Convolutional Neural Network.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "6a257acb-a09e-462f-b78f-e926360d0dd4",
   "metadata": {},
   "source": [
    "## Digit Recognition"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d7e28c0e",
   "metadata": {},
   "source": [
    "CNN for Kaggle competition.  Data avialable at: https://www.kaggle.com/competitions/digit-recognizer/overview"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1062f7b9-e88a-4c35-a2ef-e321f4908886",
   "metadata": {},
   "source": [
    "### Import Libraries and Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "b91f048f-8213-4694-b763-4cfbb84ee13b",
   "metadata": {},
   "outputs": [],
   "source": [
    "#loading libraries\n",
    "import pandas as pd\n",
    "import datetime\n",
    "import numpy as np\n",
    "import tensorflow\n",
    "import seaborn as sns\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.model_selection import RepeatedStratifiedKFold\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.cluster import KMeans\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense\n",
    "from keras.utils.np_utils import to_categorical\n",
    "from keras.layers import Conv2D, MaxPool2D, Dropout, Flatten\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "a630e009-17af-42ab-b92a-c0ab7edb5482",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(42000, 785)\n"
     ]
    }
   ],
   "source": [
    "#loading training data\n",
    "df = pd.read_csv(\"train.csv\")\n",
    "print(df.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "4c330542-c8bb-48f0-81a7-9ec94296e19c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>label</th>\n",
       "      <th>pixel0</th>\n",
       "      <th>pixel1</th>\n",
       "      <th>pixel2</th>\n",
       "      <th>pixel3</th>\n",
       "      <th>pixel4</th>\n",
       "      <th>pixel5</th>\n",
       "      <th>pixel6</th>\n",
       "      <th>pixel7</th>\n",
       "      <th>pixel8</th>\n",
       "      <th>...</th>\n",
       "      <th>pixel774</th>\n",
       "      <th>pixel775</th>\n",
       "      <th>pixel776</th>\n",
       "      <th>pixel777</th>\n",
       "      <th>pixel778</th>\n",
       "      <th>pixel779</th>\n",
       "      <th>pixel780</th>\n",
       "      <th>pixel781</th>\n",
       "      <th>pixel782</th>\n",
       "      <th>pixel783</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 785 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   label  pixel0  pixel1  pixel2  pixel3  pixel4  pixel5  pixel6  pixel7  \\\n",
       "0      1       0       0       0       0       0       0       0       0   \n",
       "1      0       0       0       0       0       0       0       0       0   \n",
       "2      1       0       0       0       0       0       0       0       0   \n",
       "3      4       0       0       0       0       0       0       0       0   \n",
       "4      0       0       0       0       0       0       0       0       0   \n",
       "\n",
       "   pixel8  ...  pixel774  pixel775  pixel776  pixel777  pixel778  pixel779  \\\n",
       "0       0  ...         0         0         0         0         0         0   \n",
       "1       0  ...         0         0         0         0         0         0   \n",
       "2       0  ...         0         0         0         0         0         0   \n",
       "3       0  ...         0         0         0         0         0         0   \n",
       "4       0  ...         0         0         0         0         0         0   \n",
       "\n",
       "   pixel780  pixel781  pixel782  pixel783  \n",
       "0         0         0         0         0  \n",
       "1         0         0         0         0  \n",
       "2         0         0         0         0  \n",
       "3         0         0         0         0  \n",
       "4         0         0         0         0  \n",
       "\n",
       "[5 rows x 785 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "6e19f034-e5dc-413b-818d-1e3aa88e7b72",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(28000, 784)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#loading test data\n",
    "df_test = pd.read_csv(\"test.csv\")\n",
    "df_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "b6923af7-b5d6-45ea-b8d6-ced10938e9c6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>pixel0</th>\n",
       "      <th>pixel1</th>\n",
       "      <th>pixel2</th>\n",
       "      <th>pixel3</th>\n",
       "      <th>pixel4</th>\n",
       "      <th>pixel5</th>\n",
       "      <th>pixel6</th>\n",
       "      <th>pixel7</th>\n",
       "      <th>pixel8</th>\n",
       "      <th>pixel9</th>\n",
       "      <th>...</th>\n",
       "      <th>pixel774</th>\n",
       "      <th>pixel775</th>\n",
       "      <th>pixel776</th>\n",
       "      <th>pixel777</th>\n",
       "      <th>pixel778</th>\n",
       "      <th>pixel779</th>\n",
       "      <th>pixel780</th>\n",
       "      <th>pixel781</th>\n",
       "      <th>pixel782</th>\n",
       "      <th>pixel783</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 784 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   pixel0  pixel1  pixel2  pixel3  pixel4  pixel5  pixel6  pixel7  pixel8  \\\n",
       "0       0       0       0       0       0       0       0       0       0   \n",
       "1       0       0       0       0       0       0       0       0       0   \n",
       "2       0       0       0       0       0       0       0       0       0   \n",
       "3       0       0       0       0       0       0       0       0       0   \n",
       "4       0       0       0       0       0       0       0       0       0   \n",
       "\n",
       "   pixel9  ...  pixel774  pixel775  pixel776  pixel777  pixel778  pixel779  \\\n",
       "0       0  ...         0         0         0         0         0         0   \n",
       "1       0  ...         0         0         0         0         0         0   \n",
       "2       0  ...         0         0         0         0         0         0   \n",
       "3       0  ...         0         0         0         0         0         0   \n",
       "4       0  ...         0         0         0         0         0         0   \n",
       "\n",
       "   pixel780  pixel781  pixel782  pixel783  \n",
       "0         0         0         0         0  \n",
       "1         0         0         0         0  \n",
       "2         0         0         0         0  \n",
       "3         0         0         0         0  \n",
       "4         0         0         0         0  \n",
       "\n",
       "[5 rows x 784 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_test.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "390d1fce",
   "metadata": {},
   "source": [
    "### Standardizing the Independent Variables"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8a3ba935",
   "metadata": {},
   "source": [
    "We are going to do a PCA.  PCA is effected by scale so we standardize the data here and will use the standarized data throughout the entirity of the analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "a5bd845a",
   "metadata": {},
   "outputs": [],
   "source": [
    "#merge datasets together and then scale\n",
    "df_temp1 = df.drop(columns=['label'])\n",
    "df_temp1['source'] = 'train'\n",
    "df_temp2 = df_test\n",
    "df_temp2['source'] = 'test'\n",
    "join_list = [df_temp1,df_temp2]\n",
    "df_comb = pd.concat(join_list)\n",
    "df_comb.shape\n",
    "source = df_comb['source'].to_list()\n",
    "df_comb2 = df_comb.drop(columns=['source'])\n",
    "del df_temp1, df_temp2, df_comb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "d9755c08",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(70000, 784)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>pixel0</th>\n",
       "      <th>pixel1</th>\n",
       "      <th>pixel2</th>\n",
       "      <th>pixel3</th>\n",
       "      <th>pixel4</th>\n",
       "      <th>pixel5</th>\n",
       "      <th>pixel6</th>\n",
       "      <th>pixel7</th>\n",
       "      <th>pixel8</th>\n",
       "      <th>pixel9</th>\n",
       "      <th>...</th>\n",
       "      <th>pixel774</th>\n",
       "      <th>pixel775</th>\n",
       "      <th>pixel776</th>\n",
       "      <th>pixel777</th>\n",
       "      <th>pixel778</th>\n",
       "      <th>pixel779</th>\n",
       "      <th>pixel780</th>\n",
       "      <th>pixel781</th>\n",
       "      <th>pixel782</th>\n",
       "      <th>pixel783</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>70000.0</td>\n",
       "      <td>70000.0</td>\n",
       "      <td>70000.0</td>\n",
       "      <td>70000.0</td>\n",
       "      <td>70000.0</td>\n",
       "      <td>70000.0</td>\n",
       "      <td>70000.0</td>\n",
       "      <td>70000.0</td>\n",
       "      <td>70000.0</td>\n",
       "      <td>70000.0</td>\n",
       "      <td>...</td>\n",
       "      <td>7.000000e+04</td>\n",
       "      <td>7.000000e+04</td>\n",
       "      <td>7.000000e+04</td>\n",
       "      <td>7.000000e+04</td>\n",
       "      <td>7.000000e+04</td>\n",
       "      <td>7.000000e+04</td>\n",
       "      <td>70000.0</td>\n",
       "      <td>70000.0</td>\n",
       "      <td>70000.0</td>\n",
       "      <td>70000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>-3.823623e-15</td>\n",
       "      <td>-1.137216e-15</td>\n",
       "      <td>2.825315e-15</td>\n",
       "      <td>-3.964385e-16</td>\n",
       "      <td>1.831964e-15</td>\n",
       "      <td>1.212366e-15</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000007e+00</td>\n",
       "      <td>1.000007e+00</td>\n",
       "      <td>1.000007e+00</td>\n",
       "      <td>1.000007e+00</td>\n",
       "      <td>1.000007e+00</td>\n",
       "      <td>1.000007e+00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>-3.295091e-02</td>\n",
       "      <td>-2.338733e-02</td>\n",
       "      <td>-1.675050e-02</td>\n",
       "      <td>-1.063784e-02</td>\n",
       "      <td>-8.339086e-03</td>\n",
       "      <td>-5.342334e-03</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>-3.295091e-02</td>\n",
       "      <td>-2.338733e-02</td>\n",
       "      <td>-1.675050e-02</td>\n",
       "      <td>-1.063784e-02</td>\n",
       "      <td>-8.339086e-03</td>\n",
       "      <td>-5.342334e-03</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>-3.295091e-02</td>\n",
       "      <td>-2.338733e-02</td>\n",
       "      <td>-1.675050e-02</td>\n",
       "      <td>-1.063784e-02</td>\n",
       "      <td>-8.339086e-03</td>\n",
       "      <td>-5.342334e-03</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>-3.295091e-02</td>\n",
       "      <td>-2.338733e-02</td>\n",
       "      <td>-1.675050e-02</td>\n",
       "      <td>-1.063784e-02</td>\n",
       "      <td>-8.339086e-03</td>\n",
       "      <td>-5.342334e-03</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4.236282e+01</td>\n",
       "      <td>5.965323e+01</td>\n",
       "      <td>9.086908e+01</td>\n",
       "      <td>1.619808e+02</td>\n",
       "      <td>1.634635e+02</td>\n",
       "      <td>1.932091e+02</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 784 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        pixel0   pixel1   pixel2   pixel3   pixel4   pixel5   pixel6   pixel7  \\\n",
       "count  70000.0  70000.0  70000.0  70000.0  70000.0  70000.0  70000.0  70000.0   \n",
       "mean       0.0      0.0      0.0      0.0      0.0      0.0      0.0      0.0   \n",
       "std        0.0      0.0      0.0      0.0      0.0      0.0      0.0      0.0   \n",
       "min        0.0      0.0      0.0      0.0      0.0      0.0      0.0      0.0   \n",
       "25%        0.0      0.0      0.0      0.0      0.0      0.0      0.0      0.0   \n",
       "50%        0.0      0.0      0.0      0.0      0.0      0.0      0.0      0.0   \n",
       "75%        0.0      0.0      0.0      0.0      0.0      0.0      0.0      0.0   \n",
       "max        0.0      0.0      0.0      0.0      0.0      0.0      0.0      0.0   \n",
       "\n",
       "        pixel8   pixel9  ...      pixel774      pixel775      pixel776  \\\n",
       "count  70000.0  70000.0  ...  7.000000e+04  7.000000e+04  7.000000e+04   \n",
       "mean       0.0      0.0  ... -3.823623e-15 -1.137216e-15  2.825315e-15   \n",
       "std        0.0      0.0  ...  1.000007e+00  1.000007e+00  1.000007e+00   \n",
       "min        0.0      0.0  ... -3.295091e-02 -2.338733e-02 -1.675050e-02   \n",
       "25%        0.0      0.0  ... -3.295091e-02 -2.338733e-02 -1.675050e-02   \n",
       "50%        0.0      0.0  ... -3.295091e-02 -2.338733e-02 -1.675050e-02   \n",
       "75%        0.0      0.0  ... -3.295091e-02 -2.338733e-02 -1.675050e-02   \n",
       "max        0.0      0.0  ...  4.236282e+01  5.965323e+01  9.086908e+01   \n",
       "\n",
       "           pixel777      pixel778      pixel779  pixel780  pixel781  pixel782  \\\n",
       "count  7.000000e+04  7.000000e+04  7.000000e+04   70000.0   70000.0   70000.0   \n",
       "mean  -3.964385e-16  1.831964e-15  1.212366e-15       0.0       0.0       0.0   \n",
       "std    1.000007e+00  1.000007e+00  1.000007e+00       0.0       0.0       0.0   \n",
       "min   -1.063784e-02 -8.339086e-03 -5.342334e-03       0.0       0.0       0.0   \n",
       "25%   -1.063784e-02 -8.339086e-03 -5.342334e-03       0.0       0.0       0.0   \n",
       "50%   -1.063784e-02 -8.339086e-03 -5.342334e-03       0.0       0.0       0.0   \n",
