ZhiyaoShu/street-view-housing-number-digit-recognition.ipynb

## street-view-housing-number-digit-recognition.ipynb
{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/ZhiyaoShu/c3e067801e14ab5659ac4e0c6691c40f/1-street-view-housing-number-digit-recognition.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Q91KqmCRu64D"
      },
      "source": [
        "# **Deep Learning - Street View Housing Number Digit Recognition**\n",
        "\n",
        "\n",
        "--------------\n",
        "## **Objective**\n",
        "\n",
        "The ability of deep learning to process visual information using machine learning algorithms can be very useful as demonstrated in various applications. The objective of this project is to practice predicting the number depicted inside the image by using Artificial or Fully Connected Feed Forward Neural Networks and Convolutional Neural Networks. I go through various models of each model, and finally select the one that is giving the best performance.\n",
        "\n",
        "**Datasets**\n",
        "-\n",
        "The SVHN dataset contains over 600,000 labeled digits cropped from street-level photos. It is one of the most popular image recognition datasets. It has been used in neural networks created by Google to improve the map quality by automatically transcribing the address numbers from a patch of pixels. The transcribed number with a known street address helps pinpoint the location of the building it represents.\n",
        "\n",
        "\n",
        "## **Process**\n",
        "\n",
        "- **Images visualization and Data preparation:** Print the shape and the array of pixels for the first image in the training dataset. Normalize the train and the test dataset by dividing by 255. One-hot encode the target variable.\n",
        "-**Build two ANN model:** After fix the seed for random number generators, Develop a function returns a sequential model, and visualize the training and validation accuracy with importing matplotlib library.<br>\n",
        "Try another model with higher complexity and observe the performance of the model. Also plot the training and validation accuracies to compare with the first model.\n",
        "-**Predictions on the test data:** Make predictions on the test set using the second model. And print the obtained results using the classification report and the confusion matrix.\n",
