc: 0.9980 - val_loss: 0.1531 - val_acc: 0.9748\n",
            "Epoch 61/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 0.0167 - acc: 0.9948 - val_loss: 0.0973 - val_acc: 0.9833\n",
            "Epoch 62/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0017 - acc: 0.9995 - val_loss: 0.0871 - val_acc: 0.9845\n",
            "Epoch 63/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 5.2449e-04 - acc: 0.9999 - val_loss: 0.0885 - val_acc: 0.9848\n",
            "Epoch 64/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 8.7720e-04 - acc: 0.9996 - val_loss: 0.0893 - val_acc: 0.9850\n",
            "Epoch 65/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0055 - acc: 0.9983 - val_loss: 0.1236 - val_acc: 0.9810\n",
            "Epoch 66/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0039 - acc: 0.9986 - val_loss: 0.0988 - val_acc: 0.9843\n",
            "Epoch 67/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 0.0014 - acc: 0.9996 - val_loss: 0.0974 - val_acc: 0.9842\n",
            "Epoch 68/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 7.7300e-04 - acc: 0.9998 - val_loss: 0.1052 - val_acc: 0.9838\n",
            "Epoch 69/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 0.0074 - acc: 0.9979 - val_loss: 0.0994 - val_acc: 0.9827\n",
            "Epoch 70/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0024 - acc: 0.9993 - val_loss: 0.0886 - val_acc: 0.9860\n",
            "Epoch 71/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 0.0022 - acc: 0.9993 - val_loss: 0.0901 - val_acc: 0.9840\n",
            "Epoch 72/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0015 - acc: 0.9997 - val_loss: 0.0885 - val_acc: 0.9868\n",
            "Epoch 73/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0018 - acc: 0.9994 - val_loss: 0.1054 - val_acc: 0.9830\n",
            "Epoch 74/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0046 - acc: 0.9984 - val_loss: 0.1014 - val_acc: 0.9840\n",
            "Epoch 75/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0035 - acc: 0.9989 - val_loss: 0.1098 - val_acc: 0.9812\n",
            "Epoch 76/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 0.0029 - acc: 0.9990 - val_loss: 0.0863 - val_acc: 0.9867\n",
            "Epoch 77/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 1.7192e-04 - acc: 0.9999 - val_loss: 0.0871 - val_acc: 0.9865\n",
            "Epoch 78/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 2.5593e-05 - acc: 1.0000 - val_loss: 0.0845 - val_acc: 0.9872\n",
            "Epoch 79/80\n",
            "54000/54000 [==============================] - 3s 53us/sample - loss: 1.3263e-05 - acc: 1.0000 - val_loss: 0.0841 - val_acc: 0.9872\n",
            "Epoch 80/80\n",
            "54000/54000 [==============================] - 3s 52us/sample - loss: 1.0498e-05 - acc: 1.0000 - val_loss: 0.0837 - val_acc: 0.9872\n",
            "++++++++++++++++++++++++++++ 9 of 10 ++++++++++++++++++++++++++++++\n",
            "Dense [1000], validation split 0.1, epochs 80, batch_size 100\n",
            "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_21 (Dense)             (None, 1000)              785000    \n",
            "_________________________________________________________________\n",
            "dense_22 (Dense)             (None, 10)                10010     \n",
            "=================================================================\n",
            "Total params: 795,010\n",
            "Trainable params: 795,010\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n",
            "Train on 54000 samples, validate on 6000 samples\n",
            "Epoch 1/80\n",
            "54000/54000 [==============================] - 3s 62us/sample - loss: 0.2361 - acc: 0.9315 - val_loss: 0.0981 - val_acc: 0.9727\n",
            "Epoch 2/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0909 - acc: 0.9730 - val_loss: 0.0724 - val_acc: 0.9800\n",
            "Epoch 3/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0568 - acc: 0.9834 - val_loss: 0.0870 - val_acc: 0.9732\n",
            "Epoch 4/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0393 - acc: 0.9880 - val_loss: 0.0616 - val_acc: 0.9827\n",
            "Epoch 5/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0278 - acc: 0.9915 - val_loss: 0.0663 - val_acc: 0.9812\n",
            "Epoch 6/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0189 - acc: 0.9947 - val_loss: 0.0619 - val_acc: 0.9843\n",
            "Epoch 7/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0143 - acc: 0.9959 - val_loss: 0.0556 - val_acc: 0.9838\n",
            "Epoch 8/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0120 - acc: 0.9964 - val_loss: 0.0740 - val_acc: 0.9798\n",
            "Epoch 9/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0101 - acc: 0.9971 - val_loss: 0.0671 - val_acc: 0.9828\n",
            "Epoch 10/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0070 - acc: 0.9981 - val_loss: 0.0693 - val_acc: 0.9828\n",
            "Epoch 11/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0091 - acc: 0.9970 - val_loss: 0.0829 - val_acc: 0.9815\n",
            "Epoch 12/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0091 - acc: 0.9971 - val_loss: 0.0669 - val_acc: 0.9837\n",
            "Epoch 13/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0049 - acc: 0.9987 - val_loss: 0.0851 - val_acc: 0.9817\n",
            "Epoch 14/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0085 - acc: 0.9971 - val_loss: 0.0735 - val_acc: 0.9833\n",
            "Epoch 15/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0045 - acc: 0.9986 - val_loss: 0.0817 - val_acc: 0.9832\n",
            "Epoch 16/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0011 - acc: 0.9999 - val_loss: 0.0691 - val_acc: 0.9863\n",
            "Epoch 17/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 2.6435e-04 - acc: 1.0000 - val_loss: 0.0680 - val_acc: 0.9862\n",
            "Epoch 18/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 1.6264e-04 - acc: 1.0000 - val_loss: 0.0679 - val_acc: 0.9867\n",
            "Epoch 19/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 9.8267e-05 - acc: 1.0000 - val_loss: 0.0693 - val_acc: 0.9868\n",
            "Epoch 20/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 7.9603e-05 - acc: 1.0000 - val_loss: 0.0696 - val_acc: 0.9862\n",
            "Epoch 21/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 6.4869e-05 - acc: 1.0000 - val_loss: 0.0704 - val_acc: 0.9862\n",
            "Epoch 22/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 5.3513e-05 - acc: 1.0000 - val_loss: 0.0703 - val_acc: 0.9872\n",
            "Epoch 23/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 4.5214e-05 - acc: 1.0000 - val_loss: 0.0708 - val_acc: 0.9863\n",
            "Epoch 24/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 4.0215e-05 - acc: 1.0000 - val_loss: 0.0739 - val_acc: 0.9857\n",
            "Epoch 25/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0302 - acc: 0.9906 - val_loss: 0.0866 - val_acc: 0.9817\n",
            "Epoch 26/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0040 - acc: 0.9987 - val_loss: 0.0809 - val_acc: 0.9833\n",
            "Epoch 27/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0020 - acc: 0.9995 - val_loss: 0.0686 - val_acc: 0.9863\n",
            "Epoch 28/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0016 - acc: 0.9998 - val_loss: 0.0739 - val_acc: 0.9848\n",
            "Epoch 29/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0043 - acc: 0.9986 - val_loss: 0.1020 - val_acc: 0.9795\n",
            "Epoch 30/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0082 - acc: 0.9974 - val_loss: 0.0966 - val_acc: 0.9820\n",
            "Epoch 31/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0035 - acc: 0.9988 - val_loss: 0.0843 - val_acc: 0.9847\n",
            "Epoch 32/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0050 - acc: 0.9985 - val_loss: 0.0946 - val_acc: 0.9830\n",
            "Epoch 33/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0037 - acc: 0.9988 - val_loss: 0.0850 - val_acc: 0.9858\n",
            "Epoch 34/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0014 - acc: 0.9995 - val_loss: 0.0877 - val_acc: 0.9847\n",
            "Epoch 35/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 3.4230e-04 - acc: 1.0000 - val_loss: 0.0813 - val_acc: 0.9865\n",
            "Epoch 36/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 5.7443e-05 - acc: 1.0000 - val_loss: 0.0809 - val_acc: 0.9862\n",
            "Epoch 37/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 3.5278e-05 - acc: 1.0000 - val_loss: 0.0813 - val_acc: 0.9863\n",
            "Epoch 38/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 2.7539e-05 - acc: 1.0000 - val_loss: 0.0815 - val_acc: 0.9862\n",
            "Epoch 39/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 2.2654e-05 - acc: 1.0000 - val_loss: 0.0810 - val_acc: 0.9863\n",
            "Epoch 40/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 1.8566e-05 - acc: 1.0000 - val_loss: 0.0811 - val_acc: 0.9863\n",
            "Epoch 41/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 1.5300e-05 - acc: 1.0000 - val_loss: 0.0816 - val_acc: 0.9862\n",
            "Epoch 42/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 1.2687e-05 - acc: 1.0000 - val_loss: 0.0818 - val_acc: 0.9868\n",
            "Epoch 43/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 1.0718e-05 - acc: 1.0000 - val_loss: 0.0825 - val_acc: 0.9865\n",
            "Epoch 44/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 8.7895e-06 - acc: 1.0000 - val_loss: 0.0820 - val_acc: 0.9865\n",
            "Epoch 45/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 7.3092e-06 - acc: 1.0000 - val_loss: 0.0826 - val_acc: 0.9865\n",
            "Epoch 46/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 6.0903e-06 - acc: 1.0000 - val_loss: 0.0843 - val_acc: 0.9867\n",
            "Epoch 47/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 4.9337e-06 - acc: 1.0000 - val_loss: 0.0844 - val_acc: 0.9865\n",
            "Epoch 48/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 4.1051e-06 - acc: 1.0000 - val_loss: 0.0848 - val_acc: 0.9863\n",
            "Epoch 49/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0295 - acc: 0.9932 - val_loss: 0.0961 - val_acc: 0.9832\n",
            "Epoch 50/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0023 - acc: 0.9993 - val_loss: 0.0956 - val_acc: 0.9840\n",
            "Epoch 51/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 5.9864e-04 - acc: 0.9999 - val_loss: 0.0822 - val_acc: 0.9847\n",
            "Epoch 52/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 7.4880e-04 - acc: 0.9999 - val_loss: 0.0866 - val_acc: 0.9837\n",
            "Epoch 53/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0023 - acc: 0.9993 - val_loss: 0.1146 - val_acc: 0.9818\n",
            "Epoch 54/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0074 - acc: 0.9977 - val_loss: 0.1138 - val_acc: 0.9822\n",
            "Epoch 55/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0026 - acc: 0.9990 - val_loss: 0.0848 - val_acc: 0.9858\n",
            "Epoch 56/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 0.0023 - acc: 0.9992 - val_loss: 0.0901 - val_acc: 0.9858\n",
            "Epoch 57/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0010 - acc: 0.9996 - val_loss: 0.1058 - val_acc: 0.9825\n",
            "Epoch 58/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0045 - acc: 0.9988 - val_loss: 0.0973 - val_acc: 0.9825\n",
            "Epoch 59/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0042 - acc: 0.9986 - val_loss: 0.1010 - val_acc: 0.9833\n",
            "Epoch 60/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 0.0021 - acc: 0.9994 - val_loss: 0.1090 - val_acc: 0.9840\n",
            "Epoch 61/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 4.3271e-04 - acc: 0.9999 - val_loss: 0.0955 - val_acc: 0.9837\n",
            "Epoch 62/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 2.4570e-05 - acc: 1.0000 - val_loss: 0.0910 - val_acc: 0.9852\n",
            "Epoch 63/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 1.3686e-05 - acc: 1.0000 - val_loss: 0.0915 - val_acc: 0.9850\n",
            "Epoch 64/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 1.0778e-05 - acc: 1.0000 - val_loss: 0.0917 - val_acc: 0.9852\n",
            "Epoch 65/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 8.6816e-06 - acc: 1.0000 - val_loss: 0.0920 - val_acc: 0.9853\n",
            "Epoch 66/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 7.2108e-06 - acc: 1.0000 - val_loss: 0.0918 - val_acc: 0.9855\n",
            "Epoch 67/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 5.9891e-06 - acc: 1.0000 - val_loss: 0.0925 - val_acc: 0.9852\n",
            "Epoch 68/80\n",
            "54000/54000 [==============================] - 3s 58us/sample - loss: 5.0320e-06 - acc: 1.0000 - val_loss: 0.0919 - val_acc: 0.9853\n",
            "Epoch 69/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 4.1762e-06 - acc: 1.0000 - val_loss: 0.0930 - val_acc: 0.9853\n",
            "Epoch 70/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 3.5200e-06 - acc: 1.0000 - val_loss: 0.0929 - val_acc: 0.9853\n",
            "Epoch 71/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 2.9340e-06 - acc: 1.0000 - val_loss: 0.0932 - val_acc: 0.9850\n",
            "Epoch 72/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 2.4388e-06 - acc: 1.0000 - val_loss: 0.0930 - val_acc: 0.9853\n",
            "Epoch 73/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 2.0625e-06 - acc: 1.0000 - val_loss: 0.0940 - val_acc: 0.9853\n",
            "Epoch 74/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 1.7117e-06 - acc: 1.0000 - val_loss: 0.0941 - val_acc: 0.9852\n",
            "Epoch 75/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 1.4386e-06 - acc: 1.0000 - val_loss: 0.0947 - val_acc: 0.9857\n",
            "Epoch 76/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 1.1965e-06 - acc: 1.0000 - val_loss: 0.0950 - val_acc: 0.9855\n",
            "Epoch 77/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 1.0085e-06 - acc: 1.0000 - val_loss: 0.0956 - val_acc: 0.9857\n",
            "Epoch 78/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 8.3801e-07 - acc: 1.0000 - val_loss: 0.0965 - val_acc: 0.9855\n",
            "Epoch 79/80\n",
            "54000/54000 [==============================] - 3s 55us/sample - loss: 7.0863e-07 - acc: 1.0000 - val_loss: 0.0960 - val_acc: 0.9858\n",
            "Epoch 80/80\n",
            "54000/54000 [==============================] - 3s 56us/sample - loss: 6.0198e-07 - acc: 1.0000 - val_loss: 0.0969 - val_acc: 0.9853\n",
            "++++++++++++++++++++++++++++ 10 of 10 ++++++++++++++++++++++++++++++\n",
            "Dense [2000], validation split 0.1, epochs 80, batch_size 100\n",
            "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_23 (Dense)             (None, 2000)              1570000   \n",
            "_________________________________________________________________\n",
            "dense_24 (Dense)             (None, 10)                20010     \n",
            "=================================================================\n",
            "Total params: 1,590,010\n",
            "Trainable params: 1,590,010\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n",
            "Train on 54000 samples, validate on 6000 samples\n",
            "Epoch 1/80\n",
            "54000/54000 [==============================] - 4s 73us/sample - loss: 0.2153 - acc: 0.9364 - val_loss: 0.1075 - val_acc: 0.9688\n",
            "Epoch 2/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 0.0802 - acc: 0.9759 - val_loss: 0.0763 - val_acc: 0.9782\n",
            "Epoch 3/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0504 - acc: 0.9841 - val_loss: 0.0716 - val_acc: 0.9795\n",
            "Epoch 4/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0330 - acc: 0.9898 - val_loss: 0.0745 - val_acc: 0.9800\n",
            "Epoch 5/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0231 - acc: 0.9934 - val_loss: 0.0690 - val_acc: 0.9813\n",
            "Epoch 6/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0186 - acc: 0.9944 - val_loss: 0.0678 - val_acc: 0.9822\n",
            "Epoch 7/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0160 - acc: 0.9949 - val_loss: 0.0721 - val_acc: 0.9833\n",
            "Epoch 8/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0103 - acc: 0.9969 - val_loss: 0.0839 - val_acc: 0.9788\n",
            "Epoch 9/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0120 - acc: 0.9959 - val_loss: 0.0811 - val_acc: 0.9808\n",
            "Epoch 10/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0120 - acc: 0.9959 - val_loss: 0.0767 - val_acc: 0.9812\n",
            "Epoch 11/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0076 - acc: 0.9975 - val_loss: 0.0770 - val_acc: 0.9847\n",
            "Epoch 12/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0092 - acc: 0.9970 - val_loss: 0.0786 - val_acc: 0.9842\n",
            "Epoch 13/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0039 - acc: 0.9989 - val_loss: 0.0748 - val_acc: 0.9855\n",
            "Epoch 14/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0116 - acc: 0.9966 - val_loss: 0.0750 - val_acc: 0.9840\n",
            "Epoch 15/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0040 - acc: 0.9991 - val_loss: 0.0859 - val_acc: 0.9800\n",
            "Epoch 16/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0079 - acc: 0.9976 - val_loss: 0.0842 - val_acc: 0.9840\n",
            "Epoch 17/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0072 - acc: 0.9977 - val_loss: 0.0955 - val_acc: 0.9812\n",
            "Epoch 18/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0047 - acc: 0.9983 - val_loss: 0.0828 - val_acc: 0.9845\n",
            "Epoch 19/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 0.0041 - acc: 0.9986 - val_loss: 0.1015 - val_acc: 0.9827\n",
            "Epoch 20/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 0.0076 - acc: 0.9975 - val_loss: 0.1057 - val_acc: 0.9808\n",
            "Epoch 21/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 0.0057 - acc: 0.9978 - val_loss: 0.0928 - val_acc: 0.9835\n",
            "Epoch 22/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 0.0016 - acc: 0.9996 - val_loss: 0.0854 - val_acc: 0.9857\n",
            "Epoch 23/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0038 - acc: 0.9989 - val_loss: 0.1058 - val_acc: 0.9823\n",
            "Epoch 24/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0083 - acc: 0.9974 - val_loss: 0.0851 - val_acc: 0.9845\n",
            "Epoch 25/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 0.0050 - acc: 0.9983 - val_loss: 0.0749 - val_acc: 0.9862\n",
            "Epoch 26/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 0.0035 - acc: 0.9989 - val_loss: 0.0907 - val_acc: 0.9830\n",
            "Epoch 27/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 0.0025 - acc: 0.9993 - val_loss: 0.0842 - val_acc: 0.9863\n",
            "Epoch 28/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0029 - acc: 0.9992 - val_loss: 0.0911 - val_acc: 0.9845\n",
            "Epoch 29/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 0.0093 - acc: 0.9974 - val_loss: 0.0965 - val_acc: 0.9830\n",
            "Epoch 30/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0052 - acc: 0.9984 - val_loss: 0.0925 - val_acc: 0.9843\n",
            "Epoch 31/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 7.8728e-04 - acc: 0.9998 - val_loss: 0.0762 - val_acc: 0.9880\n",
            "Epoch 32/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 1.0334e-04 - acc: 1.0000 - val_loss: 0.0786 - val_acc: 0.9872\n",
            "Epoch 33/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 2.2232e-05 - acc: 1.0000 - val_loss: 0.0773 - val_acc: 0.9882\n",
            "Epoch 34/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 1.5176e-05 - acc: 1.0000 - val_loss: 0.0774 - val_acc: 0.9880\n",
            "Epoch 35/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 1.2114e-05 - acc: 1.0000 - val_loss: 0.0779 - val_acc: 0.9880\n",
            "Epoch 36/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 9.9500e-06 - acc: 1.0000 - val_loss: 0.0777 - val_acc: 0.9880\n",
            "Epoch 37/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 8.1897e-06 - acc: 1.0000 - val_loss: 0.0778 - val_acc: 0.9878\n",
            "Epoch 38/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 6.8193e-06 - acc: 1.0000 - val_loss: 0.0781 - val_acc: 0.9877\n",
            "Epoch 39/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 5.6694e-06 - acc: 1.0000 - val_loss: 0.0783 - val_acc: 0.9878\n",
            "Epoch 40/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 4.6739e-06 - acc: 1.0000 - val_loss: 0.0784 - val_acc: 0.9878\n",
            "Epoch 41/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 3.8923e-06 - acc: 1.0000 - val_loss: 0.0791 - val_acc: 0.9880\n",
            "Epoch 42/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 3.1932e-06 - acc: 1.0000 - val_loss: 0.0796 - val_acc: 0.9882\n",
            "Epoch 43/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 2.6403e-06 - acc: 1.0000 - val_loss: 0.0798 - val_acc: 0.9882\n",
            "Epoch 44/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 2.1606e-06 - acc: 1.0000 - val_loss: 0.0802 - val_acc: 0.9883\n",
            "Epoch 45/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 1.7854e-06 - acc: 1.0000 - val_loss: 0.0804 - val_acc: 0.9880\n",
            "Epoch 46/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 1.4764e-06 - acc: 1.0000 - val_loss: 0.0816 - val_acc: 0.9883\n",
            "Epoch 47/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 1.2282e-06 - acc: 1.0000 - val_loss: 0.0817 - val_acc: 0.9880\n",
            "Epoch 48/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 1.0157e-06 - acc: 1.0000 - val_loss: 0.0817 - val_acc: 0.9880\n",
            "Epoch 49/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 8.3823e-07 - acc: 1.0000 - val_loss: 0.0826 - val_acc: 0.9883\n",
            "Epoch 50/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 6.9792e-07 - acc: 1.0000 - val_loss: 0.0828 - val_acc: 0.9883\n",
            "Epoch 51/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 6.4109e-07 - acc: 1.0000 - val_loss: 0.0833 - val_acc: 0.9873\n",
            "Epoch 52/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 2.0512e-06 - acc: 1.0000 - val_loss: 0.0951 - val_acc: 0.9858\n",
            "Epoch 53/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0375 - acc: 0.9919 - val_loss: 0.1269 - val_acc: 0.9813\n",
            "Epoch 54/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0037 - acc: 0.9990 - val_loss: 0.1080 - val_acc: 0.9852\n",
            "Epoch 55/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 0.0015 - acc: 0.9996 - val_loss: 0.1015 - val_acc: 0.9850\n",
            "Epoch 56/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 2.1240e-04 - acc: 0.9999 - val_loss: 0.0912 - val_acc: 0.9865\n",
            "Epoch 57/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 1.3845e-04 - acc: 1.0000 - val_loss: 0.0981 - val_acc: 0.9862\n",
            "Epoch 58/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0098 - acc: 0.9976 - val_loss: 0.1321 - val_acc: 0.9823\n",
            "Epoch 59/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0101 - acc: 0.9972 - val_loss: 0.1063 - val_acc: 0.9847\n",
            "Epoch 60/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0031 - acc: 0.9991 - val_loss: 0.1041 - val_acc: 0.9858\n",
            "Epoch 61/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0019 - acc: 0.9995 - val_loss: 0.1344 - val_acc: 0.9810\n",
            "Epoch 62/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0052 - acc: 0.9987 - val_loss: 0.0995 - val_acc: 0.9847\n",
            "Epoch 63/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0020 - acc: 0.9993 - val_loss: 0.0996 - val_acc: 0.9858\n",
            "Epoch 64/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0015 - acc: 0.9996 - val_loss: 0.1006 - val_acc: 0.9862\n",
            "Epoch 65/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0085 - acc: 0.9984 - val_loss: 0.1082 - val_acc: 0.9858\n",
            "Epoch 66/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 0.0023 - acc: 0.9994 - val_loss: 0.0968 - val_acc: 0.9865\n",
            "Epoch 67/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 7.8732e-04 - acc: 0.9998 - val_loss: 0.1044 - val_acc: 0.9855\n",
            "Epoch 68/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0084 - acc: 0.9981 - val_loss: 0.1386 - val_acc: 0.9825\n",
            "Epoch 69/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0052 - acc: 0.9987 - val_loss: 0.1250 - val_acc: 0.9838\n",
            "Epoch 70/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0052 - acc: 0.9986 - val_loss: 0.1209 - val_acc: 0.9833\n",
            "Epoch 71/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0027 - acc: 0.9992 - val_loss: 0.1141 - val_acc: 0.9852\n",
            "Epoch 72/80\n",
            "54000/54000 [==============================] - 4s 68us/sample - loss: 0.0017 - acc: 0.9996 - val_loss: 0.1237 - val_acc: 0.9843\n",
            "Epoch 73/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0011 - acc: 0.9997 - val_loss: 0.1419 - val_acc: 0.9830\n",
            "Epoch 74/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0022 - acc: 0.9994 - val_loss: 0.1259 - val_acc: 0.9843\n",
            "Epoch 75/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0037 - acc: 0.9989 - val_loss: 0.1304 - val_acc: 0.9807\n",
            "Epoch 76/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0063 - acc: 0.9985 - val_loss: 0.1337 - val_acc: 0.9833\n",
            "Epoch 77/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0036 - acc: 0.9991 - val_loss: 0.1205 - val_acc: 0.9830\n",
            "Epoch 78/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0022 - acc: 0.9995 - val_loss: 0.1076 - val_acc: 0.9863\n",
            "Epoch 79/80\n",
            "54000/54000 [==============================] - 4s 69us/sample - loss: 0.0016 - acc: 0.9995 - val_loss: 0.1336 - val_acc: 0.9825\n",
            "Epoch 80/80\n",
            "54000/54000 [==============================] - 4s 67us/sample - loss: 0.0025 - acc: 0.9992 - val_loss: 0.1449 - val_acc: 0.9830\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "T2k8erbGufpZ",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Результаты обучения"
      ]
    },
    {
      "metadata": {
        "id": "H77-7JxkuiEU",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f70532c6-259c-48c4-f28c-44520c9e1df6"
      },
      "cell_type": "code",
      "source": [
        "for res in results:\n",
        "  print(\"{} | {}\".format(res.history.history[\"val_acc\"][-1], res.config))"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.9700000286102295 | Dense [100], validation split 0.1, epochs 80, batch_size 100\n",
            "0.9853333234786987 | Dense [100, 200], validation split 0.1, epochs 80, batch_size 100\n",
            "0.9851666688919067 | Dense [200, 100], validation split 0.1, epochs 80, batch_size 100\n",
            "0.984666645526886 | Dense [400], validation split 0.1, epochs 80, batch_size 100\n",
            "0.9851666688919067 | Dense [400, 100], validation split 0.1, epochs 80, batch_size 100\n",
            "0.9860000014305115 | Dense [400, 400], validation split 0.1, epochs 80, batch_size 100\n",
            "0.9829999804496765 | Dense [800, 800], validation split 0.1, epochs 80, batch_size 100\n",
            "0.9871666431427002 | Dense [800], validation split 0.1, epochs 80, batch_size 100\n",
            "0.9853333234786987 | Dense [1000], validation split 0.1, epochs 80, batch_size 100\n",
            "0.9829999804496765 | Dense [2000], validation split 0.1, epochs 80, batch_size 100\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "HxM4uRkNvPWN",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Графики обучения"
      ]
    },
    {
      "metadata": {
        "id": "2ulDxnHMvR4Y",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1150
        },
        "outputId": "9c90949c-f12d-4e2d-90d4-d69923eb1653"
      },
      "cell_type": "code",
      "source": [
        "figure = plt.figure(figsize=(20, 20))\n",
        "\n",
        "i = 0\n",
        "cols = 3\n",
        "rows = math.ceil(len(results) / cols)\n",
        "\n",
        "for res in results:\n",
        "  history = res.history.history\n",
        "  \n",
        "  cell = plt.subplot2grid((rows, cols), (i // cols, i % cols), fig = figure)\n",
        "  cell.plot(history['acc'], label='Train')\n",
        "  cell.plot(history['val_acc'], label='Validation')\n",
        "  cell.set_title(res.config.plot_title())\n",
        "  cell.legend()\n",
        "  \n",
        "  i += 1\n",
        "  \n",
        "plt.show()"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABIUAAARtCAYAAADLW/VCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzs3XdYXPeZ6PHvFGAYGHrvCNBBSKh3\nybIly723uMRJnO4Ur7Ob7Td790k2N7vZzW6cTeLETpzETlzlItmWu9VlNdQAAYfeexlmgGHquX8M\nQkJCEiAkkHg/z+NH5tR32vmd8/6aTtM0hBBCCCGEEEIIIcTMop/qAIQQQgghhBBCCCHE5SdJISGE\nEEIIIYQQQogZSJJCQgghhBBCCCGEEDOQJIWEEEIIIYQQQgghZiBJCgkhhBBCCCGEEELMQJIUEkII\nIYQQQgghhJiBjFMdgBCjURRFA6oALxACHAP+n6qq+6YwpgygYiiuR1RVPaIoyo3AX4D/VVX1x6dt\nuwD4DRADdAKPq6paOLTuIeAHQABQDHwFyAJeGvo3R1XV2sv0soQQYsrNtGu+qqq9Fzj3nKHjxQMe\n4F9VVX1zaN0G4GdAKFAHfFlV1cahdd8Dvom/0m838G3gJuC/AEVVVd1Fvi1CCDElroRyAmgDfgtk\nAzrgF6qq/mZo20ktJ4b2+8LQMR9XVfUvpy2XckKMi7QUEtPZdaqqKkAq8DywRVGUdVMcU5OqqrlD\nDwePAP8KHBllu1eA/1RVdTbwH8CLAIqipAG/BG4dem21+Au0I6qq5gJNl+NFCCHENDRjrvljOO8m\n4HlVVefgf9B4QVGUcEVRQobO9bWhc72D/wEERVFWAk8Cq4BcIAL4K1VV3xkqX4QQ4ko3rcsJ4Bng\n8NC1ewPwE0VRlKHtJrWcUBTlH4EHAPWM5VJOiHGTlkJi2lNVVQM2KYoSjv8iulpRlCD8Ge2bgUDg\nWVVVfwKgKEot8O/AV/EXGi+pqvp9RVGM+C+K1wAGoBB4TFVVm6IodwE/xl/zUIm/VrjzAqGVAeuB\nZ09fqChKPhChqurmofjfVhTld0M1vxuBT1VVrR/a/DlgO/DdCb05QghxlZnp13xFUQzAvwFvDB2v\nSFEUF5A59Pqqhx4+AP4A/ExRFAv+h4NXVVW1Dh3nD/iTWD+7wOsSQogryjQuJ54B9gzF2KwoSg0w\nR1GUQCb/2WA78NOhf0+3ASknxDhJSyFxJXkbWKEoSjDw90AekA/MBe5XFOX207Zdhz8LvgR4QlGU\nFPxNIzPxZ8ZzgBPAKkVRZgF/Bh5WVXUW/ovrby8UzFDrHtcoq2YD1Wcsqx4672z8TUxPqgLiFEWJ\nvND5hBBihpmR13xVVb2qqr6qqqoHQFGUFUOrys88nqqqfUAX/q4Ko51Lan6FEFez6VZOvKOqag8M\ntwCajb916aQ/G6iqemAoOXYmKSfEuElSSFxJbPi/sxbgDuBpVVWdqqr2Ay8A95627UtDN9bN+Pv3\npgId+AuLewCzqqr/oqrqh/hrFHaoqlo8tO9vgTuHamsnwgwMnrHMgb+mYcQ6VVWdgDa0TgghxCkz\n/pqvKEoq/vHmnlBVdWA85zptuRBCXK2mZTmhKEoE/taePxlqAXQ5nw2knBDjJt3HxJUkA3ADVvx9\nYH+uKMpPhtYFAQdP2/b0wdm8gEFV1X2KojwBPAE8ryjKO/gHV4sA1imKUnbG/tFA+wTi7AdMZywz\nA31nrlMUxYR/ILq+CZxHCCGuZhnM4Gv+0DgU7wH/rqrqi+M912nLhRDiapXBNCsnFEVJAN4H3j7Z\nfY3L+2wg5YQYN0kKiSvJ/fiz9i5FUZqBn6mq+u54DqCq6uvA64qiROHvY/t3+GcN+ERV1fsnKc4y\n/LOIAaAoig5/k80SIAm49rRtc4CWk317hRBCDJux13xFUZKBD4C/V1V10xnnevC07cKBSPyvqWzo\nvKefq2QCr0cIIa4U06qcUBQlDPgQ+JOqqj8/bdXlfDaQckKMm3QfE9Oeoig6RVHuB74H/PPQ4i3A\n1xRFMQyt/4GiKDdf4DhfVhTlXwBUVe3Gf2HU8F+8rxnqP4yiKMsVRfnFRONVVbUE6BiaqQbgS0Cd\nqqrlQ3Fff9pMBH8DvDzRcwkhxNVGrvmAf4rhp85ICIF/XIt0RVHWDv3918C7Q10lXgMeVhQlfmjw\n1CfHeC4hhLiiTONy4sfAtjMSQpf72UDKCTFu0lJITGc7FEXxAOH4s9i3qapaMLTu1/ibjJ7A38Sy\nAHjqAsfbAvxBUZQKwIM/Y/6YqqrdiqJ8HXhraHYAO/5C5ryGRuxfDSQCLkVRHgV+parqr/BPIfw7\nRVF+iL/f8ucBVFVtUhTl28DmoYvxEfxNVoUQYqabUdd8RVGWA/+mqupNZ5wnCf/YGLmKonzrtFV/\np6rqO4qiPAT8WvFPO1wJPDZ0rgJFUX4G7B56jz7Gn1wSQoirxbQuJ4BvAs2Kotxy2rKnVFX9LZNY\nTgyt+xBIB9KA2Yqi/AD4J1VV35JyQoyXTtNGG7RcCHEmRVEy8DdRzbiE56gFrlNVtfZSnUMIIcSF\nXaZr/suqqj58qY5/xrk0VVV1l+NcQggxE0g5Ia4W0n1MCCGEEOIyUxQlGnh+quMQQggxPUk5IS4X\n6T4mxPgkD81E8Iiqqkcm66CKoizGP+1w8mQdUwghxEW7JNd8AFVVu/APJn1JKYpyB/Bfl/o8Qggx\nQ0k5Ia540n1MCCGEEEIIIYQQYgaS7mNCCCGEEEIIIYQQM5AkhYQQQgghhBBCCCFmoGkzplBHh33C\n/dgiI8309AxMZjiTYjrGNR1jAolrvKZjXNMxJrg64oqNtchMEFx95cR0jAkkrvGajnFNx5hA4hov\nKSfGT8qJy0PiGp/pGNd0jAkkrvGarHLiqmgpZDQapjqEUU3HuKZjTCBxjdd0jGs6xgQSl/Cbju/3\ndIwJJK7xmo5xTceYQOIar+ka19VqOr7f0zEmkLjGazrGNR1jAolrvCYrrqsiKSSEEEIIIYQQQggh\nxkeSQkIIIYQQQgghhBAzkCSFhBBCCCGEEEIIIWYgSQoJIYQQQgghhBBCzECSFBJCCCGEEEIIIYSY\ngcY0Jb2iKPOALcDPVVX91RnrNgI/AbzAe6qq/tvQ8p8DKwENeFJV1UOTGbgQQojpQ8oJIYQQ5yPl\nhBBCTE8XTAopihIC/BL49Byb/C9wE9AE7FQU5Q0gFshRVXWVoihzgD8AqyYnZCGEENOJlBNCCCHO\nR8oJIYSYvsbSUsgJ3Ar8w5krFEWZBXSrqtow9Pd7wPX4L+KbAVRVLVUUJVJRlDBVVW2TFrm4ajR1\n9nOkqovungHcXh9er8YSJZbE6JAL7qtpGt02J932QXrsTrptTgZdHnyaf51P0zAFGgk1GQkJDiAq\nzERWUhg6ne6cx7P2uahvs1Pf3ofeoGfA4ZrQ61qUE0t2cvgFt/NpGnWtdqx9Tnr7Xdj6XTjd3vPu\nYw4OnFBcBr2etfkJxEWaL7jtscpOKhqtYz72RGMaTURoENcvSUF/js8J/J9VdbON4ppuXJ5zv1+T\nGdd4mAKNrM1PJNISdNnPPQWknBBiBvJ4feh0/rJlPLw+HwdL22ns6LtEkY3PVJUTZ8pNiyR/VvRU\nh3GpSDkhZpxu2yCHytq5fV32hPb3adp574XF2DhdXuwOFxGhQRgNMnrOaC6YFFJV1QN4FEUZbXUC\n0HHa3+1AFhADHD5tecfQtue8iEdGmjEaDWMIeXSxsZYJ73spTce4xhuTpmm4PT7sAy4a2uw0tPXR\n0GZn0OUhMymcWcnhZCWHE2oOHPdxt+yq5vmtJ/B4tRHrPi5o5Km/ufa8yQu1rpvfvFlIVWPvuM77\n+ZtzeeiGs7/PL35Qxgf7a7HaneM63rl8eriJX/3tehJjzp3c8nh9/L8/HqSgtG1SzjkW+0va+J/v\nrSPSYhp1vcvt5XdbivlgX+1li2k0cdEhbFyeftby2hYbnx6qZ29hMx09jimIbOze/ayWm1amc/+G\nHKLDg89aPx2vDxMh5cTETceYQOIar+kY16WMqb1ngHd2V/Ph/jpS4kL58eOrMZsCLrif2+PjaHU3\nmz4tp7Vr4JLFd6WqaraxYUXGiGXT8bs1EVJOTNx0jAmmX1xWuxOX2zst4urocbBpWzkfH6jD49Vo\n6hrgH764bMz7O5we3thWweZdVdyyKoOv3jnvgvvUtdg4XNZOj32Q7t5BrH1OUuMtXL8sleyUiFEr\nxM/1Xmmahs+nYbhECZR+h5u6Vht1LTa8Po0bVqQTFHDqdzuRz9Dr06hvtVHT3Et1k//f1u4Bevuc\nOF3+yuO1C5LG9DlsK2jgRHUXGYlhZKdEkJkcNuG4zhmv18fBkjaOlrezPC+BJblx52y0cJKmaWga\n6PUjt5uMuMY0ptA4nOuVXDDF2dMz8ZuD2FgLHR32Ce9/qUzHuGJjLdQ1dLOnsIXtR5vw+jRuWZHG\n2vlJBBj9P3yH08POY83sLmzG1u9i0OXF69NGPd72w43D/5+eYGH9omRW5MUP/7D7HG4Oq+2UN/SS\nkWBhQXY0cZFm+hxu/rC1lGOVnYSZA3j0ljng9WE06qlvtbN5Tw0/fu4A//j5xcNxndTncPPGzip2\nHWtGAxZkRZMUE0KkJYioMBPBQUb0Ov8PRoeOQbeHPoebfoeH9w/U8fKHKumxISNa8ew63swrH6tY\nzAEsyokhPd5CWryF1ORwrBP4blY323j50wr+58UC/vbhRaNm+TVN4/kPVApK25idEs7CnFjCQgII\nCwnEFGg8748mItI8obiOVXaydV8dP/r9fv7+4UVnZcvbugd4enMxDe19pMSG8tD12SMu0ucz0ZjO\nNOj28qs3inju7RNkJVgIDT71oFHeYOW/Xj6K16cRHGRg9bwElsyOJSzk3AnJyYprvBo6+nhvXx3v\n7qnhg3113LoyjbuvmTW8fjzXh+lwgzOJpJwYMh1jgpFxdfY6cLp9JEWbL3izcjnjmky9fU7MJiMB\npz1IappGa/cAxdXdGA061i1MOmdrmHPFdayyk4KydtbkJ5KbdvYNucPpQa/TERR44WvssYpOyup7\nUFIjUNIiMZvOf/t2vveqf9BNV+8gBoMeo0GHUa8nwhI4ptY+TZ39vLO3hoKyDnyaRmCAnooGKz/8\n3T6+98CCc9bAaprG/pI2tuypob3HgdGgY8PiZJbPicegn/pa8KkqJ86UGG0e8blJOTHm5cOknLg8\npltclU29/Ozlo0SHB/PNO/NIi5+a30O3bZCt++rYdbwZr08jLjIYnU7HnuPNbCxuvmBcPp/GnqIW\n3tpVTW+/v/Xi5p1VhJmMXLco+Zz7ldZ289Trhbg9vhHLCys72bq3huSYENbkJ7J+cfLwff25PsOK\nRivPvl1Cb7+LpGgzybGhJMWYcTi9tFsddPQ46B90s3peAjctTyM4aPTy6GRPiOLqLjp6B7H3u7AN\nuLD2ueg5owJ+X2Ez3703H6NBPxyXy+3l9Z1VGA16blyWSkTo6C3ve+xO9hQ2s+t4C122wRHrIi1B\nJESasYQE0N7tYM/xZtYeaWB2asToMfs0Xv60gk9Pe74F0OkgPSGM1NgQZiWFkZEQxqDLQ3uPg3ar\ngx67E00Dg16HXg8BRgNRliAiw4KIspgICQ4Yfjb1+TQKytrZcax5+H14/7Na5mZE8uCGHFLiQkeN\nrbCqk+c/UMmfFc1jt+QOL5+scuJik0LN+DP2JyUPLXOdsTwJaLnIc4kJ6HO4sfY5sfW76O130dhZ\nzacF9bjcPgKMenTAnz8q5739ddy6KoMeu5NthxsZcHoIDNATGxGMKdCAKdCIOchIfJSZpBgzSdEh\nBAYYqG+zU9dmp7bFTll9D396v4xN2ytZmZdAR6+DEzXdwwmlfSdaefnTChKizDjdXnrsTuakR/KN\nO/LIzowZ/kIvyIqmrcfBvhOtvLqtgkdv9Ncq+XwauwubeWNnNX0ON8kxITx642yUtMgxvx8psSH8\n50tHefbtE/zwK8sJDjJS02LjLx+phJiM/MsXlxITcapFR2yshQ7zhWs/zzQrKYyy+h6OVnSy82gT\n6xennLXNe/v9hUZafCjf+9wCTIFj/zleTFwdVgcHS9v584cqj92Si06nw+n2sut4M2/tqmbQ5eXa\nhUk8fH0OgWNMCF1MTKO5+5pMXt1WyabtlXz51jkA9Pa7+M2WYjQNvn57Hktz485KGF7quMYjKzmc\ntfmJfFbcyruf1XKgpG1EUmgGkXLiCtXWPcCPnj+Ew+kl0hLE3Iwo5mZGsUSJvWqaX+8tauEPW0tB\nB/GRZpJjQwgxBVBa102H9dTNZYHawbfunjciSX0+xys7+fWbRXh9Gp8Vt5IWH8pNy9JIiDZTVN1F\nUXUX1c02NA1CTEaiwkzEhJu4dWU6WWd0Oz5e2cmv3izCp2l8dKgBnQ4yE8NYmB3DmlG6qHq8Pupb\nbeh9vhFJ/faeAT481MDewhZcZzw4mIOM5GdFsyArmnmzokd9nQdL2/jDe6W43D5SYkO5aXkqS3Pj\neGbLCY5VdvLc1lK+fkfeWZUg7T0D/PlDlRO1PQQa9WxcksItK9OnVdfaqSonxDApJ6YZTdOmvCJg\nrDqtDn75RiFur4+Wrn5+/MJhPn9DDusWJI36Grw+H1s/q6O4tpvgQCNmk/8ZJyTYSKgpgFBzAKHB\ngWQnh42pBSSMngy6Y3UGK+fGU1rbw/+8dpzNu2v4q/vnn7XvoMtDWb2VE9XdFFZ30mEdJDBAz11r\nM1mqxPLTl47y4sflJESZyU0/+5lHre/hF68XomkaX7xZIS3OQkRoIKHBAZTW9bC3qIVjlZ28tr2S\n8gYr370v/5yV1R8ebOD1HVVoaKTEhtLSPUB9+8huvgFGf4XC23tr2XakiVtXprN+UTJ2h4uOHgdt\nVgcVDb0U13RhH3CP2Ndo0BEWEsjcjEiSY0NJiQ3lYGkbhVVdPPtOCd+8Mw8Aa5+TX75RRE2Lv1Hg\nJwWNrFuQyK0r0wkKNNDQ1kd9ex9qfQ/HK7vwaRpBAf7K4llJYaTFWUiODRmRsKps6uUnfz7Mpu2V\n/PMXloxaUfPM2ycorOoiOSaEL9yk0NU7SE2rjbpW/7AitS02dhdOziXIFGhg/eJkFmTF8ElBA8U1\n3fzrHw+yZl4iK+fGMzs1AqNBj8Pp4dVtlew63ozRoEM5R0LrYl1UUkhV1VpFUcIURckAGoHbgc/j\nb+75Q+AZRVEWA82qqk6fdPJVzOn2Ut5g5URNN8U13TR39p+1TXRYEBvWpHDNgiS8Po3399ex/WgT\nf/5QBSA0OIB7rslk/eKUC94AJ0SZWT4nHvBfEHcca2bX8WY+PeLPsKbFh7JiTjxzM6OobbVzvLKT\nktoeXB4v91yTyW2rMs5qAqfT6fjizQoN7Xa2HWkiKzmciNAgXv6kgsaOPoICDDywPosblqaO+8FE\nSYvk1lXpbN1Xx4sfl/PghmyefqsIr1fjG/fNHZEQuhg6nY4v3KSg1lt5bUcV+VnRxJzWfehASRtv\n7Kwm0hLEk/ePLyF0sXF9+dY5tHU72F3YQlxkMD6fxscFjfQ53AQFGPj6HXmsmptw4YNdQhuXprC3\nqJXdhS2snZ/IrKQwntlSTG+fiwfWZ7Fq3tTGN1ZGg551C5JYOz8R3zla213tpJy4MjndXn79VhEO\np5d5s6KobbGzp6iFPUUtrMlP4Ku35Z13f03TeG9/HS1dA8zNjGJeZhSWc3Qx9nh97D7eTHFNN7np\nkSyfE0/4eVr/nblvdbONyqZeOnsH6bb5//P6NPLSo1iQHY2SFjlqAvlEbTd/er+M4CAjybEhNHX0\n09rtb2UQHGRgiRJL/qxojld2crSik397/hBP3DeflNhQGjv62FvUwvHKLuZmRXPbirThWszSuh5+\n/VYxBr2/HCiu7uJweQe/e7dk+Nw6HWQlhRMUaKDbNki71UFDex9F1V08dksuq+clAv7WkU9vLsZo\n0PHojbl09jooqe2hutlGdbONt3ZXM39WNKvzE7H1uzhR001pfQ9OlxedDpKiQ0iLt+DyeDlS3oGm\nQXSYiQXZ0Wia//1zeXxUNFo5UNLGgZI29Dodi3JiuGFZKjkp4WjAW7uq2bqvjqBAA9+6ex5Lldjh\nG+rH75rLz145xoGSNsJDAvnc+mxsA/6a4OKabt79rBa3x0f+rGiefHgxeu/5x8wTM4+UE9OHpmls\n3VfHBwfqeej6HNbOTxzX/p1WB0cqOkmLC2V2WsSYx8Nxub3sGOqlkBobyp1rM0mIOjWEhKZpVDT2\n0mF1sCgndri15MCgh1+8Xoh9wM2jN85mVmok//3iYZ7/QKW8oZcHr88m7LSyp6t3kGfePkFl04WH\nngg06lk2J45rFyYPj0eqaRoDTg+d1kFauvpp7uqnqaOfwqoufzIoIpjbV2ewal78cOvLuZlR5GVG\ncayyk+pmG7OS/F2RfJrGK59UDPfcAAgKNLB2fiL3XDNrOHH+nXvm8bNXjvH05mJ+8KWlxJ32rFLe\nYOXnm47j9Wl85958FmbHjHgNC7JjWJAdQ5/DzW82F3OsspN39tZy19rMEdsNDLp5bmspRys6CQ8J\n5PG75qKkReLzabRbHbR09WMOMhIXaSY8NBCX28snBY28f6Ce17ZX8tr2yrPev/CQQNbOT2T+rGhS\n40KxmAMJDjKclYxZNieOn792nIKydoIC9DywUeHfni+gx+5k9bwEspPDeW9/HduONLH9aBPaGbfT\nafGhXLfQ31PlXK2WALKTw1mqxFKgdlCgdrAsN254XbdtkKc2FdLY0cfczCi+ddc8/3csleFnjqio\nEI6VtlLTYqO+rQ+zyUhsRDBxEcFEh5vQ6cCn+RsxOF1euu2Dw+PeOpxefD7/WLeapjErKZyVp8U7\nPyuaououXt1WOXyvFRRoIC89kob2Pjp7B0mJDeXrd+SReo6WRBdLp535zp5BUZQlwH8DGYAb/6wA\nbwM1qqq+pSjKOuCnQ5u/oarqz4b2+w9gHeADvqOq6vHznaejwz7hJ6bp1oTxpMsVl9Pl5XhVJ4fK\n2imq6hquAQw06slJCSc20kyYOYDw0CCy06NIigg6q5l4j93JruPNhJkDWJ2fOOYuQ6PxeH2U1fUQ\nHW4adbBot8eL2+MbkX0f7b06WUvtdPnwDX1P1+QncO+6rIuqYfR4ffzkz4epbbUTH2WmrXuAu9dm\ncucZF8hzxTUee4taeG5rKXMzIvnmXfMorevhRE0XnxW3EmDU80+fX3LOZoLnc7FxddsG+dHzBdiG\nmqaGmIysX5zCxiUp5+2KdSljOtPJjH5ybAj5mdF8cLCeRTkxfPfe/HHVXl0N14fYWMu0rq6TcmJi\npmNMADExofz0TwfZW9zK+kXJfOEmBZ+mUd9m57l3S2nu6ufHX1txzskAfD6Nv3yksuNY8/AyHZCZ\nFEZeRhRz0iLISg4nwKjnaEUnm3ZU0dZ9qsuHTgd5GVEsyIom0mIiPDSQsJBAIiLMVNV1Y+1z0m0b\npLyhl/IG61kD85uGumMNDo0hEBRgYNHsGO5YnTEcc2N7H//+4mHcHh/ff3AhSlrk8EQDtn4XybEh\nw5UOPk1jy+4a3vmslqAAAwnRZupa/Z+bQa/D69MICjRw55oMZiWG8dSmQjxeH0/eP595Q4MGd1gd\nbD/SRN+gm3mZ/hZXIaeVgZqmcaKmm99sOYHD6eG2Veksy43jpy8dxeX28sR9+czPOnWzPzDo4WBZ\nG7uPN1PTMvI7FB9lZu6saBpbbdS19w2PpZAeb+HmFWkszY096x5A0zSaOvr99xKl7cM1w2nxoYQG\nB1BS20NcRDBP3JdPcuzZZVafw81/vHiE5s5+9DrdcJkNEBYSyCMbc1iWG0dcXNi0/M5P19+ilBNS\nTlzOmJwuL8+9V0pBWTvgvxZ/9558Fs2OHbGdw+lBF2DE43QTMpScqWjs5eOChuHkM/groVfOTWDF\nnHiSYkLOqgSGU8mg9/fX0dvvQqcDTfOfe/W8BDYsTqGsvofdx1uGk/amQAPrFiSxYUmKvxViTTcb\nl6bwyMbZxMZaKK1s5zebi6lpsaPX6cjLjGRlXjxGg54/f6jSP+hh+Zw4vnhTLga9jgGnh/5BN/0O\nN30Oz3AX2/0lrcMtRuMigtHwlxFndtECRrQMGq0rbqvNyT8/vZe5mVF8/8GF+Hwaf3yvlL3FrcRF\nBLM8L465GVFkJYePWuG941gTL3ygkhhtZlluHB6vf7zXXceb8Xh9fPvueWd9TmeyD7j40Z8K6LIN\n8sR9+dy4ehYdHXZKarv543uldNmc5KZF8M075xJ+jq5aZ+pzuPngQD3ljVaiw0zERQQTFxlMSmwo\nafGhY75fdzg9/OyVo0OfmT+5ct+1s7h1ZTo6nQ6P18f+E23sPNaEKchIWnwoaXEW0hMsI5KHF9LW\nM8APfneA6DATP/76CowGPVXNvfzyjSJs/S7WL0rmkRtyRv0ML8dv0evzUVrbQ2F1F0VVXbT1ONDp\n4NaV6dy5JnPUCq7JKicumBS6XK62izhMflyaplHeYEWtt2IfcGN3+G9eq5ttw4mgxGgzC7NjmJcZ\nRXZK+IgxEi5FTJPlXHEdKe/g6beKyUoO46Hrc8hMDJuU87V1D/CvfzyIy+1jflY0f3X//FFrMy72\n/dI0jac2FVJU3TViucUcwDfvnEteRtSEjjsZn2NVcy+btlexKCeGdQuSzptdv1wxnelP75ex67j/\nwTIuMpj/+6VlFxxL43LENRmuppv9y+VqKyemOiafT+P9A3XsKWpl8ewYrluYTGxEMEequvnVpmNk\nJFj4p0eXjLgJOax28Ou3iliRF88375x71jE9Xh/PbS3lQEkbaXGhPHLDbCoarRRVdVHZZBtOFhgN\nOqLCTLT3ONDrdFy7MIkNi5Mprethf0kb1c1jm1woMdpMXnoUuekRxEeZibKYMJuMeLw+Kht7OV7l\nb+XTPnRjtXpuAusWJvHbLSfosTv5xp15rMwbW8vDgrJ2fr+1BI9HY96sKNbmJ5KfFU1RbQ8vvFdK\nn8PfTF6v0/Gtu+exRDn/DfpoWrr6+cXrhbT3OIYTThdqvdnQ3sdhtZ2IoS5+sRHBw98tn6bR3uPA\n6fKO+Qb9ZI386Q94Jys2ztenRgSUAAAgAElEQVR6uNs2yJ/eL2PQ7SUiNIjI0CCiw02syU8YToBN\n9Xf+XK6GuKSc8JNyYmI6ex386o0i6tv7mJ0Szs0r03lmywm8Po3vP7hguNXIzmNNvLmrmv5BD+BP\njAcHGYevf+nxFtbOT6Su1U6B2j6cnA8w6kmIMpMYbcYcZKTD6h+LpavX6e/6E2hg45IUbliWSnm9\nlS17amg6rbeD0aBnqRJLfJSZHcea6O07NVvg/Kxo/uq++ej1uuH3y+3xseNoE/tLWkckzgOMej5/\nw2yumZ94weuhT9Moreth57Fmiqq6CA4yED50bYu0BJEYbSYpJoSkmBDCQwLPe7zYWAt//7+7KK3r\n4e8eXsTu483sL2kjMzGM7z+4YEzd1F78qHy4J8ZJBr2Ox+8ae3lT12rn3/9yGINBx0++vZa3d1ay\n/UgTep2O21enc8eajHHPJjlZ+hxu/uvlo7RbHXzttrwJlaFj8dLH5XxyuJGHr88hLCSQ57aW4vX5\neGhDDhuXppzzc5yK60N7zwB6ne68PVkkKXSa6XgRh8mLy+H08FlxK9uPNo3aHSw+yp81Xp4bR3Js\nyAUvSlfae+VwejAFnt3c8GIdrejgUGk7n79x9oga27HGNVbdtkH+57XjhJiMQzXE0WQkWEatMRmr\n6fg5XoqY+hxu/s/v9uN0efk/X1w6oSaT0/G9ArnZn4irrZy4VDH1Ody43F6iwkafYRD816Xfv1tC\nWb11eJkOfzN3tcFKoFHPvz627KwbEU3T+OGfDtHQ1sePvraC5NNmV3S5vfxmczHHq7rITgnne/fP\nH3GjOzDoobzRSlldD2q9lcaOPuZnRXP/dVlntTpq6xmgutmGrd81/J/ZHIjJqCfCEkREaCAZCWFj\najWqaRpHKzp5a3c1TR2nytD7r8vi1pVnz3B4PtY+JzqdbkT3tthYCzX13WzZU8P+E608csPsi+qC\n2+dw8/RbRZTVW3l4Yw43LE0d9zEm67vV2eugtsXOotkxk/KgMB1/h3B1xCXlhN9MKif6HG5MgYZR\nW5ZUNffyxo4q+hweXEMt9MNCArl33Szyh1owgj/xsaewhdd3VNHncHPdwiQeuWE2RoOe4poufrGp\nkMAAPY9snM0nhxupa7UTHGRgVX4SPb0ObP0u7ANuUuNDuWGpv8vpyft1p9vLsYpOCqs6ae4coKWr\nf8R4ZmHmAGIjg5mTHskNS1NHdDH2+TQOlrVxpLwTJTWClXPjh+/V3R4fB0vb+PhQA4GBBv76gQXD\nFZvn6nlwoKSNth4Ht6xMI2WU1o6XWmyshX3HGvnJnw8TYNTj9vjITg7new8sGHNlp6ZpVLfYcLt9\nGA16DAYdUZagMbfqOWn/iVaefedUd+akmBC+etucSat4vxgerw9LWDCO/smZCXo09gEX//jMPrxe\nDZfHR3CQgcfvmjfidzGa6Xh9AEkKjXA1fEinG3R5qGm2Udlso6qpF7XBitPlxaDXsSw3jhV58URa\ngrCY/YOIjWWw3YuN6VKTuMZnOsZ1qWLq6h3E4/URP47moaebju8VyM3+RFxt5cR4YhrroJ81LTZ+\n8Xohtn4XeRmRXDM/icWzYwgwGtA0jf5BDyW13cNN6BflxPDojQqldd1sP9JE1VALnSfvn8+CM8Ym\nOOloRQe/fKOI5XPiePwu/zS5A4Nu/veNIsobrMzNjOK79+RfcFYtn6aNebwJuPjP0KdpHCpt54OD\n9cxJj+SB67ImpbLh9Lgma3BWn0+j2z44Yiy6icY0nUhc4yPlxPhd7eXEyVl9D5S0odZbmZ0awfcf\nWjgiMTQw6OZfnjtIj92JOchIQICeQKOezt5BNM3fsubBDdkMOD28+FE5ta12ggIMPLgh+6wZrg6W\ntvHMlhOcfFNXzU3gc+uzRkwQM1Y+TaO7dxCHy0tMuOmiW6iPZjp+hnAqrqc2HaewqovZqRE8ef/8\nS/IejMWm7ZV8eLCem5ancfc1mWf1LJlKl+MzfG9/Ha/vqCIm3MSTDywYUcE1lXFNxGSVE1PzTRSj\ncrq8bN1fywcHGvB4T2XS4yKDWbMykXULksY8+KYQV4vo8HO3dhDiauf2+PjkcAPv769H0zRiwoOJ\nCTcRFxnM0ty4ETV7h9UOfvfOCdxeH+kJFkpqeyip7SHEZCQsJJBum3N4/J0Ao54v3KRw3UL/7Cyr\n5yWyel4i9W12TOYg4iznLmsWZseQnmDhUGk7t6/uI8QUwM9fO0ZjRz9Lc+P4+u15Y6qsGE9CaDLo\ndTpW5MWzIi/+kp1jslq06vW6CSeEhBBXnx67k9d3VHGwtG14UOKI0EDUBitv7qrmc+uzh7d96ZMK\neuxO7lqbOWJA4cb2Pl7+tILCqq4RswOvzIvngfXZo7a8XD4nHpfbx4GSVm5fnTGuGX/PdKFuMDPB\nl2+dQ0FZO2vzEy9YcXIpPbA+m6/cnY+91zFlMUylm5anEhvhb6U21llFr3aSFJoGNE3jUFk7r26r\npMfuJNISxIq8eLKTw8lKCht3s0AhhBBXNk3TKFA72LS9ks7eweHETnNXP3Vt/hqh9w/Uk5FgYf3i\nZPoG3Ly+o4rAAANP3DefhdkxtHT1s6ewhc9OtGIfcBMfGUxUmInocBPXLUwadcDgtHjLBWuddDod\nd6/N5BevF/LSx+V0WAfpsg2yYXEyj2ycfVFdY4UQQpzi8fr48GA9W/bUMOjykhQTwpp5CSyfE4/Z\nZORHzxfwwYF6spPDWTw7liPlHXxW3Ep6goXbVo3sHpsSF8rfPrTQP8D/9kpMgUYe3pjD7AtMcb12\nfuK4ZyETowsPCeT6JSlTHQYApkAj06/dy+Vh0OtHzD4mJCk0pQYG3RSoHewubKaqyYbR4B/k67aV\nGVOaPRZCCDF1alpsvPxpBZWNvRj0Om5clsodazIIMQWgaRq2ATd1rTZ2HmvmWGUnf3yvDPDXGj95\n/wLSEywAJEaH8MD6bB44rQZ5sszPiiYzMWx4TKJ71s3i9lXpkz72mxBCzEQ+TaOwsovNew9R32on\nxGTkSzcrXLMgaUQry+/cPY8fv1DAc1tLCA9ZyAsflGE06Pna7XmjjjWk0+lYPDuWRTkxcr0WQgyT\npNAUKG+w8nFBA8cru4a7iS3MjuGh67OJi5zYuClCCCGubD12J2/srOKz4lYAlsyO5YH1WSPKhZOD\nHM/PimF+VgxdvYPsPN5Ma/cAD23IPu/g0pNJp9Px4IZsnttawm2rMli3IOmynFcIIa5mLreXz4pb\n+ehQA63dA+h0cO3CJO67NmvUbi4pcaF84SaF57aW8pO/HEbT4MEN2RccI0USQkKI00lS6DLyeH1s\n2VPDe/vq0PCP9r4yL56VefEzvo+tEELMVG6Pv3vAu/tqcbl9pMWF8vDGnDGN3RAdbuLedbMufZCj\nmJ0awU8fXz0l5xZCiKtBa/cANS02Gjv6aOrop6qpl/5BDwa9jjX5CTx00xxCjOdP4KzJT6SyqZed\nx5qZnRI+oRkLhRAzmySFLpNOq4Nn3j5BVbON2AgTX70tb8S0jUIIIWaeE7Xd/OWjctq6BwgzB/DI\nxtmszU+UcXmEEOIqVt5g5d19tRRXd49YHhEayG2r0tmwOIVIS9CYZxZ6ZONsspLCmZ8dLeWHEGLc\nJCl0GRyr6OR375bgcHpYmRfPF25SpmwKQiGEEJef0+Xlo4IGBgbdBBj1BBgNtFsH2VvYjE4H1y9O\n4Z51mZhNMguGEEJcrcobrLyxs4qKxl4AlNQIliixJMeGkhIbgsU8sVmGA4x6GQxaCDFhkpm4xD4u\naOCVTyoIMOr5yq1zWJOfIK2DhBBiBvFpGr/fWsJhteOsdbOSwvjCjcrw4NBCCCGuTu09A/z3q8dw\ne3wsyIrmtlUZZKeET3VYQgghSaFLxefT+N3mIt7eXU14SCBPPjCfjISwqQ5LCCHEZfbu3loOqx3M\nTo3gwQ3ZuNxe3F4fMdGhxFkCR8wkI4QQ4uqjaRp/+bgct8fH12/PY9W8hKkOSQghhklS6BJwur08\n+/YJjlZ0khwTwpMPzCcmXAaSFkKImeaw2sHmPTVEh5n49j3zCDuta8BYx4oQQghxZTusdlBc3c3c\njEhWzo2f6nCEEGIESQpNMo/Xx283F3O8qouFObF87bY5mE3yNgshxEzT2N7H798tITBAzxP35Y9I\nCAkhhJgZHE4PL31SjtGg59EbFRlGQggx7Ui2YhJpmsbz75dxvKqLuZlR/N+vrcTa0z/VYQkhhLiM\nfD6NAyVtbNpRidPt5dt3zyMtXsYMEkKImWjz7hqsfS7uXJNBfJR5qsMRQoizSFJoEm3aUcXe4lYy\nE8P4zj3zCDDqpzokIYQQl4mmaRyr7OTNXdU0dfRj0Ov43PpslubGTXVoQgghpkB9m51PDjcQFxnM\nbavSpzocIYQYlSSFJoGmaXxwsJ4PDtSTEGXmew/MxxQob60QQswUmqbxmy0nKChrR6eDNfMSuGtt\nJjERMp6cEELMRPYBF8+8fQJNgy/cqBBgNEx1SEIIMSrJXFwETdMoqu7mnc9qqGqyEWkJ4m8eXIBF\nxo0QQogZZU9hCwVl7WQlhfHYrXNIjgmZ6pCEEEJMEYfTw89fO05L1wA3L09jbmbUVIckhBDnJEmh\nCSqp7eb1HVXUtvpnjlmUE8Pn1mfLLGNCCDHDdPUO8sq2CoKDDHzr7nlEhZmmOiQhhBBTxO3x8ss3\nCqlttXPN/EQeWJ811SEJIcR5SVJoAqqaevn5a8fx+TSW5sZxx+oMUuNCpzosIYQQl5mmafzp/VIc\nTi9fviVXEkJCCDGDeX0+frP5BGX1VpYosXzp5lyZbUwIMe1JUmicbP0unt5cjE/T+OsHFzAvM3qq\nQxJCCDFFdh5v5kRtD/mzolk7P3GqwxFCCDGFtuyp5VhlJ3kZkXzjjrno9ZIQEkJMfzI91jh4fT5+\nu6WYHruTe9fNkoSQEELMYJ1WB69uqyQ4yMhjt0htsBBCzGQdVgcfHKgn0hLEd+7Jl1mIhRBXDLla\njcObu6opq7eyKCeGW1fKtJJCCDFTebw+nn2nBKfLyyMbc4i0BE11SEIIIabQa9sr8Xh9PHBdFsFB\n0hlDCHHlkKTQGB1W23l/fz3xkcF89bY8qREWQogZ7PUdVVQ29bJ8Thyr5yVMdThCCCGmUGldD4fV\nDrKTw1mRFz/V4QghxLhIUmgMGjv6+P3WUgID9HznnnzMJsn+CyHETFVQ1s5HhxpIjDZLtzEhhJjh\nvD4fL39SAcDDG3OkTBBCXHEkKXQBfQ43v3yjEKfLy1dvyyNFZhkTQogZq617gD+8568k+PY9+ZgC\npZJACCFmsl3HW2js6GNtfiKZiWFTHY4QQoybJIXO4+TA0h3WQW5fncGy3LipDkkIIcQUcbm9/Pqt\nYgZdXh67OZfkmJCpDkkIIcQUGhh089auaoICDdx37aypDkcIISZEqjjPY9P2Kkpqe1iYHcPd12RO\ndThCCCGm0IcH62ns6OO6RcmsnCvjCAkhzk3TNJxeJ17NN7xMh45go0m6F11Fdhxrps/h5r5rZxEe\nKhMOCCGuTJIUOofTx4z4+h156KUAF0KIGWtg0MNHhxoIMRl54LqsqQ5HCDFBPYNW2gY6CAu0EBEU\nRrAx+Kxt+t0D1NsaqbM30NzXikfzju3gmsaAx0Gv04bVZcPldZ21iVFvJDwwjPCgMCKC/P+e/Dsn\nYhaRpoiLfYniMvF4fXx6uJGgAAPrFyVPdThCCDFhkhQahdPt5eVPKzAa9Dxx33yZVlIIIWa4Tw83\n0D/o4b5rZ0mZIMQVxKf5aOlvo7CjhMLOYurtTSPWB+gDMAcGo/m04e373P0TPp8OHaGBIcQHx2AJ\nshCgDxhe5/V5sbv66HXZqLXV4zutFRFAmiWZf1j25ITPLS6vI+Ud9NidbFicjNkUcOEdhBDiNF6f\nlzp7Ixlhqeh1Uzuqj9zZjuKjg/X02J3cujKdhCjzVIcjhBBiCjmcp1oJbVicMtXhCCFG4fZ5sDlt\nWJ02epxWGu3N1NkbabA34vAMAqDX6cmNzCEjLJU+zwC9zl6sThse3Hi9pxI0qZZk0i0ppIWlkmpJ\nIsgw9m5BJkMQBr3hgtv5NJ8/QeS0YR2KI9WSNP4XLqbMJwWNAFy/RMoFIa5mmqZN+jFLulTerHyX\nlv42bs64njtm3TTp5xgPSQqdwdrn5L399YSZA7htVfpUhyOEEGKKfXK4kf5BD/euk1ZCQkwmt9eN\n2lNJYWcJJ7rKGHAPTOg4GuD2uUddF2+OJT8mj7lRCnnRuZgDzu4uFhtroaPDPqFzT5Rep/d3HQsK\nIw1JKlxpalpsVDb1Mj8rmsRomXRAiKuJy+uivKeK6t46anrrqLU3sCx5Po9kf+6ij93S38abFe9S\n0q2iQ0egPoAdDXvYkHoNIQFT1xhF7m7P8OauapxuLw9eny03/0IIMcM5nB4+OlhPiMkotcFCTBKn\n18Xmyq0caD2Mc2jcnZAAMwkh8RM+pslo8o/RExhGRFA4iSHxpIUljzpmkBAX65OCBgA2Lr36ygWf\n5qO6t45Z4elT3qVFiMutqa+FZwufp3Owe3hZoD6AvfUFXJewjqTQiU800uu08Z8Fv8TldaFEZnNv\n9u2oPZW8Wfku2xv2cPusG0dsv6/5EPtbCzi9odKqpGWsSlw64RjOZUxZD0VRfg6sxF8Z86SqqodO\nW3cX8APACbyiquqvFEUJBV4AIoEg4Ieqqn442cFPtrpWO3sLW0iJDWHdfGnCeyVy+zx0Obqxu+zY\nXH0MeAbIj8kjIih8qkMTVylN02jub6XCWo11sBeby47d1Ydep+OxuQ/LA8kVTloJCTG64s5S3qzc\nSkpoIl+Z9/kx71dvb+SPJ16ifaCTaFMka5NWMj92rjyAiiuGtc/JwdJ2EqPNzM2ImupwJt1nzQd5\nWX2TL+U9xPKExVMdjpjBfJoPHboJzdhY3VuHT/ORHTH2GcQPtR7lxbLXcfvcrEteRX5MHhlhaVRY\nq3i26AW2N+zh83PuH3csJ+1pPoDL6+KurFu4Ie06dDodseYYPqrbzo5Gf2uhk61Z1e5KXix7HQ0N\nHadef3rYpUlEX/AOV1GUa4EcVVVXKYoyB/gDsGponR74FbAY6ALeVxRlM3A3oKqq+k+KoiQB24Dc\nS/IKJommaby6rQINeHBDDnq9zDZ2ukGPk/KeSubFzLnkN20n+22O5wKgaRpH2gvZVLEFu6tvxLqj\n7UX81aJvTCiOc/UhbR/opLmvhdyoHExG07iPLSaPpmnn/K402Jtp9npJ0CeP+Xs76BmkwlpNSVc5\nHY5OQgNCsASGEhZowWQMgqELs0/zUWdroLS7HJtr9G4Hak8VC2PnTeh1iak36JJWQmJ62Nawm+re\nOr6c9/CYxqu5lHoGrbxe8Q7HOooAaBtoZ3X3cnKjcs67n0/z8U7ZJ7xUuBmv5uX61HXckXUzAXpJ\ntoory/YjTXh9GhuXpk7oYXW6299yGICSrnJJCk1Ar9NOkCFw6J5RTJTd1cdTR58hMiicx+c/hnEc\nZcWR9kL+eOIl9Oj4p+Xfu2ArVK/Py1uVW9neuAeTIYgv53+JBbFzh9fnx+QRHxrLwbYj3Jl1M5bA\n0HG/Ho/Pw56m/QQbTVybsmb42hFkCGRj2rVsrnqPnY17uSVzIzaXnT+VvIxOp+P7i79DZnjauM83\nXmN5d68HNgOoqlqqKEqkoihhqqragBjAqqpqB4CiKJ8CG4FOYP7Q/pFDf09rJXU9lNVbmZ8VzdzM\nqy/rf7E+rNvGR3XbuTF9PXdl3XLB7etsDcSZY8bVSsLq7GVHw172NB9gcdx8Hsm9b0z79QxaebX8\nLYo6SwnQG1mRsITIoHAsQRYOtx1D7amkylpLVkTGmI7n9rrZ2fQZH9Vux2g0okRkMydqNhlhqag9\nlRxoOUKNrQ6AYGMw1ySv5LqUNYQHhY35tV4JWvvb6XR0kRSaQGRQxFk3Pj7NN2qixe7qo6mvhX73\nwHAyJSwwlGBj8Dlvnnyab0QCTqfTXTCJ02Bv4pP6nRxtLyLFksSKhCUsiVtASICZkm6VT+p2Um6t\nAiDGFMWGtHWsSlxKgD6ADkcXpd3lVFircXqdw8cc9AxSa2s4a0aY87EEhLIsfjG5UdnEmWOwBFio\ntzfwhxMv0drfDrFjPpSYZgqruugf9HDbqnRpJSSmTNtAB29VbsWn+VAis7kmeeUlPZ+maVidvdTb\nG6mzNdI20IGGdnIlZT0VOL0uZoVncE3ySp4veYUtVe8xO/KJ8163N5VvYVfTPiyBoXxpzkPMiZ59\nSV+HEJeC1+dj57EmQkxGVs+deDeS6arT0TV8j1tprT5vxZs4W0mXyjNFzxOgN7ImaQXXpawh0hQx\n1WFdcbw+L88V/4XW/jZa+9vYVPE2Dyv3jmnfY+1F/PHESxh0etw+Dy+VvcH3Fj9+3vLp4/qdbG/c\nQ0JIPN+Y9wXiQ+JGrNfr9Nyas54/Hn2N3U37uDXzhnG/pmMdxdhcdjakXkOQIXDEumuSV/Fx/Q62\nNezm2pQ1vFDyKjaXnXuyb7ssCSEYW1IoATh82t8dQ8tsQ/9vURQlB6gF1gM7VFX9qaIojymKUok/\nKXTbpEZ9CWw77J9B4I41GVMbyCSyuew097UyOzLrolr3aJrG0fZCAD6q205GWCoLztP6oaKnmqeO\n/paEkHj+ZvG3LjhoVnNfK5uq32J33UG8mhfwN11dn7qWxAtkdg+3HeOlsjcY9DrJiZjFI7n3EWc+\n9RSeGprM/xx5mvdqPuaJRV8/77F8mo+CtmO8U/0h3YM9wwmtg61HONh6ZHg7HTpyI3NIsSSxv6WA\nj+q2s61+F4viFjAvJpfcyBxCA8c+6OCA20GtrZ54cxxRprOTL1Ohtb+N/yr4NYNe/4wtZmMwyaGJ\n6NBhc/dhd9rp9wxgMpgICwzFEmghxGSirqeR3nO0mgkPtJAbNZu8qNnkRGYNJ2ZKu8uptzWeeujA\nn2j5x+VPjtrtr6KnivdrP0XtqQQgyhRJva2ROlsDr1e8TWRQBF1D/YBzI3NIioxjV+0BXivfzNaa\njzAZTMPrz6RDR5olhTlROcyJVkgJTWLAM4DNZcfmtOMaGvvipPiQOJJDE8/6fZ18La397WN5u8U0\ndaS8A4BluXEX2PLSGnA7eLboeZJCE7lj1o0zvkui3dXHbwv/xOqkZaxJWnHW+n0tBRxoKeBLeQ9N\n6Gbc5XUTaJg+00u/VfnucBL+vZqPWZ6w+KwbyvFwe93U2f3XzDpbAw19TTg9p5LjHp+Xfs+5B3wO\nMZq5L/cOViUuQ6/Tc6KrjIK2YxxtL2RJ/MJR9ynuLGVX0z7SwpP5dv5XJ1TLKsR00NDeh23AzboF\niQQFTm2rvUvhUOsxwN96ocdppWuwh5hgqSwfi9Lucp4peh4dYNQb+aR+J9sadrMkbiHfWv3IBff3\naT6eK/4LJoOJh3PvHVfLmDNVWmt4veJtsiMyuTf79iuua+4ble9SYa1mfsxcuga72dO0n9TQJNZe\noFLkeEcxz514kQC9ke8u/Bqf1u/iWEcxnzUfPO++R9sLMegMfH/xtzCf47l1feYqXil6m12N+7gh\n7ToCxnmfsLPxM8CfADqTyRjEhtR1vFP9AU8d/S1NfS3kRStsSL1mXOe4GBP5tg0/saqqqimK8iX8\nXcp6gRpApyjKo0C9qqo3K4qyAHgOOO+ISJGRZozGiV9cY2MtE9630+rgeGUnWSnhrJifPO6Hco/X\nw9vqxyxKnEdmZOpZcXl8Xkray/H4vMPLQwKDSQtPJjhgfF2PrI5e+twDpIQljrq+vLOag03HOd5a\nQp3Vn+i6M/cGHl1wKrs63veq3tpEh6OL7KgM6nub+HPZa+SlziLJMnrC5unibYA/sfD7khf4l+ue\nJMg48gZW0zRKOip4u+xjjrYUA5BkiecOZSPBAcE8te/3bG/ZyZOrvnrOuIrbVJ4veYVAYyCPL36U\n9Zmrz/rsYmPzyW9SKGpT6dF1MDtm1vC6QfcgBc1F1FkbqbM2UmttxDpow6g3cruykXvzbiYkwEyd\ntYnCthIqumrJic5gbdpyosz+B43HPPeys/YA76qfcKjtCIfajqBDx6zINNakL+PG7HXnfbhwe908\nte1pKrprAQgJCCYtIoVocySnv5LgABMRpnAiTBZCAs3sKekYiruJjoEuLIGhhJssRJjCiDFHkRaR\nTEZECmnhSfQ67RxvLeF4aykVXTU8OO8Ors9ae86YbM4+nj3wPIPeQW7MWofN1UedtZFKaw0aGqGB\nIUSaw0kLSmLA5cDqtNNhq0Xr1Yg2R7I4Jp+MiGTCg8KwOfuwDtroGeylsquGA62HOdB6eMT5DDo9\n2VHpmIZ+Cw73IJXdtexs283Xljw8YtvankZ+cexZNE0jP17hDuVGFiTMwTpoY2/9IXbW7KfB1sLa\ntGXckXvD8O/xofw7+aBiBx9V7sLhdbAiZRELEuaQH59LhOlU4kmv0437Ij+aaF8IAQeNdLk6z/t7\nu5jrlri03B4vx6u6iAk3kRo3tQ+we5sPUGGtpsJazdH2Qu7LuYMlcQumRQJ5KrxVuZVaWz1NfS3k\nRGQRZ44ZXtc20MGr6pu4fR6eLXqBv178rTEneLw+L6+ob3Kw9QhfmffoiKbjJzX3tRJkCCI6OHLS\nXs/5lHaXU9RZSk7ELLIiMvng/7N3n3Fx3Xei/z9nGgwMvffOoQhJqFu9WZJtybJlO+7dsZ3ESXY3\n2Xbv3m33f282+9+8kjjeuMY1bnGR3GTJVpfVJRASCA69M3SGDlPOfTBoBKIIyZJA8Hs/gjPtBwNz\nzvmebynfzd6qg2yIXXPZz6WqKtmNZ/m48HMs/e2u7UadES/9hQsZ7joNiX7xRHtFEuMdSbhnGLpB\nJWsXj1zfFL+e7IazfF66k1lBM4adyHT0d/Lngo/QSVp+tuhxjFYREBJuXEVVFgCSIqde9oeqqpyo\nz0an0XFz9Cq+LNtJUV0ka4MAACAASURBVFvpDREUcqgOdpbvZY4jhRBNxDV9raLWErIazpAekEKK\nfxI6jQ6lpZiXz7wJwDMZj5HoG8eJ+mx2Vx7gRH0Woef82RCxbsznPVWfw+lG5zlRt62Hp2Y8NGa5\nsNVupam3hWBjoOt+/XYrX5TuYG/Vd6ioVHXUYHXYuC/5zis+Zui3W9lRvhvZLxHZP/GKnmMkdoed\nUw05lFkqkf0SSAtIwaDVs7f0MPurDxHmGcKjaffSZe3m1yef5y+FnxHmGTqs8qPb2k1ZeyXFbWXs\nrjyATqPjx7OeJN4nFv9kPwpaitlWsp2MwLQRqzra+ixUd9aS4pc0akAIwF3vzpLwheyq3M/J+tPc\nFD4fuPD3sCziplGbUFd11FBqKSctQB5yzDLYisjF7K7cT01nHT4Gbx5Jvfe6BvPGExSqxZkZdF44\nUHf+G0VR9gPLAGRZ/hXOjKEVwM6B23NkWQ6XZVmrKIqdUbS2XtkYUvj+o0S3HijFocKyjDCamjov\n/YCLH1/8Fbsq97Oz8AD/a9EvXVfwgoK8aGho59Wzb5PTlDfiYwPd/YnwCmdh6Jwxs2/MXfXsqjzA\nCXMWdtXBY+n3M++iK3L7qg/xUeFngDNCneKXRENPE18U7CLRI4kE39ghv6uzTecotVQQ5hlCpCmc\nEI+gET989pQdA2B52GLsoQ7eOvcBv97/In8776fDrlYWt5WR11BIqn8yJr0nJ+qz+Y99L/JMxqNo\nNVp6bb2cbszlQPURKjqckxvifWK5K2MD0fpYNJIGVVWJNIVzuPIUq8NWEuo5/Cp9Q3cj/3XyZUDi\nmRmPkuSVMOp7tzZiNWfrFd7N/oznZj8FOD8AXjj9GnVd9a77+bn5sih0HrfErSXQ6E+PxYEpSMLT\n5sNNATdxU4AzsmvvgsauC39vs31mM3P+TKo7al2ZLyVt5ZS0VvBlwW42xa9nXsjsEf+xP1C2UtRS\njuyXiKfeg5rOOgoai4dkzYzFXetOiDGIblsPVZZaSlsrL/mYHcoBZnrPGvE2m8PGH06/Sn1XExti\n17ApZr3rtn67FY0kjXjlwqE68PFzo6Nt5JHA5+9T3VlLQXMRxZYyAtz9SPFPJtkvAeOgvkx2h51/\nP/Zf7C45xNKgxQQMHIyoqsqr2e+jqio/nPEws4MzAAbedw0L/Rey0H/hhbI2GzQ2dhAU5IW1Q2JN\n6CpWhawAuPBe9EB7Tx9D9V7qVzguwR5BVLebqW+wjPjeX87nlggeXX/nylvp67ezcnb4hAZf7A47\n+6oPYdAaWBu1nG8r9/FG3nscrTvJ4+kPjJiJmdescK65wPW9hMT80ExivKOG3fdGU9zmDC6b9J50\nWrt4r+Bjfp75DJIk4VAdvFfwMVaHjWivSCo7qnlf+YRHUu91vYf99n4O1hzF392PWUHprv9Nq93K\nG3nvufbVb517n1/OfW7IAV5ecwEvn3mLIGMA/7TwF9f878LusPNp0ZdISNyVtIlAYwDf1Rzl24r9\nLA1fdFkZqU09zXxYuI1zzQo6ScvyiMUk+MYS7RVJkDHge/0sgcYAlkUsYl/1IQ7VHmdF5GLXbaqq\n8r7yKR39ndyZeBvRvhHXffS7IFxNRdVtACRFTa6gUKe1i7fOfYCX3kScTwzxPjGEeYZc1kllVWcN\n9d0NZAZlkBGYypdlOyluLR33pCNVVWnobnReSPaNv649dfZXH+bLsp2ca8vnF5nPXbPXqWiv4o85\nr9PvsHKg5ggeOiPpASmcbsxFVVWenvmoqzR2cfgC5odk8ncH/5Ws2twxg0J2h50vy75BK2mJ8Y7k\nTFMer+e9yxPpD44aGPpA2cpR80kMGj0x3lHEekdzpimP+u5Ggo2B3JW0ic9Ld/BdzVE0aPhB8uYh\nn/X9dislbeWUWsops1Rg0Bp4MPWeYX3evijdwZ6qg3xTsZeN8etZH7Pqe+0zemw9HKo9zt6q72jr\ncwZZD9Qcxk1rIC0ghbNN5zDqjDyd8SjuOnfcde48mf4QL+S8xqu5b7MycgmtfRbaetto7GmmvrvR\n9dxuWgM/mvm4q7m0r5sPmxNu4cPCrXxU+BlPZTw8bD25TfkAzAhMveTaV0YuYU/VQfZUHSTBN5Zt\nJV+TMxDIO1x7nI3x61kTvXzY/92BgSyhlZFLRn1uo86d9bGr+bJ0J4+n33/dM2rHExT6Bvg34GVZ\nlucAtYqiuPbosix/DTwKdAGbgN8AEcBC4BNZlmOAzrECQhPJZndwIKcWo5uOhamXPwo1v7mQXZX7\n0UgaWvva2FG+e0jPnayGM+Q05RHjHUVmUIZre3t/B9WdddR01pLTmEtOYy5Lwhdyd9ImDAOBFlVV\nKbGUs6tyH2cH/mCDjYG093fyZt77SOBK1d5ffZiPCj/Dy2DiwZS7kf0SMWgNlLSV89usF3k7/0P+\ncf5fAc6Ty+9qjvK+8umQn0Wn0bExbh03x6wcsj2nMRedpCU9IAV3nTvl7ZXsrz7Mu/kf8Vj6/UP+\n8LeXfQvAbXE3E+UVQae1i7zmAl7Pew+9RsfpxlysDisSErOCZrA2egXxPjFDTpAlSeKWuLW8evZt\ndpTv5rH0odki3dZuXjzzBt22Hh5MuYckv4Qx36NE3ziSfRPIbymkzFKJh97IH7JfpbWvjcVhC5gf\nmkmEKeySZW5j0Ugaor0jifaOZH3sarqs3eys2MP+qkO8de4D9lQdZEvibST7XYiwH6s7xcGaI4R7\nhvLszMdc73u/vZ8u64UgqYpKj63XVcLUZesmPiQCL7sv/u5+rg9mVVXptffR1NPs+tuq6TRj1LmT\n6p9Eil8y7xZ8RHFbGZ39XcNOKM4fvBe3lZEZlMFtF9XLjnW1XSNpcNe708HoQSGNpCHaK5Jor0jW\nsWrU+2k1WjbGrePNc++zvWwXD6f9AHDW4ha1lTIjIMUVEBrtda7ktqst1COYms462vos+Ltfn6yC\niTTVplSeLx2bkzyxTaGyG8/S1mdhReQSbotfx8KwuXygbCW/pZCd5XvYkrRxyP17bX28kfcuPbah\nwc2j5lP808K/uaEnMdocdj5UtgLw7MzH2Fmxl7NN5zhce5wlEQs5VHuM4rYyZgXN4PG0+/lt9ksc\nN2cRZQpndfRy8poVPlS2uspHI0xhbIpfT6JvPK+ceYvCthJkv0TmhWTybsFHvHL2Lf5u3k/x0HtQ\n0FLEK2ffxq7aMXc3UNtlJsI0csbueHT0d7K1+CtS/JNGbeR6uO44tV1mFofNJ8rLeeV7Q+waPi76\nnJ0Ve7gradO4Xuu4OcsVLEvxS+Je+Y4hZdZXw4bYNRytO8n2sm9ZGDrHNYDhmPkUOY25JPrGXdc0\neEG4FlRVpajGgo/JQJDP5BoycrbxHOeaFQBXRravmw9/PedH4870OWHOBmB+aCbhplA8dEaK2kov\n+bhjdafIajhDWXuF6/g1whTGj2c9cVn7HLvDzjFzFqn+SZdV+lvf1cBnJdsBKGurosfW873LrK12\nKzqNbkjwo6mnmRdz3sDqsHF30u0097SQ1XCGE/XZaCUtP8x4mPSAoXOV9Fo9Kf7JnGnKo6G7adQs\nkSN1J2jqaWZ5xGLuTLyVP+a8zunGXN449/6IAwaaelo4Xp+Fj8EbT70HxW1lrvdqVeRSbk/YgEFr\nINY7mt9nv8yBmsNIEiT6xlNmqaDUUkFVZw12x9DTc51Gx4Mpd7t+7qLWEvZWfUeguz821c4XpTuo\nbK/i4bR7h1zQvZhDdVDeXkXWQPbT4JJkm8OGQ3Vg0BpYGbmEmYHp5LcUktVwhuyGM0iSxLPpjw35\nXcn+zvHtHxd9zhelFw4V3bXuyH6JrkBonHeMa3rXeUsjFnKiPovsxrOcacxj5kVZwLnNA0GhgEsH\nhfzcfZkTPJOT9af596P/hYpKvE8s80Nms718F9tKtpPTmMfDqfe4+hJ1Wrs4UZ9NoDGAVP+xe+mt\niVrOisglEzKA4ZKvqCjKYVmWT8myfBhwAD+RZfkxwKIoylbgVZyBIxX4laIoTbIsvwy8Lsvy/oHX\nePaa/QTfU3ZRE5auftbOixyzNrhyoP5+Yehc18l7e38Hb+V/gFbS8rPMp3kz7312Ve5nYegcQj1D\n6Ozr4qPCz9BrdDyWdv+IHwTnx1m/de4DDtUeo9RSzuPpD9DQ3cSuyv2UtzszP+K8Y7g5ZgUZgWlU\ntFfzwunXePPcB0iShs7+Lv5SuA0vvYm/ynxmSIf1BN9Y1kav4NvKfWwr2c5Pwx7hUO0x3lc+xaT3\n5D55C5a+dmo6aznTdI4vSneSGTzTtQNp6G6iprOOGQGproO8LYkbqeqo4VRDDka90ZWSWNxWhtJa\nTKp/MnE+MQA8NeNhnj/9imtKSaAxgIWhc1gQOodAY8Cov++ZgWlEmMI4WX+aW+LWEjJwAGt32Hkt\n9880dDexJno5iwdS9y7l1ri1FGaX8FHhZzT3ttBp7WJT/HrWx6y+Jld7PfUebEncyIqIxXxeuoOT\n9af5ffYrzAhIYXPCrThUB+8rn2DUufPDjEdcf1MABq1hyPfnRXDhBGSkLBNJkjDq3InyinCdQFws\nzV+mqK2UgpZC5oVmDrntu9pjHK07SbRXJI+kXd+UxYvNDZnFNxV7OWY+xc0xKwlw92Nr8ZdoJA1b\nxnkiNNHO7wzMXQ1TPig01aZU2h0Osoua8PE0kBAx+gHttxX72Fd9iL+d99w1CbaoqsqeqoNISKyK\ndJZ8BhoDeHbmY/yvw7/iUO1xbo1bO2QC4lHzSXpsvayOWsbC0LmAswRpW8l23jn3F34y+8kh/9u1\nnWYO1x7HTeeGj8EbHzdvZhoTkbj2V3hrO81oNVp83XzG1SNnR9HegSDJAuJ8YrhPvpOi1lI+Lf6K\ncFMY24q3Y9S5c2/yHei1ep7OeIRfn3ieT4u/4txAFqdG0rA6ahkd/V2crM/mpTNv4q51o9fex+yg\nGTyW/gB6jY7Gnia+qdjL63nvsS5mJS+deRNUlWURN3Gw5ghZ9TmXDAoVt5WR11zA0vBFQ8rNqjtq\neenMm7T2tZHdcIYk3/hhJ0Dd1h6+LP0Gd60bmxI2uLYvjVjE3qrvOFB9mGURN9Fn76eyo4r67kYy\ng2YOa0hZainnz/kfYdAaeCjlHuaGzL4m+zwvg4m10Sv4suwb/ueh/+ssdzZ4U95eibvW7bqnwQvC\ntdBo6cXS2c+8lOBJV757/nzh8fQH6LP1ca5F4XRjLvurD40rgOxQHZyqP41RZyQtIAWNpCHBN46z\nTedo6W0d9TimoKWIt/M/BCDA3Z80fxkVlZP1p/nNqT/y3KwnhzXtHYmqqvylcBvf1R4jyTeev5oz\nvlNHu8POW/kfDskQLWkrH1fWx2hqOuv4zan/xt/dj9VRy5kfmkmvrZf/Pv0nOqyd3Jt8J8sjndUD\nW5I2UmapxKA1EOUVPuLzzQhM4UxTHnnNBQR7DG/fYLVb+bp8N3qNng2xqzFoDfxo1hP8MedPZDec\nwdfNm7uTbh/ymN2VB3CoDu5IvJUFoXPosfVQ3l6Fl95E5KB1mAye/CzzaX6f/TL7qw+7+tpoJA1x\nvlFEe0YR5xNNlFcEb+S9x5G6E0R6hbMycgm9tl7eyf8LAI+l30+gMYDXc98lpykP88k/8LPMp0c8\n9jlUe4yvy3bT2ufMqvPQGQkddCFCI2mZFZjO0oiFrnIt2T+RzQm3UNlRjY+vEV/H8HPmlZFLiPKK\noNfWi5+7L35uvhh17pf8X9RIGu6X7+JXJ37HZ6U7hkzS7rdbUVqKCfUIJshj9PPSwdZELyer4Qx+\nbr7ckXgrmUEZSJLEnJBZ/EXZxqmGHP792H8R4hFEnE8MNocNq8PGioibLrkflCQJvTQxg03G9aqK\novzDRZtyBt32KfDpRffvBH7wvVd3HezNcvbdWZU5ev1pVUcNv896mV57HzvK97Axfj0LQjJ5+9yH\ndPR3siVxI4m+cdyTvJlXzr7Fh8o2fpb5NG+f/oQOayd3JNw6amRYkiQiTGH87dzn2Dowiu7/Hv+t\n6/aZgemsjV4xpH4yziea52Y/yQunX+ONvPdwqA5Mek9+PueZEUfu3Ra/jtzmfA7WHMF4Us83JQec\n9898ZkhqfLI5eyA741seSbsXwJUSN3istk6j45mZj/F89isDKYkSP0i+g6/LdgHOAMx57jo3fjLr\nSQ7VHiPRN5447+hx7Ug1koZbY9fyau477Cjfzf3yXRwzn2JP5QEaeprICEzjjoRbL/k85yX5JZDk\nG09RWykSEvfLWy7ZrOxqCDD683j6A6yOWsbW4q/IbS4gr1nBQ2fE6rDxRPqDo/5tXAupATKflX7N\nuYuCQs6TzwPoNTqemfnoiEGp60kjadgYv45Xzr7NV2XfEGkKp7m3ldVRy1wBwsku1GMgKNTdQFqA\nPMGrueam1JTK4moLnT1WVmZGoBnl86q+u5EvS3diU+3sqTw4LGPnaihrr6CivYqZgelDDlZ0Gh3L\nIxbzZdlOjtSdZFWU8yDToTrYV/UdOo2OdTGrXKnHEaYwitpKyWsu4EDNEVf6cklbOS+eeX1YVpEm\nV8NtcetYF7Ny2AGM3WFHI2ku+Tlu6Wvn89IdADwg3zXsKueRupP8eeBgE5xX+wKMfswPyeSmsPnD\nMhnb+iz8JfdLPPUebE50ZuP6uvmwJfE23lM+4bdZL2JX7TyYcrerZ4Cvmw8/zHiE32e9RH5LIfE+\nMdwnb3EFc9bFrGR72bdkN55lcdh87pO3uNa5KX49NZ115DUXUNBShEbS8MOMh0n2S+RY3UmyGs6w\nMX79qL+H0425vJH7LjbVzu7KAywJX8j62FWUt1fx1rkP6Lf3kxYgc65Z4fPSHTyadt+Qx28t/pJO\naxebE27B23ChfFSv0bEpfj1vnnuffzv6n0Mec6D6ME/OeIiMwDTX7+zVs++govLUjIcuOTL++1od\nvZymnhYqOqqw9LVjHijPfiT1XlcZsCDcyIqqBkrHIidfxmVZeyV6jY7MoAy0Gi0Lw+byPw/9H46b\ns9iccMslmxYXtpZg6e9gSfgCV5ZCkm88Z5vOUdRaysKwucMe41AdbC3+CoBfzP0x8T6xgPOYMswz\nlC9Kd/CbrD/yo5lPXHKC0s6KPXxX62xXUdRWSqmlgviBC8xj+bZyPxXtVcwLmc2i0Hm8kPMaxW1l\nVxwUcpYhf0KfvZ/67kbeLfiIz0u/xlPvSUNPE+tjVrsCQsBA8Cx2zOc8nz2U25Tv2l8PdrDmCG19\nFm6OXunafznLoJ7gP08+z76qQ8wLmU2st/N32NHfyZG64/i7+zE32NkOwqgzjpqF4mUw8bPMp/m6\nbBd+7r7E+8QS7RVBRGjAkAvMT2c8yq9PPM8nRV8Q7hnKCXM2zb2trI9Z7brY/9zsp9ha8hV7q75j\na/FXPJ4+tIG2uaueD5StGDR651TgkFnIfonjapotSRIx3lEEBYzcXkGSJFdZ2OUKN4UyPySTY+ZT\n5Dblu7KFitpK6HdYSQ8c/zXJaK9I/v2mf8BkMA3J6DHpPXlixoNkNszku5qjlLdXcrTuJAAGjZ5F\n4yzDnCjTesZubVMXBZVtpMb4ERYwcm1+Q3cT/336T/TZ+1kQOofshjP8Of8vfFGyA0t/O2kBsusf\nfGZgGjMCUshtLuAD5VO+qz3mTFsfR8q0XqvnB8mbkf0S+azkaxJ8YlkTvXzEfjoAcT4x/GT2U7xw\n+lX0Gj0/z3xm1Eldeo2OR9Pu4z9P/oFvSg7gqffgZ5lPD2uGdT4747g5i3UxKwn1DOF0Yy4aSUNG\nUNqQ+5r0nvxs9tM8f/oVDtQcoaW3lYLWIlL8klw7hfM89R6sixm9XGg0M4PSXR9K55oVOq1d6CQt\nS8IXsCVx02Vfdbwz8TY+UD5lXcxqMscoP7oWYryj+HnmM+Q1F7C1ZDvmrno2xKwelsJ4rUWYQvEy\nmDjXogwZKV/RUUVDdxNzg2dNmvKSmYHpxHhFkdVwhrNN+Zj0ntwSu/bSD5wkQgdlCk0D12VK5fUa\nSLD1UDkAq+ZHj/gYVVV5af/r2FQ7Bq2eQ3XHeGjeZjwNl1+COtaa3ik8CsCdGeuG3e8O77XsrNzD\nwdrD3D17PRqNhpM1OTT2NLM6bjHxEUOzWH625DF+ueN/s61kO4sTZtPY1cwLOa9ic9h5et4DhHmF\n0NrTRlN3KzuK9vFF6Q5KO0t5buFj+Bt9KWou45viAxypOoW7zo14/xgS/GNI9I8h3j8Gf6Mz08Xh\ncPBNyQHeP/sZPVZnsMnfy5vHMu9xraXKUstfCrfioTeyIHI2rT0WWnss1HXUs61kO1+VfcNNUXOZ\nHZZGe18nLT0W8uoVem19PDPvQeLCL+y7NgeuIaf1LHkNhcwIlrl95tDsz6CgGfwPr5/S3tfBoqg5\nQ/YbQUFezIpLoru/Bw/D8FKDv13+NP9j168xdzby14ufYkGks1x7bsRMjlSdolvfTqxf5LD3cU/p\nYV7L/TMGrYG7U29jb9kRDtQc5kjdcawOG246N3655BnmRczkH7/5D46bs7hjxs0kBsQCcKD8GIfr\nThDnF8UPMm8Z1vx+Q+BSzlnyaehqIsE/lgT/aPQaPa+eeo9Xzr7NTxY8io9/Jq/n/5n2/g4ey7yH\nZckjl6hdbX8TemE4RL+tn15bH97uQ/92J2uPNLEu4VKKqp39T5InWZPpPns/tZ1m4nxiXIFtnUbH\ngtA57Kk6SG5T/phl9wAn6gdKx0IuXDBM8nMOZyluGzkodMKcTXVnLfND5gw59pckiQ2xq/E2mHiv\n4BN+l/UicT4xF0p8fGIwDWpuf7TuJF+U7nRmXiTcwhvn3ufbin08M/PRIa9X22nmVEMO3gYv/AaO\nVbeXfYuPwZt7k+9AI2nRSJpLlry19Vn4QNlKS28rz2Q8NiST80DNEcrbK5kXMps7Em5lX/Uhvqs5\nhrm/noWhc9kUv36MZx6Zr5sPcX5RFLWV0mvrHZLd22vrZWfFXty17qyNWTHkce46N+6Xt/C77Jd5\nr+AT/n7ez9BqtOyrPoTVYWNN9PIxG1EP5m3w4l75zjHv4+fuy1MZD/P77Jd5+cyb9Nr7iDCFDbnY\nr9Vo2ZK4keK2Mk7Wn2Z11LIh/Qq3lWzHoTp4JO2+EYc1TKS10Ss4Zj7FNxX7yAhMQ5IkVz+hjHGU\njg02VnljZnAGmcEZOFQHdV31lFkqCPYIHLOJ9WQwrYNCe7NrgNGzhCx9Hfz36deGpAreHr+BL0p3\nctychbfBa0hKtCRJ3JO8GeVYMd/VHkMjaXgg9e5x/8MCzApKH/c/UbxPDP+y6O/QarRDPlxHEuUV\nwd1Jt3Ok/jgPJf9gxLT3wdkZX5Z9y12JGylvr0T2Sxzx+U0GZ2Do99kvkzvQ1PSWuKt30q6RNNwW\nv45Xz76NXXWwLmYVKyOXjNg5fjxivKP4+/k/v2rru1ySJDEjMJVU/2TM3Q2Ee47cof5a0kga0vxl\njplPUdNpdqW6HjdnAYza22IiSJLEpvj1vJDzGlaHlbuSNg2rE57Mgj2CkJBcV8unmWsypfJ6DCRQ\nVZVDOTV4uOkI83Eb8TGn6nM4W19Aqn8ysl8i20q2sy1nF+tiRw9+m7sa+FPunzEZTMR4RRLjHcXs\nmGQcXboR9xHNPS0cq84m0hROEKEjrmNBSCaHao+zp+AYs4JmsDX3GwBuClo0wv013Je8hVdz3+H/\n7P0DLX1taCUNT2c8Qoa3M+gf7BGG7AFr4pfw/HdvktOQxy+//v/wc/elurMWcJavoaoDEw3PuZ7d\nx+BNjHcUbX1tVHbUYNQZuSdpMwdrjrC9cA+B2iAWhs2lz97P/3/yFfrtVh5NvW/IiUq3tZujZmev\ntQMVxzhQcWzITzArNJUZXhnDfrb7E+9mj9tBVkctG3HgQIgmnBAjNDd1jfr+dDHy38bfZP6Yjv5O\ngt2CXK+b5pPKkapT7FaOcHvChiF/W7sq97O1+Cs8dR78ePYTxHpHszjgJo6ZT7G9bBcaScMzMx8l\nwi2M5qYuNsfdxu+yX+LV4x/wi7k/pr67kVdOvucsuZLvp62ll5Ga3z8qDx9v/NysH/LimTf4w7E3\n2F64h5LWChaGzmWe77wJbe7c2HHhtb/vYJBrZSqsa7IHj6ZC37mi6jbc9Foig8ff5P16qGyvQkUl\n9qJhAjeFzWdP1UEO150YFhTaU3kApbXE9X1hazF+br4kDMrEiDSFY9S5jxhk6bP183npDnQaHbcn\njBwoWRy+AG+DF5+VfD2k5w1AsEcg8d6xBBr92V6+Cw+dkedmP0mIRzD7qg9xpikPc1e9qwKiy9rN\nH3Ned5UkDfZg6j2uE+4Ev2hKWivptfUNa3StqirHzVl8VPQ5PbYeAH6f/TJ/PedZ/Nx9ae1t4/OS\nr/HQGbkraRPeBi/uTLyNDbFrKLdUkuyXcMVlg3PDMyhrraKgpWjIe7Gn6iCd1i5ui7t5xHOtJL8E\nFoct4HDdcXZXHWB5xE3srz6MSe/J4rDxtdC4HIm+cdyTtJkPC7eilbQ8mnbfsCwfjaRhS+Jt/D77\nFT4t/pK/ynwWSZIobC3mbFM+ib5xzAxMG+UVJk64KZSMwFTONuVTYiknwSeW3OYCjDrjsISGq0Ej\naYgwhX2v/oPX07QNCvX12zmcW4evycDspOHlO629bbx45g2aelu4JXatK1XQz92XR9LuZUPsGgxa\n/bDO4IHGANbHrOHLsp1slNcS7RV5TX+OywmQrIhczN2Z68c8wDifnZHdcAZ3rfPDdPYYU9HO16q+\nlvsO/u5+V5zWN5rZQTP4+3k/I9gjcEhk/Uam1Wgn9AMi1T+ZY+ZT5DcrRHmFY3PYOFWfg5fedMkG\naNdbin8SmUEZdNl6WBK+YKKXc1n0Gh2BRn/M3dMiU+i6TKm8HsrNHbS093FTeig67fBsxF5bL58U\nfYFOo+MHyXfgZfBkR/ke9lQfZFXU0mFZHeBsHv+n3D9T22WGLufBNwC5zslgJr0nvm7Ofj4+bt74\nGLyp7qxDRWV1sj+pPQAAIABJREFU1LJRD0JXRi51BoWqDuLv7k9RWykpfkmjjkSdHZzBwtC5HDOf\nwl3rzrMzH3NdCR7My83EDzMe4bvaY3xS9AW1XWZmB2WwLGIRsl8ikiTRae2iqr2Gio4qKtqdPffO\nDEzumh+SyZakjXgbvEgLSOY/T77Ae8onhHgGcbDmKOauelZELhl2kuKh92B11DJWRS6lsLWE2i4z\nPm7ezt+NwZvkqKgRAzt+7r7jbrp8uYw647CGpTMCUjBo9GQ15Ay5anyw5ghbi7/C182H52Y/5cre\n1Wq0LA5fwKKweaiqOiQImOQXT2ZQBtmNZzlSd5K9VQfpt/dfUWlxgm8sfz3nWV44/RolrRXEeEdx\nv7xl0vU+EaafqdB3rrPHSl1zN2mxfmg1E9Mfq7qjFqvDNqwUq7zdOc037qJyq3BTKDHeUZxrVmjr\ns7gywU/Vn+aT4i+HPf8tF/U80Uga14nz4McDfFW4m7Y+C+tiVo3ZN3FGYCozAlOdPW8sVc5JV+2V\nlFkqOWp2ltacb0txPgB0c8xKXjn7Nt9W7OfhtB+gqirv5P+F1r42Z2aKVyStfRZa+yyEeYaQPqhE\nPzU4maKWctf47/O6rd38Of8jcprycNMauH+gp+r28l38Lvtl/irzGT4s3EafvZ8HU+4ZUrZr1Lm7\nJopdqTlhGXyct52zzReythq7m/mmYh9eetOYVSV3Jt7K2eZzbC/7ltZeCz22HjbGrbtmrR6WRSxC\nK2nwcfMe9Xwl2S/RFWA503SOjMBUPh0oJdySuHHS7nfWxazibFM+31Ts5Y6EW2npbWVu8KzLSuCY\nqqZtUOhEQQM9fXbWzo1yHfj32vrIaczlmPkUha0lqKgsDV84bAoTMObB2vrYVcT7xLA4aRbNzaNf\nmZyMJEliU8J6Xjj9GkfqnBdxLlXi5GUw8ddzfnTN1hTtfW0Da9NNin8SEhLnWhTWxa5yleatilw6\n6T4UJUkacXzkjSLUM5izTfl09Hde99GS19mUmVI5eOpYQ3cjL515C5Pek5lBacwMTOdgzREs/e3c\nErvWtR9YFrGIbyv3cdycxZKIhcOe86PCz6jtMrM8YjGb4tdT1eEMpjRbm6hvb8bS1059dyNVA9k4\n5/kYvJgbMmvUtYabQkn1Tya/pZD3lU8AWB09drnyPcmb8Xf3JTN45pjBaUmSWBaxiIzAVCQ0+LgN\nzUIw6T1JDUgecqDc1meh324dsn8M9gji8fQHeDHndf6Q/Rq99l6ivSK4M3H0akFJkpD9E5H9E4ds\nnyyNig1aAxmBaZxqyKG6s47gYG+qOmr4uPBzPPUeo0770UiaQTl0F9yReBtnm87xbsFHgLOZ9Fjv\n+1giTGH8Yu6PyW47zQL/+SMGKQVhAtzwfeeKB0rHkq5y6ViXtRsJZwB6rBNppaWYF8+8jiRp+I+l\n/zykOX/ZQJPpizOFABaHzaeivYqjdSfZELsGS187HyrbMGj0/O28n+I/UAYjSZoRG/4n+saT21xA\ncWupqxdle38H2/J3YtJ7su6iicWjMeqMQ/YZ50tryi2VhJlChmRqZASmEeIRzIn6bDbGr+N0Yy5n\nm86R7JfInYm3jbkvSAtK4vOCbyhqKx0SFNpa/BU5TXkk+cbzcOoPCDD6o6oqDlR2lO/mP078nk5r\nF0m+8dx0DXq/xPtH42UwkddcgEN1ICHxgfIpVoeVh1LuHvPCt4feg3uSNvN63rscqDmMQWtgeeTi\nq77G8yRJGvFY5mJ3JNxKXrPCtpKv6LZ2U9VRw7yQ2UPKySabeJ9YEnxiyWsucE1P+z5NyaeSaRsU\nOpBTiwQsm+U8KC5oKeLVs+/Qa3emacf7xLIobC43hc2/7GinRtIg+yeimaArCd9Xil+SqylznHfM\npOkxI1wdXgYT0V6RlFjK6bX1TsrSsaki1COEs+Rj7mqY0kGhqTSlMquwEYNOQ0ykgT+cfpGm3hYa\nkCixlLkaaga6+w/pk7Yyagl7qw6yq2o/N4XPH3LAetycxeG6E0SZwtmSeBt6rd4V8BhcGqKqKr32\nXix97bT1tWPpayfKK+KSzRlXRS0jv6WQivYqQjyCLpntZ9S5s/EyeiJczuf/aPdND5DZnHAL20q2\n465154n0hyZk3OrVlBk8k1MNOWQ15JASFc2fcv+MTbXzdNp94x7/fF6g0Z/V0cv5pmIvEaYw7k78\nfllPgcYAHoi+Y1KWQwnT1nXpOwfXrvdczTFn4GV+ethVKdWz9LbzUe5X7Cr9DofqwE3nRoDRlxBT\nIOsTV5IZlu46/2hQ63jprHMUOkBlfxmLo52BC1VVqeyows/dh+TIqGHnLOt9l/Jp8Zccrz/Fg/Nu\n57WDb9Nl6+aJOfcyK+7SzecXaDLYVrKdqt4qbglaTkNXMx8XbKPX1seTc+4kOuzSk8VGE4IPsxl5\nn7UlfT0vnniHT8o+52x9Ad5uJn6x7Cn8jGPvkzytzhKvis4K1/vU0t3GsfoswryC+d83/2LI+dnj\nQXfhbtSxLX8neo2Onyx+hGCvK2tTcSlzIzLYV3aEDm0rtR31FLQWkRmWzoYZo2cEn7c+cAmnW3LI\nqsvl5oRlxIaP3Ef2cn2fv+WgIC/WNi119hFUPnFO255/N0Ge3///41qWw94z81b+4+AfOVl/2nkB\nLHku3m7jO0afrGW6V2NdN/ZR2RWqaeqiuMbCjDh/An2MKC3FvHTmTVTVwS2xa1kYOnfcY+mmIkmS\n2JxwC7/LfnncI9+FG0tqQDIVHVXOqy/N+YR6BI86xl64cq5m090NI5bpTCVTYUpla0cfdc3dzEjw\n4fVzb9PU28KG2DWsjFwykCKdR2V7Ffen3IVhUAaGr5sPC0LncLjuBGca81yp4fVdDbyvfIq71o0n\nZjw0ZtaGJEmuUqWRpkiOJs0/mVCPYMzdDayKWjppsmkutjZ6BUadOxGm8Cmxf00PkDFoDWQ1nKHz\nZAeNPc3cHL3SNWXmct0SuxZvgxezg2aI7B5hOrgmfefg2vWeyylsRCNJ+HvqvlfA1Wq3sq/6EDvK\n99Br7yXYI5BgYxBtfRZae9uo7agnuy6PFL8ktiRtxM0k8av9L2BXHdyRcCvbSrazt/gYSUZnFkxr\nbxutvRZmBaaP2FMNYHZQBsfMp/jtgdfJNueS4pdEpk/muH4Ok8MXN62BUzW5/LrzJbIazqCiEu8X\nzSzvWdcs+JzimYqvmw/Zdc4pyA+n3IutU0Nj59ivFxTkRZQpguKWCmrMzRi0Bj4t+hq7w87qiOUj\nVnCsDV2Np+qNt8GEvteDxt6r/zMFBXmR5JnIPo6wI/8ApxpyMGj03Bm7adT37WL3JNxJsCGYFSHL\nrsrv/Wr0UlsdtpID5cfotfexOmo5UreBxu7v95zXusdbpC6GcM9QarvMxHnH0Neu0jhKX8Hrua4r\ndbV6z03LoNDBHGeK/vJZ4RS2lvDimTdQVQc/zHhEpJANiPOJ4b+W/du4RggKN540f5kd5bvZWvwV\nNoeNhaFzJ239743sfFCofnpMILvhFVS0Ag46g49R31HFwtC5bIxbhyRJLA6fP2aQfG30CueY9YKP\n+LzU2Qu1s7/zivvDjJckSdwr38EJczYLQyfvuFNJklgasWiil3HVGLQGMgJSOdWQQ1NlM/E+MVc0\nlebC8+lHHFUsCFPEDd13rt9qp6yunagQE+6GsY+LSy0VKC3FrIleNqzni0N18Nusl6joqMJT5ywJ\nWhaxaEjpfm2nmU+LvyS/pZBfHf8dOq0Oh+rgqRkPMTMonWPmU+Q1F9Bj68Woc79QOjbGyPfF4Qs4\nZj7l6if3UOo9476AoNVoifeJJb+lkFMNbUSYwlgTtZwN6UtpbekZ13NcCZ1Gx9roFXxc9Dk3R68c\nUgp2KUm+8VR2VFNqqSDaK4Lvao/iY/Bm/igZ8ef38ddain8SWknLgZojgLP3TsBlZJZ6G7zYlLDh\nWi3vingZTNyVtGlgcvXlT5qeCJIksT52NW/kvTdm39zpZtqd8VttDg7nmjF56HAPaOHFnHdwqA6e\nFgGhYcTVyqkr1jsKo86dTmsXEhLzQzMv/SDhsoV4XMgUEia//MpW9NEF1NsrSfFL4oGUu8YdLA3x\nDGZF5GJO1efQbXVeqdZoNNwad/MV94cZr2S/RJL9Ei99R+GqmjNQQmYyePJ4+gOTriebIEwiN3Tf\nuXJzB3aHSlLk6KVLzT0tbCvZTlbDGQAcOIb1JM2qz6Gio4qMwFQeSb13xBHV4aZQfjLrSc61KHxa\n9CWNvc08MeNBV3/PucGz+LLsG8405rEwbC7llvP9hEYPCiX4xBJsDKShp4l7km8fc5z2SDbErsHb\n4MWC0DmuYQM67bU/hVwRuZgY76gReyWNJckvnt1VByhuK6W8vZI+ez+3xt084WXL7jp3knzjKWgt\nItorgpWRSyZ0PVfL4vAFLL7BhsHMC5lNkDGASFP4RC9l0phWQaEySyVf5B+kP7YCg1cXL561opW0\n/DDjYREQEqYVrUaL7JfE6cazJPnGX/YBgjA+Rp07vm4+mEWm0A0hv9qMLqGKYGMgT2U8fNmZkvck\nb+ae5M3XaHXCZDMjMJWVkUtYlbwQf0afviMI092N3neuqNo5Bj15hCbTNoeNr8q+ZU/VQWwOGzFe\nUTT3trCn8gArIhe7xow7VAdfl+9GI2m4O+n2EQNC50mSRHpACqn+yXj66Ohpd7humxPiDAplNeQ4\ng0LtlUhIY047liSJh9PupbqjhoWhcy/750/0jbvq04XHQyNpiL9ootp4JPjEOQeqNBfS3NuCUefO\nkvBLN06+HhaHL6Cuy8wDKfeICwkTbDI3xJ4I0yYoVN/dyPOnX6Hf3o/GJOHvHkiMTwRLwheIK6zC\ntDQ7aAanG8/ecNH9G02oRzAFrUX02nrHnC4hTKwmSw9tmioMksri8AWuqRSCMBqdRsc9yZsnbZ8B\nQZhMbuS+c0UDk8cSL8oUsjvsvJ77LjlNefi5+bI54Rbmhsxif/VhPi76nG8r9rkmLZ5uzMXc3cCi\n0HkEGsfXV00jaTC5edIzqN9JiEcQUaZwzrUU0tHfSWVHDeGmUNx1bmM+V7xPzBUFWG5EHnojEaYw\nKjqqAFgfs3rS7NPnhsy65tnDgnAlpkVQyOaw8Ubee/Tb++kvSyfePY1/fFA0UBamt3khs4nyCneV\nOAnXRoinMyhU390orkpMYkplG1r/esDZlFMQBEEQAGoau/A1GfA1XQi82B12Xs97j5ymPJL9Enl2\n5mOuke5Lwxeyq3I/+6sPsSpqKd4GL74u24WEs5fJ9zU3ZDZVJdvZXrYLq8N62eVV00GSbzzVnbXo\nNTpWRk2NMi1BuJYm55iSq+yzkq+p6qghTJKxN0axcpb48BQESZII9QwRDaavsdDzfYVECdmklldR\nj8a7iWD3kCkxHUsQBEH4/uwOB60dfQT6GAdts/PmufddJfiDA0Lg7Ml5a9xarA4bO8v3cKYxj9ou\nM/NDM6/K0IE5wTMB+K72KACx3tMjA+hyyP7OKpCbwubjbZicY8QFYTKZ8kGhvOYC9lQdJMQjiI7C\nZNwNWubKQRO9LEEQponBY+mFyUlVVfLbCpA0KvPDZk70cgRBEIRJorWjD4eqEujjLD/q7O/ijbz3\nyGo4Q4JPHM/OfHxIQOi8RaHzCDYG8l3tMbaVbEdCYkPM988SAggw+hPrHY1DdfYaEplCw80ISOXp\njEdd5XuCIIxtSgeFLH0dvH3uQ3SSli2xd9PYaiUl2g+DXjT2EgTh+gjzDAFEptBk1mjppce9BoDM\nYBEUEgRBEJyaLb0A+Hnr2VW5n389+muyG88S7xPLj2c9PmovH61Gy21xN+NQHTT2NDM3ZBYhnlev\nXP98Xxp3rbvr4pNwgSRJzApKxzBCwE4QhOGmbFCoz97Pq2ffotPaxR2Jt2FpdEb4U6LFlCVBEK4f\nk94TT50H5u76iV6KMIrc8gY0Po14afxcQTxBEARBaLL0ovFq5gQfsbX4KyQk7k66nZ9nPn3J4RFz\nQmYRYQpzZgnFrrmq65oTPNM5ncs3Bo00ZU/nBEG4TqZko2m7w85rue9Q1l7JgtA5rIxcwhtnCgBI\niRFjYwVBuH4kSSLcFEpFRzWqqooeTpPQydpcJHcHGQHpE70UQRAEYRKpt3RgSDpNj8POqsil3BK3\nFs8xxskPppE0/Gjm47T2Wa76BQdfNx9+Ofcn+Lh5X9XnFQRheppyQSGH6uCd/I8416yQHpDCQyn3\nIEkSBZWteLrriAw2TfQSBUGYZn6QfAfNvS0iIDQJqapKVV8xuMOymDkTvRxBEARhElG6ziDprSwJ\nWcHdyZffn8bP3Rc/92tTpSAmmgqCcLVMqaCQqqpsLf6KE/VZxHlH8+SMh9BqtDS19dBk6SUzKRCN\nOCkTBOE6CzeFEm4KnehlCCOobe7AbjJjcJiI8oqY6OUIgiAIk4TNYaNOk4tq17IubtlEL0cQBOGa\nmVJFqAUtReypOkioRzDPzrowDaCgsg2AlGhROiYIgiBcsL80B0lrJ9aYJDK5BEEQBJeT9aexa7vR\ntkXj7yHKtARBmLqmVFCotL0CgDsTb8Ok93RtL6hsBUQ/IUEQBOECq91KTttJAJZEZU7wagRBEITJ\nwqE6+LZiH6pDIqA/baKXIwiCcE1NqfKx+oGRz2GeF8o0VFWloLIVk1FPRJDnaA8VBEEQppFeWy8v\nn32bTl0dansgcyKTJnpJgiAIwiSR25SPubsBe0s4IZ4BE70cQRCEa2pKZQqZuxswaPT4ufu4tjVa\nemlp70OO8hX9hARBEAQ6rV08n/0qha3FONqCCW5bjlajnehlCYIgCJOAqqp8U7EPAFtdHIE+Y4+e\nFwRBuNFNmaCQQ3VQ391IiGcwGunCj1VQIUrHBEEQBCdLXwe/zXqJio4qZvnPpq9wNuEBXhO9LEEQ\nBGGSKLGUU9ZeQaQhHrXHSwSFBEGY8qZMUKi5pxWbw0aoR/CQ7a5+QtHXZhykIAiCcOPYWbEbc1c9\nqyKXstC0DtAQ5u8x0csSBEEQJok9VQcBiGQWAAEiKCQIwhQ3ZYJC5u56AEI9LwSFVFVFqWzDy0NP\neKDoJyQIgjCdWR02TppP42UwcWfibdS3dAMQFiD2D4IgCAK09Vk423SOSFM4tg7nBeUAH+MEr0oQ\nBOHamjpBoYEm04MzhRpae2jt6EOO9hOjhgVBEKa53KZ8umzdLAiZg1ajpW4gKBQqMoUEQRAE4Ejt\nCRyqg6URi2hp7wMg0FtkCgmCMLVNnaBQ90BQaFCmUP5A6ViqKB0TBEGY9o6ZnePnF4bNBaCuuRsJ\nCPEXV4EFQRCmO4fDwaHa47hpDcwPmU2TpReTUY+bQQwiEARhapsyQaH6rgY0koYgY6BrW3ldOwBJ\nkSIoJAiCMJ219baT16wQ7RVBhCkMAHNzF4G+7uh14oBfEARhujttzqO1r435IZm4ad1oae8V/YQE\nQZgWpkRQSFVVzN0NBBkDh4wVbmzrBSDYT1wFFgRBmM6+q3CWBCwMnQdAZ4+V9m6r6CckCIIgAPBt\nibPB9NKIRbR39WO1OcTkMUEQpoUpERRq7bXQY+slzHPo5LHGth58TQYMenEVWBAEYbpSVZV9ZUfQ\nSlrmhcwGwCz6CQmCIAgDWnvbyKrLJcYriiivCJranReWRVBIEITpYEoEhWrazcDQJtM2u4Pm9l6C\nfEWWkCAIwnRW3VlLpaWGjMBUTAZnZlBdcxcAYQEiKCQIgjDdHao9jqqqLI1YBECzxRkUChBNpgVB\nmAamVFAoZFCmUEt7L6qKCAoJgiBMc8fqTgGwMHSua5u5WYyjFwRBEMDusHO49jhGvTtzQ2YB0GQ5\nnykkziMEQZj6pkRQqLq9Dhg6eex8PyERFBIEQZi+bA4bJ+qz8XYzkR6Q4tpeNxAUChWZQoIgCNNa\niaUMS387y2MW4qY1AIMyhUT5mCAI04BuPHeSZfm3wCJABX6uKMqJQbdtBv4J6AM+UBTlhYHtDwJ/\nB9iAf1YU5aurvHYXV6aQx+CgUA8AQb7iw1wQBGG6auxpptPaxUZ57ZBBBHUt3Xi66/Ay6idwdYIg\nCMJEC/UMYWn4Qu5M24C907mtSZSPCYIwjVwyU0iW5RVAkqIoNwFPAs8Puk0DvADcCiwHNsmyHCnL\ncgDwL8BSYCOw+Rqs3aWm3Yy/u58rug+Dg0IiU0gQBGG6CvUI5qezf8i9Mza5ttnsDpraeggL8ESS\npAlcnSAIgjDRvA1e3J9yF/5GX9e25vZePNx0eLiP6/q5IAjCDW08n3RrgG0AiqLky7LsJ8uyt6Io\n7UAg0KYoSiOALMu7gbVAD7BLUZQOoAN4+pqsHui29tDW206avzxkuwgKCYIgCJIkkeKfhJvOgDOh\n1bl/sDtUUTomCIIgDKOqKk2WHkL9xD5CEITpYTxBoVDg1KDvGwe2tQ987SXLchJQDqwC9g3cz0OW\n5c8BP+BfFUXZPdaL+Pl5oNNd/uj4wqZGAOICIwkK8nJtb+3qx6DTkBgbMKFXggevabKYjGsCsa7L\nNRnXNRnXBGJdwlB1ribT4oBfEARBGKqzx0q/1SH6CQmCMG1cSU6kK8KiKIoqy/KjwOuABSgbdHsA\ncCcQA+yVZTlGURR1tCdtbe2+gqVAfm05AD6SL42NHa7tdY1dBPoaaWrqvKLnvRqCgryGrGkymIxr\nArGuyzUZ1zUZ1wRTY10ieHR1ucbR+4vJY4IgCMJQTaLJtCAI08x4gkK1ODODzgsH6s5/oyjKfmAZ\ngCzLv8KZMWQEDiuKYgNKZFnuAIKAhquz7AvMXfXA0HH0Xb1WuvtsJEX6XO2XEwRBEG5wZpEpJAiC\nIIyiWYyjFwRhmhlPUOgb4N+Al2VZngPUDvQKAkCW5a+BR4EuYBPwG8ANeFOW5V/jLB8zAU1Xee0A\nmLudcaah4+hFPyFBEARhZHUt3Wg1EoFiOqUgCIJwETF5TBCE6eaSQSFFUQ7LsnxKluXDgAP4iSzL\njwEWRVG2Aq/iDBypwK8URWkCkGX5Y+DowNP8VFEUx7X4AcxdDfi4eWHSXygDaGpzfpiLoJAgCIIw\nmKqq1DV3E+xnRKu55ABOQRAEYZppaT9fPuY2wSsRBEG4PsbVU0hRlH+4aFPOoNs+BT4d4TEvAy9/\nr9VdQr/dSktvK6lBiUO2i0whQRAEYSTtXf309NlIjfGb6KUIgiAIk1Bvvx0ADzcxjl4QhOnhhr5M\nqqLipjWQFpw0ZPuFoJBI+xQEQRAuEJPHBEEQhLFY7c7iBp32hj5NEgRBGLcbOgTupjXwq6X/THiI\n35ApY+eDQoEiU0gQBOGak2X5t8AinGXEP1cU5cSg2zYD/wT0AR8oivLCwPYHgb8DbMA/K4ry1fVY\na12LMygU6i+CQoIgCMJwNpszKKTXiaCQIAjTww3/aWfQ6pEkaci2xrZefDwNuOm1E7QqQRCE6UGW\n5RVAkqIoNwFPAs8Puk0DvADcCiwHNsmyHCnLcgDwL8BSYCOw+Xqtt6O7HwBfL9ErQhAEQRjufKaQ\nCAoJgjBd3NCZQiOxOxw0t/cSF+Y90UsRBEGYDtYA2wAURcmXZdlPlmVvRVHagUCgTVGURgBZlncD\na4EeYNfAJMsO4OnrtVjb+YN9URYgCIIgjMBqE+VjgiBML1MuKNTa3ofdoYp+QoIgCNdHKHBq0PeN\nA9vaB772kmU5CSgHVgH7Bu7nIcvy54Af8K+Kouy+Hou1irIAQRAEYQxWuwNJAq1GuvSdBUEQpoAp\nFxQSk8cEQRAmlOsoWlEUVZblR4HXAQtQNuj2AOBOIAbYK8tyjKIo6lhP7OfngU535WXBQUFe6PTO\n3V5IkBdBQV5X/FxXy2RYw0jEui7PZFzXZFwTiHVdrsm6rqnMZnOg12qGtacQBEGYqqZeUMjSC4ig\nkCAIwnVSizMz6LxwoO78N4qi7AeWAciy/CucGUNG4LCiKDagRJblDiAIaBjrhVpbu694kUFBXjQ2\ndtDe4dxHdHT00Kib2AP+82uabMS6Ls9kXNdkXBOIdV2uy1mXCB5dPVa7Q2STCoIwrUy9oJDIFBIE\nQbievgH+DXhZluU5QO1AryAAZFn+GngU6AI2Ab8B3IA3ZVn+Nc7yMRPQdD0WaxU9hQRBEIQxWG0O\ndCIoJAjCNCKCQoIgCMIVUxTlsCzLp2RZPgw4gJ/IsvwYYFEUZSvwKs7AkQr8SlGUJgBZlj8Gjg48\nzU8VRXFcj/WeHzUsDvgFQRCEkdjsDnHhQBCEaWVKBoV0Wg0+JsNEL0UQBGFaUBTlHy7alDPotk+B\nT0d4zMvAy9d4acO4Gk2LA35BEARhBFabA5NRP9HLEARBuG6mYFCol0AfdzSiOZwgCIJwEVf5mMgU\nEgRBuG5kWf4tsAhn1ujPFUU5Mei2zcA/AX3AB4qivDCw/UHg7wAb8M+Konx1PdZqszvEOHpBEKaV\nKfWJ191ro7PHKkrHBEEQhBFZbWLUsCAIwvUky/IKIElRlJuAJ4HnB92mAV4AbgWWA5tkWY6UZTkA\n+BdgKbAR2Hy91mu1iUbTgiBML1MqU6jJcr6fkPsEr0QQBEGYjM4f7ItRw4IgCNfNGmAbgKIo+bIs\n+8my7K0oSjsQCLQpitIIIMvybmAt0APsGhhc0AE8fT0WqqoqNrsqMoUEQZhWplRQqLHNOWo40Edk\nCgmCIAjDiQaigiAI110ocGrQ940D29oHvvaSZTkJKAdWAfsG7uchy/LnOKdU/quiKP+PvTuPj6q+\n9z/+muwLAQIkhB0U+CJiVVzqLu5LpV6ty+3trXq11+Wq9dr+7q231dalLa3Wai1tXVrbemurXkWr\ndakVBVtxQVAUhK+ssiRAIHsCySzn98eZCZNkMpkJSeZk5v18PHyYzJlz5p1J+H7OfM/3+z0Le3qh\n0tIicnKyex10eGkxAMWFuZSVlfT6OH3JKzk6U67keDGXFzOBciWrL3KlVadQS6sfQIvDiYhITLrV\nsIhIyrVyZk+dAAAgAElEQVQP1bTWOsaYy4FHgXpgY9T2kcAFwCTgDWPMJGutE+/AtbUtvQ5VVlZC\n1fZ6AJyQQ3V1Y6+P1VfKyko8kaMz5UqOF3N5MRMoV7KSyRWv8yitOoUid5XJy9UJv4iIdOXXSCER\nkYFWiTsyKGIsUBX5xlq7GDgRwBgzD3fEUCGwxFobANYbYxqBMmBnfwb1B90+J108EJFMklYtXvut\nhtWQi4hIDFpAVERkwL0KXARgjJkNVIbXCiL82MvGmHJjTDEwF3gtvM+pxpis8KLTQ4Bd/R3UHwgC\n6OKBiGSUtBop1KZOIRERiUNrComIDCxr7RJjzDJjzBIgBFxvjLkCqLfWPgs8gtsJ5ADzrLW7AIwx\nTwPvhA9zo7U21N9ZA+GRQrk5uhmBiGSOtOoUap8+th8LzImISPrSSCERkYFnrb2l00MrorYtABbE\n2Och4KF+jtZB+6yDbH2WEJHMkVZnxu1DPnXCLyIinYTCtxpWjRARkVgCQc06EJHMk1YtnqaPiYhI\ndwLhGqEFREVEJBZ/e53Q9DERyRxpdWa8b/pYWv1YIiLSB/yRK8BaU0hERGLYN31MdUJEMkdatXj7\n7j6mecAiItJRQKNJRUQkjsjFA40oFZFMklYtnm5JLyIi3dEVYBERiSegOiEiGSitWrw2LTQtIiLd\n8GsBURERiUMjhUQkE6VVi+f364RfRERi82uhaRERiUMjSkUkE6VVi+cPhsjJziLLpzsGiIhIR5pi\nLCIi8eiW9CKSidKqxWvzh9SIi4hITAHdfUxEROLQSCERyURp1eL5gyHdjl5ERGLSSCEREYlHdUJE\nMlFatXj+QFCNuIiIxKQrwCIiEk9kRGmO6oSIZJC0avE0fUxERLqju4+JiEg8GikkIpkorVo8d/pY\ndqpjiIiIB+nuYyIiEo8uHohIJspJ5EnGmPuAYwAHuMlauzRq2/nArUAr8IS1dn7UtkJgJXCXtfZ3\nfZg7Jr9GComISDf8WmhaRETiCAQ0fUxEMk+PLZ4x5mRgmrX2WOAq4IGobVnAfOBc4CRgrjFmfNTu\ntwI1fZq4G8FQiJDjqFNIRERi0rQAERGJR3VCRDJRIi3eacBzANba1UCpMWZoeNsooM5aW22tDQEL\ngdMBjDEzgJnAi32eOoY2vxpxERHpXkAn+yIiEodfC02LSAZKpMWrAKqjvq8OPxb5usQYM80Ykwuc\nAowOb7sX+EZfBe1JpGdft6QXEZFYdPcxERGJRyOFRCQTJbSmUCe+yBfWWscYcznwKFAPbAR8xpjL\ngLettRuNMQkdtLS0iJz9WCS6ZFih+/8h+ZSVlfT6OH3NS1kivJgJlCtZXszlxUygXOJqvwKsk30R\nEYkhEHQAdQqJSGZJpFOokn0jgwDGAlWRb6y1i4ETAYwx84BNwAXAAcaY84DxQKsxZqu19rXuXqS2\ntiXp8BFlZSVs39EAQDAQorq6sdfH6ktlZSWeyRLhxUygXMnyYi4vZoL0yKXOo76hK8AiIhKPPxAE\nNKJURDJLIp1CrwJ3AA8ZY2YDldba9k8yxpiXgcuBZmAucK+19omo7bcDm+J1CPUFTR8TEZF4dPcx\nERGJJxB0yM7ykZXl6/nJIiJposdOIWvtEmPMMmPMEiAEXG+MuQKot9Y+CzyC23HkAPOstbv6M3B3\ndAVYRETiUZ0QEZF4/IGQFpkWkYyT0JpC1tpbOj20ImrbAmBBnH1v71WyJLXpZF9EROLYd/ex3q9f\nJyIi6csfDOmzhIhknLRp9dqnj+XqZF9ERLrad6thTQsQEZGuAoGQaoSIZJw06hTSwnAiItI9TR8T\nEZF4NFJIRDJR2rR67Sf7uWnzI4mISB9Sp5CIiMTjD4Q0xVhEMk7anBm3rymkkUIiIhKDPxgiy+cj\nO0t1QkREuvIHNX1MRDJP2pwZa00hERGJJxDQtAAREeme6oSIZKK0afU0LUBEROLRFWAREelOMOQQ\nDDmadSAiGSdtWr22yELT6hQSEZEY/LoCLCIi3YjctCZHdUJEMkxOqgP0lfbpY2rIRUQGlDHmPuAY\nwAFustYujdp2PnAr0Ao8Ya2dH7WtEFgJ3GWt/V1/51SnkIiIdCeg9UlFJEOlTaun6WMiIgPPGHMy\nMM1aeyxwFfBA1LYsYD5wLnASMNcYMz5q91uBmoHKqrvKiIhId9r0WUJEMlTatHr7GnKd8IuIDKDT\ngOcArLWrgVJjzNDwtlFAnbW22lobAhYCpwMYY2YAM4EXBypoIBjSFWAREYnJr5FCIpKh0mj6mDsP\nWNPHREQGVAWwLOr76vBjDeGvS4wx04BNwCnAovDz7gVuAC5P9IVKS4vI2Y+O/0AwRGFBDmVlJb0+\nRl/zUpZoypUcL+byYiZQrmR5NVc6avNrTSERyUxp1CmkIZ8iIh7Qfnsva61jjLkceBSoBzYCPmPM\nZcDb1tqNxpiED1xb29LrUCNGDiEQdMBxqK5u7PVx+lJZWYlnskRTruR4MZcXM4FyJSuZXOo82n+B\noEYKiUhmSptOoTa/FpoWEUmBStyRQRFjgarIN9baxcCJAMaYebgjhi4ADjDGnAeMB1qNMVutta/1\nV0h/+x0qNcVYRES60kghEclUadMp5A9qpJCISAq8CtwBPGSMmQ1UWmvbL20bY17GnSLWDMwF7rXW\nPhG1/XZgU392CIFGk4qISHxtWlNIRDJU+nQK+YP4gBw15CIiA8Zau8QYs8wYswQIAdcbY64A6q21\nzwKP4HYcOcA8a+2uVOSMdArlZPt6eKaIiGQiXTwQkUyVPp1CwRC5OVn4fDrhFxEZSNbaWzo9tCJq\n2wJgQZx9b++nWB1EpgXoZF9ERGLxR6aP6QKziGSYtGn12gIhneyLiEhM+64Aa00hERHpSktRiEim\nSptWz+9Xp5CIiMTWPlJIV4BFRCSGyE1r9HlCRDJN2rR6keljIiIinekKsIiIxNN+l0pdPBCRDJM2\nrV6bP0iepgWIiEgMfr8WmhYRke5poWkRyVRpt9C0iIhIZ20BLTQtIpIqxpj7gGNw70R5k7V2adS2\n84FbgVbgCWvt/KhthcBK4C5r7e/6M2Nb+8UD1QkRySxp0eo5jqM1hUREpFv71orQiFIRkYFkjDkZ\nmGatPRa4CnggalsWMB84FzgJmGuMGR+1+61AzUDk9Ad18UBEMlNatHqBYAgHyFMjLiIiMfg1UkhE\nJFVOA54DsNauBkqNMUPD20YBddbaamttCFgInA5gjJkBzAReHIiQmmYsIpkqLaaP6QqwiIjE075W\nhKYFiIgMtApgWdT31eHHGsJflxhjpgGbgFOAReHn3QvcAFye6AuVlhaR08vPA5E6UV5WQllZSa+O\n0R+8lCWaciXHi7m8mAmUK1l9kStNOoV0BVhERLrXFj7Zz8nRFWARkRRrb4ittY4x5nLgUaAe2Aj4\njDGXAW9bazcaYxI+cG1tS69DRdaea2rcS3V1Y6+P05fKyko8kyWaciXHi7m8mAmUK1nJ5IrXeZQe\nnULhk31NHxMRkVj8kYsH2RpRKiIywCpxRwZFjAWqIt9YaxcDJwIYY+bhjhi6ADjAGHMeMB5oNcZs\ntda+1l8hIyOFNH1MRDJNenQKaaSQiIjE0aZbDYuIpMqrwB3AQ8aY2UCltbb90rYx5mXcKWLNwFzg\nXmvtE1Hbbwc29WeHEOiW9CKSudKiU6i1vVNIV4BFRKQrvy4eiIikhLV2iTFmmTFmCRACrjfGXAHU\nW2ufBR7B7ThygHnW2l2pyNl+kVlrz4lIhkmLTiG/Xz37IiLSPX9QdUJEJFWstbd0emhF1LYFwII4\n+97eT7E60EghEclUadHqRXr2taaQiIjE0n6XSl0BFhGRGPatKaQ6ISKZJS1avcjdAnJz0+LHERGR\nPhapEzm6eCAiIjG0dwqpTohIhkmLVk9XgEVEJB5NMxYRkXjaAkFysn1k+XT3MRHJLGlxdhxZaDov\nVwtNi4hIV+1rRejigYiIxOD3hzR1TEQyUkILTRtj7gOOwb0rwE3W2qVR284HbgVagSestfPDj98N\nnBh+jXnhReT6hT+guwWIiEj32qcZa6SQiIjE4A8GVSNEJCP12PIZY04GpllrjwWuAh6I2pYFzAfO\nBU4C5hpjxhtjTgFmhfc5G7i/P8JHtOpWwyIiEodGComISDxtGikkIhkqkZbvNOA5AGvtaqDUGDM0\nvG0UUGetrbbWhoCFwOnAm8DF4efUAcXGmH6b2xVZKyJPC02LiEgMkbtU5uRorQgREenKHwjpArOI\nZKREpo9VAMuivq8OP9YQ/rrEGDMN2AScAiyy1gaB5vDzrwJeCj/WrdLSInJyetdv1ObfBsCokUMo\nKyvp1TH6i9fygDczgXIly4u5vJgJlEvck/3sLB/ZWTrhFxGRrvyBIIX5eamOISIy4BJaU6iT9sus\n1lrHGHM58ChQD2yM3h5eb+gq4MyeDlpb29KLKK7I9LGW5laqqxt7fZy+VlZW4qk84M1MoFzJ8mIu\nL2aC9MilzqP95w8EdZthERHplj+g6WMikpkS6RSqxB0ZFDEWqIp8Y61djLugNMaYebgjhjDGnAV8\nBzjbWlvfR3lj0loRIiIST1sgpBohIiLdatP0MRHJUIm0fK8CFwEYY2YDldba9svbxpiXjTHlxphi\nYC7wmjFmGHAPcJ61tqYfcnew75b0ashFRKQrv18n+yIiElswFCIUcnTxQEQyUo8jhay1S4wxy4wx\nS4AQcL0x5gqg3lr7LPAIbseRg3vr+V3GmKtxF6F+yhgTOdRl1trN/fFDtI8U0gm/iIjE0BYI6mRf\nRERiCgQcQJ8lRCQzJbSmkLX2lk4PrYjatgBY0On5DwMP73e6BO27JX2/3eBMREQGsTZ/iKFFuamO\nISIiHuQPaikKEclcadHyRW41nKfefRERiSEQ1ELTIiISW2TWgeqEiGSitGj5/H5NHxMRke61+bXQ\ntIiIxKaRQiKSydKi5Wv1B/H5IDvLl+ooIiLiMcFQiGDI0YUDERGJSSOFRCSTpUXL5w8EycvJxudT\np5CIiHSkBURFRCSeQEAjhUQkc6VFy9eqWw2LiEg3NC1ARETiidSJnBxdYBaRzJMWZ8ht/qA6hURE\nJKbItADVCRERiUUjhUQkk6VFy+cPqFNIRERia78CrJN9ERGJoX1EqT5PiEgGSouWr9Uf0u3oRUQk\nJo0UEhGReDRSSEQyWVq0fH5/kNyc7FTHEBERDwqoU0hEROLQSCERyWSDvuVzHIe2gBaaFhGR2HSy\nLyIi8bTfkl4jhUQkAw36li/SiGv6mIiIxKKTfRERiUcXD0Qkkw36lk+NuIiIxKM1hUREJB7VCRHJ\nZDmpDrC/2vxqxEVEUskYcx9wDOAAN1lrl0ZtOx+4FWgFnrDWzg8/fjdwIm4dmmetXdBf+fxaQFRE\nROII6C6VIpLBBn3LFxkplKeFpkVEBpwx5mRgmrX2WOAq4IGobVnAfOBc4CRgrjFmvDHmFGBWeJ+z\ngfv7M6M/GAR08UBERGLTSCERyWSDvuXz+3WyLyKSQqcBzwFYa1cDpcaYoeFto4A6a221tTYELARO\nB94ELg4/pw4oNsb0W89+IOAAqhMiIhKbRgqJSCYb/NPH1LMvIpJKFcCyqO+rw481hL8uMcZMAzYB\npwCLrLVBoDn8/KuAl8KPxVVaWkROL0aFFhTuAmBEaRFlZSVJ79+fvJYnQrmS48VcXswEypUsr+ZK\nNxopJCKZbNB3CqkRFxHxFF/kC2utY4y5HHgUqAc2Rm8Przd0FXBmIgeurW3pVaCauj0A7Glpo7q6\nsVfH6A9lZSWeyhOhXMnxYi4vZgLlSlYyudR5tH+09pyIZLK06RTSLelFRFKiEndkUMRYoCryjbV2\nMe6C0hhj5uGOGMIYcxbwHeBsa219fwb0BzTNWEREuhfQ3YxFJIMN+pZv30ghLTQtIpICrwIXARhj\nZgOV1tr2S9vGmJeNMeXGmGJgLvCaMWYYcA9wnrW2pr8DBoLhNYV0BVhERGLQzAMRyWSDfqRQm64A\ni4ikjLV2iTFmmTFmCRACrjfGXAHUW2ufBR7B7ThycG89v8sYczXuItRPGWMih7rMWru5PzLqZF9E\nROLxhy8eaKFpEclEg75TSNPHRERSy1p7S6eHVkRtWwAs6PT8h4GHByAasK9O6GRfRERiCejigYhk\nsLTpFFIjLiIRP//5fVi7mpqa3bS1tVJRMZahQ4fxwx/eE3e/l156geLiIZx88ikDlFQGgl9rRYhI\nlOgasXfvXiZPnkRBQbFqRAbz65b0IhIl0+rEoO8UatOaQiLSyY033gy4DfP27Vu48sr/SGi/c8+d\n25+xJEW00LSIRIuuERs2rOeOO25L6C5fqhHpa9+IUl8PzxSRTJBpdWLQdwpFTvbzcnWyL+I1T72+\njqVrdvbpMY+aUc4lp05Ner/ly9/niSf+QEtLCzfccDMffLCMRYsWEgqFOPbY47nyyqv5zW8eYvjw\n4UyZciALFjyFz5fFZ59tZM6c07jyyqv79OeQgdO+0LQ6hUQ8xyt1QjUis/kDIfJysvD51Ckk4jWq\nE/0vDTqFwiOFNNxTRHqwfv06/vSnBeTl5fHBB8v45S9/TVZWFpdccj6XXvovHZ77ySer+OMfnyEU\nCnHxxXM93ZBLfJpmLCKJUI3IXIFgSDVCRHqUrnVi0HcKtU8f00ghEc+55NSpvRrV01+mTp1GXl4e\nAAUFBdxww9VkZ2dTV1dHQ0NDh+caM4OCgoJUxJQ+poWmRbzLS3VCNSJz+QMhcnO1FIWIF6lO9L9B\n3ymkkUIikqjc3FwAtm+v4sknH+fRRx+nqKiIr371ki7Pzc7WyWG6aF9TSHVCROJQjchcGikkIolI\n1zox6Fu/fWsKDZ43XURSq66ujtLSUoqKirB2Ddu3b8fv96c6lvQTfzBETraPrCytFSEiPVONyDyR\nNYVERBKRbnVi0Ld+keljashFJFHTpk2nsLCI6667koULX+X88y/k3nt/nOpY0k8CAUd3qBSRhKlG\nZB53pJDqhIgkJt3qRNpMH8tRp5CIdHLuuXMpKyuhurqR2bOPZPbsIwF3OOdPfzo/7r6R5wK8+OLC\nfs0p/cuvaQEiEkP0rYNVI/qXMeY+4BjAAW6y1i6N2nY+cCvQCjxhrZ0ffvxu4ETczyvzrLUL+iuf\nP6A6ISJdZUqdGPStn18jhUREJA5/IKgaISKSIsaYk4Fp1tpjgauAB6K2ZQHzgXOBk4C5xpjxxphT\ngFnhfc4G7u+vfI7juNPHtBSFiGSoQX+WrFsNi4hIPLqrjIhISp0GPAdgrV0NlBpjhoa3jQLqrLXV\n1toQsBA4HXgTuDj8nDqg2BjTLw15MOTgoJsRiEjmGvTTx9oCQbKzfGRnqSEXEZGuAkFHI4VERFKn\nAlgW9X11+LGG8NclxphpwCbgFGCRtTYINIeffxXwUvixuEpLi8hJcm2glr3u4rC5uVmUlZUkte9A\n8GImUK5keTGXFzOBciWrL3Il1CnUy3nA3e7Tl/x+DfcUEZHuaaSQiIintN8K0lrrGGMuBx4F6oGN\n0dvDnzOuAs5M5MC1tS1Jh2lsaQPcWQfV1Y1J79+fIusieo1yJceLubyYCZQrWcnkitd51GOnUPQ8\nYGPMQbiN9rHhbZF5wLOB3cDLxpjngAO726ev+YMh8nJ1BVhERGLzB0KaFiAikjqVuCODIsYCVZFv\nrLWLcReUxhgzD3fEEMaYs4DvAGdba+v7K9y+9Ul18UBEMlMiZ8m9mQccb58+1aaRQiIi0o1gKETI\ncXTxQEQkdV4FLgIwxswGKq217Ze2jTEvG2PKjTHFwFzgNWPMMOAe4DxrbU1/hvMHtT6piGS2RKaP\nJT0PuId9YurNHGCAoONQkJOd1nP8+poXM4FyJcuLubyQ6dJLL+W2225j1qxZ7Y899tjDlJaWcuWV\nV3Z47rvvvsvjjz/OAw88wHXXXcevfvWrDtv/8Ic/UFtby4033hjztdasWUN+fj5Tpkzh5ptvZt68\neRQUFCSc1QvvV7rbdzMCXTwQEdc11/wbN9/838yYcVD7Yw8+OJ9hw4bz5S//a4fnLl/+PgsWPMX3\nv383t9zyDX70o5922P7MM09SV1fHVVddE/O11q1bS15eHhMnTuJ73/sfvv3t75Gfn3idSAfW2iXG\nmGXGmCVACLjeGHMFUG+tfRZ4BLfjyMG99fwuY8zVuBefnzLGRA51mbV2c1/nC+imNSLSSabVid4s\nNJ3wPOBY+3SnN3OAAVrbApSW5A/6OX4DxYuZQLmS5cVcXsk0Z87pPPPMnxk9ehLg5nrppVf4+c8f\n7JKvrq6F1lY/1dWN3Hnn3V22NzXtpbm5tduf67nn/sKMGTMZMmQU3/72nTQ2+mls9CeUs6/mAEt8\ngaADoJFCItLujDPO4vXX/9bhZH/Rotf5+c8fjLtf5xP9RCxe/DozZsxk4sRJ3HHHvKT3TxfW2ls6\nPbQiatsCYEGn5z8MPDwA0dpHCmnmgYhEZFqdSKRTqDfzgAvi7dOX2vwhzQEW8agF6/7CBzs/7tNj\nHl5+CBdOPa/b7aeddibXXXcV//EfXwdg5cqVlJWVsWnTRm699Vvk5uZSUlLCnXf+qMN+X/jCabz4\n4kLef/89HnjgXkaMGMnIkaMYO3YcgUCAH/zgdqqrd7Jnzx6uvPJqKirG8Oc/L2Dx4tcpLS3lu9/9\nHx577EmamhqZN+9O/H4/WVlZ3HLLbfh8Pn7wg9sZO3Yc69atZfp0w7333t2n74vEprUiRLxNdUJ1\nItUCAffigUYKiXiT6kT/14lEWr+k5wH3tE9fCYUcgiFHPfsi0q60dARjx47jk09WAvDyyy9zxhln\n09jYyPe+933mz3+YoqJi3n337Zj7P/TQfG677S7uv/+X1NfXAdDY2MDRRx/D/PkPc+ed8/jNbx7i\nwAOn8vnPH8s119zAzJn7pqr9+tcPct555zN//sNccMFFPPqoe6HT2tVcc831/PrXj/H222/R0NDt\nbFrpQ/6AewdjneyLSITqhETbVyf0eUJEXJlWJ3ocKdSbecDArs779EnaTtqvAGtagIgnXTj1vLi9\n8P3ljDPOZuHCvzFz5ixef/115s//NevWfcqPf/x9gsEglZXbOOKIoygqKuqyb1VVFdOmTQfgsMNm\n09raSknJUFavXsXzzy/A58uioaH7m6BYu5prr70BgNmzj+R3v/s1AOPGTWDkyFEAjBpVRmNjI3l5\n/bL+vkTxa60IEU9TnVCdSDV/ZJqx6oSIJ6lO9H+dSGhNoWTnAXezT5/THGARieXkk0/hscce5Ywz\nzmLy5MkMHTqUefPu4p577mfy5Cn89Kc/7nbfrKx9J4WO454o/u1vr9DQ0MAvfvFrGhoa+NrXvhrn\n1X3t+/n9AXw+93jZ2R3bqchzpH+pTohILKoTEqGLByISSybViUHd+rX53eGeWitCRKIVFRVz4IHT\neOyx3zJ37lwAmpubGD26gsbGRpYvX4bfH3tB6FGjyti8eROO4/DBB+5NFOvq6hgzZixZWVksXvx6\n+74+n49gMNhh/4MOmsny5e8D8OGHyzosUCcDT2tFiEgsqhMS4Q//fnJ18UBEomRSnRjUZ8lFBTkM\nKcxl8lgNrRWRjs4442yWLn2XU089FYALL7yY6667irvv/gFf+cpl/OEPv2P37l1d9rv66v/g1lu/\nxbe+dTPl5aMBmDPnVJYs+Ts33XQdhYWFlJeX89vfPsKhhx7O/fffw/vvv9e+/9e+di2vvPISX//6\ntbz00l+6vf2kDIzhQ/LIz81mYoXqhIh0pDohAKOGFZKXk8WE8iGpjiIiHpMpdcLnlaGp1dWNvQoS\nDIWoGD3ME7fC7swrt+iO5sVMoFzJ8mIuL2aC9MhVVlbi6+c4g0K61Yl0+NscSMqVOC9mAuVKlupE\n8lQnBoZyJceLubyYCZQrWX1VJwb1SCGA7KxB/yOIiEg/Up0QEZF4VCdEJJOpBRQRERERERERyUDq\nFBIRERERERERyUDqFBIRERERERERyUDqFBIRERERERERyUDqFBIRERERERERyUDqFBIRERERERER\nyUDqFBIRERERERERyUDqFBIRERERERERyUA+x3FSnUFERERERERERAaYRgqJiIiIiIiIiGQgdQqJ\niIiIiIiIiGQgdQqJiIiIiIiIiGQgdQqJiIiIiIiIiGQgdQqJiIiIiIiIiGQgdQqJiIiIiIiIiGQg\ndQqJiIiIiIiIiGSgnFQH2B/GmPuAYwAHuMlauzTFeWYBfwbus9bON8ZMAP4XyAaqgK9aa1sHONPd\nwIm4v+t5wFIPZCoCfgeMBgqAu4AVqc4VzlYIrAxnWpjqTMaYOcD/AavCD30M3J3qXOFsXwH+GwgA\n3wU+SnUuY8xVwFejHjoSOB74FW478ZG19rqBzBTONQR4DCgF8oE7gO2pzpUJVCcSyqQ6kVw21YnE\ns6lOJJ5LdSJFvFQnvFgjwrlUJ5LLpjqReDbVicQy9WuNGLQjhYwxJwPTrLXHAlcBD6Q4TzHwc9x/\n+BF3Ar+w1p4IrAOuHOBMpwCzwu/R2cD9qc4UNhd431p7MnAJ8FOP5AK4FagJf+2VTIuttXPC/93o\nhVzGmJHA94ATgPOA872Qy1r7m8h7Fc73e9y/+5ustccDw4wx5wx0LuAKN549BbgI+JlHcqU11YmE\nMqlOJE91IgGqE0m7AtWJAeelOuHFGhHOpTqRPNWJBKhOJOUK+rFGDNpOIeA04DkAa+1qoNQYMzSF\neVqBc4HKqMfmAM+Hv34BOH2AM70JXBz+ug4o9kAmrLVPWmvvDn87AdjqhVzGmBnATODF8EMpz9SN\nOaQ+1+nAa9baRmttlbX2ao/kivZd4MfAlKirfqnKtQsYGf66FPdEwQu50p3qRM9UJ5KgOpEU1Ynk\nqE6khpfqhBdrBKhOJEV1IimqE4nr1xoxmKePVQDLor6vDj/WkIow1toAEDDGRD9cHDXcbScwZoAz\nBcYOao4AACAASURBVIHm8LdXAS8BZ6UyUzRjzBJgPG7P8GseyHUvcANwefj7lP7+osw0xjwPjMAd\nKuiFXJOBonCuUuB2j+QCwBhzFLAFdyhqbdSmlOSy1j5hjLnCGLMO9/2aC/wi1bkygOpEz5lUJ5Kj\nOpG4yahOJEx1ImU8Uye8WCPCuVQnkqM6kbjJqE4kpL9rxGAeKdSZL9UBepCyfMaY83Eb8Rs6bUrp\ne2atPQ74IvCHTlkGPJcx5jLgbWvtxm6ekqr3ai1uw30+bnH5DR07c1OVy4fbW30h7nDG35Li32En\nX8OdZ95ZSnIZY/4V2GytnQqcivs3Hy3V71em8Pr7rDrRiepEQlQnekd1QmLx8vuc0myqEz1TnUia\n6kSC+rtGDOZOoUrcnvyIsbiLUXlJU3ihMYBxdBwOOiCMMWcB3wHOsdbWeyTTEeGF87DWfojbKDWm\nONcXgPONMe/gNgC34YH3ylq7LTw81rHWrsddUKw01bmAHcASa20gnKuR1P8Oo80BluBe8RsZ9Xiq\nch0P/BXAWrsCKARGeSBXulOdSIDqRMJUJ5KjOpEc1YnU8HqdSHkbA6oTSVCdSI7qROL6tUYM5k6h\nV3EXWcIYMxuotNY2pjZSF68BXwp//SXglYF8cWPMMOAe4DxrbWSxs5RmCjsJ+CaAMWY0MCTVuay1\nl1prj7LWHgP8GvduASl/r4wxXzHG/L/w1xW4d1j4bapz4f77O9UYkxVeJC7lv8MIY8xYoMla22at\n9QNrjDEnhDdfmKJc64DPh/NNwi16qz2QK92pTvRAdSJxqhNJU51IjupEani9TqT834zqROJUJ5Km\nOpG4fq0RPsdx9jthqhhjfoTbIISA68O9ZqnKcgTuHNLJgB/YBnwFd8hZAfAZ8G/hP6qBynQ17tzM\nT6Mevhy3kUpJpnCuQtxhixNweznvAN7Hvc1eynJF5bsd2ITbG5vSTMaYEuCPwHAgD/e9+iDVucLZ\nrsEdRgzwfdzbk3oh1xHA962154S/nwk8hNsJ/q619hspyDQEeBS3COfgXjnanupcmUB1osdMqhO9\ny3c7qhOJZFOdSDyT6kSKeKVOeLFGhHOpTvQu3+2oTiSSTXUisTz9WiMGdaeQiIiIiIiIiIj0zmCe\nPiYiIiIiIiIiIr2kTiERERERERERkQykTiERERERERERkQykTiERERERERERkQykTiERERERERER\nkQykTiERERERERERkQykTiERERERERERkQykTiERERERERERkQykTiERERERERERkQykTiERERER\nERERkQykTiERERERERERkQykTiERERERERERkQykTiERERERERERkQykTiERERERERERkQykTiER\nERERERERkQykTiERERERERERkQyUk+oAIt0xxjjAeiAIFAMfAj+w1r6dwkyTgbXhXP9irV0efrwY\nWAX8zlp7e/ixU4GfAEOAz4B/s9ZuDW/7T+Aa3I7ZvwP/AZwF3AMYa61v4H4qEZHUyrT23lrb1sNr\n+4D/B/wQOMVa+4+obf8M3ArkAiuBK6219eF95gEXAA7wrLX2f4wx1wL/BWRbayfv59siIpISqhPd\nZjgUeB84w1q7KN7xjDF5wC+Bk3Dfx19Zax8wxtwFXA5ssNbO2Y+3RAYpjRQSr5tjrTXABOD3wJ+N\nMSelONM2a+2MSMMfdnv0E8LF4Anga9ba6cALwIPhbccANwHHAjOA4cDXrbUvWGtnDEB+EREvypj2\nPoHX/RUwHdjZ6bUmAj8Hzg2/V5uAH4Q3XwrMAT4X/m+OMeYia+2DwGmJ/bgiIp6mOtHxuFm49WJ7\n1GPxjvcNYET48c8D/2mMOdJaextwWeI/sqQbjRSSQcFa6wD/Z4wZBvwIOM4Yk487suZsIA942Fr7\nQwBjzCbcK6ZX4RaOP1prv2mMycFthE8EsoGPgCustQ3GmPOB7+NefViH2+O/q6dsxpjP4Z5w/yHq\n4VNxe9sjBeJR4CfGmBLgYuBJa21deP9Hge/hXj0QEcloau8B+L219u3wzxbtfGChtXZz+PvfAG8A\nN4Rf63fW2tbwa/1v+LGne/q5REQGE9WJdtfijpiKHlUU73gXA9+x1oaABmPM0+HH3k/gtSSNaaSQ\nDDbPA583xhQC/w3MBA4BDgYuMsacF/Xck3B7yY8AbjTGjMedojUFt4d8Gu7QzmONMQcA/wt82Vp7\nAO5J9oM9hQkP1/8V7gl5IGrTdNyhpABYa5uA3cDUztvCX2uEkIhIRxnb3seZDhHreOXGmNLevpaI\nyCCWsXXCGFOBOyLo2502xTue6oTEpE4hGWwacP9uS4C5wC+tta3W2mbgMeDCqOf+0VobtNZWAjtw\nrwxU4xaMC4Aia+1t1tq/4l5VWGStXRne90Hgi8aY7B7yXAN8Yq1d0unxImBvp8f24F5t6Lwt8riI\niOyj9r6rDscLjwpy+um1RES8LpPrxP3AnZERQXFeK/p4qhMSk6aPyWAzGfADdbhzZO8zxvwwvC0f\neC/qufVRXwdxF9l82xhzI3Aj8HtjzAu4izwPB04yxqzptP9IOq3pEGGMGQ3cjHvVobNmoKDTY0VA\nU4xtkcdFRGSfyai9j/taxpgCwNdPryUi4nWTycA6YYw5CxhprX08gdeKPp7qhMSkTiEZbC7C7blv\nM8ZUAj+x1v4lmQNYa58GnjbGjMCd0/tfuHcOeM1ae1EShzodKAc+McaAezeByHDO53EX/ST82DCg\nNPw6a3CHi0ZMAz5J5mcQEckAau+7WgOc3Ol4VdbauvCHl6nA3/rotUREvC5T68QFwOHGmMgC0yOA\nBeG7jsU7XmTb2iReSzKApo/JoGCM8RljLgL+k31zZ/8MfM0Ykx3efqsx5uwejvNvxpjbAKy1NbiN\nowP8FTgxPIcYY8zRxpifxTuWtfZxa22ptbbCWluBu4DbT6y11+LOPZ5kjDkh/PSbgb+Eh7M+BXzZ\nGDM6vMDdTcCfkn1PRETSkdr7uP4MnGbCnzhw7yQTOd5TwNXGmGJjzBDg6v18LRERT8r0OmGtvdZa\nOyrqtZYAF1prH+vheE/hrqeUbYwZA/wz8GS815LMoE4h8bpF4auflcB1wBestZEV8n8BfIa7KNwa\n4CDgHz0c78/AEcaYtcaY1bjziH9qra0C/h14Nvz4fPajkbTW7sFtaH9hjFkHHANcH972Pm6h+Duw\nGvgUd1E6EZFMllHtvTHmgvBdYbowxqwMvxfjgMeNMWuMMUdba7fhTm14zhizFnfo//fCr/U08Aru\nnWiWA89Ya1/o7c8lIuJBqhM9v1a8zxk/w33vLG5H1Z3W2hW9/bkkffgcx0l1BpFBwxgzGXeY6uR+\nfA3HWuvrr+OLiEjP+ru9N8bkAr+x1l7WH8fv9FqT6efaJSKSadKsTswBbrfWzunv1xLv0UghERER\nkYE3EfhlqkOIiIhnqU7IgFCnkEjyxoWH8s/uy4MaY+Z2usuBiIikVr+09wDW2vXW2nf6+ridGWOu\nBRb29+uIiGSodKgTdwGP9ffriHdp+piIiIiIiIiISAbSSCERERERERERkQyUk+oAEdXVjb0eslRa\nWkRtbUtfxukTXszlxUygXMnyYi4vZoL0yFVWVqKFx0m/OuHFTKBcyfJiLi9mAuVKlupE8lQnBoZy\nJceLubyYCZQrWX1VJ9JipFBOTnaqI8TkxVxezATKlSwv5vJiJlAucXnx/fZiJlCuZHkxlxczgXIl\ny6u50pUX328vZgLlSpYXc3kxEyhXsvoqV1p0ComIiIiIiIiISHLUKSQiIiIiIiIikoHUKSQiIiIi\nIiIikoHUKSQiIiIiIiIikoHUKSQiIiIiIiIikoE8c0t6ERERERFJT8aYWcCfgfustfM7bTsd+CEQ\nBF6y1t4Vfvw+4BjAAW6y1i4d2NQiIukvoU4hNeIiIhKP6oSIiHTHGFMM/BxY2M1THgDOArYBi40x\nzwBlwDRr7bHGmIOAR4FjByKviEgm6XH6WIKN+JeA44EzjTEzjTEnE27EgavCzxERkTSkOiEiIj1o\nBc4FKjtvMMYcANRYa7dYa0PAS8Bp4f+eA7DWrgZKjTFDBy6yiEhmSGSkUKQR/1bnDdGNePj7SCNe\nRlQjbowpNcYMtdY29Fly8TTHcdhe08Inm2r5ZFMNG6saCIacbp+fleUjFGd7X8nLyebGLx3CxNEl\ncZ+3bVczT7+xjpqmVva2BggEHQLBUL/nS1Qy79dxsyq49NRpMbf5A0HWb2tgzeZa7OY6qnY309vf\nwkD9DpPVl7lmTRnBv889OOY2fyBIVpaP7Kx9fe276/eydmsdn26tZ2hRLv904gF9ksODVCfEs1rb\ngoQch8J878yYdxyH9dsaqK7fQ6s/SFtbEJ/Px3GHVFBckJvqeP3OcRy27Gxi/bZ6qna3UFXTwo6a\nFlr9wZTk8Ur9mj29jMvPnpHqGP3CWhsAAsaYWJsrgOqo73cCBwKjgGVRj1eHn9ttnSgtLSInJ7vX\nOcvK4p8fpoIXM4FyJctruZpa2ti4s5mVG3azasMu8nKzue3Kz1PkgRrktfcqoqysBH8gyM7aPYwZ\nWUxWlq/HfZr2+PEHgpSWFMTcHgyG2Fm7h+27m9m+u5m2QIhzj5tCbk7iSz/3xfvV4xmSGvH948Vc\nsTI17/Gz9JPtjBpeyMEHjMTni/9HvnL9Lpat2UlpST4jhxUyclgBDS1trN9Sx7qt9azbWktNQ2v7\n80cNK6CkOLUn5P5AiO27W/h4Uy1HzBrbzXOCPPXaWp5+/VMCQYeSojzycrPIy82mqCCXrEG2NPvO\n2j0s/rCS6y46jOzsjuE/2bib7z78Nq1t7km4zwcVI4vJye65gctENfV7eXvVDq750qGMHFbYYduG\nbfV882d/JxAMUZifw5CiXEIhh931e9ufM3nMUP79wkM77OfF9qE3VCd6z4uZYP9zBYIhnnljLVu2\nNzF1wjCmTyzlgHHDKMjbvzqQTK7ahr08//cNvLRkI9lZWfz3V4/gsOnl+/X6+5srEAzxjxWVPPvG\nOjZU1nfZ/snmWu68+riETjQTzfTJxt28t2o7xYW5DBuSz/Ah+UybOLzbE9Tu7G0L8Pr7W1i9qQa/\nP4Q/ECIQDHHEjHLmnnhAj+cNADtqWli8oopFy7ewZUdTh20jhuYzvCQ/qUzppmxEcZe/Ja+2Ef2s\nuz+mHv/Iamtbev2iZWUlVFc39nr//uDFTKBcyYrkqtrdzKqNNZx82Fhy9+N8Zn84jsOf/7GRvyzZ\nROe+8HseW8p1/zQrofY8GXtaA6xYt4ula3ayobKB4w6p4IITDyAnu+sHq1T+DgPBEGu31vPh2l3s\nqt/DkMLc9v/2BEKsWr+bLTsbCQQdDppUytcv+hz5ubF/j/5AiFeXbuaFJZvw+0PMnFzK8Z8bw+xp\nZbT6g3y4dhfLP61m1abaLgMPPt1Uw+Vnm4R+D8m8X/HqSV9/SlcjHqU3uRzHYfOOJsaMLCIvxh/Z\n6k01fLyxhr2tAfa0BdnbGmDYkDwOmjSCgyaVMrQ4L6FMjuMQDDms3VrPPz6qZJmtpi3g/kEeOHYo\nXzh2Mp+bOpKsGH+Miz7cxh/++ikhp/urasOG5HHUjHJmTi7loMkjKB9e2O1zo3P1p9a2IDfc/yYf\n2J184fMTu2zfWNXAIy98wvaaFkpL8vnXM6Zz5vEHDOq/rcdeWcOiDyt5f2UVB4ztOOL61SWbaG0L\ncuLnxnDYtFFMnzB8v65Qp9O/w1hefW8zT7y+jsVLN3PioR07FV/+xwYCwRCTKkoIhRxa9gYIhkIc\nPm0U08YPZ/qE4UwcPaRDjr5qxAch1YkwL2aC/c+1q24PDz2/ivWVbv/e4g+2ApDl83HBSVP4wrGT\nezzG2q11rNpYQ8WIIsaMLKZiRBHjxw1PKFddUysvvLWJv39URSAYoqQol5a9fr778NtcPGcqZx09\noU9PeBN5v6rr9vDe6h0s+mAbuxta8fngqBnlHDS5lPzcbPJzs1n04TZWrN3FE6+s5oyjJnR7rJDj\n8NZHVeTmZHH0QaNjdiCVlZVQWVXPs3/fwF/f3dxlBGhxQQ7fuexIKkYU9fjzNTS3sXDZVt74YBtN\ne/xdti+3O6nc2ciFJ3XfMVRdt4enXl/Hsk/d/uGc7CyONGUcOnUUY0e5v99UjuTy0r/FDK0TlbgX\nBSLGhR9r6/T4WKBqAHNJBqhramXBmxtoavEzZexQDhg7lCkVJX06YsZxHN74YBtPLlxLWyDEjto9\nfOWM6R2eE3Icnlm0nqY9fg6eMoKZk0cwpLBvR+2EHIc/vbaWhcu2Ul5ayDEzRzN9wnAmVwzlgadX\n8L6t5m9Lt3Dm0V0/J4HbafLS25/x1soqpo0fzsmHjWXquGHdtv3Ne/08/rdPeX9NdXvHR35uNi+/\ns5m1W+q55osHM3LYvgsU1XV7WL21nk831bC9poXtNS34gNIS96JB6ZB8Sks6/tfdxaaQ44BDhxrp\nOA6btjeydM1OPtlYQ3Z2FiVFuRQX5BIIhli1sYaW1kC37192lo/x5UPIzvKx+rNaHnj6o5gdQx+t\n38UfX1vLzto9lBTlMqF8CKs21bJqUy0Fedm0+oNEPkaPKytmYnkJZcMLKC8t5NX3tvDmikomlA/h\ntCPGd5ulr+1vBVYj3ofqmlp57BXLh+t2MWxIHnOPm8yJnxtLbk4WO2tbePL1dXywdlfMfd9c4b69\nE8qHcOZREzhuVkWHf6D+QJAnXl/HWx9V4Q+EupwglpcWctzBFXy2o5EP1u7igWc+YnxZMaccPo6j\nDhrNkMJcQo7DgsUbeOmdzxhSmMtlZ7mjAmobW6ltaqUgN5tJFSVMrihh2BDvXe3Lz8tmypihbKhs\nYE9roMMJqOM4/Oq5leyu38tpR4znwpMO8NRUg96aPmE4iz6s5NMtdV06hVZvriUvN4uvnmVi9tRL\nR4ccOJInXl/Hxxt2d+gUchyHD9ZWk5+bzbf/dXbKrvx4mOrEIBVyHNZvq2dIYS5jRhYntM8yW81v\nX1pNS2uAYw4ezReOmcSW6iY2Vjby7ifb+fM/NnLswRWMGNr9KJXPtjfy0ydXdJlKNHtGOZedOZ2h\nRd1f/Kiu28M9f/qAXfV7GTWsgHOOmcTxsyrYvLOJXzz7MU+9sY5N2xv4t3MOIj+v67/VZbaal97Z\nxLTxwzn0wJFMmzC8vX1s3utnZ+0e9+Q0gRrX0NzGO6u28174yihAXm4Wpx0xnjOPmkBZp4slB44b\nxnd/8y7/t2g9M6eMYNyoru95dd0eHn1xNXZLHQDPv7WJfzpxCkfOKO9wEeezqgZ+/Nj7bNnZRPnw\nQv75tGlkZflobGlja3UTf31vCz/7vxV857IjO3zocBz3YtFnOxqp3NXMtupmNm1vJBAMUVyQw9zj\nJnPcrAoK83PIyc6iZa+fe5/8kBff/gzHgS+d3LFjqLUtyIvvfMYr724mEAwxfeJwjj+4giNMOUUF\ng7/GSt+w1m4yxgw1xkwGtgLnAV/BHVF6B/CQMWY2UGmt9UbvnXhKbWMrTy9az+Rxwzhx1uiERqU6\njsM/Pq7iyYXr2jsCPly373PWyKH5VIwoomJEMWNHFXHkjHJKOtWfbdVNPPH6OnAc/v2LB8esT40t\nbTz02/d4d9V2igtyGDYkj4XLtnLwlBEcNnVU+/OeWbSel9/dDMDfP6rCB0yqKGHmZPfC/9Txwzp0\nPgRDIfa0BgmGHEIhh2AoRHFBbrefX4KhEL97eQ1vfbydcaOK+eH1JxBs3dfRf+0/zeKO3y7lqTfW\nM3nMUKZPGN5h/y07m/jNi5+weUcT2Vk+lqzczpKV2xkzsog5h43jpEPHdqirO2pauP/pj9hR08KY\nkUUcfdBojpxRzoiSfH7/yhreW72T23/7HheedAA7avfw8YbdVO3uePEvUn83bY/9z94HTKwoYdaU\nERw8eQQjhhWw5rNaVm6sYfWmGva0BsOzWgooLcln/bZ6doVH8edkZ7UPkogYOTSfYw+u4LBpo5gw\negh79gZo3OOnaY+fCWOGMTQ/i9ycbALBEL96bqX7mTncMZTl87Hs050sWr6NT7fWk+XzcfqR4/mn\nE6ZQVJBL1e5m3vp4O0vX7GBcUTGzTRmzp5cxurTjxRkzoZS7fr+UP722ljEji5g5eUT7tpDjxByw\n0Rd8TpzRHh0CGnM7sCvGXWVWAV/AbcTfJqoRt9aeEW7EH7DWnhDv+NXVjb2ezO2lKzzRyspK2LGj\nAbuljk1VDWRl+cjJziIn28eQwjzGlxVTNrwQnw/e/WQHj//tU5r3Bpg0uoSqmmba/CFGDs3n4Ckj\nWbKyikDQYdr4YXzxhCmUDsmnMD+H/NxsdtS2sPozd+2eT7fUEwiGwvPSDSVFeeys28Mvn/2YzTua\nKC8tZFhxHlk+H1lZPkYOK+CEQ8Ywbfy+Xt6t1U289M5nvPfJTkKOQ062j8OmjiLkwPJPqxldWsh/\nXnJolz/i/X2vBuJ3+Mzi9bz49mfcfMmhHHLAyPbHt+5s4ruPvsfRB5Vz7fmzBjxXshLNVdvYyjd/\n8RaHHjiSmy7eN3WprqmVb8x/i1kHjOAblxw2oJkGWl/lchyHbz34Ns17A/zs6ye0F6qq3c1855F3\nOWJ6GddfeEi/5CorKxkUc/pUJ5LjxUwATk42Lyxaxz8+rmJX/V58wPGfG8OXTjogbof/35Zu4U8L\n15KXk8VXzpzOCYeM6dBB8PePKvntS2s4bfZ4vnLm9JjHqG1s5a7fL6W+qY1LT5tGdpaP7btb2FBV\nz8aqRkYOzeeGCz/HpIquoyK217Rwz58+oLaxlS8eP5m5x0/usMZXXVMrv3xuJeu21jOpooSbLzm0\nwwn86s9que+pDwkE9/2ZFeZnM7q0iF31e9tHyWRn+Tj24ArOOWYiY0YWd/k9bt3ZxKtLt/DOJ9sJ\nBN0TuIMmDefog0Yz25TFHZG5/NNq5i/4mImjh3DrZUe2tzOO4/DmikqeeH0drW1BDp82ipKiPP7x\nURUhx2FcWTHlwwtpbPHT0NLG7vq9BEMOJx82lktPndrlA1KkFpoJw/nmPx9GTnYWu+v38vu/rmHl\nhpr252X5fIwdVcTJh43jhEPGxOxIq21s5e4/LmdH7R7OOWYix88aw4bKBjZUNbBi3S5qG1spLcnn\n4jkHct7JU9m1q6nLMVLNq/8W06lOGGOOAO4FJgN+3LuMPQ9stNY+a4w5Cfhx+OnPWGt/Et7vR8BJ\nQAi43lq7It7rqE4kr765jbc+ruLgySNitq3J5Fq1qYYHn1vJlDFDOeXwcXxu6sj2dnhPawC7pQ4f\ncGhUZ0hPahtb2VjV0H5R9/MzR3f43LJ0zU4ee2UNzXvdjp1hxXlceNIBHH/ImJgjKR3HYVt1M0++\nsY5VG2vIz8vmkjkHcti0MjZWNbS/VtXuZuqa2tr3y8vN4qRDx3L20RMpKsjh+bc28belW9o7FcpL\nC7k56vOR4zgsXbOTPy1cS31TGzMmDudr582keW+Au37/PgV52dxx5dGUluSz+MNt/P4Vy+gRRVxx\ntmHt1npWbaxh3bb69uPnZPuYUF5CWyBIQ3MbTS3+Lhf5fcC4siFMHT+MqeOGUpSfS6s/2D5d6cN1\nu5gypoSbLzmMKRNHdPkd2s213POnDykpyuWblx5GWyBEbeNeNlQ18Op77s96wiFjuPS0qWze3sji\nFZUs/7SaQNBhaFEuZ39+EqccPo4NlfX88rmVNO8NcM7nJ/Klkw/sMmLnzRWV/PG1tfjDM1XycrOY\nOWkERx5cwdD8bEaPKGLk0AJ8PmjeG3AHITS2UtfUSk3DXuqaWtles4f1Ue9RtBFD3ZFENQ2t1DW2\n4uAOEDh86iiOmlHOrANGkJOdxd62IM17/ARDDuWlhd2Oeur8Nx/dMTSxfAi1Ta00trjnCbMOGMEl\nc6YyvnxIzGP1ZO3WOu7+4wcU5GVz08WHUrWrmQ/X7WLVphqOmVnBFefsW3uur+pEj51CasST5zgO\n6ysb+HhTLW8u30p9c1u3z83NyaK0JJ+dtXvIy83iklOmMufwcTS2+Hn5nc94ffk2AsEQI4bmc8kp\nUzlqRnncYe+76vfw67+s5tMtdQwtzuP0I8bz8rub2dMa4KRDx/D1Lx9BQ11iUzBqG1t5Z9V23lq5\nncpdzQBMHT+MGy88pEtP+f4aqN/hqo013Pvkh5xzzEQunjO1/fG/LNnEgjc3cPXcmRxz8L7BC177\n24pIJtctD75N4x4/P7/pxPYG+e1V23nkhU+4+JQDOefzkwY800Dqy1z/+6rljeXb+Na/HI6ZWArA\nS+98xtOL1nPVFw7i+EPG9Esuney70qlOgPcyOY7DU2+s49WlW3Ac9wTtSFPOZzsa2VbdTH5eNucd\nO4kzj5rQZUTc7vq9fOeRdyjIy+a/vnw448q6nggFgiG+/fA71DW18eNrj6W00/oxe9sC/Ojx5Wze\n0cQlp0zl7KhpviHHYdFHVTz+8hpycrK4/GzDcbP2/XvbtquZn/zpA+qb27rs2znDY69Y/vFxFWNG\nFvHNSw9jxNACNu9o5EePLycQDHHjlz6Hzwcr1u1u79QYNbzw/7N33+Ftl/fC/99almTJQ7a895Qd\nxyM7ISFkEyCEFSiQFihQWiilPW3PeXrO6ekZv/a0z3NON6W0UKClQEtCIIwkkL134hlb3nvbkrdl\nS/r+/pAt7NiJ7WDHjnO/rsvXlVjrli3r+9Xn/gyCDFqMPhoulVvcKe0ZCUbiIw1Y2nro63fQ0Nrj\nzuIJMmhZuyCcxXOCrprddLnXduVzLLuOdQvDiQ3xprDKSkGllfrWbrRqJVvXJ7AsxZUN3Gjp5sPj\n5ZzMq0eSXEEcL08VAX6e3Lk4koyE0T94OQeyY8+bm1iRGkJsqDfvHiymt89BykDfgzCjnmA/z3E1\nu7R02Ph/71ykoXX4+YWHUs6GxRHcuTQKjYdyxr3mB82Gdc3048T1Io4T49fW1cee0xUcvFBDlA4f\ncAAAIABJREFUn92JUiHnq3cmsSwleMzbjrauivoOfvb2BfqGlMYYvNSkxxupbOigvK7D3XpirPfp\nggoLFwqbyC5tGdajdFBYgI5VGWGU1rZzMq8eD6WcB1fH45TJeO+Aq0QrPEBPaqwf/j4a/L01KBQy\nckpaySxuosnqyhaZG+vH47cnDSthGqrHZqfB0o250spnZ6uwdNhQyGXoNErau/sx+mh4dF0ipXVt\nfHzCVUnx7S1paDwUvLW3kIJKK0qFjK0bk7k15fNy333nqnh7XxHJUQbuWBrJr97NxlOj5F8fWzBs\n0723z05RdRv5FRbyKyxUNnSg9VDirfPAW+eBTuPK2lTIZchkMlrbeymra3e3BLmcKcKXF7akoVVf\n+f14z+lK3j1YPOL7Bi81j280kRY3/LjS0e0qMd57rooemwMvTxVdPXZkMnh8YxIr0q58blzd1Elm\nUTMxId4kRvigUiom/Jrv7bNTWGUlt6wVS4cNU4QvKTF+BPt5uj8z2x1OrB02fPQe15zRP9q67A4n\nL72fS2ZxMzqNkhVpIazKCCNoHKXZYzmaVcvruwuGfS/Yz5P7VsayKOnz/ojXLSh0vcyWN/GCCgs7\njpZSXO1qIKnXqlhoCiAlxh+5DPodrsaMbZ19VDd1UdPcSX1LN3FhPjy+0UTgZdk3lg4b5XXtzInx\nu2Ijq8s5nRKfnqlkx5FSHE4JlVLOVzaYWJEWcs19jsrrO6ho6GD53OApKY+5Xr/Dwb5CUcFe/PCx\nhe7v/+TNc5TVdvCrF1YMS6WfSa+toSayrtc+yedYTh3/8dVF7qlrgx86/v2JRePeFZrMNV1Pk7mu\nrOJmfr09mzuXRrFlVRwA//3meUpq2/jVt1ZMKFgqTvYnbrYcJwZdjzV1dPex+3QldrvTXd57pYkZ\nn5ws573DpYQadWxYFMGipEC0aiUOp5OjWXW8f7SUju5+UmP9+faWtGH38dL7OZwzN40ZHD2SVcsb\nuwtYtzCcR9d9ni3kdEq8uCOHzOJmbssI5bHbRzZYDAjwYv+pMv7w4SV6bHa8dR5o1Uo81QoaLT10\n9drZuj5xzBp8SZLYdrCEPWcq8fNW88TGJP70ST7tXX18/Z4UFicHDbv+5enaTkniYmEzu06VU1Y3\n8veXFOnLhsWRpMWN3pdvLD02O//+2hl3iju4ejCkxvrx8NqEUUvvBrOYPDVK5DLZuF5btn4HP3vr\nAhUDaflatZKH18aPyPAaL0uHjW0Hi5HJZMQO9OWICNQPK0+eiX+HMDvWJY4TLuI4MTZJkvjweDm7\nT1XQZ3di8FKzPDWY/eer6bE5uGNpJA+sjLtiw/v61m4UHkqMOpX7vaLZ2sNP3jxPe1cfz947l2A/\nTw5erOFEXj22PgcKuYyYUG9MEb6cyK3H0mHjkbUJw/qnNVi6+fhEORcKm+kZKOfSa1XEh/m4e/3I\ngMOZruyUweyQmBAvvnZ3CsF+ngQEeFFY2syOwyUcz60fdf0aDwVzY/1ZnBTIAlPAuN/v7A4np/Ia\n2H26giZrDxuXRHHXsij357NDmTW8+akZpUKOwyHhlCTS4/x5eF0CcxODhv0eJUniN9uzySppGaje\ngO8/PG9EydblxlM+ZHc43dMd7Q4JtYcCtUqOTqNiTrSfO9B/pdeWJEm8f7SURksPvkN6+MyN8b9q\n2W9nTz97z1ax91wVSoWc5+9PHfP5jGYm/h3ClddldzgpqWkjJsR71H7AX8RHJ8opqLCQGutPRoJx\n1D6Ak3WcEAXdk6S42tXQMb/CAkBGvJF7V8cT6qsZs1+LJElXfENy/SEGTGgtcrmMO5ZGkRLjx6GL\nNayeH07ENaavAchkMmJCvIkJ8R77yjPcaH2F2rv6KK1pJyHCd9Ibus0EpkhfjuXUYa6yEhnkhSRJ\n5Jdb0GmUX+h1cTNKijKgVMjJLmlhy6o42rr6KKlpIyHCd9Kz5wQBXAGCE7n1nC1oJCbEi7tviRlx\nUtZs7aGysZP4MB/3sAFJkjiRW8/fDxSPaA6sVilYMz+Mzctj3OVAp/Lqee9wKQYvNf/93HKcfZ83\nWlTI5ayaF8bi5CBe3plLTmkL2w+X8NBqV7ZlXnkr58xNxIf5sGzu1XeYb5kbzEfHyzmcWcudS6Pw\n1avp6u3nTx/nk1ncTEqMH1vXJ17xmJgWZ+RHjy/krX2FNFl66LHZaW3vRS6T8fhGE7dlhI35M5XJ\nZDy0Jh69p4rth0r4xbuuBLVH1iWMCAgBI07A5TIZC0wBzE80UtnQiVanpqfLhtpDgadaOebAh7Fo\n1Uq+eV8qn56tJDLQC1Okq1G94irjL6/l2KVWKXjhgTR+vS2LAIOWR9cljsjemgiDl5pnNqdc8+0F\nQZh6kiTx1t5CDlyoweCl5qFlUe7+pctSgvnN9mx2n6qkpqmLr29OGdGfptHSzY//fI5um51Qo461\n88NIizPyi3ezaOvq45F1CSwcyGL4yu0mtqyKo6api/BAnbuMdUVqCD97+wLv7C9CLpexKCmQj46X\ncyizBodTws/bFaRakBhAfLjPiPe+OdF+tHXaOJFXj4dSwW0ZocM+axm81Dy1aQ5bVsfTaOmmpa2X\nlvZeemwOkqMMmCJ9r6mXplIhZ0VaCLekuo5zlx8bVmWEYdCreXlnHgYfNQ+vSxjWM2gomUzGk3cl\n86PXztDW2cdTd84ZVwBlPBsNSoX8C31uk8lk3L8ybsK302tV3LcyljuWRiJJzIrerOOhVMjd1QOT\n7e5born7lugpue/L3Ry/rSkiSRLZJS3sPl1J4UC6+NwYP+5bGUtMiPe4I3eTPfZvUGSQF49tTBr7\nijcZU6QvxTVtFNe0kRrrT3ZJCxKQHu8/5m1vRIMHmcJKK+sXRtA0cHBckBgwKWOPbyZqlYKkSF93\nimpuqeu1c6WDviAM5XA6r/rBfqgmaw97z1ZxLKeO3j5Xw+XCKisn8xr40up4lqYEUdPUxa7Tn/d/\nG2xKmRrrT1G1q+RIrVLw8Jp4YsN8qKjvoLy+nUvlFnafruRMfgOPrEtEq1byp0/y0aqV/MND6fj7\naEc9dnlqlHzjnhR+/Jfz7DldSUSAnkXJgby9txCZDLauTxzzhFWpkHPXsij+8qmZPacrWTIniN9/\nkEtzWy/JUQaevWfumCfrQX6eI3qhXW1z5UruXBqFTqPkrb1FbFwSyfqFV574NRqZTEZUsNeU7GpG\nBXvxzN1TH2AxeKn5jycXT/njCIIw/SRJ4u29RRy4UEN4gI7vPzJvWGlriL+OHz6+kJd35pFd0sKv\ntmXxDw+lu4M5tj4HL+7IodtmJz3BSG5JC29+VggUArBx8cj3Ua1aSXy4z7DvBfl58k+PzOP/vn2R\nt/YWsv1QCbZ+B4G+WrasihtX9o6PXj1m+wMfnQc+Og8SJnmA09WOc+nxRn7x/HJUSvmYxzIvTw/+\n5csL3Me/2WI8Tb6FmUf81q7RxcImdhwtpabJ1Wtnbowfm26JvqY0OeH6Soo08MnJCgoqXel4WSWu\nSQOz9YO90UeDn7cac5V1IEvI1UQ0OXr2HICup9RYf3LLWskpbSFzYBrgvCv07RCEQYM9axaYAnjs\ndtNVy3DrWz/fifXVe7gnaB3PqeOTkxW88vElPjhW6u6JEBagY15CAMXVVoqq29xTOjLijWxdn+ju\nlRAf5joxt/U7+PhEOXtOV/LijhwUA8Hh5+9PJXyUXkBDeWpUfOuBVH78l/O8vruASxWt1LV0s3p+\n2LhLUZenhvDRiXIOXKhh//lqnE6Jzcuj2bw85poD1de6uXJbRhjLU0PEBEZBEG4oDa3d/H5nLnGh\nPmzdMHZAfjAgtP9C9agBoUE6jYrvPJjGKx9d4kx+I7/Zns23H0zHQynn9d35VDd1sXpeGN/98kKK\ny5o5nFnL0ew6kqMNbFk9/uySEH8d//jIPP7n7Qs4JdiyKm5Exs+NaiIZMgG+2hGTKAVhOoig0DU4\nV9DI7z/IRSaTsTQliI2LI929WoSZLz7MB4VchrnSSr/dSW5ZK4EG7ah1mrOBTCYjMcKXU3kN1LZ0\nu0scZ9OuxPWUFufPO/uLOG9uwlxpIcTfc1Iaygmz257TlfT2OTieU099SzfP359KQMDI40Z3r53f\nbM+m22bn4bUJrJkf5j5Jvnt5DMtSgnlnfxEXi5qJD/fhrqVRpMX5u4MiPTY7BRUWNB4KkqIMowZL\n1CoFD9wWx7KUYP76mRlzlZWnN80Z93tCiL+Or29O4dfbsjieU+9KGb81dtw/C5XSlS30188K8fZU\n8bXNKaQMGbl6vc2GDyGCINw8yuvb+eW7WXR091PZ0IlSIefhtfGjvt83WLrJKWnhQmETBZVWwq4S\nEBqkkMt5etMcHE6J8+YmfrM9m+QoA2fyG4kP8+GRdQmAK1tn84oYNq+IuabnEWbU8bNvLEMuk016\nLxZBECZGBIUmqKDCwh8/ysPDQ8E/PTJvVvTZudkM7SuUXdKMrc9Beppxysr4ZgLTQFDIXGmhoMKC\nr95j1gbBplqQnyeBvlpySlsAmJcwsZ5fws3H0mHjbEEjYUYdEUF6TuU18F9/Pse/P70Ub/XnJ8JO\np8QfPsyjvrWb2xdHsGHRyHImo6+Wbz2QRl+/Y9STaK1aybzE8b0mQ42undoem+OqzSNHkxbnz0Nr\n4tl2sIQvrYmfcE+bVRlh6LUqEiN88b3KmHtBEISb1eHMGs4UNJESbWBxciBGHy2Xylv57Y4c+voc\nPLQ6nmM5dew9V4W3TsVdy6IB6Lc7OHSxlgMXa4ZNBEwM9+G5+1PHNQ1RqZDz9c0p7pHb+RUWfHQe\nPHvv2CW+EyFKjQRhZhB/iRNQ2dDBb3dkI0muNHsRELpxDfYV2nGkFICMWdpPaNBgA7RDF2to7+5n\nWUrQrA6CTbXUOH/2n68GuOLIZ0EYdPBiNQ6nxLqF4axMDyXMqGPH4VL+6bdHWWAKYGlKMHOiDew4\nXEpOaQtzY/x4cFX8Ve9zsnZVZTLZhANCg25fHMnK9NBraiYpl8tGbeosCIIguAJCf95jBiC/vJXt\nh0qIDfWmssFVHvzsvXNZmBTI4uRA/vuv53nvcCmeGhUyXBOLLB02PFRy5iUYSYvzJzXWf9TJhVej\nVMh59t65/P6DXPLKW3nuvrlfqBm9IAgzlwgKjVOztYdfbsuix+bg69Oc6i58cYN9hepautGqFSTM\n8l5QQQYt3joPqgd6YCVHidfvF5Ea6woKeXuqiA0VwWHhygZ3bHUaJUtTgpHJZNy1LJowo553DxVz\nMq+Bk3kN6LUqOnv6CTJo+fo9KTdME/ibZbqIIAjC9XIit46/7DGj16r40VNLuVTSxOlLDRRUWFB7\nKPjWA2nucl8/bw3f+1IGP/3rBd781BVE8lDKuWNJJBuXRH7hyaiDo8Vt/Q6R1SMIs5j46x6HggoL\nL3+YR3tXH4+sTWDJHLG7eaMb7CvkcErMjfGf9T0lZDIZpghfzhY0AqKf0BeVFOlLRKDeNcFNZFwJ\nV3Eqr4HOnn7uXBqFekh2T0aCkXXLojmdVcPJvHrO5Dei0yh5YUsaOs3Ex4sLgiAIN74z+Q386ZN8\nPDVKvv9wBskxfhj1Klamh9LWaQNcvXyGCvHX8Z0H03l9Vz5JkQbuuiVqUstyZTKZCAgJwiwn/sKv\nwilJ7D5VwY4jpchlMh5Zm8D6UXo8CDeewb5CxTVts3bq2OUSB4JCgQatexqRcG08VAr+U4xxFi7T\n1+/A7pDc5ViSJLH3XDVymYw188NGXF8mkxEX5kPcQONOu0MaFjgSBEEQZi+7w8n+89VUN3bS2mHD\n2mmjvrUbjYeC734pY8QQm8uDQUPFhnrz/z29ZKqXLAjCLCWCQlfQ3WvnlY/yyCppweCl5tl75hIf\n7jPdyxIm0W0ZodgdTtJneT+hQSkxfshkkB53cwTBBOF6Mlda+ONHl+js6efWtBA2Lo6kua2X6qZO\nFiYFjtnLQSGXM8sTFgVBEIQBPTY7v3s/h0vlFvf3dBol0cFePLIuUfQtFQThuhJBoSt4Z38hWSUt\npEQb+NrmlHF16hduLMtTQ1ieGjLdy7hugv08+a+nlmAUWUKCMGkcTicfHS/noxPlyJDho/fgwIUa\nDl2sxUfvOm6sXxg+zasUBEEQZgpLh41fvptFdVMnGfFGvrQ2HoNeLcayC4IwbURQaBRlde0cz6kn\nPEDPdx5KRyEX27fC7BBm1E33EgRh1rB02Hh5Zy5F1W34e6t5ZnMKMSHenM1vZNepCmqau4gJ8SI+\nTGSZCoIgCK5Jxr95L5vWdhur54WxdX3iDTNYQBCE2UsEhS4jSRLv7C8C4JF1CSIgJAiCIIzgdEq8\n9H4OJbXtLEwK5ImNJjwHGkQvmxvMkpQgiqvbCPDVIhPNyAVBEG5K/XYneeWt5JW5vupbuwHYsiqO\nO5ZEiuODIAgzgggKXeZMfiPF1W3MTwwQE5oEQRCEUX12toqS2nYWJwfy9c0pI07s5TIZiRG+07Q6\nQRAEYbrZHU5+/vdMCqusgGvISUa8kVvTQ5iXEDDNqxMEQficCAoNYet3sO1QMUqFjIfWxE/3cgRB\nEIQZqL61m/ePluLlqWLr+kSx0ysIgiCM8Lf9RRRWWUmL8+eOJZHEhfmgFBMFBEGYgURQaIhPT1fS\n2m7jjqWRBPpqp3s5giAIwgzjdEq8tiuffruTr22ag5cYQiAIgiBc5nhOHQcu1BAeoOPZe+ai9hBN\npAVBmLlEuHqApcPGrtMVeOs82LQserqXIwiCIMxA+85XU1zdxqKkQBYmBU73cgRBEIQZpry+nT/v\nMeOpVvL8/akiICQIwow3rkwhk8n0S2ApIAHfNpvNZ4dcdg/wQ8AG/M1sNr9oMpnkwMvAXKAP+IbZ\nbC6Y7MVPpp3Hyujrd/Louli0apFAJQiCIAzX0NrNjsMl6LUqtm5InO7lCIIgCDNMe3cfv9uRg8Ph\n5Jn7Uwk0eE73kgRBEMY0ZqaQyWS6DUgwm83LgKeA3wy5TA68CNwJrATuNplM4cA9gI/ZbL5l4Db/\nOwVrnzR1LV0cy64jxN+T5anB070cQRAEYYaxO5y8/GEefXYnX96QiLcoGxMEQRCGcDol/vhhHi3t\nNu69NYa0OP/pXpIgCMK4jKd8bC3wAYDZbM4HDCaTyXvgMiNgNZvNTWaz2QnsB9YBCcCZgduUAFEm\nk2nG5k6+f6QUpyRx/8o4MYJeEARBGGHHkVIq6jtYnhrM4uSg6V6OIAiCMMN8cKyUS+UWMuKN3HVL\n9HQvRxAEYdzGUycVDJwf8v+mge+1D/zby2QyJQDlwGrgEJAN/IPJZPoVEA/E4gogNUzWwidLWV07\n58xNxIR4Mz/RON3LEQRBuOHM9hLj3LIW9pyuJMigZet6UTYmCIIgDJdZ3MzHJyoI8NXw9KZk5GIq\npSAIN5BraZ7jfpczm82SyWR6HHgNaAPKAJnZbN5tMpmWA0dwBYjyh95uNAaDJ0rltScTBQR4XdPt\nfvNeDgBP3zuXwEDvMa49cde6rqk0E9cEYl0TNRPXNRPXBGJdU2loibHJZErGdTxYNnDZYInxfKAF\n2G0ymT4AFjFQYmwymeKAXwObpuUJjKG9q49XP85HIZfx9XtS0HiInnOCIAgTNZs3DxqtPbz60SVU\nSjnfvC8VT41qupckCIIwIeM5u63FlRk0KBSoG/yP2Ww+DNwKYDKZfoorYwiz2fzDweuYTKYSoPFq\nD2KxdI93zSMEBHjR1NQx4dvllbeSWdTE3Bg/Qnw013QfU7GuqTQT1wRiXRM1E9c1E9cEs2NdMzx4\nNKzE2GQyGUwmk7fZbG5nSIkxgMlkGiwxDmRIibHJZIoymUwKs9nsmJ6nMDpJco2fb+/q46HV8UQH\nT/7GgSAIwmw3mzcP+u0OXno/h26bnSfvTCYyaEYfrwVBEEY1nqDQZ8B/An8wmUzzgVqz2ez+JGMy\nmXYDjwNdwN3Az00mUzquXYAnTSbTRuDCQM+hGUOSJN47VALAA7fFTfNqBEEQbljXrcT4emeU5hQ3\nk13SQkZCAFvvnINcPvnlADM14CfWNTEzcV0zcU0g1jVRM3VdEzRrNw+OZNVR2dDJrWkhrEgLme7l\nCIIgXJMxg0Jms/mEyWQ6bzKZTgBO4Jsmk+kJoM1sNr8PvIIrcCQBPzWbzc0mk6kVkJtMpjNAL7B1\nyp7BNapq7KS8voN5CUaigmfFAVcQBGEmmJISY7j+GaUfHikGYOPiCFpaOq/5sSdzTdeDWNfEzMR1\nzcQ1gVjXRM2ijNJZuXngcErsv1CNSinnmfvT8fVSX/PjTtaariexromZieuaiWsCsa6Jmox1jas5\ngtls/sFl38oactkOYMdl13cCT3zRxU2lC4VNACyZI6bICIIgfAHXpcT4euvu7ee8uYkgP08Swn2m\nezmCIAizyazYPDhX0Eh9Sze3ZYTS39tHU2/fNT/2ZK3pehHrmpiZuK6ZuCYQ65qoydo8uGk7Zl4o\nbEKpkJEa6z/dSxFG4XA6qOyopthaRm1XPUmGBBYGZaCQj777I0kSXf3dtNos2OyfH5RlMhlBngF4\neeiv19IF4WYzK0uMT11qoN/u5Na0EGRiiowgCMIXMes2DyRJYs+ZSgA2LIqY5tUIgiB8MTdlUKjR\n0k11Uxdpcf5o1Tflj2BG6rH3ktmUy/mGTIqtZfQ7+92Xnam/wMdln7E+8jaWhCyksbuZEmsZxW1l\n1HbWY+m10Dfk+pcL9DQS7xODyS+BBYHp4kOeIEyS2VpifCSrFrlMxvK5wWNfWRAEQbiaWbd5UFTd\nRmltOxnxRkL8ddO9HEEQhC/kpoyIXChsBmBBYsA0r0SwO+1cajGTXZTLuZos+p12AEJ1wcT5xhDv\nE02gZwCn6s9xovYMfy/8gHcLdyIhue/DU6klyDMAP40Bg8YXrVLjvswhOanqqKGsrYITdWc5UXeW\nHnsPt4Ytu+7PdTI5nA7yWsycrb9AZUc14fpQ18/LNwZPpZaStnJKrGWUtVWQ7G/inrg7pnvJwiw2\n20qMK+o7qGzoZF6CER/91PSIEARBuFnMxs2DTweyhDYuiZzmlQiCIHxxN2VQ6HxhIzIZpCcYp3sp\nNyVJkihpK+dsw0UuNmTTZXfVfwd5BrAoaD6LgjMwaoeX9UV6h7Mxei0Hq45R0FpIqD6EeJ8Y4nxj\nCND6j5n543A6qOio5pcXfs/xmtMzOijU77TT3f95TbyEREdfF629Fiy9Vuq6G8hpzqPN5tpkU8lV\nNHQ3cb4xa9T7q+tqYEPUKrRK7XVZvyDc6I5m1wJwa1roNK9EEARhdphNmwd1LV1kFjUTG+otes4J\ngjAr3HRBIWunjZKadpIiffH29Jju5dxUHE4HFxuz+bTiILVd9QB4e3ixOmIFG5JW4GU3XDW44+3h\nxT1xd1xT1otCriDWJ4q5/slkN+dR1VFDhFfYNT+XayVJEu19nbT3tSNJn2c7dfR3UmItp9haRkVH\nFfaBjKkr8fLQsTLsFhYFzyPaO4LmnhaKra7soB57DzE+UcT7xnCpxcyu8n1kN11iSciCqX56M15X\nfzdOySl6TAlX1Nfv4GReAz56D1Lj/KZ7OYIgCMIMs/dsFRKwcXGkaEcgCMKscNMFhS4OTB2bJ0rH\nrhun5ORk7Vk+qzxEc08LMmQsCExnWegiTIZ45DI5AX7Xp6P7spCFZDfncbLu3JQEhRxOB/mthZxr\nyMJqs7q/75QkOvo6aLVZrxrwkSEjTB9CkOfw16feQ4dB7YufxoCfxsD82CQsLZ9nEwV6BhDoGcAt\noYuG3c5T5cmu8n1caMyaFUGhxu4mdpXtJ1gXwNqIlagUqnHdrs/Rz96Kg3xWeQi7006g1kicbwxx\nPtHoPT7vBSBDho/aGz+NAU+lVpzs3YTOFzbRY7OzZn4UCrl8upcjCIIgzCC2fgfHc+sx+miYLz5L\nCIIwS9x0QaHBUfSin9D14ZScvJW/nVP151DKFKwIW8r6yNtGlIddLyn+SXh56Dlbf4H74u4cd1Bh\nKLvTTlVHDbWd9TiH9Daq7aznQmMWnf1do97OS6UnVBeMn8aAj9obhezzD5xqhZoYnyhifSLHVeal\nvMIUtssFeQYQrg8lv7WI7v5uPFWe47rdTGNz9PFp+QH2Vx7GLjkAOFl7li2Jm0k1zsEpOSmxlnG2\n4SJVHTXuHktxPjGU15Ty2rm/09JrwcfDizB9KKVtFZysO8vJurNXfEwPhQfxPjE8NXcrmiF9qoTZ\n7WiWq3RsRVrINK9EmO3a+zrQKrWo5DfdqZgg3LDK69rptztZYApALhcbR4IgzA431ZlIV28/BZVW\nooO98PO+fh/yJElCQkIum/5d56ElS19EXVcDtZ11zL/KJC9JkthR9DGn6s8R5RXBM2mP4aue3tpr\nhVzBkuAF7Ks8THbzJRYEpV/xumfqL9BZ3053tw1w9fqpaK+ivL1q2GS0obxUelaFL2dR8DyivCJm\nRKbJgsB0dpbuJqv5EstCFl73x+/s6yKrKZdehw2Dxhc/jS++ah967b209Fqx9Fpo7bXS2mvFYnP9\n22a34avxwaD2xaDxJaf5ElZbGwa1L/fG3UFlRw0Hq4/xcvYbxPvG0NJjwTKQmSVDRmVHDSeGBHzk\nMjnrI1exMXoNGqUGp+SkprOe8vYKbI4+9/WcTifWvjZaey3UdzVyqdXMkeqTbIhefd1/bsL112Oz\nY66yEhfqTZDhxgygCjeGYmsZv818hUCtkW/P+/qwjEVBEGau4po2AOLDRC8hQRBmj5sqKJRV3IzD\nKbHANPVZQrWd9Zy3niezOp8Saxl9Tjs/WPRtjNrJ7VHR7+gHmWxcO417yg9w4NgRvjf/myPKkyRJ\n4nD1CcK9Qon3jbnq/fTabbyY+SpWWxvF1nIeTNw8asBrV/k+DlYfI0QXxHMZT6JXzYyT3mUhC9lX\neZhTdeeuGBQqsZbz50t/G/F9GTJC9cHE+UQT6RU+7Ofu5eFFvG8MinFm8Vwv8wLT2FlRws23AAAg\nAElEQVS6mwuNWVcNCtmddpp7WmFI9pNepZ/whxVJkujo76TIUsKZ+otcajXjlMY3RXawfEvvoae5\np4WazjoAlDIFG6PWsCF6DWqFBwuD57EsdBHvFu6k0FKMRqFhachCFgXNI943hrquBlePprYyPDUe\nrAlZRbAu0P04cpmcCK9QIryu3Ei4u7+HH538KfurjrAy/BY0ymufQnWpxUxLb+uMbnAuQElNG5IE\nSVGG6V6KMIs1dDXyx+w/Y3faqe2q57eZr/Dtec/csJmcgnAzKa52BYXiRFBIEIRZ5KYKCg2Oop/q\nGuCc5ku8nP2G+/9apZYeew+flh9ga/KWSXucms46Xsx8FS8PPd9f8DweVymF6rH3sLfiEL2OXrYV\n7uSb6U8Ny2I515DJtqKdxPnE8N0Fz171cfeU78dqa8ND4cGRmhP0Ofp4NOkBdzDEKTk5UHWUXWV7\n8df48XzG0zMmIAQQrAsixjuK/NZCLL1WDBrfYZdLksTOkt0AfHPx46j6XeVcMpmMUF3QDXfiHuDp\nT6RXGAWtRXT1d6Mbsn5Jkihtq+BMw4Vhk+CGCvIMIM7HNe4+zjcGf83whuCWXivnGjIpySulob15\nRN+kCH0oC4PnYdT40WqzYum1YrG1oVVoBnok+Q58GfBV+7hfR5Ik0WPvoaXXireHHh+197B1heiC\neCHjazT2NGNQ+w57/Ud4hRHhFcaqiOUEBFxbvypPlZbV4SvYVb6PozUnWR+1asL3Aa73gz/m/AVv\nDy9WhC6dEdljwujMVa5ss8QI3zGuKQjXpqOvk5eyXqPL3s3WpAcpb6/keO1pXsz8E9+a97SYEikI\nM5hTkiiuacPoo8FXf+0bRYIgCNfC4XQgk8mmpPropgkKSZJEYZUVo4+GEP+pDVCca8gE4LGMLcRo\nYjFq/fjx6Z9zqv4cG6PX4D8J2ULl7ZX8LvNPdNt7aO/r4IOSXTyUeM8Vr3+0+hS9jl60Sg35rYVk\nNeWSEZgKuE5StxXtBKChu/Gqj1vf1ciBqqMY1L58b8FzvJL7Jqfqz2Fz9rExag3nG7M4W38Ri82K\nj4cXL8z72rSXjI1mWehCytorOF1/no3Ra4dddqm1kJK2Mub6J3NbzNLr0gB7qs0PTKeyo4asplxu\nCV0MQH5rIe8U7KCltxUALw89S4MXolQMvC1IEi29FkrbyjlRd4YTdWcA8FX7EOcTTZg+hILWIoqs\npUgD2UWf903yJUQXzIKgdEJ0Qde0ZplMhqfK86pBOJlMNiLrbTKtjljBgapj7Ks8zMrwW1ArJjax\nsNBSwqu5f0UhU/BkylYREJrhiqqsyGSiLECYHJ39XZS1VeDt4YW/1g8PuQd/yH6D5t5WNkav5ZbQ\nRSwNWYDD6eBU/Tl+l/kaz2c8JXqYTYCrDNlCS0/rsFJgGRCiDyZEFzShk2dJkii2lmF32vHXGjBo\nDKjkShxOB1ZbGy29Frr6u/FV+2DU+qFX6cZ8X3dKTtps7e51tvS20tJjoaW3lY6+zmHXTQtIuaYJ\nq8L10dDaTVevndS46emLKQjCzcvhdPDj0z8n1iear8x5aNLv/6YJCjW19dLZ08+c6KktC7A77eQ2\nF+CnMXBX4hqam10H/I3Ra/nzpb/xacVBHk164As9RpGlhN9nv06fo59HTPdzsPo4h6uPM9c/iTn+\nphHX73f0c6D6KBqFhn9b9W3+bf//sr3oI5L9TagVHmwv+pCu/m5UchWd/V0jskkGSZLEtsKdOCQH\nDyZuxqDx5YWMr/H77Ne52JjNxcZsADQKNUuCF7Axes20NZQey/zAdLYXfsjJ2rNsiFrtPml0Sk4+\nKtmNDBmb4zZO8yonz7zAND4o2cWFxmyWhSzicPUJ3iv+CDkyFgfPZ3HQfBINcaOWvjmcDmo66yhu\nK6PEWkaxtYzzjVmcb8wCIM4nmkXB81iffAu97ZPTs2qm8FR5sipiOXvK93Os5hRrI1eOej27087p\n+vMY1L7E+kShUWqoaK/i5ezXkSSJr6c9Tpxv9PVdvDAh/XYHpXXtRAZ6oVXfNIfGSXW85jRnGi6w\nJWHzlEx3vJEUWUp5Pe8t2vo+31RQyhTYJQcLgzLYFLMBcJWybk3egkNycrbhAvurjnJXzPrpWvaM\n5nA6qOioptBSQqGlmOrOWrr6R2a3DqVX6UgwxBHtHUFXf/dAUMaCU3KyICidJcEL8PLQI0kSeS1m\nPin9jIqOKvftZcjQq3R02btHLYP2kKvw8tDjCkON5JScdPR1uAckDCVDhqdKi2zIbcd6PsL0Giwd\nExsHgjB+g/1sxcboF3Op1UxjTzPJo3zWnww3zZlvWW07ADEh3mNc84spspbS6+hlWcjCYS/+BYHp\n7C7bx6k6V7aQn+baglMFrUW8nP06TkniqblfZl5gKlHekfzPud/yZv67/Ovi747oAXOq/jwdfZ2s\nj1xFvH80ayNX8lnFQfaU7yfWJ4pzDZlEe0cS4x3JwepjNHQ3EesTNeKxLzblUGApYo6/iTRjCgAa\npYZvpj/FO+Yd9NptLAzKINU456qlbDOBVqlhQVAGJ+vO8ofsN3gi5VG0Sg2ZTblUddayMCiDMP3s\nmT5k1PoR5RWB2VLMm/nvcrr+PF4qPc+kPT7q73oohVxBpHc4kd7hrIm4FUmSaOxuorqzjmjvCHfm\nm5daTy83flbV5dZE3MqhqmPsrTzErWFL8RglW+jDkj3srzoCuE70w71Cae2x0Ofo58m5W0cN1goz\nS2ltO3aHJErHrlFmYw7vmHcgIfG/53/HlxLvdWcl3kwkSWJf5WE+LN0D4H7PbO5tpaWnlQivMB5J\nemDY+YFcJudh071cbMwipynvpgwKSZJEbks+5e1V7mwaS28bziHBlB57L31DhjwEaP2J9ArHX+uH\nUeOHdkiGlV1yUNFeRaGlZNimFbh+3q6BBNXsLNlNqnEOvc5uzC2lgGsTJVQX5M7msdraMGr98dca\nMGr80Kk8sQxkDbX0tA4Ecq68IRLmFYpR44e/1g8/jWHg359nIQk3DtFkWhAmpr6rkd9nvUZHfydG\nrT9GrT9BngGsibh1IKAujNfp+gsALAmePyX3f9Mcjcrqrk9QKLspD4C0gDnDvq+QK7g9eg1v5r/L\nZxWHeNh034Tv2yk5+bv5fZwDmQcp/kkARHiFcnfs7XxQsot3zO/x9NyvuE84nZKTfZWHUcqVrI5Y\nAbiyls7WX2R/5RFOqTxRyBRsTdpCaVs5wKhBoV67jfeKPkIpU/Bgwj3DTmg9FB48PufhCT+f6fZA\nwiastjZyWwr4+fnf8UzqY3xc+ilymZy7BnZxZ5P5QWlUdFRxuv48YfoQvpH2xDUFJ2UyGUG6QIKG\nNG6ezXQqT24LX86nFQc4VnuaNRG3Dru8oLWI/VVHCNQayQhMpdhaRmV7FXbJwdakLcwPTJumlQsT\nUTjF/YSckpMLDVnM8TdNeV+y7YUf0tjTzNfmfgXVdQjQl7VV8sald1ApVGyO3ciusr28VbCdkrZy\nvpR476iB1Nmo197LXy79nazmPHw8vHhy7pfHHNwwSKPUkGCIu2Kvu9nM5ujjnYL3ONtw0f09uUyO\nj4c3asXnfVv0HnpifKIwGeJJ8I0d1wcKSZJo6mmmurMOL5Ueo9YPH7U33fYeztVncqLuDFlNuQCk\nGVO4K2Y94VcZQCDc3Ipr2lB7KAgPEB9mBWEsll4rL2a+isVmJUQXRNOQATIOycH98ZumeYXXh7m1\nmOzmPJp7WmjqaaW110KoPphbQ5eyICh9XOdI3f095DRfIsgzkEiv8ClZ500VFJLJICrIa1zXL2gt\n4lTdeR5I2DTuSKZTcpLdfAlPpZY4n5EngouC5rG7bB8na89we9TqCZ/0mVuLaexpZknwAndAaNDa\nyJXktuST2ZTL+8WfsCF6NXqVjouNOTT3tLA8dIm7Ua9a4cGWhLt5JfdN2vs6uDNmPaH6YLr6uwBo\n7G4a8dgn6s5gtbWxMXotgZ7GCa17ptIqtTyb9lV2FH/Moerj/OT0L7BLDlaELpk1z3GohUEZ7Ks8\nTLxvLF9OevALTdO62ayJuJWD1cf4pHQvwZ6B7syf7v5u3sx/F7lMzhMpjxDlHQG4SjY7+7tuqg92\nN7rBoFBCxNTsAGc25fL6pXeY42/iubQnpyyNuru/m8M1J3BKTrYV7eTRpMkbbjCa5p4WXs5+HbvT\nwTfSvsJcYzJpxjm8mvsmp+rOUdfVwD/Mf/a6ZUTUdNaxq2wv1Z11PGK6nyS/hOvyuA6ng1dz/0p+\nayEJvrF8NWUrPurxnW8MmmtMJr+1kNyW/Bk7qfD94k8wW4rZmvTgVac3jldDVyOv5L5JXVcD0d6R\n3B17OwFa/2FDB74ImUxGoGcAgZf1ntOrdKyKWM6qiOXUdNZh9PNC3Sc+6AtX1tnTT11LN3OiDcjl\nogxmqLyWAlRyFQm+saJE6AZhtbVRVl1CtMfU/M46+7vcAaHNsRu5PXoNkiTR1tfOf576H3KaL015\nUOj1vLdxOB08nfqVKX2cqzlRe5a3C7a7e696KrUEaP2pbK/mr+3beK/4Y5aGLGBTzIar9hO82JiN\n3WlnSfD8KfsbuymCQg6nk4r6DsKMetQeY59kZDXl8lruW9glBz5qL+6Lv2tcj1PVUYPV1saS4AWj\nnsy4soXW8lbBNnaV7eORpPsn1ADxcM1xAG4Lv2XEZXKZnMeSH+bn519kf9URjtaeYkXoEsyWYmTI\nWBd527DrpwfMZWFQBm22dm6PWg1AoKcr86NhlKBQWVsFALeEzK5yAIVcwYOJ9xCiC+LvhR+gkiu5\nI2bddC9rSviqffjv5T+cko71s53eQ8dXkh/iz5f+xktZr/FAwt2sCl/O38zvY7W1sSlmgzsgBKBS\nqDAoREDoRuFwOimuaSfUqMPbc2qyWs4MpP1eajFztuEii6co/TenOR+n5EQhU3C89gzR3pFTVsbV\n3d/DS1mv0dnfxcOm+5hrTAbAX+vHd+c/x5v573K+MYuPSvdM6OSvtddCoaWENls7fhoD/lo//DV+\ndLS2cqoim0JLCWXtFfiqfUg0xGMyxGFQ+7Kv8jAXGrPdJ18vZr7KptgNw/rGTQVJkvib+X3yWwtJ\n8U/i66mPX1NAI9U/mW3sJKd58oJCbbYOiizF1Hc3DitwCtUFMz8wbUInl8XWMvZVHgbg5+df5KHE\n+7gldNGE1yRJEo09zeS1FPBx6afYHH3cFr6c++PvQjkN5VRh+hACfK5tSqVw8yitFaVjoym0lPBS\n1msABHkGcmvYUpYEL8BTdX0nKTolJ3ktBaRoY5Ez/mb9kiRxqbWQXWV7MWr9eCz5S5MSkJ6pJEni\nTP0FthXtpMfey3PpT5EyyS0ObI4+Xs56nfruRlZHrGDDwOdMmUyGr9qHZL9EsppyaehqnLKqgyJL\nqXvwU3NPy7T0uD1UfZxthTvRKT15IuURor0j3Jnirb0Wjtec5njdGQ5WHRuY5H3lTbzT9ReQIWNR\n8LwpW+9NERSqbe6mz+4kJmTsXbsz9Rd4M/9dlHIlOpma47WnuSN63YisCqfkRIZs2AnV56VjKVe8\n/yXB8/m0fD8n6s5Q2lbOhqjVLAzKGPMNqLmnldzmAqK8I4Z9+BzKX2vg35f9H47XnmZ/5REOVB0F\nXPXxl2e+yGQyvpry6LDveXvo0Sg0NHSNnEBW1VGDTumJ3yzNfFgRtpQYnygcTseMnJY2WURA6NrN\nD0zDoPbhDzl/ZnvRh2Q15VJkLSXGO9J9wBNuTJUNndj6HSSGT+xv/7OKgxyqOsY/LvzWVbPCuvq7\nudRixqjxo72/k+2FH5Lkl0AAE8skGY+LTTkAfCPtCV7Pe5u/F35AuD6USO8rpxs7Jec1vTccGuhB\ntybi1hFBDJVCxaNJW6jqqGF/5RHm+JlGZO302m0097S4erP0tlLX2UChtYTmnpYxH9uo8aO5p5W6\nLteghUER+lA2xd6OTuXJn3Lf4qPSTylrq+CxOQ+POkDhcpZeK+caMlkYlDHuTL+dBZ9xou4MEfpQ\nnkzZes0fKPy1foTqgjFbirE5+iY87XBQr72XNy7u4UJ1LvVXmSh6rPY0j5oeIMBz7JNlh9PBu4Uf\nALA5diP7Kg/zVsE2StvKeSjx3nH1EazvauRvJdvJriugrc9V0u+h8OCrKY+yMChjnM9OEKaH6Cc0\nkt1p5+/m95EhIz0ghdzmfLYXfcjOkt08NXcrqcY5Y9/JJMhvKeT9kk+o6azDx+zFd+Y9O66M/8qO\naj4o3oXZUgy4JjsrZAq+nPzgrDxf7ujr5J2C98hqznM/v4LWwkkPCr2Vv42y9koWBc3j/vhNIzYf\nUo1zyGrKJbv5EuunKCi0p3y/+9+ZTbkjkiMud7ruPJ+U7cVX7Y1R60+A1kiA1g+jp6sPkk45sbL/\nvRWH+KBkF14eel7IeIZQffCwy/00Bu6O28jGmHX89MyvOFl3jnWRq0Z93Tb3tFDSVkaib9w19yQe\nj5siKOTuJxR69X5Cx2pO8Tfz+2iUap5Lf4qC1kI+KdvLqfpzrApf7r5en6OPX5x/CblcwfPpT7mj\nflnNeajkSpL9Eq/4GAq5gm/Ne4ZdZXs523CRv+T/nY/LPiPeN8Y9gUIpV7AqfMWwF9CxmlNISNwW\nNjJLaCi1wsN9gn62/gI5zflsjr396j+gAYPjvas7a3E4He4T2x57D009LSQZEmZ1WuhsaiwtTI0Y\nnyj+z8IXeDn7DYqspQP9tB6Z1btKNwNz5cT7CVl6rewq20u/086BqqM8kHD3Fa97sTEbh+RgRdhS\nVHIV24p28m7hTv457NkvvPaheu295LcWEqoLZo6/iSdSHuH3Wa/zau6b/NOiF9CrdCNuc6L2DDtL\ndvPVlEcnVGrllJycqD2LWuFxxR5sGqWaJ1Ie4X/P/44389/lnxd/B71KR7+jnz0VB9hbcQjHZVOZ\nNAoNqcY5JBriCND609prdY/w9vPyJlITSaIhDh+1N/1OOxXtVZgtxTR0NTI/KJ10Y4r7OPWDRd/m\njUvvkNtSwK8v/oEfLPr2VU/0bY4+Xsp6jdquej4u+4wVoUvYELXmqmVg5xoyeTvvA3zVPnwj/atf\nuCx3rjGZzyoOYm4tuuoG09XsKt/H/sojeMhVJPslYjLEE+kV7n6fckoO9lceJbcln5+c+Tl3xWxg\nTcStV30fO1pziprOOpaFLOL26DUsCMrg1dw3OVl31j150V/rR4DWjzURKwm+7ETf7rTzx5w/09Dd\nhF6lY0FgOgmGOFKNybN6I0aYPYqr25ABsWN8lvginJKTQksJiYa4GyIocaDqKPXdjawMW8aXTPfR\n0dfJybqz7Crby9sF7/GjpTFolVOXMVTbWc97RR9RYClChgyTIR6zpZgXM1/lewuec7fNGM3BqmNs\nL/oQgDl+Ju6IWcf2og85XX8evYduRva7sTvtFFvLJvz6sNraOF5zmiM1J+ns7yLBN5aHTffzs3O/\ndgfEJkuPvZeLTTmE6oL5SvJDo65zrn8SMmTkNF9ifdSqYZftqzxMZUEl90XfPebGTL+jn9K2CuJ9\nY4Ydv8raKiiwFBHlFUFlRzVZYwSFiq1lvFWwHXBl8JQM9NgdSqvUoFfrcDpcUyhlMjn3xd1JRmDq\niOseqT7JByW78FX78MK8Zwi6rHx5KJVcyabYDfwp9698UvbZiIQN+DzTfHHIgivez2QYV1DIZDL9\nEliKa7zCt81m89khl90D/BCwAX8zm80vmkwmPfAXwACogf80m82fTvbix6t0YPJY7FWaTF8cmJyi\nV+l4PuNpIrzCCND682nFQQ5WHWNl2DL3C/uTsr1UddYC8FLW6zyf8TTtfe3UdTWQakwec3fPqPXj\nsTlf4q6YDeyvOsyJ2jPuX/igrKY8vjv/WYJ0gfQ5+jlRewa9SjfuprUquZJbQhdPuGwg0DOAio4q\nWnot7mhldYfrud7sI4YFAcCg8eW7C55jT/l+En3jxrXLLsxs19Jk+sPSPfQ77QNlWq6M0iuly59t\nuIgMGQuDMvBRe3O+MZOLjdmcqc4kRh03Kc8BILelALvTTkbAXABS/JO4I2Ydu8r28ualv/ONtK8O\nC+y32dp5r+gjeh02Xst7ix8s+va4d6HMrcVYbFZuCVl81UBIlHcEm2I28GHpHt4p2MHqiBW8XbCd\nhu4mDGpf5hqT8R8oEQvQGgnVBV0xOBEQMLzERyVXEu8bc8VmznoPHc+lP8lreW9zsTGbvJaCK+5c\nS5LE2wXbqe2qJ8U/ifquBg5VH+d47RnWR63izuh1IzZFGroaeTP/XbRKDc+lPzkpwY1U4xw+qzhI\nTnP+sKCQJEnYnfYxG4fbHH2cqD2Lj9qL/1j6gytm8CT4xnGhMYt3C3fyQckuClqL+EbaE6Pef3tf\nBx+VfopWqeWeuDsA13nM9+Y/x67yfRRZSmnpbaXQUkyhBQpai/nnxd8ZNg3sQNVRGrqb2BC3ks2R\nd83qDSZh9rE7nJTWtRMaoMNTM3XN+0/VneOtgu2sjVw5I4MSQzV1tbC7bB96lY67BzafvTz0bIha\njcPp4OOyz/ikdC9bEjdf82NIkkR1Zy1h+pARwQWrrY1fXXyZrv5ukv0SuTfuTsK9QjnceIR3cz/m\nd1l/4jvzvnHF4/KRmhOoFR48k/q4e0PkubQn+cWF37O/8gheKv2IgMV0ckpOXst7m6ymXB6f8/C4\nStDL2irYW3mYnOZLOCUnGoWGB+I3sSpiBXKZnCRjLDkNZjr6OidtElihpRin5CQ9IOWKx3KvgaEB\npW0Vwx7b0mvlo5I92CUHhc1lPJN65QnJVlsbf8z+CxUdVSwKms9jcz4PQO0eyBK6L/4uPi77lBJr\nOW22jlE3eCy9Vl7NeRMJiefTnybON5rWXgtNPS009bTQPPDV1NOKXerHKbkKsa29rXxacWBEUMg1\n4OkQHgoP/mH+sxgHJjRfTUbAXCL0oZxvyGJD1OphSQqD5X4quYp5A+d2U2XMMKPJZLoNSDCbzcuA\np4DfDLlMDrwI3AmsBO42mUzhwBOA2Ww2rwa2AL+e/KWPX1ldOyqlnFDjyF1ScDVW/mv+u3jIVbww\n7xl38MPLQ8/ioPk097SQ05wPQEV7Ffsrj2DU+LEwKIOy9gpeyfkLFwbGnaYZx/8L89caeCjxXn62\n4kf817J/dn89mHAPnf1d/DbzVVp7LZxvzKLL3s3y0CVTPkkmWOeKZg5tNl3ZUQMwKU0lBWE2UCs8\nuCfuDpL9r5wVKNwYnJJEUbUVo48GP+/x9SGobK/mTP0FwvWhbIrdgM3Rx7GaU6Net7XXQrG1jHjf\nGAwaX+QyOVuTtqCUKXj1/Ds0djdP2nPJbHSVjg09Sbkjei1JhgRyWwo4UnNy2PXfL/6EXoeNZL9E\nuvq7eSXnTfod/YzH8bozAOPaeFgftYp43xgym3L45YXf09jdzKrw5fxwyXd52HQf66NWMT8wjQiv\n0EnPupPL5NwRvRZw7QxfyeGaE5xryCTGO5JnUh/jR0v/kYdN96NTebKrbC9ZzXkjbvNByW7sTjvP\nLv7KpGWaRntHoFfpyG1x9YYC127obzNf4d9P/l+strar3v503Xl67D2sj1951ZIumUzGgqAM/m3p\n95nrn0SBpYhXct+k32kfcd0PinfR6+hlc+ztwz44qBQq7om7g+8v/CY/XfFv/PK2n7Au8jZaelvZ\nVrjTfT1Lr9X94fGRtHtEQEi44VQ3ddLX75zy0rGsgTYUByqPUjrQy3OmeuPiNvqc/dwfv2nERM11\nkbcRqDVyqPo4VQMby9fiaM0pfnb21/w1f5v7/RBcH7zfyHuHrv5uHki4m+cznnZPDXxgzp2sDLuF\nms46Xs5+g75Rjmnd/d00djcT4x01LENW76Hj+Yyn8FX78EHJLk7Wnh1xW3BVjNSP0mpjqrj61u1w\nT0osspSMeZuOvk5+ffEPZDXlEqYL5lHTA/xk+b+yJnKlO3gyN9A1tKhwHPc3XpdazADugSxXkmac\ng4REXkuB+3v7Kg9jlxwsDE2js6+LX194mVN150bctry9kv939jdUdFShU3lytuEC7xd/giRJVLZX\nk9dSQJxPDAmGWNID5iIhkT3KMbzP0c8fc/5MR38nD8TfjckvHqVcSaBnACn+SawKX86WhM18I+2r\n/NuS7/HS3T/hx8v/hR8v/xdS/JOo7KihtrN+2H2WtlXQ0mthXkDquAJC4DpPuTtuIxISH5UOz6Ep\na6+kqaeF9ICUqzaingzjyT1bC3wAYDab8wGDyWQaTLkxAlaz2dxkNpudwH5gHdAMDG6fGwb+Py1s\n/Q5qmrqICvJCqRj5dPscfbya+1d6HTYeTdoy4sRucIz7waqjOJwO3hroIP5o0hYeS/4SqcZkCixF\nfFz6GTJkpA402pwIjVKDv9bg/loVsZx7Yu/AYrPy28xXOFB5BBkyVoQtubYfwgQMTugY2my6yh0U\nEplCgiDMLrXNXXT12jGNM0tIkiR2FH8MwP3xm7g1bCkahZqD1cdG/UA92OhwaM+UYF0Qm2Jvx9rb\nzn+f+YWrjMrpGHHbiehz9JHXUkCgp5FQ3eelx3KZnK/MeQidypP3iz+mrqsBcDVhPNtwkUivcJ5L\nf5IlwQuo7Khm20A6/dV09HWS3ZRHqC6Y6Cv0uBtKLpPz+JyH8VLpCdEF8b0Fz/Fg4j1TfoIzKEwf\nQoJvLGZL8YgTOIDStnLeK/oIvUrHU3O/jFKuRClXcmvYUl6Y9wwKmYLthR9ic/S5b1NkKSG7OY84\nnxiWhE9e40e5TM5c/2Ta+zqo6qjh/2fvzgPkqst8/7+rqvd9q+5OZ+lsnW8SEhKykQBJCGE3bILg\nyCAo6uggF53f6Dgz/FSu916GuYM4yqgMozKDOqgIEWXfEkNCIAuEbHyzddZOr+l9re3+UUt639JL\npfvz+qurTtWpp5twzqnnfJ/n8Qf8PL3/t9jqQ9S21fFbu55AINDte/0BPxtObnXvWkMAACAASURB\nVMblcHH1jJX9+ryU2GS+MP+zzM027K36mJ/t+SXe0L/jZm8zb594h/dKdzA5pYDLJi7vdV9xrlhu\nmH4NU1In8V7pDnaU7QLg9wf/SJvfw80zP0Fy3MD6MohEg0Mnh7+fUJvPg60+FOl99sv9v+02odGT\nJk8TP9vzS145+ia1rb03Tff5fTx74IVI79GB2l25j22ndjEjfVq3K1ZiXbHcPutmAgT4jX2uQ0Kn\ns9rWOho9TV2eDwQCkX5x75XuYP3hlyLbXj36FgdrjrDAPY81ky7r8D6Hw8GnZt3I4twFHK4t7tBz\nLuxo3QkApqZP6bItKyGT+xd+geSYJH5tfx/pF3s23nr+7/bH+d57/xJZEDAUAoEADZ5GjtWd6NJX\n709HXmVzSbBvXYIrodvyps4O1hzB4/dyTeEV/N3SB7h04sVdVvXOywsmbgZaQuYP+Klo6tr7L9y0\nOzEmkcLU3q8Nwqt2d1fuA4J/180l75EZn8HfXPol7lt4L7GuOJ7e/1t+9MGT/O7AH3j7xDu8cXwj\nj+38KXVtDXxy5jq+vfwb5Cfl8taJTbx+fEOkl9B104I3gxaEFmuEE2rtY/31x89yvP4UKyYs7XaI\nU28uDpVyda70ee/0DoABDxOZm2WYnj6V3ZX7IsOdTtaXRJJEF+cPb+kY9C8plA+0H0dVEXou/HOq\nMabIGBMLrAHyrLXPAFOMMYeAPwN/O4QxD8jxsnr8gQDTeigd+82B9ZxqOM3KiSu67ehdkJLPnKxZ\nHKw5wn/ue4ZTDae5ZMIyTNZMXE4X917wl8zKmEGAANPTpw7Z8rurp67hqimXU95USUljKRe6LxjW\n5lJheT0khRJcCaPSuV1EZDgNtHTso8p9HKw5wrzsOZismSTGJHLZxOXUtdWzrdPFAQSTQi6Hi4s6\nlf5eOWU1X1txL/GueNYffon/u/1HHK8/OejfY1+Vpc3vYaF7fpeVGBnx6dw5+zY8fi+/2PtrWn1t\n/PbAehw4uMPcjNPh5NPmk0xKKWBzyXu8fWRLr5/1fulOfAEflxQs6/eqj6yETL53yd/zj8v+hmk9\nLAcfTuEbPBtOdlwtVNtax3/s/iWBQIB7593ZpYdBXpKbtVNWUd1aw6tH3wKCF8S/DyUGby3q2kTz\nXIWnuO2u3MefjrzGjvJdTE8vZGbGNHZV7o00E+/s4zMHKWsqZ3HeAjIS+//lNdYZw5fmfZbZmUXs\nrtzHf+x5mv/c9wx//87/4tmDLxDrjOEO079pqTHOGO654C+Ic8by3/Y5tpRs44OK3UxPL+TiYZq4\nJzLcIk2mBziMYCCCX+I9LJ+whNWTLqGsqYKXil/v9/t3ln/EzvKP+OORV3lwy//mZ3t+ycHqI11e\nF1518vbJd3jh8Cu0eFsHFGcgEGD9oZdwOZx82tzS4/FvTvYsFuVeSHHd8V5W3Hh4eNsP+P6OH3e5\nMXKoppjSpnLmZc8mLymXN4//mdePbeBQTTEvFr9OZnwGd86+rdvPdzqc3D7rZgD2nznQZXtx3XGA\nHm9q5Cfn8ZUFnyPG4eLne3/FoZpiILjq8Qc7f0JJYykOHDy97zecajjdw1+qf94v3ckj2/6Vb2z6\nDn+36SH+efuP+M67j/Cddx/hGfs8zx36E68cewt3YjZ/vfBepqcXUtZUQX1bQ6/7Da8mmpczu8f/\nRtMzp5DgSug2KVTf1kCTp7nL883eZn760VN8d+sjbCv9oMO28qYKzrRUMzurqM9Vv3lJbnITc9h3\n5gAen4c3T2zE4/dydeEaYpwu5mTN4ptLvsqklAI+rj7IhpObefbgCzx/6EVinTF8ZcHnWTtlVaTt\nS2Z8Bn84/DK7KvcyNW0KszODK8CyEzOZnDoRW32ow++z8dQWtpV9wLS0KdzRy7/jnszPnkNiTALb\nyj6IJD3bfB4+qPgoNBl1YK0BHA4HN06/FoBn7PM8uuPfeHjbDzhQfYjC1MmYzJkD2t9gDKbRdOSv\nZq0NGGPuBn4O1ALFgMMY85fAcWvttcaYBcDPgCW97TQzM4mYmMEvG3e7u28EuXlfcHnfhSa3y2ve\nOrKFrae3MyOzkC+v+IseS7NumXcN+/98gB3lu8hMSOeLy+/ocLfrH7O/yq92Pc/yyYu6fEZPcfXH\nF3JuxxHr563iLdw6/9pz2ld/Y0rPjMfxvoMznjO43am0eFspa6pgjnsmebnDu2x2qH6/oaa4+i8a\nYwLFJT07GLoDHE4KvXbsbRJc8azq5q6Rz+9j/eEXcTqc3DLz+sjzayZfFrmDtXzCksiX55KGUk41\nnObCnAu6TL5yOBxcMmUJE2Im8fzBF9laup1/3vYjFuct4PppV/XamLC6pYYXjrzCxfmLI0vfw4mC\ni9xdmx4CLHDP49KCZWwueZ9/3v4jShvLuLRgGVPTgndK41yxfHH+XfzTth/y79t/xS0z13H5pEu7\nXCgFAgG2lLxPjMM14NGow13+3Jv5OXPJTsji/dKd3DjjOlJik4Njcz/6BbVtddw843pm9XDRde3U\ntWwr/YA3j2/k4gmLOVZ3ghP1p1iSt7DHaaDnYk5WETEOFxtObqbZ24I7MZu/mn8PTd4m/s/7j/Fb\nu55ZmTO6NA5/O5Tw6nznvD9iXbH81YV38+NdP4+Uy+ckZrNiwlIuzl/U70lsELzYv63oRn5tf8+v\nPv4dDhzcPuuW86Jxrkh3jpU1kJwQQ27GuTVNDgQCvHrsbVLjkrm0oOPq/71Vwf/v5mXPoTBtMnsq\n9/PG8Y1cPmsZGfQ9SSv8xf66qVeyq2JPJEm0OHcBd5hbIuegF468wpbT24hxxuDxe/iocu+AVjUU\n1x2jtKmcS6Ys6TJRqbNbi25gb9XHrD/8Egty53U5Zn1Q/hH1bQ3UtzXwXumODuXI75QES7KvKlxD\nVkIGj+74MesPv0RSqHH1PRf8Ra8TJVPikilIzudI7TE8fi+xzrNfeY9GkkJdVwqFTUsv5AvzP8tP\nP/oFP/3oF9w153aePfhHzrRUc3XhGgrTJvPk7v/iiY+e4ptL/gcpcd23KOmNx+/l2YMv0OxtITfJ\nzcyMaeQkZHOmtQZ75hCbQiXfaXGpfHXhF0mLS2VGxlT2nbEcrj0a6R/YnQPVh4lzxfW6YsfldFGU\nOZ3dlfuoaq4mOzG4+KDJ08z/fu/7tPnbuGrKGq6YspJ4VxzlTRX89KP/pCw01fKN4xtZkrcwcp2w\nL5SAm5vV9zQzh8PB/Jy5vHniz+ws/4hNp7aSHpfGigln0wW5SW7+ftnXaPI0hfr7nKG2rY752XM7\n9PPMTMjgqwu/wPd3/phGTxPXTr2iw7XLQvc8TtSfYk/VfpblL+JEfQnPH/xTaKXsXR3+bfRXrCuW\nRbkXsrnkfQ5UH2Z2VvCmSrO3hZXt+hAPRFHmdOZkzWL/mQM4cDA3y7By4nLm5cwZkfNnf/4KJZxd\nGQRQAETSotbajcBKAGPMw8BRYDXwamj7LmNMgTHGZa3tcX18dXXXpYP91bn5ZHt7DgVXvOSkxHZ4\nzZaSbTxjnyMpJpG7Z/8FNWdagJZu91HgmsSE5DxON5bxqaKbaKr10UTHz7upMNgQrv1n9BZXf91U\nuI5rJl5FAvHnvK/+xpSVkMHJ2tNUVNRzpPYoAQLkJ+QPyeefS1yjQXH1XzTGBGMjrvMheXS+DiQo\nO9NEbIyT3MzEYLLl8Cs4HA7mZs/uUg++ueR9ypsqWTlxBfnJeZHnM+LTWZp3EVtLt7OnXYPgbWXB\nu2i9JU9SYpO5a+7tLM2/iPWHXmR72YfsKNvFsvxFrJt+dbcrRP945FXeL93J+6U7WZ6/hBtnXMue\nyv1kJ2T2WuZ7a9GNHKw+QmljGckxSdw4/boO23MSs/nyhffws71P8+zBFyiuPcZnZt/aocwr/IVg\nce6CbqeZRSunw8nqSZfw3KE/seXU+1xZuJpf7P0Vx+tPccmEpb1OJol3xXFr0Q38x56n+Y19nvKm\nSmKcMV3+fkMlISaBoswZ7D9zgOTYJP56wedJiUsmJS6ZddOv4flDL/L7g3/k7rmfjrynrLGcfVWW\n6elTmZI2aVCfG+eK48sXfo53SrZSmDo5OBV1kKugLilYxp6qj/moci+rJq1QT0I5b/n9Aapqm5mc\nm3rOqwLfPb2dPx55hRiHi3nZcyONbwOBAHsqPybBlcCM9Km4nMGx6D/44Al+/P5/8TcL7+u1oX94\nallGfDqfmHYVn5h2FYdrj7L+0IvsKN/FoZojfGb2bZQ3V/LasbfJTczhL+fczvd3/pjtZR8OKCm0\nJbTq54ppfZfbZMSnc/20q3j+0Iu8dXwTN864tsP2d0q24sCBy+nipeI3WJq/iFhnDPVtDXxQvpv8\n5DxmpE/F4XDw1YVf4LEdP6HR28S6aVf3OGCgvVmZMyhpLOVY3YnI6wOBAMfqTpCdkNVndccF2YbP\nzrmDp/b9N/+++78AWDftmkjS4fppV/FS8ev8bM8v+erCLwy4J96H5btp9DRx5ZTV3DLzEx22+fw+\njtWf4FBNMQtyLohcj8xID/4eh2uKe0wK1bXVU9pUzpysWX3GZDJnsrtyH7b6EJckLgXgzRN/pt7T\nQIzDxZ+KX2XTqS1cWnAxG05upsnbzNrJq6hsOcOuij0crj0a+due7SfUv36b4aTQbw48T5uvjRun\nX9vtzaOk2CQKY5N6vQmTn5zL/7forzlRf4p52R1buSxwz+OPR15lV8UeFrjn8Yu9v8Ib8HHXnNvP\naUDEsvzFbA4Ni5qdVcT7pYMrHWvvL+d8ih1lu1jgvmDEK3T6k3Z6jWCzaIwxi4ASa23km4wx5mVj\nTK4xJhm4AXgDOARcHNpeCDT0lhAaTsUldR2y+/6An98f/CO/+vh3ka7z2X00gnI4HHz+gjv53AWf\nYcEwd/7uzrmOuB2o3CQ39W0NNHub2zWZVj8hEene+TyQoLK2hey0BBwOB9vLPiRAAH/Az+vH3u7w\nujZfG68cfYM4ZyzXT7uyy37WTlkFwK8+fpZHtv0rj2z7Vzae3EyCK77LBUp3ZmcV8XdLH+CL8+5i\nQnIe75Xu4Ac7n+jS+LmmtZbtZR+Sk5DF5JQCtpZu57vvPkKLr5UF7nm9fmmJd8XxuQs+Q16SmzvM\nzd3e2ZyZMY1/vvofmZ4+lR3lu/jn7Y+zt8pS1XwGn9/H5pL+N5iONismLCXOFcfGU1v43YE/sLty\nP7Mzi/i0+WSfX/YWuucxJ2sWtjo4de2KySsjd1WHw6UFF5MZn8Ffzb8n0usPgquApqRO4v3SnWwp\neZ+q5mp8fh8bTgZL/sJlcoOVEBPPlVNWU5Q5/Zy+ADscDj479w4+O+cObp7xib7fIBKlahpa8foC\nuDPOrQdaaWMZvzuwHgBvwNeh101ZUzlVLWeY067spihzBpdPupRTdaX8x56nI72+ulPSUEqDpxGT\nOROHw4HD4WBmxjS+vugr3Dj9Who8Tfzko1/w+4N/JD0ula8uDE5Zmpw6kf1nDtDgaezX79DibWFH\n+S6yEjIj/Wj6smriJaTGprDx5Gaa2vUOOtVwmiO1x5iTPYvVEy+hurUmMrBh6+nt+AI+VhYsjxyH\nJiTn8bVFX+aOWbdwzdQr+vXZRaESngPtyqMqmqto9DT1qx8eBG/qfGrWTcQ5Y7l15jqum7Y2EtN1\nU9eywD2PAzWHef7Qi/3aX3ubQr9v51VjEFzFMz19KlcXriEvOTfyfGHaZFwOF4drjva433Dp2KyM\nvkuYwmVJtvogECwbe+vEJtLiUvmfl/wD105dS5O3hZeOvkGbr4275tzOJ4vWccXkYN+6DaEBDm0+\nDwdrDlOQnN/vRMv09EKSY5Jo9bWRGpvCped4XZGXnMuS/Iu6nLvyk3LJS3Kzr8pGpp9eMXllpFR7\nsGakTyU7IYsPKnZT2XyGfWcOMCV1IhPa3TQcqIz4dNZOWTUqLVv6TApZa7cAO4wxWwhe6N9njLnH\nGHNL6CVPEkwcvQM8bK2tBJ4AphpjNgK/Br48LNH3oaHZQ3lNM1MnpOFwOGj2tvDTj57irRObyEvK\n5RtL7qcoc3q/9lWQkt+hUehYlp8UPPiUNVWoybSI9Md5OZCgpc1LQ7OHnPTgxf62sg9wOpxkJ2Sy\n9fR2qltqIq/986l3qW2r5/LJl5EW13XlVkFKPpcWLKPN76G0qYLSUF+2NZMv63UKVHsOh4OFufP5\n+2VfY9XEFVS1nGFDpyaZb594B1/AxzVTr+AbS+7n5hnXEyDYeHhRp75F3ZmSNolvL/8Gi3s5n2Ul\nZfC1i/6KKyavpKypnB/v+hnffvef+NrGf+S90zvITsgccL18NEiKTWR5/hJqWmv586l3KUjO5wvz\n/7Jfd3eDzUtvwuVwkRKbzNXDPKr4otz5/K9L/4EZGVM7PB9eQeB0OPnVx8/y7Xcf5msb/5FNp94l\nMz6DBTkXdL/DUZAYk8DFExb3+9+/SDQqrw72IXGfQ+mYx+fh53t/TZvfw2fn3EFKbDKbTr0b6eez\nJzSB6YJOX1I/OXMdiybMY/+ZAzy9/7c9NmwOl451Pi67nC6umXoF31xyPxNTJpASm8x9C78QuRm+\nJG8h/oCfD8q771PW2c7y3bT52jqUSfclzhXL2imraPG18na781k4AbSyYDlXF64h3hXHq0ffosXb\nwjunthLrjO2y4qIgJZ9Vk/pfmlOUMR0Hjg69lSKlY900me7J5ZMu5V9W/U+uCN38CXM6nHx2zu3k\nJ+fx9sl3Ir2H+qOkoZTDtcXMziwiN6nv8sCwOFcsU1IncaLhVI/9oA7UBH/fon6cpyck55Eal8KB\n6sOh8sa3aPO1cc3UK0iPT+WG6dfw3RXf5LqpV/L1xV9heai8a0b6VCalFLCrcm9oymqwsXVfU8fa\nczldXJATnIC2dsoq4lxx/X7vQDgcDha459Hm97C97EMmp07kxhnnvtLX4XCwLH8Rbb42ntr7a/wB\nP8tGoCH0cOlXEZ219ludntrVbttzwHOdXt8A3H7O0Z2jo6frACJNpn++91fsq7LMzTJ8ft5nSIw5\nt9rgsSoygawxmBSKdcb22t9CRMa9fGBHu8fhgQR1tBtIQLC8eA2wwVr7SOgGwyGCSaE+lxMMde+5\nY6XBc8Sk/DRa4uo51XCaxQXzWTZxIT/Z9jTvVGzh84vuoMnTzBvvbCApNpFPL/pEj70DHnB/7pxj\nCrsn41Z2vvgRrx5/i3XzLictIZWmtmY2l7xHRkIa189bRawrls/k3cCVs1dQUl/OwglzB/z5PcnP\ny+DLeZ/h0rJF7C23lDdUUdFYxZnmGm674BPD3mOuJ+daSvnJhKvZVPIuGfFpPHjF/eQk9W9kLICb\nVL6T9HUSY+OZkpHbcdsIlni63ak8lPo3fHB6D+UNVZQ3VlHdXMOn5q0jPy+jw+uikeIamGiNa6DO\n1xLjippzTwo9f/hFTjWc5rKCi7l4wmIqW87wUvHrvHt6G2smX8aeUB+vCzp9oXY5XXz9ki/ynTe+\nz/ayD0mOTeZTRTd2WQkRXgnTUzPaSakF/P3Sr+H1ezuU5yzOXcDzh15kR9mHrOxjuiDAu6ffx4GD\n5fm9tontYuXEFbx+fANvn3iHKyavxIGD90t3khGfzgXZs3E5XaydvIqXjr7Bk7ufprLlDCsmLCUp\n9ty+pyXHJlGQks+RumN4fB5iXbFnJ4/10k+oOz3dPEiISeDO2bfx6I5/4zf2eb619IF+3Wh4p+Q9\ngH793TubmTGN4rpjHK07Hukr2N7B6sPEu+KY0o8b+g6HA5M5k+1lH/LxmYNsOrWVrITMDquXMuLT\nWTf96i7vu3zSpfzy49+x6dRWPP7gqubO/4b7ct3UK8lOyGT1pEsH9L6BWuiex2vH3ibeFcfnL/jM\noPoIdWdZ/iJePvoGxXXHcTqc5/UCkqH5i0SpI5GkUCo1rbXsrzrA1LQpfPnCewZc9zmehBNApxpO\nc7qxjMLUyWoQKSIDMSwDCYa699zB4uBI1eQ4J6/tD97BXJA5nznJc8lKyOTNw++wKnclm0u2Ut/W\nyLpp19Bc66eZoelR1VdfqWsL1/LswRd4esd6bp91M68f20Czt4Wrp6zp0AfPQTwTYyYPWe+s9nHl\nOyeSn9/1wnI0+nQNRX+wWJL4xuKvkh6fRqAxlorGge0vm1zwDH3/wIHKIpe1+V1LKMJxjIVeaiNp\nLMQVzcmj9iXGxpg5BM8HK0LbwiXGi4Aq4GVjzHrgZoIlxn9vjCkA3gJmj3TsFbXBpNBgmkx7/V62\nlGxj48ktTEjO49aiGwFYPfESXj+2gbdObGJp/kUcrj1KYerkblehxsfE8ZULP8djO3/KxpObSYtL\n5dp25VM+v4+DNUfITcrptSG8w+Ho0q8lMyGDGenTOFRTTHVLTYf3BwKBDsmn0sZyjtQeY3Zm0YBL\nZxNi4rli8kr+eORVNp18l+TYJFp8raydsiryfeyKKavYeHILH4fKmAaTLOnOrMwZnGo4zdG64xRl\nzuBo3XFcDheTU4auz9n09EIumbCMLaffZ8PJzZFy8p60+tp4v3QH6XGpkdHsAzEjYyqvHw/2Feqc\nFKptraOsqYK52abf33XDSaGn9v03Xr+X66dd1a+kyZK8haw//BKbT71HYmwica44pqdPHdDvkpuU\nw7rp1wzoPYMxJXUSn5h2FdPSCzuUZJ+r3KQcpqUVUlx3jLlZZsimkI+GMf1Nv7Qq+AVisjuFD8p3\nEyDAsvxFSgj1IS85+D/LhxW78Qf8Kh0Tkb70OZDAWrvSWruOYGLoKHAp7QYSAAXGmBE9OFfWBpMq\n2WkJbC/7gHhXHPNz5uJyuri68HI8fi8vHHmZN49vIiU2mTWTh/dOVmcrJy7HnZjNplNbOdVwmrdP\nvEO8K47LhuhiebwqTJt8Ts0lRWTAzssSY4CKmuB5YiArhaqaq3nh8Cs8uOX/8JsDzxPnjOXzF9wZ\nKaVMiUtmxYQlnGmp5r8//j3+gD9SRtOdpNgk7lt4L5nxGbxY/BplofJkgGP1J2j1tWEyu64Y6Y8l\neQsJEGBHebAIJBAI8Nqxt/nbP3+HPxx+GU+ol9G7p4MLuy4pWDqoz1k96RISYxJ488Sf2XByM06H\ns0NvusSYBK6eugYItqyYkjq4hvmdhfvqHKg+jMfv5VR9CRNTJgz5NMybZl5HcmwSLxa/1qH0vDs7\nynbR7G3hkoJlg/pOOi29EIDDtUe7bBtIP6Gw8AqzBk8jeUluluX1b7JorCuWSwsuptHbRGVzFSZz\nJjFDtAJnqIUbg8/J6l8T7IEIJzAvm9i1N9T5JDr/yw2RyroWHA7ISI1n5+GPcOBgYQ/jeuWs9Lg0\n4l1xVLVUA+onJCJ9eg14CHiip4EEwN1AI8GBBI8CEwkOJPj9aA0kqAolhdriKqlqqWZZ/qJITfvy\nCUt55ehbbD29HYBbp63rMIVrJMQ4Y7h5xvU8uedpfvThk9S3NXDF5JXnvKReRGSEjUiJMQx9mXFN\nQxsxLgdF03NwOftuvv7+yQ95dOu/EwgESI5LYt2stVxdtJr8lI6rE25LvI5NJVv5sGIPACtnLsad\n1f1qL7c7FTepfJ7beXTzv/NGydt8bcW9AGwsD5ZDLS2cN6jVYlelruB3B//Arqrd3Lrgan667Ze8\nczyYAHrt2NvYmgN8ZdldbCv/gOS4JK6YszyS3BrY56Vy/awr+P2+l2jwNLJs4kKKJnVM/NyaeTUN\ngTounbKU3Ny0HvbTt/ZxXZw+n3/f7eBo4zEaY2rwBnzMyZsx5Cvr3KRy18JP8tNtv+RPJ17hby75\nIn6/nyPVxzlYVcwk/wRMzgziXLFs/fB9HA4HN8y7gpzkgcfhJpXJaRM4Wn+CzOwkYtollo4fDf57\nWDZtPu7svvcd/reV91EOZY2VfGbhTR3KkPtyc/KVvH58A/6An4sLLxyyv2u0rnzsLq5P5Kxm+cwL\nB1SOPtSG4u81ppNCVbUtZKbGU++p40jtUYoypkfGP0rPHA4HuUnuSJPp/tSkisj4Za3dYowJDyTw\nExpIANRaa5/n7ECCAKGBBMaYJ4CfhwYSxDAKAwkqQ2UBxa3BJp9L290di3XGcOWU1Tx78AUy4tNZ\nOXHFSIcHBEepzkifxuHaYpwOZ2Tih4jIeWxYSoxh6MuMT1U0kJ2WwJmqhj7fHwgE+M2uP+LAwZ2z\nb2Nx3sJgAqUZKpo77tdFAgtz5vFBxW5S41JI8WZ0WyrYPqZpcTOYkjqRLce3szrvMialFvDByb04\ncJDvmjjoEsjZWUXsq7J885WHKWksZVpaIZ+dewdvHt/IOyXv8a3X/wmA1ZMupTZUujyYksuLs5fx\nJ9cbtPraWJazpNv331x4AzD4EuXu4pqUWsCByiNsKw4m4PJi84elXPSClHlMSytk64mdfPeNH3C0\n9jhN3ubI9lhnLFPTJnO45hjzc+YQaIqlomlwcRSmTOFE3Wk+LLYdRrXvPv0xCa6EHv89tdf+b3V9\n4VUcrTvB9PiZA/zbxLA4dwE7yz9iSty0Ifm7np/lvAMvRx8qQ1VmPGaTQl6fn5qGVmZOTOeDimBX\n/UW5C0Y5qvNHXigpFONwndNoPREZH87HgQSVtS3EuGBv9R5SY1O6NOm8tOBiTtSfYnHegiFfat5f\nDoeDW4vW8S87/o1l+Yt67RkhIhKl+iwxBlYCGGMeJrhiaDXtSoyNMQXGGNdIrihtbg1OqJya378b\nysfrT3KioYQF7nms6EeZ1VWFl/NhxR4WuOf1q3enw+HghunX8m+7fsYfj7zKvfPupLj2GJNSC0iO\nTepXjN1ZkruQfVWWksZSlucv4dOzP0msM4a/mH0rF7rn8av9v6OurZ5LJgyudCwsOTaJTxXdRHHd\nMUxW902xh0NRxnRO1J9i48ktwMAmjw2E0+Hk0+YWHtn+Q/ZVWTLjM1jons/0jKnU+KrYeWovB0OT\nwVZNvOScPmtGxjTeKXmPwzXFkaRQTWst5c2VzAs17x6IJfkXsSS/f2Vj1XqmDwAAIABJREFUnX1m\n9q1cP+3KAfeakugyZpNC1fWtBAKQnZ7AzrJ3g6VjufNGO6zzRrjZdEHKBPVgEpExqbK2hbQJtTR6\nmrh80qVdjnVxrlg+O/eOUYrurMK0yXxn+TdJ76YJqYjIeeC8LDHuPHmsydNERXMVVS3VTE2bTFZC\nxy/Bm9qNWe+PwrTJ/MOyr3fZT2/mZM1iZsY09lTt5/XjG/EGfF1G0Q/Uwtz57K36mBkZ01g1cUWH\nBtMXZBv+/+V/S21rLflDcJN4RcHSfiXMhtKszBm8dWIT1a01JMUkkpvY/xHwAxWe9OZ0OMlLckf+\nlm53KtdNuoaa1lpqWmsHPP2ssxnp0wA4VHuUKwg2tj4Q6ifUn1H0QynOFTekzZtldIzZpFC4V0RS\nqodddccxmTO77eov3QsnhdRPSETGopY2Lw3eetLdwdKxJf1srDhachJHr1ZdRORcnK8lxuEm01VJ\nu/jGn3/doRQoJzGbf1z29UgfuiZPM9vLPiQnIWtAq2AKUvL7flE74dVCj+38CS8XvwEw6CbTYfGu\nOD4/784etyfGJJA4wj31htLMjGk4cBAgQGHa5A5Jr+HQ23/TjPj0IRl0kJWQQUZ8Oodritle9iEA\n20o/ADjnJKGMT2M3KVQXPJA3xp+AZliUe+EoR3R+mZM1i7nZZtBTBkREotn2k3tJmLeZtlgPF+cv\nZmq7mnwRERla52OJcUVNMzh8HPF+SKwzhnnZc3AnZlPdWsOHFXt44cgr3BYaM/9+6U48fg+XTVze\nr1KwczEzYxpzsw37qiwuh4uZGdOG9fPOd4kxiUxOncjx+pPnvEInWjgcDooyZrCtbCe/2PvryPPJ\nsUlMSikYxcjkfDXmkkKVzVVkxmdEkkKlvkM4HU4WuFU6NhBJsUnct+De0Q5DRGRI+QN+Xj36Fn86\n+hq4HMyNXcldc9YN+51DERE5v1TUNuNMO4Mv4OXyiZfwyZnrAGjzeSjZ9hgbTmxmoXs+M9Knsqlk\nKy6Hi+UT+uyFPSRumH4N+6os09MLiQ+tVpKezc4q4nj9SWZkTB3tUIbMLTM/wcyMqQQIRJ6bkjpp\n2JOSMjaNqaTQ7sp9/PSjp0iNTSGxeTLOzATK204zO7OI1LiU0Q5PRERG2caTW/hT8WskOlKo3j+P\nxWuXKiEkIiJdVFQ348qoAGBe9pzI83GuWO6aczvf3/ETfrX/d3xq1k2UNpaxJG/hiH3fmJI6if+x\n8Etq7ttP1xSuoTB1ErPPsdQumqTHp3LZxP71rxLpy5hKJX5UsRcAb8BLecx+4ouCtZWL8lQ6JiIy\n3gUCATaXvEeMw8XCwM0EGjPISU8c7bBERCQKldc0EZNZQWJMAjPSp3bYNj19KmsmX0Z5cyX/sedp\nAC7rZ4PpoWKyZpKTmD2in3m+SohJYGHufN0EEunBmEoK2epDJMYk8vBl3ybx1AocNRMpypjORW4l\nhURExrvj9Sc53VjG/Jy51NUFT3856edv80wRERkefn+AqtZKiGtmdtasbifx3jD9GnITc2j1tZGf\nlKvePiJy3hozSaHK0IhIkzkDl8NFXWkmuXWX8LVFXyYpVneCRUTGu/dKdwBw8YTFVNU2E+Nykpas\nXgwiItJRdX0rjrRw6djsbl8T54rjrrm3E++K46rCy7UKRUTOW2MmKfTxmYMAmMyZ1De24fX5yU7T\nHWAREQGvz8v2sg9JjU1hbpahsraF7PQEnLqIFxGRTipqmnFmlANwQQ9JIQiWkT266nsj1mBaRGQ4\njJmkkK0+BASTQpWhyWPZKgsQERFg5+k9NHqaWJp/EV4v1Dd5VDomIiLdOnmmGmdqDVkxeX02j9YK\nIRE5342JpJA/4OdA9WEy4tPJTXJTVRtKCmmlkIiIABuObgVg+YQlkRsHSgqJiEh3DtYewuEIUJQ6\na7RDEREZdmMiKXS8poQGTyMmcyYOh4MqrRQSEZGQ+rYGPijZzaSUAiamTKCqthlQUkhERLp3srUY\ngIvyLhjlSEREht+YSArtLvsYCJaOAVopJCIiEdvLPsQX8HPxhMUAVOocISIiPfAH/NQ5TxLwxDEn\nr3C0wxERGXYx/XmRMeYxYDkQAB6w1m5rt+0m4EGgFXjGWvu4MeZe4K52u1hire29IPcc7CkPJYWy\nOiWFdBdYRGTce+/0dlwOJ0vzLgLOJoVy0jWZUkREOjpefxK/q5W4xkJiXF1H0YuIjDV9JoWMMauB\nImvtCmPMHODnwIrQNifwOLAIqAJeNsast9b+DPhZu/ffPkzx4/V72VdxiPykXDLi0wGoqmslPs5F\nckK/cl4iIjJGVTZXcaKhhMUF8yPNQit140BERHrwYdk+ALKYPMqRiIiMjP6Uj60F1gNYa/cDmcaY\ntNC2HKDGWlthrfUDbwJXdnr/t4HvDVG8XRytO0GrtzWySgigqq6FnLQETQMQERnnkmOTuTh/MXfM\nuyHyXFVtMzEuB+kpcaMYmYiIRKODZ44CMDlp6qjGISIyUvqTFMoHKto9rgg9F/451RhTZIyJBdYA\neeEXGmOWAiestaVDFG8X9szB4GdlFgHQ1OKludWrO8AiIkJiTAKfnXsHUzPP3vGtrG0hOy0Bp24c\niIhIO4FAgNNNJfhbkpiQnjHa4YiIjIjB1FdFrqKttQFjzN0ES8pqgeL224EvAE/1Z6eZmUnExAy8\nbvfIR8U4HA5WzLyQ5LgkiktqAZiYl4rbnTrg/Q21aIihs2iMCRTXQEVjXNEYEyguOau1zUd9k4cp\nucPW5k5ERM5TVS3VtAZa8Dfm456hvnMiMj70JylUwtmVQQAFwOnwA2vtRmAlgDHmYeBou9deDtzf\nn0Cqq5v687IO2nxtHKgqZkZmIU21Ppqo59DRMwAkx7moqKgf8D6HktudOuoxdBaNMYHiGqhojCsa\nY4KxEZeSR0Onsi7cT0gX+yIi0tHx+pMABBrTcWeo6kBExof+lI+9BtwGYIxZBJRYayPfZIwxLxtj\nco0xycANwBuh5wuABmtt29CHHeR0OJmRPpVriy6PPFdVp1HDIiLSvaraZgByVGIsIiKdHK8LJoX8\njWkkxGnymIiMD32uFLLWbjHG7DDGbAH8wH3GmHuAWmvt88CTBBNHAeBha21l6K0TgPLhCTsoxhnD\n1xZ9ucMdd42jFxEZecaYx4DlBM8FD1hrt7XbdhPwINAKPGOtfdwYcy9wV7tdLLHWDntN19lx9DpH\niIhIR8fqw0mhdGIH0dZCROR81K+eQtbab3V6ale7bc8Bz3Xznh3AdecU3SBUaqWQiMiIMsasBoqs\ntSuMMXMI9plbEdrmBB4HFgFVwMvGmPXW2p8BP2v3/ttHItbGZg8AqUmaPCYiImf5A35O1J8kzpdG\nsz+GGJeGEYjI+NCf8rHzSlVtCy6nRg2LiIygtcB6AGvtfiDTGJMW2pYD1FhrK6y1fuBN4MpO7/82\n8L2RCNTj8wMQGzPmTn8iInIOyhoqafa2EO/NAiDGpfOEiIwPg5k+FtWq6lrISovXqGERkZGTD+xo\n97gi9Fxd6OdUY0wRwUEEa4AN4RcaY5YCJ6y1pX19yGCnVIa53anExsUGf85JiYoG3tEQQ3cU18BE\nY1zRGBMoroGK1rjGosNnjgEQ05YJ6OaBiIwfYyop5PH6qGtsY/aUjNEORURkPItk5a21AWPM3QRL\nymqB4vbbgS8AT/Vnp4OZUhkW7j1XFyoxbqhvGfWpdGNhMt5IUlz9F40xgeIaKE2pHFlHwkmh1kwc\ngMupG8wiMj6MqaTQmbpWQE2mRURGWAnBlUFhBcDp8ANr7UZgJYAx5mGCK4bCLgfuH/YIQ1Q+JiIy\n8s6HYQSHq4/jwIGjJZ2YmFYcqjoQkXFiTCWF1GRaRGRUvAY8BDxhjFkElFhrI7e3jTEvA3cDjcAN\nwKOh5wuABmtt20gF6vEGk0JqICoiMjLOh2EE/oCf4urj5Cfn0uZxqp+QiIwrY+qIV9cQ/F6RkRo/\nypGIiIwf1totwA5jzBbgh8B9xph7jDG3hF7yJMHE0TvAw9baytDzE4DykYzVG1kppFHDIiIjJOqH\nEZQ3VdDibWVK6iQ8voBWk4rIuDKmVgqFywLidbEvIjKirLXf6vTUrnbbngOe6+Y9O4Drhjm0DsIr\nhWK1UkhEZKSMyDACGPxAgv2N+wC4oGAm+wJtxMe5oqZPU7TE0ZniGphojCsaYwLFNVBDEdfYSgp5\n1StCRER65lVPIRGR0TYswwhg8AMJ9pw6CECWI4eWthMkxrmiovn4WGiCPpIUV/9FY0yguAZqqAYS\njKmr4kivCF3si4hIN8JJIZf6RYiIjJQ+hxFYa1daa9cRTAwdbffay4Etwx3g8fqTOB1OJqYU4PX6\ndeNARMaVMXXE83h9gO4Ai4hI9zxePzEuB05NlRERGSmvAbcB9DSMwBiTa4xJJjiM4I3Q8yMyjMDn\n93GivoTJ6QXEuWLx+vxqNC0i48rYKh/zBQCI1YFcRES64dHFvojIiLLWbjHGhIcR+AkNIwBqrbXP\nc3YYQYBRGEZQ2lSOx+9hRuYUAoFA8OaBbjCLyDgyppJCXvUUEhGRXnhUFiAiMuKieRjBmZZqAGbl\nTMfnDxBAN5hFZHwZU0e8s1NlxtSvJSIiQ0RlASIi0t7cLMOXL7yHVVOXaxiBiIxLY+qI5/Gpp5CI\niPTM4/XrxoGIiES4nC7m58wlxuk6O7RG5wkRGUfG1BFPI+lFRKQ3Xl9A5wgREemWN9yfVOcJERlH\nxtQRTyPpRUSkN8HpYzpHiIhIVx5feKWQJlSKyPgxpq6MvZo+JiIivfD61GhaRES651V/UhEZh8bU\nEc/jVU8hERHpnt8fwOcP6A6wiIh0S1UHIjIe9WskvTHmMWA5EAAesNZua7ftJuBBoBV4xlr7eOj5\nO4FvAl7g29baF4c49i48Xj8OwOXUBb+IiHTkiUyVcY1yJCIiEo0i08e0UkhExpE+j3jGmNVAkbV2\nBXAv8MN225zA48D1wCrgBmPMJGNMNvAd4DJgHXDTMMTehSdUFuBwKCkkIiIdedUrQkREenH2PKGk\nkIiMH/1ZKbQWWA9grd1vjMk0xqRZa+uAHKDGWlsBYIx5E7gSaAbesNbWA/XAl4Yl+k48XvWKEBGR\n7mlCpYiI9EblYyIyHvUnKZQP7Gj3uCL0XF3o51RjTBFwFFgDbAi9LskY8wKQCXzXWvvmEMXcI4/X\nr4O4iIh0Sw1ERUSkNx6Vj4nIONSvnkKdRNbdW2sDxpi7gZ8DtUBxu+3ZwC1AIfC2MabQWhvoaaeZ\nmUnEnEOfB7c7FT8QHxeD25066P0MtWiKJSwaYwLFNVDRGFc0xgSKS4LO9hTSxb6IiHQVmWSs84SI\njCP9SQqVEFwZFFYAnA4/sNZuBFYCGGMeJrhiKBHYYq31AoeNMfWAGyjv6UOqq5sGGnuE251KRUU9\nra1ekhNjqaioH/S+hlI4rmgSjTGB4hqoaIwrGmOCsRGXkkdDI1IWoDvAIiLSjfAkY/WeE5HxpD9J\nodeAh4AnjDGLgJJQryAAjDEvA3cDjcANwKNAPPCUMeYRguVjKUDlEMfehcfn13JPERHpVnilkMqM\nRUSkO1opJCLjUZ9JIWvtFmPMDmPMFsAP3GeMuQeotdY+DzxJMHEUAB621lYCGGOeBbaGdnO/tdY/\nHL9Ae2o0LSIiPVFPIRER6Y1WlIrIeNSvnkLW2m91empXu23PAc91854ngCfOKboB8AcCeH0BJYVE\nRKRbWikkIiK98arRtIiMQ4NpNB2VvBohKSIyKowxjwHLCa4YfcBau63dtpuAB4FW4Blr7eOh5+8E\nvgl4gW9ba18c7ji93lBZgC72RUSkG+GVQrrJLCLjyZg54imzLyIy8owxq4Eia+0K4F7gh+22OYHH\ngeuBVcANxphJxphs4DvAZcA64KaRiNWr6WMiItKL8HlC5WMiMp6MmZVCyuyLiIyKtcB6AGvtfmNM\npjEmzVpbB+QANdbaCgBjzJvAlUAz8EZoaEE98KWRCFTnCRER6Y3KjEVkPBp7SSFl9kVERlI+sKPd\n44rQc3Whn1ONMUXAUWANsCH0uiRjzAsEJ1R+11r7Zl8flJmZREyMa9CBJiTFBfeTkYTbnTro/Qyl\naImjM8U1MNEYVzTGBIproKI1rrFK3ydEZDwaO0khlQWIiEQDR/gHa23AGHM38HOgFihutz0buAUo\nBN42xhRaawO97bi6umnQQbndqZwJvb+luY2KivpB72uouN2pURFHZ4prYKIxrmiMCRTXQA0kLiWP\nhkZ4JL1WConIeDJ2kkJqNC0iMhpKCK4MCisATocfWGs3AisBjDEPE1wxlAhssdZ6gcPGmHrADZQP\nZ6Bne0U4+niliIiMR97ISiGdJ0Rk/BhzSSEt9xQRGVGvAQ8BTxhjFgEloV5BABhjXgbuBhqBG4BH\ngXjgKWPMIwTLx1KAyuEOVD2FRESkN2crDwZfqiwicr4ZM0khTZURERl51totxpgdxpgtgB+4zxhz\nD1BrrX0eeJJg4igAPGytrQQwxjwLbA3t5n5rrX+4Y9WUShGR0WGMeQxYTvBc8IC1dlu7bTcBDwKt\nwDPW2sdDz98JfBPwAt+21r443HGGVwppRamIjCdjJimkO8AiIqPDWvutTk/tarftOeC5bt7zBPDE\nMIfWQaTMWEkhEZERY4xZDRRZa1cYY+YQ7DO3IrTNCTwOLAKqgJeNMesJTqn8DrCY4GrSh4BhTwp5\nNJJeRMahsZcU0kFcRES6oYEEIiKjYi2wHsBau98Yk2mMSbPW1gE5QI21tgLAGPMmcCXBpNAboXLk\neuBLIxGobjKLyHg0dpJCutgXEZFeeLVSSERkNOQDO9o9rgg9Vxf6OdUYU0RwEMEaYEPodUnGmBcI\n9p77rrX2zb4+KDMziZhz6AfkcDpwOCA/Lw2HIzpKyKJ1spziGphojCsaYwLFNVBDEdfYSQpp+piI\niPRCNw9ERKJCJNtirQ0YY+4mWFJWCxS3254N3AIUAm8bYwqttYHedlxd3TTooNzuVJpbPMS4nFRW\nNgx6P0PJ7U6loqK+7xeOMMU1MNEYVzTGBIproAYSV2/JozGXFFL5mIiIdMfrC36X0EohEZERVUJw\nZVBYAXA6/MBauxFYCWCMeZjgiqFEYIu11gscNsbUA26gfDgD9XgDOkeIyLgzdpJCugMsIiK9UK8I\nEZFR8RrBRtFPGGMWASWhXkEAGGNeBu4GGoEbgEeBeOApY8wjBMvHUoDK4Q7U4/PrHCEi486YSQp5\ndbEvIiK90Eh6EZGRZ63dYozZYYzZAviB+4wx9wC11trngScJJo4CwMPW2koAY8yzwNbQbu631vqH\nO1av10+sxtGLyDgzZpJCKh8TEZHenO09pwt+EZGRZK39VqendrXb9hzwXDfveQJ4YphD68Dr85MQ\nN/hG1SIi56Mxk0E5Wz6mA7mIiHTl8flxOhy4nGPm1CciIkPI41X5mIiMP2PmqKdeESIi0htd7IuI\nSG+8Pr8aTYvIuNOv8jFjzGPAcoK1vg9Ya7e123YT8CDQCjxjrX3cGHM58Dtgb+hlu6219w9l4J2F\nVwrFqA5YRES6EbzY1zlCRES6CgQCeHx+YnTzQETGmT6TQsaY1UCRtXaFMWYO8HNgRWibE3gcWARU\nAS8bY9aH3rrRWnvb8ITdlVYKiYhIbzxeXeyLiEj3fP4AgYD6k4rI+NOfo95aYD2AtXY/kGmMSQtt\nywFqrLUVoYkAbwJXDkukfTg7fUw9hUREpCuvz6+LfRER6ZZuMIvIeNWfo14+UNHucUXoufDPqcaY\nImNMLLAGyAttm2uMecEY844x5qohi7gHOpCLiEhvvOopJCIiPYhMqNTNAxEZZwYzkj7SkMFaGzDG\n3E2wpKwWKA5tPwg8BPwWmA68bYyZaa1t62mnmZlJxJzDKh9H6AA+IS+NhPjB/FrDw+1OHe0QuojG\nmEBxDVQ0xhWNMYHikiCPVgqJiEgPPF4foBvMIjL+9Cd7UsLZlUEABcDp8ANr7UZgJYAx5mHgqLX2\nFPCb0EsOG2NKgYkEk0bdqq5uGljk7bjdqTQ2BfNNNTWNUTNu2O1OpaKifrTD6CAaYwLFNVDRGFc0\nxgRjIy4lj4aGxxtQTyEREenW2ZVCGkggIuNLf66OXwNuAzDGLAJKrLWRbzLGmJeNMbnGmGTgBuAN\nY8ydxpi/DW3PJ1hSdmrIo2/H4/PjdDiiJiEkIiLRIxAIqKeQiIj0KNKKQucJERln+jzqWWu3ADuM\nMVuAHwL3GWPuMcbcEnrJkwQTR+8AD1trK4EXgNXGmE3AH4Cv9FY6NhQ86hUhIiI98PpCd4B1nhAR\nkW5EVgrpPCEi40y/mu9Ya7/V6ald7bY9BzzX6fX1BFcNjRg1EBURkZ60eXQHWEREetYW7imk84SI\njDNj5qinlUIiItIT3QEWEZHeaPqYiIxXY+aop6kyIiLSE90BFhGR3kR6CunmgYiMM9Ezu/0cebx+\nEqNoFL2IyHhhjHkMWA4EgAestdvabbsJeBBoBZ6x1j5ujLkc+B2wN/Sy3dba+4czRm/kYl9TZURE\npCuPJ3jzQCuFRGS8GTNZFK0UEhEZecaY1UCRtXaFMWYO8HNgRWibE3gcWARUAS8bY9aH3rrRWnvb\nSMXZprIAERHphcenlUIiMj6NmaOeGk2LiIyKtcB6AGvtfiDTGJMW2pYD1FhrK6y1fuBN4MrRCNIT\nLh/TeUJERLpxtqeQVpSKyPgyJlYK+Xx+fP6ALvZFREZePrCj3eOK0HN1oZ9TjTFFwFFgDbAh9PNc\nY8wLQBbwkLX29b4+KDMziZgY16CCrCiuAiAtNQG3O3VQ+xgO0RRLe4prYKIxrmiMCRTXQEVrXGNR\nZEqlvk+IyDgzJpJCagwnIhI1IrdYrbUBY8zdBEvKaoHi0PaDwEPAb4HpwNvGmJnW2rbedlxd3TTo\noDyhi31Pm5eKivpB72coud2pURNLe4prYKIxrmiMCRTXQA0kLiWPzp3Xq55CIjI+jYmkULhXhHoK\niYiMuBKCK4PCCoDT4QfW2o3ASgBjzMPAUWvtKeA3oZccNsaUAhMJJo2GRaRXhM4TIiLSDZ0nRGS8\nGhNHvXCviBitFBIRGWmvAbcBGGMWASXW2sitbWPMy8aYXGNMMnAD8IYx5k5jzN+GtucDecCp4Qyy\nzaPzhIiI9EyVByIyXo2Jo55HK4VEREaFtXYLsMMYswX4IXCfMeYeY8wtoZc8STBx9A7wsLW2EngB\nWG2M2QT8AfhKX6Vj50rnCRER6U24p5DKx0RkvBkb5WMeTZURERkt1tpvdXpqV7ttzwHPdXp9PcFV\nQyNG08dERKQ3qjwQkfFqbCSFtNxTRER6cXbUsM4TIiIjzRjzGLAcCAAPWGu3tdt2E/Ag0Ao8Y619\n3BhzOfA7YG/oZbuttfcPZ4zqKSQi49WYSAp5NEJSRER6oZsHIiKjwxizGiiy1q4wxswhOJFyRWib\nE3gcWARUAS8bY9aH3rrRWnvbSMUZ/j6hlUIiMt6MiaNem0ZIiohIL7RSSERk1KwF1gNYa/cDmcaY\ntNC2HKDGWlthrfUDbwJXjkaQZ3vPOUbj40VERs3YWCmkO8AiItILXeyLiIyafGBHu8cVoefqQj+n\nGmOKgKPAGmBD6Oe5xpgXgCzgIWvt6319UGZmEjExrkEFGT5P5OelkZ2eOKh9DAe3O3W0Q+iW4hqY\naIwrGmMCxTVQQxHXGEkKhRqI6g6wiIh042yj6cF9WRARkSETyc5bawPGmLsJlpTVAsWh7QeBh4Df\nAtOBt40xM/uaVFld3TTooMKVB7U1TfjbvIPez1Byu1OpqKgf7TC6UFwDE41xRWNMoLgGaiBx9ZY8\nGhNJoTb1FBIRkV5EysditFJIRGSElRBcGRRWAJwOP7DWbgRWAhhjHgaOWmtPAb8JveSwMaYUmEgw\naTQsVGYsIuPVmDjqadSwiIj0ps2jFaUiIqPkNeA2AGPMIqDEWhu5tW2MedkYk2uMSQZuAN4wxtxp\njPnb0PZ8IA84NZxBetWOQkTGqX6tFBroGMl22xKBPcD3rLVPDWHcHWiqjIiI9ObsSiGdJ0RERpK1\ndosxZocxZgvgB+4zxtwD1FprnweeJJg4CgAPW2srQ72Efh36nhEHfKWv0rFz1eb14QBcTq0oFZHx\npc+k0GDGSFprT4be/iBwZlgibydcPqblniIi0p2zjaZ1nhARGWnW2m91empXu23PAc91en09wVVD\nI8bj9RMT48ThUFJIRMaX/lwdD2qMpDFmNjAXeHHIo+5E5WMiItIbrRQSEZHeeLx+3TgQkXGpP0e+\nfILjIsPCYyTDP6caY4qMMbEEx0jmhbY9CvzNUAXaG90BFhGR3rRpSqWIiPQivFJIRGS8Gcz0sT7H\nSBpjPgu8a60tNsb0a6eZmUnEDHJUcLiBqNud0uuotdEQbfFAdMYEimugojGuaIwJFJe0u3mgC34R\nEemGx+sj1qXSMREZf/qTFBrwGEngFmC6MWYdMAloNcactNa+0dOHVFc3DTj4sPDFfmN9CxUV9X28\neuS43alRFQ9EZ0yguAYqGuOKxphgbMSl5NG5C58n1EBURES64/H6idONAxEZh/qTFHoNeAh4oqcx\nksDdQCPBhnCPWmufabf9u8DR3hJC50rTx0REpDcer49YNRAVEZEeeLx+khMGU0QhInJ+6/PIN5gx\nksMZcHfC5WOaPiYiIt3xeP06R4iISI/adJ4QkXGqX+nwgY6R7PTe7w4qsgHwaqWQiIj0os3j1zlC\nRER65PX61GhaRMalMXHka9NIehER6YXH51cDURER6ZbP78cf0IRKERmfxsSRr00j6UVEpBcej2/Q\nEy5FRGRs04RKERnPxsSRz+MJHsi15FNERLrj8WqlkIiIdM/rCwDqTyoi49OYOPK1eX24nA6cmioj\nIiLdaPOqp5CIiHRPK4VEZDwbE3MXPWogKiIyaowxjwHLCU6hfMBgvnnqAAAgAElEQVRau63dtpuA\nB4FW4Blr7ePttiUCe4DvWWufGq74AoFAsIGo7gCLiEg3PL5Q1YFWlIrIODQmrpA9Pp+SQiIio8AY\nsxoostauAO4FfthumxN4HLgeWAXcYIyZ1O7tDwJnhjtGnz8QbCCq84SIiHTD+//Yu+/wKuv7/+PP\nczIImQSSAGGF+WEJCKjgYKl1gXxVHHVXW62jtevb2t9XW621to7aWu2wde+9t1hwoOw9PqyEFSAJ\nkEnWGb8/zklIQsY5Mck5yXk9rovr4tzznZOT+32f92fcmp9URCJYl7jy6VHDIiIhcyrwJoC1diOQ\naoxJ9q9LAwqttfnWWg8wHzgNwBgzEhgNvNfeAbpqW4CVJ0RE5Gi1eULfJ0QkAnWN4WMutyr7IiKh\n0QdYXud1vn9Zsf//ScaY4UAOMBNY4N/uAeBm4KpAT5SaGt+qJ4gVlVYCkBAfS3p6UtD7t6dwi6eG\n4gpOOMYVjjGB4gpWuMbV1dQMH9P3CRGJRF2iKFRV7SG+W5f4UUREOrvaCRmstV5jzFXA40ARkA04\njDFXAl9ba7ONMQEf+NChw60K6FCJryjkdXvIzy9p1THaQ3p6UljFU0NxBScc4wrHmEBxBSuYuFQ8\n+nZqho+pR6mIRKIuUUmpcnl0ERcRCY1cfD2DamQCe2teWGsXAqcAGGPuwddj6DxgiDFmNtAfqDTG\n7LbWftoeAVZr+JiIiDSjtqeQho+JSATq9EUhr9frGz6mi7iISCh8DNwJ/MsYMxHItdbWNm0bYz7A\nN0SsDJgDPGCtfbHO+juAnPYqCIEeNSwiIs1zubyAGg9EJDJ1+qKQ2+PFq6fKiIiEhLV2kTFmuTFm\nEeABbjLGXA0UWWvfAP6Nr3DkBe6x1hZ0dIwaFiAiIs1RTyERiWSdvihUrUdIioiElLX21gaLVtdZ\n9zrwejP73tFOYdXSzb6IiDTHpR6lIhLBOv2VTzf7IiLSnCM9hRwtbCkiIpHoyNxzyhMiEnk6fSVF\nlX0REWmOGg9ERKQ51RpmLCIRrMsMH9NFXEREGuPSMGMRkZAyxjwITME3v9wt1tqlddbNBW4DKoEX\nrbUP11nXHVgH3GWtfbK94nOp8UBEIlinv/LpqTIiItIc9RQSEQkdY8x0YLi1dipwLfBQnXVO4GHg\nbGAaMMcY07/O7rcBB9s7xtqikBoPRCQCdforn272RUSkOepRKiISUqcCbwJYazcCqcaYZP+6NKDQ\nWptvrfUA84HTAIwxI4HRwHvtHaDyhIhEsoCGjwXb5dMYEw88CfQG4vB1+Xy3jWMH1FNIRESaV9MC\nHK08ISISCn2A5XVe5/uXFfv/n2SMGQ7kADOBBf7tHgBuBq4K9ESpqfFER0cFHWBMrO8rUXpaIunp\nSUHv357CLZ4aiis44RhXOMYEiitYbRFXi0Whul0+jTGjgMeBqf51NV0+JwIHgA+MMW8CJwHLrLX3\nGmMGAZ8A7VMUUndPERFpRrXmFBIRCSe1j/iy1nqNMVfh+35RBGQDDmPMlcDX1tpsY0zABz506HCr\nAiouqQSgtKSC/PySVh2jPaSnJ4VVPDUUV3DCMa5wjAkUV7CCiau54lEgPYXqdfk0xqQaY5KttcXU\n6fIJYIyZD5zWYCK4AcDugCJthdrunmoBFhGRRrjcXkB5QkQkRHLx9QyqkQnsrXlhrV0InAJgjLkH\nX4+h84AhxpjZQH+g0hiz21r7aXsEqB6lIhLJAikKtbbLJ8aYRfgu5LPbJtyj6akyIiLSnGqXG9Aw\nYxGREPkYuBP4lzFmIpBrra1t2jbGfIBviFgZMAd4wFr7Yp31dwA57VUQgro9Sh0tbCki0vW05pH0\nLXb5rLP+RGPMBOBZY8x4a623qYO2dgxw3M5CAHqmxoflOD/FFDjFFZxwjCscYwLFFemq/T2F1Hgg\nItLxrLWLjDHL/Y3FHuAmY8zVQJG19g3g3/gKR17gHmttQUfHeOSR9MF/FxER6ewCKQoF3eXTGDMJ\nyLPW7rLWrjLGRAPpQF5TJ2ntGOCD/v0qyqvCbpxfOI49DMeYQHEFKxzjCseYoGvEpeLRt+PSAwlE\nRELKWntrg0Wr66x7HXi9mX3vaKewatXMURqtnkIiEoECuUP+GJgH0FSXT2NMhjEmAV+Xz0+BacDP\n/et7A4lAu1T9jzx9TJV9ERE52pGbfRWFRETkaHokvYhEshZ7CrWmy6cx5p/AY8aYL4DuwE3WWk97\n/AB6qoyINPS3vz2ItRs5ePAAVVWV9OmTSXJyCn/4w33N7vf++++QkJDI9OkzOyhS6QjV6ikkInXU\nzREVFRVkZQ0iLi5BOSKCHRk+pjwhIpGXJwKaUyjYLp/W2nLg0m8dXQBqW4Cj1d1TRHx+9KOfAr4L\n8759u7jmmhsD2u/ss+e0Z1gSIi4NCxCROurmiO3bt3HnnbcHNJxXOaLrcrk8OBwQ5VSeEJHIyxOt\nmWg6rKinkIgEYsWKZbz44rMcPnyYm2/+KStXLmfBgvl4PB6mTj2Ja665jsce+xc9evRg8OChvP76\nyzgcTnbsyGbGjFO55prrQv0jSCtpmLGItEQ5IrJVuz3ERDlxOFQUEpHGdeU80emLQi7d7IuErZc/\n28rSTU3OL98qx43M4KJZw1q177ZtW3nhhdeJjY1l5crl/P3v/8HpdHLRRXO5+OL6nRs3bFjP88+/\nhsfj4cIL54T1hVyaVzssQD2FRMJOOOUJ5YjIVe3yEhOj7xIi4Uh5ov11+qJQtcYAi0iAhg0bTmxs\nLABxcXHcfPN1REVFUVhYSHFxcb1tjRlJXFxcKMKUNqaJpkUkEMoRkcvl9ui7hIi0qKvmic5fFNIE\noiJh66JZw1rdq6c9xMTEALBv315eeuk5Hn/8OeLj47niiouO2jYqSi2GXUVNj9Jo5QmRsBNOeUI5\nInKpKCQSvpQn2l+nv/ppTiERCVZhYSGpqanEx8dj7Sb27dtHdXV1qMOSdlLt9hAd5cCpuSJEJADK\nEZGn2uUhVkUhEQlQV8sTnf7qd+TpY53+RxGRDjJ8+Ai6d4/nhhuuYf78j5k793weeOBPoQ5L2km1\ny6N550QkYMoRkcfXU0h5QkQC09XyRKcfPuZSTyERacLZZ88hPT2J/PwSJk6czMSJkwFfd84///nh\nZvet2Rbgvffmt2ucnZ0x5kFgCuAFbrHWLq2zbi5wG1AJvGitfdgYEw88CfQG4oC7rLXvtld8LrdX\nwwJE5Ch1Hx2sHBHZqt0eNTCLyFEiJU90+qufJpoWEQkdY8x0YLi1dipwLfBQnXVO4GHgbGAaMMcY\n0x+YAyyz1k4HLgL+3J4xVrvcGhYgIiJN0vAxEYlknf7qVzOnULQeNSwiEgqnAm8CWGs3AqnGmGT/\nujSg0Fqbb631APOB06y1L1lr7/VvMwDY3Z4B+noKaViAiIgcze3x4PWqgVlEIlenHz5WU9l3aAJR\nEZFQ6AMsr/M637+s2P//JGPMcCAHmAksqNnQGLMI6A/MDuREqanxRLeiuONye4mJcZKenhT0vu0t\nHGMCxRWscIwrHGMCxRWscI2rK3G5vABqPBCRiNUlikKq7IuIhI3aCr211muMuQp4HCgCshusP9EY\nMwF41hgz3lrrbe7Ahw4dblVAVS43MdFO8vNLWrV/e6mZ7yrcKK7ghGNc4RgTKK5gBROXiketp6ko\nRCTSdfqrX7XbQ0yMKvsiIiGSi69nUI1MYG/NC2vtQmvtKdba2fgKQznGmEnGmAH+9avwNVCkt1eA\nLpeHWLUAi4hII2qmolBRSEQiVae/+rk0gaiISCh9DMwDMMZMBHKttbVN28aYD4wxGcaYBHwTTH+K\nb9Lpn/vX9wYSgYL2CM7j9eL26OljIiLSOJd6ColIhOv0V79qTSAqInVcf/332LRpY71l//znw7zw\nwrNHbbtixTJuu+2XANx668+OWv/aay/x2GP/avJcW7duYefOHQD89re/prKy4tuE3ilZaxcBy/3z\nAz0E3GSMudoYc55/k3/jKxx9CdxjrS0A/glkGGO+AN4DbvJPRN3mXDUPI9DNvoj4KU9IXUd6Cun7\nhIj4RFqe6BJzCsXG6GZfRHxOP/0MPvvsE0aOHFW7bMGCz/jb3/7Z7H5//GPwT0VfuPAzRo4czcCB\ng7jzznuC3r+rsNbe2mDR6jrrXgdeb7B9OXBpB4RWO1eEepSKSA3lCanLpTwhIg1EWp7oGkUhVfZF\nxO/UU7/DDTdcy403/hiAdevWkZ6eTk5ONrfd9itiYmJISkrid7/7Y739zjnnVN57bz7Lli3hoYce\noGfPXvTqlUZmZj9cLhd3330H+fl5lJeXc80119GnT1/eeut1Fi78jNTUVH7zm1/z9NMvUVpawj33\n/I7q6mqcTie33no7DoeDu+++g8zMfmzduoURIwwPPHBvY+FLG3OpBVhEGlCekLo00bSINBRpeaJT\nF4W8Xi8ut0fDAkTC1Otb32Vl3to2PeaxGcdw/rCmn2CemtqTzMx+bNiwjtGjx/LBBx9w+ulnUlJS\nwm9/+3syM/tx112/YfHir4mPjz9q/3/962Fuv/0uhg8fwS9+8WMyM/tRUlLM8cdP4ayzZrNnz25u\nv/1WHn/8WU44YSozZpzK6NFja/f/z3/+yezZczn11O/w3/9+yuOPP8q1116PtRu5884/kJrak/PO\nO5vi4mLqPIhL2okmEBUJb8oTyhOhpmHGIuFNeaL980Snvvqpu6eINOb0089k/vxPAPjss8+YMeNU\nevTowZ/+9Htuvvk6Vq5cTnFxUaP77t27l+HDRwAwYcJEAJKSktm4cT033HANd999R5P7Ali7kWOP\nnQTAxImT2bLFAtCv3wB69UrD6XSSlpZOSUn4Pf64K1ILsIg0RnlCahwZZqwepSJyRCTliYB6Chlj\nHgSmAF7gFmvt0jrr5gK3AZXAi9bah/3L7wVO8Z/jHv+8Em2qpgU4Vo+kFwlL5w+b3WwVvr1Mnz6T\np59+nNNPP4OsrCySk5O55567uO++v5CVNZg///lPTe7rdB4pHni9XgA++eRDiouLeeSR/1BcXMz3\nv39FM2d31O5XXe3C4fAdLyqq/nWqZhtpX8oTIuFNeUJ5ItRcLt/7rMYDkfCkPNH+eaLFq58xZjow\n3Fo7FbgW39NlatY5gYeBs/E9YniOMaa/MWYmMNa/z5nAX9ok2gaq3bqIi8jR4uMTGDp0OE8//QRz\n5swBoKyslN69+1BSUsKKFcuprq5udN+0tHR27szB6/WycuVyAAoLC+nbNxOn08nChZ/V7utwOHC7\n3fX2HzVqNCtWLANg1arl9Saok47nqskTUcoTInKE8kTHMsY8aIz52hizyBhzXIN1c40xS40xXxpj\nbvYvizfGvGyMWWiMWWyMabdvhHokvYg0JpLyRCBXv1OBNwGstRuBVGNMsn9dGlBorc33P054PnAa\n8DlwoX+bQiDBGNPmzbRut1qARaRxp59+JkuXLmbWrFkAnH/+hdxww7Xce+/dXHbZlTz77JMcOFBw\n1H7XXXcjt932K371q5+SkdEbgBkzZrFo0RfccssNdO/enYyMDJ544t+MH38sf/nLfSxbtqR2/+9/\n/4d8+OH7/PjHP+T999/l2muv75gfWBpVe7Ovp1SKSAPKEx2jNQ3MwBxgmbV2OnAREPwjfQJ0pCik\n7xMiUl+k5AlHS12OjDGPAu9Za9/yv/4CuNZau9kY4wCygdOBHOBtYIG19k919r8OOMVa21z/KFwu\ntzc6yIux1+vlsbfXc+K4vowe3CuofUVEOhHNNArk55cE3Ue2strNy59t5cLTDXFhVhdKT08iPz/8\n5gxRXMEJx7jCMSZQXMEKJq709KSwzRPGmN8BO621//G/3gQcb60tNsZkAPOttcf41/0SyLPWPlln\n/5OB31lrZ7V0rtbkiaLSSt76Mpurzx2Lp8oV7O7tqit8NjuS4gpcOMYEiitYbZUnWvP0sdqDWWu9\nxpirgMeBInwFotr1/vmGrgW+09JBDx063IpQYO6Jg7rEL6mjhGNMoLiCFY5xhWNM0DXiSk9Paudo\nuq5uMVFccYYJ28+BiEgE6AMsr/M637+s2P//JGPMcHwNzDOBBTUbGmMWAf2BgIaPpabGE2wjc3p6\nEj8fnBbUPh0pXO8BFFdwwjGucIwJFFew2iKuQIpCufgu3DUygb01L6y1C/FNKI0x5h58F3SMMWcA\n/wecaa1temptERERERGJFAE3MFtrTzTGTACeNcaMt9Y22xOotY3MEJ6NSOEYEyiuYIVjXOEYEyiu\nYLVVI3Mgnek/BuYBGGMmArnW2tozG2M+MMZkGGMS8I3//dQYkwLcB8y21h4MKEoREREREelqWmxg\nttaeYq2dja8wlGOMmWSMGeBfvwpfQ3Z6B8YsIhIxWiwKWWsXAcv93TcfAm4yxlxtjDnPv8m/8RWO\nvsT36PkC4GJ8k1C/bIxZ4P83sH1+BBERERERCVNBNzDjm3T65/71vYFE4OjZXEVE5FsLaE4ha+2t\nDRatrrPudeD1Bts/Cjz6raMTEREREZFOy1q7yBhT08Dswd/ADBRZa9/gSAOzF38DszHmn8Bj/gfc\ndAdu8j/pWERE2lhrJpoWEREREREJSCsamMuBSzsgNBGRiBdmD+gVEREREREREZGOoKKQiIiIiIiI\niEgEUlFIRERERERERCQCqSgkIiIiIiIiIhKBHF6vN9QxiIiIiIiIiIhIB1NPIRERERERERGRCKSi\nkIiIiIiIiIhIBFJRSEREREREREQkAqkoJCIiIiIiIiISgVQUEhERERERERGJQCoKiYiIiIiIiIhE\noOhQB/BtGGMeBKYAXuAWa+3SEMczFngLeNBa+7AxZgDwDBAF7AWusNZWdnBM9wKn4Ptd3wMsDYOY\n4oEngd5AHHAXsDrUcflj6w6s88c0P9QxGWNmAK8A6/2L1gL3hjouf2yXAb8EXMBvgDWhjssYcy1w\nRZ1Fk4GTgH/gu06ssdbe0JEx+eNKBJ4GUoFuwJ3AvlDHFQmUJwKKSXkiuNiUJwKPTXki8LiUJ0Ik\nnPJEOOYIf1zKE8HFpjwReGzKE4HF1K45otP2FDLGTAeGW2unAtcCD4U4ngTgb/j+8Gv8DnjEWnsK\nsBW4poNjmgmM9b9HZwJ/CXVMfnOAZdba6cBFwJ/DJC6A24CD/v+HS0wLrbUz/P9+FA5xGWN6Ab8F\nTgZmA3PDIS5r7WM175U/vqfwfe5vsdaeBKQYY87q6LiAq33h2ZnAPOCvYRJXl6Y8EVBMyhPBU54I\ngPJE0K5GeaLDhVOeCMcc4Y9LeSJ4yhMBUJ4IytW0Y47otEUh4FTgTQBr7UYg1RiTHMJ4KoGzgdw6\ny2YAb/v//w5wWgfH9Dlwof//hUBCGMSEtfYla+29/pcDgN3hEJcxZiQwGnjPvyjkMTVhBqGP6zTg\nU2ttibV2r7X2ujCJq67fAH8CBtdp9QtVXAVAL///U/HdKIRDXF2d8kTLlCeCoDwRFOWJ4ChPhEY4\n5YlwzBGgPBEU5YmgKE8Erl1zRGcePtYHWF7ndb5/WXEogrHWugCXMabu4oQ63d3ygL4dHJMbKPO/\nvBZ4HzgjlDHVZYxZBPTHVxn+NAziegC4GbjK/zqkv786Rhtj3gZ64usqGA5xZQHx/rhSgTvCJC4A\njDHHAbvwdUU9VGdVSOKy1r5ojLnaGLMV3/s1B3gk1HFFAOWJlmNSngiO8kTgslCeCJjyRMiETZ4I\nxxzhj0t5IjjKE4HLQnkiIO2dIzpzT6GGHKEOoAUhi88YMxffRfzmBqtC+p5Za08EzgWebRBLh8dl\njLkS+Npam93EJqF6r7bgu3DPxZdcHqN+MTdUcTnwVavPx9ed8QlC/Dts4Pv4xpk3FJK4jDGXAzut\ntcOAWfg+83WF+v2KFOH+PitPNKA8ERDlidZRnpDGhPP7HNLYlCdapjwRNOWJALV3jujMRaFcfJX8\nGpn4JqMKJ6X+icYA+lG/O2iHMMacAfwfcJa1tihMYprknzgPa+0qfBelkhDHdQ4w1xjzDb4LwO2E\nwXtlrd3j7x7rtdZuwzehWGqo4wL2A4ustS5/XCWE/ndY1wxgEb4Wv151locqrpOAjwCstauB7kBa\nGMTV1SlPBEB5ImDKE8FRngiO8kRohHueCPk1BpQngqA8ERzlicC1a47ozEWhj/FNsoQxZiKQa60t\nCW1IR/kUuMD//wuADzvy5MaYFOA+YLa1tmays5DG5DcN+DmAMaY3kBjquKy1F1trj7PWTgH+g+9p\nASF/r4wxlxljfuH/fx98T1h4ItRx4fv7m2WMcfoniQv577CGMSYTKLXWVllrq4FNxpiT/avPD1Fc\nW4ET/PENwpf0NoZBXF2d8kQLlCcCpzwRNOWJ4ChPhEa454mQ/80oTwROeSJoyhOBa9cc4fB6vd86\nwlAxxvwR3wXBA9zkr5qFKpZJ+MaQZgHVwB7gMnxdzuKAHcD3/B+qjorpOnxjMzfXWXwVvotUSGLy\nx9UdX7fFAfiqnHcCy/A9Zi9kcdWJ7w4gB181NqQxGWOSgOeBHkAsvvdqZajj8sd2Pb5uxAC/x/d4\n0nCIaxLwe2vtWf7Xo4F/4SuCL7bW/iwEMSUCj+NLwtH4Wo72hTquSKA80WJMyhOti+8OlCcCiU15\nIvCYlCdCJFzyRDjmCH9cyhOti+8OlCcCiU15IrB42jVHdOqikIiIiIiIiIiItE5nHj4mIiIiIiIi\nIiKtpKKQiIiIiIiIiEgEUlFIRERERERERCQCqSgkIiIiIiIiIhKBVBQSEREREREREYlAKgqJiIiI\niIiIiEQgFYVERERERERERCKQikIiIiIiIiIiIhFIRSERERERERERkQikopCIiIiIiIiISARSUUhE\nREREREREJAKpKCQiIiIiIiIiEoFUFBIRERERERERiUAqComIiIiIiIiIRCAVhUREREREREREIpCK\nQiIiIiIiIiIiESg61AFI12eM8QLbADeQAKwC7rbWfh3CmLKALf64LgVKgX8CmUA1cL+19in/trOA\n+4FEYAfwPWvtbv+6nwDX4yuwfgHcaK2tauHc5wB3A3HAAeCn1tolzR3PGBML/B2Yhu99/Ie19iFj\nzF3AVcB2a+2Mb//OiIh0DpGWW4AzgPsAY611dNxPJSLSeXSG3GCtXWGM+Q7wLPCQtfb3dbYdD/wD\nSAMKgB9aa9f4110C3AbEAOuAa6y1RcYYB3APcB7gBd6w1v7aGPND4H+BKGttVof8sNIpqaeQdJQZ\n1loDDACeAt4yxkwLcUx7rLUjrbUrgCeAV621I4HvAA8YY0YYYxKAF4HvW2tHAO/gu8HHGDMFuAWY\nCowEegA/bu6ExpgewPPAlf5z3QW8FsDxfgb09C8/AfiJMWaytfZ24Mq2eDNERDqhiMkt1tp3/McR\nEZHmhXVuMMZcCvwWWNHIdi8C9/pzwx+B5wCMMQOBvwFn+3+2HHyNzAAXAzOAcf5/M4wx86y1/wRO\nbb8fSboK9RSSDmWt9QKvGGNS8F3oTjTGdMPX+nkmEAs8aq39A4AxJgdf5ftafBf25621PzfGROO7\ngT4FiALWAFdba4uNMXOB3+NrHdiKryJf0EJoxwDz/THuNcZsBkYDBl8vnJqL9uPA/caYJOBC4CVr\nbaE/1sfxXeDvb+Y8Q4DDNRV/4DOgv79Y1NzxLgT+z1rrAYqNMa/6ly1r4ecSEenylFtERKShMM4N\nm4CZwKN1FxpjjgF6WGvf9Mf/tjHm38aYUcBpwHxr7U7/5o8B/wVuxpc3nrTWVvqP84x/2atBvmUS\nodRTSELlbeAEY0x34Jf4bpKPAcYA84wxs+tsOw1fi+kk4EfGmP74utEPxteKOhxYD0w1xgwBngG+\na60dgu9i+c8A4pkPXGKMcRpjDJAFfAOMwNfVEwBrbSm+IV/DGq7z/7+lVtyNgNs/bABgHrDMf/Pf\n3PFacy4RkUgTqblFRESaFla5wVq7oonpJkYA2xss2+4/b2O5IcMYk9rEOuUNCZiKQhIqxfg+f0nA\nHODv1tpKa20Z8DRwfp1tn7fWuq21ucB+fJX7fHwX9POAeGvt7dbaj/BV/RdYa9f59/0ncK4xJqqF\neH6Cr1WgANgA/N5auw+IByoabFuOrzWg4bqa5U2y1pYD1wHvGWMO4psn6Ef+1c0dL+hziYhEoIjM\nLSIi0qxwyw1NCTg3+HsFeRtbh/KGBEnDxyRUsvBNulmIb76EB40xf/Cv6wYsqbNtUZ3/u/FNlva1\nMeZH+AoqTxlj3sE3EWcPYJoxZlOD/XsBec3E8zrwG2vtk/4Wgc+NMSuBMnwTQtcVj2/y0IbrapY3\nyRiTia+75/HW2rXGmBnAG8aY4S0cL+hziYhEoCwiMLeIiEizsgiv3NCUgHODMSYOcDS2DuUNCZKK\nQhIq8/BV1quMMbn4nsjybjAHsNa+CrxqjOmJbz6G/8U3s/+n1tp5gR7HGJMGTMQ/kZu1drcxZhFw\nMr7uoRfX2TYFSPWfZxO+rv41huNrCW7OifjmkVjrP9cCY4wbGNXC8WrWbQniXCIikSZSc4uIiDQt\nbHJDCzYBQ2te+J8qNgxfDsgEptfZdjiw11pb6C9KDQM+qbNOeUMCpuFj0qGMMQ5jzDx8Xer/n3/x\nW8D3jTFR/vW3GWPObOE43zPG3A5grT2I7yLqBT4CTvGP8cUYc7wx5q8thHUQX7fQOf59UvEVb9bh\nGxs8yBhzsn/bnwLv+rubvgx81xjT2z8B3S3ACy2cazMwxv9oSowxE4EUfGN/mzvey/jGNUcZY/oC\nlwAvtXAuEZGIoNwiIiINhWluaJK1dgOQb3xPJwO4Cthhrd3sj/tU//x04Hsycd3vCdcZYxKMMYn4\npqpQ3pCAqaeQdJQFxhgXvgLIBuAca23Nk7Mewdetcz2+bpDLgL+0cLy3gMeNMVsAF75K/dXW2oPG\nmB/gG5IVC5TgSwRNstZ6/AnjfmPMPf4YnrTWfgBgjLkEeA/GnR8AACAASURBVMT4HiG8Fbjav98y\nY8z9wBf+fT4B/uHf5zxgjrX2mgbnWmOMuRX4wBjjBCqBy/0J5mBTxwP+im/COOv/eX9nrV3dwnsk\nItLVRVRuERGRgIRtboDap0qeCPQFqowxlwMPW2sfBi4F/m2MuRPfnEaXAVhr9xhjbgTe9DcYrMA/\nL6m19lVjzCRgFb5i1fPW2ndaikOkhsPr9YY6BpEO5++ps8Bam9VOx48BHrPWXtkex29wrhnAHdba\nGe19LhERaVp75xb/ObzWWkd7HV9ERNpWR+SGcDy3dB4aPibSPgbie7KYiIiIiIiISFhSUUgiWT9j\nzCb/vD5tylq7zVr7TVsftyFjzF34HqUpIiLhoV1yizFmToMn3IiISOfRbt87mmKM+SEwv6POJ52X\nho+JiIiIiIiIiEQg9RQSEREREREREYlAKgqJiIiIiIiIiESgsHkkfX5+SavHsaWmxnPo0OG2DKdN\nhGNc4RgTKK5ghWNc4RgTdI240tOT9KQhul6eCMeYQHEFKxzjCseYQHEFS3kieMoTHUNxBScc4wrH\nmEBxBaut8kSX6CkUHR0V6hAaFY5xhWNMoLiCFY5xhWNMoLjEJxzf73CMCRRXsMIxrnCMCRRXsMI1\nrq4qHN/vcIwJFFewwjGucIwJFFew2iquLlEUEhERERERERGR4KgoJCIiIiIiIiISgVQUEhERERER\nERGJQCoKiYiIiIiIiIhEIBWFREREREREREQiUECPpDfGjAXeAh601j7cYN1pwB8AN/C+tfYu//IH\ngSmAF7jFWru0LQMXEZHwoTwhIiLNUZ4QEQlPLRaFjDEJwN+A+U1s8hBwBrAHWGiMeQ1IB4Zba6ca\nY0YBjwNT2yZkEREJJ8oTIiLSHOUJEZHwFUhPoUrgbOBXDVcYY4YAB621u/yv3wdOxXcRfxPAWrvR\nGJNqjEm21ha3WeQiImGuqtpNYWklhaVVpCTE0rtnfKhDai/KE51EyeEqdu4vpdrtaXKblPwyiorK\n2+R8/dISSO/RPej9DpVUsmN/SbvF9W0M7pNESmK3oPdzezyUHq6motpNZZWbymo3LrcXj8eL2+PF\n4/W2WYyheq+cDhgxoAdxsQF1RO/yKqvdlJRVcbjSRXmli4oqN839lgdmJNIzOa7D4utgyhMh5vF6\nKSuvpqisiuKyKqpcTeeBlgRzjUmMi2FY/5RWn6sjFB+u4mBxBdUuT+2/trsid9w1uX9aAmmtyLnh\npKLKxYGiCkrLqyktr272fqUjhcs9yNDMZJLiY9v8uC1mbWutC3AZYxpb3QfIr/M6DxgKpAHL6yzP\n92/b5EU8NTWe6OioAEJuXHp6Uqv3bU/hGFc4xgRdKy674yAx0VEM6Rd8Esw7eJg9+aX06ZVAemp3\noqMan/orHN+vcIwJmo7rQFE5PZLiiHI62uxcFZUuXvzE8vHinZQcrqpd3iOxG0/fcQYOx5Fzhev7\nFayOyhMSPI/Xy5KN+1m3/SDb9hSx/1DH3tD0SIzlgZtOqve5b86OfSV8tHQnSzfm4fa05S1520mI\ni+bnl0wgq09yi9tWVrlZl32AFZsLWLOtgLIKVwdEGFoTR6Rz8/nHBLStx+tl+55iVmzJZ+32A0RH\nORmQnkj/jEQG9U5k+IAeOAP47FRWudldUMr+g4fZd7CcA0UVzDy2X9BfRKuq3WzZU0RstJMBGYmt\nLm55vV4+W7GHV/67Nagv3ll9kvjN1ce16pzhTnkidDxeL89/spmFq3JDdl39vysnMTQzsL9Hl9vD\nnvwytucWUVRWxfQJ/UhNCr4QH8h51m47wBdr9rJm24E2LcyHSq/kOO678cRmtzlQVMHbX2WzcksB\nHo8Xp9NBlNNBWkocZ08ZxIThaQHn7MZUVbuJinIQ5Qx86uJql5tvNuxn/vLd7Nxf2upzR4JJI9K5\nKcAcG4y2bspp6hPU4ifr0KHDrT5penoS+fklLW/YwcIxrnCMCbpOXB6Plze/zObdRTkATBnTm3nT\nhwbc8ldWUc1t/15MUZmvmOB0OOiV0o3k+FjiYqOI6xZNcnwsl589GofbHfTP01oer5fKKjfduzV+\nyVi5JZ/XP9+Ox+MlJtpJbHQU/TMSuXjWMLrFHF3s9Xq9lFW4OFhcwYGiCqpcHo4dnkZsI9s2F9O6\n7Qf4Zv1+Sg5XUV7lpqLKTUyUkwtmDGHs4F5A47/DwxXVvLJgGwtX5TImK5UfXTCu2XNXuzws2bif\nwtJKBmQkMrB3EikJsUclzVVbC3ju480cKK4gJSGWMVmp9EjsRo+kbowY0IOCgiOJLpjPVlcpHvm1\nOk90xcaD9owpZ28xj7yyik07DgEQHxfNsSPSGTEwlfi4mHY7b43PV+1m2+4iorrF0Cul+ZbL3IJS\nHn55NWu3FQAwoHcS04/tR8y3+H23h+KySt5YsJX7X1zFb66dwpghR64zDX22bBePvLKqtiiQlhLH\nBJNBfLdo4rpFExcbRUyUE2eUA6fD4S9Ot12BOhS+WL2HFZvzyc4r4/gxfRrdpua9+njxDp75YCOF\nJZUAxMZE4fF42bHvyHVx1uQB/OSSY5v9grIh+wD3PLmUwtLKess37y7kkf+dRUL3+p91t9vDtj1F\nR74AemH51hyWbtjPqi35VFX7cqvDAZlpvsLURaeNYEDvwP5WSw5X8dBLK/lm3T6S4mM5eUJv4uOi\nSYiLoXu36GZ/ltFDeh71WQrH61YHUJ6o49vE5PV6efSNtXy2Yg+9e8YzpF9K7X1JR/To27ankM9X\n7qG4wt3iz5Gzt5jH3l7HhuyDtX+HAMs35/OHG08O+F66pfNUVLl4a+E23v0qu/b6M6x/CmOGpBEb\n4yQ2xndt/jaFkVBYsGIX2bnFRMfFkJp09HtVWFLJK/M38/6iHFxuD2k9upPYPQa3x4vb7WH73mL+\n9vpahvZP4bIzRjJ5VO+A34PSw1Us2bCPRWv2ssLmER8XzQ0XjOekcZnN7ldUWslbn2/jo292UFxW\nhdPpYNywNHomx5GcEEtSQiyxYXYfEGqTRmW0S574tleDXHwV+xr9/MuqGizPBPZ+y3NJJ1RZ7ebT\nZbv4ev1+Tj6mL6dN7t9kz5fOrvhwFY++vZ4NOYdIS4kjIS6Gb9bvZ4XN54zjB3L2lEF0i23+wvbi\n/C0UlVUx2aQTEx1FfmE5eYXl5Owrqde6k3vgMP97yQScDXq4fLh4J+uyD3D8qN4cNzKjySJOc0oO\nV7Fu+0HWbD9AbkEZxYerKD1cjdvj5biRGVw/d0y9ltuisiqeeH8T5ZUuuneLptrlocrlZuueIvbk\nl3LLvPHExx2JY/XWAp7+yHKopP4N/NB+ydwybzyJ3Zv/slpe6eLLtXuZv3w3eXV6PUQ5HcTFRlFR\n5ebPL63mtEn9mTdjaL19vV4vy20+z32ymaKyKuJio1ifc4iHXlvDjy4Yd1QB63CFiwWr9vDJ0l21\nhboayfEx9EyOI6F7DAlx0ZSVV7M+5xBRTgfnTB3E7BOzGi2IRaA2yxNdrfGgvWKqrHbzzlc5fLRk\nJ26Pl8kjMzj3xCwy0xMC6nXRVnEdOFTGtt1FrLH7a4u0jfF4vPzxmeVk7y1mzOCenHHcAMYM7nnU\nzWi4/A7Tk7vx73c28JtHF/GjC8Yx47hBjcb18dfZVLk8zD5xEBNHpDOod1KHfckI1Xs1pE8idzy+\nhL+/uprMHnFH5byauFxuD4++uRaAaeP7MmF4OqMHpRIV5WD/wXJ255fy4eKdfLZsF3HRzqOu5TW+\nWruXpz7chMcDM4/tR/+MRPqkdmddzkE++GYnf39lFVefNbJ2+2qXLz/YXYWNHi8zLYFxQ3rh9njZ\nub+EnXklLFhRyuL1+7jpvLGMzurZ6H4ut4cDxRXs2l/KC/O3cKikkpEDe/CDOWOC7uFQ9/cWQY0H\nyhNN+LYxvbZwG+99vYP+6Qn88tKJLd5jtXVcaYkxfL5yD9m7C8kf0vjfj9vj4cPFO3nzi2zcHi/9\n0xMY2i+FIX2T2Z1fxifLdnHrw1/wy0snkpLQ/LCZ5uLyer0s3rCfVxZs41BJJQlx0Zw2qT8nj+vL\nwACLvq3VEZ+t/IOlZOcWs3rTfsY0uFblF5ZzxxNLKK90k5YSx9yTBzNnxnAOHjjSWJlbUMbbX2Wz\ndGMev3tsMRNHpHPj/4w96ruGx+tlQ85Bdu0vZXd+GbkFZezOL639ntK3VzwFRRX88amlnDC6N5ed\nPuKoz53X62Xxxv08/8kWSsurSYiL5uwpg5g1sR9maHrY/R1CeF0f2iNPfKuikLU2xxiTbIzJAnYD\ns4HL8HX3vBP4lzFmIpBrrQ2PdzGCHK5w8fnqXKZPyGxVceDb8Hi8LFq3jze+2F775f/l/27ly7V7\nufz0EYwclNo25/F6cbt9vVO+jQNFFbz++TbKKlzMmtifY4Yc+UJSVFbFh4t3sGjdPs48YSBnnTDo\nqP235Rbx9zfWcaikkgnD0rh29ii6d4tm0dp9vPb5Nt5ZlMOidfu44gzDuKGNfzlau/0AX63dx6De\nSVw/d0y9bpderxeX20N5lZvnP9nMko15zF+xm9MnD6i3/8v/3QrAhpxDPP/JZiaZdI4Z0oteKXH0\nSo6jR2K3oy7uNZZtyuOjJTvZnltcO466W2wUKQmxpPWNo7TcxdJNeQzsncg5U7Nq93vuY0tpeTU/\n+J+xTB2ZAfhukP/z7gaWbMzjvhdW8rOLx9O9WzSvLtjGx0t3ER3lZPxQf1wpcWTvLWHZpjz+8Mxy\nfnbR+Hrjob1eL7vzy1iXfYD12QfZvKsIl9tDdJSTk4/py6xJ/eiXllj7Gdixr4RH31nPp8t3s2HH\nIS4/cxQ79xax7+Bhdu4vYXtuMdFRTs6fNoTTJw/gX2+vZ9XWAh56dQ0/njeO2Ggn23OLWbR+H1+v\n20dFlZvu3aI464SBDO2Xwq68UnbuL2FXXil7CsqorjM0YET/FK44w9AvPbGpj1rEUZ4IXlFZFX97\nbQ1ut5czTxjIcSMzmvy7bWhPfimPvLGOfQcP0ys5jivOGMG4oWntHHHj+vv/DvbklzVbFPrvyj1k\n7y3mhNG9uf7cMR0VXqsdP6o3sTFR/P2Ndfz1ldX0TI2nb0r9Vlmv18uO/aVk9OjO+dMaL2h0Rf3S\nEjjzhIG89/UO3l6UzYUzhjW6XfbeYiqr3Myc2I8rvlN/OFFmWgKZaQmMGpTKH55dwfvf7CAlMbZe\nvvN4vLy6cBsfLt5JfLdobjhvbL0vQcMH9GDttoN8vjqX40ZlMCarJx6vl/+8uxG7q5BRg1LJ6nvk\n5nhAnxSG9Ekko8FcHB6vl8Xr9/PEBxt58OXVXHmm4ZRxmXi9XjbvKuSLNXvZvKuQA8UV1HQ8cjoc\nnDdtCOdMGRTw322kU55onWqXh+goR5PF5g++2cF7X+8gI7U7P794QpsVhIJRcz9X0MRcLPsOHuY/\n725ge24xKQmxXHXWSCYMO5KzvF4vTid8tGQX972wkl9+91iSmygMLbf5JOYWk5HUrV4xtqConI07\nDvH5qly25RYTHeXg7CmDOGfqoA7/ftSeanNuXulRRaHVWwsor3Rz5vEDOX/6EKKjnEdNnZCZlsAP\n545l9omlPPORZcXmfN78Mpvzpw2p3cbj9fLEexv5at2+2mXRUU4G9k7i2OFpTDLp9O2VwN4DZTz2\n3kYWb9jPph2HmDWpP/3TEshMTyDK6eDZjzezZtsBYqOdXDRzGLMm9gtqtIC0vUCePjYJeADIAqqN\nMfOAt4Fsa+0bwA3AC/7NX7LWbgY2G2OWG2MWAR7gpvYIXprm9Xp54v2NLN+cT0WVi/85ZUjLO7WR\nksNV/Pml1ezYX0JMtJNzpg5i2vhMPvhmBwtX5XLvCys5flQGc08eTN9eCc0ea++BMv76yhoy0xKY\nOrYP44f2IjYmioPFFXyxZi9frsnlYEkl/dMTGdovhaGZyfROjfcNYYpxEhXlJO/gYXb6v8gfKKpg\nWP8UJpkMhmQm43Z7+WjJTt5dlFPbxX/NtgP0T0/kjOMHcKB0B+9/lV277pX/bqO62sO5Jw+ujfGb\n9ft4/P1NuD0eLpg+hLOmDKptjT95XF8mj0zn3UU7+GjJTv7yympOGN2b7546vF5SK6908dSHm4hy\nOvje2SOPGofrcDiIiY4iJjqKS08fwcYdhby+cDvHDk8jLaU7h0oq+c+7G4iOcnDjecewa38JX63d\nx9fr9/P1+v21x4lyOjh2eBrnnjSY/hm+5FFWUc1zH2/mmw37cTocDO+fwrhhaYwb2ot+aQm1NxvF\nh6u484mlvP75doZkpjBqUCpLN+WxzOYzvH8Ks08awgF/i0N0lJPr5owhLjaKz1fv5U/PryQ22knO\nvhL69Iznh3PH1GuV8Xi9vJoSx4eLd3L3M8u55pxRHCiuYNOOQ2zaWUhxnV46AzISOW5kBtMnZDY6\n0dqgPkn89urjeGXBNuYv3829zy478j4CY7JSufw7pnbS5xvPG8s/3lzHyi0F/PG5FZRXuMgr9N28\npCTGMufELKZP6Ffb22niiPR656uqdlNW4aLa5Sa9R/dO1924LShPtJ2DxRXc9+Iq9h/0tXb/6+31\nvPHFds6eMogTx/ZptqflNxv28eQHm6iq9nDa5P5cMG1oi70T21O/dN/1fU9+WZPbHCqp5PXPtxHf\nLZpLZjVeQAhHE4al8eN5x/Dnl1bz7pfZ/OCcUfXWHyyupLS8us0aQDqT2SdmsXjDfj5esoupY/rU\nflGpa0OOb0jj6Gben6T4WH520Xj+8MxyXvx0C3ExUXSLjcLuLGRDzkH2Hyqnd894bpk3jj4NJvGP\njnJy7TmjuOupZTz5/iZ+d+3xvPVlNks35TGifwo/uXBcvaGJTbW0Oh0Opo7tQ8/kbjz8+lqeeH8T\n67MPkrO3pDZPJCfEMrxfCuk9upPeozvjhvUKaL6pSKM80bbyCsu555nlDOqTxI8uOOao+8Yv1uTy\nyoJt9Ezuxi8umdCqyfHbQs/kbjgdDgoKKxpd/+DLq8gvrGDK6N5c2kiPEofDwUUzh+H2ePl02W7u\nf3Elv7jk6MLQ/OW7ee6TzbWv01LiGJCRyJ78stq/VYDJIzO4cMbQVj0AIdzVNEjuyj96Tp6aYbkn\njevb4oiN/umJ/OiCcdz11FLeXZRDVp8kJo5Ix+v18tL8rXy1bh+D+yZxztSs2odJNCyA9+2VwK8v\nn1jbA+yNz7cfdZ5Rg1K56kxDRmqXfQhLp+LwhsmkWvn5Ja0OJJy6c9UVyri+WJPLE+9vAiAjtTv3\nXDcFh8NRL6aCwnK25hbVPvnE64WRg1KPailrTHmli8pqNz0aJBmPx8uDL69ifc4hjh+VwUUzh9Ub\nA5y9t5hnP7Zk7y3BAUwamcHsqYOYNDaz0ffqb6+tYeWWgtrX3btFMSA9kS17ivB6fT1Z+qUlsCuv\ntF6PjaY4oLYXTGpSN6KjHOQXVpCcEMtFM4fSLy2RD5fsZMnG/bWtfqlJ3Zg9dRAjB6Xy4MurKSiq\nYPaJWfzPKYN564ts3lmUQ/duUdwwdyxjhzTdGr4rr5QnP9hI9t4SEuKiOXFsX44fncGQvsk88/Fm\nFqzcw5wTszhvWssFvLU7CnnwhRUcM6QXt8wbx/0vrmTTzkK+e9rw2tZUr9fLtj3F7MzzFcMOFFfU\ndvMEmGTSGT80jdc/30ZhaRVDM5O5dvboo26u69q6p4g/PbfCP9Hqsdz/4koqqtz87prjGWt6H/U7\n9Hq9vPTZVj5euguAk8b24bLvjGhyHPsnS3fx4vwt9Z74kJIYy6hBqYwd3JMxWT2DurHZtOMQuYXl\nJMZG0adnPL17xjc6pMvl9tQWhmJjnEwakc7UMX0YlZUa1ER5wQiyu2fkVZoa0dXyRMOY8g4d5r4X\nVnGguIKzpgxk2vhMPly8k6/W7sXl9jKsXwo/uuCYo4qhLreHl+ZvZf6K3cTFRnHN2aOY7O+11xZx\ntZbb4+GGBxYyICOR269qfPLcf7y5jqWb8rjyDMOMY/t1SFxt6eePfAXA/TeeWK8gvGJzPg+/vpYL\npg+p17Oyo4T6vVq9tYC/vrqG4f1T+NVlE2sbSmri+uOzy9myu4iHfnIKCS3Mb7VjXwl/en4FFVVH\n5hjpFhPFhOFpXP6dEc3uf2ToTCK780vp2yueX18+6agvnoG8X/sOHuYvL68mr7Cc2Ggnk0dmcMq4\nvowY0KPdGgOUJ4LXmfOE2+Mhe28Jg/sm1d57NBZTtcvN3c8sr52Q96wTBnLhzCNF9c27CrnvhZXE\nxUbx/66Y1GIDbGsE81798h+LqHZ7ePDmk+stLy2v5sd//YJjhvTipxeNb/YYXq+X5z/ZwvwVu+nb\nK55fXHJsbW+glZvzefiNtSTFx3LutCGs21LAlt2FlFW46N4tCjMglVGDUhkzuCeZaW3/XgSiIz5b\nNTm3X3oiv20wYf3tjy0mv7Ccv/90em0Bp6WYdu4v4Q/PLMfpdHD7VZNZtimPN77IJjMtgVsvC3wo\nYmFpJTl7S9hT4Othf7C4khPH9uGUcX0bvXaG+u+wKV0hrubyRNfpMxcBDpVU8tSHmzhQVMGvL5/Y\n5ESheYXlPP/pFrp3i2JQ7yQ27Swke28JQzKPtFx5PF7uf3FVveo5+Ios3z9nNJNMesPD4vV62Z5b\nzMLVuSzZuB+Px8s154xiyugjw73f/DKb9TmHGDe0F9edO+ao+SsG903m/66czAqbz7tf57BsUx7L\nNuUxc1J/Ljt1eL1K87Y9RazcUsCw/ilceYbhm/X7+WbDPjbvLmJIZjLTxmdy/KgM4mKjcbk97Mor\nZeueIopKq6hyuXH5HynZKyWOARlJDOydSI/EWNbnHGL5pjxWbimgsMrFaZP78z8nD6ntCXL9uWM4\nf9oQvlyzl4GZKYzLSq0dmnTrZRO59/mVvLsoh5Vb8tmTX0Z6jzh+PG88/VpINAMyEvm/KyYzf8Vu\n3v4ym0+W7eKTZbvoldyNA8WV9EtLYPaJWc0eo8bMSf35+Jsc1m4/wAMvrWLTzkImDEvjtEn9a7dx\nOBwM659S7+krXq+XtdsP8NaXOSy3+Sy3+UQ5HZw/bQhnTRnYYgFkWL8ULpo5jBfmb+Gup5bicnu5\neNawJh+17nA4uHjWMPqlJxDfLZpJpvkvqqcfN4D0Ht1ZsSWfwX2SGDkolT4941t9wz1yUCqnTB7Y\n4sUyOsrJjeeNZevuIgb1SdLjlKVdeDxeHA6O+jxXVrvZsquQx97fSFFpFedNG8LsqYNwOBxcdeZI\nzj1pMC/O38JS/xDLn140vrZlze48xLMfb2ZPQRn90hO46bxjmi3sdqQop5O+vRLYU1CGx+s9Kh+s\n2XaApZvyGNovmWkTmp+MMlwNyUxmuc3nQHEFaXUm065plR3Up1PP89Jq44elMWlEOss357PC5tcr\nUlZUudiWW8ygPkktFoTA9x7eMm8cnyzbzeC+SYwcmMqgPkkBzU947kmDWbmlgN35paQkxvLTi1qe\nt64pfXrGc9tVk9m04xCjs3rWmytP5Ns6XFHNP95az/rsg0wYlsYP545pcjjN859uYef+UqaM6U12\nbjEfLN5JVt9kjhuZQUFROY+8sRavF24875h2KQgFKy0ljk07C6l2uev10Nvn7xHb0v0z+PLmpacP\nJybayYdLdnLPs8v53+8eS/HhKv719npiop3cMm8cx4/rx6zxmXi8XgpLKklJjG23xr1wE+V0ktkr\ngdyCstqnioHvHiO3oIyhmSlBDWkd2DuJq88eyaNvb+CPz62g5HA1aSlxQQ9F7JHYjQnDuzFheGiG\nsktglNE6iWWb8njqw021j7N968scvnva8KO2c3s8/OedDVRWufnB7NF0j4tm085Cvtmwr15RaMXm\nfPIKyxk3tBcTR6QT5XRQWl7NG19s55E31nLuSVmce/JgnA4Hh0oqWbxhP4vW7WW3fxhAWkocZRXV\nPPr2BvIOlTPnxCzWbDvAu4tySEuJ4wdzRjc5oanT4WDyyAwmmXTWZR/ktYXb+O/y3aTEx3DuSb5h\nWV6vl1cXbANg3vSh9E9PZN6MRM6fPoTS8mqSG7SUR0c5Gdw3mcF9W+6yPWFYGhOGpeFy+4pGjY0n\nTu/RnfOmDTmq+tozOY5fXTaRe19YyZ78Mkb0T+Gm849uuW+K0+ng9MkDmHlsP9ZnH2TJxjxWbsn3\nDxsbFfDcSA6HgyvPMNz+2GI27jhEalI3rjlnVIvFE4fDwbihaRwzpBdrtx9kxeY8Zk3sH9QEe6dN\n7s+WPUUs83+ZqzvPQ1PnPKWFpw/UNWF4WkgSR5TTiRkYeUM9pP1VVLl4Z1EOnyzdTWy0k9494+nT\nM560nvGs31bAjjoTyX/31OGcflz9v6nUpG5cP3cM6T268/43O7j7meVce84olmzMY9G6fTiAGcf2\n4+KZw0I6XKwx/dJ9PTkLiirq9UKtrHbz7MeWKKeDq84YGdAE2OFoaGYKy20+23OL6xeF9vvyRntP\nXhrOzp8+hOWb81m4ak+9otDmXUW4Pd4mJ21ujBmY2qrrc0y0k+vPHcPbX2Yz56Sser+j1kjsHvOt\neuGJNCavsJy/vrKavQcOk9g9hlVbC/jzy6v58QXjjtp20bq9LFyVy8CMRK4+cyT5heX8/unlPP7e\nRnomd+PpDy0lh6u5/DsjGBUmw1d9f3eFFBRV1CtS1QyTzugZ2N+lw+Hgwpm+YdFvfZnNH59bQbXL\nQ7Xbw48vGFfvO4DT4Qj4aWVdSf+MRHbmlbL/0OHa93p3Xileb+saKaaM7kPO3hI+XrqL5IRYfn7J\nhKAnz5fOQUWhMFdZ5ebZTyxfrd1HbLSTS08bzqfLdjN/+W6mTcg8qrr+/tc72LqniONHZTBlTG/c\nHi+J3WNYsjGPi+vM1/DRkp0AXDxrWL0L9OisnvzttTW8/VUO23KLwetlw45DeL2++Wgmm3SmTchk\ndFZP9haU8ZdX1vDmF9nszi9jQ/ZBYqKd3HTeMQG12GeAPAAAIABJREFU/jkcDo4Z0ovBfZP53VPL\neOuLbIb1S2F0Vk/WZx/E7ipk3NBejBjQo3Yfp8NxVEGotaKjnK16ElpqUjd+fflE1mcfZLLJaNUk\n19FRTsYPS2P8sDSqqt2UV7qCHu+d3qM7l542gje/2M4P544JqmrvKw71anLS65b2/d5ZIxnSN5kT\nRvfWRJoiTfB6vfx/9t47uq3zytd+DjpIgmDvXSRBiqKa1YttyUUush2XxE6c2E7s9EwykynJvV8m\n3zdzv5lMSybVScaJ4ziOe42LLFvFKlYlJVaRYO8ESYAE2AAQ5dw/QEKi2GVRpKj3WUtriQenbBwA\n57zn9+7926equnn5YB19A24iDVp0GiUtXQM0dvYDgWtaeoKB3FQjq7NjpnzwVUgSD9y4jOhwLc99\nWMNPXykDID3ewBd2mcaJ/ouJsXtUe8/gOFGotM6K1eHi1vWpQX+zq5FlyYHzHrjvxgeXN3cNEBWu\nvWz3q6uRxOhQslOMnGvqw+pwBgWZc029ACzPuDIPrKlxYXzzvsIrciyBYK7Uttn5xWvlDDo93LIu\nlftvyOJ371ZRVN3Nf7xwhn/5eqDkasTjo6VrkGf3mgOWBfeuQKNWkhwbxuN35vPkmxX865+KkeXA\nJMHOtSkzHPnKERMREGcuFoXGMoUS5uApI0kS92zLRKtWBpurPHKbiVXZIgsFxjd4GDvXTaOZqxmX\nmLn66R3LiI/Uszwjinjh/7NkEaLQImDQ6UGnUU4QKPx+mV+/VUFZvY30eANfuXs5idGhxBj1/Py1\nMl7cV8N3H1wdzA4pqu7mraNNRBq0fGGXCUmSUCkl1ufFcfBsO1XNfSTEG6lrd1Df0c+qZdET0kpT\n48L44WPr+fWbFVQ2BgZuy5LD2VyQwPq8uHEZMcmxYfzg0XX8/NUyiqq7AfjiHXlzVqLD9Gq+98g6\nvv/Lo/zPXyr5f7+4gdcOBQzJ7puFv85CEB6iYXNBwswrzgKNWnnJjvvXr0qasiZ3PtFrVdy2Me2K\nHlMgWOy09QxSVm+jb8CNfdBNV+8wbT1DqJQK7tqSwR2b09Gqlfj8fmz9bpQaFWEqxZyye3asTSEy\nXMdbRxvZVpjIjjXJi1qYTb5ggLom53xZ8pgwsHF5/KTbXS2kxxtQKSUaOvqDy+yDbhyDI6wRqfJs\nX5lIXZuDo2WdwYYXVc19qJQKspONM2wtECxdBoZHePvjJg6ebUeW4Qu7TOwY9VX72t0F/Emn4lBJ\nB1/7t334ZRmn+7yn1jfvLRz3cL4uL47bN6Wx50QLeWkRfG6SSoKFJHZUELZeZFkxlik0lQXBdNy2\nMY3YCD0jHh+bV1ye8fhSIGW0wUNbz2Awq/GTljMrFQp2LCKRUTA/CFFogalts/OTl0qJNur47mdW\njUt1fO1QPWX1NlZkRvHtB1YGRaNV2dGsyIyiorGXklora3JjOVbRye/frUKrVvL1T60Yl6mzqSCe\ng2fbOVHZxY4NGcEsoV0bJn+oD9Or+e6DqzhbYyUtPmxaV3hjqIZ/+NwaXj1YT3iYZk5lQheSlx7F\np3dk8+L+Wv7/Z4voG3CzcXn8NZ16P1uuxW5XAsFio7zBxi9fLx9neK9USKzOjuGhm7LHXUeVCgVx\nEfpLNi0cK4G9GkgZyxSynu9AJssylY19hOpUpF/l13iNWklmkpHGDgcerx+1SnF+AH6Vv7fLwfq8\nOJ7fV8vH5Z3cvS0T+4Cb1u5B8tMjRfthwTWJx+tjX3Eb7xxrxun2Ehuh45Hb8sa1EFcoAhYBEWFa\nPq6woFMrMYZpCA/RUJgVNanv5/3XLyM/LZLsFOMlZcHPJ9HG85lCF2LpdaJVK4kIu7SMysnOw7XO\nWOZta/f5DmRNlgE0KgWJ0SLLRzA1QhRaQBo7+/npK6WMjBqA/etzxXz3M6tJignleIWFPSdbiB9t\n333hBV6SJD57cw4//P0pXjxQi63fxQv7atFrVXz3wdUTygiyk43EGHUU1/TQbOnnTE0P6fEGTGkR\nF4cURKlQzLpuXqtW8vCtuZd2Ei7glnUp1LTaOVMT8Nj51PbMmTcSCASCBabY3M1v3qpEoZB47PY8\n0uMNRBi0GELUV61XzuUiyqhDq1HSfkGL3G67E1u/i+tMsYs6y2m2mNIjqW2109w1QHay8byf0DVq\nMn0hOo2KjflxHC7tpKqpD6U6MOy8UqVjAsFiwdI7zJGyDo6VW3AMjRCqU/HQzmx2rE2Z1IZgrEzq\niXtXzmryQKGQpu2Au5CMtX/vuUAU8ssy3X3DJERfejMRwUSMoRrC9GraRz1gx54xM5MM14zhtuDS\nEKLQFcDSO0xJrZXCZdFBf4W27kF+8lIJrhEfX72nAKvDxasf1fOj54q574ZlQZHn2/cXTtplLDE6\nlJuuS+GD0608v6+WML2av3to9aSZNZIksXF5PO8eb+ZHz5xClmHXhtRFdxGWJIkv3ZGHe8Qr6lYF\nAsFVwfEKC79/twq1WsF37l9J3iIx9lwsKCSJ5JhQmi0DeH1+VEoF55r6AOZkNLyYyUuP4p2jjTS0\nOwKikMgUGse2lUkcLu3kSFkHEeGBh8Ol8tkLBDNR3dzHG0caqG1zABAyWn5/5+b0WflvLgWMYRpU\nSsW48jH7gJsRr3/RdMtcKkiSREpsKOYWO64Rb7D7Z0b84vQdFCwehCg0zww6Pfz4xRJs/S5ePlhH\ncmwo1+XG8lFJB0MuL1+6Iz9oTmnQq3nm/Wr+tNeMJMG375m+leTdWzM5Xd2NX5b5u4fWTNvScVNB\nAu8eb6a9Z4hIg3bRds8I0an524fWLHQYAoFAMCPHKjr5/TtV6LUq/ubBVSxLEh4pk5EcE0pDRz9d\nfU6SY0KvuNHwfGMaFQLrOvq5FWjpGiA8VHPJJRFLjWVJ4SRGh3CmpofwUA0h2qu/bFAgmA2VTb38\n7JUyfD4/yzMi2b4yibW5MePasl8LKCSJaKNuXPnYmMm0mAC+/KTEhlHdYqfdOkTLJ/QTElw7CFHo\nMmC1O3n7WBMFmVHjuo/4/DK//Usltn4XWwsTGHZ5KW+w8ZePmwB4+JZctq1MDK6/fVUSYSFqnvug\nhjs2pc+YBhqiU/HPj29ApZjZqDQ5JpS00TaFN69LWXT1xgKBQHA10WwZ4Jk9ZkJ0Kv7+s2uE/9k0\nnDebHiQxKoTq5j6iw3XjupFdzcRHhRAeqqGhw8HA8Ai2fjeFWdGLLht3oZAkie0rk3j5YB29/W7W\n5i6NskGBYDoCXcXKAJnvPriagsxrOzsuxqijq3cYp9uLXqsKmkyLTKHLz5ivUHvP0CfuPCa4dhCi\n0CfAL8vsL27j9UMNuD0+jpR1Utfu4DM7slEpFfz5/SoqG3tZuSyaL96Rj0KSGHZ5OVvbg1atnDRb\nZ01O7LgOLTMxl9TTe7ZncvxcNzeuTp71NgKBQCAYz5DLw6/eKMfr8/Ot+1YIQWgGkmPH2tIPERsx\nwJDLy9rc2CUjmkiSxLKkcM7WWimtswGQnhC2wFEtLrasSOC1Q/X4/DL5osRSsMQZ8wz1+WS+eW/h\nNS8IAcSOmk3bHC5S4sKw9AZKyS6l85hgesba0rd1D9JsGUCtUpAYI86zYHqEKHSJdNqG+MN71dS1\nOwjVqbjv+hwOlXawr6iNFssAm1ck8Mr+WuIi9Hz5ruVBs9EQnYqthYkz7H1+WJMTy61bsi6p241A\nIBAIApMBv3v7HFaHi7u2ZLBy2dXRBWwhubADmUYdyFJdap4yy5KNnK21cvBsGyD8hC4mPFTD2txY\nis3drBAPyIIlTKdt6Lxn6N0FrM4R9wiAmKDZtJOUuDC6+sba0S+NjNHFRHJMKBLQaOmn3TpEeoIw\nmRbMjBCFLoH2nkF+9NwZht1e1uXF8fAtuRhDNWxflcjT71VTVN1NTZsDrUbJN+8rvGaM5AQCgWCp\n897xZkrrbRRkRHLPNtEhcTaEB7uhDOJ0ewGWXLbIstGun42dwr9hKh69zcRDu/KI1Iuhp2Dp8s6x\nJoZcXh69zTTOUuJaJ2asLb094Ctk6R3GEKIWz0jzgFajJDZCT317PyDuR4LZIWTDOdI34Oa/Xyll\n2O3lsdvz+ManVmAMDZhJ6jQqvn5PAZ/esYwwvZrvPLiG1DiRQi4QCARLgbEuMlHhWr5yd4HwRZkl\n0mgHsu4+J7VtDlLjwggPXVomzBkJ4cGM4FCdiuhw3QJHtPgI0anJTVtaYqBAcCHDLi9F5h7iIvVc\nvyppocNZVIy1pbc6XHh9fqx2lygdm0dSLnj+zBCZq4JZIEShKfD6/FQ19+Ee8QWXOd1efvpKKb39\nbu6/IWvSC74kSdy+MZ2ffXsb24V3j0AgECwJhl1efvfuOSQkvn7PCgwhS0vUmG+SY0ORCdxbl0rX\nsQvRapSkxAXK5NITDEvGL0kgEMyeU1VdeLx+tq9MFNeAi4geyxRyOOmxO/HLMgmi89i8kRJ7viO1\nyBQSzAaRwzsFe0628MbhBrRqJWtzY9hUkMAHp1po7R5kx5pk7tiUPu324mYgEAgES4fn99XQ2+/m\n7q0ZLEsWrefnylgHMlh6fkJjLEsy0tI1KPyEBIJrlKPlnUgSbFmxMN6hixmDXo1WraTH7qIraDIt\n/ITmizGzaZVSQVJM6AxrCwRCFJqSU+e6UCkVhIeqOV7ZxfHKLgBWZ8fw8C25QvQRCASCa4Si6m6O\nVVjISDCwe0vGQodzVZI8OihVKiRyUyIWOJr5YVV2NAfPtrNcGCkLBNcc7T2DNHT0s3JZNJEG7UKH\ns+iQJImYCB1WhxOLaEc/74yVj6XGhaFSisIgwcwIUWgSOqxDtFuHWJMTw7fuK6S+vZ/jlRZcIz4e\nuc0kfCQEAoHgGsE+6ObZvWY0KgVfvmu5GFxdIimxoQFBKDUCrUa50OHMCyuXxfCzb28TpYUCwTXI\n0fJOALYtUIfhq4FYo572niEaOhyAaEc/n8RF6rlpbQp56UtzEkZw+RGi0CQUm7sBWJcXhyRJZKcY\nyU4R5QICgUBwLSHLMs/sqWbQ6eHhW3JJjBYp2JdKiE7N9z63lqjwpT2DLgQhgeDaw+vzc6zCQphe\nLVrQT8OYr9C5pj4kIC5ClI/NFwpJ4uFbcxc6DMFVhBCFJqHY3INKKbFqmbiwCwQCwbVKc9cAZfU2\n8tMj2bFWNA74pIjJFYFAsBQpq7cxMOzh5nUpIpt0GmJHRaFht5focB0a9dLMGhUIrkbElesiuvqG\naekeZHlGFCE6oZkJBALBtUp5Qy8AN6xOCrYbFwgEAoHgQo6WidKx2RBzQWZQgjCZFggWFUIUuohi\ncw8A60xxCxyJQCAQCBaSigYbkrR0u2UJBAKB4JPhGHRTVm8jPcFAmug8OC0xo5lCIPyEBILFhhCF\nLqKouhulQhI1wQKBQHANM+zyUt/eT1ZiOGF69UKHIxAIBIJFyKnqbvyyLLKEZkGM8Xx2kBCFBILF\nhaiPugCr3UmTZYAVmVHiIUAgEAhmiclk+m9gEyAD3zGbzacveO0e4AeAG3jRbDb/0mQyKYDfACuA\nEeBrZrO5+spHPjVVzb34ZZkC0V5cIBAIBFNwqqoLSQo0pxFMT4hORahOxZDLK9rRCwSLDCEKXUDR\nWOmYuLALBALBrDCZTDcAOWazebPJZMoHngY2j76mAH4JrAVswB6TyfQmsB4wms3mLSaTaRnwM2D3\ngryBKahoDPgJFWZFL3AkAoFAcPWzFCcPbA4X9e395KdHYgwVnQdnQ4xRz5BrQGQKCQSLDFE+dgHF\n5m4kCVE6JhAIBLPnJuBNALPZXAVEmkym8NHXYgC72WzuMZvNfmA/cDOQA5wa3aYeSDeZTIumDYks\ny1Q02AjVqchMDJ95A4FAIBBMyYWTB8DjwM8veG1s8uAO4HrgLpPJlALcw+jkweg2/3XFA5+B09Xd\nAKzPF5PJs2VdXix5aRHEhOtmXlkgEFwxrtlMIVmW2XOyhbo2ByNeHyNeP/UdAbU/PESo/QKBQDBL\nEoDiC/7uGV3WP/p/g8lkygGagB3AR0AZ8Dcmk+mnQDaQRUBA6pruQJGRIahUl64dxcbOzgS0tWsA\nW7+bbauSiI+fX1FotjFdaURcc2MxxrUYYwIR11xZrHHNkXGTByaTKdJkMoWbzeZ+Lpg8ADCZTGOT\nB3FcMHlgMpnSTSaT0mw2+xbmLUzkVFUXCkniutzYhQ7lquHOzRncuTljocMQCAQXMStRaCmmfJ48\n18WrH9UH/5YAvVbJzrXJCxeUQCAQXP0Ee7ebzWbZZDI9SqCkzAE0ApLZbN5jMpm2AocJCERVF243\nFX19w5ccVGysgZ6egVmte6S4FYCcpPBZbzPfMV1JRFxzYzHGtRhjAhHXXJlLXItcPFpykwed1iGa\nLAOsNcWRlT6/ZcaL9bMVcc2NxRjXYowJRFxz5XLENaMotBT9IoZdXl46UIdapeCHj60nLkKPSikh\nSTM+kwgEAoFgPB0EBvdjJAGdY3+YzeZDwHYAk8n0IwKDfsxm8w/G1jGZTPVA9xWIdVaUN9oAWCH8\nhAQCgWA+uOonD/YebwJg9bJoMXmwiBBxzZ7FGBOIuObK5Zo8mI2n0JLzi3jzSAOOoRF2b04nOSYU\ntUohBCGBQCC4ND4AHgAwmUxrgQ6z2Ry8O5lMpj0mkynOZDKFAncB+0wm0yqTyfT06Ou3AWdG7yEL\nzojHR02LneTYUCIN2oUORyAQCJYCM04emM3m7WazeTcBYahpdPkPzGbzVrPZ/HUgkkU0eXCqqhuV\nUmJtrvAhFQgEVz+zKR+7IimfVyrds6HdwYEzbSTFhPKF3QWoP8ExL2dcV5LFGBOIuObKYoxrMcYE\nIq75xGw2HzOZTMUmk+kY4Ae+aTKZHgMcZrP5DeApAsKRDPzIbDZbTSZTL6AwmUynABfw8AKFP4Ga\nNjsjXj+FmSJLSCAQCC4THwD/BPx2qskD4FFgiMDkwY9NJtMqApYVX1pskwedtiFauwdZnR1DiE69\n0OEIBALBJ+ZSjKbnJeXzSqR7+mWZn790Br8Mn70pB/snOObljOtKshhjAhHXXFmMcS3GmGBpxLXY\nxSOz2fz9ixaVXvDa68DrF63vBx6b/8jmTkVDoBV9QVbUAkciEAgES4OlNnlwukp0HRMIBEuL2YhC\nS8Yv4mhZJ/Xt/azLi6MgUwz4BQKBQDCeysZeNCoFuSnGhQ5FcJkZ9jg5ZTlDUddZwjUGtiZvJD8q\nF4U0m0p6cPtGePbcSySFxnNn1q3zHK1AsLRYSpMHp6q7UasUrM4WpWMCgWBpMBtRaMmkfO491YJG\npeChndkLHYpAIBAIFhnDLg/t1iHy0yPnvbT4asAv+2l0tKBSKEkzpCxK770Rn4dDbR9T3FVCQmg8\neVE55EXlEKE14vaNYHP20uO0UtpTyZnuMjx+DxISMjKl1kqidJFsSVxPpjGdWH00EVojSsXEz97r\n9/JU+bNU9dZgGe4WopBgzji9Lipt1Xj93inXSQ9PJTE0/gpGJZgr3XYnHdYh1uTEoNdeSsGFQCAQ\nLD5mvJotlZTP/qEROm3DrMiMIipct9DhCAQCgWCR0WgJzHdkJYXPsObSpt81wIfNH/Fxx0l6nIFO\nbMlhiWxL2sj6hDWM+DyY++qo7q2lZaANg8ZAjC6KGH0U8SGxZEdmEaYOndcYfX4fJyxFvNe4D7vb\ngYRE62AHp7vOAhCqCmHIO75EPEYXxdbkjWxKXEefy87R9pMUdZfwTuMHwXUUkoK4kFhuSt3OxoTr\nUCqU+GU/z5x7kareGgDcXve8vjfBwtLS30akLgKDJuyy7M/j93Kk/Th7mw4w6Bmadt00QzLfW/+d\ny3JcwfxQ3dwHwPIMUXEwhizLi3LSQCAQzJ5ZSdxLIeWzts0OQG5qxAJHIhAIBILFSFNnPwCZideG\nKOTyutjfeoQzXaX45fPJvDZ3Hz6/D7VCxYaEtYz4PJRZK3mp5k1eq30br+wLrqtRaugc6qLmgv1K\nSKQaksiLyuW6uFWkGJIuOUaP30tpTwXHO07j8Djw+QJxOn0uBkYGUSvU3Jq+g1vSbsDu7qe6t4bq\nvjp6hq2kGpKJ1kcRq48mzZBCTmRWsFQsXGMgPTyV+3J2U249R/dwDz1OGzZnL62DHfy5+lX2txzm\nrmW3UWmt5mx3GdkRmTjc/Qx5Lo8fYVN/CyEqPXEhsZdlf5cDt2+EBnsTKYakSUURy1A3bQPt45ZF\n6aNIN6RMmmF1tdE11M1/FP2CgmgTX1/1pTlv/+eqV+gY6iJGH0WMLgqdSsfh9uP0uvrQKXXcnnET\n0bqpxYT08NRPEr7gClAVFIUiFziSxUGjo4XfVzxHfEgsD+c/QJROnBeB4Grkmsl7NLcKUUggEAgE\nU9PQcW2IQl6/l6PtJ9nTtI9BzxAapQa9Uht8PTU8kXWxa9mYsJYQdQgADvcAxztPc6a7lHCNIVCm\nFZlDUlgCXr+PXlcvVmcvrQMdmPtqaXA00zLQzscdJ/m3bT+ctW/PGHa3gwOtRzjZWRzMrojQhSP7\nZQCUkpJtSRu5PfNmIrQB/6cQdQhJYQnsTLt+1sfRq3RsSFg7blmfy857jfs43nmap8qfBQIZHF9b\n+UV+VfI7el32Ob2XyThlOcMfz70IQE5EFtuSN7EqdgVqxfwNy/yyn3M2M8c6T+P0OMmNzCYvKoc0\nQzKW4W6Otp/klOUMLp8LlaRkVewKtiVvIiM8jZKeco62n6Te0TjpvnVKHTmRWeRF5bAx4Tr0qqsz\nI/tg28fIyJzrrWHYMxz8/s8Gh3uAY52ngYDgN4ZKUrIzdTu7MnbOewadYH6RZZmq5j6MYRoSomb/\n3VjsDHucnOkupWPIQo/ThnXYxpBnmE2J67gtY+eUv4NTljP8ufpVvH4vfW47/3Lyv/lM7j1sSFgr\nMocmoc9l552GD6jsrSbdkEJ+lIn86Fzi9DHTni+n14XVaaPHacPh7idCayRWH02MPhqdSjvldgLB\nXLhmRKHaVgcqpURm4uLu4iMQCASChaGxs5+IMA2RhqU5yBocGeKEpYjDbcewufrQKjXcmXkLO1O3\no7vgIX6yznhGrYHbMnZyW8bOCfvVKBUkhMaTEBrPiph8bs+8CZfXzR8qn6fCVoXVaZtTNszgyBA/\nKf41NlcvYepQbkq7nq1JG1mRnnVFOglG6iJ4OP8Bbkq7nncaP2BwZJDHV3wevUqHVqnFJ/vw+r2o\nLlHAKeup5E9VL6NX6UkJS6TW3kCtvYFQdQhphhRi9NHE6KOI1BpRSOezb5xeFzanDeuoAOfyTV3G\nplGpMKgMwX05vS6Od5ymz31e0Kqx1/NO4140Sg0jvhEAjJpwNiSspaavjuLuUoq7S1FKSnyj2WF5\nkTkUxiwPZgXJskzHkIXq3hrKrecot56j19nHfTm7L+ncXAmKu0op6irh0eUPjvveD3mGOdlZBAQE\ntHJrFRsTr5v1fhv7mwHYnbmLjYlrsTp7sbsdZEdkiuyJJUKHdYj+oRE2FcRfdaKHzdlLvaOJaF0U\nMfpowjVhtA12crjtGKe7zuLxe4LrhqlDUUgK9rce5njnaW7PvJnrkzcHr3l+2c8LZW/xRtX76JQ6\nvrLqURzufl6tfYtnq16izHqOB02fIlwz92euSls1aYaUy1a+ORt8fh+/Lf8j/e5+8qJyWR6dS5Yx\nAwkJm6uPHqeNfnc/Rm04sfoYonQRKBVKhj3D9IyKNU6vc9w+w9RhQeEGZF4sP8jb1fvw+D3oVXoq\nbNVU2KqhNiCqK6b4PvllGZfPNWXsBs3548Tqo8mPMpFpTLucp0dwjXBNiEJOt5eW7gFyko3CPFQg\nEAgEE+gbcGMfHGFNztLrJtPc38r+lsOU9lTglQNlYTtStrErY+e8Dbx1Ki2myGVU2KpoGWiftSjk\n9Xt5quJZbK5ebk67gd1Zu+Y1e2Y6EkLjeGLF58ct047Oyrp9I5ckCtX01fH7yj+jkpR8Y9WXyDKm\n0zXcw8cdJymylAR9i2ZCKSmnzcbxjfho8Ywv89IqNWxL2si25E1E6iKo6aunureWOnsj0fpItiVt\nZEV0PkqFElmWqXc0cbT9JC0DbayMWc7WpI3EhkRPecyOQQv/cuonWEd9qBYjsizzdsP79DhtvNPw\nAQ/k3h187eOOk4z4PWxN2sjHHSc521M2J1GowdEEQHZEBlG6SCEELUHGSsfy06++z/bpyufHZbCp\nFeqgEBSti2Rb8ibyonKI1UejV+nx+Dwcaj/G+037ea32bfY07kM7mlHqlb0MjAwSq4/maysfI2HU\nHD03chnPnnuJkp5yqnrN3JK2g5vStqNRamYVY529kSdLnyYpNIG/W/cttLPc7pPyYctHVNqqAWgd\n7ODDlo9QK1T4ZP+40uoxFJICrVKD0zu1WHMhKkmJV/Zh1BjYnfUpNiVeR5/LQVWvmareGrqHrVNu\nK0kS4RoDsfoYYvVRGLXh2N39wcyhnmErTf2tNDgCovT+liP867YfXLFzJ1g6XBOiUF27A1mGHFE6\nJrjK8ct+znaXcX3EuoUORSBYUjQuUT8hq9PGT4qfxCv7SAiJY1vyJjYkrCV0DmUxl0raqD9KS38b\n6+JXz7i+LMu8ZH6TOnsja2ILuWfZ7XMuO5tvxgbaLq97zuew0dHCb8qeAVnmK6seI8uYDkB8SCz3\nZe/mvuzduLwubK4+rE4bfW5HoIXHKBqlJuBVo48iQmuc9tzExhpo7ujG6rJhdfbi9/tYEZM/LjNm\nbdxK1satnHR7SZLIjsgkOyJz1u8vMTQetUIViHuR0uBoDpqnf9T2MRsS1pIWnoLX7+NQ2zE0Sg2f\nWnYHjY5mqnprcXld487ZdDQ6mlFIiuD3XrD0uFpFoY5BC039LaQZUjBFZo9mt1iJ1EawPXkTy6NN\nE64naqWam9NuYFPiOt5v3E+5rQpZDlyQNJKGjSlruD/znnHXwRh9FH+99qscbT/Ju40f8E7jXo52\nnODurNtmJbAeaT8eiHfIwnNVL/OlgocvW0ZG4L0tAAAgAElEQVRWg6OJPU37+dSyO0gOSwwu7xi0\nsKdxH0aNgX9Y/23aBjqo6q2h1t6AVqkldjQDJyDGOLCOdrQc9rpYZswYzdCJIUwdAmOxyjKOkYFA\nKd5oyde2zHVsit4UvIdE6wNC3LbkTZ/4vfn8PmyuPva3HOJox0lKusvnJGjPlfbBTv5U9TIe3/ns\nskhdBF8pfGTWAuC1hGWoi/6RAaJ10UTqzt+73b6RwL3eZSdcawgKsrIs0+O0cq63hureGrx+H5sT\n17EqdsUlZyjPhmtCFKoRfkKCq4iBkUF8si/ok3EhR9tP8lLNG9h8Vm5NunkBohMIliZBUWiJdR7b\n23QAr+zjwdx72Z686YqWPKSEJSEh0TLQNqv1P2r7mGOdp0gNS+ILyx9cdIIQEJwpd09TunUxdrcj\n6FEkyzJPrPg8+VG5k66rU+lIDksc99ByqYSo9aSpU0gzpHzifc0GSZKI0BrHlajNFq/fS1N/K11D\n3aPlcTYGRga5L3s3aeGXL/4To+Vhu9J3srf5AC+YX+Pv1/0VJ9vOYHc7uCFlKyFqPatjV/Be0z4q\nbNWzEjQ9fi8t/W2khCWKGfolit8vU91iJzZCR4xRv9DhzImx7/0t6TdOKQRPRZg6lAdy7+YB7h63\nfLIyYwhk0Vyfspn1CWvY1/wR+1sP82zVS0RojZiisqc8zuDIECXd5cSHxBKmDuVMdxmphmRuTd8x\np3in4o26d2lwNNPc38pfrf4yqYZk/LKf56pfwSv7eMh0HxFaIxFaIyti8i/LMS9kqvN1OVAqlMSF\nxHBz2o0c7TjJ8c7T8yoKHWo7RutAO6GqECRJwuv3YhnuprSnkvUJa+btuNMxJrAEM6hGvbE0Sg15\nUTnkR+USFxLIBPf5ffS5HfS6+oKl0ZMRpg4lOSxxwljE6/fSMtA+bhyglBREaiODpYVev5eSngoO\ntx0f58WnkpRE6SJx+dz0j0z8PoSqQ1Ar1Ngvmlyp6q0hXGNga9IGbkjZOi9Z3teMKCRJkJ088SFb\nIFhsPFn6NDZXLz/c+PeEac6bUnr8XvY2HwDgVHuJEIUEgstIUBRKWDq+c1ZnLycsxcSHxLEteeMV\n98DQqbTEh8TSOtCOX/ZPKfJ4/F6OtB3j9bp3MWjC+OrKxxbtg/VYXO5RD57pcHld7G0+yMHWo3j8\nHuJDYrk/5y4KovPmO8wFI0JrpNbeMCvPpT6XnbM95VT31lJrbwj6Gl3IwbajPLr8ocsS24hvhDPd\nZURojezOuhW728FJSzGH2o5RaitDQuLGlK0ArI4r5L2mfZR0l89KFGodaMcr+8g0ZlyWWAWLj+au\nAZxuL+vzFk+3wNng8/s4ZTlDqCqEwpjlV+y4epWOu5bdRk7kMn5R8hQlPeXTikInLcV4ZR/bkjay\nLmEN/3765/yl/n2Sw5IoiDZ9olhaBtpocDQTrYui19XHz87+D99a/Th19kaa+1tZF7+albEFn+gY\ni4HYkGhyIrKotTdgddpG/YwuLz6/j9KeCgyaMP516w9QSAq6hrr555P/xamuM5dVFPLLft6oe5de\nV1/QMylaF8WQd5ie4TEByBrIxppEYBmjzFoJQLQuCqWkwOrqnbQscDLC1KFBUUltlzjVXEaNvX7S\n+xUERNEoXSRur5sBzyAA+VG5pBtSsLp6g11OtUoteZE5Af9AXQT9F2SWOT0uVscWsjwql7yoXHyy\nlyPtJzjeWcSepv10DnXx5cJH5ng2Z2bJi0Ier4/Gzn7S4gzotUv+7QqucoY8w8FZ9bcb3uezefcH\nXzvecRq724FCUtA50I1lqJuE0LiFClUgWDL4ZZnGzgHio0II0akXOpzLxt6mA/hlP7dn3LRgWTdp\n4SlYLN30DFuJv+h65Zf9nLac5d3GD7C5+tCrdHy18FEidYs3q1c3y0whj9/Lr0p/T4OjmQitkTsy\nb2ZTwrol0bZ9OsY+O4e7n2j91K3XGx3NPFn6NMOj5qzxIXHBTmgx+miidZH8uPhJKqxV+Py+Ceft\neMdpirpKiNZHEqOLJlofRaohidhpuviU9lTi8rm4IWULCknBvdl3UmGt4s26d/HKPgpjlgdnkpNC\nE4jTx1Bpq2bENzJjScSYn9BYSaBg6XG+dGzq7/VipNJWzYBnkBtSti6IP1tORBZ6lZ5yaxWfyf3U\npL9PWZb5uOMkKoWKjYnrCFWH8JXCR/jJmV/zh8rn+V/r/5po/cwle0OeYSJ9E7O4DrUeA+BB070M\neYZ49txL/OLsU/hkH2HqUD6dc88nf6OLhE2J66i1N3Cis5jdWbdOuZ5f9mN3OoC5jQ3qHY0MeobY\nnrw5OK6ID40j3ZBKdW8t/SMDl2QwPhnvNe7jQOuRadeRkIjSRQYFlhh9NLEhMecFJM9QsAzL3FeP\nUlKQbkghRh9DtD5y2t+E1dlLVW8NRV0lFHWVBJfHh8SSF5Uz7n16/d5g6XfPsA0/fnambmd78qY5\nNdqYivtz7uKurF2U9VQSP+rhdblZ8ipJQ0c/Xp9MTqrIEhJcOSxD3WiVmjk/3DSOGsUBfNxxiq3J\nG0kzpODxe/mg+SBqhZo7M2/hzfr3KLNWClFIILgMdPUO43R7WZ19+WfVLpWBkUEUkuKSvX9szl5O\nWIqID4nluvhVlzm62ZNmSOGU5QzNA23jRKFBzxC/PPsUrYMd51t2p+8clx25GJlNppAsy7xQ/RoN\njmbWxq3kC/kPolEuHbFxOsbKnvvcjilFoereWn5b/ke8fi/3Zt/JdXGrJr1XFsYs53D7MersjeMy\nDLx+L2/Uv8uQZxj6xm8TqY0gPyoHU1QOK2MKxp33k5ZigGBZhUETxqey7+DP1a8CsDN1e3BdSZJY\nHVfIB80HOWczszquEICSngrKeip5IOeucW26x+7dQhRauoyJQnkz+AnJsozb5561F9V8M1Y6tjlx\nYbwolQolBdEmirpKaB/sJMWQNGGdWnsDXcM9rI8/73eXHp7Kvcvu5JXatzjTXcot6TdOe5yBkUH+\n6cR/sCwqna8WfDEoWAyMDFLUXUKcPob8qBwUkgKVQsUfKp/HL/t5ZPlDi/6+MxfWxK3k5Zo3OdFZ\nxB2ZN086IVTVW8Mbde/SPtjJHZm3cEfGzbPOJD7bXR44TmzhuOXrE9bQXNtKcVcpO1K3feL3UdJT\nwZ6mfUTrIvnGqscZ9AxhddqwufoIVYcEvZ6idJHTZqXqVFq2J29i+yV6N8myTOdQF+a+OqKMBlLU\nabMSKOcDjVLDunksz1vyolBNW6Amz3SF/YQc7gFiWTplCIuV4x2naepv4dO598yr+dZs8ct+9jYd\n5N3GD0gxJPH99d+Z0/b1o7ONt6TdyIctH/FKzVt8d+03ONEZaCW8M3U7mxPX81bDHsp6zl22Wuur\nFZfXjd3tIEYftSg+f8HVSVNnIO04IzGc/pEBFJKCMPXCDRKLukp4ruoV4kJi+F/r//qSyr72Ngey\nhG5bwCwhIOhn0zrQzoaEtcHlJzqLaB3sYHXsCu7LvmvBBllzZTaeQvtaDnHSUky6IfWaEoQAIkdF\noYv9EMY41VbCr0ufBuCJFZ9nVeyKKfe1KraAw+3HKLNWjhOFKmzVDHmGuSFlC9uTNwc9JBrsTZj7\n6jjWeZpjnadJD0/lm6seJ1QdQp/LTnVvLZnh6cRfMGu7KXEdFdYqFGqJnIisccdfExsQhc72lLMy\ntoB3Gj4IlnDH6mO4PfMmIPDQMJYRFqldvFlugkvH4/VT22onOSYUY+j0WWPHOk7xvPk1ciKyuD5l\nC6tiChYsQ3BgZJByWxUpYUmkGpIXJAaAwuh8irpKKLdWTSoKfdxxEoBtyRvHLV8VW8ArtW8FO2tN\nx+G2Yzi9Liq6zXwQdpDbMgK/z2Mdp/D6vVw/miEIAZN9gzqMHqdtgrhxtaNVargubhXHOk9T01dP\nXlRO8LWOQQtv1L3LuV4zEhIGbRjvNX6IZaiLL+R/ZsaMSL/sp6SngjB16IQmBOviV/N63TucspyZ\nsyh0cTZom6OTZ8+9iEah5iuFjwYnwOfS+OByIUkSSWEJJIUlzKsv1GJgyT9F1Y6aTOekXLkbdXVv\nLb8oeYq/2fIE2brJzSTnA1mWOdB6hPyoXJLCEq7YcReKcus5/lz9KjIyepWeT2XfsaDxDHmG+eO5\nF8+3tRxoZ9jjJEQ9e0PCensTEhK7MnbS47RR0lPO8c4i9jYdRK1QcXPajYRpQsmLyaa6p+6ypmle\nDfhlPx80H6S2rwHLcHfwwUOv0lEQncfKmOUsj86btlXzhciyPOUDd4OjmQprFZsT10/bhllw9dNw\ngZ/Qfxb9Eq/fy3fXfuOKf+5+2c9b9XvY13IICHT4qHc0zXkgZHP2cryziLiQmFn5ocwnKYaA2XRz\n/3iz6TNdZSgkBZ813X9VzdKezxSaXBQqt57jrfo9RGiNfGXlI9eUIATnM4UmE4UqrFX8tvyPqBQq\nvlr46LiHlckIlJ3oKO2p5IGcu4PX6rHMh61JG0kMjSdxNJV+Z+p2/LKf1oF29rccpri7lJ+e+Q3f\nWv0EpyxnkJHZdJH5qkJS8JWVj0462E81JBOti6TCWsWvy/7AOZuZGH00gyNDHG4/xs3pN6BWqLC5\n+ugfGWBN3Mor7tsluDI0dDgY8fpnlSV0oO0oEhK19gZq7Q0YNeHclrGT61O2XKFoz3PKcga/7GfT\nAmUJjTHW3azcdi4opo4xZjCdEBLHsos8uSJ1EURqI2hwNE07XhvxeTjcfpwQlR6dWsu7jR+SG5lN\nuiGFI+0n0Cg1E377OZFZ5ERmTbq/q51Nies51nma452nyYvKwev38m7jh3zY/BEyMqbIbO7N3k12\nUhI/+ujXnOkuo8dpm7F8u8HRTP/IAFuTNkwQOg2aMPKjcqm0Vc/a3qJ1oJ036t6lpq+elLBE8qNN\n5ERk8Xr927h9I3yp4OFJRUTB/LD4WntcRnx+P7XtDhKiQgifQdm/nNTbAy7jBxqOXbFjQiDL5PW6\nd4IzWUsZy1A3z1S+gEqhJFIbwYctH1HdWzunfdjdDkq6y4MtNj8JLf1t/Pvpn1FpqyY/KjfYYrKp\nv2XW+wi42beSFJaAXqXjvuw7UStUvGB+jT63ne3JmzFqAwLQ+uSVyMhUWKum3We9vYn2wc5Lf2Nz\nwOf3UdNXx6s1f+GHx/6N/+/4v1Nnb5x5w1kiyzIvml/n7Ya9VPcFPuu8yBw2JKxFr9JT1FXC05XP\n84/HfkT3cM+0+7I5+3il5i3+7vAPecH8+gTDuab+Fn5R8hR7mw/wTyf+g99VPEdL/+w6KAmuPho7\n+1EqJDSGIXpHH/B+UfLUlNkO88GQZ5gnS59mX8sh4kJieMh0H3B+BnU2DIwMUmQ5y7NVL416CU2e\nOn4l0So1JIbG0zrYHvydWZ02mgdaMUVmX1WCEIBWNZYpNLF8zOq08YfK54Oix2QdJJc6EbrR8jHX\nxA5kxzpO4Zf9fHPV4zMKQhAoO1kRnU+f207bYAcQ+I5X2qpJDUuatEObQlKQHp7KYwWf5YaUrXQM\nWfjJmV/zcccp1AoVa+NmX0opSRKrYlfg8rk5ZzOTH5XL99b9FVuTNtA/MsCZrlJA+AldC4yVji2f\nQRRqHmjFMtTF6rhC/nHj33FjylbcvhFeqnlzxnHJTMiyzK9Ln+bfTv8Mj987q/VPdBahlJSsj1+Y\njlBjhKhDWGbMoLm/FYd7vPh6wlIUMJieojtmljGdQc8QPU7blPs/aSkO+tz81aYvIssyz1Q+zynL\nGfrcdjYlXIdedXV1jPskZBnTiQuJobSngnp7E/9Z9Es+aD5IlC6Sr6/84mj3tSTCdQa+vebLbElc\nT+tAOz8ufpJhz/CU+z3bXQbAmtjJO9htGP2enbacmTa+PpedZ8+9xL+f/jnmvjriQ+PoHOrig+aD\n/Kr093QOdnNL2o0LWvp+LbKkM4Vauwdxj/iueCt6y3A3AOVd1QxmD12xQW91bw3AtBfOpcCwx8lv\ny5/B5XPz2PLPEhcSw38V/4pnz73I/97w3Vmf77fr93LCUsQtaTdecpaRLMsc6zjFyzVv4ht9CLsj\n82bO2cwcbT9Bg6OZ5bPsmtA60I7H7w3OlETro7glfQfvNX4YzBIaY13yKp4teY0yayVbkjZMuj+H\ne4Cfl/wPOqWWf9r8vXmtby/pqeCFo68xODIEgE6pY8Q/ws/O/pZ7lt3OTanXTznD45f9dA9bZ5xV\n+EvD+3zcEWhX/a3VXx73OcuyTPtgJ8c6T3Go7RhH2k9wf85dE/ZhGerm/ab9FHeXBrshHW0/wdDI\nEI8WfBa1QkXnUBdPljyNx+fh9oybqLBWcba7jLPdZayPX8ujyx+ccTb4nM3M+00H+GLBZxe1aa4A\nvD4/LV2DpMSGUd8fEDEzwtNGhcHf8Tdrv3ZJpWROr4u/1L9Pj9MaXKZSKLkzcxepk8x8vWR+g6re\nGlZE5/NYwUPolDr2txzibHcZn865e5x/ycWcs5l5s/69cQJwZng6183hAXg+STUk0zFkoXu4h4TQ\neM50BQaWc3lAXywEM4W8EzOFqnprcPtGuD/nrsvaRv1qYqx8ajJBtctpRa/WTcgGmI6VsQWc7jpL\naU8lqYZkTnedxS/72ThD5oNCUgR+Nyode5r2A3Bd3Ko5Ze4CbEnawJnuMjYkrOWurF0oJAU3pGzh\nQOsRDrYeYUPCWuEndA1gbrEjAblp09/Pj1/g35MQGsenc+8hKSyB56tfo9xaxU1pl244W2atpGI0\nE31/y2Fuy9g57vWm/haeLH0ajUJDrD6acK2BjiELa2ILF4X4XhiznFp7A5W2quC4dcQ3wpG246gU\nqnHlxReSaUynuLuURkdz0Aj+QvyynwOth1FJSm5I2UJ2XDK70nfwfvOBoF/YQmRpLSSSJLE5IWA1\n8ZMzTwKwJXE99+fcNeFZQKVQ8bm8BzBoDOxtPsCrtW/zyPIHJ+xzrHQsRKUnN3LZpMddGVuAVqnh\ndNdZ7sy6dcKklF/2c6jtGG/V78Hj95Aclsi92XeSH5WL2zdCbV89Vb01GEL1osPyArCkM4VqWgIz\nVblX2GS6a3Q2IPADKr9ixx3LlLE5e2e1vs/vwzuL2YbLwVSt++aKX/bzzLkX6B62cnPaDaxPWEN6\neCp3Ze3CMTLAc9UvzzrzZ2x278OWj/iw+aM5xzLiG+FPVS/zvPk1tEotX1/1JXaPXgQzjGnAeOPo\nmagPzjZmBJfdknYj+VG57M7aFcwSAkgIiyUxNJ7q3topDU8/ajuK1+9l0DPEvpbDc35/s6V1oJ1n\nKl/A4/dyffIWvrX6Cf59+w/59uqvEKYO5Y26d/ldxZ9wel0TtvXLfv5c9Sr/5+R/cbD16JTH2Ndy\niA+aDxKnj+Gbq5+YMMCRJIkUQxL3Ze8mTB3KKcuZCd9tp9fJj4t/xemusySExPFI/oP8aNs/khOR\nxdmecn5T+gc6Bi38suR3DHmH+VzeA+zO2sX31n8nMKsSlsTprjOcneE37Zf9vFL7FvWORl6re2cO\nZ1KwELT1DOL1+clMCqemrx6ALxU8zI7UbVhGBcKhaWbOJqNruIf/LPolh9uPUdVbE/xXbq3io0m+\n5z6/j0qbmRhdFF9d+Sh6lR5JktiStAGP38vpC7peXEzrQAdPlT+LZagbU2Q292TdzvfWfZvvXvf1\nRdPpakwgGSshK+4uRSkpWX0VtgA+7yk08brr9ASucXH6iQ8u1wqh6hBUkpK+i0Qhv+zH6rSRFBY/\npxKr5VG5qCRlsKXwic4iFJJiVmWRkiSxO2sX9+fchVFjYMcFRtKzJTE0nn/Z+v9wz7Lbgw840foo\nVseuoHWwgzp7Aw2OZtQKFSlhosxhqWLpGyYmQkfoNN0pR3weirtKMGrCyY86bx2xIjofCJSWTkZJ\nTwUftX0c/Heis4gRn2fcOj6/jzfr3gs2H3i/aT8253mXdafXxR8qnmfY40RGpsZeH+yWtDVpvE/P\nQlEYM3Yezme4v1b3DlZXL9uTN03ZVGFMbG3on3w8XWGtonvYyrqENRi14QDckXkLmeFpwVKpxHnq\n1rSY2ZC4FrVCjUEdxlcLH+Xh/E9POTksSRJ3Zt5CmiGZk5bioA3GhTT1t2J3O1gZO7VHlkapYXVs\nITZX3wQfqD6XnV+W/I5Xa/+CVqnh83mf5vvrvxP8rWiVGlbE5PPp3Hv4wur7FzzL+VpkaWcK9QwC\nkJkYfsWO6Zf9dA33EKE1Ync7KO4qDZYSzSfDHidN/a1AoKuL0+ua0Vfl12V/oNdl539v+Os5m/R2\nDnXRMWiZVWrfactZ/njuRbIjMrklfQfLo3Ivqe7eL/t5ofo1Km3VLI8ycc+y24Ov3Zx2A1W9tZRb\nqzjSfnzGWYFBzxDdTiuphmQGRgZ5s/49QtUhU2bdXIzV2cv/lP+R9sFO0gwpPLHiC+OMUsPUocSF\nxNDU3xrMSJmJsQvosoiM4DKNUs23Vj8x6forYwrY23yA6t6aCWadTq+LI+3HMajDkCSJ/a2Hx5Wf\nXS4GRgb5bdkf8fg9/MOWr5GuOV+fnROZxffX/zVPVz5HSU8FHYMWnij8QjDlX5Zl3qh7lxOWwMza\nOw0fcF38qgkeScc7i3ij7l0itEa+tfrLGDRhU8ajUqhYn7CGg61HqbBVs/qC83K0/STDXie70ndy\nV9au4HfwG6se5+nK5yi3VvGvp/4bGZl7s+9kS9J6IHCzzIvK4UsrPs+/nPwxr9e+Q0F0XjBj4GKK\nukroHrYiIXG2uwxzb904k1TB4qKxI+AnlB4fytv2+kA7bH0k92XvxulxccJSxPeP/jMZ4WnkReWQ\nH5VLRnjqlL/pCmsVf6h8AZfPxc7U7ezO2oUCCRmZHxz7V6r7aid4I7QNduDyubguftW4/W5KXMfb\nDXv5uOMk1ydvnnDddLgH+G3ZM4z4PXy58JFx3/fFRPqo2XTLQBsZxjTaBjtYEZ03bfbTYmU6o2mn\nLyAKXUtlChejkBQYtUbsrvGiUJ/LjtfvJcEwt0wJnUqHKSqHSls1pT0VtA92siqmYNr7wMXsTN3O\njpRtl9XvZ0fqds72lLOnaT/tg51kGTNEs4MlisfrxzE4Qt5olpBf9vPbsmdINSSzO2tXcL3Sngqc\nXhfb0zePu44bteGkG1KpdzQx7Bked92rszfyVPmzE455pruMrxY+Gnz4PtJxgm6nleuTN5MRnsaz\nVS/xWt3bfKXwEQBeMr+J1dXLrek7uGfZ7Yz4PFidNjx+D+nhqfNyXuZKXEgs8SGxVPfW4PF5qOwN\nZNQnhSZwT9btU26XEpaEWqGecpJ1bNLzptTrg8uUCiVfLPgcL9W8ye0Z12bGSYTWyD9u/FtC1PpZ\n3ZOUCiWfz/8M/3b6Zzxf/Ro/2Pi3454jz5eOTW/MvSFhLSctxXzQfDBYuTLkGeL9pv04vS4KY/L5\nXN4D15Qf6tXCkpbh+gYCg7bo8CvXFrLX1YfX7yUnIgtTdBa19oYJ9bPzQY29Hhk5eCOyzpAtZBnq\noqq3hq7h7uBswmxxed384uxTPF35Z07NUDfaPzLAyzVvIkkB070nS3/Pj07/NNjScLZ4/F5+X/Fn\njnWeJtWQzBcLPjvupquQFDy6/EFCVHrebtg7Y2ZS86iAtiI6j79a/QSh6hCer36N4x2nZ8yeah/s\n5CfFv6J9sJNtyZv47nXfmLRzTlZ4Bi6fC8tQ97jlftlPla1mnI+NLMs02JuI0BqJ0s2uC8/K2OUA\nlPVMnH36uOMkTq+LG1O3cUfmLYz4Rni/ad+s9jtGr6uP7mHrlK/7/D5+X/EcfW47uzNvZV3yRIHQ\nqDXw7dVf4ea0G+h2WvnPol8GvzN7mw9yoPUICSFx7M68FZfPxVt1e8ZtX2dv5PnqVwlVhfCt1U/M\nqkPR5sSAmHOi83Rwmcfv5WDrEXRKLTen3TDu4UCjVPPlFY+wIWEtMjK3pu/g5rQbJuw3LiSGm9Ju\noM9t58Pmg5Me2y/72dO0D4Wk4PEVn0dC4pXat/D5fTPGLVgYGkc7j4VEDuP0usiNCKRFKyQFn8u7\nn7uzbiPVkEyjo5n3Gj/kx8W/4p9O/CcfNB9kYCQw8eDyuii3nuP56tf4Tdkz+GQvjy5/iPtz7kKr\n1KBWqtEoNZgis7G7HRO8Jcx9dQATUrLDNQZWxiynfbCTloHxnlYen4enyv9In9vOXVm3LVpBCCA5\nLAmFpKBloD3ow3I1lo7B+fIx1ySikMs7JgotjlbUC0WE1kj/yMC46173aBllomHuM/YrYwL3uuer\nXwOYsXRsMi63AXSWMZ10QyrmvjpkZFE6toTpGwz81iMNgd+1w91Pha2aPU37OdZxfpwxZoA+malz\nYcxy/LKfczbzuOVH208AcG/2nTy+4vM8vuLzQbPeP1e/iizLDI0M817jh+iUWu7IvIUNCWtZZsyg\ntKeCSls1pyxnON11hvTwVHZn3goExjVJYQmLRhAaY0VMPiN+D6csZ3i+6lXUChVfWvEw6mkM+ZUK\nJenhKXQMWiZknDf1t1DvaGR5lGlCg51ofRTfWPUlMkcz969FovVRc5qkSA5L5Lb0ndjdDt6seze4\n3Ol1cba7HP2oSD8duZHLiNAaqbRV81zVyzxX9TJv1L2LX/bzubz7+WrhY0IQWqQs6WmN3n43YXo1\nGvWVS6EfEwASQuMoMGZjtjVwtqeMG1O2XtL+Wgba2Nt0YMYOLVWjfkIrYwoo6SnH6rRN6lsxxonO\n4uD/97UcYkPC2lmn6u1tPoBjJDC7/pL5DTLD06fs0vNyzVsMe518Ovcelhkz2NdyiOKuUn5X8Se+\ntfqJcSm2U+Hyunmq/Fmq+2rJicjiqysfm3TQHaE1cn3KFt5v2s9py1m2Jk+dMtvkCBhAZ4SnkRAa\nzzdXPc7Pzv6W56pf4bW6t1keZWJlbAE3GMff3BsczTxZ+jROr5P7c+5i5zTp6JnGNE5Yimh0NI+7\nWR1qO8artX/h9oyb2Z0VuIH3OK0MeGk1kjkAACAASURBVAbn5AGSZkjBqDFQYasal43k9Xs50HIE\njVLD9cmb0Cq1HGg9zNGOk+xI3UZcSCyyLFPSU8GJziJuz7yJjPDxN82W/jZ+dva3SJKC/7Pl+5Pe\nVF6tfZtaewOrYwvZdVFt+4UoFUruzb6TTGM6fzr3Mn889yInOosw99URqY3gW6ufIFxj4GxPOScs\nRWxL3kSmMY1eV19wBu3LhV+YdfpvclgiaYZkKm1mHO4BYjFw2nIGx8gAN6VdP6mnhFKh5JH8B9md\nuWta4enW9B2ctBTzYcshNiWuI0Y//ns/liW0NWkDa+IK2Zy4nmOdpzjcfnzOLToFVwZL7zAKSaJX\nbgfAdIEwo1Qo2ZWxk10ZOxn2DFNjb6Csp5Iz3WW8Vb+Hdxo+ICk0nvYhS1DkjdJF8uXCLwRbsV9I\nXmQOZ7rLqB41VhxjrGxtsjr9LUkbKOmp4OOOU8EBvizLvGB+ncb+FtbFr2ZX+o7Ld0LmAY1STWJo\nPG0D7Qx7hlEpVKy8CkvHYIbyMSEKARCpMyI7ZPpHBoKeamMTDEmGmTvSXExhTAEvmt9g0DNEmDqU\ngln69M0nkiSxI3Ubz5x7AQj4ngiWJn39gd91tDHw27e5zpdtvWR+ncTQeMI1Bsx9dWQZM4gPmZgN\nVxiTzzuNeym3VbEuIWDGO+gZ4mxPeWDC6QLfxeVRJn5+9n84aSnGqA1Hp1cz5BnmnqzbgxlyD5ru\n5d9O/4yXzG8w5BlGp9TyxeWfWzQlw1NRGL2c/S2HecH8OjIyD5nundXYLsuYQZ29keb+1nEm9fvH\nsoTSrp9qU8Ec2ZWxk5KeCo52nEStUNMy0E5jf3PAyy3hOtQzZEQqJAV/tfrLNF7QaEciML6Z7aS3\nYGFYsplCsizTO+AiyqC9oscdM5mOD4ljU+paJCSKR2dGL4XXa9+hpKeC013/l707j4+zPO/9/5lF\nGm0jaSSNJEvebfm2zW7sgFliE0hCCMYlBdKEJCShTX4JoXRNOD2cpvx6EpqekyY/Qk9LCJTTpgnZ\ngIQSEgIBEmIgxoBZbG6877b2xVpneX5/zCLJHsmSPJJm+b5fL15I88yMLgvzPM9c93Vf16vjPu/t\n9h0UeYqS27lax2k2HYlGeOnoFkq8xZxXezZHeo+dtHoxlua+Fn69/zcEfJV81PwhA5FBHtz2/ZSV\nEFtb3uTV5tdZXLGQdzeujVf4fJS/PP8WAH6664mTpj6d6EDPIe5+7du83bGDs2pW8vlzbh73hvvS\nxgtxu9w8c/D5cXsLJbbaJT5kLSifx1+d/wXWz72YYm8xW5q38m9vfY+bf/rXfOvV+/jNwU1sOfYa\n33r12wxGBvnEig+PmxCC4ZvEkftqE42pIdbLKFHRtSvZqHLhuO85ktvl5qyalRwP9fLjHT9LVjht\nPvYaXUPdXNJwASUFJXjcHq5Z/AGiTpTHdv+S1v52/uX1f+M7b/4Hb7Zt5/979dujJrcd623mn7fe\nz0BkkP5wP88fOnn60Rut2/jNoU00lNbz8RU3TCiheG7wTL605lYay+ZgO3ZSVlDKref9CYGiSjxu\nD9c3bQTgR+/8lMHIEN9+4985HurluqZraBqjqd1YLpyzhqgT5fdHtxB1ojy1/zk8Ls+4/81cLtcp\nK5GKvD6uXfpBwtEwP9kxul9Q1Inyi71P43a5ef+CWJLsmiVXUuwt5vE9TyarSiSztHUPEPAXsqMz\nlphpCqTe6ldSUMK5wTP5xMoP89WL7+D6po3UltRwML6F9MoF7+HPzvssX77wr1MmhIDkCpsd8f9b\nOBpmZ+ee5AeLE62oWkbAV8nLx16lY6CTZw48z//8/T/x0tEtLCifx43Lr8+KMdjz/XMZioY42tfM\nGVUmaxMnwyPpx04KTWdT/2yQaDbdMTg8gawlnhSqL5t8UqjC508uXKypOy9jtmmdV3sWFYWx9gSq\nFMpdbfGkUFV810Gib+d5tWcTcaLc98a/86v9sVHfa8eoYmssm0PAV8m2Npu8X/79kS2Eo2Eubrhg\n1Dm8yOvjc+d8itqSGp7c9wyP2V8R8FWyfsTCUmPZHNbNvYi2gQ4GIoN82Fw75uJsJllcsYASbzEO\nDufUnMElDRNrr5HsKxTvvQmx3n2vNr/BvLIGzBjXbZk8r9ubvK9/5uDz7O7ayzx/Ix9YeDnXL9s4\nofeoL61l7ZzVyX8unLNaCaEskBlX1mnQNxhmKBRNnsRnyrHe2LaA+tJaqoorWVq5iB2du+kY6Jz0\nFKI9XfvZ0bkbiPWpGKvSoLW/ndb+Ns6uOSO5QjFeUmhbu6V7qId3N17EJY0X8Grz6zy571nOjDeB\nG89PdjxG2InwoaarWVV7Njs697D52Cv8fM+v2LDkyuTz+kL9/MA+gtfl4cbl141KGiyqmM+auvPY\nfOxVXml+/aSGkY7jYDt28qt9zyZHj19Qfz43Lr/ulKsglb4KVtWezcvHXuOdjl0pe7k4jsPe7v3U\nFFeP6kvQUFbP9cs2cl3TNRw8foTXW99ie8fbvN2xIxmH1+3lT878+IRWueeU1lHk8bFnRHO8/T0H\nOdx7lIrCcrqGunl453/xmbM+we7OvQAsrpzcjeX7F76HnZ17eO7gJg70HOLTZ9zIU/uexe1yj0qA\nnBs8kwXl83il+XXeaN1GKBrGBJZybvBMfrLjMf5l6wN86oyPsqB8Ht967TscD/Vy7dIP8sSep3jm\nwG9ZP++S5OpA1Iny6K4ncOHiU2d8lCLvxBOvtSVB/ur8W/jtoRc5o9qMWlFrCixmdd25vHzsNb62\n+W6O9TVz0Zw1vLtx7aR+JwCr687l4Z3/xYtHXmbpoXkc62vhwjmr0zIi+vzac/jtoRd4vfUtXji8\nmTX1sQ8pW45t5VhfCxfNWUN1cRUA/sIyrl70Pn6046f8bNcT3Lji+tP++ZnKGPMN4ELAAW6z1m4e\ncWwjcAcwCDxkrb3HGFMG/DsQAHzAndbaX85kzOFIlM7jgyxtLGNX5x7qS+sm1HerpKCY9fMuZt3c\ni4g60QmvztYUV1FTXI3t2JX8YLC3+wChaIhlY9zUul1u1s5Zzc/3PsX/2HQXDg4el4fVdefyh00b\nKByn7D6TzPfP5YX4ls5VWTxm1uP2UOD2pu4pFB7AhWvMfmP5InGe7Rjogvgp91h/7N5ojr+Wvq7J\nb6e9qGENB48fHrcCeKZ53V5uPvNjtPa3TarHkWSXtu7RrSjaBmJJoYvnvIsF/rk8uuvn/PbQCxS6\nC1hVm3pUt8vl4qyaFfzm0Avs7trL0srFPH/4JbwuDxfWn5xI8heW8YVz/pivb/lnuoZ6uGbJlSed\n6z+46H3s7tzHksqFY07uyjQet4dLG9fyZtt2PrriugkvaCwqP3mR9cl9z8S2/C98T1YsjGST+eVz\nueWcm+kN9WGqlk5pAqtkn5xNCrV3J/YAz3ylkNvlTm4rOb/uHHZ07uaV5tcnXd741P5ngVjTyh2d\nuxkID6RcgUyMol9e1UR1UezD6Hg9hZIjMxtW01g2h5XVhm1tlt1d+8Zd7XqzdTtvtr3NssolyUZj\nHzZ/wO6uvfxy3zNUFQeSK4QvHnmZrqEeNiy+MuWo8asXv59Xml/nsV2/GNULo3uoh/ve+I/kasCy\nyiVcsWD9pJpTr597MS8fe41nDj6fMinU3N9KX7h/zFHxLpeLef4G5vkb+OS7PoQ9sJ/XW7exs3MP\n6+dezNLKRROKw+1ys7B8Pm937KA31EdpQUlyz/lHln+IJ/c9y9aWN9ne/g67u/ZS6CmksXTOhN47\noaoowF+vvpXvvf1jtjRv5f996X8zFBnigvrzRyUhXS4X1y65im++ei9FniJuXL6B1XXn4nK5qC0J\n8u03/i/fefO7VPjK6RzsYuPiD3DF/HV0D/Xw9P7fsPnoK8km3C8e2cLR3mNcNGfNSXu4J6LQUzjm\n/wvXLv0gr7du41hfM4vKF3CDuXZKF/vSghLOqTmDLc1beeCVHwDw3hR9gqbC5XJxfdNGvvby3Xz3\n7R/x4x2PcUa1YV/PwViV0Alb6S5tvJBNR37P663buDEtEWQeY8w6oMlau9YYswJ4AFgbP+YG7gFW\nAW3AE8aYR4E/AKy19r8ZYxqAXwPLZzLujp5BHAeKAscZioZGbR2bCJfLhcc1uXL95YGlPH/4Jfb3\nHKS+rjLZT2i8n31Rw7t45uDz+AvKuLjxAi6oPz/rPoTOL28EoMBdkJzGk618Hl/KSqGByAA+jy/v\np6ZUFsUyQSPH0rf0tVJWUEppYQl9TL7P4to5a7hwzuqM+90uqVw4ajiE5J72EyuF4tvHqooDLK9q\n4kDPIbY0b+W82rPHrRI8s2Ylvzn0QnL61rG+FlbXnTtma4jq4ir+4vzP0+Y001R88v1qsbeIL665\n9bT+bLPhmiVXcs2IReSJKCsspba4hr3d+4k6UToGuvj90VeoK6nN6H562Wz5KXoHSe7J4aRQ4iQ+\nPUmh3xx8AZ+nkAvmnD/q8WN9zdQUVSWrKs4NnsUP3/kpW5q3TiopdKyvha0tb7HAP48V1cv4xd6n\nebtjZ8qTX2Lrz/KqJoq8PvyFZWNWCnUNdPNG6zYay+Ywryx2k/6++evZ1mZ5av9zyUkGJwpFQvx4\nx89wu9xcv2xj8oN6sbeIT53xEf7plX9JNoFMaCybM+YH8ZriKt7duJZnDj7P84de4vq6K2nrb+db\nr91HS38bZ9Ws4MqFJ/e6mYhFFQtYUD6PN1u309rfdlLfl5H9hCaiqijA+rkXT6kv1KKKWFJoT9c+\nTGApm4+9RkWhn5VVhkpfBV/bfDcP2Udo7W9jeaBpSvvBi7w+PnXGR1lcsZCH4yPQU/1dawos4b+/\n6y8IFFWM6hG0vKqJ2877LP+89X46B7u4fN67ee+C9UBsasuzB37HU/EeOuFohMf3PEmB28sH4/2Q\n0qnSV8EfLbuWl45u4aaVf3TKvcvjWTtnDVuat9Le38lZNSupT+NI0rn+Bv569Rf4/dFXeL1lG1ua\ntyZ/5ol/3zxuD3963mfom+RY8yxzOfAogLV2uzEmYIwpt9Z2AzVAp7W2BcAY8zRwBdAKJJZVA/Hv\nZ1RbV+w6ESlpBYcxq3XSyVQ18fzhl3i7fSfv4kze6diJCxdNlYvHfE2gqJKvXfJl3C531q6INpY1\nUOmr4Mzq5ZOqLsxEPk8hg+HUjaazdVtcOgV8o5NCkWiEtoGOKV3PE1wuFy6y8+++xGRjNSmM2D4W\nX2Ruj4+Cr/JV4nK5+NiK65nnbzyp6v1EyyoXU+gp5I22bcm+nJecYlx8TXE1K4ILaWmZ/oE1mW5R\nxQJeOrqFo73N/PbQi0SdKO9fcFnGJYpFslXOJoUSk8eq/Om/QesP9/OjHT+l0F3A6rpzkx/ke4aO\n0xvqG9UXxl9YxrLKJbzdsYOdnXsmXGXy9P7ncHC4YsE6Ar4KfrH3ad5q3X5SUijqRJMNe2uLawCo\nKapmX88BItHISUmG3+7bTNSJsnbOmuSHi6WVi1lQPo/XW95iX/cBoo7Dsb5mjvW1JP9p6Wsl4kRY\nP/fik6pDFlUs4LbzPsuuzj3JxzxuD2vqVo2b5Hj/wvfwwpHNPLH3Kc5dYPj6ln+la6ibKxe8h6tH\njAyfivVzL+b/bnuI5w5u4g+bNow6lugndDo3qBOV6Cu0p2tfskfPJfPX43F7mOdv5OLGC5LTJ06n\nJ4HL5WL9vItpCiyme6gnOfb9RGNV9iwon8cXV/8p+7r3s6r2nOTvvtJXwZr683jxyMu80bqN5r5W\nOge7eN+Cy9KyFSuVC+acf1KydSpM1VIqfRV0Dnbx3vnrTz+wE8z3z2W+fy5/uHQDh3uPsrtrL+fX\npr4pLCsozfXy23pgy4jvW+KPdce/9htjmoC9wGXAs9barxljPmmM2Unspv+Dp/ohgUAJXu/UG2kG\ng6O3hr2+N3Zz31/YjGvQxdolZ1Pmm97/TheXn8sDb/4nu4/vZjA8xJ7u/SwKzGNBQ/qSlulw4u8q\nHf71mq+Ci9O6iZ+OuCarxFdMe3/nqFiCQT8D0UGqiiszIkaYvd+VN77g1EcvwaCfwz3HiDpR5gfm\nzGpcp6K4pk+2VpNCbOdBaZGXYl/sI1PbQDsVheXJiVmFnsLkQtp4CjwFrKhaxtaWN2ntb6eupJal\n4ywGyGiL40mh11reYNOR31NdVHXKRJyITFzOJoXae2KZ/enYPmbbdxJ1ogxEBtndtY+mQOykfiw+\nZri+ZPR2qSsXXs47nbv4zpv/wZdW/+kpewt1Dfbw0pEtBIurk0mgsoJS3mp7e9SUKYg1Yu4L93Nu\n8MzkB/ma4mr2dO+jY7BzVNWC4zg8s2cTHpeHNXXnJR93uVy8d/56vvPmf/CPL3/rpHiKPEXM9Tew\nwD+Pqxe/P2XMSysXTTjhleAvLOOK+ev5rz2/5I6n/hcODh9aenVapgisqj2bR3Y+zqbDm/ngoveN\nWpne270fr8vD3HGms6XLonjiaXf3/mQyamQjwg2L388rx7bSF+5ncRpK0BvL5tDI5LagJcT6nVSd\n9Ph756/jxSMv88Sep2gd6KDEWzwtSZZ0c7vcfGzF9fTQOa3l/S6XK/Z7HyMRl6eSGV1rrWOMuYnY\nh4AuYA/gMsZ8DNhvrb3SGHMOcD8w7qzpjo6pV1sFg/6TVlv3HuoEV4TmwUPM9TfQ3x2lfwpbWyZr\nnr8R27qbrUe3EYlGWOxflFErwal+V5kgU+LyOl4GQoPJWIJBP83N3fSHBigoLsyIGGfzdxV1Yuff\nY11ttLT0YFtjfUDK3bGFhEz4/ZwoU/5unWgycWV48iirqknb+jvY0bmLd9Wtoq17gNrKWHV1JBqh\nY7CLhVMc9X5W9Qq2trxJ1IlyScO7srbyczYkFtx/sffXRJwI712wPuOnrYlkk9xNCsV7Ck3H9rFt\n7XbU18mkUHwcfd0JPXSaAov50NKr+fGOn/HtN/6dP1/1uXGbgz578HnCToTL569LJoDOqF7OS0e3\ncLDnMPPLhyfbbE9uHRse7Z74YN/a3z4qKbS/5yAHug5zbvCsk/YwnxM8gwvqz6cndJy6kiB1JbXJ\nf5cXlk3bhes98y/lN4c20RM6zsfM9axtWJOW9/W6vVzaeCGP7/kVLxzZnGzSPRQJcfD4Yeb7557W\n1qSJKikoob6klj1dewlHIyyuWDDq70dZQSkfW3E9m4+9xtKKySXVZkp9aR1n1azkjdZtQKzvT6qx\n7ploRdWyjL3ZzzGHiVUGJTQARxLfWGufAy4FMMbcRaxiaB3wy/jxrcaYBmOMx1o7+S60U9TWNYDb\n30mUSMpx8NNleVUT+3sO8uO3HgdmZtuapI/P4yPiRAhHw8lJWIORIaJOVNvHiCWEKgrLk9PHmuML\nZsGSmtkMS2bXjFSTQnoqSn/0+0d4ds8LLA4uYHAoQn1NaSz529tG1InSUFk3pSTcOv9q/vPtH+N1\ne7jqzHX4fRPrDZepCb+ZjKu6ppTiV4voDw0QKK5gw1nrk9VasxnXZGRiXJkYEyiuyUpHXDmbFEps\nH0t3pZDjOGxre4dibzGhyBDb2ywbl3wAGB5HXz9iolLC+rkXc7DnMC8efZnvvf0Tblr54ZSJlgM9\nh/ntoRfwF5RxYf3wFppEUujNtu2jkkJvt7+DC9eohsrBeCKopb+N5Qw3Ctt8NDbWPtXITLfLzSdW\nfnhSv4t08HkK+cvzP09JuZeSUHq3I13auJan9/+Gx/f8ilW151Dh83Pw+CGiTnTKqzxTsahiQfLv\nxto5Jye9zgmeyTkZ3ijvvfPX80brNgK+StY1XjTb4UjmeRK4E7jXGLMKOGytTWbijDFPADcBvcAG\n4OtAI3AB8BNjzALg+EwmhCDWK8JdHuu/NpMjbZcHmnhy3zPs7Yw1J18yYsuxZD5fvPJ0MDKUTAoN\nROLj6D3Z3S8pXQJFlezt3k8kGqE53uOwLsW9keStaakmhfRUlO5ujVV2b94VWwQuK/LS0tLDjo7Y\n42WUTXGxycU1i6+kuKCIgW6HgQlUpmbqwtZsxLXQP5/t7e/wnsZL6WwfAAYyIq6JyMS4MjEmUFyT\nla6K0pztztXePYC/pICC01gtSOVoXzMdg52srFrGkspFHDh+mK7BnuQxgLqSk6dtuVwu/shcy4Ly\neWw+9gqP7HycnZ176Bk6HhvB3r6Te177Dv+w+Zv0hwd434LRGfAVVctwu9y82fZ28rGXjmxhR+du\nFlcsGNWvJFEd1HbCBLLtHTvweQozrqN8TXE1CyrnnvqJk+QvLGPjkg/QH+7nh+88Cky+yXQ6LKqI\n/azxxpVmuiWVC/nY8uv5k7M+PubKjOQva+0mYIsxZhNwN3BLfIX32vhT7iOWOHoeuMta2wrcCyw0\nxjwHfA/4f2Y67rauAQoqOmY8MbO4YkGyUnFh+fysb7ycbxIj5wdGNJvuD8c+nKhSKCbgqyDqROkJ\nHR+uFDqhCb/klVNWk1prL7XWXk0sMbQXuJgR1aRAgzFm2vcLRZ1o8n5+f/dhYMQ4+vh9dVVxYMrv\n/76Fl3Fp49rTjDI/vWfepaypW8XFjRfOdigiOScnK4Ucx6GjZ5A51elvGLqtLbZqsLLacDzUi+3Y\nydvt73DBnPM51ttMeaF/zK01BZ4CPnPWJ/jHzXfz9IHf8PSB3wDxSSbx8bZNlYu5Yv46zqge3Uuv\npKCYJRUL2dG5m+6hHtoHOvie/QnF3iJuXHH9qOcmto+1jJhA1jXYw9HeY5xTvzK5spkPLmm8kJeP\nvcZrLW/wavMbM9pkOqGpcjEuXKyuO2/ccaWZLl1b+yQ3WWtvP+GhrSOOPQw8fMLzjwM3zEBoKUUd\nh7aePgpKuphbNmdG/98s8BSwtHIx29vfmdFta5IePk+iUmg4KTSQTAplx9ba6ZYYRNAx0EVzXyuV\nvgoK48k0yUtZU03aMdDFUPyevLn/GFCTbEWRGEdfXXRy/0WZfiurDSurzWyHIZKTJpQdyLYxkr0D\nYYbC0ZO2jkWdKBEnelq9ZBJJoRVVy+gN9fEIj7Ot3XJe7Vm0D3Sestlypa+CL625jW1tNjnZq7mv\nhTmldVw+f12yqiSVM2tWsKNzNy8efpnnDm0iEo3wmbNuOqkku7zQT4G7gLYRSaEdHTsBOKsuv06m\nbpebG5dfx1c3f5MfvPMIblyUFZSmbKg8XWpLgvzNu/58Rn+miIyvpy9EtKgTXFGWzEI/r/OCZ/F2\nxw7Orlk54z9bTk+iUiixmAPDlULZnPhPp8qieFPp/lY6BjtZVqnkZz6z1m4yxiSqSaPEq0mBLmvt\nIwxXkzrEq0mNMfcCD8SrSb3MUDXp0b5jya87wrHe1olKoXYlhUQkR50yO5KNYyTbu2M3Zyc2mf7O\nm9/lrdbtNAWWcE7wDM6qWTmpsdqDkSF2du5mblkDFb5yygv9VPoq2N72Dkf7mnFwTmoynUqFr3xK\nVRdnVq/gkZ2P89PdTwDwB0uu4owUGXOXy0VNcRUt/e04joPL5cLGk0Jn1hqY0a4ds6+utJarFl7B\nz3b/AoAzq5fP+MSHsUbBi8jsiDWZjt3gp2Py32Rd1PAuLlv+LoZ6NH0m2wwnhbR9bCyJe6sdHbsB\nqFWT6byXLdWkR3pjSSGv20so2g/eoWRSqLW/HRcuAkXp7YEpIjLbJtJTaNQYSSBgjCmPH0uOkbTW\nRoGRYyQTm8dndIwkQHtPYvLY8M1ZOBqOjXTHYXv7OzxkH+G//+4r/GTHYynfIxQNs6NjF1Enmnxs\nR8cuwk4kWbrocrlYWWXoDffx+6OvACePo0+nupIgNfHVidV153LF/HVjPremuJqByAC94VjDPdux\ni2JvMQsrZ67Bcia5Yv465pbFRtDP5NYxEclMbd3DSaHZaPTscrmoKCo/9RMl4xR5hhtNJyS2j6nR\ndEwgnhR6J74gpcljki2OxicJn1EVu9f3lBynoiyWCG4f6KDSV5FXbRhEJD9M5Kw2I2Mk0zFCMiG0\nI5aDWtBQkXx8Z9tewtEw71v6bjYufx8vH3qdn+94hl8f+C0XLT6Ps+tXJF/vOA7ffOF+Xjiwhfcu\nuZQ/Pv8juFwu9uzfA8BFi89Nvu+FA+ew6cjvefHIywCYhgWjYkn36Lo/OucaXjvyFp9d8zF83rH3\n58+rqueN1m1EfAM4Pi9tA+2saTwHt9ud0+P0xvPnl9zMd7c+wvtXXkKwbOI/K19/X1ORiTGB4pKT\ntXb24y7rwO+poMKn5IxMXKqeQv0RVQqNlKgUah2INebV5DHJFkd7j+F2uTkneCZbW9+iNNCHx+0m\nHA3TOdjFklmoLBURmW5TSXVPyxjJdIyQTNh/uAsAL07y8VcObAdgTmED9BWyOrCauhVz+MeXv8W/\nvPQf/M27/iI5AWbT4c28cGALLlz8atdv8UQK2bD4/Ww59AZFHh8BJ5h83zmeubhd7uHS8dBwLNMx\num5F6UpWLF1Jd8cgsTZOqZUS+7C548iBZMO8haULAbJ+nN5UFeHnj1d8AvqhpX/io/vy9fc1WZkY\nE+RGXEoepd/B7qO4vGHmly2Y7VAky6TaPjag7WOjlBf6ceHCwQEgWKxKIcl8juNwpLeZ2pIgjfHq\n8oKyXiDWgNrBUT8hEclJE9k+ljVjJBPau0/ePranax8Ai8qHPwDM8zdyxfx1tA108F97Yn2wj/U2\n86N3HqXYW8Rfnn8LNcXV/GLv0/z4nZ/R0t+GCSwdVTZaUlDMovh2pEJPYcasOCe2mbX2tyX7CZnA\n0tkMSUQkYxwZjE0iNNWLZzkSyTY+78nbx4YbTWv6GIDH7UneD7lwadCCZIWO/i4GIgPMKanFFynH\nibpwimKLN23xqreqoqmPoxcRyVQTSQo9CVwHMNYYSWNMrTGmlNgYyaeAncTGSDKTYyQTOnpiN2eV\nZcN7+3d37Us5deoDC6+gtqSGZw/8jh0du/m3bd9nKBriI+YPWVQxn1vP/RMqCv08c/B5AFakaOyc\n6DFUVxLE7ZrIr3T6BYtjLZ1a62XMIQAAIABJREFU+tt4p2MX/sKyae13JCKSTTqdowCsrFFSSCYn\nWSkUVqPp8ST6ClUXBdSDRbLCwe7Ymnd9aR2dx0M4A2UMejqJOtFkUqhaCU4RyUGnzGBYazcBiTGS\ndxMfI2mMuTb+lMQYyeeJj5EE7gUWxsdIfo8ZGiOZ0N4zSHlpIQXe2B+vc7CLjsFOFlUsOGnqVKGn\ngI+a63BwuOe1+zjQc4i1c9Zwft05ANQUV/GFc/+Ekvjq38qqk5NCZ1THBqs1lGbOhKmq4ipcuNje\nZuke6sEEls74xC0RkUw1WNCCK1I4oYmRIiP5xmk0raTQsERfoVr1E5IskUgKzSmtpa17gGhfGVHC\ntPa309afGEevSiERyT0TWrrJljGSENsP3NEzSGNNafKx3fGtY4vLU/eOaAos5tLGtfz20AvUFtdw\nXdM1o443lNXzl+d/nmN9rVQXn3wxmOdv5LNn3cQ8f2Ma/ySnp8DtpdJXQcdgJwDLAktmOSIRkcxw\ntLsNfP0UDzZmTHWnZI9EUmggxUh6TR8bVhkf263JY5ItDnYNVwpt3TOI0+8HjnC49+hwpZB6ColI\nDsq5et6e/hChcJSAf/jGLNlPqGLsUeR/sOQq/IVlrK47N9lweqT60jrqS+vGfP3ZwTNOI+rpUVNc\nlUwKmUDTLEcjIpIZ3jgW67NW7Zkzy5FINkrVaLo/MkChuwCPe8baJ2a84UohJYUkOxzsPoILF7Ul\nQdq6dxHtiw15OHz8CO0DHbhdbiozpHeoiEg65dwSaUfKJtP7cbvcLCifN+brirw+PrjovTk1NrUm\n3leouiigJo8iInE7OvYAMLdk7GuCyFjG2j6mrWOjnRc8m3OCZ3Je8KzZDkXklBzH4UD3EYIl1RS4\nvbR3DeD0lwFw+PhR2vrbCfgqlPgVkZyUc0mh9niT6ary2E1bKBrmQM9B5pbNoTC+upcvEkmhZZo6\nJiKSdKj/AE7UzaJKJYVk8oYrhUZPH9PksdGqiwN85qxPZMxUVpHx9ISO0zvUx5yS2K6A9p5BCiml\n2FvEvp6DdA31aOuYiOSsnEoKhaNhtra+Du5IcvvYwZ5DhJ0IiypS9xPKZUsqFgJwXu3ZsxuIiEiG\n6A8P0BlpJdpbQW1l2WyHI1nI4/ZQ4PaO2j6mSiGR7Ha09xhAslVEe/cA1eXFNJTOoX0g1mS6KkVf\nURGRXJBTSaGtLW+yuf+XFCx6g0B8HH2iyfSiMZpM57KmwGK+se5/ckb1yRPTRETy0d6u/YBDtKeS\n6nJ9iJep8Xl8yZH0Q5EQYSeipJBIFjvS2wxAfWkt/YNhegfCVJX7aCwbnixco0ohEclROZUUah+I\nNVX2Vh/lQPhtYGST6fxLCgF5t2VORGQs4WiYX+77deyb49VUlmlSlEyNz1OY3D7WF+oHNHlMJJsl\nKoXmlNbR3hNL+FaXF9EwIilUpXH0IpKjciop1DXUDYDjwM8PPE5zXyt7uvfjLyyjWidyEZG85TgO\nP3znp+zo3I27u54KpxG32zXbYUmW8nl8ye1j/aFYL0NVColkryO9x3Dhoq4kSHt3oj9pEQ2lw1Mq\nqzW0RURyVE4lhboHewDwNq9kMDLEv77+IJ2DXSyuWIjLpZt/EZF89Ysdz/K7wy/RWDqHvnfOpKZc\nTYFl6lJWCikpJJK1jvY2U1taTaGnkLZ4Uqi63EdDWV3yOVpgFpFclVNJoa7BbhwHaoZWckH9+Rzr\ni+0PXlQ+f5YjExGR2bK97R0efO1H+AvLuH7hh3GiXqor9AFeps7n8RFxIoSj4WRSSJVCItnp+FAv\nPaHjNFbEqoLak0mhIoq9xVQVBfC6PJqkJyI5yzvbAaRT52APhAup9hdzw7KN7OrcQ+tAO4vjU7hE\nRCS/HA/1cv9b38Xj8vCZs25isCP2wV1NpuV0+Lyx/kEDkUH6oomkkKrPRLLR0fgi8tzyWFKopy8E\nQHlprC/ndU3X0Bfqw+3KqbV0EZGknEoKdQ914wz5CJT7KPIW8flzb2Zbm2VxnjaZFhHJd47jUFsS\n5Noz3s/i4gU8v+8IgCqF5LT44kMcBsND9IW1fUwkm4WisSTQsupFse/DUQAKPLEk0DnBM2YnMBGR\nGZIzSaHByBBD0SGcUDlVtbEbs7qSIHUlwVmOTEREZou/sIwvrr6VYNBPS0sPXb2x5sCaPCanwxef\nNDYYGRzePqbpYyJZaUXVMv7uwi+xonEBra3HCUfiSSGvKoNEJD/kTFKoazA+eSzko7QoZ/5YIiJZ\nwRjzDeBCwAFus9ZuHnFsI3AHMAg8ZK29xxhzM/DxEW+x2lpbNt1xJlaAfQW62ZepS1YKRYboD8f6\nj6hSSCR7BUuqk0NpEtcJr5JCIpInciZ70j0UmzzmhHwU6GZfRGTGGGPWAU3W2rXGmBXAA8Da+DE3\ncA+wCmgDnjDGPGqtvR+4f8Trb5iJWHWzL+kwnBQapG9IjaZFckkoMnr7mIhIrsuZs92opJBO4iIi\nM+ly4FEAa+12IGCMSYxpqQE6rbUt1too8DRwxQmv/1vg72ci0BN7RYhMRVGq7WNKConkhLAWD0Qk\nz+RMpVBy+9iQT3uARURmVj2wZcT3LfHHuuNf+40xTcBe4DLg2cQTjTFrgAPW2qMzEah6RUg6DPcU\nGqIvFNs+puljImPLli3GEKsU8rhduOPbyUREcl3OJIUSlUKEfHi1AiwiMpuSd9LWWscYcxOxLWVd\nwJ6Rx4E/Bh6cyJsGAiV4vZ4pBxUM+nHHX19XW06wqmTK75UuwaB/tkNISXGNr6avAoCCYhd9bbFK\noXn1NRR4CmYzrFEy5Xd1IsU1OZka12Rk0xZjgHDYUZWQiOSV3EkKDY7YPqYTuYjITDpMrDIooQE4\nkvjGWvsccCmAMeYuYhVDCeuBWyfyQzo6+qYcYGL62PH49LHurj7ckciU3y8dEjFlGsV1akN9sYqz\ntq5u+kP9eF0eOtsHgIHZDSwuk35XIymuyZlMXBmePBq1xdgYEzDGlFtruxmxxRjAGJPYYvzgiNf/\nLXDjTAUbikS1xVhE8krOnPG6hhLTxwpVKSQiMrOeBK4DMMasAg5ba5OfZIwxTxhjao0xpcAG4Kn4\n4w3AcWvt0EwFmuwppMUDOQ3JRtPhQfpCA5o8JjK+emJbiRMSW4wTX/uNMU3GmAJiW4zrEk+c6S3G\nEOsppGuEiOST3KkUGurB4xRA1KsTuYjIDLLWbjLGbDHGbAKiwC3GmE8CXdbaR4D7iCWOHOAua21r\n/KVzgOaZjDUccQC0eCCnJdFTaCDeaFpNpkUmZVq2GEN6thlHHAdfoSdjqq8yJY4TKa7JycS4MjEm\nUFyTlY64ciYp1DXYjdeJNXlUUkhEZGZZa28/4aGtI449DDyc4jVbgA9Mc2ijhMKxLWPaGiCnY3Sj\n6X7qSoKzHJFIRpuRLcaQnm3Gg0MRfAWejNhSmAtbG2eS4pq4TIwJFNdkpWubcU7cFYejEXpDfcNJ\nId3si4hICuGIE5sq49ZUGZm6xPaxvnA/g5EhijR5TGQ8WbPFGGI9hbweXSNEJH/kRKVQ90APDg6e\nSOymTBMDREQklVA4qmuEnLZEpVDXYKyfobaPiYwtm7YYg3oKiUj+yYmkUMdAFwDuSOymTJVCIiKS\niqbKSDokKoUSSaGieJJIRFLLli3GUcchEnV0nRCRvJITZ7zOgdhNmTsSuylTdl9ERFLRCrCkg8ft\nwev20jUYW5RSpZBIbgjHJ1RqGIGI5JOcOON19MduyggV4QI86hUhIiIpqFJI0qXI4yPsxBqXKykk\nkhvCkVhSSIsHIpJPJrR9zBjzDeBCYnt9b7PWbh5xbCNwBzAIPGStvccYczPw8RFvsdpaW5a+sEfr\njG8fc0I+vF43LpeSQiIicrJQOEqxLyd2Tsss83kKOR7qBaBISSGRnBBSpZCI5KFT3hkbY9YBTdba\ntcaYFcADwNr4MTdwD7AKaAOeMMY8aq29H7h/xOtvmKb4geFKoeiQTyvAIiIyJlUKSbr4RvQRUqWQ\nSG4IqVJIRPLQRM54lwOPAlhrtwMBY0x5/FgN0GmtbbHWRoGngStOeP3fAn+fpnhTSvQUigwVaqqM\niIiMKRyO4vWqmlROX6LZNECxRtKL5ARVColIPppIDX09sGXE9y3xx7rjX/uNMU3AXuAy4NnEE40x\na4AD1tqjp/ohgUAJXq9nwoGP1PlaFx63ByIFFBW6CAb9U3qf6ZBJsSRkYkyguCYrE+PKxJhAcUmM\npspIOo2sFNL0MZHcEI44gCqFRCS/TKWxQnKJ1VrrGGNuIralrAvYM/I48MfAgxN5046OvimEEn/t\nQDflBX6OD8V6RbS09Ez5vdIpGPRnTCwJmRgTKK7JysS4MjEmyI24lDxKj+RUGd3sSxr4vNo+JpJr\nko2mtXggInlkIme8w8QqgxIagCOJb6y1z1lrL7XWXk0sMbR3xHPXA5tOP8yxOY5D50A35T5/bNSw\nTuIiIpJCSDf7kkajt48pKSSSC5Lbx7TNWETyyETujJ8ErgMwxqwCDltrk8vbxpgnjDG1xphSYAPw\nVPzxBuC4tXYo/WEP6wv3E46GqSgsjzUQ1QqwiIikkKgU0nVC0mHU9jElhURyghYPRCQfnXL7mLV2\nkzFmizFmExAFbjHGfBLostY+AtxHLHHkAHdZa1vjL50DNE9P2MO6BmNNpssL/YTCUQo8yuyLiMjJ\nEivAutmXdFCjaZHcE9I2YxHJQxPqKWStvf2Eh7aOOPYw8HCK12wBPnBa0U1A91CsaKmsoAzQCrCI\niKSmUcOSTomkkNvlptBdMMvRiEg6hLV4ICJ5KOvPeIlKoTJvLCmkEZIiIpKKRg1LOiW2jxUXFOFy\nqUpZJBckGk2rUkhE8knWn/ESlUIlXlUKiYjI2FQpJOmUGENfUqCtYyK5QtuMRSQfZf0ZL5EUKnaX\nAsrsi4hIamFVCkkaJbaPKSkkkjvCWjwQkTyU9We8xPaxYlcJoMy+iIikpkohSSefV5VCIrlG24xF\nJB9l/Rmve6gHFy4K3fGkkG72RUQkhXDYAXSzL+kxXCmkcfQiuUKLByKSj7L+jNc91IPfV0o0Emvy\nqJt9ERFJRTf7kk4+9RQSyTmqFBKRfJT1Z7yuwR4qiyq0B1hERMaVHDWs64SkQaWvAhcugqXVsx2K\niKRJOBKrKNV1QkTyiXe2AzgdQ5EhBiIDBIrLNS1ARETGlawU0nVC0qDCV86da7/E0rlz6Wjrm+1w\nRCQNwrpOiEgeyu6kUDQEQG1pjVaARURmiTHmG8CFgAPcZq3dPOLYRuAOYBB4yFp7T/zxG4EvAmHg\nb621j093nMltAbpOSJpUF1fhdXtmOwwRSZPh7WOuWY5ERGTmZPWdcVlBKbed91k+fOYG3eyLiMwC\nY8w6oMlauxa4Gbh7xDE3cA9wFfBuYIMxZq4xphr4MnAJcDWwcSZiVUWpiIiMR73nRCQfZXWlEMCy\nwBLKi/yEwkcB3eyLiMywy4FHAay1240xAWNMubW2G6gBOq21LQDGmKeBK4B+4ClrbQ/QA3xmJgLV\nzb6IiIwnrEbTIpKHsj4plKCbfRGRWVEPbBnxfUv8se74135jTBOwF7gMeDb+vBJjzM+AAPB31tqn\nT/WDAoESvN6pb9Up9MUueTXVpQSD/im/TzplShwnUlyTk4lxZWJMoLgmK1PjylX6PCEi+ShnkkLK\n7IuIZIRkIwZrrWOMuQl4AOgC9ow4Xg1cCywAnjHGLLDWOuO9cUfH1Jv5BoN+uroHAOg9PkhLS8+U\n3ytdgkF/RsRxIsU1OZkYVybGBIprsiYTV6Ynj7Ku95w+T4hIHsmZpFBIIyRFRGbDYWKVQQkNwJHE\nN9ba54BLAYwxdxGrGCoGNllrw8AuY0wPEASapzPQsBqIiojMuJG954wxK4gtFKyNH0v0nlsFtAFP\nGGMeJbbN+MvA+UAZcCcw7UmhsCqFRCQP5U5SKBwBlNkXEZlhTxK7Wb/XGLMKOBzvFQSAMeYJ4Cag\nF9gAfB3wAQ8aY75GbPtYGdA63YFqW4CIyKzImt5z4XAUF+Bxa/FARPJHDiWFdLMvIjLTrLWbjDFb\njDGbgChwizHmk0CXtfYR4D5iiSMHuMta2wpgjPkx8GL8bW611kanO9awpo+JiMyGrOk957hcFHjd\n1NaWT/k90i1TtwYqrsnJxLgyMSZQXJOVjrhyJikU1vYxEZFZYa29/YSHto449jDwcIrX3AvcO82h\njaJKIRGRjJCxvef6B8J4Pe6M6TGVC/2uZpLimrhMjAkU12Slq/dcziSFQloBFhGRcSQbiCopJCIy\nk7Km91woEtU1QkTyTs4khRKN4XQiFxGRVJKVQlo8EBGZSVnTey4cjuoaISJ5J2eSQsOVQmoMJyIi\nJwurUkhEZMZlU++5UCRKsS9nPh6JiExIzpz1hkdITr25nIiI5K5QJIrH7cLt0uKBiMhMypbec+Fw\nlIISLRyISH7JmbNesleEKoVERCSFUDiqJtMiIjKmcCRKgVefJUQkv+TM3bGmyoiIyHjCEQevekWI\niEgKjuMQCkd1nRCRvJMzZ71kTyElhUREJIVQOKJrhIiIpBSJOjjos4SI5J+cOeuFIlFcLvC4c+aP\nJCIiaRSOOJoqIyIiKQ2FIgCqFBKRvJMzZ72wekWIiMg41FNIRETGol0HIpKvJjR9zBjzDeBCYqMi\nb7PWbh5xbCNwBzAIPGStvSf++I3AF4Ew8LfW2sfTHPsooUhUK8AiIjKmUES9IkREJLXkJGNdJ0Qk\nz5zyrGeMWQc0WWvXAjcDd4845gbuAa4C3g1sMMbMNcZUA18GLgGuBjZOQ+yjhMNRvMrsi4jIGFRR\nKiIiYxkKxScZ6zohInlmIpVClwOPAlhrtxtjAsaYcmttN1ADdFprWwCMMU8DVwD9wFPW2h6gB/jM\ntEQ/giqFRERkLJGoQyTqKCkkIiIphcKxnkL6PCEi+WYiSaF6YMuI71vij3XHv/YbY5qAvcBlwLPx\n55UYY34GBIC/s9Y+Pd4PCQRK8Ho9kwp+pEjUoaSkgGDQP+X3mA6ZFg9kZkyguCYrE+PKxJhAccnw\nzb62j4mISCpD8Z5Cuk6ISL6ZUE+hE7gSX1hrHWPMTcADQBewZ8TxauBaYAHwjDFmgbXWGetNOzr6\nphBKTDDoZygUwQ20tPRM+X3SLRj0Z1Q8kJkxgeKarEyMKxNjgtyIS8mj06cGoiIiMp6wrhMikqcm\nkhQ6TKwyKKEBOJL4xlr7HHApgDHmLmIVQ8XAJmttGNhljOkBgkBzesI+WUg9hUREZAyh5Aqw6xTP\nFBGRfDSUrCjVdUJE8stEkkJPAncC9xpjVgGH472CADDGPAHcBPQCG4CvAz7gQWPM14htHysDWtMc\ne5LjOIQjjvYAi4hISkOheK8ILR6IiEgKqigVkXx1yqSQtXaTMWaLMWYTEAVuMcZ8Euiy1j4C3Ecs\nceQAd1lrWwGMMT8GXoy/za3W2uh0/AFgxAqwTuIiIpJC8mZfiwciIpKCrhMikq8m1FPIWnv7CQ9t\nHXHsYeDhFK+5F7j3tKKboCGdxEVEZBxaPBARkfGEQqoUEpH8lBNnveQISZ3ERUQkBY0aFhGR8YQi\nmlIpIvkpJ856icy+TuIiIpLKkHpFiIjIOIZCqigVkfyUE2e9UEQ3+yIiMjZtCxARkfGop5CI5KsJ\n9RTKdMmpMjqJi4jMOGPMN4ALiQ0cuM1au3nEsY3AHcAg8JC19h5jzHrgR8Bb8ae9Ya29dTpjDIW1\nLUBERMaWvE5o8UBE8kxOJIU0QlJEZHYYY9YBTdbatcaYFcADwNr4MTdwD7AKaAOeMMY8Gn/pc9ba\n62YqTm0fExGR8ejzhIjkq5w46w1PlXHNciQiInnncuBRAGvtdiBgjCmPH6sBOq21LdbaKPA0cMVs\nBJm8TqhSSEREUtD2MRHJVzlRKaTtYyIis6Ye2DLi+5b4Y93xr/3GmCZgL3AZ8Gz865XGmJ8BVcCd\n1tpfneoHBQIleL2eKQUZ2tUGQHWghGDQP6X3mA6ZFMtIimtyMjGuTIwJFNdkZWpcuUgVpSKSr3Ii\nKTTcaHpqHxZERCRtkiWb1lrHGHMTsS1lXcCe+PEdwJ3AD4HFwDPGmKXW2qHx3rijo2/KQSVWgPv7\nhmhp6Zny+6RTMOjPmFhGUlyTk4lxZWJMoLgmazJxKXl0+tR7TkTyVW4khZIj6bV9TERkhh0mVhmU\n0AAcSXxjrX0OuBTAGHMXsNdaewj4Qfwpu4wxR4FGYkmjaaFRwyIisyc7BhLo84SI5KfcSArFM/sq\n9xQRmXFPEqv6udcYswo4bK1NLm0bY54AbgJ6gQ3A140xNwJzrLX/2xhTD9QBh6YzyFBE24xFRGZD\ntgwkUKNpEclXOXHWG1IDURGRWWGt3QRsMcZsAu4GbjHGfNIYc238KfcRSxw9D9xlrW0FfgasM8b8\nFvgp8LlTbR07XYmKUt3si4jMuKwYSKAepSKSr3KjUiikSiERkdlirb39hIe2jjj2MPDwCc/vIVY1\nNGM0fUxEZNZkx0CC+HWivr6CsuKCKb3HdMjUflGKa3IyMa5MjAkU12SlI67cSApFtAIsIiJjG9I2\nYxGRTJGZAwninye6OnrpP54Zw2tyoQn6TFJcE5eJMYHimqx0DSTIiaRQooGoyj1FRCQV9YoQEZk1\nWTGQIKSBBCKSp3LirKebfRERGY+myoiIzJongesAxhpIYIypNcaUEtta/JQx5kZjzF/Fj8/MQIJw\nBI/bhdul64SI5JecyKIkpo+pV4SIiKSSbCA6xV4TIiIyNdkykGAoHFWVkIjkpZzYPqZKIRERGU/y\nOqFKIRGRGZctAwnUikJE8lFOnPk0kl5ERMajxQMRERlPOBzVNUJE8lJOnPmGNJJeRETGkUgKebR4\nICIiKQyFI6oUEpG8lBNnvrBWgEVEZBxD4QhejxqIiohIaiH1FBKRPJUTZ74hNZoWEZFxhMJRXSNE\nRGRMIVUKiUieyokzn3pFiIjIeELhiK4RIiIyplilkKpJRST/5MQd8vBUmZz444iISJoNhdRAVERE\nUos6DuGIo88SIpKXcuLMNxSK4HG7cLuV3RcRkZNp+5iIiIwl0Z9UPYVEJB/lxJkvFNHNvoiIjE3b\nx0REZCzhiHYdiEj+yokzn7YFiIjIeFQpJCIiY1F/UhHJZ96JPMkY8w3gQsABbrPWbh5xbCNwBzAI\nPGStvccYsx74EfBW/GlvWGtvTWfgI4XDSgqJiMjYhnSdEBGRMYTilUJaPBCRfHTKpJAxZh3QZK1d\na4xZATwArI0fcwP3AKuANuAJY8yj8Zc+Z629bnrCHm0oHMHrUT8hERE5WSQaJRpVA1EREUlNlUIi\nks8mcua7HHgUwFq7HQgYY8rjx2qATmtti7U2CjwNXDEtkY4jFI5S4PXM9I8VEZEsEA47gG72RUQk\ntXAkdp1QpZCI5KOJbB+rB7aM+L4l/lh3/Gu/MaYJ2AtcBjwb/3qlMeZnQBVwp7X2V+P9kECgBO8U\nEzuhcIQiXzHBoH9Kr59OimniFNfkZGJcmRgTKK58p20BIiIynmSlkK4TIpKHJtRT6ATJfVrWWscY\ncxOxLWVdwJ748R3AncAPgcXAM8aYpdbaobHetKOjbwqhxAyForhwaGnpmfJ7TIdg0K+YJkhxTU4m\nxpWJMUFuxKXk0enRtgARERlPYvqYRtKLSD6aSFLoMLHKoIQG4EjiG2vtc8ClAMaYu4C91tpDwA/i\nT9lljDkKNBJLGqVVNOoQUa8IEREZg0YNi4jIeELJ64R6lIpI/pnIHfKTwHUAxphVwGFrbXJ52xjz\nhDGm1hhTCmwAnjLG3GiM+av48XqgDjiU9ugZsS1AmX0REUkhUSmk64SIiKQyXFGqHqUikn9OWSlk\nrd1kjNlijNkERIFbjDGfBLqstY8A9xFLHDnAXdba1ngvoe/Fx9UXAp8bb+vY6dAKsIjI7DLGfAO4\nkNh14DZr7eYRxzYCdwCDwEPW2ntGHCsG3gT+3lr74HTFp14RIiIynnBi8UCVQiKShybUU8hae/sJ\nD20dcexh4OETnt9DrGpo2qlXhIjI7DHGrAOarLVrjTEriPWYWxs/5gbuAVYBbcATxphHrbUH4y+/\nA2if7hiHe0XoZl9ERE6W3D6mzxMikoey/synFWARkVl1OfAogLV2OxAwxpTHj9UAndbaFmttFHga\nuALAGLMcWAk8Pt0B6johIiLjGa4U0nVCRPJP1p/5NC1ARGRW1QMtI75vYXg4QQvgN8Y0GWMKgMuI\n9ZgD+DrwFzMRYFgrwCIiMg5dJ0Qkn01lJH1G0QqwiEhGSe7RstY6xpibiG0p6yI2gdJljPkE8IK1\ndo8xZsJvHAiU4J1CE9DiY8cBqKwoIRj0T/r10ynT4klQXJOTiXFlYkyguCYrU+PKNfo8ISL5LPuT\nQsrsi4jMpsMMVwYBNABHEt9Ya58DLgUwxtwF7AWuBRYbY64G5gKDxpiD1tqnxvtBHR19Uwqwrb0X\ngMGBIVpaek7x7JkTDPozKp4ExTU5mRhXJsYEimuyJhOXkkenR9OMRSSfZX1SSHuARURm1ZPAncC9\nxphVwOH4sAEAjDFPADcBvcQGEHzdWvvQiON/B+w9VULodGgFWERkdmlKpYhI5sr6M58qhUREZo+1\ndhOwxRizCbgbuMUY80ljzLXxp9xHLHH0PHCXtbZ1pmNUrwgRkdkzckolcDOxa0XiWGJK5VXAu4EN\nxpi5I14+Q1MqHUCVQiKSn7K+UiikSiERkVllrb39hIe2jjj2MPDwOK/9u2kKK0nXCRGRWTVqSqUx\nJmCMKbfWdjNiSiWAMSYxpfJBTakUEZkZWX/mS57EldkXEZEUVFEqIjKrsmZKpSqFRCQfZX2lkLYF\niIjIeJK953SdEBHJBBk3pdJbEHtNXdBPsKZ00q+fTpnaRFxxTU4mxpWJMYHimqx0xJX1SSGVe4qI\nyHiSlUK6ToiIzIaMn1KcycyBAAAgAElEQVTZc3ww9u/ufrxOdErvMR1yYTLeTFJcE5eJMYHimqx0\nTanM+qRQojGcKoVERCSVcFjXCRGRWZT5UyoT28c8rlM8U0Qk92R9UkgNREVEZDyqKBURmT3W2k3G\nmMSUyijxKZVAl7X2EYanVDrM0pRK9SgVkXyWA0mhCKCTuIgM+9a3voG122lvb2NoaJD6+gbKyyv4\n6lf/17iv+/nPH6O0tIx16y6boUhlJoTUQFRERhh5jRgYGGDhwgUUFZXqGjGNMn1KZbLRtBYPRIT8\nu05kf1Iovn1M5Z4iknDrrX8OxE7MR48e4NOf/vyEXnfVVRumMyyZJaoUEpGRRl4jdu/exZ13/o8J\n9WTQNSJ3hcJR3C7wuPV5QkTy7zqR9UmhcLLcc/KTBkQkf7zyyss89NB36evr4wtf+HNefXULzz77\nNNFolLVrL+bTn/4M999/L5WVlSxatISHH/4hLpebffv2sH795Xz605+Z7T+CTJFGDYvIqegakd/C\nkSherweXS0khEUktl68TWZ8UCmkkvUjG+uGvd7L57ea0vuea5bXc8J6lU3rtrl07+f73H6awsJBX\nX93C//k/38HtdnPDDRv58Ic/Ouq527a9xfe+9xOi0SjXX78ho0/kMj5VColkrky6Tugakb9CYUef\nJUQylK4T0y/7k0JhTQsQkYlZurSJwsJCAIqKivjCFz6Dx+Ohs7OT7u7uUc81ZjlFRUWzEaakWViL\nByIyAbpG5K9QJEqhrhEicgq5ep3I+qSQbvZFMtcN71k65aqe6VBQUADA0aNH+MEP/pMHHvhPSkpK\n+PjHbzjpuR6PtqTmCi0eiGSuTLpO6BqRv8LhqD5LiGQoXSemX9af/bQtQEQmq7Ozk0AgQElJCda+\nzdGjRwmFQrMdlkyTUCR2s69eESIyEbpG5J/YdSJ7PsCJyOzKtetE1mdSkkkhZfdFZIKampZRXFzC\n5z73aZ5++kk2bvwQX//612Y7LJkmWgEWkcnQNSL/6DohIpORa9eJnNk+5lWlkIic4KqrNhAM+mlp\n6WHVqtWsWrUaiJVz/tM/3TPuaxPPBXj88aenNU6ZXrFeEVoBFpHRRo4O1jUivyUqSkVERsqX60TW\nn/2SvSJ0IhcRkRRC4aiuESIikpLjOITDUQoLtHggIvkp6++Sw5EoXo8bt3pFiIhICpoqIyIiY4lE\nHRzUn1RE8lfWn/1C2gMsIiLjUK8IEREZS7I/aYGuEyKSn7L+7BeKRCnUSVxERMYQikQp0LYAERFJ\nIdGfVIsHIpKvsv7sFwpHVe4pIiJjCocdbR8TEZGUEpVCGkggIvlqQtPHjDHfAC4EHOA2a+3mEcc2\nAncAg8BD1tp7RhwrBt4E/t5a+2Aa404KRaKUFBVMx1uLiEiWi0SjRB1HK8AiIpKSKoVEJN+d8uxn\njFkHNFlr1wI3A3ePOOYG7gGuAt4NbDDGzB3x8juA9rRGfIJwWA1ERWTYZz/7Kd5+e/uox/71X+/h\n+9//7knPfeWVl7njji8CcPvtf3HS8Z/85Afcf/+9Y/6snTt3sH//PgC+/OX/xuDgwOmELtMg2StC\nK8AiEqfrhIwUijiAJhmLyLB8u05M5Ox3OfAogLV2OxAwxpTHj9UAndbaFmttFHgauALAGLMcWAk8\nnvaoRwhF1EBURIa9973v59e//tWox5599tdcccX7xn3dP/zDP036Zz333K85cGA/AHfeeRc+X9Gk\n30OmVzh+s6/rhIgk6DohI4W1fUxETpBv14mJbB+rB7aM+L4l/lh3/Gu/MaYJ2AtcBjwbf97XgS8A\nN00kkECgBO8UTsbhiEOB10Mw6J/0a2dCJsaViTGB4pqsTIwrE2K64YYP8ZGPfIQvf/m/A/Dmm2/S\n0FBPZ+cx7rzzbygoKKC8vJxvfvObVFaW4PMVEAz6ueCCC3jppZd44YUX+OpXv0pNTQ3BYJB58+YR\nCBTzpS99iWPHjtHX18ett95KQ0MDjz32CL/73XMsXjyXP/uzP+Oxxx6jp6eHv/mbvyEUCuFyufjK\nV76Cy+Xi9ttvZ968eVhrWbFiBV/5ylcy4veVDpPdYmyMKQEeBOr4/9m77/C4qzP9/+8ZjXrvxZa7\nfNwrBtMNphgwEFoSShJaCAnJJrvZ726SHyQhjU0hEEJCDaEkARJCr6GagLHBNi6y5WPJTZYlq/c+\n5ffHjGQVy5aE5Rlb9+u6uJDmU+bWWNbxPDrnORCFf4nxyyORTb0iRKSvpUvP4etfv4FvfOM/AP84\nkZ6ezq5dO7n11v8lPDyc+Ph4fvKT/+t13QUXLOWVV95mzZqPueeeO0lJSSU1NY2cnDG43W5+/vMf\nU1lZQWtrK9dffxNZWdm88MKzrFjxDsnJyfzwh9/n8cefpqmpkTvu+AmdnZ04nU6+973bAuPFj8nJ\nGUNRUSFTpxruvPNXwXh5Rp1OLR8TkT5G2zgxqJ5CfTi6PrDW+owxXwEeAeqBnYDDGPNl4CNr7U5j\nzKBuWlvbMuQgHq8Xr9ffK6KysnHI14+09PT4kMsViplAuYYqFHMdKNOzRS/zacWmw/o88zNmc+mU\n5Qc5I4LMzGxWrPiIGTNm8dprr7Fkydns2VPOD35wOzk5Y/jpT3/IK6+8SUxMDO3tnVRWNuLz+ais\nbOSXv/wV3//+j8nLm8p///d/kJKSwY4de5k79zjOO285e/eWcNtt3+ORR/7CokWLWbJkKdnZE/F4\nvFRVNXH33b/hnHMuYOnSc3j33bf4zW/u4oYbvkZ+fj633vpTkpNTuOSS82loaKC93XGQr2O/UC4e\n9VxibIyZjn88ODFwrGuJ8QKgGnjNGPM8cDKwxlr7K2PMeOBNYGSKQvrHvkhIC8Y4kZycQk7OGLZs\nye8eJ84+exmNjY386Ec/6x4nVq/+iJiYmH7XP/DAvdx220+7x4mcnDE0NjZw/PGL+40TJ5xwIkuW\nLGXGjFnd1z/88P0sX35x9zjxyCMPcsMNX8PaAm6//Re9xoke/+yWEbL/lwcaJ0RCkcaJkR8nBlMU\nKsU/M6hLDlDW9Ym1dgVwKoAx5g78M4YuASYZY5YDY4F2Y0yJtfatz5y4h+4f4tpqWER6OPvsZbz9\n9pvMmDGLd955h3vvfZiiom388pc/w+PxUFq6l4ULFx3wh3hZWRl5eVMBmDdvAe3t7cTHJ1BQsJkX\nX3wWh8NJQ0P9gM9tbQE33/xNABYsOI5HH30YgDFjcklNTQMgLS2dxsZGIiISBrzPUaTXEmNjTLIx\nJsFa20CPJcYAxpi3gbP6bDyQC5SMVLjunkLh+se+iOyncUK6dDWaVk8hEelpNI0TgykK/Qu4HXjA\nGLMAKLXWdk8HMMa8hn+JWDNwIXCntfapHsd/DOw63AUhUK8IkVB36ZTlh5jVMzJOP/0MHn/8Ec4+\n+1wmTJhAQkICd9zxU37967uZMGEiv/3tLwe81unc//PE5/P/jHnzzddpaGjgD394mIaGBm688UsH\neXZH93WdnW4cDv/9wsJ6F6+7zjkGDHeJMcaYlfh/cTCob5LhLDNuaPcA/nEiFGdchWImUK6hCsVc\noZgJ+uf6WvqVwJVHPMell17IZZddxhVXXMKECROYPHkMN930ZR588EEmT57MT37yE+Ljo3otM3Y4\nHKSnx+Pq0bYgJiYClwtWrXqPjo5W/v73p6mrq+Pyyy8nPT2eqKhwEhOjSU+PJyzMSVpaHC5XGCkp\nsaSnx+PztRIe7iIlJbb7ecBfoPD5fCH753gs0S+ZRUKb3k+M/PuJQxaFrLUrjTFrA/949wK3GGOu\nBeqttc8BD+EvHPmAO6y1VYcl2SDs31VGRSER2S8mJpbJk/N4/PE/c8klFwHQ3NxEZmYWjY2NrFu3\nlsmT8w54bVpaOsXFu8jNHc+nn65l5szZ1NXVkZ2dg9PpZMWKd+js7ATA4XDg8Xh6XT99+gzWrVvD\n2WcvY/36tUybNn1kv9jQc8glxj2On2SMmQf8xRgz11p70JFtOMuMK6qaAH9PoaNhyWUoUK6hCcVc\noZgJQi/XhAmTueeeP3DJJRdRWdlIQ0Mj4eHx7NhRyocffkROznhiY1v6LTNOSUlj7dpN5OaO54MP\nVjJz5mz27NlHcnI61dXNvPDCS7S1tVNZ2Uh7u5uamiYqKxu7lxnn5U3jzTffC/wW+n2mTDHU1DTj\ndnu7X5+u5seDfb1UPBo+bUkvIgcymt5PDKqnkLX2e30e2tDj2LPAswe59sfDSjYI4S4nDgckxUWO\n1FOIyFHq7LOX8bOf/Yjf//5uGhs7ufTSK/j6128gN3ccV1/9ZR555EFuuukb/a676aZvcOut/0tW\nVjYZGZkALFlyJt/73n+xZUs+F1xwERkZGfz5zw8xd+587r77172mjd54483cccdPeeml53G5wvn+\n92/D7XYfsa87CIa8xNgYsxCosNbusdauN8a4gHSg4nCHiwz85jdR44SI9KFxQmD/DCGNEyLS12gZ\nJxyhsoShsrJxWEGKyxuZPiWd5sa2wx3pMwu134hBaGYC5RqqUMwVipng2MiVnh4fsp1GjTEnAbdb\na88OLDG+x1p7So/jPZcYrwaWAF8Cxltrv2OMyQQ+ASZYa70He67hjBM+n49d+xqZNz2L+rqhzzQa\nScfC9+aRpFyDF4qZQLmG6lgZJ46k4YwTXq9/nDhudg411U0jEWvYjoXvzSNJuQYvFDOBcg3V4Ron\nhrP7WEgZlxlPTFR4SBaFRESOdcNZYmyMuR/4kzHm30A0cMuhCkLD5XA4mJidoF4RIiJBZIy5C1iM\nfyz4trX2kx7HLgZuBdqBp6y19xpjYoBHgUwgCviptXZEdql0Oh1MykkgzKm6moiMTkd9UUhERIJr\nqEuMrbWtwFVHIJqIiASZMeZ0IM9ae6IxZjr+PnMnBo45gXuBBUA18Jox5nngZGCNtfZXxpjxwJvA\niBSFRERGO3VUExERERGRkbIUeB7AWlsAJBtjuvZQTgPqrLWVgRmjbwNnWWufttb+KnBOLlBypEOL\niIwWmikkIiIiIiIjJQtY2+PzysBjDYGP440xecAu4Azgva4TA0uTxwKD2o86OTkGl2v4y4VDcRe3\nUMwEyjVUoZgrFDOBcg3V4cilopCIiIiIiBwp3c17rLU+Y8xX8C8pqwd29jl+kjFmHvAXY8xca+1B\nG0nX1g5/Q4FQbCQbiplAuYYqFHOFYiZQrqEaYqPpAY9p+ZiIiIiIiIyUUvwzg7rkAGVdn1hrV1hr\nT7XWLsdfGNpljFlojMkNHF+P/xfZ6Ucws4jIqKGikIiIiIiIjJR/AZcDGGMWAKXW2u5fbRtjXjPG\nZBhjYoELgbeA04DvBo5nAnFA1ZEOLiIyGqgoJCIiIiIiI8JauxJYG+gPdA9wizHmWmPMJYFTHsJf\nOPoAuMNaWwXcD2QYY/4NvALcEmhELSIih5l6ComIiIiIyIix1n6vz0Mbehx7Fni2z/mtwFVHIJqI\nyKjn8PkO2q9NRERERERERESOQVo+JiIiIiIiIiIyCqkoJCIiIiIiIiIyCqkoJCIiIiIiIiIyCqko\nJCIiIiIiIiIyCqkoJCIiIiIiIiIyCqkoJCIiIiIiIiIyCqkoJCIiIiIiIiIyCrmCHeCzMMbcBSwG\nfMC3rbWfBDnPLOAF4C5r7b3GmFzgCSAMKAO+ZK1tP8KZfgWciv/P+g7gkxDIFAM8CmQCUcBPgQ3B\nzhXIFg3kBzK9HexMxpglwD+AzYGHNgG/CnauQLargf8B3MAPgY3BzmWMuQH4Uo+HjgNOBu7D/3Ni\no7X260cyUyBXHPA4kAxEArcD+4KdazTQODGoTBonhpZN48Tgs2mcGHwujRNBEkrjRCiOEYFcGieG\nlk3jxOCzaZwYXKYRHSOO2plCxpjTgTxr7YnADcA9Qc4TC/we/1/8Lj8B/mCtPRUoAq4/wpnOAGYF\nXqNlwN3BzhRwIbDGWns68HngtyGSC+BWoCbwcahkWmGtXRL471uhkMsYkwr8CDgFWA5cHAq5rLV/\n6nqtAvkew/99/21r7clAojHmvCOdC7jWH8+eAVwO/C5Ech3TNE4MKpPGiaHTODEIGieG7Fo0Thxx\noTROhOIYEcilcWLoNE4MgsaJIbmWERwjjtqiELAUeB7AWlsAJBtjEoKYpx04Hyjt8dgS4MXAxy8B\nZx3hTO8DVwQ+rgNiQyAT1tqnrbW/CnyaC5SEQi5jzDRgBvBK4KGgZxrAEoKf6yzgLWtto7W2zFp7\nU4jk6umHwC+BiT1+6xesXFVAauDjZPz/UAiFXMc6jROHpnFiCDRODInGiaHROBEcoTROhOIYARon\nhkTjxJBonBi8ER0jjublY1nA2h6fVwYeawhGGGutG3AbY3o+HNtjulsFkH2EM3mA5sCnNwCvAucG\nM1NPxpiVwFj8leG3QiDXncA3ga8EPg/qn18PM4wxLwIp+KcKhkKuCUBMIFcy8OMQyQWAMWYRsAf/\nVNTaHoeCksta+5Qx5lpjTBH+1+tC4A/BzjUKaJw4dCaNE0OjcWLwJqBxYtA0TgRNyIwToThGBHJp\nnBgajRODNwGNE4My0mPE0TxTqC9HsAMcQtDyGWMuxv9D/Jt9DgX1NbPWngRcBPylT5YjnssY82Xg\nI2vtzgFOCdZrVYj/B/fF+AeXP9G7mBusXA781epL8U9n/DNB/jPs40b868z7CkouY8w1QLG1dgpw\nJv7v+Z6C/XqNFqH+Omuc6EPjxKBonBgejRNyIKH8Ogc1m8aJQ9M4MWQaJwZppMeIo7koVIq/kt8l\nB38zqlDSFGg0BjCG3tNBjwhjzLnA/wecZ62tD5FMCwON87DWrsf/Q6kxyLkuAC42xqzC/wPgNkLg\ntbLW7g1Mj/VZa7fjbyiWHOxcQDmw0lrrDuRqJPh/hj0tAVbi/41fao/Hg5XrZOANAGvtBiAaSAuB\nXMc6jRODoHFi0DRODI3GiaHROBEcoT5OBP1nDGicGAKNE0OjcWLwRnSMOJqLQv/C32QJY8wCoNRa\n2xjcSP28BVwW+Pgy4PUj+eTGmETg18Bya21Xs7OgZgo4DfgugDEmE4gLdi5r7RestYustYuBh/Hv\nFhD018oYc7Ux5r8DH2fh32Hhz8HOhf/v35nGGGegSVzQ/wy7GGNygCZrbYe1thPYaow5JXD40iDl\nKgJOCOQbj3/QKwiBXMc6jROHoHFi8DRODJnGiaHROBEcoT5OBP3vjMaJwdM4MWQaJwZvRMcIh8/n\n+8wJg8UY83/4fyB4gVsCVbNgZVmIfw3pBKAT2AtcjX/KWRSwG7gu8E11pDLdhH9t5rYeD38F/w+p\noGQK5IrGP20xF3+V83ZgDf5t9oKWq0e+HwO78Fdjg5rJGBMP/A1IAiLwv1afBjtXINvX8E8jBvgZ\n/u1JQyHXQuBn1trzAp/PAB7AXwRfba39ryBkigMewT8Iu/D/5mhfsHONBhonDplJ48Tw8v0YjROD\nyaZxYvCZNE4ESaiME6E4RgRyaZwYXr4fo3FiMNk0Tgwuz4iOEUd1UUhERERERERERIbnaF4+JiIi\nIiIiIiIiw6SikIiIiIiIiIjIKKSikIiIiIiIiIjIKKSikIiIiIiIiIjIKKSikIiIiIiIiIjIKKSi\nkIiIiIiIiIjIKKSikIiIiIiIiIjIKKSikIiIiIiIiIjIKKSikIiIiIiIiIjIKKSikIiIiIiIiIjI\nKKSikIiIiIiIiIjIKKSikIiIiIiIiIjIKKSikIiIiIiIiIjIKKSikIiIiIiIiIjIKKSikIiIiIiI\niIjIKKSikIiIiIiIiIjIKOQKdgCRLsYYH7Ad8ACxwHrg59baj4KYaQJQGMh1FVAO3A9MARzA76y1\n9wXOnQvcB6QBVcDN1tqNgWNfBG4FwoF84Hprbb0xxgHcAVwC+IDnrLXfN8bcDPw/IMxaO+HIfLUi\nIiIiIiIymmimkISaJdZaA+QCjwEvGGNOC3KmvdbaadbadcADwFpr7XTgTOAXxhgTOO8p4FfW2qnA\n/wF/BTDGjAN+D5wf+Np2AT8PXPMFYAkwJ/DfEmPM5dba+4GlR+KLExERERERkdFJM4UkJFlrfcA/\njDGJ+AssJxljIoFfA8uACOBBa+0vAIwxu/DPuLkBf0Hpb9ba7xpjXPhn9pwKhAEbgWuttQ3GmIuB\nn+GflVQEXGWtrTpEtAeADwIZS40xO4HpxpgIIMla+3zg2IvGmIeMMdOBs4C3rbXFgXv8CXgX+CZw\nBfCotbY98HU8EXjsmWG+dCIiIiIiIiKDoplCEupeBE4wxkQD/wPMAGYDM4HLjTHLe5x7GnAisBD4\nljFmLHAuMBGYBuQBm4ETjTGTgCeAK621k/AXae4/VBhr7UvW2lrongE0FVgX+P+OPqfvCDzvVPzL\nz7psBzKMMckDHJt2qBwiIiIiIiIin5WKQhLqGvB/n8YDFwJ/tNa2W2ubgceBS3uc+zdrrcdaW4q/\n908uUIm/kHQJEGOtvc1a+wb+2UbvWWvzA9feD1xkjAkbTChjTBLwT+AXgRlAMUBbn9Na8c9C6nUs\nMCvId6BjPa4RERERERERGVFaPiahbgLQCdQBScBdxphfBI5FAh/3OLe+x8ce/E2aPzLGfAv4FvCY\nMeYl4BuBe51mjNna5/pUoOJggYwxWcBrwItdy9eAZiCqz6kxQFPfY8aYKPxNqvsd63GNiIiIiIiI\nyIhSUUhC3eX4Z/R0GGNKgd9Ya18eyg2stc8AzxhjUoBH8O/qVQi8Za29fCj3MsYkAG/g7wN0V49D\nW4HJPc5z4N+hbAuQA5ze49w8oMxaWxcoSk0B3uxxbMtQMomIiIiIiIgMh5aPSUgyxjiMMZcD3wF+\nEHj4BeBGY0xY4Pitxphlh7jPdcaY2wCstTX4izc+/IWdUwO9hTDGHG+M+d0gov0MeKdPQQhr7Rag\n0hhzVeChrwC7rbXbArmX9til7L+AJwMf/x24yRgTa4yJA27qcUxERERERERkxGimkISa94wxbiAR\n/4yZC6y1awLH/oB/Odlm/Muv1gB3H+J+LwCPGGMKATf+GULXWmtrjDFfBZ4L7BzWiL8AdShfA0qN\nMef1eOzuwBbyVwEPGWNux9/T6GoAa+1eY8w3gOcDu6Gtw7+cDWvtM8aYhcB6/MWqv1lrXxpEDhER\nEREREZHPxOHz+YKdQSRkGWMm4F++NmE0PbeIiIiIiIgc+7R8TERERERERERkFFJRSOTQxhhjthpj\nFhypJzTG3Ay8faSeT0REREREREYfLR8TERERERERERmFNFNIRERERERERGQUCpndxyorG4c9ZSk5\nOYba2pbDGeewCMVcoZgJlGuoQjFXKGaCYyNXenq8Y4TjiIiIiIjIKHRMzBRyucKCHeGAQjFXKGYC\n5RqqUMwViplAuURERERERAZyTBSFRERERERERERkaFQUEhEREREREREZhVQUEhEREREREREZhVQU\nEhEREREREREZhVQUEhEREREREREZhVQUEhEREREREREZhVyDOckYMwt4AbjLWntvn2NnAb8APMCr\n1tqfBh6/C1gM+IBvW2s/OZzBRURERERERERk+A5ZFDLGxAK/B94e4JR7gHOBvcAKY8w/gXQgz1p7\nojFmOvAIcOLhiSwiIiIiIiIiIp/VYGYKtQPnA//b94AxZhJQY63dE/j8VWAp/qLQ8wDW2gJjTLIx\nJsFa23DYkssxxev1UV7Twp6KJkoqm2hs7aTT7cXt9tLp9uILUq7ISBft7e4gPfvAeuaKjXJx1dlT\niQwPG/T1nW4PhSX17ChtoMPtxev14fX6SIiN4Jzjc3E6HCMVfcjcHi8+H4S7tNpVRERERETkcDpk\nUcha6wbcxpgDHc4CKnt8XgFMBtKAtT0erwycO2BRKDk5Bpdr8G9q+0pPjx/2tSMpFHN1ZdpWXEtN\nQxsnzMzCEaQigNfr46HnN/HmJ8W0d3iCkuFYcPrCXE6YlX3Qc9ra3by+ahdrCyrYsrOaDrf3gOdN\nGpfMyXNyhpVjMN/vHZ0eCvfUMWlMItGRA/8Iqqhp4dWVO3lj1W4S4yK46z+XHPT8Q+XqdHtYU1BO\n8b5G6ps7qG9qp6mlk5z0WGZNTmPWpFQS4yKHdX/wf10OB7jCnIP++xSKPx9ERERERGT0GN47rIEN\n9E7okO+Qamtbhv2k6enxVFY2Dvv6kRKKuboy1TW1c+tDq2ht9zBrYgrXnjeNlISoYd3T6/NRuKeO\nxpZO4mPCiY+JICE2grjo8ENe+9z7O3h55S4ykqOZnJdIbkYcuRlxJMZFEO5yEuEKwxXmCFrRKi0t\njqqqpqA898F05dpQVMWfXimgaHcNkzLjDniuz+dj9ZZy/vHedmob2wHIzYhj+vhkTG4SMVEuwsKc\ntLR18rtnNvK31wvIy4ob1Gve6fawblsVH2wqo6KulWWLclkyf0y/axuaO1hfVMWGoio276qho9PL\npJwE/t8X5xMZ0bsYvLeqmeff38G6wkp8Pn+RpamymfufWc815xywOH1QDe0eXnp/O6s276O5rf+s\nr3UWXv5gJwBj02M5fd4YTpmTPaiZVy1tnay1lawuKKdgdy0+n/+HXbjLSXSki/FZ8UzKTmBSTgKT\n+xTBhvLzQcUjEREREREZCZ+1KFSKfwZQlzGBxzr6PJ4DlH3G5zpmHGw5jMfr5YONZczLSycxNmLE\nMjz1diGt7R6yUmLI31nDrQ+v5otL8zh1TvagCzBVda18mL+PDzeVUVXf1u/4xadM5OJTJg54/Yeb\nynhp5S7Sk6L47XdOp6O1Y9hfz0iJj4mgbRDFrSOtK1duhr8QVF7XesDz9lQ08cS/LEUl9bjCnCw/\naTxnLcwlYYDvrUXTMvi4oIJNO6qZMzltwOfv6PTwzIrtrNy0j5bAMrYIl5Mn/rWNtdsque686aQm\nRlFW3czrq4v5aPM+3B7/IsDs1BjiosMpLKnnvhfy+eals3GF+f8ubN1dy++f3Uhru4dxmXGcfVwu\n8/PS+cVf1vLOugFSFo4AACAASURBVL0cZzKYNj75kK+Pz+dj884aXvxwF0V76wFIiI1g2fHjmD4h\nmYSYCOJjwomOdFFS2cTW4jq2Fddi99Tz1ze38cIHO1m6cCxLF449YHGzrLqZ5/69k/WFld1f18Ts\neGKiwuns9NDp8dLQ3MnG7dVs3F4NQHpSFL+8+aRDZhcRERERETlSPlNRyFq7yxiTYIyZAJQAy4Gr\n8S8fux14wBizACi11obWlJkg8fp8/ObJT6lv7uAnN5zQrzD04aZ9PPa65b1PS/n+NQuIGEKfGPAv\nx3I6D17Uyd9ZzccFFUzKSeAH1yzkw01lPPVOEY++tpVte+q44YLpBy0M7Sxr4KUPd7G+qAqAyPAw\nTpmdzdiMOJpaO2hs8b8ZfunDXczPS2NcZv9ZDra4lkdf20pMpIvvXDGXxLhIKkOwKBTq0pOiAais\n7V8U8vl83P2PDdQ2trNwajqfP3NK9/kDWX7iBD4uqOCllbuYPSl1wO+Dl1bu4q01JSTGRnDevHGc\nMiebMdmJ/Pava9m4vZofPrKayWMS2byjBh+QmRzNGfPHMDcvjczkGNweL/c8s5GN26t57LWtXH/B\ndD7ZWsHDL2/B54Mbl0/nxB7LGq8/fzo/f2INj7xawE9vOKF7dpHH62Xbnnp8Ph9RES4iI8Koqmvl\npZW72FHqX626aEYmi6dnMHtSanfxqae8sUnkjU2CkybQ0NzBW2tLeHddCS98sJNXV+1moUnnlNnZ\nTBufTHNrJy9+sIt3P92L1+djTHosi2dkcsL0TNIO8No2NHewo7SBHWX1JMYOf2maiIiIiIjISBjM\n7mMLgTuBCUCnMeZy4EVgp7X2OeDrwJOB05+21m4Dthlj1hpjVgJe4JaRCH80WrlpH9tK/DMXPtq8\nj9Pm7u/d4vX5eOPjYgB2lzfy+Bv2kAUagPrmDtYXVrJuWxUFu2s4fnom150/jTBn/zfA7Z0e/vLG\nNhwO+PK5BqfTwalzc5g1KZV7n93Iyvx9TMpJ4MwFY/tdu6O0gRc/3Nk982FyTgKnzcth0bQMoiJ6\nfytt2lHNXX/fwBP/snz/moW9GheX17Rw77ObALjlkllkp8YO5qWTA4iOdJEQE07FAWYK1TV1UNvY\nzvy8NG65dPag7jc2I475eWl8WljF1uI6ph9gVk55bQtvfFxMSkIkP79xcXeBJjUxmm9fPocPNpbx\n5NuF5O+oYWJ2AucvHsf8vPRexUpXmJNvXDKLXz/5KR/m76O2qZ2CXbVERoTxzUtnM2NCSq/nnJST\nwLITxvHaqmKeWbGdLy6dwqrN5by8chflByiIASycms6FJ09g4aycQS/TSoiN4NLTJnH+4nG8v6GM\nd9eVsGpzOas2l5OaEEVLu5vWdjcZydF84YwpzMtLO+jfz4TYCOblpTEvb+BZVyIiIiIiIsEymEbT\na4ElBzn+PgfYbt5a+73PlOwY1Nru5pkV24lwOfH6fLy6ajenzM7ufrO8saiasuoWjp+eQWVdGyvz\n9zExO4GlC/sXaMA/Y+fZ93ewZWdN9+5csVEuVubvo9Pt5asXzug3M+KZtwupqGvlnEW5vWbwJMdH\ncssls/nxnz/hybcKmZidwMTsBMC/3O3Jtwp599O9AEzNTeLikycwbXzygG+IZ09K5TiTzhpbyQcb\ny7qLX6VVzfzmqU9pbnNz7XnTmN7nzb8MXXpyNLvKGvF4vb0KgfuqmwEYkz60otvykybwaWEVL6/c\ndcCi0NNvF+H2+PjCmXn9+gE5HPuLjLWN7UzMjh/weyQqwsW3r5jLL55Yy5ZdtSTGRvCfn597wJll\nAJ87ZSLrC6t4e20JG4qqqKpvI8zp4NQ52aQmRNHW4aGt00OYw8Fp83K6l9YNR1SEi3MW5XL2cWMp\nLKnng01lfFJQgSvMwZVL8zhjwZgDzjoSERERERE5mhzuRtNyEC+v3EVDcwefO3UitY3trFhfyhpb\nwfHTMwF4PTBLaPmJE4iJcvGTRz/hqbcLGZseixm3/815ZV0rz76/g9VbygGYMiaR40w686emExcd\nzu+e2cgnWyvweH3cfPFMXGH+IlRRST3PvFNIcnzkAXv9pCRE8bWLZvLbp9fzx+fy+dF1i3B7vPzx\nuXyK9tYzNj2Wq86aOqieLgBXnjWVTTtr+Me7RczPS6OmoZ07n15PU2snXzhzSq9ZUjJ8GUnRbN/b\nQHVDOxk9ljCV1fibt2enDK0oNDE7gZkTU9i8s4aivfVMGZPYfWzj9mrWF1UxbVwSx5n0Ae+RHB9J\ncvyhl0slxETw31+cx7vr9nLG/DEHXILVJdwVxvUXTOcXT6ylrqmdM+aP4fzF40lNHF6D9MFwOBxM\nzU1iam4SXzrHdO8uJiIiIiIicixQUWgEeLxetuyq5YS4/W9Wy2ta+Ncne0hNiGLZ8eOoa2rn/Q2l\nvLxyN4umZbCzrJFte+qYNSmFsYEZDl//3Cx+89R67n12E2PSYvF4fXi8Pkoqm3B7fIzPjOfzZ0zu\nN9vmP6+Yyz3/3Mi6bZXc9fcNxMeEs2VXLU2tnQBcuTRvwK29Z05M4aJTJvLCBzu5958bKa9rpb6p\ngxNmZHLtsmn9ZoYcTHJ8JJ87ZSJPv1PEgy9uZkdZA23tHr6yzHD6vDFDfVllABnJMQBU1Lb0LgpV\nBYpCaTFDvufyE8ezeWcNT75VyJfPNYzPivfPGHu7EIcDrjpr6mHbES4tMZorzpgyqHMn5yRy+3XH\nExsdPqii0+F0oMbwIiIiIiIiRzMVhUbAsyt28NrqYmJe3Mzpc3M4e1EuT71diMfr4wtnTiEiPIyM\n5BhOmJ7Jqi3lbNxezYf5+wBYdvy47vuYcclcc85U/vZWIdtK6glzOghzOkhLjOaiUyZw/PTMXr16\nukRGhPHty+fw+2c3sXlnDeAv0Jw8K4szFo0bcOvyLheeNIGivfVs3lmD0+Hgi2dO4exFucMqAixd\nOJYPN5WxeVctYU4HN100kxNmZA75PjKwjJ7NpntMACur8S8fy0oZelHIjEtm3pQ01hdVcfujn5A3\nNpGMpGjKa1pYumBsd+EyGIL53CIiIiIiIscSFYUGqam1k8q61u4+OwMpKqnn9dXFJMdH4gNeW13M\nvz7Zg8frY9q4JBb2WHJz/uLxrNpSzj/e205ZdTPjMuL69XA5fd4YTpubM+SCTER4GP9x2Rw2FFUx\nJj2WrJQYHA4H6enxh2y663Q6uOnCGTz/wU4WDXIL8IG4wpxcd/50nnyrkAtOHM/cKWq4e7ilJ/uL\nQn2bTZdVt5CSENmvCfhgffOy2eTvqOatNSXk76yhsKSeuOhwLj61/9JDEREREREROfqoKDQIXp+P\nO59az+7yRi47fRLnLx5/wCJNe4eHh1/ZAsDXLprJotk5vPBeIa+vLqausZ0r+yy5GZsR1z0bA2DZ\nCeMOeN/hLtMJdzk5blrGsK6Nj4ngS+eYYV3b18TsBH7wpYWH5V7SX0ZXUajHLlyt7W5qG9uZOWH4\nBT2nw8GcyWnMmZxGWXUzH27ax4wJycRFh3/mzCIiIiIiIhJ8KgoNwurN5ewu98+u+eeKHdQ3d/DF\npXn9lm49s2I7FbWtnHt8LlNzk4gID2PJvDGcNieH9k7PAfv4XHDieNYXVZGSEDnsAo6MbvHR4URF\nhPWaKbQv0GQ6K3VoTaYHkp0ay+VLJh+We4mIiIiIiEhoUOfUQ+h0e3j2/e24whz84JqFjEmL5a01\nJTz80hbcHm/3eQW7anh7bQnZqTFcetqkXvdwOh0DNnaePCaR686bxs0XzdKuRjIsDoeDjORoKmtb\n8fl8AJQFtqPPSR16PyEREREREREZHTRT6BDeWltCdUM7y44fx5Sxifzv1Qu455mNrNpSzqYd1URF\nuIgId1LX1IHT4eDG5TMIdw1+hy6AU7U1u3xGGUnRFJc3UdfUQXJ8JGXVh3emkIiIiIiIiBx7NDXl\nIJpaO3l55W5io1xccNJ4AOKiw/nuF+dxyuxs4mIiAB/NrZ04HfCFM6ccshG1yEjoajZdGVhCti9Q\nFNJMIRERERERERmIZgodxMsrd9Ha7uYLZ04hNmp/c93I8DCuv2B6EJOJ9JaZ7C/+lNe2MDU3idLq\nZqIjXSTERgQ5mYiIiIiIiIQqzRQawPbSet5eW0JaYhRnLhgb7DgiB5WetH+mkNvjpaK2lezUmGHv\nXCciIiIiIiLHPs0U6qHT7WWNreDddXsp2lsPwOVLJhPuUu1MQltG0v5t6avq2/B4fWRr6ZiIiIiI\niIgchIpCAUUl9fzhuU3UN3fgAGZPSmXpwrHMmZwa7Ggih5ScEIkrzElFbStlVf6dx7LVZFpERERE\nREQOQkUh/NvJ3/PPTXS6vZx7fC5nzB9DRrJmWcjRw+lwkJ4URWVdK2U1/ibTmikkIiIiIiIiBzPq\ni0Ibt1fzh+c24fP5uOWSWcyfmh7sSCLDkpEUTVl1C0Ul/qWPmikkIiIiIiIiBzOqi0JrbSX3v5CP\n0+ngW5fNYfYkLRWTo1fXtvRbdtcQ5vTPHBIREREREREZyKgtCrW0dfLgS5txhTn59uVzmDY+OdiR\nRD6TrmbTHZ1ectJiCXOqQbqIiIiIiIgMbNS+a9xQVE2n28v5i8epICTHhJ59sLJT1E9IRERERERE\nDm7UFoXWbasEYIHJCHISkcMjI7B8DCA7TUUhERERERERObhRWRTq6PSwaWc1mSkx5GiHJjlGpCVG\n4XD4P85OUZNpERERERERObhRWRTavLOGjk4vC6am4eh6Fy1ylHOFOUlN8DeXzlKxU0RERERERA5h\nVBaFupeOaft5OcaMTY8j3OUkW0UhEREREREROYRRt/uYx+tlfVEVSXERTMxOCHYckcPqmnOm0tDS\nQVTEqPurLSIiIiIiIkM06mYKbSuuo7nNzfyp6Ti1dEyOMSkJUUzIUrFTREREREREDm3UFYXWFVYB\nWjomIiIiIiIiIqPboNaYGGPuAhYDPuDb1tpPehy7GLgVaAeestbea4xxAvcDs4AO4GZr7dbDHX6o\nfD4fnxZWEhPpwuQmBTuOiIiIiIiIiEjQHHKmkDHmdCDPWnsicANwT49jTuBe4HzgNOBCY8xY4GIg\n0Vp7UuCa34xA9iHbta+RmoZ25k5JxRU26iZJiYiIiIiIiIh0G0xlZCnwPIC1tgBINsZ0NS1JA+qs\ntZXWWi/wNnAWkAd8HLhmOzDeGBN2uMMPlXYdExERERERERHxG8zysSxgbY/PKwOPNQQ+jjfG5AG7\ngDOA94CNwH8aY+4GpgCT8BeQygd6kuTkGFyu4deN0tPjD3lOQXEd4S4nSxaNJyryyOzONJhcR1oo\nZgLlGqpQzBWKmUC5REREREREDmQ4lZHuLbustT5jzFeAR4B6YCfgsNa+Zow5GXgff4GooOd1B1Jb\n2zKMKH7p6fFUVjYe9JxOt5fdZQ1MyIqnsaGVg599eAwm15EWiplAuYYqFHOFYiY4NnKpeCQiIiIi\nIiNhMEWhUvwzg7rkAGVdn1hrVwCnAhhj7sA/Ywhr7a1d5xhjtgMVnz3u8O2tasLj9TEuS2+uRERE\nREREREQG01PoX8DlAMaYBUCptbb719vGmNeMMRnGmFjgQuAtY8xcY8wjgePLgHWBnkNBs3ufP/L4\nTBWFREREREREREQOOVPIWrvSGLPWGLMS8AK3GGOuBeqttc8BD+EvHPmAO6y1VcaYGsBpjPkYaAOu\nHrGvYJB2lzcBKgrJ4dfmbqOitQqfz0du/Bicjt61VrfXzZ7GUlKikkiMTBjgLn4+n4/ylgqq22p7\nPOpgYsI4YsKjRyC9iIiIiIiIjFaD6ilkrf1en4c29Dj2LPBsn/O9wLWfNdzhVFzeSJjTQU5abLCj\nSB+dnk7cHvcBj1W11vDqzjdxOcPIiEknIzqN1OgUwhwDT3KLDY8lLjwWh+OgbawOyuvzUttWT9m+\nEraVFVPRUklVazUe3/4Jbx2eTqpaq6jv2N8XJi48lpmp05iVNp1OTyebqgsoqN5Gm6cNgHHxY5mV\nNp0ZKVOJdkV1X1fbVk9+dQH5VQVUtdX0y5McmcQ3591AVmzmsL8mERERERERkZ6OzBZcQebxetlT\n0cSY9FjCXYNZMSdH0gObHqO4qYSrzRXMTZ/Z/fiexr38ccMjNHQMvUlwtCuKjOh0MmLSAv/5P06O\nTKK2vY6KlioqWiqpb2/A132Vj+bOFipaqqhsraLTe+BCVRcHDlKikpieMpWMmHQ6PZ1sri5g9b61\nrN63f8O+1KgUjsucS2VrNYV1OyhuLOHVnW8e8J5RYZHMz5hDblxOd1Grtq2O9/d+xJ1r/8jNc65j\nctKEIb8eIiIiIiIiIn2NiqJQWXULnW4v47R0LOTUtddTULMNgAc3PcYZuafwucnnU1S3k4c2PU67\np4PL8y7CJE+hotVfyKltq8Pbo5TTi89HY2cz5S2V7G0qZXfjniFnigyLIDs2k/ToNCamjyXOl0Bm\nTDpp0amEh4V3n+fEQZgzrNe1Xp+XksZSNldbXM4wZqVNJysmo7vA0+pupaCmkKK6nXh8nu7rosOi\nmJaSx5Skibic/f9ajk/I5a9bn+H36x/kuplXcVb6iUP+ukRERERERER6GhVFITWZDl2bqgoAOGfK\naWwo3cq7ez5ga00hFS1VOIDrZ13Ngow5AOTEZR3kTv15fV5q2moDs4KqqGitpLatnuSoxO5ZRMlR\nSTjZv8wsyhVFQkR8dxFnqNuZOx1OxiWMZVzC2AMej3ZFsyBjTvfXNFiLs48jPiKeh/Of4KFNTxAb\nH8GUqKlDuoeIiIiIiIhIT6OiKFTc1WRa29GHnE1VWwC4yJzNsjHn8LR9jtX71hLtiuKm2V9havLk\nYd/b6XCSFp1KWnQqM1LN4YocNDNTDd+Z/zV+9+kD3PfxE/zvcf9BRkx6sGOJiIiIiIjIUWpUNNjZ\nXd6IA8hNjwt2FOmhzd2OrS0iJzaLjLg0IsMi+PKML/CteV/le4u+/ZkKQseq8Qm5XGkuo83dzp/y\n/0qnp/Mz3a/D03GYkh0+Pp/vkP2cRERERERE5LM75mcKeX0+issbyUqNITIi7NAXyLDkVxUQERbO\n1OQpg75ma20hbq+bOWkzej0+LSXvcMc7pizKmk9xWzHv7PiQZ4te5gvmkmHd56PST/jr1mfIjM1g\ndup0ZqVNZ2LCuH59ko6k3Q17+Me2F9jbvI/rZ17F7D7fG6HK4/Xw772rWJQ1n9jwmGDHERERERER\nGZRjvihUWdtKW4dHS8c+ozXl63lx+2t8c95XyYhJ63Ws1d3KQ/lP4MTBbYv/m5So5EHdc1Olf+nY\n7PSj441/KLlu/ufZWr6d9/d+RF7y5CH3KNrbVMbT254jPCyc6tYa3ix+jzeL3yMzJp3/d9w3iXZF\nH/bMzZ0tbKm25FcXsK12O2nRKcwKFKMSIxJ4ccfrrCz9GB8+whxhPLDxMa6cdikn55xw2LMcbqv2\nreEfhS9Q017LpVOWBzuOiIiIiIjIoBzzRaHd5f4mweMyVBTqq7GjCbfXTXJU0kHPq2mr5cmtz9Lm\naWNl6cd8bsr5vY5vrNyCO7Dc55nCl7hp9pcP+dxen5f86gISI+IZF3/gpswysEhXBDfMuppffnIP\nfy34B5MSx5MUmTioa/1Lz/5Cp9fNzXOuxSTnsa22iBUlK9lSY8mv2sqirPkHvcfepjL+se0FLpp8\nHpMSxx/03A5PJ48XPM2Gyny8Pi8A8RFx7KwvZkf9bl7c8TpOhxOvz0tWbCafz7uYiLAI7tv4CH/b\n+k/q2hs4f8JZ3c2/Q9G68o0AbK7aqqKQiIiIiIgcNUZNUSiUZgp9uHc1jZ1NnDv+zKC90fV4Pdy1\n7j5a3K387KQfHHAbdPD3d/lrwTO0edpw4GBN+XoumrwMp2N/O6p1Ff43xFmxmWyozCe/qoBZadMP\n+vw76nfT1NnMyTkn9LqXDF5WbCYXTDqH54peIb+qgFPGLD7kNT6fj6e3PUd5SyVn5p7avTxrVtp0\nkiIT2VJj2Vx98KLQvuYK7vn0QZo6m1lV9slBi0KdXjcPbXqcLTWWMXHZLMiYy+y06eTEZtHibmVL\ntWVT1RZKm/dxYvYilow9uXv52ncX3sIf1j/MqzvfpNXdyuV5Fw3xFToyGjua2Fa3HYB9LRVUtdaQ\nFp0S5FQiIiIiIiKHdsy/Gy/u3o4+NJpMry5by9/sP3lpxxu8v/ej4OXYt5bylkoaO5rYWlM44Hkf\nlq5ma20hM1OncULWQmrb69hRv7v7eEtnK1trtjEmLpsbZl6N0+Hk79teOGQD465dx/r2E5KhmZU6\nDYDCuh0HPP7xvnW8uP11VpWtYUf9bv699yM+3reO8Qm5XDz5vF7njonL7i4Mdc3o6auypbq7IOR0\nONlet2vAbB6vhz9v/htbaiwzU6fxP8d9i2UTzmRMXDYOh4PY8BgWZc3n+llXc+sJ32XpuNN69TPK\njEnnuwu/SUZMGu/t+ZDatrohvjpQVLeTV7e9g8frGfK1g9U1AyorJgOAzdVbR+y5REREREREDqdj\nuijk8/nYXd5EelIUMVHhwY7D9rpd/G3rM0S7ookLj+XZwpcobiw54jk6vW5e3fkWDvyzlLpm+vRV\n3VrDs0UvE+2K4qppl3XPHvlk37ruczZVbcHt87AgYy45cVmcmXsq1W01vLH73YNm2FS1hQjn0BpT\nS3+ZMRnEh8dRWLsDn8/X61hLZytPFPydN3a/wxMFf+fOtX/g6W3PE+2K4vqZV/ebHeZwOJiZamju\nbGFXQ3G/56ptq+Oe9Q9S39HAZVOWk5c0iX0tFTR3tvQ71+vzdi8Zm5o8hRtnfWnA2WgHkxgZz9Lc\n0/Dh45N9nw7p2saOJh7Y+CiPfvoPfrvuPqpba4b8/IPxacUmAK6cdhkA+dUFI/I8IiIiIiIih9sx\nXRSqaWinqbWTcZlHdulYWXM5v/j4Lh7/9Blq2moBf4HlwU2P4cXHjbOu4cszvojb5+FP+X+l1d02\n7OeqaKmkurV2SNesLP2Y2vY6zsg9heTIJDZWbe63BbjX5+UvW5+h3dPBFXkXkxSZyNTkySRExPNp\nxabuHkLrKjYAsCBjNgDnTTiLpMhE3tr9HuXNFQd8/vLmCspbKpmeMpWIsOAX645mDoeDvORJ1Hc0\nUNla1etYQc02vD4vJ2Uv4kpzKUtzT2Ne+ixumHnNgMubZqb6l/1truo926XT08nv1z9ETVstyyee\ny5njTmNy4gQAdtTv6nefl3a8wZry9UxKHM/XZn/lM/05L8iYi8vpYtW+Nf0KXwfzXNErtLhbGZ80\nll0Nxdzxye/YUJk/7BwH0tjRhK0tYmLCOKYkTSQnNovC2u2HnCknIiIiIiISCo7polBxVz+hI1wU\n+vfeVextKuPlbW/zo49+ySP5f+W+jX+mqbOZK/IuZlpKHjNTDWePW0JVazVPbv3nkN7sdnF73dy5\n9o/8Zu29tLpbB3VNu7uD13e9TURYBOeMP4P5GbNpdbextWZbr/PWlW9gW20Rs9Omc3zWAgCcDicL\nM+fS7G6hoGYbLZ0tFNQUMjYuh4yYdACiXJFckXcRbp+HJwr+fsBlO+sCMyuOlu3GQ11e0iQACmt7\nLyHrWsZ06tgTOWXMYi7NW85XZ3+Z6alTB7yXSZ6CyxFGfp8lUKv2raG8pZJTxixm2YQzAZiUNAGg\n13JC8BcUV5Z+THxEHN+Yez1RrsjP9PXFhEczN20m5S2V7GrYM6hrCmt3sHrfWsbG5fB/Z3+Pq6dd\ngdvr5sFNj/PS9tcPeM2exr08sPGxIRVZ11fm48PXvfvbzNRpdHrdbKvdPuh7iIiIiIiIBMsxXRQK\nRpNpr8/Lhsp8YlzRfH3Rl8iKyWBtxQbKmss5fezJnDb2xO5zL5x0LpMSx7O2YgMryz4+4P26lnC1\nudv7HdtaU0hTZzMNHY28tONfg8r3RtEKGjoaOXPsKcRHxLEgYy7QewmZx+vhlZ1v4nQ4uTzv4l7N\nsBdldi0h+5QNVVvw+Dz9tkOfmz6L4zLnsbOhmBd39H4Dvr1uF6/veotoV7SKQodJXvJkoHdfIa/P\ny+bqrSRExDM2LmfQ94pyRTIlaRIlTaXUtdcD/u+HN3e/R7jTxQUTz+7+fpiYMA4HDrbX7ex1j531\nxTR1NjMnbeZh29r+hOzjAH9xqq++s9w8Xg9Pb3sOBw6+aC4lzBnGSTmL+J/jvkVGdBqv736Hf+9d\n1eua2rY67tvwCBurNrOydPWgc3X9vZnfoygE9CuqiYiIiIiIhKJjuihUVu3vdTImLfaIPWdxYwl1\n7fXMTpvBGZNO4gfH/yffmvdVrph6MZf12ao6zBnG9TOvJtwZzjt7Pjjg/d4sXsHbxe/zQemqfse6\n3pDGumJ4v2TlIfsTtbnbeKHgDaJdUSwddxoAExJy/UvIKrfQ6ekE/M2JK1qrOCnn+H7LjMbFjyUj\nOo2NVVv4qPQTYP8b4i4Oh4MrzaVkRKfxVvEK8qv8PVb6LqGLizhyfy7HsqyYDOLCYyms299XqLix\nhKbOZmamThvy7m5dO8d1zTRaU76e6rZaTsw+noSI/QXWKFcUY+Ky2d1Y0qswMxJNxKen5JEYkcDa\n8g3d36fgn6nz3RW3cefaP7K+YhNen5d3Sz6grLmck3KOZ2LiuO5zc+KyuGXejcSFx/L3bc93N1hv\n93TwwMZHqe/wF5E3BvIfSkNHI4W125mUOJ7kqCQAJiWOJ9oVzebqrcOa/SciIiIiInIkHdNFoYq6\nVlxhTpLih758pc3dxobKzQPuwjSQ9RX+niXz0mcB/gLJtJS8Xltt95QclcTU5Mnsay7vt7uSz+dj\nS7UF/H2Aer7J7PS62Vi1meTIJK6bdRU+fDxlnzto3vdKVtLY0czS3NOJCY/pzrcgcw5tnjYKarbh\n9rp5dddbuJwulo0/s989HA4Hx2XOo9Pbyfb6neTGjyEjJq3feVGuKK6fdQ0up4vHC56mrLmc+zc+\n2msJnRweQ3WlZwAAIABJREFUDoeDvKRJ1LXXUxVoppwf6AnUtTvZUHTNdtlctRWvz8sbu9/F6XBy\n9vjT+507OWkCbq+bPY17ux/bOAJNxJ0OJ8dnLaDV3dpdtCluLOGxzU8C/r5GD+U/we0f/YpXdr5J\nXHgsF01e1u8+adEpfHX2l3Hi4OH8v7CvuYLHtzzFnqZSTs45nhmphtLmfd2vY09lzeVsrt7a3S9o\nfUXX0rG53eeEOcOYkTKVmrZayprLD9vXLyIiIiIiMhKO6aJQZW0r6UlROHssfxqMNncbv1//MA9u\nemxI20v7fD7WV24iIiyCaSkD923pa0aKAWBLje31eGVrFdVt/jen5S2VFPVYprO1Zhut7jYWZMxh\nespUFmbMZXfDHlaWHngZmsfr4f2SlUSHR7Ek9+Rexxb2WEK2svQTatpqOXXM4u7ZD30dF9iFDOi3\ndKyn3PgcLptyIc2dLfzfx3dT2ryv3xI6OTz2LyHz97LZXL2VMEcYZhjFt4yYNDKi0yioLWRd+QbK\nWyo4PmsBKVHJ/c6dFGg23bWErLylkvKWihFpIn5C9kIAVu9bS117PfdveJROr5sbZ13DbSd8l5Nz\nTqC+o4EOTwefm3w+ceEHnok2JWkiV027nFZ3K79ccw/rK/PJS5rEF6Zewpy0mcD+2U5d3F4393z6\nIH/c8Aj/8+8fc9+GR1hR8iEA8wNN1rt0F9V6/Oxo6Ggcsd3PREREREREhuuYLQo1tXbS0u4mPWlo\nPU06PB3ct/HP3Vty25qiQV9b1lxOZWs1M1PMkN4Qzwg0/t1S3bvZc9fnXX18PuxR8OlaOrYg01+U\nuTRvOVFhkbyw/TUaO5r6PceGqs3UdzRw+oTFRLuieh0bFz+W1KhkNlZt9jehdoZz7gFmCXXJjEln\nfHyu//kPUhQCOHXMYuZnzMHt8zA9ZWq/JXRyeHQ1m95Wu4P69kaKG0uYnDSx35/1YM1Mm0aHp4On\nAr15zhm35IDn7d+BzN9suquYMjt95rCe92CyYzMZH5/LlmrLHzc8Qn1HAxdPPo856TPJis3kqmmX\n8dOTfsB35n+NxYEeRAM5IXsh54w/gw5PB+nRqdw4+0uEOcOYHVg617cotL5iEw0djUxMGE96dBr5\n1VvZ11LB5MQJJEUm9jp3RqrBgYPN1VspaSzl8S1Pc+uHv+BXa35/eF8QERERERGRz8gV7AAjpbLO\nvxtXxhCKQp2B3YmK6nYyL30Wm6u3sq1u8LsIdW133bV0bLDSo9NIjUrB1hbi8Xq6l5l1zRy6aPIy\ndjfsYX3lRlo6LyLcGc7Gyi2kRCV3F2eSIhNZPun/b+/e4+Ou63yPv2YyadO0KU3b9JJCL9DwpRVR\nCgiFxRYKKki3wiJeEEHg4KIoR49nxT0o4GWreBDXxd1lFXA56oK7VnTFqlykrlsvpSCK1K8tUG5t\nJb2kTe9zO3/MJKSlTSZt0plMXs/Hg8djZn6X+UzS8njMu5/v5/tm/mPlD7hv1Y+4ZOZFe7zHz19c\nCsCbp8+BvWZWJxIJZo17HQ88/wi7srt505QzaBgyotuaL33NO1m/YwNjh43p9rxEIsElMy7iuLEz\nOW7szH0uodPBmzC8MFdoVdszPLXhwJeOdTh2zAx+9sIvOrvRxg8ft8/zGutG0Th0FM9sXk0+n+f3\n658iQeKg3rs7p0w8gefaX+ClrWs5ZeKJnDV5zyVtDUNG9Phnt8P8I9/MpBETmT5qWmdX0aihhzG5\n4XBWtj3D9vQO6msL//9Y8lLh7897Z76DcfVj2bBjI3HT00wv7sC2dw1TRh7ByrZnWLjsywCMrx/H\neUe+6UA/tiRJkiT1i6rtFHp5UyEUamosLRRKZ9Pc+eS3WLHxTxw7Zgbve827mTpyMmu2rmNbentJ\n9/ht65OkEjW8pthtUKpEIsHMMYEdmZ2dW26ns2lWbnqaCcPHM7qukVOb30A6l+E36x5nxcY/sTNb\n+LLedWewN06azeEjmvnVukf3WGq2Zus6VrY9wzGNLUwaOWGfNXQsgamrqXvVF+19GV/f1LlMpidD\na4bwhgmzqDvArhX1LJlIMn3UkWza1cbPiwFGqb+ffTlq1DSG1AwB6LZrrHDuVLamt/GnDc/wdNtq\nph02peRgprdOGP96hqWG0TLqSN4VLtjjz39vJRNJThz/+ld1+hw3dia5fK4zlH2hfQ3PbH6OmWNC\n5/ysMcNGc2rzSYyrb9rnvTu6+0LjdK4+7n1cf/JHe+yqkyRJkqRDrWpDod50Cr28vZX/u/yr/G79\nHzimsYUriwOSWxqPIk9+j4Blf9bv2MCLW9dw9OjpB7RkZ0ZxBlHHF9GnN69mdy7NzOLrp0w8kZpE\nDf+95tcsf/kJ4NVLt2qSNbwznA/AvfF7ZHNZAH7+0i8Bup3lM7nhcN405QwunnEhw4tDqDWwdCwh\ne779JcbWjWb8fgKLUtQmU1zYMp+3HXUuhzd0v6V9x1yhf/vd98mT79Ndx/Y2vLaeT8/+OB96/f8g\nleyfRsfXFuv/XesfgFe67OZMOrXke8w5/FRuPv1GPnz8VRw7dkavd4CTJEmSpEOhar+pvNwRCvXQ\nKfTousf5/LK/58Xi7kPvP+4yaovzgI4ufsleWcISst8e4NKxDqHxKJKJJCuKc4Q6wqGOIdQNQ0Zw\n3NiZrNm2jsde/h1j6kYzueHwV91n2mFTOK35DazZto6fvfgLdmR28Ot1y2kcOopjx+y/gymRSLDg\nqHPsZhjAWhqP7Hz8mrEzDqqLBuC05pM5e8rcHs/rCIWeai1s8f7afgyFAOpr6/t1GeKkERMZXdfI\nUxsj7bu3suzPjzO2bjQzx4SS75FIJAxXJUmSJFW8qg2FWjftIAGMPWzfoVAun+Oe+D3ueqqwpfX7\nZr6Ldx9z4R4DoqeOnEwqmWLVpmd6fL8nWp8kQaJz96LeqkvVcdRhU3m+/UXad29lxYY/UZtMcdSo\naZ3nnNZ8cmftey8d6+ovjzqHEbXDuf/ZB/jJ6p+xO7ub0yed4jyfKjdx+PjOIKK/Zvrsy6QRE6ir\nGQrAuGFjD6pDqRIkEgleO3YmOzI7+eaK75DOpTn98Nl2+0iSJEmqOlX7Leflth00jhxKbWrfH/GH\nz/yU/3rpl0waMZHrTrp2j23WO9TW1DJt5GRe3LqW7XvNFcrn86zb9mceeO4Rbn3sn3hm83NMHzXt\noGapzBwdyJPnV2sfZc22dbSMOmqPkCqMns6Y4rbg3XX0jKgdztuOOpfd2d088PwjpBI1nNr8hgOu\nSwNDMpHk+HHH0Th0VOdSskP1vtMOmwIUuoQOtkOpEnQsgXtywx+pTaaYPfGkMlckSZIkSX2vpKEc\nIYRbgVOAPHBtjHFZl2MLgOsp7Gl1T4zxthDCCOBuoBEYCtwUY/xJXxe/P+lMlrb2XYTJo/Z5/Ndr\nl/OT5x6madgYrj3+/d0u82gZdSQr255hVduzHFfcZjuXz/FPT9zVucQrQYJpIydzQcvBbbc+Y8zR\nfP+ZxfzkuYc7n3eVTCR59zEXsnrLCxzRMKnbe5088QSWrl3GM5tXM2v86/pt8K8qyzuPPp/80flD\n3tUyc0zgjxtXMmt8dSw/nD5qGnU1dezM7uTE8ce7FEySJElSVeoxFAohzAFaYoyzQwgzgDuB2cVj\nSeA2YBawAVgcQrgPeBsQY4yfCCE0Aw8Dh2w9S2vbTvJA0z6GTD/dtppv//E/GJaq4+rj3tfjl72W\nxiNhNaxse6YzFPrNusd4amNk6sjJvHHSbGaOCX0SukwaMZGGISNo370VeGWeUFfHjG7hmNEtPd4r\nmUjynmMuZNGqH/KWqfMOujYNDIlEggSHvlNn7uGnMS+cQn57bc8nDwCpZIrXjp3Jo39+vNsB7ZIk\nSZI0kJXSTjAPuA8gxrgCaAwhjCweGwu0xRhbY4w54CHgLGA9MKZ4TmPx+SGzvyHTG3Zs5F9+/6/k\nyHPlsZcwfvi4Hu81deQUUokaVrYV5grtzu7mP5/5CbXJFFce+x5OnnhCn3XhJBPJzl3IGoeOOujZ\nLOOHj+Pq110+4Ge8qPIlE0nGDh9d7jL61EVH/yX/+8Rr9jnQXZIkSZKqQSmh0ASgtcvz1uJrHY8b\nQggtIYRa4AxgfIzxHmByCGEV8HPgY31Yc49aNxVCoY5OobZdm/n+04tZuOzv2ZrexttbFpTUbQMw\npKaWKSMn82L7Grand/DQ8z+nbddmzjzijTTW7Xt52sHo6A6aOeboqpjNIg1U9bX1TBl5RLnLkCRJ\nkqR+U9JMob10JhUxxnwI4VIKS8o2A88CiRDCe4DnY4xvCSG8DrgDOLG7mzY21pNKHfjuWE1NDZ2P\n23dlABjXXMO9T3+X/37hUbK5LCOHjuCiYy/kraF3y6leP2kGT29+lpU7/sQDLyzhsKENvPuE+Qyr\nretVXaV40+hT2Z7cytxppzC2vnfXlqq3NR0q1tU7lVhXJdYE1iVJkiRJ+1JKKLSGVzqDAJqBtR1P\nYoxLgNMBQggLgdXAHOAnxeNPhBCaQwg1Mcbs/t5k06bt+zvUo6amBlpb2zufP792C9SkuesPd7B+\n5wYm1I/jzMmnc9L4WQypqd3j3FJMGlJYPnLXY4Xtqc8/6q1sbUuzlXSv6irVnHGnk98Grdt6f21P\nDrSm/mZdvVOJdVViTVAddRkeSZIkSeoPpYRCPwVuAm4PIcwC1sQYO7/JhBAWA5cC24D5wC3AJOBk\n4LshhCnA1u4Cob72cts2hh39W9bv3MC8yW/kbUede1C7MU07bDI1iRrSuTQTho/nVLenliRJkiRJ\nA1yPSUmMcSmwPISwFPgK8MEQwmUhhPOLp3yNQnD0C2BhjHE9cDswNYSwBPg28Nf9Uv0+ZHM5No5c\nDg0bOG7saw46EAIYUjOEqSMnA3DB9POoSR74MjdJkiRJkqRKUNJMoRjjdXu99ESXY4uARXudvxW4\n6KCrOwA/fvrn1DS9QF22kUtnvvOgA6EOFx29gBe3ruE1Y169TbwkSZIkSdJAcyCDpivW8+0vsviF\nxeR3D2FW3TnUpYb22b0Pb2jm8IbmPrufJEmSJElSOfVNG02FeLptNXnypF8ITG4cV+5yJEmSJEmS\nKlZVhUIbd24CILdzOOMah5W5GkmSJEmSpMpVZaFQGwD53cNoGmUoJEmSJEmStD9VFgptgnySVK6O\nUQ19N09IkiRJkiSp2lRfKLS7jqZRw0gmEuUuR5IkSZIkqWJVTSi0O5tma3ob2Z3DGOfSMUmSJEmS\npG5VTSi0qThkOr+7jiaHTEuSJEmSJHWrakKhziHTu4bRdJihkCRJkiRJUneqJxTa9UqnUH1dqszV\nSJIkSZIkVbbqCYW6dAqlaqrmY0mSJEmSJPWLqklPNnbOFDIUkiRJkiRJ6knVpCebOjqFdtdRm3I7\nekmSJEmSpO5UTSi0cecmhjIc8klq7RSSJEmSJEnqVlWkJ7lcjk27NjOUEQCkUlXxsSRJkiRJkvpN\nVaQnm3ZuJpfPMSQ/HMCZQpIkSZIkST2oivSkddtGgM5QyOVjkiRJkiRJ3auK9GT99g0A1GaLoZDL\nxyRJkiRJkrpVFenJ+u2F7eiTWZePSZIkSZIklaIq0pPWbYVOoWSmHnDQtCRJkiRJUk+qIj1Zv70w\nUyixexgAtTWJcpYjSZIkSZJU8aojFNq2kWGpOvLZFOBMIUmSJEmSpJ4M+PQkn8/Tun0jo+sayWRz\nANQ4U0iSJEmSJKlbAz492Z7Zwc7MLkbXjSKdyVGTTJBMuHxMkiRJkiSpOwM+FNq4s7DzWOPQRtLZ\nnEOmJUmSJEmSSpAq5aQQwq3AKUAeuDbGuKzLsQXA9cAu4J4Y420hhCuAS7rc4sQY44i+K/sVG3e2\nATC6bhSZbJ5al45JkiRJkiT1qMdQKIQwB2iJMc4OIcwA7gRmF48lgduAWcAGYHEI4b4Y4x3AHV2u\nv6if6u/sFBpd10gms90h05IkSZIkSSUoJUGZB9wHEGNcATSGEEYWj40F2mKMrTHGHPAQcNZe138K\n+Ewf1fsqr4RCowrLx9yOXpIkSZIkqUelLB+bACzv8ry1+NqW4uOGEEILsBo4A3ik48QQwknACzHG\ndT29SWNjPalUTcmFd9j+p20AtEw6gmzuBUbU19LU1NDr+/SXSqqlQyXWBNbVW5VYVyXWBNYlSZIk\nSftS0kyhvXS24sQY8yGESyksKdsMPNv1OHAl8I1Sbrpp0/YDKAXWbm4llUyxux12p7Mk8tDa2n5A\n9+prTU0NFVNLh0qsCayrtyqxrkqsCaqjLsMjSZIkSf2hlFBoDYXOoA7NwNqOJzHGJcDpACGEhRQ6\nhjrMBT50sEV2Z+POTYypbySZSJLJ5pwpJEmSJEmSVIJSEpSfAhcChBBmAWtijJ3/vB1CWBxCGBdC\nGA7MBx4svt4MbI0x7u77sgt2Z9O0p7fSVD+aXD5PJpsn5e5jkiRJkiRJPeqxUyjGuDSEsDyEsBTI\nAR8MIVwGbI4xfg/4GoXgKA8sjDGuL146EXi5f8ouyOQyJEhw+MiJZLM5AFJ2CkmSJEmSJPWopJlC\nMcbr9nrpiS7HFgGL9nHNcuCcg6quB/W1w/jEG/4nRx9+BK3rdgJQa6eQJEmSJElSjwZ8gjJpxETq\na4eRsVNIkiRJkiSpZFWToKQzhVCotibRw5mSJEmSJEmqmlCos1PI5WOSJEmSJEk9qpoEJe3yMUmS\nJEmSpJJVTYLS0SnkoGlJkiRJkqSeVU2C0jlTyE4hSZIkSZKkHlVNgpLJOFNIkiRJkiSpVFWToKSz\neQBS7j4mSZIkSZLUo6oJhTpnCqVqylyJJEmSJElS5auaUKhzppCdQpIkSZIkST2qmlCoo1PImUKS\nJEmSJEk9q5oEJd0RCrn7mCRJkiRJUo+qJkHJdC4fq5qPJEmSJEmS1G+qJkGxU0iSJEmSJKl0VZOg\n2CkkSZIkSZJUuqpJUNLZPAApdx+TJEmSJEnqUdWEQh27j9WmaspciSRJkiRJUuWrmlAonenYkt5O\nIUmSJEmSpJ5UTSj0SqdQ1XwkSZIkSZKkflM1CUqms1Ooaj6SJEmSJElSv6maBCVtp5AkSZIkSVLJ\nqiZBSdspJEmSJEmSVLKqSVAyxS3paw2FJEmSJEmSelQ1CUrHoOlUyt3HJEmSJEmSelI1oVA6myOZ\nSFCTrJqPJEmSJEmS1G9SpZwUQrgVOAXIA9fGGJd1ObYAuB7YBdwTY7yt+PrFwN8AGeBTMcb7+7j2\nPaQzObuEJEmSJEmSStRjW00IYQ7QEmOcDVwBfKXLsSRwG3Au8EZgfgjh8BDCGOAG4C+A84AF/VD7\nHjLZnPOEJEmSJEmSSlRKp9A84D6AGOOKEEJjCGFkjHELMBZoizG2AoQQHgLOAnYAD8YY24F24Kp+\nqb6LTCbnzmOSJEmSJEklKiUUmgAs7/K8tfjaluLjhhBCC7AaOAN4pHhefQjhB0AjcGOM8aHu3qSx\nsZ5UqqZXxXeVA4YOTdHU1HDA9+gPlVYPVGZNYF29VYl1VWJNYF2SJEmStC8lzRTaS+fgnhhjPoRw\nKXAnsBl4tsvxMcD5wBTgZyGEKTHG/P5uumnT9gMopaCpqYFduzLU19XS2tp+wPfpa01NDRVVD1Rm\nTWBdvVWJdVViTVAddRkeSZIkSeoPpYRCayh0BnVoBtZ2PIkxLgFOBwghLKTQMTQMWBpjzABPhxDa\ngSbg5b4p+9XS2bzLxyRJkiRJkkpUSij0U+Am4PYQwixgTXFWEAAhhMXApcA2YD5wCzAU+EYI4QsU\nlo+NANb3ce17yGRz1Lr7mCRJkiRJUkl6DIVijEtDCMtDCEspjO75YAjhMmBzjPF7wNcoBEd5YGGM\ncT1ACOE/gF8Vb/OhGGOuPz4AQD6fJ5Nx9zFJkiRJkqRSlTRTKMZ43V4vPdHl2CJg0T6uuR24/aCq\nK1EmmycPpFKGQpIkSZIkSaWoihQlnckCOFNIkiRJkiSpRFWRoqQzhZVpLh+TJEmSJEkqTVWkKJls\nMRRy+ZgkSZIkSVJJqiJF2Z0uhEIuH5MkSZIkSSpNVaQonTOF7BSSJEmSJEkqSVWkKB0zhVI1iTJX\nIkmSJEmSNDBURyjkTCFJkiRJkqReqYoUJZ129zFJkiRJkqTeqIoU5ZXlY1XxcSRJkiRJkvpdVaQo\nnYOmDYUkSZIkSZJKUhUpijOFJEmSJEmSeqcqUpTdaUMhSZIkSZKk3qiKFMUt6SVJkiRJknqnKkKh\njDOFJEmSJEmSeqUqUpTOmUKGQpIkSZIkSSWpihTFmUKSJEmSJEm9UxUpyiszhari40iSJEmSJPW7\nqkhR0h0zhewUkiRJkiRJKklVpCjOFJIkSZIkSeqdqkhR0sWZQnYKSZIkSZIklaYqUpSOmUK1NYky\nVyJJkiRJkjQwVFUo5KBpSZIkSZKk0lRFipLOOmhakiRJkiSpN6oiRdmddtC0JEmSJElSb1RFipLp\nmClkp5AkSZIkSVJJUqWcFEK4FTgFyAPXxhiXdTm2ALge2AXcE2O8LYQwF/h34A/F034fY/xQXxbe\nVceW9DVJB01LkiRJkiSVosdQKIQwB2iJMc4OIcwA7gRmF48lgduAWcAGYHEI4b7ipUtijBf2T9l7\nSmeypGqSJBKGQpIkSZIkSaUoZb3VPOA+gBjjCqAxhDCyeGws0BZjbI0x5oCHgLP6pdJu7E7nXDom\nSZIkSZLUC6UsH5sALO/yvLX42pbi44YQQguwGjgDeKT4eGYI4QfAaOCmGOMD3b1JY2M9qVRNL8sv\nSGdyDK2toamp4YCu70/WVDrr6p1KrKsSawLrkiRJkqR9KWmm0F4612jFGPMhhEspLCnbDDxbPL4S\nuAn4DnAk8LMQwvQY4+793XTTpu0HUEpBOpsjmYTW1vYDvkd/aGpqsKYSWVfvVGJdlVgTVEddhkeS\nJEmS+kMpodAaCp1BHZqBtR1PYoxLgNMBQggLgdUxxpeAe4unPB1CWAdMohAa9blMcaaQJEmSJEmS\nSlNKkvJT4EKAEMIsYE2MsfOft0MIi0MI40IIw4H5wIMhhItDCB8rHp8AjAde6vPqi5wpJEmSJEmS\n1Ds9JikxxqXA8hDCUuArwAdDCJeFEM4vnvI1CsHRL4CFMcb1wA+AOSGE/wK+D1zd3dKxg5XO5uwU\nkiRJkiRJ6oWSZgrFGK/b66UnuhxbBCza6/x2Cl1Dh0Q6k6PWUEiSJEmSJKlkAz5JyeZy5HJ5UjWJ\nnk+WJEmSJEkSUAWhUCaTB6D2ALezlyRJkiRJGowGfCiUzuYA7BSSJEmSJEnqhQEfCmWKoZC7j0mS\nJEmSJJVuwCcpmUxHp9CA/yiSJEmSJEmHzIBPUtJ2CkmSJEmSJPXagE9S0nYKSZIkSZIk9dqAT1Iy\n2eLuY4ZCkiRJkiRJJRvwSUrHoOlUyt3HJEmSJEmSSjXgQ6GO5WN2CkmSJEmSJJVuwCcp6c5OoQH/\nUSRJkiRJkg6ZAZ+kuCW9JEmSJElS7w34JKVjppDLxyRJkiRJkko34JOUzplCLh+TJEmSJEkq2YBP\nUjp3H7NTSJIkSZIkqWQDPklJZ/OAoZAkSZIkSVJvDPgkpXOmUCpR5kokSZIkSZIGjlS5CzhYnTOF\n7BSSVPQP/3ArMa5g48YN7N69iwkTmhk58jD+7u++2O11P/rRfzJ8+AjmzDnjEFUqSZIkSeUz4EOh\nzplCDpqWVPShD30EKIQ869a9wOWXf6Ck6849d35/liVJkiRJFWXAh0IdnULOFJLUnccee5R77vkm\n27dv55prPsLjjy/nkUceIpfLMXv2aVx++VXcccftjBo1imnTjmLRou+QSCR57rlnmTt3HpdfflW5\nP4IkSZIk9akBHwp1zhQyFJIqznceXsWyP77cp/c86ZhxXHTm9AO69umnV/Fv/7aIIUOG8Pjjy/nH\nf/w6yWSSiy5awDve8e49zn3qqT/w7W9/l1wux9vfPt9QSJIkSVLVGfChUGenkMvHJPVg+vQWhgwZ\nAkBdXR3XXHMVNTU1tLW1sWXLlj3ODeEY6urqylGmJEmSJB0SAz4UslNIqlwXnTn9gLt6+kNtbS0A\n69at5d57v8Wdd36L+vp6LrnkoledW1NTc6jLkyRJkqRDasAnKelsHrBTSFLp2traaGxspL6+nhj/\nyLp160in0+UuS5IkSZIOqZI6hUIItwKnAHng2hjjsi7HFgDXA7uAe2KMt3U5Ngx4EvhMjPEbfVh3\np0znlvSJ/ri9pCrU0nI0w4bVc/XVl/Pa176eBQsu4JZbvsBxx72u3KVJkiRJ0iHTYygUQpgDtMQY\nZ4cQZgB3ArOLx5LAbcAsYAOwOIRwX4zxxeLl1wMb+6XyonTW3cck7du5586nqamB1tZ2Zs06kVmz\nTgQKS8O+9KXbur2241yA++9/qF/rlCRJkqRyKCVJmQfcBxBjXAE0hhBGFo+NBdpijK0xxhzwEHAW\nQAjhGGAmcH+fV91FxkHTkiRJkiRJvVZKkjIBaO3yvLX4WsfjhhBCSwihFjgDGF88dgvw0b4qdH8y\n2RypmgTJhMvHJEmSJEmSSnUgu491pi8xxnwI4VIKS8o2A88CiRDCe4FfxhifDSGUdNPGxnpSqd7v\n9pNPJKhNJWlqauj1tYdCJdZViTWBdfVWJdZViTWBdUmSJEnSvpQSCq3hlc4ggGZgbceTGOMS4HSA\nEMJCYDVwPnBkCOE84HBgVwjhxRjjg/t7k02btve6eIAdO9OkampobW0/oOv7U8csk0pSiTWBdfVW\nJdZViTVBddRleCRJkiSpP5QSCv0UuAm4PYQwC1gTY+z8JhNCWAxcCmwD5gO3xBjv6XL8RmB1d4HQ\nwchkcwypdZ6QJEmSJElSb/QYCsUYl4YQlocQlgI54IMhhMuAzTHG7wFfoxAc5YGFMcb1/Vnw3jLZ\nPEOH9H7ZmSRJkiRJ0mBW0kyhGON1e730RJdji4BF3Vx74wFVVqJ0JseI+tr+fAtJA8j73/8+PvKR\nv+F5BHerAAAJoElEQVSYY2Z0vvbP/3wbhx02ine96z17nPvYY4+yaNF3+Oxnb+a66z7K5z//pT2O\nf/e799LW1sYVV7x/n++1atVKhgwZwuTJU7jhhk/wt397A0OH1vX9h5IkSZKkfjDg112lszlqa+wU\nklRw9tlv5uGHH9jjtUceeZizznpTt9ftHQiVYsmSh3nhhecBuOmmhQZCkiRJkgaUA9l9rKJkMjlq\nnSkkqWjevDdx9dVX8IEPfBiAJ598kqamJlavfpbrr/84tbW1NDQ08OlPf36P69761nncf/9DPPro\nb/jKV25h9OgxjBkzlubmSWQyGT73uRtpbX2ZHTt2cPnlVzFhwkS+//1FLFnyMI2NjXzqU5/g7rvv\nZevWdhYu/DTpdJpkMsl1132SRCLB5z53I83Nk1i1aiVHHx245Zaby/HjkSRJkqROAzoUyuXzZHN5\nalOGQlIlWrTqhzz+8u/79J7Hj3stF0w/b7/HGxtH09w8iaeeepKZM49l8eLFnH32W2hvb+eGGz5L\nc/MkPvOZT/HrX/+S+vr6V11/++238clPfoaWlqP52Mc+THPzJNrbt/CGN5zCOeecx0svvcgnP3kd\nd975TU4+eTZz585j5sxjO6//+tf/mfPOW8C8eW/iZz97kDvv/BeuuOL9xLiCm276OxobR3P++eey\nZcsWINGnPxtJkiRJ6o0BnaZkszkAamsG9MeQ1MfOPvstPPRQYQnZww8/zNy58xg1ahRf+MJnueaa\nq3j88eVs2bJ5n9euXbuWlpajAXj962cB0NAwkhUr/sDVV1/O5z53436vBYhxBccffwIAs2adyMqV\nEYBJk45gzJixJJNJxo5tor29tO3oJUmSJKm/DOhOoXSmGAqlnCkkVaILpp/XbVdPf5kz5wzuvvtO\nzj77zUydOpWRI0eycOFn+OIXv8zUqdP40pe+sN9rk8lXQuZ8Pg/AAw/8mC1btvDVr36dLVu2cOWV\nl3Tz7onO69LpDIlE4X41e80+6zhHkiRJksplQLfYpLOFL1XOFJLUVX39cI46qoW7776L+fPnA7Bt\n21bGj59Ae3s7jz22nHQ6vc9rx45t4vnnV5PP53n88eUAtLW1MXFiM8lkkiVLHu68NpFIkM1m97h+\nxoyZPPbYowD89rfL99gFTZIkSZIqyYBOUzKdnUID+mNI6gdnn/0Wli37NWeeeSYAF1zwdq6++gpu\nvvlzXHzxe/nmN7/Bhg3rX3XdVVd9gOuv/zgf//hHGDduPABz557J0qX/xbXXXs2wYcMYN24cd931\nNV73uuP58pe/yKOP/qbz+iuv/Gt+/OMf8eEP/zU/+tEP97udvSRJkiSVW6JSljC0trb3upB0Jsfn\nv7WcBXOmc9zUxv4o66A0NTXQ2lpZc0MqsSawrt6qxLoqsSaojrqamhqcSC1JkiSpzw3omUK1qSSf\nvPSkiv3SJ0mSJEmSVKlcdyVJkiRJkjQIGQpJkiRJkiQNQoZCkiRJkiRJg5ChkCRJkiRJ0iBkKCRJ\nkiRJkjQIGQpJkiRJkiQNQoZCkiRJkiRJg5ChkCRJkiRJ0iBkKCRJkiRJkjQIJfL5fLlrkCRJkiRJ\n0iFmp5AkSZIkSdIgZCgkSZIkSZI0CBkKSZIkSZIkDUKGQpIkSZIkSYOQoZAkSZIkSdIgZCgkSZIk\nSZI0CKXKXcDBCCHcCpwC5IFrY4zLylzPscD3gVtjjLeFEI4A/h9QA6wFLokx7jrENd0MnE7hd70Q\nWFYBNdUD3wDGA3XAZ4Anyl1XsbZhwJPFmh4qd00hhLnAvwN/KL70e+DmctdVrO1i4G+ADPAp4Hfl\nriuEcAVwSZeXTgROA/6Jwv8nfhdjvPpQ1lSsawRwN9AIDAVuAtaVuy5JkiRJg9uA7RQKIcwBWmKM\ns4ErgK+UuZ7hwD9QCBI6fBr4aozxdGAVcPkhrukM4Njiz+gtwJfLXVPRfODRGOMc4CLgSxVSF8D1\nwMbi40qpaUmMcW7xvw9VQl0hhDHADcBfAOcBCyqhrhjjHR0/q2J9/0rhz/21McbTgMNCCOcc6rqA\nywrlxTOAC4G/r5C6JEmSJA1iAzYUAuYB9wHEGFcAjSGEkWWsZxdwLrCmy2tzgR8UH/8ncNYhrunn\nwNuLj9uA4RVQEzHGe2OMNxefHgG8WAl1hRCOAWYC9xdfKntN+zGX8td1FvBgjLE9xrg2xnhVhdTV\n1aeALwDTunQRlquu9cCY4uNGCsFjJdQlSZIkaRAbyMvHJgDLuzxvLb62pRzFxBgzQCaE0PXl4V2W\nz7wMTDzENWWBbcWnVwA/At5czpq6CiEsBQ6n0GnyYAXUdQtwDXBp8XlZf39dzAwh/AAYTWHZUSXU\nNRWoL9bVCNxYIXUBEEI4CXiBwtK2TV0OlaWuGOM9IYTLQgirKPy85gNfLXddkiRJkga3gdwptLdE\nuQvoQdnqCyEsoBAKXbPXobL+zGKMpwJ/CXxzr1oOeV0hhPcCv4wxPrufU8r1s1pJIQhaQCGsuoM9\nw9xy1ZWg0PlyAYWlUXdR5t/hXq6kMLdqb2WpK4TwHuD5GON04EwKf+a7KvfPS5IkSdIgNJBDoTUU\nOoM6NFMYbltJthYHFwNMYs+lZYdECOHNwP8Bzokxbq6Qmk4oDuEmxvhbCiFHe5nreiuwIITwKwqB\nwiepgJ9VjPGl4nK7fIzxaQrDiRvLXRfwZ2BpjDFTrKud8v8Ou5oLLKXQQTimy+vlqus04CcAMcYn\ngGHA2AqoS5IkSdIgNpBDoZ9SGNhKCGEWsCbG2F7ekl7lQeCvio//CvjxoXzzEMJhwBeB82KMHcOT\ny1pT0RuB/wUQQhgPjCh3XTHGd8QYT4oxngJ8ncLuY2X/WYUQLg4hfKz4eAKFHdvuKnddFP7+nRlC\nSBaHTpf9d9ghhNAMbI0x7o4xpoE/hhD+onj4gjLVtQo4uVjfFAoh2ooKqEuSJEnSIJbI5/PlruGA\nhRA+TyFgyAEfLP4LfLlqOYHCTJqpQBp4CbiYwhKWOuA54H3FL6mHqqarKMx6+VOXly+lEHqUpaZi\nXcMoLIM6gkLHxE3AoxS27C5bXV3quxFYTaGzo6w1hRAagG8Do4AhFH5Wj5e7rmJt76ewLBHgs8Cy\nCqnrBOCzMcZzis9nArdTCMF/HWP8aBlqGgHcSSHUS1HoRFtX7rokSZIkDW4DOhSSJEmSJEnSgRnI\ny8ckSZIkSZJ0gAyFJEmSJEmSBiFDIUmSJEmSpEHIUEiSJEmSJGkQMhSSJEmSJEkahAyFJEmSJEmS\nBiFDIUmSJEmSpEHIUEiSJEmSJGkQ+v+y7lbJliyCUgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 1440x1440 with 10 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "yqyS73Ghvh9y",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Проверка на тестовой выборке"
      ]
    },
    {
      "metadata": {
        "id": "Wynvz6Cuvm9e",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def predictInTest(model, n_rec):\n",
        "  x = x_test[n_rec]\n",
        "  x = np.expand_dims(x, axis=0)\n",
        "\n",
        "  prediction = model.predict(x)\n",
        "  prediction = np.argmax(prediction)\n",
        "#   print(\"Predicted: [{}], expected [{}]\".format(prediction, y_test_org[n_rec]))\n",
        "  return prediction == y_test_org[n_rec]\n",
        "  \n",
        "def predictAllInTest(model):\n",
        "  correct = 0\n",
        "  n = len(x_test)\n",
        "  for i in range(0, n):\n",
        "    correct += 1 if predictInTest(model, i) else 0\n",
        "      \n",
        "  return correct"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "KY4HuhMyxW8I",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 187
        },
        "outputId": "7e175b2a-3751-426d-b77a-12b59a1bc3f7"
      },
      "cell_type": "code",
      "source": [
        "n = len(x_test)\n",
        "test_acc = []\n",
        "for res in results:\n",
        "  print(\"Checking \", res.config)\n",
        "  model = res.model\n",
        "  correct = predictAllInTest(model)\n",
        "  test_acc.append(correct / n)\n",
        "  "
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Checking  Dense [100], validation split 0.1, epochs 80, batch_size 100\n",
            "Checking  Dense [100, 200], validation split 0.1, epochs 80, batch_size 100\n",
            "Checking  Dense [200, 100], validation split 0.1, epochs 80, batch_size 100\n",
            "Checking  Dense [400], validation split 0.1, epochs 80, batch_size 100\n",
            "Checking  Dense [400, 100], validation split 0.1, epochs 80, batch_size 100\n",
            "Checking  Dense [400, 400], validation split 0.1, epochs 80, batch_size 100\n",
            "Checking  Dense [800, 800], validation split 0.1, epochs 80, batch_size 100\n",
            "Checking  Dense [800], validation split 0.1, epochs 80, batch_size 100\n",
            "Checking  Dense [1000], validation split 0.1, epochs 80, batch_size 100\n",
            "Checking  Dense [2000], validation split 0.1, epochs 80, batch_size 100\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "Lj-gp0_0Ec2f",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 221
        },
        "outputId": "e570aff6-fe82-4896-afcb-d4f8d7dfdf93"
      },
      "cell_type": "code",
      "source": [
        "print(test_acc)\n",
        "print(\"Accuracy on test set;Accuracy on training set;Validation accuracy;Model\")\n",
        "for i in range(0, len(results)):\n",
        "  config = results[i].config\n",
        "  history = results[i].history.history\n",
        "  print(\"{};{};{};{}\".format(test_acc[i], history[\"acc\"][-1], history[\"val_acc\"][-1], config))  \n"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[0.9688, 0.9823, 0.9805, 0.9837, 0.985, 0.9825, 0.9832, 0.9839, 0.9849, 0.9843]\n",
            "Accuracy on test set;Accuracy on training set;Validation accuracy;Model\n",
            "0.9688;0.9990925788879395;0.9700000286102295;Dense [100], validation split 0.1, epochs 80, batch_size 100\n",
            "0.9823;1.0;0.9853333234786987;Dense [100, 200], validation split 0.1, epochs 80, batch_size 100\n",
            "0.9805;0.9993888735771179;0.9851666688919067;Dense [200, 100], validation split 0.1, epochs 80, batch_size 100\n",
            "0.9837;0.9999259114265442;0.984666645526886;Dense [400], validation split 0.1, epochs 80, batch_size 100\n",
            "0.985;1.0;0.9851666688919067;Dense [400, 100], validation split 0.1, epochs 80, batch_size 100\n",
            "0.9825;0.9985740780830383;0.9860000014305115;Dense [400, 400], validation split 0.1, epochs 80, batch_size 100\n",
            "0.9832;0.9998148083686829;0.9829999804496765;Dense [800, 800], validation split 0.1, epochs 80, batch_size 100\n",
            "0.9839;1.0;0.9871666431427002;Dense [800], validation split 0.1, epochs 80, batch_size 100\n",
            "0.9849;1.0;0.9853333234786987;Dense [1000], validation split 0.1, epochs 80, batch_size 100\n",
            "0.9843;0.9992407560348511;0.9829999804496765;Dense [2000], validation split 0.1, epochs 80, batch_size 100\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "RG_QremWSEUJ",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Графики точностей нейросетей, выдававших точность больше 90%"
      ]
    },
    {
      "metadata": {
        "id": "0pCT65t1LLZ_",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 592
        },
        "outputId": "bd7395aa-d209-4cfb-981e-857a36c05822"
      },
      "cell_type": "code",
      "source": [
        "x_titles = list(map(lambda x: x.config.dense, results))\n",
        "y = test_acc * 100\n",
        "figure = plt.figure(figsize=(20, 10))\n",
        "\n",
        "acc_list = list(map(lambda x: x.history.history[\"acc\"][-1], results))\n",
        "val_acc_list = list(map(lambda x: x.history.history[\"val_acc\"][-1], results))\n",
        "xticks = x_titles\n",
        "\n",
        "plt.plot(acc_list, label=\"Training\")\n",
        "plt.plot(val_acc_list, label=\"Validation\")\n",
        "plt.plot(test_acc, label=\"Test\")\n",
        "plt.xticks(range(len(xticks)), xticks)\n",
        "plt.legend()\n",
        "plt.show()\n"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABIoAAAI/CAYAAAAYxjIJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n3nJRAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 1440x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "SERe1QjwSZKX",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Распознавание пользовательских изображений"
      ]
    },
    {
      "metadata": {
        "id": "FRGP1yocSg_f",
        "colab_type": "code",
        "colab": {
          "resources": {
            "http://localhost:8080/nbextensions/google.colab/files.js": {
              "data": "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",
              "ok": true,
              "headers": [
                [
                  "content-type",
                  "application/javascript"
                ]
              ],
              "status": 200,
              "status_text": ""
            }
          },
          "base_uri": "https://localhost:8080/",
          "height": 108
        },
        "outputId": "e991944e-7444-4959-a04c-b4e6c016ab06"
      },
      "cell_type": "code",
      "source": [
        "files.upload()"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "\n",
              "     <input type=\"file\" id=\"files-2fb2e9d8-ccba-423a-bdaf-fbff9d9adb59\" name=\"files[]\" multiple disabled />\n",
              "     <output id=\"result-2fb2e9d8-ccba-423a-bdaf-fbff9d9adb59\">\n",
              "      Upload widget is only available when the cell has been executed in the\n",
              "      current browser session. Please rerun this cell to enable.\n",
              "      </output>\n",
              "      <script src=\"/nbextensions/google.colab/files.js\"></script> "
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Saving 8.png to 8.png\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'8.png': b'\\x89PNG\\r\\n\\x1a\\n\\x00\\x00\\x00\\rIHDR\\x00\\x00\\x00\\x1c\\x00\\x00\\x00\\x1c\\x08\\x02\\x00\\x00\\x00\\xfdoH\\xc3\\x00\\x00\\x00\\tpHYs\\x00\\x005\\xd4\\x00\\x005\\xd4\\x01^e\\xe5\\x08\\x00\\x00\\x03\\xbdIDATH\\x89\\x8dV;KsM\\x10\\x9e\\xddsr\\x85$\\x10\\x14\\xbc5\\x06\\x14\\xad\\x14\\x0b\\xb10D\\xb1\\x14\\xc4* \\xf6b\\x91\\xc6\\x1f AE\\x02)R\\x88\\x88A\\xd2\\x88Xh\\x10,l$\\x08\\x16\\xb1\\x10D-,\\x82\\x1a\\x14A1h\\x82F0\\x97\\xbd|\\xc5\\xbc\\xef\\xba\\xef1\\xca7\\xd5\\x9egg\\x9f}vfv\\xf6\\x10)%42)%!D\\x08A)\\xc5O\\x00 \\x84\\x00\\x00\\x82RJ)\\xa5\\x9a\\xc5)4\\xf2\\x13\\xa9\\xc5\\x14\\xbb\\xbe\\r!D\\xdfL\\x19\\xfd\\x85\\x881\\xc69\\x07\\x00\\xce\\xb9\\x10\\x02\\x89(\\xa5\\x88##!\\xc4\\xc2\\xf8\\x9bR\\xce\\xb9a\\x18j\\xc09\\xbf\\xbe\\xbe\\xce\\xe5r\\x85B\\xc1\\xeb\\xf5\\x0e\\x0e\\x0evvv*\\xb1*\\x0e_\\xb1\\xfb\\xc98\\xe7\\xb7\\xb7\\xb7KKK###\\x0e\\x87\\x03\\x00l6[ss\\xb3\\xdb\\xedv\\xbb\\xdd\\xa9TJ\\xb9Y\\x16\\xfeF\\x1a\\x8b\\xc5(\\xa5\\xed\\xed\\xed\\xb3\\xb3\\xb3\\xbb\\xbb\\xbb\\xe7\\xe7\\xe7\\xaf\\xaf\\xafR\\xca\\xf7\\xf7\\xf7\\xb9\\xb9\\xb9\\xee\\xeen\\xc6\\x18cLJ)\\x84\\xf8_\\xa4\\xd9l\\xd60\\x8c\\xf5\\xf5\\xf5j\\xb5*\\x84\\xa8\\xd7\\xeb\\x88\\x0b!\\x84\\x10\\xd9l\\x96\\x10R*\\x95\\x1a*\\xfd1Q\\x9f\\x9f\\x9f\\x9c\\xf3\\xe9\\xe9i\\xbb\\xdd.\\xa54MS\\xc5\\x8a\\x10\\xa2\\x0bl\\xb0X\\xdfA?E>\\x9f\\x07\\x80L&\\x83\\xa0\\x10B)\\xaa\\xd5j\\xe1p8\\x14\\n}\\xd7\\xf8u\\xfcZ\\xad\\xa6C\\xca5\\x1c\\x0e\\x0f\\x0f\\x0f3\\xc6\\xf4\\xcd\\x9e\\x9e\\x9e\"\\x91\\x08!\\xe4\\xe8\\xe8H\\xe1\\xd6\\x98b\\x8c\\xa4\\x94z\\xd4\\x907\\x9b\\xcdRJ\\x0f\\x0e\\x0e\\xa4\\x94\\x8c\\xb1b\\xb1\\xb8\\xb6\\xb6\\xe6\\xf5z[[[\\x8f\\x8f\\x8f\\x1b\\xd2})\\xd5MQ\\xe3\\x82\\xa9\\xa9\\xa9\\x9e\\x9e\\x9e\\xc5\\xc5\\xc5x<\\xde\\xd2\\xd2\\xe2p8\\xe2\\xf1x\\xb1X\\xc4Y\\x8c\\xac\\xfc\\x96+\\xe0\\x9c\\xe3\\r\\xd1\\xe7\\xd4\\xfewww~\\xbf\\x9f\\x10\\xe2r\\xb9\\x16\\x16\\x16\\x9e\\x9f\\x9f\\xd1M\\x17\\xf8],\\xa5\\x94\\x1a\\x86A)U\\x97\\x9a1\\x867\\xef\\xe2\\xe2brr\\xb2T*\\x05\\x83\\xc1J\\xa5\\xf2\\xf2\\xf2\\xe2\\xf7\\xfb9\\xe7\\xe8V\\xaf\\xd7\\xb1\\x154\\xbe\\x8e\\xb83\\xca\\xc4\\x13q\\xce\\xb7\\xb7\\xb7\\x9dN\\xe7\\xd8\\xd8X>\\x9f\\x97R\\xa6\\xd3i\\x9b\\xcd6??\\xaf\\x9fTiTq\\xb0\\xc6T?{*\\x95\\x02\\x80h4\\x8a\\x95\\x8f\\xf8\\xc6\\xc6\\x06\\x00lnn*g\\xbd\\xce\\xac\\xa4:\\x9d\\x10bee\\x05\\x00\\x12\\x89\\xc4\\xf7xE\\xa3Q\\x9f\\xcf\\xf7\\xf8\\xf8\\xa8\\xa6,\\x1a\\xbf\\x12\\xa5\\xf3nmm\\x01@2\\x99\\xd4\\x19\\xd5\\xe0\\xed\\xed\\xad\\xa9\\xa9)\\x1a\\x8d\\xea\\xf8o\\r\\x85s~rrb\\x9a\\xe6\\xf2\\xf2\\xb2^\\xf0J\\x11F?\\x91H\\xb4\\xb5\\xb5)\\x81\\xbf)E\\x9b\\x98\\x98\\x18\\x1f\\x1fG\\xbf\\x9fj\\xf0\\xf0\\xf0\\x10\\x00\\x8a\\xc5\\xa2\\xee`\\x89\\x12U\\xaf\\r\\x00\\xe4r\\xb9`0\\x88\\xe5e\\x18F\\xb5Z\\x05\\x00\\xfcTu\\xe2t:\\x9dN\\'\\xf6f|\\x17\\xe4\\xbf\\x0f\\x14\\x00\\x98\\xaa\\xc3\\x0b!\\xfa\\xfb\\xfbOOO\\xd1CJ\\x89\\x8d\\x19\\xfe\\xbez\\xe8\\xb6\\xb3\\xb3\\x13\\x08\\x04\\x1c\\x0e\\x87i\\x9a\\nW5\\xfe\\xd5\\xa5TC\\xd9\\xdf\\xdf7\\x0c\\xe3\\xf2\\xf2\\xb2a\\x068\\xe7777\\x84\\x90t:\\xad\\x83\\xb2aCQ\\t\\xa9\\xd5j\\xa3\\xa3\\xa3CCCggg:\\x8e\\xe3r\\xb9\\x1c\\x89D:::*\\x95J\\xc3>b%U\\xa2\\xae\\xae\\xaeB\\xa1\\x90i\\x9a\\xc9d\\xb2\\\\.#x\\x7f\\x7f\\xbf\\xba\\xba\\xda\\xd5\\xd5\\xe5\\xf5z\\xf7\\xf6\\xf6,\\x14\\x8dK\\xcaR\\xc6\\x1f\\x1f\\x1f\\xb1X\\xccn\\xb7\\xfb|\\xbe\\xde\\xde\\xde@ \\x00\\x00\\x1e\\x8fgff&\\x9f\\xcf\\xeb\\xce?\\xe9%R\\xcaz\\xbdn\\xb3\\xd9\\xe0\\xdf\\xdf\\x90\\x87\\x87\\x87L&S(\\x14\\\\.W__\\xdf\\xc0\\xc0\\x80\\xc7\\xe3\\xc14`Z\\x90Q%YO\\xd4\\x9fw\\xff\\xfb\\x0f\\x88j7\\x84\\x10\\xd5\\xbdL\\xd3d\\x8c\\x19\\x86A\\x08\\xc1\\xb2\\xb1\\xe6\\x1d\\x00\\x00\\xfe\\x03`\\x96\\x0b\\x88\\x08P\\xbem\\x00\\x00\\x00\\x00IEND\\xaeB`\\x82'}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "metadata": {
        "id": "7dOfilSdSp8T",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 435
        },
        "outputId": "30f447db-f9be-4a04-abbe-c3e701a5c228"
      },
      "cell_type": "code",