       "75%   -1.063784e-02 -8.339086e-03 -5.342334e-03       0.0       0.0       0.0   \n",
       "max    1.619808e+02  1.634635e+02  1.932091e+02       0.0       0.0       0.0   \n",
       "\n",
       "       pixel783  \n",
       "count   70000.0  \n",
       "mean        0.0  \n",
       "std         0.0  \n",
       "min         0.0  \n",
       "25%         0.0  \n",
       "50%         0.0  \n",
       "75%         0.0  \n",
       "max         0.0  \n",
       "\n",
       "[8 rows x 784 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#apply scaling to joined test and train data\n",
    "col_list = df_comb2.columns\n",
    "scaler = StandardScaler()\n",
    "model2=scaler.fit(df_comb2)\n",
    "df_standard=model2.transform(df_comb2)\n",
    "df_standard = pd.DataFrame(df_standard,columns=col_list)\n",
    "print(df_standard.shape)\n",
    "df_standard.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "7f499448",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(70000, 785)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_standard['source'] = source\n",
    "df_standard.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9e45f175",
   "metadata": {},
   "source": [
    "### Accuracy Function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "fa8f82b4",
   "metadata": {},
   "outputs": [],
   "source": [
    "#bulding a function to evaluate model accuracy\n",
    "def score(predict,actual):\n",
    "    return sum(predict == actual) / len(actual)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b5c6adb2",
   "metadata": {},
   "source": [
    "### Neural Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "d55e7f42",
   "metadata": {},
   "outputs": [],
   "source": [
    "#need to create the hotkey encoding\n",
    "y = to_categorical(df['label'], num_classes = 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "004ed2ed",
   "metadata": {},
   "outputs": [],
   "source": [
    "#reshaping the data\n",
    "X = df_standard[df_standard['source'] == 'train']\n",
    "X = X.drop(columns=['source']).values\n",
    "X = X.reshape(-1,28,28,1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "aaa29131",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(33600, 28, 28, 1)\n",
      "(8400, 28, 28, 1)\n",
      "(33600, 10)\n",
      "(8400, 10)\n"
     ]
    }
   ],
   "source": [
    "#going to recreate the cross validation\n",
    "#because of the new hotkey encoding and reshaping\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=0)\n",
    "print(X_train.shape)\n",
    "print(X_test.shape)\n",
    "print(y_train.shape)\n",
    "print(y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "92c49814",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(33600,)\n",
      "(8400,)\n"
     ]
    }
   ],
   "source": [
    "#also going to create the train_test split for y values at are not hotkey encoded.  \n",
    "#this will making testing accuracy a little easier later on\n",
    "#using the same random_state as above so we get the same splits\n",
    "y_actual = (df['label'])\n",
    "temp1, temp2, y_train_digit, y_test_digit = train_test_split(X, y_actual, test_size=0.2, random_state=0)\n",
    "print(y_train_digit.shape)\n",
    "print(y_test_digit.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "97bb9b95",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv2d (Conv2D)             (None, 24, 24, 32)        832       \n",
      "                                                                 \n",
      " flatten (Flatten)           (None, 18432)             0         \n",
      "                                                                 \n",
      " dense (Dense)               (None, 10)                184330    \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 10)                110       \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 10)                110       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 185,382\n",
      "Trainable params: 185,382\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#building and compiling model 1\n",
    "#one convulution layer and two other hidden layers\n",
    "#10 nodes in each dense layer\n",
    "\n",
    "model1 = Sequential()\n",
    "\n",
    "model1.add(Conv2D(filters = 32, kernel_size = (5,5), activation ='relu', input_shape = (28,28,1)))\n",
    "model1.add(Flatten())\n",
    "model1.add(Dense(10, activation = \"relu\"))\n",
    "model1.add(Dense(10, activation = \"relu\"))\n",
    "model1.add(Dense(10, activation = \"softmax\"))\n",
    "\n",
    "model1.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "\n",
    "model1.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "62e51950",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv2d_1 (Conv2D)           (None, 24, 24, 32)        832       \n",
      "                                                                 \n",
      " flatten_1 (Flatten)         (None, 18432)             0         \n",
      "                                                                 \n",
      " dense_3 (Dense)             (None, 100)               1843300   \n",
      "                                                                 \n",
      " dense_4 (Dense)             (None, 100)               10100     \n",
      "                                                                 \n",
      " dense_5 (Dense)             (None, 10)                1010      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 1,855,242\n",
      "Trainable params: 1,855,242\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#building and compiling model 2\n",
    "#one convulution layer and two other hidden layers\n",
    "#100 nodes in each dense layer\n",
    "\n",
    "model2 = Sequential()\n",
    "\n",
    "model2.add(Conv2D(filters = 32, kernel_size = (5,5), activation ='relu', input_shape = (28,28,1)))\n",
    "model2.add(Flatten())\n",
    "model2.add(Dense(100, activation = \"relu\"))\n",
    "model2.add(Dense(100, activation = \"relu\"))\n",
    "model2.add(Dense(10, activation = \"softmax\"))\n",
    "\n",
    "model2.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "\n",
    "model2.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "234f2f0e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv2d_2 (Conv2D)           (None, 24, 24, 32)        832       \n",
      "                                                                 \n",
      " flatten_2 (Flatten)         (None, 18432)             0         \n",
      "                                                                 \n",
      " dense_6 (Dense)             (None, 10)                184330    \n",
      "                                                                 \n",
      " dense_7 (Dense)             (None, 10)                110       \n",
      "                                                                 \n",
      " dense_8 (Dense)             (None, 10)                110       \n",
      "                                                                 \n",
      " dense_9 (Dense)             (None, 10)                110       \n",
      "                                                                 \n",
      " dense_10 (Dense)            (None, 10)                110       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 185,602\n",
      "Trainable params: 185,602\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#building and compiling model 3\n",
    "#one convulution layer and four other hidden layers\n",
    "#10 nodes in each dense layer\n",
    "\n",
    "model3 = Sequential()\n",
    "\n",
    "model3.add(Conv2D(filters = 32, kernel_size = (5,5), activation ='relu', input_shape = (28,28,1)))\n",
    "model3.add(Flatten())\n",
    "model3.add(Dense(10, activation = \"relu\"))\n",
    "model3.add(Dense(10, activation = \"relu\"))\n",
    "model3.add(Dense(10, activation = \"relu\"))\n",
    "model3.add(Dense(10, activation = \"relu\"))\n",
    "model3.add(Dense(10, activation = \"softmax\"))\n",
    "\n",
    "model3.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "\n",
    "model3.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "9f2f73f8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_3\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv2d_3 (Conv2D)           (None, 24, 24, 32)        832       \n",
      "                                                                 \n",
      " flatten_3 (Flatten)         (None, 18432)             0         \n",
      "                                                                 \n",
      " dense_11 (Dense)            (None, 100)               1843300   \n",
      "                                                                 \n",
      " dense_12 (Dense)            (None, 100)               10100     \n",
      "                                                                 \n",
      " dense_13 (Dense)            (None, 100)               10100     \n",
      "                                                                 \n",
      " dense_14 (Dense)            (None, 100)               10100     \n",
      "                                                                 \n",
      " dense_15 (Dense)            (None, 10)                1010      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 1,875,442\n",
      "Trainable params: 1,875,442\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#building and compiling model 4\n",
    "#one convulution layer and four other hidden layers\n",
    "#100 nodes in each dense layer\n",
    "\n",
    "model4 = Sequential()\n",
    "\n",
    "model4.add(Conv2D(filters = 32, kernel_size = (5,5), activation ='relu', input_shape = (28,28,1)))\n",
    "model4.add(Flatten())\n",
    "model4.add(Dense(100, activation = \"relu\"))\n",
    "model4.add(Dense(100, activation = \"relu\"))\n",
    "model4.add(Dense(100, activation = \"relu\"))\n",
    "model4.add(Dense(100, activation = \"relu\"))\n",
    "model4.add(Dense(10, activation = \"softmax\"))\n",
    "\n",
    "model4.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "\n",
    "model4.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "03953e8c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/2\n",
      "336/336 [==============================] - 6s 15ms/step - loss: 0.5611 - accuracy: 0.8196\n",
      "Epoch 2/2\n",
      "336/336 [==============================] - 5s 14ms/step - loss: 0.1579 - accuracy: 0.9543\n",
      "Epoch 1/2\n",
      "336/336 [==============================] - 7s 20ms/step - loss: 0.2221 - accuracy: 0.9320\n",
      "Epoch 2/2\n",
      "336/336 [==============================] - 7s 20ms/step - loss: 0.0743 - accuracy: 0.9774\n",
      "Epoch 1/2\n",
      "336/336 [==============================] - 7s 19ms/step - loss: 1.2414 - accuracy: 0.5588\n",
      "Epoch 2/2\n",
      "336/336 [==============================] - 6s 17ms/step - loss: 0.3670 - accuracy: 0.9008\n",
      "Epoch 1/2\n",
      "336/336 [==============================] - 8s 21ms/step - loss: 0.2417 - accuracy: 0.9262\n",
      "Epoch 2/2\n",
      "336/336 [==============================] - 7s 21ms/step - loss: 0.0848 - accuracy: 0.9734\n"
     ]
    }
   ],
   "source": [
    "#fit each model\n",
    "\n",
    "start1=datetime.datetime.now()\n",
    "model1.fit(X_train, y_train, epochs=2, batch_size=100)\n",
    "end1=datetime.datetime.now()\n",
    "\n",
    "\n",
    "start2=datetime.datetime.now()\n",
    "model2.fit(X_train, y_train, epochs=2, batch_size=100)\n",
    "end2=datetime.datetime.now()\n",
    "\n",
    "start3=datetime.datetime.now()\n",
    "model3.fit(X_train, y_train, epochs=2, batch_size=100)\n",
    "end3=datetime.datetime.now()\n",
    "\n",
    "start4=datetime.datetime.now()\n",
    "model4.fit(X_train, y_train, epochs=2, batch_size=100)\n",
    "end4=datetime.datetime.now()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "7c62107e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "263/263 [==============================] - 1s 3ms/step\n",
      "263/263 [==============================] - 2s 6ms/step\n",
      "263/263 [==============================] - 1s 4ms/step\n",
      "263/263 [==============================] - 1s 4ms/step\n",
      "1050/1050 [==============================] - 4s 4ms/step\n",
      "1050/1050 [==============================] - 6s 6ms/step\n",
      "1050/1050 [==============================] - 5s 5ms/step\n",
      "1050/1050 [==============================] - 7s 7ms/step\n"
     ]
    }
   ],
   "source": [
    "#make a prediction on holdout data and translate to label\n",
    "\n",
    "prediction1 = model1.predict(X_test)\n",
    "prediction2 = model2.predict(X_test)\n",
    "prediction3 = model3.predict(X_test)\n",
    "prediction4 = model4.predict(X_test)\n",
    "\n",
    "prediction1 = np.argmax(prediction1,axis = 1)\n",
    "prediction2 = np.argmax(prediction2,axis = 1)\n",
    "prediction3 = np.argmax(prediction3,axis = 1)\n",
    "prediction4 = np.argmax(prediction4,axis = 1)\n",
    "\n",
    "#make a prediction on training data and translate to label\n",
    "\n",
    "prediction5 = model1.predict(X_train)\n",
    "prediction6 = model2.predict(X_train)\n",
    "prediction7 = model3.predict(X_train)\n",
    "prediction8 = model4.predict(X_train)\n",
    "\n",
    "prediction5 = np.argmax(prediction5,axis = 1)\n",
    "prediction6 = np.argmax(prediction6,axis = 1)\n",
    "prediction7 = np.argmax(prediction7,axis = 1)\n",
    "prediction8 = np.argmax(prediction8,axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "498d82ae",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0 1 2 3 4 5 6 7 8 9]\n",