        "-**Build two CNN model and make predictions:** Practice the same datasets on the CNN model to observe the outcomes compared to that from Ann.\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "C8U3DUa3eNsT"
      },
      "source": [
        "## **Importing the necessary libraries**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "-dVzeuF3eQx1"
      },
      "outputs": [],
      "source": [
        "!pip install tensorflow==2.8.0\n",
        "!pip install numpy\n",
        "!pip install pandas\n",
        "!pip install scikit-learn\n",
        "!pip install matplotlib\n",
        "!pip install seaborn\n",
        "!pip install h5py\n",
        "\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.preprocessing import LabelEncoder, OneHotEncoder\n",
        "from sklearn import model_selection\n",
        "from sklearn.compose import ColumnTransformer\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "from sklearn.impute import SimpleImputer\n",
        "from sklearn.model_selection import train_test_split\n",
        "import warnings\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from sklearn.pipeline import Pipeline\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.model_selection import RandomizedSearchCV\n",
        "import tensorflow as tf\n",
        "import random\n",
        "import h5py\n",
        "from sklearn.model_selection import train_test_split\n",
        "random.seed(1)\n",
        "np.random.seed(1)\n",
        "tf.random.set_seed(1)\n",
        "warnings.filterwarnings(\"ignore\")\n",
        "\n",
        "tf.__version__"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OCHjymHPPAAk",
        "outputId": "95089be2-b6a2-4338-db51-608cd5e4f18e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8lsux2ZwyTTR"
      },
      "source": [
        "## **Load the dataset**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "BApX9qgNsqV0",
        "scrolled": true,
        "outputId": "a8e493d8-86f4-4230-9852-b79a38450631",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "['X_test', 'X_train', 'X_val', 'y_test', 'y_train', 'y_val']\n"
          ]
        }
      ],
      "source": [
        "# Open the HDF5 file\n",
        "df='/content/drive/MyDrive/04-Projects/0.MIT PE/01. Deep learning/SVHN_single_grey1.h5'\n",
        "with h5py.File(df, 'r') as file:\n",
        " keys = list(file.keys())\n",
        " print(keys)\n",
        "\n",
        "# Load data into the train and the test dataset\n",
        "with h5py.File(df, 'r') as file:\n",
        "  X_train = file['X_train'][:]\n",
        "  y_train = file['y_train'][:]\n",