      "source": [
        "img_path = '6.png'\n",
        "img = image.load_img(img_path, target_size=(28, 28), color_mode = \"grayscale\")\n",
        "plt.imshow(img.convert('RGBA'))\n",
        "plt.show()\n",
        "\n",
        "# Преобразуем картинку в массив\n",
        "x = image.img_to_array(img)\n",
        "# Меняем форму массива в плоский вектор\n",
        "x = x.reshape(1, 784)\n",
        "# Инвертируем изображение\n",
        "x = 255 - x\n",
        "# Нормализуем изображение\n",
        "x /= 255\n",
        "\n",
        "for res in results:\n",
        "  prediction = res.model.predict(x)\n",
        "  print(\"{} | {}\".format(np.argmax(prediction), res.config))"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPoAAAD4CAYAAADFJPs2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "3 | Dense [100], validation split 0.1, epochs 80, batch_size 100\n",
            "3 | Dense [100, 200], validation split 0.1, epochs 80, batch_size 100\n",
            "3 | Dense [200, 100], validation split 0.1, epochs 80, batch_size 100\n",
            "3 | Dense [400], validation split 0.1, epochs 80, batch_size 100\n",
            "3 | Dense [400, 100], validation split 0.1, epochs 80, batch_size 100\n",
            "3 | Dense [400, 400], validation split 0.1, epochs 80, batch_size 100\n",
            "3 | Dense [800, 800], validation split 0.1, epochs 80, batch_size 100\n",
            "3 | Dense [800], validation split 0.1, epochs 80, batch_size 100\n",
            "3 | Dense [1000], validation split 0.1, epochs 80, batch_size 100\n",
            "3 | Dense [2000], validation split 0.1, epochs 80, batch_size 100\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "rn9WIBwmyWQT",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Результаты"
      ]
    },
    {
      "metadata": {
        "id": "6ho0LqzRy0ND",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Точность обучения и распознавания"
      ]
    },
    {
      "metadata": {
        "id": "xWgamMlWvjna",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### 18 epochs"
      ]
    },
    {
      "metadata": {
        "id": "xHzUpI2AQIIH",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Batch size 100"
      ]
    },
    {
      "metadata": {
        "id": "uJtbWscExLkt",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "| Accuracy on test set | Accuracy on training set | Validation accuracy | Model                                                               | \n",
        "|----------------------|--------------------------|---------------------|---------------------------------------------------------------------| \n",
        "| 0.9831               | 0.9998148083686829       | 0.984499990940094   | Dense [400], validation split 0.1, epochs 18, batch_size 100        | \n",
        "| 0.9827               | 0.998962938785553        | 0.9828333258628845  | Dense [2000], validation split 0.1, epochs 18, batch_size 100       | \n",
        "| 0.9817               | 0.9961481690406799       | 0.9818333387374878  | Dense [800, 800], validation split 0.1, epochs 18, batch_size 100   | \n",
        "| 0.9812               | 0.9970740675926208       | 0.9836666584014893  | Dense [400, 400], validation split 0.1, epochs 18, batch_size 100   | \n",
        "| 0.9811               | 0.9970555305480957       | 0.9826666712760925  | Dense [800], validation split 0.1, epochs 18, batch_size 100        | \n",
        "| 0.9804               | 0.9980370402336121       | 0.98416668176651    | Dense [1000], validation split 0.1, epochs 18, batch_size 100       | \n",
        "| 0.9798               | 0.9972777962684631       | 0.9823333621025085  | Dense [200, 100], validation split 0.1, epochs 18, batch_size 100   | \n",
        "| 0.9784               | 0.9976111054420471       | 0.9800000190734863  | Dense [100, 200], validation split 0.1, epochs 18, batch_size 100   | \n",
        "| 0.9778               | 0.9973148107528687       | 0.9773333072662354  | Dense [100], validation split 0.1, epochs 18, batch_size 100        | \n",
        "| 0.9763               | 0.996833324432373        | 0.9786666631698608  | Dense [400, 100], validation split 0.1, epochs 18, batch_size 100   | \n",
        "| 0.9285               | 0.9447592496871948       | 0.9455000162124634  | Dense [10, 10, 10], validation split 0.1, epochs 18, batch_size 100 | \n",
        "| 0.9171               | 0.9370185136795044       | 0.9448333382606506  | Dense [10, 10], validation split 0.1, epochs 18, batch_size 100     | \n",
        "| 0.9115               | 0.9407036900520325       | 0.9468333125114441  | Dense [10], validation split 0.1, epochs 18, batch_size 100         | \n",
        "| 0.1981               | 0.3379444479942322       | 0.3551666736602783  | Dense [1], validation split 0.1, epochs 18, batch_size 100          | \n",
        "| 0.1135               | 0.11318518221378326      | 0.10499999672174454 | Dense [1, 1, 1], validation split 0.1, epochs 18, batch_size 100    | \n"
      ]
    },
    {
      "metadata": {
        "id": "3djmzlwKQWYg",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Batch size 200"
      ]
    },
    {
      "metadata": {
        "id": "KbW6-1DvQZbZ",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "| Accuracy on test set | Accuracy on training set | Validation accuracy | Model                                                               | \n",
        "|----------------------|--------------------------|---------------------|---------------------------------------------------------------------| \n",
        "| 0.9857               | 1.0                      | 0.9865000247955322  | Dense [2000], validation split 0.1, epochs 18, batch_size 200       | \n",
        "| 0.9835               | 0.9999814629554749       | 0.9853333234786987  | Dense [1000], validation split 0.1, epochs 18, batch_size 200       | \n",
        "| 0.9824               | 0.9999629855155945       | 0.984666645526886   | Dense [400], validation split 0.1, epochs 18, batch_size 200        | \n",
        "| 0.9822               | 0.9993888735771179       | 0.9838333129882812  | Dense [400, 100], validation split 0.1, epochs 18, batch_size 200   | \n",
        "| 0.9812               | 0.998462975025177        | 0.9815000295639038  | Dense [800], validation split 0.1, epochs 18, batch_size 200        | \n",
        "| 0.9796               | 0.9964814782142639       | 0.9789999723434448  | Dense [400, 400], validation split 0.1, epochs 18, batch_size 200   | \n",
        "| 0.9783               | 0.9977592825889587       | 0.9778333306312561  | Dense [100, 200], validation split 0.1, epochs 18, batch_size 200   | \n",
        "| 0.9779               | 0.9973703622817993       | 0.9800000190734863  | Dense [200, 100], validation split 0.1, epochs 18, batch_size 200   | \n",
        "| 0.9778               | 0.9947222471237183       | 0.9779999852180481  | Dense [100], validation split 0.1, epochs 18, batch_size 200        | \n",
        "| 0.9764               | 0.9978148341178894       | 0.9818333387374878  | Dense [800, 800], validation split 0.1, epochs 18, batch_size 200   | \n",
        "| 0.9204               | 0.9399444460868835       | 0.9409999847412109  | Dense [10, 10, 10], validation split 0.1, epochs 18, batch_size 200 | \n",
        "| 0.9137               | 0.9415925741195679       | 0.9466666579246521  | Dense [10, 10], validation split 0.1, epochs 18, batch_size 200     | \n",
        "| 0.9084               | 0.9356111288070679       | 0.9453333616256714  | Dense [10], validation split 0.1, epochs 18, batch_size 200         | \n",
        "| 0.1979               | 0.3476296365261078       | 0.34299999475479126 | Dense [1], validation split 0.1, epochs 18, batch_size 200          | \n",
        "| 0.1135               | 0.11318518221378326      | 0.10499999672174454 | Dense [1, 1, 1], validation split 0.1, epochs 18, batch_size 200    | \n"
      ]
    },
    {
      "metadata": {
        "id": "jYl__UuYQo15",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Batch size 300"
      ]
    },
    {
      "metadata": {
        "id": "IlmyjNqNQrdA",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "| Accuracy on test set | Accuracy on training set | Validation accuracy | Model                                                               | \n",
        "|----------------------|--------------------------|---------------------|---------------------------------------------------------------------| \n",
        "| 0.9835               | 0.9999629855155945       | 0.9860000014305115  | Dense [1000], validation split 0.1, epochs 18, batch_size 300       | \n",
        "| 0.9832               | 1.0                      | 0.9861666560173035  | Dense [2000], validation split 0.1, epochs 18, batch_size 300       | \n",
        "| 0.9824               | 0.9999814629554749       | 0.9853333234786987  | Dense [800], validation split 0.1, epochs 18, batch_size 300        | \n",
        "| 0.9821               | 0.9962592720985413       | 0.9825000166893005  | Dense [800, 800], validation split 0.1, epochs 18, batch_size 300   | \n",
        "| 0.9805               | 0.9996111392974854       | 0.9823333621025085  | Dense [400], validation split 0.1, epochs 18, batch_size 300        | \n",
        "| 0.9794               | 0.9976111054420471       | 0.9826666712760925  | Dense [400, 100], validation split 0.1, epochs 18, batch_size 300   | \n",
        "| 0.9782               | 0.9982222318649292       | 0.9794999957084656  | Dense [400, 400], validation split 0.1, epochs 18, batch_size 300   | \n",
        "| 0.9768               | 0.9988703727722168       | 0.9801666736602783  | Dense [200, 100], validation split 0.1, epochs 18, batch_size 300   | \n",
        "| 0.9747               | 0.9970555305480957       | 0.9756666421890259  | Dense [100, 200], validation split 0.1, epochs 18, batch_size 300   | \n",
        "| 0.9746               | 0.992981493473053        | 0.9789999723434448  | Dense [100], validation split 0.1, epochs 18, batch_size 300        | \n",
        "| 0.9193               | 0.9384444355964661       | 0.9461666941642761  | Dense [10, 10], validation split 0.1, epochs 18, batch_size 300     | \n",
        "| 0.9166               | 0.9363889098167419       | 0.9401666522026062  | Dense [10], validation split 0.1, epochs 18, batch_size 300         | \n",
        "| 0.9117               | 0.9333148002624512       | 0.9390000104904175  | Dense [10, 10, 10], validation split 0.1, epochs 18, batch_size 300 | \n",
        "| 0.1952               | 0.3007592558860779       | 0.3100000023841858  | Dense [1], validation split 0.1, epochs 18, batch_size 300          | \n",
        "| 0.1135               | 0.11318518221378326      | 0.10499999672174454 | Dense [1, 1, 1], validation split 0.1, epochs 18, batch_size 300    | \n"
      ]
    },
    {
      "metadata": {
        "id": "GFfFilTHvs-o",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### 30 epochs"
      ]
    },
    {
      "metadata": {
        "id": "eIU9ZrqPvw08",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Batch size 100"
      ]
    },
    {
      "metadata": {
        "id": "aZ0r5_0bIizO",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "| Accuracy on test set | Accuracy on training set | Validation accuracy | Model                                                               | \n",
        "|----------------------|--------------------------|---------------------|---------------------------------------------------------------------| \n",
        "| 0.9858               | 1.0                      | 0.9883333444595337  | Dense [2000], validation split 0.1, epochs 30, batch_size 100       | \n",
        "| 0.9834               | 0.9993888735771179       | 0.9829999804496765  | Dense [1000], validation split 0.1, epochs 30, batch_size 100       | \n",
        "| 0.9827               | 1.0                      | 0.9856666922569275  | Dense [400], validation split 0.1, epochs 30, batch_size 100        | \n",
        "| 0.9826               | 0.9976296424865723       | 0.984333336353302   | Dense [400, 400], validation split 0.1, epochs 30, batch_size 100   | \n",
        "| 0.9814               | 0.9986851811408997       | 0.9826666712760925  | Dense [200, 100], validation split 0.1, epochs 30, batch_size 100   | \n",
        "| 0.9813               | 0.9952407479286194       | 0.9818333387374878  | Dense [800], validation split 0.1, epochs 30, batch_size 100        | \n",
        "| 0.9809               | 0.9977962970733643       | 0.9791666865348816  | Dense [800, 800], validation split 0.1, epochs 30, batch_size 100   | \n",
        "| 0.9808               | 0.9975740909576416       | 0.9828333258628845  | Dense [400, 100], validation split 0.1, epochs 30, batch_size 100   | \n",
        "| 0.9780               | 0.9998703598976135       | 0.9825000166893005  | Dense [100], validation split 0.1, epochs 30, batch_size 100        | \n",
        "| 0.9752               | 0.9996851682662964       | 0.9768333435058594  | Dense [100, 200], validation split 0.1, epochs 30, batch_size 100   | \n",
        "| 0.9257               | 0.9486851692199707       | 0.9471666812896729  | Dense [10, 10], validation split 0.1, epochs 30, batch_size 100     | \n",
        "| 0.9093               | 0.9460370540618896       | 0.9474999904632568  | Dense [10], validation split 0.1, epochs 30, batch_size 100         | \n",
        "| 0.8822               | 0.9405184984207153       | 0.9403333067893982  | Dense [10, 10, 10], validation split 0.1, epochs 30, batch_size 100 | \n",
        "| 0.1963               | 0.4028148055076599       | 0.40966665744781494 | Dense [1], validation split 0.1, epochs 30, batch_size 100          | \n",
        "| 0.1963               | 0.39407408237457275      | 0.4065000116825104  | Dense [1, 1, 1], validation split 0.1, epochs 30, batch_size 100    | \n"
      ]
    },
    {
      "metadata": {
        "id": "OfH1fp6RIk-j",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Batch size 200"
      ]
    },
    {
      "metadata": {
        "id": "HFQl1e0DIxNR",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "| Accuracy on test set | Accuracy on training set | Validation accuracy | Model                                                               | \n",
        "|----------------------|--------------------------|---------------------|---------------------------------------------------------------------| \n",
        "| 0.9857               | 1.0                      | 0.9861666560173035  | Dense [400, 400], validation split 0.1, epochs 30, batch_size 200   | \n",
        "| 0.9854               | 1.0                      | 0.9871666431427002  | Dense [2000], validation split 0.1, epochs 30, batch_size 200       | \n",
        "| 0.9844               | 1.0                      | 0.9854999780654907  | Dense [1000], validation split 0.1, epochs 30, batch_size 200       | \n",
        "| 0.9838               | 1.0                      | 0.9868333339691162  | Dense [800], validation split 0.1, epochs 30, batch_size 200        | \n",
        "| 0.9821               | 0.9991481304168701       | 0.9838333129882812  | Dense [800, 800], validation split 0.1, epochs 30, batch_size 200   | \n",
        "| 0.9819               | 1.0                      | 0.984499990940094   | Dense [400], validation split 0.1, epochs 30, batch_size 200        | \n",
        "| 0.9772               | 0.9953333139419556       | 0.9789999723434448  | Dense [100, 200], validation split 0.1, epochs 30, batch_size 200   | \n",
        "| 0.9771               | 0.9861666560173035       | 0.9803333282470703  | Dense [400, 100], validation split 0.1, epochs 30, batch_size 200   | \n",
        "| 0.9754               | 0.9996296167373657       | 0.9786666631698608  | Dense [100], validation split 0.1, epochs 30, batch_size 200        | \n",
        "| 0.9746               | 0.9976296424865723       | 0.9773333072662354  | Dense [200, 100], validation split 0.1, epochs 30, batch_size 200   | \n",
        "| 0.9274               | 0.9452962875366211       | 0.9456666707992554  | Dense [10, 10], validation split 0.1, epochs 30, batch_size 200     | \n",
        "| 0.9184               | 0.9420555830001831       | 0.9508333206176758  | Dense [10], validation split 0.1, epochs 30, batch_size 200         | \n",
        "| 0.8905               | 0.9467037320137024       | 0.9449999928474426  | Dense [10, 10, 10], validation split 0.1, epochs 30, batch_size 200 | \n",
        "| 0.1965               | 0.37405556440353394      | 0.38316667079925537 | Dense [1], validation split 0.1, epochs 30, batch_size 200          | \n",
        "| 0.1135               | 0.11318518221378326      | 0.10499999672174454 | Dense [1, 1, 1], validation split 0.1, epochs 30, batch_size 200    | \n"
      ]
    },
    {
      "metadata": {
        "id": "iYf4MzHpIzFw",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Batch size 300"
      ]
    },
    {
      "metadata": {
        "id": "JQdbdwZEI2O2",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "| Accuracy on test set | Accuracy on training set | Validation accuracy | Model                                                               | \n",
        "|----------------------|--------------------------|---------------------|---------------------------------------------------------------------| \n",
        "| 0.9851               | 0.999740719795227        | 0.984499990940094   | Dense [800, 800], validation split 0.1, epochs 30, batch_size 300   | \n",
        "| 0.9839               | 1.0                      | 0.9865000247955322  | Dense [2000], validation split 0.1, epochs 30, batch_size 300       | \n",
        "| 0.9836               | 0.9999444484710693       | 0.9863333106040955  | Dense [400, 400], validation split 0.1, epochs 30, batch_size 300   | \n",
        "| 0.9834               | 1.0                      | 0.9853333234786987  | Dense [800], validation split 0.1, epochs 30, batch_size 300        | \n",
        "| 0.9829               | 1.0                      | 0.9851666688919067  | Dense [1000], validation split 0.1, epochs 30, batch_size 300       | \n",
        "| 0.9821               | 0.9993333220481873       | 0.9823333621025085  | Dense [400, 100], validation split 0.1, epochs 30, batch_size 300   | \n",
        "| 0.9812               | 1.0                      | 0.9836666584014893  | Dense [200, 100], validation split 0.1, epochs 30, batch_size 300   | \n",
        "| 0.9805               | 1.0                      | 0.9821666479110718  | Dense [400], validation split 0.1, epochs 30, batch_size 300        | \n",
        "| 0.9774               | 0.9993518590927124       | 0.9815000295639038  | Dense [100, 200], validation split 0.1, epochs 30, batch_size 300   | \n",
        "| 0.9755               | 0.9986851811408997       | 0.9773333072662354  | Dense [100], validation split 0.1, epochs 30, batch_size 300        | \n",
        "| 0.9317               | 0.9493333101272583       | 0.9523333311080933  | Dense [10, 10, 10], validation split 0.1, epochs 30, batch_size 300 | \n",
        "| 0.9107               | 0.9434629678726196       | 0.9480000138282776  | Dense [10, 10], validation split 0.1, epochs 30, batch_size 300     | \n",
        "| 0.9098               | 0.9384258985519409       | 0.9434999823570251  | Dense [10], validation split 0.1, epochs 30, batch_size 300         | \n",
        "| 0.2120               | 0.30335184931755066      | 0.30799999833106995 | Dense [1], validation split 0.1, epochs 30, batch_size 300          | \n",
        "| 0.2112               | 0.3151296377182007       | 0.3306666612625122  | Dense [1, 1, 1], validation split 0.1, epochs 30, batch_size 300    | \n"
      ]
    },
    {
      "metadata": {
        "id": "bIREKCPes-pE",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### 40 epochs"
      ]
    },
    {
      "metadata": {
        "id": "o2ALLlVvtCZW",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Batch size 100"
      ]
    },
    {
      "metadata": {
        "id": "81D0coeWtE13",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "| Accuracy on test set | Accuracy on training set | Validation accuracy | Model                                                               | \n",
        "|----------------------|--------------------------|---------------------|---------------------------------------------------------------------| \n",
        "| 0.9856               | 1.0                      | 0.9868333339691162  | Dense [800], validation split 0.1, epochs 40, batch_size 100        | \n",
        "| 0.9852               | 1.0                      | 0.9865000247955322  | Dense [1000], validation split 0.1, epochs 40, batch_size 100       | \n",
        "| 0.9852               | 0.9992407560348511       | 0.9854999780654907  | Dense [2000], validation split 0.1, epochs 40, batch_size 100       | \n",
        "| 0.9832               | 0.9999444484710693       | 0.98416668176651    | Dense [400], validation split 0.1, epochs 40, batch_size 100        | \n",
        "| 0.9819               | 0.998370349407196        | 0.981166660785675   | Dense [400, 100], validation split 0.1, epochs 40, batch_size 100   | \n",
        "| 0.9817               | 0.9978333115577698       | 0.9828333258628845  | Dense [800, 800], validation split 0.1, epochs 40, batch_size 100   | \n",
        "| 0.9813               | 0.9993703961372375       | 0.9793333411216736  | Dense [400, 400], validation split 0.1, epochs 40, batch_size 100   | \n",
        "| 0.9782               | 0.9969629645347595       | 0.9801666736602783  | Dense [100, 200], validation split 0.1, epochs 40, batch_size 100   | \n",
        "| 0.9754               | 0.9996481537818909       | 0.9818333387374878  | Dense [100], validation split 0.1, epochs 40, batch_size 100        | \n",
        "| 0.9730               | 0.9977037310600281       | 0.9771666526794434  | Dense [200, 100], validation split 0.1, epochs 40, batch_size 100   | \n",
        "| 0.9250               | 0.953499972820282        | 0.9524999856948853  | Dense [10, 10], validation split 0.1, epochs 40, batch_size 100     | \n",
        "| 0.9197               | 0.951796293258667        | 0.949833333492279   | Dense [10, 10, 10], validation split 0.1, epochs 40, batch_size 100 | \n",
        "| 0.9054               | 0.9457963109016418       | 0.9451666474342346  | Dense [10], validation split 0.1, epochs 40, batch_size 100         | \n",
        "| 0.1964               | 0.4018518626689911       | 0.40933331847190857 | Dense [1], validation split 0.1, epochs 40, batch_size 100          | \n",
        "| 0.1135               | 0.11318518221378326      | 0.10499999672174454 | Dense [1, 1, 1], validation split 0.1, epochs 40, batch_size 100    | \n"
      ]
    },
    {
      "metadata": {
        "id": "-xSaT3b6tOj7",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Batch size 200"
      ]
    },
    {
      "metadata": {
        "id": "B4kvVfPBtQ-8",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "| Accuracy on test set | Accuracy on training set | Validation accuracy | Model                                                               | \n",
        "|----------------------|--------------------------|---------------------|---------------------------------------------------------------------| \n",
        "| 0.9862               | 1.0                      | 0.987333357334137   | Dense [2000], validation split 0.1, epochs 40, batch_size 200       | \n",
        "| 0.9853               | 1.0                      | 0.987500011920929   | Dense [400, 400], validation split 0.1, epochs 40, batch_size 200   | \n",
        "| 0.9842               | 1.0                      | 0.9851666688919067  | Dense [400, 100], validation split 0.1, epochs 40, batch_size 200   | \n",
        "| 0.9838               | 1.0                      | 0.984499990940094   | Dense [1000], validation split 0.1, epochs 40, batch_size 200       | \n",
        "| 0.9835               | 1.0                      | 0.9856666922569275  | Dense [800], validation split 0.1, epochs 40, batch_size 200        | \n",
        "| 0.9829               | 0.9988148212432861       | 0.9838333129882812  | Dense [800, 800], validation split 0.1, epochs 40, batch_size 200   | \n",
        "| 0.9828               | 0.9987778067588806       | 0.9848333597183228  | Dense [400], validation split 0.1, epochs 40, batch_size 200        | \n",
        "| 0.9815               | 1.0                      | 0.984000027179718   | Dense [200, 100], validation split 0.1, epochs 40, batch_size 200   | \n",
        "| 0.9787               | 0.9999814629554749       | 0.9833333492279053  | Dense [100, 200], validation split 0.1, epochs 40, batch_size 200   | \n",
        "| 0.9764               | 0.9994999766349792       | 0.9786666631698608  | Dense [100], validation split 0.1, epochs 40, batch_size 200        | \n",
        "| 0.9226               | 0.9508888721466064       | 0.9513333439826965  | Dense [10, 10, 10], validation split 0.1, epochs 40, batch_size 200 | \n",
        "| 0.9170               | 0.9472963213920593       | 0.9456666707992554  | Dense [10, 10], validation split 0.1, epochs 40, batch_size 200     | \n",
        "| 0.9154               | 0.9455925822257996       | 0.9483333230018616  | Dense [10], validation split 0.1, epochs 40, batch_size 200         | \n",
        "| 0.2128               | 0.34983333945274353      | 0.3684999942779541  | Dense [1], validation split 0.1, epochs 40, batch_size 200          | \n",
        "| 0.1964               | 0.4054444432258606       | 0.40683332085609436 | Dense [1, 1, 1], validation split 0.1, epochs 40, batch_size 200    | \n"
      ]
    },
    {
      "metadata": {
        "id": "8ATJasIZtWgp",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Batch size 300"
      ]
    },
    {
      "metadata": {
        "id": "n1Xy9lkLtZBL",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "| Accuracy on test set | Accuracy on training set | Validation accuracy | Model                                                               | \n",
        "|----------------------|--------------------------|---------------------|---------------------------------------------------------------------| \n",
        "| 0.9850               | 1.0                      | 0.9858333468437195  | Dense [800, 800], validation split 0.1, epochs 40, batch_size 300   | \n",
        "| 0.9845               | 1.0                      | 0.9853333234786987  | Dense [400, 400], validation split 0.1, epochs 40, batch_size 300   | \n",
        "| 0.9833               | 1.0                      | 0.9838333129882812  | Dense [400, 100], validation split 0.1, epochs 40, batch_size 300   | \n",
        "| 0.9832               | 1.0                      | 0.9865000247955322  | Dense [2000], validation split 0.1, epochs 40, batch_size 300       | \n",
        "| 0.9831               | 1.0                      | 0.9860000014305115  | Dense [1000], validation split 0.1, epochs 40, batch_size 300       | \n",
        "| 0.9826               | 1.0                      | 0.9848333597183228  | Dense [800], validation split 0.1, epochs 40, batch_size 300        | \n",
        "| 0.9821               | 1.0                      | 0.98416668176651    | Dense [400], validation split 0.1, epochs 40, batch_size 300        | \n",
        "| 0.9809               | 1.0                      | 0.9826666712760925  | Dense [200, 100], validation split 0.1, epochs 40, batch_size 300   | \n",
        "| 0.9772               | 1.0                      | 0.9823333621025085  | Dense [100, 200], validation split 0.1, epochs 40, batch_size 300   | \n",
        "| 0.9762               | 0.9998703598976135       | 0.9804999828338623  | Dense [100], validation split 0.1, epochs 40, batch_size 300        | \n",
        "| 0.9146               | 0.9444074034690857       | 0.9491666555404663  | Dense [10], validation split 0.1, epochs 40, batch_size 300         | \n",
        "| 0.9031               | 0.9453703761100769       | 0.9473333358764648  | Dense [10, 10, 10], validation split 0.1, epochs 40, batch_size 300 | \n",
        "| 0.8998               | 0.9497777819633484       | 0.9524999856948853  | Dense [10, 10], validation split 0.1, epochs 40, batch_size 300     | \n",
        "| 0.2116               | 0.3800370395183563       | 0.3774999976158142  | Dense [1], validation split 0.1, epochs 40, batch_size 300          | \n",
        "| 0.1135               | 0.11318518221378326      | 0.10499999672174454 | Dense [1, 1, 1], validation split 0.1, epochs 40, batch_size 300    | \n"
      ]
    },
    {
      "metadata": {
        "id": "Ssu4ucOTwbOa",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## 20 лучших конфигураций"
      ]
    },
    {
      "metadata": {
        "id": "hq6uGZkGwmDr",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "| Accuracy on test set | Accuracy on training set | Validation accuracy | Model      | Validation split |  Epochs | Batch size | \n",
        "|----------------------|--------------------------|---------------------|------------|------------------|---------|------------| \n",
        "| 0.9862               | 1.0                      | 0.987333357334137   | [2000]     | 0.1              | 40      | 200        | \n",
        "| 0.9858               | 1.0                      | 0.9883333444595337  | [2000]     | 0.1              | 30      | 100        | \n",
        "| 0.9857               | 1.0                      | 0.9865000247955322  | [2000]     | 0.1              | 18      | 200        | \n",
        "| 0.9857               | 1.0                      | 0.9861666560173035  | [400, 400] | 0.1              | 30      | 200        | \n",
        "| 0.9856               | 1.0                      | 0.9868333339691162  | [800]      | 0.1              | 40      | 100        | \n",
        "| 0.9854               | 1.0                      | 0.9871666431427002  | [2000]     | 0.1              | 30      | 200        | \n",
        "| 0.9853               | 1.0                      | 0.987500011920929   | [400, 400] | 0.1              | 40      | 200        | \n",
        "| 0.9852               | 1.0                      | 0.9865000247955322  | [1000]     | 0.1              | 40      | 100        | \n",
        "| 0.9852               | 0.9992407560348511       | 0.9854999780654907  | [2000]     | 0.1              | 40      | 100        | \n",
        "| 0.9851               | 0.999740719795227        | 0.984499990940094   | [800, 800] | 0.1              | 30      | 300        | \n",
        "| 0.9850               | 1.0                      | 0.9858333468437195  | [800, 800] | 0.1              | 40      | 300        | \n",
        "| 0.9845               | 1.0                      | 0.9853333234786987  | [400, 400] | 0.1              | 40      | 300        | \n",
        "| 0.9844               | 1.0                      | 0.9854999780654907  | [1000]     | 0.1              | 30      | 200        | \n",
        "| 0.9842               | 1.0                      | 0.9851666688919067  | [400, 100] | 0.1              | 40      | 200        | \n",
        "| 0.9839               | 1.0                      | 0.9865000247955322  | [2000]     | 0.1              | 30      | 300        | \n",
        "| 0.9838               | 1.0                      | 0.9868333339691162  | [800]      | 0.1              | 30      | 200        | \n",
        "| 0.9838               | 1.0                      | 0.984499990940094   | [1000]     | 0.1              | 40      | 200        | \n",
        "| 0.9836               | 0.9999444484710693       | 0.9863333106040955  | [400, 400] | 0.1              | 30      | 300        | \n",
        "| 0.9835               | 0.9999814629554749       | 0.9853333234786987  | [1000]     | 0.1              | 18      | 200        | \n",
        "| 0.9835               | 0.9999629855155945       | 0.9860000014305115  | [1000]     | 0.1              | 18      | 300        | \n"
      ]
    },
    {
      "metadata": {
        "id": "0AjcuAjHwv5V",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## 20 худших конфигураций"
      ]
    },
    {
      "metadata": {
        "id": "U8yqZspTw0ed",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "| Accuracy on test set | Accuracy on training set | Validation accuracy | Model        | Validation split |  Epochs | Batch size | \n",
        "|----------------------|--------------------------|---------------------|--------------|------------------|---------|------------| \n",
        "| 0.1135               | 0.11318518221378326      | 0.10499999672174454 | [1, 1, 1]    | 0.1              | 18      | 100        | \n",
        "| 0.1135               | 0.11318518221378326      | 0.10499999672174454 | [1, 1, 1]    | 0.1              | 18      | 200        | \n",
        "| 0.1135               | 0.11318518221378326      | 0.10499999672174454 | [1, 1, 1]    | 0.1              | 18      | 300        | \n",
        "| 0.1135               | 0.11318518221378326      | 0.10499999672174454 | [1, 1, 1]    | 0.1              | 30      | 200        | \n",
        "| 0.1135               | 0.11318518221378326      | 0.10499999672174454 | [1, 1, 1]    | 0.1              | 40      | 300        | \n",
        "| 0.1135               | 0.11318518221378326      | 0.10499999672174454 | [1, 1, 1]    | 0.1              | 40      | 100        | \n",
        "| 0.1952               | 0.3007592558860779       | 0.3100000023841858  | [1]          | 0.1              | 18      | 300        | \n",
        "| 0.1963               | 0.4028148055076599       | 0.40966665744781494 | [1]          | 0.1              | 30      | 100        | \n",
        "| 0.1963               | 0.39407408237457275      | 0.4065000116825104  | [1, 1, 1]    | 0.1              | 30      | 100        | \n",
        "| 0.1964               | 0.4054444432258606       | 0.40683332085609436 | [1, 1, 1]    | 0.1              | 40      | 200        | \n",
        "| 0.1964               | 0.4018518626689911       | 0.40933331847190857 | [1]          | 0.1              | 40      | 100        | \n",
        "| 0.1965               | 0.37405556440353394      | 0.38316667079925537 | [1]          | 0.1              | 30      | 200        | \n",
        "| 0.1979               | 0.3476296365261078       | 0.34299999475479126 | [1]          | 0.1              | 18      | 200        | \n",
        "| 0.1981               | 0.3379444479942322       | 0.3551666736602783  | [1]          | 0.1              | 18      | 100        | \n",
        "| 0.2112               | 0.3151296377182007       | 0.3306666612625122  | [1, 1, 1]    | 0.1              | 30      | 300        | \n",
        "| 0.2116               | 0.3800370395183563       | 0.3774999976158142  | [1]          | 0.1              | 40      | 300        | \n",
        "| 0.2120               | 0.30335184931755066      | 0.30799999833106995 | [1]          | 0.1              | 30      | 300        | \n",
        "| 0.2128               | 0.34983333945274353      | 0.3684999942779541  | [1]          | 0.1              | 40      | 200        | \n",
        "| 0.8822               | 0.9405184984207153       | 0.9403333067893982  | [10, 10, 10] | 0.1              | 30      | 100        | \n",
        "| 0.8905               | 0.9467037320137024       | 0.9449999928474426  | [10, 10, 10] | 0.1              | 30      | 200        | \n"
      ]
    },
    {
      "metadata": {
        "id": "GDiAcWylygGM",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Выводы\n",
        "\n",
        "* При малом количестве нейронов точность распознавания оказывается низкой, добавление слоёв только ухудшает результат\n",
        "* При количестве нейронов в слое от 100, добавление слоёв может как повышать, так и понижать точность. \n",
        "* При увеличении числа эпох обучения качество распознавания многослойных сетей увеличивается\n",
        "* Наилучшие результаты получаются у сети с одним слоем в 800-2000 нейронов и двухслойные, обученные на 30-40 эпохаз\n",
        "* В большинстве случаев наилучшие результаты получаются при значении batch_size = 200\n",
        "* Точность распознавания тестового набора данных примерно соответствует точности при обучении (см. последний график)\n",
        "* При обучении заметны провалы точности на некоторых эпохах, которые происходят сразу и на тестовой, и на проверочной наборах. Стабильного падения точности на проверочных наборах, которое говорило бы о переобучении, не наблюдалось."
      ]
    }
  ]
}