      "[0 1 2 3 4 5 6 7 8 9]\n",
      "[0 1 2 3 4 5 6 7 8 9]\n",
      "[0 1 2 3 4 5 6 7 8 9]\n",
      "[0 1 2 3 4 5 6 7 8 9]\n",
      "[0 1 2 3 4 5 6 7 8 9]\n",
      "[0 1 2 3 4 5 6 7 8 9]\n",
      "[0 1 2 3 4 5 6 7 8 9]\n"
     ]
    }
   ],
   "source": [
    "#just checking that we are getting reasonable outputs\n",
    "print(np.unique(prediction1))\n",
    "print(np.unique(prediction2))\n",
    "print(np.unique(prediction3))\n",
    "print(np.unique(prediction4))\n",
    "print(np.unique(prediction5))\n",
    "print(np.unique(prediction6))\n",
    "print(np.unique(prediction7))\n",
    "print(np.unique(prediction8))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "f0f9e149",
   "metadata": {},
   "outputs": [],
   "source": [
    "#evaluate accuracy\n",
    "score1 = score(prediction1,y_test_digit)\n",
    "score2 = score(prediction2,y_test_digit)\n",
    "score3 = score(prediction3,y_test_digit)\n",
    "score4 =score(prediction4,y_test_digit)\n",
    "\n",
    "score5 =score(prediction5,y_train_digit)\n",
    "score6 =score(prediction6,y_train_digit)\n",
    "score7 =score(prediction7,y_train_digit)\n",
    "score8 =score(prediction8,y_train_digit)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "2b40f9ea",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Model</th>\n",
       "      <th>Nodes</th>\n",
       "      <th>Layers</th>\n",
       "      <th>Time</th>\n",
       "      <th>Test Accuracy</th>\n",
       "      <th>Train Accuarcy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Model1</td>\n",
       "      <td>2</td>\n",
       "      <td>10</td>\n",
       "      <td>0 days 00:00:10.978455</td>\n",
       "      <td>0.957976</td>\n",
       "      <td>0.970595</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Model2</td>\n",
       "      <td>2</td>\n",
       "      <td>100</td>\n",
       "      <td>0 days 00:00:13.993535</td>\n",
       "      <td>0.974167</td>\n",
       "      <td>0.989405</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Model3</td>\n",
       "      <td>4</td>\n",
       "      <td>10</td>\n",
       "      <td>0 days 00:00:12.839884</td>\n",
       "      <td>0.922024</td>\n",
       "      <td>0.928810</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Model4</td>\n",
       "      <td>4</td>\n",
       "      <td>100</td>\n",
       "      <td>0 days 00:00:14.901011</td>\n",
       "      <td>0.971786</td>\n",
       "      <td>0.985327</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Model  Nodes  Layers                   Time  Test Accuracy  Train Accuarcy\n",
       "0  Model1      2      10 0 days 00:00:10.978455       0.957976        0.970595\n",
       "1  Model2      2     100 0 days 00:00:13.993535       0.974167        0.989405\n",
       "2  Model3      4      10 0 days 00:00:12.839884       0.922024        0.928810\n",
       "3  Model4      4     100 0 days 00:00:14.901011       0.971786        0.985327"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#creating the experiment cross-table\n",
    "d = {'Model':['Model1','Model2','Model3','Model4'],'Nodes':[2,2,4,4],'Layers':[10,100,10,100],'Time':[end1-start1,end2-start2,end3-start3,end4-start4],'Test Accuracy':[score1,score2,score3,score4],'Train Accuarcy':[score5,score6,score7,score8]}\n",
    "cross_t = pd.DataFrame(d)\n",
    "cross_t"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "9fd02029",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(28000, 28, 28, 1)\n",
      "875/875 [==============================] - 4s 4ms/step\n",
      "875/875 [==============================] - 5s 6ms/step\n",
      "875/875 [==============================] - 5s 5ms/step\n",
      "875/875 [==============================] - 5s 6ms/step\n",
      "   ImageId  Label\n",
      "0        1      2\n",
      "1        2      0\n",
      "2        3      9\n",
      "3        4      9\n",
      "4        5      3\n",
      "(28000, 2)\n",
      "(28000, 2)\n",
      "(28000, 2)\n",
      "(28000, 2)\n"
     ]
    }
   ],
   "source": [
    "#building a kaggle submission for each model\n",
    "\n",
    "df_submit = df_standard[df_standard['source'] == 'test']\n",
    "df_submit = df_submit.drop(columns=['source']).values\n",
    "df_submit = df_submit.reshape(-1,28,28,1)\n",
    "print(df_submit.shape)\n",
    "\n",
    "sub1 = model1.predict(df_submit)\n",
    "sub2 = model2.predict(df_submit)\n",
    "sub3 = model3.predict(df_submit)\n",
    "sub4 = model4.predict(df_submit)\n",
    "\n",
    "sub1 = np.argmax(sub1,axis = 1)\n",
    "sub2 = np.argmax(sub2,axis = 1)\n",
    "sub3 = np.argmax(sub3,axis = 1)\n",
    "sub4 = np.argmax(sub4,axis = 1)\n",
    "\n",
    "\n",
    "sub1 = pd.DataFrame(data=sub1,columns=['Label'])\n",
    "sub2 = pd.DataFrame(data=sub2,columns=['Label'])\n",
    "sub3 = pd.DataFrame(data=sub3,columns=['Label'])\n",
    "sub4 = pd.DataFrame(data=sub4,columns=['Label'])\n",
    "\n",
    "sub1['ImageId'] = sub1.index + 1\n",
    "sub2['ImageId'] = sub2.index + 1\n",
    "sub3['ImageId'] = sub3.index + 1\n",
    "sub4['ImageId'] = sub4.index + 1\n",
    "\n",
    "sub1 = sub1[['ImageId','Label']]\n",
    "sub2 = sub2[['ImageId','Label']]\n",
    "sub3 = sub3[['ImageId','Label']]\n",
    "sub4 = sub4[['ImageId','Label']]\n",
    "\n",
    "print(sub1.head())\n",
    "\n",
    "print(sub1.shape)\n",
    "print(sub2.shape)\n",
    "print(sub3.shape)\n",
    "print(sub4.shape)\n",
    "\n",
    "sub1.to_csv('NN_1.csv',index=False)\n",
    "sub2.to_csv('NN_2.csv',index=False)\n",
    "sub3.to_csv('NN_3.csv',index=False)\n",
    "sub4.to_csv('NN_4.csv',index=False)"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "iVBORw0KGgoAAAANSUhEUgAABo4AAAHwCAYAAACYF51OAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAAFiUAABYlAUlSJPAAANpOSURBVHhe7N19XFVVov/xb1kcM8LRro2VZo3YzEW9M6CTA1cniHxqNMrCnKuDvzTSfCo1S7EiekCyfMinaky9MjqZlA3JpGkGpheyFGYm4U6JUyqlNyabkKzDxPDba58NnHM4KAim1Oc9rzOttc5+WHvvc/jjfF1rnVNlEQAAAAAAAAAAAL73znX+CwAAAAAAAAAAgO85giMAAAAAAAAAAADYCI4AAAAAAAAAAABgIzgCAAAAAAAAAACAjeAIAAAAAAAAAAAANoIjAAAAAAAAAAAA2AiOAAAAAAAAAAAAYCM4AgAAAAAAAAAAgI3gCAAAAAAAAAAAADaCIwAAAAAAAAAAANgIjgAAAAAAAAAAAGAjOAIAAAAAAAAAAICN4AgAAAAAAAAAAAA2giMAAAAAAAAAAADYCI4AAAAAAAAAAABgO6fK4pQbbN/+A04JAAAAAAAAAAAALUG3rl2cUv1OKTgCAAAAAAAAAADAdw9T1QEAAAAAAAAAAMBGcAQAAAAAAAAAAAAbwREAAAAAAAAAAABsBEcAAAAAAAAAAACwERwBAAAAAAAAAADARnAEAAAAAAAAAAAAG8ERAAAAAAAAAAAAbARHAAAAAAAAAAAAsBEcAQAAAAAAAAAAwEZwBAAAAAAAAAAAABvBEQAAAAAAAAAAAGwERwAAAAAAAAAAALARHAEAAAAAAAAAAMBGcAQAAAAAAAAAAAAbwREAAAAAAAAAAABsBEcAAAAAAAAAAACwERwBAAAAAAAAAADARnAEAAAAAAAAAAAAG8ERAAAAAAAAAAAAbARHAAAAAAAAAAAAsBEcAQAAAAAAAAAAwEZwBAAAAAAAAAAAABvBEQAAAAAAAAAAAGwERwAAAAAAAAAAALARHAEAAAAAAAAAAMBGcAQAAAAAAAAAAAAbwREAAAAAAAAAAABsBEcAAAAAAAAAAACwERwBAAAAAAAAAADARnAEAAAAAAAAAAAAG8ERAAAAAAAAAAAAbOdUWZwyvrPcOlSQr0NfmnJrdY4IV+c29hsBuUsKtOejr+1y6ysjFNHJZZelMhXvLNSndrmtukWGqUMruxJQ2Qd52uvZWJd0j1RoO0+5uXgf3zgd5wAAAAAAAAAA4PuE4Oh7oVQvjonSzBxPrfOYdG19IFLVcZC/0pcSdc19no1j5uZq5a0d7LJUoNQfDddypxaTmq2VIzo5tbry07rplt96yonr9ympt6fcLI5s1Jjrpyn7uFO3NPs5AAAAAAAAAAD4nvleTVVXXvGlio8eUNYH2cr86xs1L1M37eb974NDKxOVmuN2aqcuO2ma0vc7lW9VmTbPmeUTGgEAAAAAAAAAgKb7TgdH/6z8p978ME+TX3tEP3tmqLovvUGxqxM08Y8Pa8qmR2tepm7azftmO7O92c/s/93kVvqkWdr8uVM9ZQVKHr9MhZVO9VvizlmgezY2PfgCAAAAAAAAAAC+vpPB0ZHyv+v+rU+q57Ihuv0PM/Xq+9v0+ddlzrsnZrYz25v9zP7mOOZ43znHN+qe+zaq1Kmesv0LdM+iAn1rMc7xPM17aI3nfG1GKWlmpN3cHNyfl6nMebkbGIadyj4AAAAAAAAAAJytvlPB0TH3l5r5xpOKej5e6/Zm6atvvnbeOTVmf3McczxzXHP87xL3tlm6/5UmR0cqXjxLSwq+jejIrfznUrS8xJRdips7VTe1r2+lpgaqLFPhSym6pVc3/aRXL/3Uef3kp1Eak7ZRxYGmw6ssVe7iRF3Tw2+fbj10TeIy5Xrf0sr9Wh7XTVf9yLx6KHW30+7DrEFVvU03zdzGaCoAAAAAAAAAwJnxnQiOqqz/bS5+S31XjtAL72WpsupfzjvNwxzPHNcc35zHnK8lS3g0WTF2ya3sB5P0oh3ENFakkh4dq852eb+W3LdA+ad5zSF3wTLdv9izqJJr6Hw9ckOIXT5llSXKvDtGQ+5bo3z/afuOlyr7t9PUPy7N97qsfV4cH6ORC3JUWud63SrdtkAjByXVTgPYqqsGDQ93Km6lv1HglL0cydHrOU65zSjd2LeJYRgAAAAAAAAAAKeoxQdHZh2iR3KWaEJWsv5Rz3R0Qa3OV2TncM25/l5tG52u9ydv0YGp22tepm7azftmO7N9IOb45jzmfC15/SPXj0cpZW607HjieI6SUzJ0qNHTrLkUEjtVC+/s6qnuX6H7T+eUdZVFWn7fMhWbcptoPTFrgJoYG6nwt3fontecz0ybcE16OlNv7dmjt9Yn67YwJ7yxruu/nsyrua6y1+fXjgjqOkors/fqw7/t04d7s7VytHMvPs/QPYtq9+nc90ZFOGX3a3l11oQqezdH2U7ZFT9AvciNAAAAAAAAAABnSIsOjr6sOK67Nz+mlQUvBRxldNlFP9RTA2bqvQl/1LpbF+q/eg5VaPsuan2e7y/zpm7azftmO7O92c/s78+cx5zPnNecv6XqfHOyUmI998G9LUnJGacy7MiliClzlOjkJcW/naZ5eacnOipe87DmeQYbKeaBZMV19JRP2ecbteRJ54AKV0rmek0fGqbO7ULUufcopa1fooQ2nnfdqzOU7eRLxX/Z6ClYYsZNUEwX57PUppPVr6V6ZvIETZqZqmcGXFIbonUZqOHRTrnEOtb7TtlWptw3tjhllxJ+FekJ9AAAAAAAAAAAOANabHBU5i63w5s/flA9x1etH7QOsYOfnWPXKb774DpB0cmY7c1+Zn9zHHM8f+a85vymHy1Sq0667dE5inHCkeykaUqvzlEao024ps+doFC7UqLliWnKbu48rSRDqSnOFG/hyUqK7+QpN0FZXrY2O2Xd8Bvd5IRfNdpEa9L6dK1NN6/fqLvzEQq5NMxTsOS+/roKvae4M9PSTZ2q6XfGKyayq9eIqA4afPMAp1yi9duLnLKl7G1trs6iOk1QXPWsdgAAAAAAAAAAnAEtMjgy08Q9nLNIW/f/j9NSK/rKa7TlN6vs4KfVOU27PLO/OY45njmuP3N+048WO21dx6FaOHeoM8KlQMnjl9WZRq0hXOFTtXCGk7wcX6O7Zm1R4EkDT0WpMlNSnKncumr6o6MU2squNMmhA7VrDUX1DQ847V2HsEjrPfMKV2cnOAodeEdN2ObelqIhvbrpJ1FxGjk9TekbC1RcGnjEVUjfQRrklA9tzvdMuWdx5+cp0yl3jo9W92a4NgAAAAAAAAAATlWLDI6e2/OiXvnfrU7N49xzztXonw3T8zem6ofB/+a0Ng9zPHNcc3xzHm+mH6Y/LVXIDXP0zEgnFdm/QPcsKtAXnlqjdL9zqVKc0TLujdN0zyulnkoTlb6SpPudNYVCJ89RYu2AnyZxf1E7NZ+rMSPSOg7VMxvMVHm1+7iPFCn3lRVKvnu4+vfpoZ8mrFCh/6irdtdq0A1O+b1XlXvAFNzK3bbGbpI6aXhsM10cAAAAAAAAAACnqMUFR3s+2avn97yof3mtaWTCnCl9EpQcPVnntzrfaW1e5rjm+OY83uGR6Yfpj+lXy+RSzKzlSnRmfytePEv3r68eD9MIrboq4elUxdgVt7IfnKjH37UrTVCg9AdzatYK+vS1h3VLXJyGeL1un5/nvCu9OMtp/23taKL6uNrWTnf36ZeNGx/lunqUFubu1V+zM7Vy7kwl3hyp0HbOm5aynWm65e6N8o3OQhRzw1CnXKCcd613K4uU94rT1HOsBpEbAQAAAAAAAADOsBYVHH31zdd6+u3V+vxr3x/6b/73/pp0zagmT013Mub45jzmfN5Mf0y/TP9apDaRmv7EWHW2K/uVv7t2NE6jdIpXSmq0p3y8QPknz29Oyu01cqdsf5EK3/N7HamdGq7m/aNOwwl07lK7mFDhu0UBptZz61BBnnJ3mleBDgWYgc7VJUwxt45V0rx0bd2zT39OH+us9WTtve1V5foNugqJvVEJTjn7tRzr+K8r3bm+iOHRNfsCAAAAAAAAAHCmtKjg6I8f5GjHwd1OzSPi0u56OHrKaRtp5M+cx5zPnNeb6ZfpX0vlipyqhXc66xQ1Qef4ZKXFNmLqt5Po0DNM3U/w8h7pE9LVaW/vNBjHS5S9doVe3O03/icypmbNIb22VEt2+yVDx/P0/G8SNDLBvF511iSyjrV4gZLvMiObUrT5c7uxRkjfgc6IK4+v/deLckVq4Gjn3uRs0bzfve6MpgpXXN/aEVAAAAAAAAAAAJwpLSY4KnOXa/WfNvhMUdfm/Naa8Z93KMQV7LQ0v39W/lOLdq3W68U7nBbZ5zPnNeevZvpl+mf62TK5FDFljhKbmh216qTbkpMV08apN0m4EjMzlXWC1xPxzqaW2+Y47XdWjyYqUfpvYjTmwTTNHG7913vdpXZDNWlG9cXu1/L/N1wz1+ap+EipineusbafVDMayDV6gKLsvKeDXEdXKP11M7Jpje4Zt0C5B8rkrpTcn5cod8Ecpdt7WLr+Qr06OuUaLvWKjbf+38hR5kZnZFd0vAZ38RQBAAAAAAAAADiTWkxw9M7Hf1FR6X6n5nHjj69XZOfaKceamwmN5uWt1IK8/9b0LXN8wiNzXnN+b6Z/pp8tVptwTZ87oelTpnWK1xOPRjsByZlUqpKa6fLcyn7fdwq+7nc+Xzs66niRXnwwQf2jotQ/IUUvFjkjkLqO1e9nRDrX4lLUjN/VhGvu3cs0MqaXftKtm37SK0YjFxd4RhC1CdOkOaMC3kdX5I1K8AvVBt08UB2cMgAAAAAAAAAAZ1KLCI6++dc3yijcZP+3WrvWIRr9s5t1jvW/06E6NHpu9zp7NNEx95c+4ZE5rzm/6Ue1QP1saVzhU7WwZiTOqetw8zwtHHqmo6MwDZoa5gl92oQrJd4vZDSjo57NVdbcUYrwmvLO1qaDYu6cr62ZMxXhHfRYx0l6OVML74xWhzqjqkIUcfNMrX0jU9N713PtrcIVN857WroBGtS39jMEAAAAAAAAAMCZdE6VxSmftT459qluXT9JH5f9n9MiDQr9pZb+KlnnnXue09J8/EMjb3E/uV5P9r9frvOC7IBo4h9TtLn4Ledd6fKQH+ql4Ut02UWXOC3wcSRDY6KSlO1UTyxaabnLdVudKd8ax11WJl0YIlcrp6EeZjsz7ZzhCjn59qp0q6ysdm2kBu0DAAAAAAAA4KznLinS5u2vq/hTp+HCToqMHaiork37R+B1jqsOCr0+WoPCOjXst8XKMhXnva7sd0tU5jR1CB2omAFh6tyQf8ff1P0/36/cbTnKO1Szd+P6X8OtQ7vzdehrp2pc2Fm9wq3jOFWcOS0iOMr5aJfGZibVjOQ595xz7dDohm7Rdr05nSg0+tXV0XZodGFQ7VCT1/bl2OFR9bYmyFoRl6roK/vYdfg5A8ERAAAAAAAAADTI8f1Kn/UbpW4s9SxL4ccVNlbPrJqpmMauPWEd98XHpil5XVHA46rdAKX8bo4SwuoPpsoKVuiuO9KU+7nT4K1NmG57eL5Sbu1ab/DSpP0ry5S/YpJuT8urCZy8ucLilTI/Wbdd3bDYx707TUOGr1CxU7dFp+qdlfEs63EWaBHB0bzclVq0a7VTky658GKtH75IV/3Ae8qvpmtsaGR8+I8SDV8/RZ9++ZnTIk3pM1rTo8Y4Nfhwlyj/3UPyDpLr11qdfx7esKQbAAAAAAAAAJrieIFS44Zrue9S+3W1iVba5uW6raE/Tzf0uOqqxPWZSgqwDEbAoKUOl2LmbtbKW+t2rGn7u5WfFqdbfnuSC2jofXEXaN6Q4VrifziCo7NGi1jj6MN/HHJKHle0vUwd2vgvStM0pxIaGaYfpj/e/PsLL65OiugbqagGvQiNAAAAAAAAAHwb3Mp/bpZPuOPqPVZpy9O1Nn2+kgZ6rQt/PEczH8pQqVM9mcLVvscN6TtVK7fm6s97svXyo/HqXvOz834tn7ZA+f5Dkqzzpf4/79AnRFEjZ1p9m6+UMUO99ncr+75pSj/gVKs1cX93Tpr+yzs0ahephJmpWvl0shKHOmvMG+a+3L1GJ/t1vHDlrLqhEc4qZ31wVF7xpQ4f8/0KXnpRBwUHXejUmu5UQyPD9MP0x5vpr+k3AAAAAAAAAKAF+HyLli/2SjO6TtXLL8zUbbHmH7gPVeIzmVo50utfuecs04vvOeUTcedp/dLa47piU5W1aoJiunZQSLtOihiZqpf/e6xCnfdVskJLNvr+Hl68Jk3px52KPSopV2sfHWv1bagSHpivrA0zFeG8KxXo+deKnLJH0/bfr/TH16gmy+o6Vi/vSFfKnfGKGTpKSU9nKuuBcOdNS8EKZfqe3tf+NXroyer70VURvZt3VjE0j7M+OPrmX5X66hvfic26tb/SKTVdU0Kjav79Mf01/QYAAAAAAAAAnP1Kt72qzU7ZTNmWkDxW3Vs5VZtLMXd4BywlWvLHAqd8AiX7tacmtJFuG3mjOvsc14xsitdvejoVS/brb3uNZirS5jVewdPoZE33n8ru6lFauN6MjPK80n4RUhv0NHX/otf1u5rdzX2Zqgi/n8tDR8zXy86+a9NTFRUScBUnqbJELz6eonyn2nlMsp4YXhOZ4Sxy1gdHX39Toc+/CrTcVmCvvr9Nj7211A6ETqY5QqNATH9NvwEAAAAAAAAAZzu39v45xykb8Yr+eYA1NLpEKs4r4HHnFZ5kzSDL5yUqdIqGKzjAcRUi18VO0dhWqH3V4xJKCpVX4pQtt10b4Zkazl2q4t15yt1pXkVyd/VaAiS8U+30cU3c/9Bf3vaaeq72vrhL9yvf3td6FbkVGlm7f0SnmrP7OPRKipKrb3ObUUqZFqm2ThVnlxaxxlFDmdDovi1ztSL/JT361rIThkenKzQCAAAAAAAAALQkZfr0Y6do9Oxaz9rrXRX6c6dovFeiT0828VTHropyisae971SnGruYhW945Rt+3WoesjRof3KdYpStMJ+IhW/lKQhvaLUf3iCRiaY13D179VDQx7LUal/f5q4/6HiPKdkiQ5Tj8r9ejEpTj/tM0i32Ptar+GD9NOfxik15wSrPh3ZqOSHc5yRTC7FzZ2qGH5+P2ud9cFR6/OC1O6CEKdWv+rQyEwTZ4Kg3/35D3YwFCg8Ot2hkemv6TcAAAAAAAAA4GxXov3eA44udtU7Eqb1eU7B5hXw1KdTd0V6LeOTPydF6R94TeVWWarsNKvNazo7b6Uf104zZ+x/YZKG3Jehwjrbu1W4MlH9xm/0muauqfuX6tDfnKJtv9LvjtPMdUW1U9lVO16k5WNiNOaVQDekTJvnzFK2c05X7BzNvuHkv/njzDnrg6Pzzm2lC85r7dQ89h39yCl5uL+p0Bt/y/VZC8kEQiYYWvLOGlV6hUOnIzTy74/pr+k3AAAAAAAAAOD7LEwJM4fWTh13PEfJg3rpmiFxGhIXp/7XRGnM6hJF9A53NjiRHC1fnGMdMl4p67P15z179Nb6ZN3W1Xnb4t42S0t21rPGUFP3z1mhJduk7iOS9XL2Hmv/bL38aLxqVylyK/vBZcr1292ds0D3bKxujFZK8lB1cGo4O531wVFw0IW69CLfj9HhY6Uqr/jSqUmu84I0J3a6Hfx4M8HQol3pWvT2ajs8Oh2hkemH6Y8301/TbwAAAAAAAADA2c6lEK9RQY3RugHjB0JumKPfTw6rDY/kVmlRkQrfK1Lx51LorfM1fZjvGKd6j9tprFauT1VC704KaReizr1HKW39UsXV/KztVnrGdpU5tTqauH/nMcv1cuooRXQJsfbvpIiRqXp5iXcwtkbrt3ntfTxP8x5aUzNCKSY1Wbed4r3Gt6dFrHF01Q86OyWPg198otLj1jfKiwl8TPBTX3i0IG+lnspd0ezT05l+mP548+8vAAAAAAAAAOBs1UEdaofNSO+U6JBT9FWmTw87RVtXdW7Q0BmXIqZmasf6VCUODFMH+2foEIX2Haqk9FxtnTtAl7g/tbf0qD2uq81FnoKj+9h4Rfn/jN1ugIaPcMrGtkIVO8Wm7e9SW5/dw3THryO9AjCPkIHxSnDKxub3q6fHcyt3fpKWVy/rFJ6spHhSo5agRQRHvS7rrvPOrZ088u/HP9f/llZ/9GuZ4Cft+hnq3/U/nRYPExQt3vU7Pbv7hWYNjQzTD9Ofaqafpr8AAAAAAAAAgJagg7r19IpDjhdp3xGn7K1yv/K3OWUjtrsaM4SgQ+94JT2TqXf27tOHf9ujrenzldjXJESl2rO9yLORMTSiZvq3kCtDfc5xyYWB1wbyWXvpeO1ccU3bP0Sdf+Id9FwiV6CJtlr5LjXjrjl9kXJWVqdGloIU9e/WTVf9qPZ1zX1ei0vlJOkauz1N+U4TzowWERxdffFV+mHwxU7NEwRl/nWbvvnXN05LrRBXsOYPTKoTHgXS1NDInN/0wzuMMv00/QUAAAAAAAAAtAzdfx7vNZImR+u3eQUeDnfB69p83KlYIqLDT75WjztHyV5ByS8XewVE1Yoy9LxXfjLo+l+oJt75caQGef18nVtYXDPtW60S7X3XKRregVYT9+/+84Fe9yVPRcV199aBQu1xikbMjxlV1NK1iODokgvbq+clP3ZqHrtK/qQPPvvIqflqSHjU1NDIMOc3/fBm+mn6CwAAAAAAAABoGVyR8ZrklXfkz0nRiwecilGWp3mzVnhNYRet4bFeO1SWKnfxRA2JS1DyS/trwxlXhCJucMqWQwumKXlb7Zr57pIcJd+9oGZqObUZpd8M8BoV1CpcceNqz+NenaLUHN819w+ts9oKnIrFJ9Bq6v7hN3rdF7fSU9KU7b17ZYlefMx7hFC4oiNOGqfhLHdOlcUpn9Xe+Fuuxm180GeU0YgeQ5TW/16dY/0vkDJ3ucZnPaT/OeiddzZPaFRl/W/m1qe0bm+W0+KZpu65oY/q+h9FOS0AAAAAAAAAgJag7LVpumbSRq8ROSEKHXijYtoU6g+vF6jUa7RR6IxMbb0rzKlJpS8lek271knTM7M1qadTLVqm/kO8wqF6uRQzL1srb/YLXo7nKPmaRKXXnN+lDr0H6KbItjr02qvavL/Mabd0HauXM2cqwvun7ybu785J0U/HrKm9L206KMK6L1FtDmnza1tUXLuSi0LvXK+smeFeo5ROLD+tm275rVOJTtU7K+NPPooLp12LGHFkXHP5fyisQ1en5vHq+28o75BXFOrHjDx6dsgj+s8rejktzRMaGea85vzeTP9MPwEAAAAAAAAALUvIDXP0+8lhXqFHmYpfX6Plr/iGRiE3zNfKO2tDI+NTnzX5S1R8yCuMCZuglfMG1E4/F5BLETPW6xn/0MhoE62kDcmKqflJ263S3Ru1fPEa39Cn3QClPe8XGhlN3N8VPVNZydG19+V4qfJfWaEla31Do5AbUrVyRsNDI5y9WkxwZEKg0T8bpnPPqe3y8X9+rYV5/22PLKqPd3jUXKGROd+T//O8ff5qpl+mf+Z8AAAAAAAAAICWxqWIqZnakT5VMR0DxB9twnTbXOv9p4eqcyunzdF96DQNaucpu8KmKiHWNybqfPNS7ciar8TYDnWClZDwUUrLytXLd3mHVr5cV4/SytxMLbwzWh3qBEMdFHPnfGVlL9VtXZw2P03b36XQ0cv1Tj39d3WMVuLT5r7E17kvaJlazFR1xlfffK1xrz6o7QfecVo8bgkbqCeun6HzW53vtNT1zb8q7f+ed27TPrn/rPyn7n/jSb1c9LrT4nFtl2v03I2P6oLzWjstAAAAAAAAAICWyn1kvwqLP5UZPtC2c3eFdgqR60Q/L1e6VVbmliukYdvZ2oQo5BSG6Lg/L3OmjnMppF3jD9Ck/b3777L637RxGjgLtajgyNjzyV6NzZylz7+uHUJnRvtM6ZOgKb8YrVZeI5KaW2XVv7To7dVatCtd/7LK1dq1DtGKuDnqdVkPpwUAAAAAAAAAAKDlaTFT1VUz4cwdvW7zmbLOhDgmzEnJWWyPCDodzHHN8f1DI9MP0x9CIwAAAAAAAAAA0NK1uODIGNfrNt387/2dmocJc1b/aYPueDVJ/1f+d6e1eZjjmeOa43uHRobph+kPAAAAAAAAAABAS9fipqqrVuYu17TXU7V1//84LbV+0DpED/xygoaFDWzS1HVmaroNRa/rsbeW6R9eU+NV69/1PzV/YJJCXMFOCwAAAAAAAAAAQMvVYoMj48uK45qx9Qn98YMcp8XXZRf9UNMib9fQH1+n1uc1fIGvr79xa+P7b2p+3ip9cuz/nFZfv7o6Wk/2v18XBrXMlb8qvihXhSm0ClJwcJDdBgAAAAAAAAAAvt9a5FR11Uxo8/SgBzQm/NaAI4tM6HPvljT1XPYrjXjpHv3+vY0qPnrADoa8mbppN++b7cz2Zr9AoZE5jzmfOW+LC40qy7VvQ6ruuuF6xd50kwab19AbFPvr2dpQUO5sdHbbu/B69ZuYqaNOvVEqK1RuArNKp45TVKClMdZzuGme8uv72BzJ1H3WNvdtOOI0nK2OKGuidS0LC5x6YEc3TFE/cz0Z9V+P/dmMWay9Tr2xTJhbbqe5Z5Hdi61rmqKsU3mMzmeg38BU5X/ltNWjYkeqYs22p/rdrpfzWT3J8w2sYZ+NQA6kj7E/L6d2XgAAAAAAAODMatHBkXF+q/P1UPQkLRuSYk9RF0hF5T+Vd6hAs954SrGrE/TjxQPUZcG1NS9TN+3mfbOd2T4Qc3xzHnM+c94WpfKItqeM0ZjFOxUUO1lPL1+jTX9YozWPT9awtgVaMG24Hsg823/kb6LSTXrkppu0nN9ym8cXmzRj4U7PyLXvibxlSdrwoVNpVgVabn02HzmlhOYsV/GmNuScKA46qu2vvvnd+Rx9nKGlqw46FQAAAAAAAKDlafHBkXGO9b9Bob/UzjHr9OueQ5q0rlEg5njmuOb45jzmfC3N0axUPbJDGpG2Xk9Pj1NEaEcFt+2oLlFxmvjsej0TF6ztyxZre/P+c398l10QpKBtqUrb8n350AQp+IIjWnD/c9rHqLVG2Z6xTYedch0Ht2nDbqfc0lUeUVbqc9oTe7smhjltAAAAAAAAQAvznQiOql3kulBp189Q7h0ZGtFjiC44r7Xzzqkx+5vjmOOZ45rjt0xFylpVJMVOUWKfYKfNW7B6JNyuayt26Y85viMeyt/fpKXTRilu4PXqd8Nw3f1YpvZ94bxZ7b1VGjPqYW0/clT5y2ZolD0V3hgt2FBsjyKoeD9TC8YP1+CYGxT36xla5zct3t7nR2lMyiYdLS9W1mNj7HPF3jRKD8x7U4cb+AP90dxVesQ+x/UafMt4PbJql4567WufY+oq7bHKWXOs8qhRemKTb+BxsmM0zVFtT7HO+7z1HA5ucu6H5zzV98mfd3/s+7HMtz/lO+ZZ1zFPeT7Po0L5i831LfabHuyIts622uftVLNNSjjgXj0ZF6yt8x7W1oYOlKk8ovxVs3X3LTeon/V5GDV+ttbl+u1cuklPWM9n3XtOvUaR1vk9t+rPzmFzT8fcZE93trQ6hDBTM26apwd+7Wlv+jMN1cTkcYr8IkMpywoaPELG+zvk+Vxv8vkOHd30sPW85ijLKu9Jv9f+bNqfE0uTn7Hf/Tbfv3T/+13z2SxQecEqPWBve6Lp6cq1d9l4jRnzsLI+PPldGBEfr0s/XKutdZ6nx76stdrbNl6Jg50GPyf7HtSw/n5snef5+9Nv4E26y/pbdeAEH/bT8X03Af2C4j56aEKs9VcVAAAAAAAAaJm+U8FRtY7B/6Yn+s/QexOytOqmNA0M7acQV8N+xjPbme3NfmZ/cxxzvBbtwwLlfBGkYYP6KshpqqP9YD2W/YbmDuvoNEgVuxdr/PjF2hMUq0nJj+vpyYN1ceFzGjN8irI+djYy3OXa9/ERbV84RcvL+2riA0maGF6hrMVTlPbcc3rgwZ26NG66Hn/gdkWpUEunTdEG75mcyo9oX0mRXki6V1sviNNM61yPJYTr2LZUjRr1nPaeZH2UwxumKH52hj7rNlIPPf64HhrZTZ9lzFb81Mya4OnK6MmaNLKfuljlXnFWecJk/eo/aj8TDTlG01To2KfWdb6/Vo9MyVTQgPF6/PHpGt3lqDaY++Q3ase/P/b9yJqtuF8v1l7nx/DgbqH6t4836e0Crx/vKwuVl2Wdx2rPK3TajC+KlJd7RFf2/Fkz/qDdTj3GzVJC2yKlpTbgPpnRGFPH6O6Mz3T1yHv19OP3akS3z7R69hjftZAqK/SZ9Xn6zHcpMovb0/6V1/Waz86+TVpw73M6/KN4zbzvdvW42LxxRFunDteYubukfrfrMeeZHl5nPdPZb556eNbF+nxOCLOezxwt333yo/h/h54cF6sgq21Motdz/I846/MYp15WuUu/kfZnc1K053vYpGcc4H4n9j6mF2aP0qiF3sGX89ksWKVHZu9QUOw4PTQ5Vlde4Lztw/quP3ev7s6UoqfeqyFX1fsXpdY1gzXiqnKtzgowreFXO5WVWa4eIwYrIsChGvI9sJUXaOmY8Xpk2zFF3pGkp5PH6dryDI1PWqUDzibeTsv3/eibWrKsWL2s53dte6cNAAAAAAAAaImqvkeOucur9n32UdXG99+s+sP/bq15mbppN+9/J727qKpv9OSqjYedekMc31E1f0BsVfwTb1e5nSbbN4VVq+Njq667f1vVMafJc/zYqpHPFjoNxrGq3Adjrfbbq17+m9NkHNtWlWZtO+Pl2s68t8BsZ7Wt9+tgyR+qZlh9GP/CAafB2XbCH6o+c+pVB9ZXjY8e7HM8W8n6qhnWMWe/VrNlVdVh63hW25J3nXq1xhwjkG/cVceO+dylAA5XbZxgrtN6DiVOk/GN017nmgLcj09fq0qx7kftffbse92CfKduKXy2Kt46xwyrPd7rebizH7fOfW/VlpNcStXxY1Xub5xyvfKrllj961t9XuucI/3vn3Ovvds+eeH2qr4D/K7f8sn6yb59q+85Oeet+9mJt7at+TR6/KOwastTSVWr/drdbyTbz2DLp05D9XPxvocBfPay6WP1d+izqpz7B1f1jV9U9Z7X4T19sdqcevV3aKR1bJ9Px7G3q5bEWc9nqff3pe61eZz6M7bvt//nzfLZa0lV11n3bHXNrs49CPBsfP92HKv64NlxVdcNGFe1+i9+9zsQr+d47LV7ffpWzbs98He7Id+D+j9bH60298Dv+Tb4+96wz4aH9ffuUeszMWF91Sf296cx+wIAAAAAAABnl+/kiKP6BAddqND2XTTk6hjF/eT6mpepm3bz/nfR0SONX6i9Ytdb2lARpoSEPr6jlFqFadgdfaz3t2mPz/RZHTVsgPeiHsG6yP5X9+G6+iq7wSM4TN2tzfIO+q94cp1uHlI72sl2eZwSBgdpb9aOgKMGjAM5mdp7+UhN9BopZbP2vTlW2r79TycdXdK0Y1Qof+5NGjz0Ji3NrTOeoq4+Nyn6cqdstOqoXtGhUtFBfeI02f0Jsq7dvz8dBuvXw4J1YNMO7bMbOqpnn46qyHmnZs2dA7t36LB1jsTYMB3eUVBz3z74004prK96nWgkxMEM3XWDdS3jM+pfjyaQsHF6/PaOyluc6jsSzcdB5WUdVLeR92qI9/VbLh0yXMNUoLwGjOAJyHpOQ3r7jaNqG6b+0x9Xgl97UKcu6qYifVDfB6pB2uvaB5I07ItMPbJsV91RNA7Pd+g6TRwX7vsdCu6jIdazPbyl+jmeyKk+Y8/9Doq7vc79bj94pEa3PaqsHZ7p8Gr0G64BftvWKteB9CSN31CuEY8/pYSejRu3Fhwdp2FBBdqwxftv0UFtyShQ0IC4gCN0Gv49cK51cN1r7RI3Sv2dcrXm+Jvhr2LXc1qwraOm3huvS1s5jQAAAAAAAEAL9b0Kjr6v2ne8wik1XPlRM3VYN13p99uqEdz9Z4pQgT74m9Nga6+ggFNbNVCoda4A+192Raj08VEdc+r+jpl+lmYoxawL4/Mao6W5ZoNj9f6wX62pxwgKNj+iB+uiHzRg2q6O7epMFedq5buf3Z/QK3RZgB+gu/2sn/RFoT4q9dS79Busbl9sUt77pnZE7+06oshrwtStdz/1+HiX3rNngCtSfk6FevQLt57SCQQF6SLTlYsv1kWelgbrMjJV94cV6YkH1+pAwKm+PtPhj6UDG5L87rH1SlykLdYW5eWnGBy1te69U/RRWaHDRW8q6/nFWjp7vHWu4Ro8fm0DwpoGCO6riY8MljalaGlO4H57vkM7tTTR73qtV8qGo9ZzrFCd2fgCOLVn7Lnfva641K75ClOPKOlw4T75TJAYfGE9U1ke1b5V1nNbVaRL4x9Won9I1xAX9NWQuGDfEPjgLm39MFijhwSeQrPh34MTXGvbUPXwC5Oa42+Gj68KtHreJgWNuFfDvENyAAAAAAAAoIUiOPo+uLi9Pcpi318b9XNo/VoFyaUKVZzqWiCBBJljnqILrlDPPn0U6fe6dnCcEvpdUf+6Tt5O+RhB6jF5vXa8sV4J3gOuThcnZHJX3/urwhXdtlw5uw/aa9zkF3XUL8I7Slf0Uf/LrfpfyqWiHcr6oqP69zlJgNgxTnNfe0Pb0q5r/DpIrTpqSNJkRX68SrOf9xvJ4uXiK8Lr3OPIPv00bFicfnHVKQQS9Skv0PJRN2n4xKe0ruiYgjr20bDbZ+nZpeMU6WzSVEF9xumhuGBtmDNH232XqPLSUVfXuV7rFTtYCcO6y16O6WSa8xk7rK+b1ODv7xFtKGivYf3a68DaJC09xZFh3YaMVI+PM7T9PU99X9Za7b1qpPr39NQbxf970FjN8TfDVqG9z89RekWcZn4rfwAAAAAAAACA04/g6Pvgqn4acpWUtXln/VMwHd2kB2Ku130b7OELCgo2YziO6DOf6egcB4qVp1B1CTSY4VQVFeujAD8Cf1Z6UGpbf6jkuiBYZuqvYZMnKzHQa0T4SUOQ5jiGmnF6Krs/H1v33ql7O/xhofX/XXRZB0/djB6JHBysfbsKdKDgbW1t20897VEPV6hnv2BtfedP2memNqtpP4mmXMflcZo6LkyH1z2s5bv8P2kuXdTWus294wPfY+s1LNzvLjchmDy6bZXSj4Tq/jWvac38JOv4t2tIbLi6WH1oyCifhglWjztmaUTbXXrk4Ux95LRW83yHQtX/jsDXmzj5OjXsK3Qqz9hzvz8qDZRoHdFH9seo44lHoNUI1dSnHtbE5EW6v2e51s1O0lbPn4nGuSJWw3qXa3XWTlV8tVNZmeW6Nj623nvQmO+BCcICXutXB3XAb/rEZvm+1yjUdnv0WKbuuuF69bP+hnpeo/SEyU8zZ9j16r+rAAAAAAAAQEtAcPS9cIUGjOgj7XpKjwT6AbPyiLIeXqztQdfp5gGeuemCe/fVtdqlP27x375ceVu3SVfFqle966Gcim3avsMvbCjfpS2bynXpgH7q5jT56xY1WJd+nKntdQa5HNXetWu1tbjuj8kV5b4jr07lGKeT3Z8vNmmLf/hSeVB5m4oVNKCPengFPN16xSqoaKdWb96pSwfX3iu7fccaLck54tN+Ol06LEkze5Yrfdla7XHaPMIUOSBY+7I21azVU+PoLm1Yu0kHqm9zh47qYv1n+x7fB1Kx601td8onZZ+jm670+4we3bVD+U65WQSHK3FWnC597zkt2OS0OTzfoZ3auqPu5+fAplXK2nWkzpRof/f7bFZr/DP23O/DmzYp3z/D+3Cb/lgcpGGR4U7DyQR5RiiZUWWpj2tE2yKlpWbqcKODvfa69sbrpC2Z2rDOeuk6DYuuP7pq+Pegu3pFBwW81vIdm5TllKs17/e9i4Y8/riervOarhHms9f7drue0KdhER0AAAAAAABwNiA4+p4IHjBLT8cFK2/xGI2ZuUpbcwu0r6hA+ZtW6ZFRY/TEe8Ea8fgURVb/U/v2gzV6RHvlLZuiJzYU6egX5So/Uqytc8frgW3WthMGN3C0RANdHqrDz92r5duKPec6uEvLp6Zo3VdhShx2gimgesZrYp9ya9vxtfta/cxbnKS7n8/UB//wmnSqYzf1aCtlrX3Ouv5d2uesE9SoY3wb7P5UaN1D1v3IPahy+34UaN3M8VrwcZimJvitCRPeV8OCCrR1l3RtL6971T1cQyqKlf9hsIZEfVvTaHVU/4fvVX8zlaHTUq3biCnq/0WGxo9frLziI/Z1HS1+U8vvT9GCzGIdq76oVuGKHBCkwxn36u5lm5RvPautq2Zr/POFDZvazdK+W3fr87lJ6c8X6LC5f18c0b4ND+vedUea93NrCQqfrOQRwarwv2D7OxSsrXPG64l11f04qL3rHtbsuRl662MTyVSzPpt9zPRtK7Vh2y7tPeiXgJzCM7bv91ebNGPqYuVbxzP3+/DuVXpg4ip90HOcRkSdwuc6OFwTF0xWr/cXa8bigjrP+GSCogbb17E0vUBB0b9UjxOti9bg70GQIm4fp0ifaz2iA5vmacqqY3UD7mb9vrdXl6g+iqjz8oxu0+Vhdr3H5adwrwEAAAAAAIAzhODoeyNYPe5Zr8zH4/Rvf83QI7NnaMzEGbp7bob2dIjVQ8+u1ES/Re+7jVupNZPD9P5zUxR3000a/OvxSsttr8T5dbdtsraxmvrUYB2YN95zrtGzta78Gj20fJH6ewZB1aO9rn18jZ6OkzY85uxr9fO+TdLotGf9+hmmEY+adVYyreufrTmbDjrtjTnGt8GrP7PHaLB9P2Zo9ad97fsxxP+H8FbdFdHPFPoqorvd4nHBzxTRx/pvUKwivq3cyGh/nWbOus433DKs9ofWPK7RrTbpvsRR9nXFJaZqQ6s4Pb10snrU3OYgRUxepKl9grQ3Y57utp7V/F1XaFLqOPVwtjipniP15ORr9PeMGRpu7t9NozRlSxfNsp7/lc4mzanbHU9paoCpAGu+Q6uq+zFGd60qVk/r+h4b5v3BDta1k5M0pNU7WvDYbN2V/iffUOZUnrHX/b57tDn3TRo+I0OfRU3XmgVxuvRUpyWsnpIwc7YeaOwUbK3CNSDOPOhgjR7iF4DW0YjvQcc4PbZosgaUV1/rKI1adUyjn0rSL02A4+Ns+74DAAAAAAAAZ5dzqixOGd8nX5XLzIoVFBysoJP9gFxZoXKzcasgBQc3/7+c37vwet21b7Iyl8bJTOhU8UW5Kk7lXI3op5muLijQNqf5Whutuj9BwQo+0eiMlqbC+vx9Zf33Auu6TnSbzfVXWs/iVB/FWfQ87c+1rH60PUk/rO9mhfW8T/q9bIzGfN/PRo34HlSUW/fPfGZOdp+Ns+37DgAAAAAAAJwFCI5wxvkHRwAAAAAAAAAA4MxgqjoAAAAAAAAAAADYCI5wxl0cFqeEflecZL0TAAAAAAAAAABwujFVHQAAAAAAAAAAAGyMOAIAAAAAAAAAAICN4AgAAAAAAAAAAAA2giMAAAAAAAAAAADYCI4AAAAAAAAAAABgIzgCAAAAAAAAAACAjeAIAAAAAAAAAAAANoIjAAAAAAAAAAAA2AiOAAAAAAAAAAAAYCM4AgAAAAAAAAAAgI3gCAAAAAAAAAAAADaCIwAAAAAAAAAAANgIjgAAAAAAAAAAAGAjOAIAAAAAAAAAAICN4AgAAAAAAAAAAAA2giMAAAAAAAAAAADYCI4AAAAAAAAAAABgIzgCAAAAAAAAAACAjeAIAAAAAAAAAAAANoIjAAAAAAAAAAAA2AiOAAAAAAAAAAAAYCM4AgAAAAAAAAAAgI3gCAAAAAAAAAAAADaCIwAAAAAAAAAAANgIjgAAAAAAAAAAAGAjOAIAAAAAAAAAAICN4AgAAAAAAAAAAAA2giMAAAAAAAAAAADYCI4AAAAAAAAAAABgIzgCAAAAAAAAAACAjeAIAAAAAAAAAAAANoIjAAAAAAAAAAAA2AiOAAAAAAAAAAAAYCM4AgAAAAAAAAAAgI3gCAAAAAAAAAAAADaCIwAAAAAAAAAAANgIjgAAAAAAAAAAAGAjOAIAAAAAAAAAAICN4AgAAAAAAAAAAAA2giMAAAAAAAAAAADYCI4AAAAAAAAAAABgIzgCAAAAAAAAAACAjeAIAAAAAAAAAAAANoIjAAAAAAAAAAAA2AiOAAAAAAAAAAAAYCM4AgAAAAAAAAAAgI3gCAAAAAAAAAAAADaCIwAAAAAAAAAAANgIjgAAAAAAAAAAAGAjOAIAAAAAAAAAAICN4AgAAAAAAAAAAAA2giMAAAAAAAAAAADYCI4AAAAAAAAAAABgIzgCAAAAAAAAAACAjeAIAAAAAAAAAAAANoIjAAAAAAAAAAAA2AiOAAAAAAAAAAAAYCM4AgAAAAAAAAAAgI3gCAAAAAAAAAAAADaCIwAAAAAAAAAAANgIjgAAAAAAAAAAAGA7p8rilPGd5dahgnwd+tKUW6tzRLg6t7HfCMhdUqA9H31tl1tfGaGITi67LJWpeGehPrXLbdUtMkwdWtmVgMo+yNNez8a6pHukQtt5yqes0rqOojzl5hWoxL4WyyWhGnTtAHWv6SMAAAAAAAAAADhVBEffC6V6cUyUZuZ4ap3HpGvrA5GqL2opfSlR19zn2Thmbq5W3trBLksFSv3RcC13ajGp2Vo5opNTqys/rZtu+a2nnLh+n5J6e8qnwv1BhpKnpejFIrfT4s2l7iPm65lHB6jzCYIsAAAAAAAAAABwYt+rqerKK75U8dEDyvogW5l/faPmZeqm3bz/fXBoZaJScwIFMI2TnTRN6fudyulUkqG7hiX5hEYhXcO8RjC5VbhuovpP26IypwUAAAAAAAAAADTed3rE0T8r/6kdB3frlf/dqh0H3tXnX588VmjXOkT9uvxcN/97f/W7orfOb3W+805L5jviyNZmqJ7ZMV+DAkwf19ARR7auU5W1eYK6Bxjp0zwjjkqUfkuMkgs8tdBbU7XwgXh1D/HUy4rW6J7hKco+bmqdND0zW5N62m8BAAAAAAAAAIBG+k6OODpS/nfdv/VJ9Vw2RLf/YaZefX9bg0Ijw2xntjf7mf3NcczxvnOOb9Q9921UqVM9ZfsX6J5FBWr6+KV6lBYoxwmN1GmqFs6pDY2MkLBRSpoR5tRKtDn/FIdAVbpV9nmZ51XW8KtxV+9jvdyVTiMAAAAAAAAAAC3Udyo4Oub+UjPfeFJRz8dr3d4sffXN1847p8bsb45jjmeOa47/XeLeNkv3v9Lk6EjFi2dpScFpio7aD9DCPXv0Z/PKGhtwZFPbCy9xSlLZ8Ub2o6xILz44XD/t1kM/7dXL8/pZD/0kKlGpG/cHDsQqS5W7OFHX9Oimn1TvY71+Yh3jmsRlyvW+pZX7tTyum676kXn1UOpup92HGRFWvU03zdx22mI4AAAAAAAAAABO6DsRHFVZ/9tc/Jb6rhyhF97LUmXVv5x3moc5njmuOb45jzlfS5bwaLJi7JJb2Q8m6cUSu9JIkUp6dKw62+X9WnLfAuXb08U1s1YuhbQL8bxCXE6jN7f2/jnPKbsU07OrU26Ako26KyZOM9cW1FkbyX0kR8vvHqQhaX6jqSpL9OL4GI1ckKPSOtfrVum2BRo5KEmbP3eaWnXVoOHhTsWt9Deqh095sc71evU0gm1G6ca+ga4TAAAAAAAAAIDTr8UHR2Ydo0dylmhCVrL+Uc90dEGtzldk53DNuf5ebRudrvcnb9GBqdtrXqZu2s37ZjuzfSDm+OY85nzmvC2V68ejlDI3WnY8cTxHySkZOtToadZcComdqoV3OkHN/hW6/3ROWeelZnq4AwVKTxquu9Z6zuqKTdYdkQ0MXSqLtOT2aTUBj6v3BC3MzNaf92Tr5Ufj1b2Np734t79R6s7aqyp7fX7tiKCuo7Qye68+/Ns+fbg3WytHO/fi8wzdsyiv5l507nujIpyy+7U8Ffrd67J3c5TtlF3xA9SL3AgAAAAAAAAAcIa06ODoy4rjunvzY1pZ8FLAUUaXXfRDPTVgpt6b8Eetu3Wh/qvnUIW276LW5/n+Mm/qpt28b7Yz25v9zP7+zHnM+cx5zflbqs43Jysl1nMf3NuSlJxxKsOOXIqYMkeJTl5S/Ntpmpd3uqOjAs2rnh4uZriS1xVJHcMVNzNdO56NV+cAU9kFUvbaUs2rXg4pPFlZL0xVXM9OCmnXSREjU/XyklGeYM2MEsrYXjMiqfgvG52SFDNugmK6OJ+lNp0U88BSPTN5gibNTNUzAy6pDdG6DNTwaKdckqHs952yrUy5b2xxyi4l/CrSOS8AAAAAAAAAAN++FhsclbnL7fDmjx9Uz/FV6wetQ+zgZ+fYdYrvPrhOUHQyZnuzn9nfHMccz585rzm/6UeL1KqTbnt0jmKckTXZSdOUXh2kNEabcE2fO0GhdqVEyxPTlP0t52nushIder9I+xq8XJN3WCMN+n83KtQvcHL1m6CX09O11rz+X1hNmBNyaZhTknJff12F1VPSGWZauqlTNf3OeMVEdlXtp6aDBt88wCmXaP32IqdsKXtbm6uzqE4TFFc9qx0AAAAAAAAAAGdAiwyOzDRxD+cs0tb9/+O01Iq+8hpt+c0qO/hpdU7TLs/sb45jjmeO68+c3/SjxU5b13GoFs4d6oQiBUoev6zONGoN4QqfqoUznGFHx9forllb6qwZ1HzCNX3PHv05O1Nr0+craWiYXMdLlf9KmkZGDVLq7oaMeCpRcc1SQ5GKDq8bDKpVB3XvG6ko8wrvVBMchQ68oyZsc29L0ZBe3fSTqDiNnJ6m9I0FKi4NfP6QvoM0yCkf2pyvYqfszs9TplPuHB+t7g0cMQUAAAAAAAAAwOnQIoOj5/a8qFf+d6tT8zj3nHM1+mfD9PyNqfph8L85rc3DHM8c1xzfnMeb6YfpT0sVcsMcPTPSiUX2L9A9iwr0hafWKN3vXKoUZ7SMe+M03fNKg4f/NJqrXYhCuoQpqu9QJT6dqa3V6zVpv5ZPW9GA8MutspqZ+aw9z3OKDdFxqJ7ZkKy4jtVRknW0I0XKfWWFku8erv59euinCVYf/EddtbtWg25wyu+9qtwDpuBW7rY1dpPUScNja0czAQAAAAAAAABwJrS44GjPJ3v1/J4X9S+vNY1MmDOlT4KSoyfr/FbnO63NyxzXHN+cxzs8Mv0w/TH9aplcipm1XImdPLXixbN0//rq8TCN0KqrEp5OVYxdcSv7wYl6/F27ctp1vnmMEpyySt7W3sNOuV4uhTjXK30q95dOsYFcV4/Swty9+mt2plbOnanEmyMV2s5501K2M0233L1RvtFZiGJuGOqUC5TzrvVuZZHyXnGaeo7VIHIjAAAAAAAAAMAZ1qKCo6+++VpPv71an3/tOxHazf/eX5OuGdXkqelOxhzfnMecz5vpj+mX6V+L1CZS058Yq852Zb/yd9cMx2mcTvFKSY32lI8XKL9mOrhTV/pakobExXlez9RzwG/cqp0g7pBKTzpPXieF1qwlVKS8ogA7uEuUvzNPueZVUOJ1/FquLmGKuXWskuala+ueffpz+lhnrSdr922vKtdv0FVI7I01AVf2azk6VPC60p2RSRHDo2v2BQAAAAAAAADgTGlRwdEfP8jRjoO7nZpHxKXd9XD0lNM20sifOY85nzmvN9Mv07+WyhU5VQvvdNYpaoLO8clKi62dxq2pQkJcKn6vSIXmtXSZXgyQaR16aa3SnbIUre7el3G8RNlrV+jF3d4pToiirh/glKXNi5Yp329qOXfeCv1XQoJGmteG6hFY1rEWL1DyXSbIStHmz51mR0jfgc6IK4+v/afMc0Vq4Gjn3uRs0bzfve4EUuGK61szBAoAAAAAAAAAgDOmxQRHZe5yrf7TBp8p6tqc31oz/vMOhbiCnZbm98/Kf2rRrtV6vXiH0yL7fOa85vzVTL9M/0w/WyaXIqbMUWJTs6NWnXRbcrJi2jj1JnJFxmtSdZ+O52jmoDjdsyBD2WYk0LYMpSZGqf+DtYFd6Ix4xdTkViVK/02MxjyYppnDrf96rbsUcsNETa8+7v4V+q/hSUrfuV+lR/Yrd22Sbpm0xgl1XEoYGGlWQrJ0kOvoCqW/boKsNbpn3ALlHiiTu1Jyf16i3AVzagOsrr9Qr45OuYZLvWLjnWPlKHOjk4JFx2twF08RAAAAAAAAAIAzqcUER+98/BcVle53ah43/vh6RXaumXOs2ZnQaF7eSi3I+29N3zLHJzwy5zXn92b6Z/rZYrUJ1/S5E5o+ZVqneD3xaLQTkDRRqzBNWjVfg6rXEDpepMzFSRpjRgIlJmn5ttKaaeRCb12qlXd6LxRUqpKa2e3cyn7fa7iSfdzUmoDLXZSh5IRBuiZqkEY+mKFCZwRS6J2/U1Lf6itxKWrG72rCNffuZRoZ00s/6dZNP+kVo5GLCzx9aWMde86ogPfRFXmjEvxCtUE3D1QHpwwAAAAAAAAAwJnUIoKjb/71jTIKN9n/rdaudYhG/+xmnWP973SoDo2e273OHk10zP2lT3hkzmvOb/pRLVA/WxpX+FQtnNHUYUdSh5vnaeHQZomOpE5D9cw7uVo7c6giqgMkLyHhQ5WUnqutcweocyun0RamQVPDPAFWm3ClxPuFjJ3itTI3U2kjw1X7FD1cHaOV+PRmZc0M9w3ArOMkvZyphXdGq0OdUVUhirh5pta+kanpveu59lbhihvnPS3dAA3q6392AAAAAAAAAADOjHOqLE75rPXJsU916/pJ+rjs/5wWaVDoL7X0V8k679zznJbm4x8aeYv7yfV6sv/9cp0XZAdEE/+Yos3FbznvSpeH/FAvDV+iyy66xGmBjyMZGhOVpGynemLRSstdrtv8pnxzl3mmhzNcISFy+YRFdZntdeFJtqt0q6zMGbvUymWvrXRS3vtYGtIXAAAAAAAAAC3E5/uVuy1HeYfKnIYOCr0+WoPCOjXpd0B3SZE2b39dxZ86DY09rtWv/J15yi6uXpojRJ3+I1JRfcPUuc7Pmm4dKsjXoS+dagO0vjJCEZ3q/j5atj9Pm7a9rZLqY13YSZGxAxXVtYH/ML6yTMV5ryv73RLV3NHQgYoZEKjfOJNaRHCU89Eujc1MqhnJc+4559qh0Q3dou16czpRaPSrq6Pt0OjCoNqhJq/ty7HDo+ptTZC1Ii5V0Vf2sevw0wzBEQAAAAAAAACcNpWlyn5you75bUFNwOGj3QClvDBfCVc3Mu04vl8vPjZNyeuKapbf8GGO+7s5SgirJ4g5Wb/ahOm2OUuVNtR7xqNSvTgmSjNrl4k/qZi5uVp5q9fiGqV5WjJ9kubtDHhWucLG6pnnZyrmBL/jlhWs0F13pCn3c6fBm+n3w/OVcmvX5ln+BE3WIoKjebkrtWjXaqcmXXLhxVo/fJGu+oH3F6DpGhsaGR/+o0TD10/Rp19+5rRIU/qM1vSoMU4NPtwlyn/3kL52qifWWp1/Hk7aDAAAAAAAAODbUVmiF8cP0sxtAaMdL12VuD5TSfUtWeHveIFS44Zrue8y/gHUc9wG98ulmLmbtfLW6t/OmxgclWRozKAkZTtrwterTbTSNi/XbQF+snfvTtOQ4StU7NQD8+83zqQWERxNei1FG99/06lJvS/rqdU3P6HgoAudlqY7ldDIKK/4UqNfuV+7P3nPaZGG/vg6Lbkh2akBAAAAAAAAAFqC0lcS1W96Tu2IoK4DlDRllLq3c+tQzjIlryyofa/TWL28daYiGpAdFT4zSEOerE2NQvpO1cLkePVq71bxa8v00JwMFVaHMwGOW7Zxon569xan5tn/kRk3KqaTdGjnq0p9eEHtaJ428XomN1WD7IFLZcpfu0LZNdPiBfBRjpZsLHIqUsy8XK282QRHZdp8d5Tu2lh9xS51HzFTj4y7UaGuUu3JSNE9C/JqRz9Fp+qdlfHyGqskHc9R8jWJSq8JnkIUNXKCboy+RO68bK1ft7H2uhWulOz1SujiVHHGnPXB0bcRzJxqaFTt2wi2AAAAAAAAAACnUWWRlsTEaV6JU28zVAuz5yvOKwkpXj1c/VMKnJo06Ok9emboSdb4cecpuVdCTXjiik3V1mfj1dlrPSP/UTm+08WVKjMxSvdsc6rhydq6fpRC/fbvb+1/yKknrt2npEinckJuZT/YS2PWOuFQm1FauydZUSa0OrBGt8SkKN/zjkLvXK+smeE+08kdeilR/e+rDtpcSly/V0m97Yqt+LeD1D+tOjALMJrqgxW6ZVBazTk6z8jUW3eFOTWcKec6/z1rffOvSn31je/EZt3aX+mUmq6poZHh3x/TX9NvAAAAAAAAAEALUfCqllSHRpaIWdN8QiMjdMRUJXj9XLz5lddV6pTrVbJfe7ymertt5I0+oZHh6h2v3/R0Kpbs19/2Om6IIiaka22685pzo09oZLg6dVWoUzY+/TzwekR17M/QkurQyBIxa6wnNLK4P9xfE+hIYfpNvG9oZHQeOlK3OWUTQqXvqA3VpCJtXlM7yso1OlnT/afgu3qUFq6vvba0X4TUjujCGXPWB0dff1Ohz79q4Ifc8ur72/TYW0vtQOhkmiM0CsT01/QbAAAAAAAAANAyFBfmeYUWYYr7RYD1dlwRir7ZKRs5Rdp7sqTj8xIVOkXDFewfvxghcl3sFI1thdpXMzbBpc7hkYrq67yurjvCyf3XIuU6ZTPlW0T3k4yCsrmV/d+1o33MaKPpXmsMlf3dK0XTJXIFmmDL1dYnTHLvLq4Z9aSSQuV5HeK2ayM827pLVbw7T7k7zatI7q5e1xbeqU44hW/fWR8cNYYJje7bMlcr8l/So28tO2F4dLpCIwAAAAAAAABAy1N2uHadHylCnQPkRibECe3uPQfcfn16siFHHbsqyikae973DmQc7mIVveOUbft16ETHdZep7HPrdaBI2WuTdMukNTWhV+ids3RbQ9YJOsFoI6PD5V2dkpGnouIACdmBQu1xira8/bUjpQ7t9wqzohX2E6n4pSQN6RWl/sMTNDLBvIarf68eGvJYjkqZxOuscdYHR63PC1K7C06ejlaHRmaaOBME/e7Pf7CDoUDh0ekOjUx/Tb8BAAAAAAAAAC1BqfZ94BRtLrWtZ+hL6/O838jT/iNOsT6duivSK4TKn5Oi9A+8QpjKUmWnWW1e09mdTOnG6fppr176aUycxjyYoUJr35CuA5T49OY66xAFVqbNi+ofbWTrHqE4p2hGJ6WnpCnbO8wqK1L6Y17H8FP6ce00dcb+FyZpyH2evvpyq3BlovqN33jyaf/wrTjrg6Pzzm2lC85r7dQ89h39yCl5uL+p0Bt/y/VZC8kEQiYYWvLOGlV6hUOnIzTy74/pr+k3AAAAAAAAAOD7LkwJM4fWhjnHc5Q8qJeuGRKnIXFx6n9NlMasLlFE73Bng1Pj/rJEhwqLVNyQlV+K1mjexvpHG9lCBihxhteoo/1rNKZPL/W3+mz6fU1UnJK3dbL6HXBolp8cLV+cY92KeKWsz9af9+zRW+uTdZvX4d3bZmnJzgCjmvCtO+uDo+CgC3XpRb4rkB0+Vqryii+dmuQ6L0hzYqfbwY83Ewwt2pWuRW+vtsOj0xEamX6Y/ngz/TX9BgAAAAAAAAC0BC61vcgpNopLrgaMIQi5YY5+PznMaySQW6VFRSp8r0jFn0uht87X9GFtnfc8Wp/guCG/mKC16elauzxVSSMjZebsch8p0ubfTtOQqES9GGA2vFpl2vzcMhU7tYCjjRzd73xeC2/wnhGsTMVWn02/S4+7FDFjviZFeJ2szQnGOnUaq5XrU5XQu5NC2oWoc+9RSlu/VHE1P8m7lZ6x3ToDzrQWscbRVT/o7JQ8Dn7xifWhtL5NXkzgY4Kf+sKjBXkr9VTuimafns70w/THm39/AQAAAAAAAABnsxBdcqlTtNW/xtCnpTWRiyVSnXz2q49LEVMztWN9qhIHhqmD/VN0iEL7DlVSeq62zh2gS9yf2lt6dFVn3/EUPlydwhXVN1JRsfFKfDRd72yeqRjnPTOiaWbaxvoDmIaMNqrWqpPins7W1uUzFde3qx1QqU0HdR84VguzcvXyXV2l2jEe0jWdVP3ruKuNbxLXfWy8ovx/gm83QMNHOGVjW2FtoIUzpkUER70u667zzj3PqUl/P/65/tfny+lhgp+062eof9f/dFo8TFC0eNfv9OzuF5o1NDJMP0x/qpl+mv4CAAAAAAAAAFqOzj/2HpSQp8IDTtFHqfa+6z3CJkzdOjrlBujQO15Jz2Tqnb379OHf9mhr+nwl9jUJUan2bC/ybGQMjVCoU2wI19U3aqB391+rL4Bp+GijGq1CFBo7VgvTN+vPf7P6vTdXWc/MVFxYiFRZpLxXnO0snX/eXdV5V8iVoTUhknHJhd4jl2q1rv3pXzrOVHVngxYRHF198VX6YfDFTs0TBGX+dZu++dc3TkutEFew5g9MqhMeBdLU0Mic3/TDO4wy/TT9BQAAAAAAAAC0HB0irlWEU7anTftjnvX/fo7k6PUcp2xxxf9CJx1G4M5R8o+66Srn9cvFXgFRtaIMPe913EHX/8IzuscoydBIr/2HrN7vvOHN6mlDMpfGjDbav0ZDvM5718a6Y5jKXv+d0o87FXXS8L5hTtny40gN8vrpPbewOEAXS7T3XadoxHb3CZtwZrSI4OiSC9ur5yU/dmoeu0r+pA8++8ip+WpIeNTU0Mgw5zf98Gb6afoLAAAAAAAAAGhButyohBucssW9OkXz8rzCksoSvfhgirKdqglKEn4VUbtuUWWpchdP1JC4BCW/tL82JHFFKMLruIcWTFPyttp58NwlOUq+e4HPKKDfDPAandOhk0K9fsYufDJN6UXe/SpTodXX5DynboR3qhn5U6tMmcu8ztNprGafaLTRlREa5PX25lnTfc5b9t4a3XPfxtrrDB+ruJ5O2WgVrrhxtQcw9zM1x3f+v0PrrLYCp2KJiA4P0G98286psjjls9obf8vVuI0P+owyGtFjiNL636tzrP8FUuYu1/ish/Q/B/c4LR7NERpVWf+bufUprdub5bR4pql7buijuv5HUU4LAAAAAAAAAKDFKFqm/kO8whW51KH3AN30Y7eyX9uiYq+l911Dl+qdpwfUjAwqfSlR19xXPWyok6ZnZmtSdZBS57j1cSlmXrZW3uwbnxxal6hfJnkNSbK4OoYp1NrMXVLk0y9zjISVe5QS7TeU6L1l+mXcAh1yqjFzc7Xy1hPHNGWvTdM1k7zCoXp11aSXMzU93O+cx3OUfE2i16gk535GttWh117V5v1eAVjXsXo5c6YiTv1nezSTFjHiyLjm8v9QWIeuTs3j1fffUN4hrzjSjxl59OyQR/SfV/RyWponNDLMec35vZn+mX4CAAAAAAAAAFqgsAlaOa82DDLzv5Xu3qjla/1CI2u738/x3k761Gdd/hIVH/IKReocNxCXImas1zN+oZHROX6e1k4Os7ao5T5SpML3/EOjEA2au7luaGRGGz1XGxqZ0UaThp58bE/IDXP0e7/z1tVVty15vm5oZLSJVtKGZMXU/Bzv3M/Fa3xDo3YDlPY8odHZosUERyYEGv2zYTr3nNouH//n11qY99/2yKL6eIdHzRUamfM9+T/P2+evZvpl+mfOBwAAAAAAAABomTrfvFQ7slKVEB4o5glR1NR07cicWifk6D50mga185RdYVOVEOu7v+e485UY26FOEBMSPkppWbl6+a56QppW5ryZ2rE+NeD+pl8RN8/U2tw9eibQ9HPvrdG815yyJWbKWEWcOA1yuBRhzps+U3F17odLHWKnWufcrLQb6p/yznX1KK3MzdTCO6PVwf+n+TYdFHPnfGVlL9VtXZw2nHEtZqo646tvvta4Vx/U9gPvOC0et4QN1BPXz9D5rc53Wur65l+V9n/PO7eV/d9T9c/Kf+r+N57Uy0WvOy0e13a5Rs/d+KguOK+10wIAAAAAAAAAaMncZSUq/sshfWGVW18aqu5XdpDrRD8xV7pVVuaWKySkQdvZ2oQopEEhjhfv/eVSSLvGHuAUuctU5kw7d9JrrIf78zJn6rtvsd9olBYVHBl7PtmrsZmz9PnXtcPYzGifKX0SNOUXo9XKa0RSc6us+pcWvb1ai3al619WuVq71iFaETdHvS7r4bQAAAAAAAAAAAC0PC1mqrpqJpy5o9dtPlPWmRDHhDkpOYvtEUGngzmuOb5/aGT6YfpDaAQAAAAAAAAAAFq6FhccGeN63aab/72/U/MwYc7qP23QHa8m6f/K/+60Ng9zPHNcc3zv0Mgw/TD9AQAAAAAAAAAAaOla3FR11crc5Zr2eqq27v8fp6XWD1qH6IFfTtCwsIFNmrrOTE23oeh1PfbWMv3Da2q8av27/qfmD0xSiCvYaQEAAAAAAAAAAGi5WmxwZHxZcVwztj6hP36Q47T4uuyiH2pa5O0a+uPr1Pq8hi+y9fU3bm18/03Nz1ulT479n9Pq61dXR+vJ/vfrwqA2TkvLUvFFuSpMoVWQgoOD7DYAwFmiolzlX3mKQcHBCjqFhSYBAAAAAACAU9Eip6qrZkKbpwc9oDHhtwYcWWRCn3u3pKnnsl9pxEv36PfvbVTx0QN2MOTN1E27ed9sZ7Y3+wUKjcx5zPnMeVtcaFRZrn0bUnXXDdcr9qabNNi8ht6g2F/P1oaCcmejs9vehder38RMHXXqjVJZoXITmFU69TOh8ojyF49XbEwTruM7pUBLzb24aZ7y6/sIHsnUfdY292044jScrY4oa6J1LQsLnHo9nOvpNzFDh+v7LO5erH7WNkt3O/XG+qpc5eV2NHz2cK77lK6p+p4NTFW+E6bUp2JH6mn6fjmf1ZM934Aa+NkwyouVlTJKsQOdv9HWK3boeKXv4q8FAAAAAAAAvh0tOjgyzm91vh6KnqRlQ1LsKeoCqaj8p/IOFWjWG08pdnWCfrx4gLosuLbmZeqm3bxvtjPbB2KOb85jzmfO26JUHtH2lDEas3ingmIn6+nla7TpD2u05vHJGta2QAumDdcDmWf7D/NNVLpJj9x0k5afyu++zaD8/Uw98etRuntTubq0dRrh8cUmzVi40zMK7vui6Dk9svagU2lee5+7SYPv3/TdCyYr3tSGnBNd1VFtf/XNFvw5OqKtSVP0xK5LlfjkSutv9B+06YVn9VC/o1o+815t+NDZDAAAAAAAADiNWnxwZJxj/W9Q6C+1c8w6/brnkCataxSIOZ45rjm+OY85X0tzNCtVj+yQRqSt19PT4xQR2lHBbTuqS1ScJj67Xs/EBWv7ssXazj9qP00KtHr8c3q/d5IyNz6lYZc7zfC4IEhB21KVtuX78wEMsq75g1X3Kr3IaUCDbM/YpsNOuY6D27ThVEdpnQ2KMrX8vSAlPvGkRvS+wvobHazgjqHqf9/DSmx7UEuzzlDqDQAAAAAAgO+V70RwVO0i14VKu36Gcu/I0IgeQ3TBea2dd06N2d8cxxzPHNccv2UqUtaqIil2ihL7BDtt3oLVI+F2XVuxS3/M8R11VP7+Ji2dNkpxA69XvxuG6+7HMrXvC+fNau+t0phRD2v7kaPKXzZDo+yp8MZowYZi+1/+V7yfqQXjh2twzA2K+/UMrfObFm/v86M0JmWTjpopmh4bY58r9qZRemDem/VP5eXnaO4qPWKf43oNvmW8Hlm1S0e99rXPMXWV9ljlrDlWedQoPbHJN6Q42TGa5lINeXa9Vt53ndqfYK2SQPf7QENmETTTEG6apwd+fZM9TdcJ+1+6S+tme+6zOccDy8x2RVoX4J7oi2JtnTdDY266Qf2s5zdqfKqy3j8N0xoOuFdPxgVb53pYWxs68M1M+7dqtu6+pbpvs7Uu12/n0k16wrqude859Rp1r7f6c3j44CYtGOO5j97TqtV5NrNXKb8Jg/R63Z6sqT3Ltfzhxdp7kunXang/j4E3acy0edrq/Tyc603bYpWLV+leqzxm1CrttarlO+ZZ5XnK8/n+Vih/sdlmsd8UcEe0dbbVPm+nao7ud7/Ndznd/36bET8p1n7PF6i8YJUesLedoqwT3KfDmbOt80/R8gZMxTYiPl6XfrhWW+s8T499WWu1t228Egc7DX68v+P23xj7s++86c36W2Tus/lbZu7zXSf5Hjbb344f366Vf1ijEWFOvVqr9mp/uRk969QBAAAAAACA0+g7FRxV6xj8b3qi/wy9NyFLq25K08DQfgpxBQpM6jLbme3NfmZ/cxxzvBbtwwLlfBGkYYP6KshpqqP9YD2W/YbmDuvoNEgVuxdr/PjF2hMUq0nJj+vpyYN1ceFzGjN8irI+djYy3OXa9/ERbV84RcvL+2riA0maGF6hrMVTlPbcc3rgwZ26NG66Hn/gdkWpUEunTdEG7xm6yo9oX0mRXki6V1sviNNM61yPJYTr2LZUjRr13El/VD+8YYriZ2fos24j9dDjj+uhkd30WcZsxU/NrAmeroyerEkj+6mLVe4VZ5UnTNav/qP2M9GQYzRNR3X58Yk/gxUFzv1u67kHT0+P12XW/R41Za0OnLAPR7R16nCNmbtL6ne7HnP6f3id1f/Zb9b+8G98nKn7Rs3W8r9dql/fY84xUpf+JUVTFm/TAesZfvaV1yRf5QVamjheabuDNGBcsp5+fLKG/KDIDgGbf72hduoxbpYS2hYpLbUB97zyiLKmjtHdGZ/p6pH3Wn27VyO6fabVs8f49q2yQp+Z6/Jd1szi9rT7XK/1Ody3SQvufU6HfxSvmffdrh4Xe946mjlFcePnKU/9lHif575d/WmG7h49+9RH6bXqoiFJ4xT5Rab1PSk4+fRq/s8jeZwGBBUobfwYLd3tPOXgcP3K+mwP6W6VO/TTr63ypAn9dKV5q1uo/u3jTXq7wOtMlYXKy7Ku22rPK3TajC+KlJd7RFf2/JnsT22A+53Y+5hemD1KoxZ6971Cxz61jlewSo/M3qGg2HF6aHKsrrzAedvP4S0Pa8bCffpxQpIS+7R3Wk/gmsEacVW5VmcFmNbwq53KyixXjxGDFRHgD53/d9z+G5M1W3G/Xqy93l8Sc5/HjNcj244p8o4k+z5fW56h8UmrdMDZxFuz/u1oFWSPMgryD5c/3qG3iqRrf2z+ggEAAAAAAACnWdX3yDF3edW+zz6q2vj+m1V/+N+tNS9TN+3m/e+kdxdV9Y2eXLXxsFNviOM7quYPiK2Kf+LtKrfTZPumsGp1fGzVdfdvqzrmNHmOH1s18tlCp8E4VpX7YKzVfnvVy39zmoxj26rSrG1nvFzbmfcWmO2stvV+HSz5Q9UMqw/jXzjgNDjbTvhD1WdOverA+qrx0YN9jmcrWV81wzrm7NdqtqyqOmwdz2pb8q5Tr9aYYwTyjbvq2DGfu3QSh6s2TvC7DsueJ6y2O9ZXfeLUbZ/tqHph6R+qPviHUw/kH4VVW55Kqlr9bs0TsbnfSLaf+5ZPnQbrmeQ8OLiqb/yiqvd8NrWe1aNWu89zcVftSbXa4p6q2uN72KoPlsZX9R2QVJVzkttS5T5mfaeccr3yq5ZY5+27IN9TLXy2aqT/s3Cem3fbJy/cbvXB+kyXOA2OT9ZPtq7j3qot1X2r75k75/U+pudzGG9t63fB1v354A9PVaUszff9LlR/lv9QczLPc62+lvr4Xc+x15OqrvM/r/Odqu238zwCPLs9T8RZz+nZqg++cZosdb4nNk//rvPun3W/463PyAyrPd7r++vOftznPtr32/wN8bvfn73m6fvqml2dexDg2fg/i09eT7aedXxV2utez7o+Xvsee+1e32fs8G4P/HfCuuf+f2M+fa0qxfob4/23q77P1kerzT3we74N/tvRwM+GF3dJftWe/3m7asvKpKrxg60+Prit6hOvZwwAAAAAAACcLt/JEUf1CQ66UKHtu2jI1TGK+8n1NS9TN+3m/e+io0cavwB/xa63tKEiTAkJfXxHKbUK07A7+ljvb9MenymvOmrYAO/5lYJ1kT2AIFxXX2U3eASHqbu1Wd5B/1VKrtPNQ2pHO9kuj1PC4CDtzdoR8F/6GwdyMrX38pGa6DVSymbte3OstH37n3xH3ATQtGNUKH/uTRo89CYtzT3pmJETMmveqPSgDnvf1/Z9NWJCnLq1deqBtA1T/+mPK6G374imoE5d1E1F+qD65n3xjrbvqNC1t49UD59NgxU5LE6XOjXbV9a2WyoUmTBSEX4DpbqNHK/+Fbu0vXqUSyBf7dQC654Mvmme3xRoJxE2To/f3lF5i1N9R7X5OKi8rINWP+7VEL+1oi4dMlzDVKC8E/XtRKxnPsTvPpr70y1uuh6aEO77XbjgCnWxzp/3YX2fzoYJHjBLj8WWa91jzym/vm47z2PYhHF1nl1EfLy6fbFJee87TfXqqJ59Oqoi5x3tc0bCHNi9Q4f73KTE2DAd3lFQ8z374E87rWfRV73s77DnfgfF3V7nfrcfPFKj2x5V1g6/hZr6DdeAE6zjdTQnVTPmvKMr71mk+wf4fe9OIjg6TsOCCrRhi/fftYPaklGgoAFxujbAwCX7Ox5k/T3x/453GKxfDwvWgU07tM9ucK51cN1r7RI3Sv2dcrXm+PtTn/Jdq3T37Nl6JH2XPrm8r4YMCtOlJ5jmEgAAAAAAAGgu36vg6PuqfccrnFLDlR81031105UBftMN7v4zRahAH/zNabC1V1A901E1SKh1rgD7X3ZFqPTxUR1z6v6OmX6WZijFXsvF+zVGS3PNBsdOOgVYU48RFGx+yQ/WRT8IMD9WI/QYNk7XVmzS3TfdoFHTUrVuwy4dONrAMKqyQoeL3lTW84u1dPZ4q+/DNXj8WufHcMdXx+wfsS/tEOCX9au6K9Ip2r74TCba63JFgA9A2zBFWI9l+/s+R/dzoYJN2GXdm4saeVu6jEzV/WFFeuLB+qbos/r2sXRgQ5Lf87JeiYtklvcpLz/Fn+vbWv11iv4qPi5S3oZVWj5vtu62zjVq6HgtrTfcaoxgRd6TrBHapBkLA0zBZjjPY8viMXWvOSlDB6wne+y4Z9MT6dJvsFfIdETv7TqiyGvC1K13P/X4eJfes2f5K1J+ToV69Au3vtWG5373usInWnSEqUeUdLhwn3xm7Au+sN5pMQ9vTdWUlDd1IOpezYxrXGhku6CvhsQF+wbKB3dp64fBGj0k8HSc9nc89ApdFiB46fazftb9LdRHpaZ2gmttG6oefmFSc/z9qU/7YYu0I/sN7XhjvRYNC1KWmRZwcQOmNAQAAAAAAACaiODo++Di9vbIk31/baafHFsFyaWK5l2oPcgc8xRdcIV69umjSL/XtYPjlNDvivrXdfJ2yscIUo/J6+0fdxP8F7RvrMvj9Ngf1uiZyfHqZT2vrOdna9QtN2lMypsnXiulvEDLR92k4ROf0rqiYwrq2EfDbp+lZ5eO8w2DmpN1Q074/C8IV+ILb2jHC+PUrbGjJFp11JCkyYr8eJVmP+83ksXLxVeE13lekX36adiwOP3iKv9RQ01Rrr2LR2nwqCl6ZG2RyoM6KnLYGM1ctEhTm/rMqwX3UeIDcbp0W6rSttS/aNKVP/G/XusVFasR1jVXr8d0QleFK7ptuXJ2H7TXMcov6qhfhHeUruij/pdb9b+US0U7lPVFR/Xv07DA2frqSo34W7A956AiB4cpKPcpPXKKa2V1GzJSPT7O0Pb3PPV9WWu196qR6t/TU28U6++Z4T7Vv2fN8ffnRFq1V5fBSXr89it0YMNLyvMZ6QkAAAAAAAA0P4Kj74Or+mnIVVLW5p31T5t0dJMeiLle9zk/5AYFm3EXR/RZoB8pDxQrT6HqEmgAwqkqKtZHAX64/az0oNS2/lDJdUGw/aP7sMmTlRjoNSJcJ4sQmuMYaq4ppII6qsew2zV1/hqtee0NZc7qq/KcVK3bUX/od3TbKqUfCdX9a17TmvlJVp9v15DYcHVpK7mdbWwXXGRfx+HSAMHEh4XWM/XibPvZ0QCfmMoDOlAkRVx+kg+AuSenel8uj9PUcWE6vO5hLd/l3weXLrKuLbh3fN1n5byGhfs9sVMNBYzSbUq3vhe9JqzRppef1FTr+COGXaceoe0V1IzDP4J6366ZccHaOu9hbfWfXdJ5Hlf3Hxfwes3rWu8pIesVpsjBwdq3q0AHCt7W1rb91NPe7wr17Ged+50/aZ+Zvq6m3fDc748CfW6svxEfFVr/6dLRGZ10csOSn9XE+1L1tHWteYunaHnBKdzEK2I1rHe5VmftVMVXO5WVWa5r42N9p1v0Yn/HP7b+njl1b4etz751Abqsg6dugrCA1/rVQR3wG2HWLH87vH1VrvIvAt+PLmHh1v8f07HGTP0IAAAAAAAAnAKCo++FKzRgRB9pVz3/wr/yiLIeXqztQdfpZme9keDefXWtdumPW/y3L1fe1m3SVbHqdYI1TBpvm7bv8AsIyndpy6ZyXTqgn7o5Tf66RQ3WpR9nanudgSlHtXftWm0trvsDcEW57w+zp3KM5ndU+zas0rodvve7fb9f2qOGDh89QR/sUKSbrvR7Hkd37VC+U7a1/Zmu7ROk7avWaq/3ra60numGTHsqtBrWtpG9pa2Zm+qMdqrYsU0brM/UtX1OYZqxRrh0WJJm9ixX+rK12uO0eYQpckCw9mVtqlmrp8bRXdqwdpMOVN+uDh3VxfrP9j2+D7di15va7pRPyjlHl6v8rvfILr1V7JSbRbB6jJulhLZFSntuk9PmcJ5HwPC3eJM9reFh/0Dhi/KAUzx26xWroKKdWm0d69LBtd8tu33HGi3JOeLTXn2/D2/aVHcNpg+36Y/FQRoWaUKNhrFHKJlrnbxI95vnO+c5389jg7TXtTdeJ23J1IZ11kvXaVh0/dGV/R3/YpO2+IeQlQeVt6lYQQP6qIcdcnZXr+iggNdavmOTspxyteb+23E4c4oG33RvgPW9ypX/pvV3N6ijLjvRemcAAAAAAABAMyA4+p4wC/B7/oX/GI2ZuUpbcwu0r6hA+ZtW6ZFRY/TEe8Ea8fgURVb/8/j2gzV6RHvlLZuiJzYU6egX5So/Uqytc8frgW3WthMG1/uv+0/J5aE6/Ny9Wr6t2HOug7u0fGqK1n0VpsRhJ5gPrGe8JvYpt7YdX7uv1c+8xUm6+/lMffAPr4miOnZTj7ZS1trnrOvfpX32miaWxhzjNPrknQwtfSxV6bsPqtzuQ5Gy5izSBoXpl1H