        "  X_test = file['X_test'][:]\n",
        "  y_test = file['y_test'][:]\n",
        "  X_val = file['X_val'][:]\n",
        "  y_val = file['y_val'][:]\n",
        "\n",
        "X_train, X_extra, y_train, y_extra = train_test_split(X_train, y_train, test_size=0.2, random_state=42)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "y3lwKpOefkpA",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ac751d37-2dc9-46b8-8a71-40a7b1e7adf5"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Number of images in the training dataset: 33600\n",
            "Number of images in the testing dataset: 18000\n"
          ]
        }
      ],
      "source": [
        "# Check the number of images in the training and the testing dataset\n",
        "num_images_train = X_train.shape[0]\n",
        "num_images_test = X_test.shape[0]\n",
        "print(\"Number of images in the training dataset:\", num_images_train)\n",
        "print(\"Number of images in the testing dataset:\", num_images_test)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "akTUOfLlgwoM"
      },
      "source": [
        "**Observation:**\n",
        "\n",
        "1. Both the training and testing datasets seem to be of considerable size (42,000 and 18,000 images respectively).\n",
        "2. The given dataset has a training-testing split of approximately 70-30, which is a common split ratio.\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kxODV6HKykuc"
      },
      "source": [
        "## **Visualizing images**\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Bvsc8ytHsqWD",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 363
        },
        "outputId": "54729d76-76fe-452b-88d4-cd71a2d43538"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 10 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgQAAAFJCAYAAADkLDW5AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAAEAAElEQVR4nOz9e4xt2VXfj37mYz323lV1Hn262w+MHf9weCgWICxDEAgDSQwKFxnFF6KIQG7iBEXhykIQkkgEI0XKQ4CCwAlEioEkOIkiC7gkQuQqARQRORAuCgF+PBxjA360+/R5VdXeez3m4/4x5pxr7V11us/pruN2NzWkOrvOrv1Ya6655hzjO77jO1SMMXJpl3Zpl3Zpl3Zpf6xNv9gHcGmXdmmXdmmXdmkvvl06BJd2aZd2aZd2aZd26RBc2qVd2qVd2qVd2qVDcGmXdmmXdmmXdmlcOgSXdmmXdmmXdmmXxqVDcGmXdmmXdmmXdmlcOgSXdmmXdmmXdmmXxqVDcGmXdmmXdmmXdmlcOgSXdmmXdmmXdmmXxkvAIfjwhz+MUorv+77vu7DP/MVf/EWUUvziL/7ihX3mHze7vC6funZ5bT417fK6fOra5bUReyQOwY//+I+jlOJXf/VXH8XHv+j2u7/7u3zbt30bX/zFX0zbtiil+PCHP/xiH9Zz2sv9unzP93wPSqkzP23bvtiH9pz2cr82P/mTP8k3fMM38PrXv57lcslnfuZn8u3f/u3cvXv3xT60Z