1hzTtu3W3nsUmpT9foMNmvy+OaN+Gh3XvuiN+z6i9rp08TteWZ+qu4eO1YPFiLV88T4+MGa509VWEs5VHe/W/PV5dip7TjJRM7TviOe6BTfM0fs6bujRusoY0fumsRuqo/g/fq/5mWkSnpVq3EVPU/4sMjR+/WHnFR+z7dbT4TS2/P0ULMot1rPqRtQpX5IAgHc64V3cv26R867lvXTVb458vVENmdrN18HxuNqSv0t6D5j5Y5yrK1BNT1+qT5v4R30zx93C8Lq3wv2LP87h0V6rGP5RR04/DuzP0RNI8Lc85Iu850a4M6yN9nKl1695UfsFB3/sX3lfDggq0dZd0bS+v71b3cA2pKFb+h8EaEuX7nbPv91ebNGPqYuXXnHuVHpi4Sh/0HKcRUafwHTFTEqY+rhHK1N1JmSeejjGAoKjB9nUsTS9QUPQv1eNEa6zZ3/EKrXvI+huT63y/DhZo3Uzre/BxmKYmVK+NFKSI28cp0udaPZ/7KauO1Q3Lm/lvx6VDJmtEhyI9kThe6ZuK7O/z0eJdypp7r2ZsqlDkuDGKaMpacgAAAAAAAEADEBx9bwSrxz3rlfl4nP7trxl6ZPYMjZk4Q3fPzdCeDrF66NmVmtjbd1KlbuNWas3kML3/3BTF3XSTBv96vNJy2ytxft1tm6xtrKY+NVgH5o33nGv0bK0rv0YPLV+k/icc2NJe1z6+Rk/HSRsec/a1+nnfJml02rN+/QzTiEfN2iiZ1vXP1pxN1fOBNeYYp4vVhwcW6f6oo1o9Y4wG232YogV/6aiJ81M15ET3oOdIPTn5Gv09Y4aGm/1uGqUpW7polnWtVzqb1DDrKL3wrB4acrE+2rVLebuO6OIRi/T0Hd3rTgcYNk4rn52snvue05hfe447auEO/duIJ/Xs5PCmr93SEO2v08xZ19U9l9X+0JrHNbrVJt2XOMq+X3GJqdrQKk5PL52sHjWPLEgRkxdpap8g7c2Yp7ut5z5/1xWalDpOPZwtTqqV9bl5YrIGlGbortHmPljnuneTuiSlakSzjrpzWPfdrGdTR83zWFXTj+EzVun97pO1MjVOl3pNCxgcO1mPRQdpy3OpunvaGu3xHo3Uqrsi+plCX0V0t1s8LviZIvpY/w2KVYR/Vut1v++uOXeGPouarjULfM/dKMHhSrzP2v+9xZqxuKBOQHhCrcI1IM486GCNHlId/NTH6zs+2/l+jZ6h1Z/2tf/GDPF+jh2t78gi63mXV1+r9blfdUyjn0rSL+sEhc38t8O6HxOXPq6J4RV6Ye4U+/sclzhbC3KDNPrxlZo77PSO8gMAAAAAAACMc6osThnfJ2YtjQopKDhYQSf70beyQuVm41ZBCg5u/rhg78Lrdde+ycpcGicz2VTFF+WqOJVzNaKfZrq6oEDbnOZrbZAK69mYH/ob24em9P1Ipu779WIFz/qDHhpQ98fuinLrmVQ28PPybau+XxcE64SXbe5PpXVvmvBY7c+m9d8zfh8a+v1thmuuozF/O85G1d+TIOvzcpLRO57PvXX/2jbgBjb3345G9BMAAAAAAABoTow4+r4yP7K3beAPv+aHUGvbbytICTrVczWinwFDI+NbvtaAzA/Fp9KHBvS9/P0MPbF4l8r9pgUrLyzUHgXryk6BR0iYkMAc+6wMCqrv18lul7k/TXys9mfzbLgPDf3+NsM119GYvx1no+rvSQPCGM/nvoE3sLn/djSinwAAAAAAAEBzIjgCvkeCjh3RextmK35iqjZs2qX8bZla99h4xT9m1i1K1ogTLCcFAAAAAAAAAPjuY6o6nHGHtyxW1tG++vWIcDVyRRCciiO7tG7xKm05UO6pXxKmIb8apQHRVyi4pY4iAQAAAAAAAAA0C4IjAAAAAAAAAAAA2JiqDgAAAAAAAAAAADaCIwAAAAAAAAAAANgIjgAAAAAAAAAAAGAjOAIAAAAAAAAAAICN4AgAAAAAAAAAAAA2giMAAAAAAAAAAADYCI4AAAAAAAAAAABgIzgCAAAAAAAAAACAjeAIAAAAAAAAAAAANoIjAAAAAAAAAAAA2AiOAAAAAAAAAAAAYCM4AgAAAAAAAAAAgI3gCAAAAAAAAAAAADaCIwAAAAAAAAAAANgIjgAAAAAAAAAAAGAjOAIAAAAAAAAAAICN4AgAAAAAAAAAAAA2giMAAAAAAAAAAADYCI4AAAAAAAAAAABgIzgCAAAAAAAAAACAjeAIAAAAAAAAAAAANoIjAAAAAAAAAAAA2AiOAAAAAAAAAAAAYCM4AgAAAAAAAAAAgI3gCAAAAAAAAAAAADaCIwAAAAAAAAAAANgIjgAAAAAAAAAAAGAjOAIAAAAAAAAAAICN4AgAAAAAAAAAAAA2giMAAAAAAAAAAADYCI4AAAAAAAAAAABgIzgCAAAAAAAAAACAjeAIAAAAAAAAAAAANoIjAAAAAAAAAAAA2AiOAAAAAAAAAAAAYCM4AgAAAAAAAAAAgI3gCAAAAAAAAAAAADaCIwAAAAAAAAAAANgIjgAAAAAAAAAAAGAjOAIAAAAAAAAAAICN4AgAAAAAAAAAAAA2giMAAAAAAAAAAADYCI4AAAAAAAAAAABgIzgCAAAAAAAAAACAjeAIAAAAAAAAAAAANoIjAAAAAAAAAAAA2AiOAAAAAAAAAAAAYCM4AgAAAAAAAAAAgI3gCAAAAAAAAAAAADaCIwAAAAAAAAAAANgIjgAAAAAAAAAAAGAjOAIAAAAAAAAAAICN4AgAAAAAAAAAAAA2giMAAAAAAAAAAADYCI4AAAAAAAAAAABgIzgCAAAAAAAAAACAjeAIAAAAAAAAAAAANoIjAAAAAAAAAAAA2AiOAAAAAAAAAAAAYCM4AgAAAAAAAAAAgO2cKotTxneWW4cK8nXoS1Nurc4R4ercxn4jIHdJgfZ89LVdbn1lhCI6ueyyVKbinYX61C63VbfIMHVoZVcCKvsgT3s9G+uS7pEKbecpN4m7RIVbcrS5uNRpCFGnn0drcGRXhZygLwAAAAAAAAAA4OQIjr4XSvXimCjNzPHUOo9J19YHIlUdB/krfSlR19zn2Thmbq5W3trBLksFSv3RcC13ajGp2Vo5opNTqys/rZtu+a2nnLh+n5J6e8qnxq3il1J0z8MZKjzuNHlrF6mk55coMTzEaQAAAAAAAAAAAI31vZqqrrziSxUfPaCsD7KV+dc3al6mbtrN+98Hh1YmKjXH7dROXXbSNKXvdyqn2aGXJmnIfV6hUZsO6h7WoTb8+jxPqbcMV+rupl8XAAAAAAAAAADfV9/pEUf/rPyndhzcrVf+d6t2HHhXn39d5rxTv3atQ9Svy89187/3V78reuv8Vuc777RkviOObG2G6pkd8zUowPRxDR1xZOs6VVmbJ6h7gGnimm3E0edbdE+/icq0Q6OuSlj+O6XEOn2qLFX2Y7/RmNVOghWerLdeHqXOnhoAAAAAAAAAAGiE7+SIoyPlf9f9W59Uz2VDdPsfZurV97c1KDQyzHZme7Of2d8cxxzvO+f4Rt1z30ZVrxR0yvYv0D2LCnQ6x/mU5W12QiPJNTpZSdWhkdGqg2IemK/p1TPmFWxX7hGn3FiVbpV9XuZ5lTXwirz3sV7uSqcdAAAAAAAAAIAW6DsVHB1zf6mZbzypqOfjtW5vlr765mvnnVNj9jfHMcczxzXH/y5xb5ul+19pcnSk4sWztKTg9EVHbrVW955h9uumn4fWXZupVQd1CHXK+kLuxnalrEgvPjhcP+3WQz/t1cvz+lkP/SQqUakb9wcOxQLtY71+0q2XbnkwQ8Xe6zB9sEJDftRNV5lXjzTlBwqXjmRoTPU2P0pSdqB1nAAAAAAAAAAAOM2+E8FRlfW/zcVvqe/KEXrhvSxVVv3Lead5mOOZ45rjm/OY87VkCY8mK8YuuZX9YJJeLLErjRSppEfHOlPC7deS+xYo/zSFHR1uSFVWZqb9SrvBa7RRjVKVFjtFddIlAabfq1fJRt0VE6eZawvkPybNfSRHy+8epCFpfiOqjhco9ZbA+8hqyV+bpCHDV6iwOiC6eqCGhzvl42u0ucApeynduUXZTtk1eqii2jgVAAAAAAAAAAC+RS0+ODLrGD2Ss0QTspL1j3qmowtqdb4iO4drzvX3atvodL0/eYsOTN1e8zJ1027eN9uZ7QMxxzfnMecz522pXD8epZS50Z6RO8dzlJySoUONnmLNpZDYqVp4Z1dPdf8K3X+ap6yrjzsnQ0uqw68bYhQV4pRPprJIS26fps2fe6qu3hO0MDNbf96TrZcfjVd3J7wp/u1vlLqz9soKV0/TcmdJJVdssrJ27dWHf9unv+7KVEqsZzyUuyhND62r7lQnxQyrTo7c2pxX5JSrlSk3p3oBKpdui42oO6oKAAAAAAAAAIBvQYsOjr6sOK67Nz+mlQUvBRxldNlFP9RTA2bqvQl/1LpbF+q/eg5VaPsuan2e78/ypm7azftmO7O92c/s78+cx5zPnNecv6XqfHNybcixLUnJGacy7MiliClzlOhkR8W/naZ5ed9ydFSSobsmrXECq3ClTB2qhuZGZa8t1TwnAFJ4srJemKq4np0U0q6TIkam6uUlo5wAx630jO3O6KJS7X239l4ljBul7h08W7k6hClh7u+UMnmCJj06X9P/w1UTpHWOjXdGeUmHMnJU6JRtZW8r+zWn3GaU4iKJjQAAAAAAAAAAZ0aLDY7K3OV2ePPHD6pHatT6QesQO/jZOXad4rsPrhMUnYzZ3uxn9jfHMcfzZ85rzm/60SK16qTbHp2jGGdUTXbSNKVXhyiN0SZc0+dOkGeJoRItT0z79tbnMdPMxVWvB+RSzNz5SnBCrJMrU+4bW5yyNOj/3ajQVk7F4eo3QS+np2utef2/MCdECtEll9sFW/Yfc3TIOytrF66EqVM1feRQRfXsUDtyqONAxd3glEsylO016Khs52ZlOuXO425UhF8/AAAAAAAAAAD4trTI4MhME/dwziJt3f8/Tkut6Cuv0ZbfrLKDn1bnNO3yzP7mOOZ45rj+zPlNP1rstHUdh2rh3KFOuFGg5PHLatflaQRX+FQtnOEkNsfX6K5ZWwKs/dPM7NCoepo5lyJmrNczt3ay32qYEhXXrDUUqejwAOOUWnVQ976RijKv8E7OfXIpalh1UCYVr07UL/+9m37aP053PbhCL24r0qGAwVmIYgYOcMol2vxudUrn1p68jU65k4ZHhzllAAAAAAAAAAC+fS0yOHpuz4t65X+3OjWPc885V6N/NkzP35iqHwb/m9PaPMzxzHHN8c15vJl+mP60VCE3zNEzI51xMfsX6J5FBfrCU2uU7ncuVYqzjI974zTd80qpp3I6BAiNfn9X9YighnKrrGbGOWvP85xiA5ig7PcrxyqindNgKdtfpM1r0zQzMU6/7NFDQ5K21Fk3KqTvIA1yyoWv5umQKbjzlLPWbpI6xSump1MGAAAAAAAAAOAMaHHB0Z5P9ur5PS/qX15rGpkwZ0qfBCVHT9b5rc53WpuXOa45vjmPd3hk+mH6Y/rVMrkUM2u5Ep3BOsWLZ+n+9cWeSmO06qqEp1OddXzcyn5woh5/1640r5IMjRnU1NDIcCmkZoDSp3J/6RQbqEP0TL38zl69s3W9Fj46QQkDw9TBmfbPXH/huokas6igZo0jW7trFTfUKRdsV+4R67/vva3q2LH72IHq7pQBAAAAAAAAADgTWlRw9NU3X+vpt1fr8699J0K7+d/7a9I1o5o8Nd3JmOOb85jzeTP9Mf0y/WuR2kRq+hNj1dmu7Ff+7pqhOI3TKV4pqdGe8vEC5ddMBddMDpjQqHZNo+6Tf3eKoZHRSaHOCCmpSHlFASbXc5cof2eecs2roMQ3BDJaudSha7jiRk5VyjOZeufPe7XVuv7q/hSvyFGhU/YIUczNo5xyjl7faR3/jTXOccM1PLrBCzQBAAAAAAAAAHBatKjg6I8f5GjHwd1OzSPi0u56OHrKaRtp5M+cx5zPnNeb6ZfpX0vlipyqhXc2PbjoHJ+stNhTi3JOyIRGv/INjV6eGt6g0Mh9IEfpv81QvhnhUyNEUddXrzkkbV60TPl+axO581bovxISNNK8NjijsMoK9OKCNN2TEKchCSt814Rq5VJo9ABFOVUdrxM1Wfd5gBKckUnZr81X+mvONuE3KqaLpwgAAAAAAAAAwJnSYoKjMne5Vv9pg88UdW3Ob60Z/3mHQlzBTkvz+2flP7Vo12q9XrzDaZF9PnNec/5qpl+mf6afLZNLEVPmKLGp2VGrTrotOVkxNdO2NZ17/xqv0MgIU6hytGTBAs0L8Erf7TV66MAa/VdMopLTknTL9YnK9AqPQm6YqOnV17t/hf5reJLSd+5X6ZH9yl1rbT+pejSQSwkDIz0hVUgHuXeuUObOIhXuTNM9szJUeMTaqtKtsiNFenH+CmXb+1h73Rph9dOPK0ID4524K2ejMp3BXTG/HuiM+AIAAAAAAAAA4MxpMcHROx//RUWl+52ax40/vl6RnWvmG2t2JjSal7dSC/L+W9O3zPEJj8x5zfm9mf6ZfrZYbcI1fe6EumFHY3WK1xOP1k7Z1lRlBdu9QiOjQJmLl2lJPa+cj7xG+pSWKN8p6niOCr1n4WsVpkmrUmtCLndRhpITBumaqEEa+WCGCp1zht75OyX1rb6aTp61nJx9il9K0pCoHrqqWw/9NCpOM19yPqPtBuiJKQMU4ql5cSnqV6P87s0AxcV2cMoAAAAAAAAAAJw5LSI4+uZf3yijcJP932rtWodo9M9u1jnW/06H6tDoud3r7NFEx9xf+oRH5rzm/KYf1QL1s6VxhU/VwhlNn7Kuw83ztHBoc0VHTdBzoKaHefrh6p2s4f45Y6d4rczNVNrI8Dohj6tjtBKf3qysmX5T4pl93khX0s1191GbDoq5c76yspcqrpPT5i/8Rk3yfu+GQYpp55QBAAAAAAAAADiDzqmyOOWz1ifHPtWt6yfp47L/c1qkQaG/1NJfJeu8c89zWpqPf2jkLe4n1+vJ/vfLdV6QHRBN/GOKNhe/5bwrXR7yQ700fIkuu+gSpwU+jmRoTFRSzXRuJxattNzluq2jUz1VZhq5Mimk3UmCLHs7Z7RSK5dCQhoQfHnvY3GFhMjVyqkAAAAAAAAAaJk+36/cbTnKO1S9LEYHhV4frUFhnZr2+5+7RIVbcrS5uNRpsI4cOlAxA8LUuSH/Dr/SrUNFW5T9RrGqjxDS+ReKiY1UaAP+cbq7pEibt7+u4k+dhkZeV9n+PG3a9rZKvnQaLuykyNiBiupad+6lgE7XfUWzahHBUc5HuzQ2M6lmJM+555xrh0Y3dIu2683pRKHRr66OtkOjC4NqF/B5bV+OHR5Vb2uCrBVxqYq+so9dh58zERwBAAAAAAAAQENUlir7yYm657cF8lpJvVa7AUp5Yb4Srm5IyuPNreKXUnTPw7VLZPgK0aBHf6cnRoQppL4A5cBGzZw4Sy8WeS3VUSNEUVOXaOGESHUItP/x/XrxsWlKXlfkrOvux1zX7+YoIayeAKg0T0umT9K8nQHvilxhY/XM8zMVU99vuZVlyl8xSben5QW8r66weKXMT9Ztjb6vOB1aRHA0L3elFu1a7dSkSy68WOuHL9JVP6hvLrBT09jQyPjwHyUavn6KPv3yM6dFmtJntKZHjXFq8OEuUf67h/S1Uz2x1ur88/CGJe0AAAAAAAAA0BSVJXpx/CDN3BYwWvHSVYnrM5XUu6E/XLqVnxanW37ru4Z/IKF3rq+7dIZRkqExg5L81oKvK+D+xwuUGjdcy096+nquq4HnVptopW1ertvq/GzfwOuvd39821rEGkcf/uOQU/K4ou1l6tCmeReFOZXQyDD9MP3x5t9feHF1UkTfSEU16EVoBAAAAAAAAODbUfpqipK9Q6OuA5T0dLrWpi9X2hjvMGa/lk9boPyT5UvVilbofu/QpF2kpi/frHf27NFb65N1m7NGu1H822mat9vvwJX7lX63d3DjUvehY5XytNWvmaMU5fVTefFvf6PUnb77F66e5RMahfSdqpVbc/XnPdl6+dF4da/52TvQdZVp85MpvucekayXs/foz7mbtXJqZO068MdzNPOhjJop9Kq5c9L0X37XnzAzVSufTlbi0LDa+2r2v3uN+HX9zDvrg6Pyii91+JjvR+3SizooOOhCp9Z0pxoaGaYfpj/eTH9NvwEAAAAAAAAALUBlkV5ckFM7jVuboVr4+6VKHGr+gXu0bntgvbKSw503LSUrtHxL4GnbfLmVm7FMxU7NHlWTma5JsV3VoV2IOvcepbT1v1NiV+d9lWj5sld9wpeyN1YotcCpyKWYuZuV9fRMJQy1+nVnstZmL1dCzc/XbqVnbK+dDs6dp/VLa0MbV2yqslZNUEzXDgpp10kRI1P18n+PVajzvrmuJRu9zn7gVS3fWJskhd75O72cOkoRXUIU0rGrYianK2tudG34k5Oi5budsm2/0h9fU3tfu47VyzvSlXJnvGKGjlLS05nKesDrvhasUGaRU8YZc9YHR9/8q1JffeM7sVm39lc6paZrSmhUzb8/pr+m3wAAAAAAAACAFqDgVS0pccqWiFnTFOc7XkChI6Z6BTTS5lderzO6pq4SFXsP4bl5pG7yn4qtTbiGjwpzKpacLcqtOXCZcl/LqA1ewmcq5Va/A4REa9J6MzLKecXXpFDW6fdrj9cUc7eNvFGd/dZAcvWO1296OhVL9utv11yX+8P9ynfKUph+E193Gr3OQ0fqNqdsB1c7alIuqeh1/a4mt3IpIXmqIvx+bg8dMV8vV/c9PVVRIV73C2fEWR8cff1NhT7/qiHJrcer72/TY28ttQOhk2mO0CgQ01/TbwAAAAAAAADA2a+4MK82nFGY4n4RYKEdV4Sib3bKRk6R9p404yhTyXtO0biwbZ3gxWh74SVOychRYfUQpcpC5W90ypbuN0aqsylUunXovTzl7vS8Pg2J8FoCpGvt9HGfl6jQKRqu4EBnD5HrYqdobCvUPmdcRNnfvdI0XSJXoInAXL7X5N5dXDPd3KG/vO019Vy8on/u2dJdul/5Tt9zi9wKjazue6QiOgXqI75NLWKNo4YyodF9W+ZqRf5LevStZScMj05XaAQAAAAAAAAAaFnKDnvPjxahzgFyIzNiJrR7pFM29uvTkw456qCu3ru8WxhgDR+39hXmOWWP4o+dAx/2DX6iundVad4yjbymh34Zl6CRCZ7XkF/20E8nrVGx1+giW8euinKKxp73vYMgh7tYRe84Zdt+HXJO3+Fyr9FLylNRcYCk7ECh9jhFW97+mhFLh4q9ris6TD0q9+vFpDj9tM8g3eL0feTwQfrpT+OUmnPSm4lvyVkfHLU+L0jtLqjJR+tVHRqZaeJMEPS7P//BDoYChUenOzQy/TX9BgAAAAAAAACc7Uq17wOnaHOpbT2DXlqf5/1GnvYfcYr16qQefb1SqII0Ja/e7zW6yTr7NtNWz9ClI/uV6xQN944F+q+RC5T7udPgpey1FPWPWyDvmfHUqbsivU6fPydF6R94bVBZquw0q80/cKrWPUJxTtEEXOkpacr2znfKipT+WJrXdHbeSnXob07Rtl/pd8dp5roin+u3HS/S8jExGvMK4dHZ4KwPjs47t5UuOK+1U/PYd/Qjp+Th/qZCb/wt12ctJBMImWBoyTtrVOkVDp2O0Mi/P6a/pt8AAAAAAAAAgO+37iNmKa7mJ2e3slMG6adRcRoSZ73699I1iWt0qHe4IpwtTiR98TIVt4vU9OWb9c6ePXpn63JN6u0VZu1fpsdf8h5VFKaEmUNrp5I7nqPkQdY5h3jO3/+aKI1ZXaII6/wBhQxQ4gyvUUf712hMn17qb/puva6xriN5Wydr/4BDtHzlrNCSbeZ+JOvl7D36855svfxovEKdt+178+Ay5daToeHbc9YHR8FBF+rSi3xXITt8rFTlFV86Ncl1XpDmxE63gx9vJhhatCtdi95ebYdHpyM0Mv0w/fFm+mv6DQAAAAAAAAA427nU9iKn2CguuRoyfqDdAD3x3xPU3eunZ/eRIhW+Z732l0ld47VwarzaOu/Z6j1utNIy0zUptqs6tAtRh67Rmv7Cek33ynbyn3vVZ3q7kBvm6PeTw2rDI7lVWuQ5f/HnUuit8zV9mM/Z1drr/N3vfF4Lb/CeFaxMxabv1qv0uEsRM+ZrUoRXWNWmnuFals5jluvl1FGK6BKikHadFDEyVS8v8Q621mj9Nuue4IxqEWscXfUDe7mvGge/+MT6QPqOxTOBjwl+6guPFuSt1FO5K5p9ejrTD9Mfb/79BQAAAAAAAACcrUJ0yaVO0Va7xo+/T0uLnZIRqU4++9XP1Xuqst5Yr7Q7B6h7R09MEtI1UnEz0/XO5lQNusStT+1Wj9DOzmCKNiHy+bV55Ejd5D+4p1WY4sZ6LaRUUqBin/67FDE1UzvWpypxYJg62D+Dhyi071Alpedq69wBusTtffauqj69rVUnxT2dra3LZyqub1drT0ubDuo+cKwWZuXq5bu6SrXjPKRrOjl99g/kwnTHryO9AiyPkIHxSnDKxub39zslnCktIjjqdVl3nXfueU5N+vvxz/W/Pl9QDxP8pF0/Q/27/qfT4mGCosW7fqdnd7/QrKGR8f/bu/vAqKo7/+Mfi018iGGlxQZFkAraDdA1wE8aFgSMPMQGB7FBXLKwRFMiCIg8iAFFECKoyEMEZVGysLBSUrNGolEQCcImjZWkLZAtEKogmKys6RJSNalpfvfcuZPMTCbkgaBE3i875Zwzd84999yHP+435xzTDtMeD9NO014AAAAAAAAAQOtw/c3eAxJydfCYk/RxSgd+6z2yJlzdwpx0Y4RF6N45q5WZc0Af/emIfr9jo1b8MlLt21g1F+z2GiU0Qr1ucpIdu/pOYXdl2zqBF8N37SXpqyon4aV9n1glvZihDw4csfa/Tzs2Pq+E/iZCdEr7dhe6NzJG9PKaPs7RJlRdo+7Xio1v6/dW2z86kKPMF+fIFR4qVRUq9z+d7SzX/7/ucsedQnX9T7yjXNcoONBEXW18l6qpYKq6b12rCBzd9IMu+lHID5ycOxCU8ced+vpvXzsltUKDQ/T8sKQ6waNAzjVoZPZv2uEdjDLtNO0FAAAAAAAAALQO7XsN9ArQVGjjm7nW//spydY72U7aEhz7MzU8hMCs29NNXX7sfG5bo4P+QZ2qQv1qrVfFdw5WP8/McKERivSOaf32oD5xkrUqdORgrpO2eAe0KrI137Nv63NbileAyKMwTS977X74HT9zjyoyjm5SjNfvH9xWdxq5snf+XRu/cDLqqNH9w5201P3/DfMKdOWqsChAVOjYQe1zksbgmxuxXhLOq1YROLrmynbqec3NTs4t78TvdPjzj52cr8YEj841aGSY/Zt2eDPtNO0FAAAAAAAAALQSne/SuDudtKViwwIty/UKklSd0K8eX6BdTtYESMb9vJdXUKRCRb9eoLEulx5MydWpmuBQsHrfOtRJW04s18OLsmu/rzihXYse0bKa2dmCNW7M0NrAjdor+m6v3xcs0fwNR32CWmW5y62y2hKfgFZwL/XyOq5Plj+i+Ttr57GrOJGt+dOWq2Z+ryvi9M9Da/euG3ppuFcc5+3HZmhjYW2/lO3fpIdnb6ttT8T9cvV00kbEXXqo5vcV2rhgiXZ5T6Nn+nXREuU7WesHGtTLe548fBsuqbY46Qvau3/K0cRtj/uMMhrTI0ZLhszUJdZ/gZRVlCsx8wn913HveGXLBI2qrf/m7HhOWw5kOiXuaerWjnhKd/y4n1MCAAAAAAAAAGgVCtdoSIxXEEXBat9nqEbeXKFdb21Xkdey+8EjVuuDlV4Bng+X6CejX6kJoAxfuU8vjnC+rSrUC8NdXsGh+gVHPa8960Y4U705qo5q4+jhml/g5C2hN1rtuvN6nc59Q29/eKo2cHPFIC15e53u9R60U+e46hOswct2af3dvoGbsrce0a0PeQWH6nWjHnotQzMifKfNq8heoH+I3+TVxvbqNewu9bviE73t169df7lVmXMivAJy+Da0ihFHxq3X/VTh7W90cm5vHHpXuZ943S1+zMijl2IW6h879XZKWiZoZJj9mv17M+0z7QQAAAAAAAAAtDLhk7R+mfdonwqd+nCb1m32CxpZ2/3H097bSWWffeYTWDl4wmstpDbheij1eQ2/2snXI7jPdL220i9oZLS5UeNeWqdxXq/Hy45u18aUV5ThEzSy9rNumW/QyKhzXIEEq9esrXrRL2hkhN75tP5jSngDwZwbde8LL9cJGhnBg+Yoc/6g2t9/cUr5//mKXvDr19A7k7V+FkGjC0GbJy1O+oIWfGmQgtp83x55ZEb7GH/929c6cbpEw7oOsL8PxJSbKev+8D+H9NMf3dwiQSMzkmnW9qU6frrYKZG+d8n3NOsfExTRoXb+RgAAAAAAAABA69H273+usUM66Iv//q1+X+I/xiZU/aav09bnY9XVL7oRfN01qsp+Qx+YOeiuHqqFT/2LfhLifGmE3qyY++5Ql7+VaN9/f6wv/uqUG1dHaNzCtUqdd6eury9qcuUNGjz2Pt165Wf6+OBhFX/llNtC1evu6Xpu7TO69+8DV+A+ri6q+p98HfzoC3kvsxQaEacFL7+shXd21KVOma9L1SHyPt3b+0p99vEfdMinX4LVPmqq1m1eqQkR9YWmLlW7W+7SuHr2Hxw2SPGLV+qlqYN0TRunEN+qVjNVnfHl119p4huPa/exD5wSt3vCh2npHbP0/Tbfd0rq+vpv7kvx0u+d25X316q/6tF3n9Vrhe84JW4DO9+qtXc9pcsvvcwpAQAAAAAAAAC0VhVlJ1T0h0902kpf1qGrut/QXsFne71cVaGysgoFh4aefTtLxZ/L3COF2gQrNLTpY2wqyqzfO9GXxuzPh9NO2xWhavLuK8pU9oU72eR9G977D7b2f27jPHAetKrAkbHv0wO6P+Mx/fmr2gW4zGifqX3HaerPxquNlT5fqqr/plW/2aBVeRv1NyvtcfVloXrF9bR6X9vDKQEAAAAAAAAAAGh9Ws0aRx4mOPNA73vtYJGHCeKYYM6C7BR7RND5YOo19fsHjUw7THsIGgEAAAAAAAAAgNau1QWOjIm979Xdfz/EybmZYM6G36XrgTeS9D/l/+uUtgxTn6nX1O8dNDJMO0x7AAAAAAAAAAAAWrtWN1WdR1lFuR55J1k7jv6XU1Lr7y4L1bzbJmlU+LBzmrrOTE2XXviOFr2/Rv/nNTWex5Ab/1HPD0tSaLD3KmcAAAAAAAAAAACtU6sNHBl/qfxCs3Ys1ZuHs50SX9de9SM9EjlBI26+XZdd2vgVvr76ukLbDr2n53NT9emZ/3FKff38pkF6dsijujKoda7cVXm6XJUm0SZIISFBdhkA4MJQWW49o80Cl0EhCrncXQYAAAAAAAB8E1rlVHUeJmizcvg8xUf8IuDIIhP0mbl9iXqu+bnG/Pph/cf+bSoqPWYHhryZvCk335vtzPbmd4GCRmY/Zn9mv60uaFRVriPpyXrwzjsUNXKkos1nxJ2Kum+u0gvKnY0ubAdW3KEBkzNU6uSbpKpS5SZgZl7GfluqSpSfkqiowedwHN8pBVpt+mLkMuXXdwmWZGi2tc3s9BKn4EJVoszJ1rGsKHDy9XCOZ8DkNBXXdy1+mKIB1jarP3TyTfVlucrL7dDwhcM57mYdk6fPhiUr/0unrB6Ve5LP0/3lXKsNnd+AGnltGKf2avV467k8wnlG33mnXFNTdaR1PKIBAAAAAADwHdCqA0fG99t8X08MekhrYhbYU9QFUln1V+V+UqDH3n1OURvG6eaUoeq8fGDNx+RNufnebGe2D8TUb/Zj9mf226pUlWj3gnjFp+xVUNQUrVy3SVmvb9KmxVM0qm2Blj8yWvMyLvQX8+foVJYWjhypdc1579sCyg9laOl9cZqWVa7ObZ1CuJ3O0qwVe92j4C4WhWu1cPNxJ9OyDqwdqehHs757gcnK95SefbajKtXuN95rvdfRlwVaN/lJpQdFa+WG161n9OvKWD1R/Uo2KzEpo/5AIwAAAAAAANCCWn3gyLjE+m9419u0N36L7usZc07rGgVi6jP1mvrNfsz+WpvSzGQt3CONWbJVK2e41KtrmELahqlzP5cmv7RVL7pCtHtNinYzBOY8KdCGxLU61CdJGdue06jrnGK4XR6koJ3JWrL94rkAg6xjPpw6UxsLnQI0yu60nSp20nUc36n05o7SugCU7/m1Np6K0JylU9SrU4j1jA5Ru3CXHn1qrDrvT9WOQ86GAAAAAAAAwHn0nQgceVwVfKWW3DFLOQ+kaUyPGF1+6WXON81jfm/qMfWZek39rVOhMlMLpaipSugb4pR5C1GPcRM0sDJPb2b7jjoqP5Sl1Y/EyTXsDg24c7SmLcrQkdPOlx77UxUf96R2l5Qqf80sxdlT4cVreXqR/Zf/lYcytDxxtKIH3ynXfbO0xW9avAMvxyl+QZZKy4uUuSje3lfUyDjNW/Zeo//CvjQnVQvtfdyh6HsStTA1T6Vev7X3MT1V+6x05tNWOi5OS7N8gxQN1XFuOijmpa1aP/t2tWvjFAUQqL+PNWaKKjMNYdYyzbtvpD1N11nbfypPW+a6+9nsY94as12htgToE50u0o5lsxQ/8k4NsM5fXGKyMg+dhzmzhs7Us64Qa19PakdjB76Zaf9S52raPZ62zdWWHL8fn8rSUuu4tux38jXqHq/nOiw+nqXl8e5+9J5Wrc65mZuq/HMYpNd7wnxN71mudU+m6EAD06/V8D4fw0Yq/pFl2uF9PpzjXbLdShelaqaVjo9L1QErW75nmZVeplyf+7dS+SlmmxS/KeBKtGOuVb5sr2pq9+tvcy9v9O9vM+JngfW7lwtUXpCqefa2U5V5ln4qzphr7X+q1uU1HDQcExurDh9t1o4659PtSOZmHWgbq4Rop8CP9z1uP2Psa9/50pv1LDL9bJ5lpp8fbOA+bKlnx5k2nTRuVLR6tHMKPNq20w+tM3HmCycPAAAAAAAAnEffqcCRR1jID7V0yCztn5Sp1JFLNKzrAIUGBwqY1GW2M9ub35nfm3pMfa3aRwXKPh2kUcP7K8gpqqNdtBbtelfPjApzCqTKD1OUmJiifUFRemj+Yq2cEq0fHFyr+NFTlXnS2cioKNeRkyXavWKq1pX31+R5SZocUanMlKlasnat5j2+Vx1cM7R43gT100GtfmSq0r1n6Cov0ZEThXo1aaZ2XO7SHGtfi8ZF6MzOZMXFrW3wpXpx+lTFzk3T593G6onFi/XE2G76PG2uYqfXTu10w6ApemjsAHW20r1dVnrSFP38p7XXRGPqODdh6nzz2a/BygKnv9u6+2DljFhda/V33NTNOnbWNpRox/TRin8mTxowQYuc9hdvsdo/973aF//GyQzNjpurdX/qoPseNvsYqw5/WKCpKTt1zDqHn3/pNclXeYFWJyRqyYdBGjpxvlYunqKYvyu0g4Atv97Q1eox8TGNa1uoJcmN6POqEmVOj9e0tM9109iZVttmaky3z7Vhbrxv26oq9bk5Lt9lzSwV7nKf47WuwyNZWj5zrYp/HKs5syeoxw/cX5VmTJUrcZlyNUAJs939dtNnaZo2fm7zR+m16ayYpImKPJ1h3ScFDU+v5n8+5k/U0KACLUmM1+oPnbMcEqGfW9d2THcr3X6A7rPSD00aoBvMV9266ocns/SbAq89VR1UbqZ13FZ57kGnzDhdqNycEt3Q8xbZV22A/k7oc0avzo1T3ArvtlfqzGdWfQWpWjh3j4KiJuqJKVG64XLnaz/F25/UrBVHdPO4JCX09Y+WBHBrtMZ0KdeGzADTGn65V5kZ5eoxJlq9Ajzo/O9x+xmTOVeu+1J0wPsmMf0cn6iFO88o8oEku58HlqcpMSlVx5xNvLXks6OD1V8JU25XByfvUXnkoPapkzpc4xQAAAAAAAAA51P1ReRMRXn1kc8/rt526L3q1/97R83H5E25+f476berqvsPmlK9rdjJN8YXe6qfHxpVHbv0N9UVTpHt64PVG2Kjqm9/dGf1GafIXX9U9diXDjoFxpnqnMejrPIJ1a/9ySkyzuysXmJtO+u12sbsX262s8q2+jXwxOvVs6w2JL56zClwtp30evXnTr762NbqxEHRPvXZTmytnmXVOfetmi2rq4ut+qyyF37r5D2aUkcgX1dUnznj00sNKK7eNsnvOCz7llplD2yt/tTJ2z7fU/3q6terD/+fkw/k/w5Wb38uqXrDb2vOiK3i3fn2ed/+mVNgnZPsx6Or+8euqt7vs6l1rp6yyn3OS0X1vmSrzPVc9T7faqsPr46t7j80qTq7gW6prjhj3VNOul751S9Y++2/PN+dPfhS9Vj/c+GcN++yT1+dYLXBuqZPOAWOT7dOsY5jZvV2T9vqO+fOfr3rdF+Hsda2fgds9c/h15+rXrA63/de8FzLr9fszH1ePcdSH7/jOfNOUvXt/vt17qnadjvnI8C527fUZZ2nl6oPf+0UWercJzZ3+273bp/V37HWNTLLKo/1un8rdi326Ue7v80zxK+/P3/L3fYNNT91+iDAufE/F5++M98617HVS97xOtf18frtmbdm+p5jh3d54OeE1ef+z5jP3qpeYD1jvJ9d9V1bH28wfeB3fhv97GjktRHIGeta9X/mAgAAAAAAAOfRd3LEUX1Cgq5U13adFXPTYLl+ckfNx+RNufn+u6i0pOkL8Ffmva/0ynCNG9fXd5RSm3CNeqCv9f1O7fOZ8ipMo4aGO2kjRFfZAwgidFMXu8AtJFzdrc1yj/uvUnK77o6pHe1ku86lcdFBOpC5J+Bf+hvHsjN04Lqxmuw1Uspm/fbuKGn37t/5jrgJ4NzqqFT+MyMVPWKkVuc0OGbkrMyaNzp1XMXe/dquv8ZMcqlbWycfSNtwDZmxWOP6+I5oCurYWd1UqMOezjv9gXbvqdTACWPVw2fTEEWOcvmOcvjS2nZ7pSLHjVUvv4FS3cYmakhlnnZ7RrkE8uVeLbf6JHrkMr8p0BoQPlGLJ4QpNyXZd1Sbj+PKzTxutWOmYvzWiuoQM1qjVKDcs7XtbKxzHuPXj6Z/urlm6IlJEb73wuWd1Nnaf+5H9V2djRMy9DEtiirXlkVrlV9fs53zMWrSxDrnrldsrLqdzlJug+vfhKln3zBVZn+gI85ImGMf7lFx35FKiApX8Z6Cmvvs8O/2Wueiv3rb97C7v4NcE+r0d7vosRrftlSZe/wWahowWkPPso5XaXayZj39gW54eJUeHep33zUgZJBLo4IKlL7d+7l2XNvTChQ01KWBAQYu2fd4kPU88b/H20frvlEhOpa1R0fsAudYo+sea2dXnIY4aY+WeP6clRnplTRXW+TSynm3u0d/AQAAAAAAAOfZRRU4uli1C+vkpBqvvNRM99VNNwR4pxvS/Rb1UoEO/8kpsLVTUD3TUTVKV2tfAX5/baeu0slSnXHy/s6Ydp5K0wJ7LRfvT7xW55gNzjQ4Bdi51hEUYl7nhuiqvwswP1YT9Bg1UQMrszRt5J2KeyRZW9LzdKy0kcGoqkoVF76nzJdTtHpuotX20YpO3Oy8DHd8ecZ+id2hfYA36126K9JJ2k5/LhPa69wpwAXQNly9rNOy+5BP7X6uVIgJdll9c1UTu6Xz2GQ9Gl6opY/XN0Wf1baT0rH0JL/zZX0SVsks71Ne3szX9W2t9jpJf5UnC5Wbnqp1y+ZqmrWvuBGJWl1vcKspQhT58HyNUZZmrQgwBZvhnI/tKfF1jzkpTcesM9uY9W86D4j2CjKVaH9eiSJvDVe3PgPU42Se9tuz/BUqP7tSPQZEWHe14e7v3p38J1AzwtWjn1R88Ih8ZuwLubLeaTGLdyRr6oL3dKzfTM1xNS1oZLu8v2JcIb4B5eN52vFRiMbHBJ6O077Hu3bStQHWF+t2ywCrfw/q41Mmd5ZjbdtVPfyCSS3x/KlfuQ6kTNXSQ1316PIpfgFDAAAAAAAA4PwhcHQx+EE7e+TJkT82/xWmjzZBClalKpu4fsdZBZk6m+nyTurZt68i/T4Do10aN6BT/es6eWt2HUHqMWWr9ry7VeO8B1w1x3UuLXp9k16cEqve1vnKfHmu4u4ZqfgF7519rZTyAq2LG6nRk5/TlsIzCgrrq1ETHtNLqyf6BoNaktUhZz3/l0co4dV3tefVieoW4GX9WbUJU0zSFEWeTNXcl/1Gsnj5QaeIOucrsu8AjRrl0s+6tORbdvMCP07RcVO1cHOhyoPCFDkqXnNWrdL0cz3nHiF9lTDPpQ47k7Vke/2LJt3wE//jtT79ojTGOmbPekxn1SVCg9qWK/vD4/Y6RvmFYfpZRJjUqa+GXGfl/1AuFe5R5ukwDenbuICzdetKTXgW7M4+rsjocAXlPKeFzVwrq1vMWPU4mabd+935I5mbdaDLWA3p6c43ifU8Myqa+zxriedPHeU6snampmVIYxYn1xn9BAAAAAAAAJxPBI4uBl0GKKaLlPn2XnvESUClWZo3+A7Ndl7kBoWYcRcl+txnOjrHsSLlqqs6BxqA0FyFRfo4wIvbz08dl9rWH1QKvjzEfuk+asoUJQT6jIlocHqnlqhDTQ2O1CcoTD1GTdD05zdp01vvKuOx/irPTtaWPfUH/Up3pmpjSVc9uuktbXo+yWrzBMVERahzW6nC2cZ2+VX2cRSfChCY+OigdU69ONt+Xhrgiqk6pmOFUq/rGrgATJ80t1+uc2n6xHAVb3lS6/L82xCsq6xjC+kTW/dcOZ9REX5nrLlBAePUTm207ovekzYp67VnNd2qf8yo29WjazsFtVAs1gjqM0FzXCHasexJ7fCfXdI5HzcNmRjweM1noPeUkPUKV2R0iI7kFehYwW+0o+0A9bR/10k9B1j7/uB3OmKmr6spN9z9/XGg68Z6Rnx80Pqnc5gzOqlho+a/pMmzk7XSOtbclKlaV9CMTuwUpVF9yrUhc68qv9yrzIxyDYyN8p1u0Yt9j5+0nmdO3luxde1bB6Br27vzJhAW8Fi/PK5jfiPMWuTZEUBxepISt5Rq1JL1mlxn6kQAAAAAAADg/CJwdFHopKFj+kp59fyFv1lH48kU7Q66XXc7642E9OmvgcrTm9v9ty9X7o6dUpco9W7Rv4Lfqd17/AIE5XnanlWuDkMHqJtT5K9bv2h1OJmh3XUGppTqwObN2lFU9wVwZbnvi+rm1NHySnUkPVVb9vj2d7sBt9mjhopLz9IGOyjSTTf4nY/SvD3Kd9K2trdoYN8g7U7drAPeXV1lndP0DHsqtBrWtpF9pB0ZWXVGO1Xu2al065oa2LcZ04w1QYdRSZrTs1wb12zWPqfMLVyRQ0N0JDOrZq2eGqV5St+cpWOe7mofps7WP7v3+Z7cyrz3tNtJN8jZR+cufsdbkqf3i5x0iwhRj4mPaVzbQi1Zm+WUOZzzETD4W5RlT2tY7L+W1OnygFM8dusdpaDCvdpg1dUhuvbessv3bNIL2SU+5Z7+Ls7KqrsG00c79WZRkEZFRjgFDbNHKJljnbJKj5rz+/Ra3+uxUdpp4F23S9szlL7F+uh2jRpUf+jKvsdPZ2m7fxCy6rhys4oUNLSvethBzu7qPSgo4LGW78lSppP2OB/PDnv9p5QiRT68SpP7EjQCAAAAAADAN4/A0UXCLMDv/gv/eMXPSdWOnAIdKSxQflaqFsbFa+n+EI1ZPFWRnveU7aI1fkw75a6ZqqXphSo9Xa7ykiLteCZR83Za206Krvev+5vluq4qXjtT63YWufd1PE/rpi/Qli/DlTDqLPOB9YzV5L7l1raJtb+12pmbkqRpL2fo8P95TRQV1k092kqZm9dax5+nI/aaJpam1HEeffpBmlYvStbGD4+r3G5DoTKfXqV0heu2fvUHadp1626diyxtfLlAxeZ3p0t0JP1JzdxS4neO2mnglIkaWJ6hB0cnanlKitalLNPC+NHaqP7q5Wzl1k5DJsSqc+FazVqQoSMl7nqPZS1T4tPvqYNrimKavnRWE4VpyJMzNcRMi+iUeHQbM1VDTqcpMTFFuUUldn+VFr2ndY8u0PKMIp3xnLI2EYocGqTitJmatiZL+dZ535E6V4kvH1RjZnaztXdfN+kbU3XguOkHa1+FGVo6fbM+tcpblJni78lYdaj0P2L3+eiQl6zEJ9Jq2lH8YZqWJi3TuuwSec+JdkN4X+lkhrZseU/5Bcd9+y+iv0YFFWhHnjSwt9e91T1CMZVFyv8oRDH9fO85u7+/zNKs6SnKr9l3quZNTtXhnhM1pl8z7hEzJWHyYo1RhqYlZZx9OsYAgvpF28exemOBggbdph5nW2PNvscrteUJ6xmT49xfxwu0ZY51H5wM1/RxnrWRgtRrwkRF+hyr+7qfmnqmbrC8hZ8dlR+m2Os/nek3Vj9vf8y+Xn0+30gQGwAAAAAAABc7AkcXjRD1eHirMha79MM/pmnh3FmKnzxL055J0772UXripbpTInWbuF6bpoTr0Nqpco0cqej7ErUkp50Snj8P0ye1jdL056J1bFmie1/j52pL+a16Yt0qDTnrwJZ2Grh4k1a6pPRFzm+tds7OksYvecmvneEa85RZGyXDOv65ejrLMx9YU+o4X6w2zFulR/uVasOseEXbbZiq5X8I0+TnkxVztj7oOVbPTrlV/5s2S6PN70bGaer2znrMOtYbnE1qmHWUXn1JT8T8QB/n5Sk3r0Q/GLNKKx/oXnc6wPCJWv/SFPU8slbx97nrjVuxRz8c86xemhLRzLVbmqjd7Zrz2O1192WVP7Fpsca3ydLshDi7v1wJyUpv49LK1VPUo+aUBanXlFWa3jdIB9KWaZp13p/P66SHkieqh7NFg9pY183SKRp6Kk0Pjjf9YO1rZpY6JyVrzPlYe8bq98UTAkTlas5Hak07Rs9K1aHuU7Q+2aUOXtMChkRN0aJBQdq+NlnTHtmkfd6jkdp0V68BJtFfvbrbJW6X36Jefa1/g6LUyz9W69Xf02r2nabP+83QpuW++26SkAglzLZ+vz9Fs1IK6gQIz6pNhIa6zIkO0fgYT+CnPl73+Fzn/ho/Sxs+628/Y3zWEAqz7pFV1vku9xyrdd2nntH455J0W51AYcs+Ow7vzdAx69/SnFTNtq5Vc736fDLNtwAAAAAAAMD5dUm1xUnjYvJlucyMbUEhIQpq6KVvVaXKzcZtghQS0vLhggMr7tCDR6YoY7VLZrKpytPlqmzOvprQTjNdXVCgbc7zsTZKpXVuzIv+prbhXNpekqHZ96Uo5LHX9cTQui+7K8utc1LVyOvlm+bpr8tDdNbDNv1TZfXNOZxW+9q0/v3W+6Gx928LHHMdTXl2XIg890mQdb2cbZSSxX3dW/3XthEdeCE8OwAAAAAAAIAWwIiji5V5yd62kS9+zYtQa9tv6mVoUHP31YR2BgwaGd/wsQZkXmg3pw2NaHv5oTQtTclTud+0YOUHD2qfQnRDx8AjJEyQwNR9QQYKPP3VUHeZ/jnH02pfmxdCPzT2/m2BY66jKc+OC5HnPmkgaGS4r/tGduCF8OwAAAAAAAAAWgCBI+AiEnSmRPvT5yp2crLSs/KUvzNDWxYlKnaRWbdovsacZTkpAAAAAAAAAMB3H1PV4VtXvD1FmaX9dd+YCDVtRRA0S0metqSkavuxcnf+mnDF/DxOQwd1UkhrHUUCAAAAAAAAAGgRBI4AAAAAAAAAAABgY6o6AAAAAAAAAAAA2AgcAQAAAAAAAAAAwEbgCAAAAAAAAAAAADYCRwAAAAAAAAAAALAROAIAAAAAAAAAAICNwBEAAAAAAAAAAABsBI4AAAAAAAAAAABgI3AEAAAAAAAAAAAAG4EjAAAAAAAAAAAA2AgcAQAAAAAAAAAAwEbgCAAAAAAAAAAAADYCRwAAAAAAAAAAALAROAIAAAAAAAAAAICNwBEAAAAAAAAAAABsBI4AAAAAAAAAAABgI3AEAAAAAAAAAAAAG4EjAAAAAAAAAAAA2AgcAQAAAAAAAAAAwEbgCAAAAAAAAAAAADYCRwAAAAAAAAAAALAROAIAAAAAAAAAAICNwBEAAAAAAAAAAABsBI4AAAAAAAAAAABgI3AEAAAAAAAAAAAAG4EjAAAAAAAAAAAA2AgcAQAAAAAAAAAAwEbgCAAAAAAAAAAAADYCRwAAAAAAAAAAALAROAIAAAAAAAAAAICNwBEAAAAAAAAAAABsBI4AAAAAAAAAAABgI3AEAAAAAAAAAAAAG4EjAAAAAAAAAAAA2AgcAQAAAAAAAAAAwEbgCAAAAAAAAAAAADYCRwAAAAAAAAAAALAROAIAAAAAAAAAAICNwBEAAAAAAAAAAABsBI4AAAAAAAAAAABgI3AEAAAAAAAAAAAAG4EjAAAAAAAAAAAA2AgcAQAAAAAAAAAAwEbgCAAAAAAAAAAAADYCRwAAAAAAAAAAALAROAIAAAAAAAAAAICNwBEAAAAAAAAAAABsBI4AAAAAAAAAAABgI3AEAAAAAAAAAAAAG4EjAAAAAAAAAAAA2AgcAQAAAAAAAAAAwEbgCAAAAAAAAAAAADYCRwAAAAAAAAAAALAROAIAAAAAAAAAAICNwBEAAAAAAAAAAABsBI4AAAAAAAAAAABgI3AEAAAAAAAAAAAAG4EjAAAAAAAAAAAA2AgcAQAAAAAAAAAAwEbgCAAAAAAAAAAAADYCRwAAAAAAAAAAALBdUm1x0vjOqtAnBfn65C8mfZmu7xWh66+wvwio4kSB9n38lZ2+7IZe6tUx2E5LZSrae1Cf2em26hYZrvZt7ExAZYdzdcC9sa7pHqmuV7vTzfLFCeXnfyKnVbr+/1nH4GkWAAAAAAAAAABoEQSOLgqn9Kv4fpqT7c5dH79RO+ZFqr64y6lfJ+jW2e6NBz+To/W/aG+npQIl/3i01jm5wcm7tH5MRydXV/6SbrrnX93phK1HlNTHnW6qU3vX6OFpy5XzZ6dAg7QkZ53uDXOyAAAAAAAAAACgRVxUU9WVV/5FRaXHlHl4lzL++G7Nx+RNufn+YvDJ+gQlZ1c4uebblfSINh51MudDWaE2PtRbt47zDhoBAAAAAAAAAIDz5TsdOPpr1V/13ke5mvLWQt3y4gh1X32nojaM0+Q3n9TUrKdqPiZvys33Zjuzvfmd+f13U4U2PvSY3j7nYEyB5ieu0cEqJ9uSTm3Tg/1cmv9WmZ0NDg9XVzsFAAAAAAAAAADOl+9k4Kik/H/16I5n1XNNjCa8PkdvHNqpP3/lDkA0xGxntje/M7839Zj6vnO+2KaHZ2/TKSfbbEeX6+FVBTr38Ut+qr5SxRfWv1eE695nMvRBxpMa7P7m3FRVqOzPZe5PWeNbXeH5jfWpOB+BMgAAAAAAAAAALgDfqcDRmYq/aM67z6rfy7HaciBTX379lfNN85jfm3pMfaZeU/93ScXOx/Tof55z6EhFKY/phYIWDx0puP90bd6VoSW/CFdoG6ewucoK9avHR+sfuvXQP/Tu7f7c0kM/6Zeg5G1HAwe+qk4pJyVBt/bopp94fmN9fmLVcWvCGuV4d93hVxTz427qYj49lig/UHCpJE3xnm1+nKRdJjAGAAAAAAAAAMAF5DsROKq2/nu76H31Xz9Gr+7PVFX135xvWoapz9Rr6jf7MftrzcY9Nd8ZvVOhXY8n6Vcn7EwTRSrpqft1vZ0+qhdmL1d+SwZC2sfqxY2T1K+9kz8XJ7bpwcEuzdlcIP9xZxUl2Vo3bbhilviNmqo6oV8lDtbY5dk6Vee4KnRq53KNHZ5UO93fTcM0OsJJf7FJbxc4aS+n9m7XLicdPH6E+l3hZAAAAAAAAAAAuEC0+sCRWYdoYfYLmpQ5X/9Xz3R0QW2+r8jrI/T0HTO1c/xGHZqyXcem7675mLwpN9+b7cz2gZj6zX7M/lrz+kfBN8dpwTODFGwyX2Rr/oI0fdLk6deCFRo1XSt+eaM7e/QVPdqSU9ad6wgjj6pCvTDhkZoAT3CfSVqRsUu/37dLrz0Vq+5O8KboX/9ZyXtrW1/2zvOas9PJ3xin9bsO6KM/HdFHB3Zp/XjnmP+cpodX5TrH3FGDR3kiRxV6O7fQSXuUKSc720kH696oXu7+BwAAAAAAAADgAtKqA0d/qfxC095epPUFvw44yujaq36k54bO0f5Jb2rLL1bon3qOUNd2nXXZpb6v7E3elJvvzXZme/M783t/Zj9mf2a/Zv+t1fV3z9eCKHc/VOxM0vy05gw7ClavqU8rwYmjFP3rI1qW22KhoxZR9tZqLTvqZCLmK/PV6XL17KjQqzuq19hkvfZCnBPAqdDGtN01I5KK/rDNSUmDJ07S4M7ONXNFRw2et1ovTpmkh+Yk68Wh19QEy66Piq1Zh+mTtGwddNK2st9o11tO+oo4uSIJGwEAAAAAAAAALjytNnBUVlFuB2/ePOwZxVHr7y4LtQM/e+/fotju0XUCRQ0x25vfmd+bekx9/sx+zf5NO1qlNh1171NPa7Az4mZX0iPa6AmwNMUVEZrxzCR1tTMntC5hyQW0dk+Zct7d7qSl4f9yl7r6jWQKHjBJr23cqM3m8y/hNaOAQjuEOykp5513dNAzJZ3R5kYNnz5dM34Zq8GRN6rm6ggbJtedTvpEmnZ5DToq2/u2Mpz09RPvUq+WGlEFAAAAAAAAAEALapWBIzNN3JPZq7Tj6H85JbUG3XCrtv9zqh34aXPJuR2e+b2px9Rn6vVn9m/a0WqnrQsboRXPjHCCJQWan7hGB5s8ZZ0UHDFdK2Y5w46+2KQHH9teZy2hb8cJFdWsNRSpQRF1A4Bq017d+0eqn/lEdKwJHHUd9kBNUK1i5wLF9O6mn/RzaeyMJdq4rUBFpwKNrArV4GFDnfQJvf1bTySuQvtyPSOYOmr0oNqgFAAAAAAAAAAAF5JWGThau+9X+s//3uHk3L53yfc0/pZRevmuZP0o5IdOacsw9Zl6Tf1mP95MO0x7WqvQO5/Wi2OdcMnR5Xp4VYFOu3NN0v2Xq7XAWeKnYtsjevg/T7kz36oKldXMwGcd46VOsjHCRujF9PlyhXlCSVZtJYXK+c9XNH/aaA3p20P/MO4VHfQbXRXaf7iGO+mDb+TqE5OoyFX2ZrtI6hirwT2dNAAAAAAAAAAAF5hWFzja9+kBvbzvV/qb15pGJpgzte84zR80Rd9v832ntGWZek39Zj/ewSPTDtMe067WKViDH1unhI7uXFHKY3p0a5E70xRtbtS4lcnOGj8V2vX4ZC3+rZ35FgUr1Dku6TNV/MVJNlLwTXFakXNAf9yVofXPzFHC3ZHqerXzpaVs7xLdM22bfEJkVw+Ua4STLtitnBLr3/2/kSe02P3+YerupAEAAAAAAAAAuNC0qsDRl19/pZW/2aA/f+U7Edrdfz9ED90ad85T0zXE1G/2Y/bnzbTHtMu0r1W6IlIzlt6v6+3MUeV/WDNMp2k6xmpB8iB3+osC5ddME/dt6aiuzigoqVC5hQEm0Ks4ofy9ucoxn4ITCjQBXXDncA3+xf1KWrZRO/Yd0e833u+s6WT9fOcbyvGJHIVq8N1xTjpb7+y16n93k1NvhEYPcqb0AwAAAAAAAADgAtSqAkdvHs7WnuMfOjm3Xh2668lBU8/bSCN/Zj9mf2a/3ky7TPtaq+DI6Vrxy3MPalwfO19Lomqnd/umVBzL1sZ/TVO+GeFTI1T97vCsOSS9vWqN8v2mlqvIfUX/NG6cxppPumek1QntSlmu+Q+6FONaoLf/7BQ7QvsPc0ZWuX3lty5UcORQjXPWR9r11vPa+JYTjoq4S4M7u5MAAAAAAAAAAFyIWk3gqKyiXBt+l+4zRd0V379Ms/7xAYUGhzglLe+vVX/VqrwNeqdoj1Mie39mv2b/HqZdpn2mna1TsHpNfVoJ5xo7atNR986fr8FO4OQbcWyT/mlwguYvSdI9dyQowyt4FHrnZM3wHNPRV/RPo5O0ce9RnSo5qpzN1vYPeUYDBWvcsEizEpKlvYJLX9HGdwp1cP8mPTxxuXKOlamiSqr48wnlLH9aG+3tLDf+TL3DnLRHcC8Ni3WCZ9nblOEM4Bp83zBnVBcAAAAAAAAAABemVhM4+uDkH1R46qiTc7vr5jsUeX3NXGQtzgSNluWu1/Lcf9OM7U/7BI/Mfs3+vZn2mXa2WldEaMYzk2qmYWu2jrFa+tQgJwjzDTh1QvlOUl9k66D3THttwvVQanJNIKuiME3zxw3Xrf2Ga+zjaTrojEDq+st/V1J/T4uD1W/Wv9cE0So+XKOxg3vrJ9266Se9B2tsSoE72HSFVffTcQH6y/r9z+P8jn+oXFHtnTQAAAAAAAAAABemVhE4+vpvXyvtYJb9r8fVl4Vq/C136xLrv/PBEzRa++EWezTRmYq/+ASPzH7N/k07PAK1s7UJjpiuFbPOfcq69ncv04oR31DoqOcwzQh37yu4z3yN9o8ldozV+pwMLRkbodqz5RYcNkgJK99W5pwI30DPFRFKei1DK345SO3rjJ4KVa+752jzuxma0aeeY4y4Sw91dNLGncM1+GonDQAAAAAAAADABeqSaouTvmB9euYz/WLrQzpZ9j9OiTS8621a/fP5uvR7lzolLcc/aOTN9ZM79OyQRxV8aZAdIJr85gK9XfS+8610XeiP9OvRL+jaq65xSuCjJE3x/ZK0y8me3SAtyVmne/2nggukqkJlZVLo1Q0Eq+ztnDWH2gQrNLQRwS3v31iCQ0MV3MbJAAAAAAAAAPhuqipTUe472vXbEypzitp3HabBQ8N1fSNeK9ar4oQObs/W20WnnILG1FuhTwry9clfnOzZXHm9ekd0bNSMUKcKc3Wk1MnYrlGP/jd6/QF+E/bruOyGXurV0bP389NunF+tInCU/XGe7s9IqhnJ871LvmcHje7sNsjOt6SzBY1+ftMgO2h0ZVDtEJS3jmTbwSPPtiaQ9YorWYNu6Gvn4ed8BY4AAAAAAAAAoIWcyl6ixBmvKP/PToGPUA1/aqtWjL2xiUGOChX9eoEefrJ2CQ1fpt5/19Ix4Qqt84frR7XRNVzz9zvZsxmUrA/Wx6rBhTNOpCl+eJJ2+bTlfr32pznq5eSsntCv4vtpTraTbYTBz+Ro/S88ez8P7cZ51yqmqtv36UGf6d9+eMXV+vv257wSTx1NDRoZph2mPR6mnaa9qMfVkXpo40ZtbtRnkvoxvRsAAAAAAACAb9Anv07QgPj6gkZGmd5+fLhiljhroTdKhfKXuDRkdn1BI8PU69I9zwaqt0wnGhN8abRTyliwwC9odD60dLvxTWgVI44eemuBth16z8lJfa7tqQ13L1VI0JVOyblrTtDIKK/8i8b/56P68NPaq3/EzbfrhTvnOzkAAAAAAAAAQKtQsk3xdzziFVC5UcPnTNY/h/9QFSW79cKTryi/5ruOStj6tpLqWwfdW+EaDYlZriIna/7AfsYz83Vvr/b66ugbWv3kEv2q0BMuClBv2XY9fMtkZdiZjho8/i5191/Q3eOaQUoIsN67t7K3HtGtD20LEKDyH3FUpvzNr2jXZ042kI+z9cK2QicjDV6Wo/V3O+OGWrjd+GZc8IGjbyIw09ygkcc3EdgCAAAAAAAAAJxfB1MGK2b5CScXLNfKXVoxwmvytKObdM+QBcp3srrzef3+hRENBDsqlLOgt8ZucMI0VwzSkrfX6d6O7qztiwIlu0Zr3VEn7z9tm88SIIO0Im+dXM2d0+2LbM2/NUEb7QBYsHr1CVf+hwX2V3UDRw2p0K7Heyt+s+fY4rR533z188S8WrLd+MZc8FPVff23Kn359VdOzq1buxuc1Lk716CR4d8e017TbgAAAAAAAABAK1FVoIy1nqCRJWKOZngHjYwbYzVjvNdIoLfeUFaJk67XCRXle43tuXusRnoHjYwrIjQ6LtzJWLK3K+eUkzaKTyjHSZpRUNc3O/hSoZznFzhBI8ug+VoxvfFhojqOpukFT9DI0uux+2uDRkaLtRvfpAs+cPTV15X685dlTq5hbxzaqUXvr7YDQg1piaBRIKa9pt0AAAAAAAAAgFbi44PK8Vrzp/tdkbreSdcKVu+BsU7ayFbhH+tO+ObLb52fK9tatdTV9sprnJSRrYM189pZqipqp5Xr2dEe4VRxolC7fv2Kli1/Rb/aWaCietdkqlXx4XLNX+8JjkVowdxYXf99J9tkFdr1b0tqR19dEacZv/CLiLVQu/HNuuADR01hgkaztz+jV/J/rafeX3PW4NH5ChoBAAAAAAAAAFqhP5/QQSdp9O7sPyzILbhruPo5aaOoxHtoUCDtdWOkkzR+e1CfOMlaFTpyMNdJuxWdrK331MeeOewsIaeUs8ilf7jNpfjZS/RCyhLNSRitIb17a+y/FqisvsmwKgr0wmOv1Kyz1HXWkxp3o5NpjoZGG1lapN34xl3wgaPLLg3S1Zc3vByWJ2hkpokzgaB///3rdmAoUPDofAeNTHtNuwEAAAAAAAAArYNPkMMSHBJoXJDlUjPuqFbOxw0FjjqqR3+vIFTBEs3fcLR2JI7l1E5T1tDIJUfuGs1fX+jze7cy5SwZrXueLQjwnXRw/WN6wXOIN07S0nivqfGarExvr2pgtJG/ZrYb37wLPnB06ffa6PJLL3NybkdKP3ZSbhVfV+rdP+X4rIVkAkImMPTCB5tU5RUcOh9BI//2mPaadgMAAAAAAAAA0H3MY3LVvHau0K4Fw/UP/VyKcVmfIb11a8ImfdInQvWtNvRJ6WdqX/P7UPWbvk47cvbp9/tytGPddPW72vnKUvSvj2ldoZPxOLpJTzzriRp1VMLCSepVT1ysUQo3adm22jBPoNFGxjm3G9+KCz5wFBJ0pTpc5btiVvGZUyqv/IuTk4IvDdLTUTPswI83ExhalbdRq36zwQ4enY+gkWmHaY83017TbgAAAAAAAABA6xB8xVVOqmmCGxOAuXqolv7bJHX3ev1cUVKog/utz9Ey6cZYrZgeq7bOdzavsQm9fpmhDw4c0UcH9un999/W5imD1DUsVKFXt1fXqEnavHW+V9DpqLbu9orAVJ3QrxYvqBkddH18smZEnkvUqExvr11TM+Xd2UYbnVO78a1pFWscdfk73yXIjp/+VKe+8F0xywR8TOCnvuDR8tz1ei7nlRafns60w7THm397AQAAAAAAAAAXttBrrnFSbkWf1DMFXemp2qCJpd/1DUzR5gjuM12Z727Vkl8OVfcwd+Am9MZIueZs1AdvJ2v4NRX6zC5163q974AK2xWhur5jgPIbB2m01zpKn+Qflaf1n6Qt0JxsJ3NFnBY8Eukz1V6TNXK0kY9mtBvfnlYROOp9bXdd+r1LnZz0v1/8Wf99yvvWdDOBnyV3zNKQG//RKXEzgaKUvH/XSx++2qJBI8O0w7THw7TTtBcAAAAAAAAA0Ip07q7BTtLIOXTCSfk6VVigT5y0We2o+40BAiL1CYvQvXNWKzPngD760xH9fsdGrfhlpNq3seot2K2DzmbSCPW6yUk2SrD9vxpVnmVdTilnuydqZPlik+J7dFOXH3t9Rr/ifGm8onvs8gT9qsQp8tH40UaNU1+78W1qFYGjm37QRT8K+YGTcweCMv64U1//7WunpFZocIieH5ZUJ3gUyLkGjcz+TTu8g1Gmnaa9AAAAAAAAAIBWpH2EBkU4aUvFpjeUUzuwxnFK773lFYi5IlaR4U66XhXa9bhXoOa2NTpY5XzlUVWoX631qvfOweoX6qRVoGTPb63PbSkBpnOrOiWfsRY/7qgmhLMar0mjjS6gdqNJWkXg6Jor26nnNTc7Obe8E7/T4c8/dnK+GhM8OtegkWH2b9rhzbTTtBcAAAAAAAAA0Jp01Mh/GeqkLV9s0vzluSrzCvJ8siVJ873iO9fH3aXeNYGTChX9eoHGulx6MCVXp2p+F6zet3rVe2K5Hl6UXft9xQntWvSIlh118tb248YMVU3cSDeq1wgnaflk7XL96piTMarKlL/mSS2rGSBl/X5Q7cpBLadMGWuW14426ni/5p51tNGF0m401SXVFid9QXv3TzmauO1xn1FGY3rEaMmQmbrE+i+QsopyJWY+of86vs8pcWuJoFG19d+cHc9py4FMp8Q9Td3aEU/pjh/3c0oAAAAAAAAAAK1GVaFeGO7yCuJIwWERGn5nd1XsfkNvHy1zSi1XjNCLe57X8Kud/IdL9JPRr8gzHmf4yn16cYQT/glQb32Co57XnnUjfEbeVBQsV8w9XlPEKVRdh92l4dedVs5b25VfUjsKKDgqWTteitX1bZyCBpz6dYJune2Jht2v1/40RwHDN/vX6DbX8ppp+gY/k6P1vzj7+KDz2W6cP61ixJFx63U/VXj7G52c2xuH3lXuJwVOri4z8uilmIX6x069nZKWCRoZZr9m/95M+0w7AQAAAAAAAACtUJtwPZTqFQyyVJQUKGP9Jr+gkbXdvz3ts13ZZ5/VBI2Mgye81kgKUG8gwX2m67WVvkEjIzhiutYv8x6FVKaidzbphfXbfIMv4ZOs7c5H8KVMGWtrg0ZmtNFDIxqeVO7bbzeao9UEjkwQaPwto/S9S2qb/MVfv9KK3H+zRxbVxzt41FJBI7O/Z//rZXv/HqZdpn1mfwAAAAAAAACAVqrjCL24K0NLxkZ4BTxqhfafrs3vZmhGH9/FfUIH/bMeCnfKrh6qGXf7LX7k1Lvil4PU3v8V9dURGvdMhj54dZK61/P6+vq7V2vPjnWaMezGuu2yfu+as1F7MqZ7rY3UgvZv0rK3nLRl8NT71avetY18favtRrO0mqnqjC+//koT33hcu4994JS43RM+TEvvmKXvt/m+U1LX139zTxh56ffOLWT516q/6tF3n9Vrhe84JW4DO9+qtXc9pcsvvcwpAQAAAAAAAAC0ahVl+uTwQX1y2kpfdo26hd9YN+jjrapCZWUVCg4NVXADr6Ir/lzmHqHUJlihoY2Mwng4+7E15/ffltba7otMqwocGfs+PaD7Mx7Tn7+qHRZoRvtM7TtOU382Xm28RiS1tKrqv2nVbzZoVd5G/c1Ke1x9WahecT2t3tf2cEoAAAAAAAAAAABan1YzVZ2HCc480PtenynrTBDHBHMWZKfYI4LOB1Ovqd8/aGTaYdpD0AgAAAAAAAAAALR2rS5wZEzsfa/u/vshTs7NBHM2/C5dD7yRpP8p/1+ntGWY+ky9pn7voJFh2mHaAwAAAAAAAAAA0Nq1uqnqPMoqyvXIO8nacfS/nJJaf3dZqObdNkmjwoed09R1Zmq69MJ3tOj9Nfo/r6nxPIbc+I96fliSQoNDnBIAAAAAAAAAAIDWq9UGjoy/VH6hWTuW6s3D2U6Jr2uv+pEeiZygETffrssubfwiW199XaFth97T87mp+vTM/zilvn5+0yA9O+RRXRl0tpXQLlyVp8tVaRJtghQSEmSXAQAAAAAAAACAi1urnKrOwwRtVg6fp/iIXwQcWWSCPjO3L1HPNT/XmF8/rP/Yv01FpcfswJA3kzfl5nuzndne/C5Q0Mjsx+zP7LfVBY2qynUkPVkP3nmHokaOVLT5jLhTUffNVXpBubPRhe3Aijs0YHKGSp18k1RVqtwEzKqc/LehqkT5KYmKGnwOx3GRsM/14JFanVfftVmizMmtox8bd90WaLW5LkYuU369h5yh2dY2s9NLnIImqiy37gE7ZHwBcc7jigIn3xROnw2OV/pxp6g+x9P0oL1tig44RS3Ffa02r96mPNPKD2Vo6Wizrzu0+kOnEAAAAAAAAGhhrTpwZHy/zff1xKCHtCZmgT1FXSCVVX9V7icFeuzd5xS1YZxuThmqzssH1nxM3pSb7812ZvtATP1mP2Z/Zr+tSlWJdi+IV3zKXgVFTdHKdZuU9fombVo8RaPaFmj5I6M1L6OZL6Nbi1NZWjhypNY15/10C7Bf+t4Xp2lZ5erc1ilEA8q15YlVym0dcc2WcTpLs1bsdY8IbGGlmUmKHrm2xQMn377j2pJZ6KQDO5K5ufUetxNwdiWu1aE27ZxCAAAAAAAA4Pxo9YEj4xLrv+Fdb9Pe+C26r2fMOa1rFIipz9Rr6jf7MftrbUozk7VwjzRmyVatnOFSr65hCmkbps79XJr80la96ArR7jUp2s0QmPOkQBvMS98+ScrY9pxGXecUowFBCmnznuYtek8XTezo8iAF7UzWku3cjE1RnPG68r90Mv6+3KvMjNZ7BZVmJGta1lX2s3r9jAFOKQAAAAAAAHB+fCcCRx5XBV+pJXfMUs4DaRrTI0aXX3qZ803zmN+bekx9pl5Tf+tUqMzUQilqqhL6hjhl3kLUY9wEDazM05vZvqOOyg9lafUjcXINu0MD7hytaYsydOS086XH/lTFxz2p3SWlyl8zS3H2VHjxWp5eZI+aqDyUoeWJoxU9+E657pulLX7T4h14OU7xC7JUWl6kzEXx9r6iRsZp3rL3VNzIaeVKc1K10N7HHYq+J1ELU/NU6vVbex/TU7XPSmc+baXj4rQ0y/fFfEN1nJsOijEvfWffrnZtnCI/Nf1wKk8bH3H6Kz5ZmYdMf5XrWHqypt1zpwYMG6n4R1J1wP881KPOOZybqvxAg8us/t+xzH3+zD4etM71sfJS7V5gtetlv9EcZgRE6lx3e5zzujHnfIxYi9YTC13qkPecFjZhRFzD57Ke47LUnAcnX3N9nzyuHc716TOtWkmetsxNrOm3+EeWaYd9zppp6Ew96wqxzsWT2tHYQ/Y7H3GJc7XF53wUaot1zc/cWGSls7TEStvHdMrKFqVpdlyi0g/ZG9YoTp9lbTNLmX5TwB1Ya/12cpqOOXnDu7/te3dN3XvH06/Fx7O0PH6kPV3jWadbO5mhhVY756VYdTlF9YmMjbWeX+8pPTvwluXZGUqvjFBCbLhT4qtRzzmbdR9mLdM8z3Vfc38G5l2v+5mWVU+9ZxcUMVUZ257VqJsDPb8BAAAAAACAlvWdChx5hIX8UEuHzNL+SZlKHblEw7oOUGhw4164me3M9uZ35vemHlNfq/ZRgbJPB2nU8P4KcorqaBetRbve1TOjwpwCqfLDFCUmpmhfUJQemr9YK6dE6wcH1yp+9FRlnnQ2MirKdeRkiXavmKp15f01eV6SJkdUKjNlqpasXat5j+9VB9cMLZ43Qf10UKsfmeq7Hkl5iY6cKNSrSTO143KX5lj7WjQuQmd2Jisubq0O1DeKwFGcPlWxc9P0ebexemLxYj0xtps+T5ur2OkZNYGnGwZN0UNjB6izle7tstKTpujnP629JhpTx7kJU+eGXvra/bBXL0xN1Zn+ie7++nKvlk59WhvXJGnu3g6KmTFfz0y4RTq4WQ9OT1Ox89P6lGZMlStxmXI1QAmzrXM4Y6xu+ixN08bP9R1dVl6g1fGJWrjzjCIfSNLK+RM1sDxNiUmbdeCE1a5yr3XBqkqUOT1e09I+101jZ2rl4plK6HNGr86NU9yKghafYu2qPhM1Z1SIctck+1539WjcuazUmc/8jsvDnIfPvI7Cvr6L9GbyTL3wWVeNf3iGpof/wP6qsiBFcffN1brPwjRqitW/Vr8NVZ4WJsZr49lnTjuLq9Vj4mMa17ZQS5Ibcf0FOB9jun2uDXPjvdZCClOkdc3fN8BMc9ZdMVb6oUku3WQuyS5d1flUkTLzvG/KEu3bWWAdd4He/9A3AJWfVaLyn3a37yXDv7/tezdzrlz3peiAd0zF9OuRLC2fuVbFP47VnNkT1MPdjXWdfE9Lp6fo8I/jNXlSXzU4OVtYtEYNDdLutJ0B7onj2p5WoKChLg2sfbzVaPRzTuU6sCJecc/s1Jl+E/XM4vl6aMBflD49SRuPOJt48a/32YlRCrLK4hP8+qURQqxzVF/AGQAAAAAAAGhx1ReRMxXl1Uc+/7h626H3ql//7x01H5M35eb776TfrqruP2hK9bZiJ98YX+ypfn5oVHXs0t9UVzhFtq8PVm+Ijaq+/dGd1WecInf9UdVjXzroFBhnqnMej7LKJ1S/9ienyDizs3qJte2s12obs3+52c4q2+rXwBOvV8+y2pD46jGnwNl20uvVnzv56mNbqxMHRfvUZzuxtXqWVefct2q2rK4utuqzyl74rZP3aEodgXxdUX3mjE8vNaC4etskv+OwuPshtnqDTze6+6v/hE3VH3/tlFkq3p3fiHN6pvrw689VL1id73sOPefg9dq9f7w+trr/UKu+E06B4+MNE+xz0395vlNSXf3pq6as7rafv5VUfbt/+wNpZH+5+2NV9X6T8Vx3U16v/rSmHwL0Y6PPpfNbr+PyqHONea5va1vfVldUf/7++uoFT3m3yfL1kepXx1ptXVVbd506A8qvfsG7rw++VD3W/1ica9i7zD4fAc7dp1unWO2eWb3da6efv2bKnD71sn95tG/7/m9n9YJBY6tnTXJV9/e+1/+0qXrCIFf1hj84ebu/A9y7n71VvcC6d72fCZ7r+4Xf1tTm8DsXJ6zr0zrXYx/f6duvAbn7zO6PP7xUHevdNg+v8jrH35TnXH3XltUnpg8C1Vvnmjnzm+oXXNb+Vvv1S4PXhhfneqzzHAMAAAAAAABayHdyxFF9QoKuVNd2nRVz02C5fnJHzcfkTbn5/ruotMRvrqlGqMx7X+mV4Ro3rq/vKKU24Rr1QF/r+53a5zPlUphGDfWeBipEV9nDBCJ0Uxe7wC0kXN2tzXKP+48LuF13x/gNB7jOpXHRQTqQucdnWixvx7IzdOC6sZrsNVLKZv327ihp9+7fNbg2zrnVUan8Z0YqesRIrc5pgbE21j4H+nTj1VZPWn4ars5eIw6CukcoUoU6dsIpCChE3Vwz9MSkCN9zeHkndb7OOgcfeXq1ULszStXBlagYv7WXOrviNMRJux1XbuZxBbkm1Nm2XfRYjW9bqsw9Zxtq08z+sq67cUsn6Kb9KVpeM4qmrpa4HgKzrm+XXz9auXYDJuiJeS518B4N0iZMnX9sHekfjzc4xdpZhU/U4glhyk0520gr9/noNnZmnfPRIWa0RqlAuR82fMQ33dLfugz2ap/T4MqC32hH2wFKmBCloLzfaL8z6q+0IE9HgqLUy7lG7f4Osu5T//5uH637RoXoWNYe+QzGsc5DTJ+zjLw7tVfLpycru/MUPTv/dt9+bUjPaI3pUq4NmXu9Rr1Z11tmhoq7jNWQnk6Rl6Y859zHGq1xLr9j7eLSmL5O2uGu93ZNnuh3zYT0VYzVV8Xb/foFAAAAAAAAuIBcVIGji1W7sE5OqvHKS83L+W66we8dqRHS/Rb1UoEO/8kpsLVT0OVOsjm6WvsK8PtrO3WVTpbqjJP3d8a081SaFthrtnh/4rU6x2xwxuslcmDnWkdQiHkRHqKr/s43rNAsba16nGRLqTxZqNz0VK1bNlfTrOOKG5Go1T6BiAqdOS3rXAeYEKxtV/XwCUh8rmLrt707dXDy3sLVo59UfPDIWQMmze6vLmP1xKRw5a5JUvpHTpmflrgeAjvL9V1apANZm7UuJVnzzL5GjtbsPc5356jz2GQ9Gl6opY9v1rGAU9a5z8ex9CS/47U+Cau03dqivLzhwFFQ39t8gkyHf7dXQYNuVbfuEYoJ2qv8g6a0XPs/KLTKI3STE9Cx+7trJ10bIMDT7ZYB0umD+tiso+Rxtuv7ZJaWT35S6af6ao5/MK5ROmlobIQqt7+vfZ7pLb/8QLuzKzUwNkqBrtimPOfqP9YQ3RDuW4G73r1aneB3TqzPgnTr7jhdad11AAAAAAAAwIWJwNHF4Aft1E2FOvLH5r0yr6NNkIJVqcoWWfvHEWTqbKbLO6ln376K9PsMjHZp3IBOfqNE6tHsOoLUY8pW7Xl3q8YFXnf/W1SuAylxio6bqoWbC1UeFKbIUfGas2qVpp+ntlqnUTrrdXFu/dVhVJIe7Vmi5Y+u1ZH69tMS10MjFWfMUvQ9iZq2dqeKK69S5+ixeih5vZ6JdjY4V23CFJM0RZEnUzX35fpHcv2gU0Sd443sO0CjRrn0sy5nGeHjcfkt6tVX2vHB76w7u1D52ZWKuaW7VX6rfjagUrv3Wfs+/YF250kxkbc2rg+t54RR0djnxId7dbhPtCKD8rTw8eatLRYyyKVRQe/pzWx36LI8O0Ppul2jBjW4SlJd5/ycC9NNdc6J9YmK1rhR3VXf8k4AAAAAAADAt43A0cWgywDFdJEy395b/zRdpVmaN/iOmsX0g0LMuIASfe4zHZ3jWJFy1VWdA/0Jf3MVFunjAC9oPz91XGpbf1Ap+PIQmemfRk2ZooRAnzER7qnezqIl6lCTR0d8A07t1EbrfPaetElZrz2r6daxjBl1u3p0bacgnxhisK5qK31cEmCc0OkiHfAZneRseyrQmKISfWxGpnQOU4Ov6ZvbXyaQ8uhERZ5O04I1eXVGDzX5XFaeSzC1QJlrCnRV9GJlvb5eT8yw6h8brV7hYc0PggZynUvTJ4areMuTWpfnfwe7z0dIn9jAx2t9RkU0ePVaQtT79r7SngIdPlSg7NN91SvCBH6C1PPWviq2yo8U/EY75Cl3s/v7pPWccPLeij+yLwZd296db9DQ+Xpx9gwtWuxSh/0pmmX1bZPPzuX9FeMK0e60nSrWcW1PK1AH10j1qme0WJOecyYQFvBYK/Xpcfdz08Ndb1cNeSDwOUmYcnvAEVAAAAAAAADAhYDA0UWhk4aaRTjyntPCQOvDVJUo88kU7Q66XXcPdU+5FNKnvwYqT29u99++XLk7dkpdotTbb02Vc7NTu/f4vRQvz9P2rHJ1GDpA3Zwif936RavDyQztrjMYo1QHNm/WjqK6AY7Kct/X0c2po1VwAnGdu/jNw1WSp/eLnLQtXJHR7VSc8VKdtXSOpW/SDiftZm07NETFWVnK949hfLRTbxYFaVRkhFNwnlzn0pxJ4SpOX6vVPsfRlHMZpms7W//k/M535JJ1ze3OdtKNYOJON3Tt7DsC58sC5Zpp8VqQGWk1p2e5Nq7ZrH1OmZv7fBzJzKo7Aqs0T+mbs3SszuV7Rmc8U7l5CfnprYqs3Kn01Dwd6Wvd3229yk/u1Aub90pe5Ybd36eztN0/oFV1XLlZRQoa2lc9GhskvNzdi0F9pujZKeb8Pq11jVifyV+3mLHq8dFmZaZmaMtHnTQmpv6hbU15zpl1oIICHWvpXu3wm5rQXa8pr/vsOJaVqsy8kqYHxQAAAAAAAIBvCIGji0TI0Me00hWi3JR4xc9J1Y6cAh0pLFB+VqoWxsVr6f4QjVk8VZGewQntojV+TDvlrpmqpemFKj1drvKSIu14JlHzdlrbTopu2b+Yv66ritfO1LqdRe59Hc/TuukLtOXLcCWMOsucZj1jNblvubVtYu1vrXbmpiRp2ssZOvx/Xq/0w7qpR1spc/Na6/jzdMSz9kpT6mhN2ruPN31jqg4ct47JOq7Swgwtnb5Zn3q9/De6jX1M49oVaWl8nBYuS9G6lBQtnzpSiSe7KsbZxqPbmKka8mWWZk1PUb5Tb/GHqZo3OVWHe07UmH7nv7/auZ7UE30DDBhqwrnsEXm7gk6naepk9/Vg3wsJVl81elazzrqpj5S7Zb12FJXY/WBft1OXKbfFuyBMQ56cqSGqrBNwsM+HdRyJiSnKddpRWvSe1j26QMszinTGqy3tftLdum/3auPLWcrPsfrHO9gU1le3hZdrR16hIm8Nrx2ZFRahn113XPlFlRo48BbfEVt2f1dqyxPWvZtz3OmDAm2Zk6jlJ8M1fVz/Zk0N2GFUsvW8krbMTaoTzGxQpwH6uXUcGzdmqLhLtCLPtsRbE55zQQPiNaen77GWFmVp9eT1Kr/ZLzhr1xuiHU8naumWAhWbek8f14EtT2ruM2l63zqm83+XAAAAAAAAAM1D4OiiEaIeD29VxmKXfvjHNC2cO0vxk2dp2jNp2tc+Sk+8tF6T+/hOadVt4nptmhKuQ2unyjVypKLvS9SSnHZKeL7utuesbZSmPxetY8sS3fsaP1dbym/VE+tWaUiAhetrtdPAxZvsl8zpi5zfWu2cnSWNX/KSXzvDNeapsepxMsM6/rl6Ouu4U96UOlqRNtbxLp2ioafS9OB465is43LNzFLnpGSN8R8tFhKhhE3rtXJcuCoL8pSbd1BnIh7T5seifQMFRrvb9cSmxRrfJkvTnHpHz0rT5/1maNNylzp8I9P2WedsXpJG1Xn73vhzGTRgqtZP7Kugj8z9MFfTUvJ0w8PJSvips0GD2mnI7MVKaLdXCxPi7H6ITkhRaexzeqSfs0lLsvp9zmO31w04eJ2P2U47XAnJSm/j0srVU9TD5xYYq8XjOunj9GWaNvdJZXtuAVuYevY1N1uYfhbhfdN1Us8BppJwq9w/qubV33Pj3X0wfpY2fNbfvndjmj0q0XpePfCYxrQt1NLpKTrQpIFHYRoU4x71NjA2qsEAd+Ofc2EakrxKj/Yr1xbnWF0JqSqf8JymB1hDqabe1FkabeodGa8HU4vUc8oqLRp11ocaAAAAAAAA8K26pNripHEx+bJcZsa2oJAQBTX0or+qUuVm4zZBCglp+b+TP7DiDj14ZIoyVrtkXr9Wni5XZXP21YR2munqggJtc56P9dti96n1b6POt48CrR48S7vHrNLWiQFGfjXlOvqmtcT10EiV5Vb/Vl0A/VBpnQ8zDd3lITrr4Zi+qbT6pZ61f5rF099B1r5bst5vSlPufaefG3u+3fefVW/b784zBQAAAAAAAN9djDi6WJkXy20b+ZLbvEi1tv2mAilBzd1XE9pZb5DgGz7Wb4rdp9an3vNd8p5WL0hTsf86OYd+pwPWPz261DPfV1Ouo29aS1wPjWQCCBdEP5igjTnmhg7H9E1LB3c8/d0ag0ZGU+59p58be77d999365kCAAAAAACA7y4CRwCkqjM6lrNWcXFztS79PeXnZCl92VzFT92swz2nKCGqzoR1AAAAAAAAAIDvIKaqw7eueHuKMkv7674xEXXX08E353SRdry8Vq8WFLvzIZ0VOfQXuscVoXYX4ogiAAAAAAAAAECLI3AEAAAAAAAAAAAAG1PVAQAAAAAAAAAAwEbgCAAAAAAAAAAAADYCRwAAAAAAAAAAALAROAIAAAAAAAAAAICNwBEAAAAAAAAAAABsBI4AAAAAAAAAAABgI3AEAAAAAAAAAAAAG4EjAAAAAAAAAAAA2AgcAQAAAAAAAAAAwEbgCAAAAAAAAAAAADYCRwAAAAAAAAAAALAROAIAAAAAAAAAAICNwBEAAAAAAAAAAABsBI4AAAAAAAAAAABgI3AEAAAAAAAAAAAAG4EjAAAAAAAAAAAA2AgcAQAAAAAAAAAAwEbgCAAAAAAAAAAAADYCRwAAAAAAAAAAALAROAIAAAAAAAAAAICNwBEAAAAAAAAAAABsBI4AAAAAAAAAAABgI3AEAAAAAAAAAAAAG4EjAAAAAAAAAAAA2AgcAQAAAAAAAAAAwEbgCAAAAAAAAAAAADYCRwAAAAAAAAAAALBI/x8kWM+92ggDuAAAAABJRU5ErkJggg=="
    }
   },
   "cell_type": "markdown",
   "id": "758adaf6",
   "metadata": {},
   "source": [
    "![image.png](attachment:image.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "38de14bc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "263/263 [==============================] - 2s 6ms/step\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# confusion matrix\n",
    "Y_pred = model2.predict(X_test)\n",
    "Y_pred_classes = np.argmax(Y_pred,axis = 1) \n",
    "confusion_mtx = confusion_matrix(y_test_digit, Y_pred_classes) \n",
    "f,ax = plt.subplots(figsize=(8, 8))\n",
    "sns.heatmap(confusion_mtx, annot=True, linewidths=0.01,cmap=\"Blues\",linecolor=\"gray\", fmt= '.1f',ax=ax)\n",
    "plt.xlabel(\"Predicted Label\")\n",
    "plt.ylabel(\"True Label\")\n",
    "plt.title(\"Confusion Matrix\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8eaee0b6",
   "metadata": {},
   "source": [
    "The model is working pretty well across most digits.  The largest problem appears to be \"7\"s which are being predicited to be \"2\"s."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "75a6d874",
   "metadata": {},
   "source": [
    "The nueral network performs well.  Accuracy on the test data outperforms all other models.  We select this model as our final model"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}