7WX+3UB+C//5b/w5V/+5dy4cYOrV6/y5je/mX/zb/7Ni31Yz2kv92vzUz/1U7z1rW/lVa96FU3T8Gmf9mm8/e1v5zd/8zcf2XfaR/bJL2N7//vfzw/+4A/yOZ/zOXz2Z382/+t//a8X+5AubWY//MM/zMHBQfm/MeZFPJpLA/gbf+Nv8KpXvYpv/MZv5NM//dP5jd/4Dd797nfzsz/7s/zar/0ai8XixT7EP5b2Mz/zM7ztbW/jT//pP10c6v/wH/4D3/RN38QzzzzDt33bt73Yh/jH1n7jN36Da9eu8c53vpMbN27w1FNP8aM/+qO8+c1v5v3vfz+f+7mfe+HfeekQPA/72q/9Wu7evcvh4SHf933fd+kQfIrZ29/+dm7cuPFiH8alzex973sfb3nLW3ae+4Iv+AK++Zu/mfe+97284x3veHEO7I+5vfvd7+aVr3wlP//zP0/TNAB8y7d8C5/1WZ/Fj//4j186BC+iffd3f/eZ597xjnfwaZ/2afzwD/8wP/IjP3Lh3/micQiGYeC7v/u7+YIv+AKuXLnCarXiS7/0S/mFX/iF+77nn/7Tf8prX/taFosFX/ZlX3YudPI7v/M7vP3tb+f69eu0bcub3vQmfuZnfuY5j2ez2fA7v/M7PPPMM8/52uvXr3N4ePicr3sp2kv5umSLMXJ8fMzLrZHnS/na7DsDAF/3dV8HwG//9m8/5/s/le2lfF2Oj4+5du1acQYArLXcuHHjZYHavJSvzXn2xBNPsFwuH1mq7UVzCI6Pj/mX//Jf8pa3vIV/8k/+Cd/zPd/DzZs3eetb33puxP2v//W/5gd/8Af5W3/rb/H3/t7f4zd/8zf5iq/4Cj7xiU+U1/zWb/0WX/RFX8Rv//Zv83f/7t/l+7//+1mtVrztbW/jp37qp571eH7lV36Fz/7sz+bd7373RZ/qS8peDtfl9a9/PVeuXOHw8JBv/MZv3DmWl7K9HK7N3J566imAlzya81K+Lm95y1v4rd/6Lf7+3//7/J//83/44Ac/yD/4B/+AX/3VX+U7v/M7H3osPtXspXxtst29e5ebN2/yG7/xG7zjHe/g+PiYr/zKr3zg9z+UxUdgP/ZjPxaB+D//5/+872ucc7Hv+53n7ty5E5988sn4V//qXy3PfehDH4pAXCwW8SMf+Uh5/pd/+ZcjEL/t276tPPeVX/mV8Y1vfGPsuq48F0KIX/zFXxzf8IY3lOd+4Rd+IQLxF37hF8489653veuhzvV7v/d7IxA/9KEPPdT7Xgx7uV+XH/iBH4jf+q3fGt/73vfG973vffGd73xntNbGN7zhDfHevXvP+f4X017u1+Y8+2t/7a9FY0z8vd/7vef1/k+Gvdyvy+npafz6r//6qJSKQATicrmMP/3TP/2c732x7eV+bbJ95md+Zrk2BwcH8bu+67ui9/6B3/8w9qIhBMYY6roGIITA7du3cc7xpje9iV/7tV878/q3ve1tvPrVry7/f/Ob38wXfuEX8rM/+7MA3L59m5//+Z/n67/+6zk5OeGZZ57hmWee4datW7z1rW/lAx/4AB/96EfvezxvectbiDHyPd/zPRd7oi8xeylfl3e+85380A/9EH/pL/0l/sJf+Av8wA/8AP/qX/0rPvCBD/DP//k/f8iR+NSzl/K12bd/+2//Le95z3v49m//dt7whjc89Ps/leylfF2apuFP/sk/ydvf/nb+3b/7d/zET/wEb3rTm/jGb/xG/sf/+B8PORKfevZSvjbZfuzHfoyf+7mf45//83/OZ3/2Z7PdbvHeP/D7H8oehZfxIJ5bjDH++I//eHzjG98Yq6oqHhAQ/8Sf+BPlNdlz++7v/u4z7//Lf/kvx6ZpYoyTJ/dsP7/2a78WYzzfc3u+9nJDCGJ8eVyXub3iFa+IX/mVX3mhn3nR9sfp2vy3//bfYtu28a1vfWscx/EFf96jtJf7dfmWb/mW+Lmf+7k7EecwDPENb3hDfPOb3/y8PvOTZS/3a3Oe3b59Oz755JPx27/92y/sM+f2olUZ/MRP/AR/5a/8Fd72trfxt//23+aJJ57AGMM/+kf/iA9+8IMP/XkhBAC+4zu+g7e+9a3nvuYzPuMzXtAx/3Gwl+N1ec1rXsPt27cf6Xd8MuzlcG1+/dd/na/92q/lT/2pP8X73vc+rH3pFzq9VK/LMAy85z3v4Tu/8zvRegKLq6riq7/6q3n3u9/NMAwlwn4p2kv12tzPrl27xld8xVfw3ve+90JFlLK9aHfj+973Pl7/+tfzkz/5kyilyvPvete7zn39Bz7wgTPP/d7v/R6ve93rACGSgUzmP/Nn/szFH/AfE3u5XZcYIx/+8If5/M///E/6d1+0vdSvzQc/+EG+6qu+iieeeIKf/dmf3dGKeCnbS/W63Lp1C+fcufDzOI6EEB4dNP1JspfqtXk222633Lt375F89ovKIQB2SsN++Zd/mfe///3nvv6nf/qnd3Izv/Irv8Iv//Iv89Vf/dWAlGO85S1v4V/8i3/Bxz/+8TPvv3nz5rMezwstB3m52Ev5upz3WT/8wz/MzZs3+aqv+qrnfP+nur2Ur81TTz3Fn/tzfw6tNf/5P/9nHn/88ed8z0vFXqrX5YknnuDq1av81E/9FMMwlOdPT0/5j//xP/JZn/VZL/nSw5fqtQF4+umnzzz34Q9/mP/6X/8rb3rTm57z/c/HHilC8KM/+qP83M/93Jnn3/nOd/I1X/M1/ORP/iRf93Vfx5//83+eD33oQ/zIj/wIn/M5n8Pp6emZ93zGZ3wGX/IlX8Lf/Jt/k77v+YEf+AEee+yxndKYf/bP/hlf8iVfwhvf+Eb++l//67z+9a/nE5/4BO9///v5yEc+wq//+q/f91h/5Vd+hS//8i/nXe9613MSPu7du8cP/dAPAfDf//t/B0Tg4+rVq1y9epVv/dZvfZDhedHs5XpdXvva1/IN3/ANvPGNb6RtW37pl36Jf//v/z2f93mfx7d8y7c8+AC9iPZyvTZf9VVfxe///u/znd/5nfzSL/0Sv/RLv1T+9uSTT/Jn/+yffYDRefHs5XhdjDF8x3d8B9/1Xd/FF33RF/FN3/RNeO95z3vew0c+8hF+4id+4uEG6UWyl+O1AXjjG9/IV37lV/J5n/d5XLt2jQ984AO85z3vYRxH/vE//scPPkAPY4+CmJDJHvf7+aM/+qMYQoj/8B/+w/ja1742Nk0TP//zPz/+p//0n+I3f/M3x9e+9rXlszLZ43u/93vj93//98fXvOY1sWma+KVf+qXx13/918989wc/+MH4Td/0TfEVr3hFrKoqvvrVr45f8zVfE9/3vveV17zQcpB8TOf9zI/9U81e7tflHe94R/ycz/mceHh4GKuqip/xGZ8R/87f+Tvx+Pj4hQzbJ8Ve7tfm2c7ty77sy17AyD1ae7lflxhjfO973xvf/OY3x6tXr8bFYhG/8Au/cOc7PlXt5X5t3vWud8U3velN8dq1a9FaG1/1qlfFv/gX/2L83//7f7+QYXtWUzG+zOTcLu3SLu3SLu3SLu2h7VO+/fGlXdqlXdqlXdqlPXq7dAgu7dIu7dIu7dIu7dIhuLRLu7RLu7RLu7RLh+DSLu3SLu3SLu3SuHQILu3SLu3SLu3SLo1Lh+DSLu3SLu3SLu3SuHQILu3SLu3SLu3SLo2HUCp8/T/9fqKCcMVx9NiapnLcWK653mxotONqJY+V9mgiRoXyWCnPgemolaNSHkNAq4AhUikHQKvG8tqVGqjK3/MjVEAX4V6oGNCEqBkwO8cZomYda3zUhJm/M0RDSN2etQrlO/P3jtHKa9B0ocKjGaOhC1V6T8QQGKPhGXfIxtf838ev5H//wauJp5bFxy0HfxSxXWT1kY7q1hq17Qm37hC7HlVZlLVQV/DYNcKRSIL+f3/5u5//1Uv2bz7wRWVMayXa45pw5nUBjU9j0MWKMRp8lPMcU58rj+h9b3zDJtSEqBijIUSFVpFKeRmLdH0BWj1Spe/Nj3Or0nUH0vsCJo1nPuZ8vCZdm/3rl4/bo8p19OnvPqry2laN5frmczLps/OcMyrgo6aLFT6q9L1yPHIuMif/H6//jQe8Aufbj//en975bI/mpjvkjltxZ1zyW3dfyXHX8sxHrnL025b6JHLl93uaDzwlH1BXRGsYPu0qtz6nZTyEzad5Vq8+YdUM/MlrT/Np7V0OTM+T1T0q5QhohmgJUXESWja+oQ+WY7egDzZdTxmrI9uzsj1XzJbPaJ/iqt7Q6pGVGjBEumgZMIzRsA4NY7Tl+sB0T2kVyn2Ux91HzV2/5LZf0YeKp4dDTlzLGAwbVxGi5qDquVptafTIa9rbXDentHrkSHdlztTIvFnHmi5WF3JdAF73Q98vv+gIGqKKoAC1+zrlFXggzv6gItHK+5RXqFFBSL+n4Yk6fXZQKC8fGwzEKspHmfR9EZRTsKcGo5x8VtTTe1SS3FEe6nsau5bnYjru8RD6x7wcG+m1UYGTY4gGsJGo45nzJKZzV88iS6MBG1AKlAloE1EqgorkNgEf+H/+/ecc+2ezL/zPf1cOZzbeISp8kP87b0R1KCpCei4ETQiKGCEGvT+Uclqz41Tp0ZiQfo+Y9KhUxGj5BKPlYjbG09oRqwNLO7CyIvNsVdrHVMSes+7ldaoPlj7IvTIEw+AtAYULuqyrVst62tqRlRmw2rMwI412O5/ZB4sLhoCi95Yxyj4Y0nhpFam0p9aOVzTHXLNrDJF3fvZ/eaDxf3Dp4pjmileMTjbhravYmIpgFAtToVWEwM5J6NkCcr9NShNkQY9gkMkaoqJSgSpd3gowSmGI8j1RNrX55u+jxqPYhIYhbQZ5oPLmoJPTUSvHqCxBa3TcPUatghyLCmg9/c2k88ubYm0ctnaMjcY3EdeCCopQG6IxqNxBLH++VqA0WEOoLg6cMbNx9Xub5ZnXZmcI2fjGaOm8OAd5fO5nYzT0IY9jLN/da1uued74A6ocS6U8lXYYIo0ekyMQqZWT61icgXhm8y7nlVawEDUejSFM5600RLkx82vysXn0GUdJxsBiYpD3zsbQR42+IOCsfGc6F41snI1yNNqxsCO9tVAHQgN+gGAVWAMhQowoJ58RTdpQ6sDRouOw7nm8PuWJ+lgc2zSu62DZhJoxGp4ejlj7hlNX80x3IAuINwzeoFSktY7GOI7qjnt+wXW75orZ8IrkXFR4auUxRFa6x8eRWnkq5ZLDUDGkeZMdsnVouOuXdLHi48NVPtEfsfUVn9gecjrU+KAZnTiYy2bgoB5ojGMbaq5Xaw5Nxw17wkr3LHXPVbMB0r0emgu5LnJR0tyab4xxelQhbdJ5Q48zn0CpyYnI71Gzz5hvrCqCls0KPW3G0aYFVUFs0uvyd0Y1OSJ6em2MyfFwimDle2L+qiCOgvJqF/fN6/a+A8Dua6bf1d7JzizIMYoDosRZITkdz+ZIPITtOwIAPihilJ+QH4NKToAiBnmeqAhO/o8copiKqHS90+2OUpEQFEpFtI6E9HetAyHI7yOmOAtWB2rtuFJ1HNmtvFbFEijq2fmHqPDIJi0bfgBqQtxdW4OWv1sdaM2IUZGFGam1OMNNesxraYgKQ8DtBcFaBXQ610oFrPaydyp/boD2bPbADoHKE8ErxsESo2I91NTa46JhYVJkpqHCY5DNR6J7QQXMOZPGRw0Khpg3/FAWfzlBMEClFAbFMJu9Xay45Q4Yoy1R7hhNWRBz9Du3SnkOTUelPEvd46PCqJgWVfH8ajw+RZItY3FYfNQYHbinllTK05qRRTtCVPiFxS80KkZ8o4mVIdrd7wZkcbCa0Jzztwu0MZriFOTNrkqRuCAxlEX9tjtgE2qZhLONeD6xPRoXDGtfS0Q+Q1sO7ECtXbkxtIq4oGdOhtw0jXZcsdvkUAV6JOLLUX921nKknqP9fUQg/54jeU3A7P0tn7dBnIFWjbsDpBxe6TORmYzdxbT4yM5V3kR91LR65NB0jNFwUPWM3mBah1vUKK/wrSFWFuUDhABenNNgIdSgF47HF2sea9a8pr3Na6rbO1H7cVhw6ls2vubj3RXuDgtOh4abJyvG0eBHQxjkuqhKory6Gbl55YArdceTi2Nev3iGpR54srrLY+ZU5joerQMr5TjUMsJ93NJFwxg1x7GhCxVdrPjIcJ1T3/Ch9WN89PQK/Wg5Plnit0Y2Gi/393HjqVqHsYHjvuXG4pSDqufTFy2HpuPJ6h4r3QNw4hcch4trtBPN7oVXeSPKm2uK2lWK/Ilpg1EQdSQapHveeftgDtBVAgHSZhMNYNL/bUDZiDKBqnEYE3DO4L2WjW3URK9AR3QlUXnwiug00SlCpYlGnBV8Pk75iUoVZ2P/+GJeg/edhLljc54zkE8oIM6JUgQvYxIVO50EL8Lyxg/iJPgUTXuv5f9eEbxJqIBchxgU9IYcJ8S8JKgov89QIKUjwajyu9YyxlprXBqj/P1N5ai1Y2lHrtdrnqyOzxxvvtdDCkrHYOkTUmajrIshKmzwVDmoSoPeGMdCD+IYaAkYjJpQ9IxUB6V29jNZVyM6B8lApT02OQKP1CEoFikXxAfFEAzKR8aocdFQxd0DyAOVF60ME9/PfIJAjPIpYhRHYW4Zxh6iYZOgzC5axiAOwalvCszt9hwCqzxGBRo1RbSBUDYXSEiGmqJJQ2RIN5BHJZgoiDdmPMYGnE2LtkkeumbmoiZL7mnUmqgv7ga633hmOByld1AEoESsIOfko06bcZzOT3t0FGhUx0hQCk0sYxDi+ZF0cSDS2Fu8bNkxTk4CipBuiJIGiAZ0ijaDeOs7G/1soTIq7jgl8xvyQSyjSXN7tnn5fMzsrcY5XZLH16qA0bJYRBOJRsmmoXWKxgQlAErUqTS0dqQxbseJzSjZGA0bX7MJNadjw+nQcNLX9F0tm8ygUUOah6PG20Af4Lhp8EFjtedqtWU0hkOz5Uh3gKSFagTWzGhdIOLTHApBMyKphWPXcuob7g0LTrtGHJGNRXV6ir6RdXsEfBU47WsqsyBEzXG1FQRBD6xNg1GBIZp0Tz8a2lNUUZyCmRq+Cshul/fWgiCcs9PCFIrn1IOa9t2YHYkM1xtxBowN1LUTyFhHxtFMULhSKB0F2tYRpbRkL9ASnWt2HBU5thTgZ29gL9WhoiLmE9o59mcbnNlrItOJxUhEkT7xkVhI+01OE5RUQUyoS/qJQRWEJc+v6TyVfJBS5T6KqPJ/lS5QjNlhyqkIQRDmqIUhlE02rx8797kK+GjkXo+R/e3YqFgcAQmeAjajAdrvoKl5PaoArzQBRaU8fcolaRWwUNIWQHEGrJ6Cq4exB3YIsketvMJvLN4YTpUsBE3laI2Tm7WCA8Srr1JkplVMUVx2CqYBNHsOA6SBjgYYaZVGK0WIEU+kj/CUP+I4tHxkeIwPd4/RB8vaNSU/czLK4ubSD6QcEpIPurE4ZWkHjmzP4/UJrR5LrkU2ekeNRHUrJQtuzqV2sWKpB8ZoOKw6biw3nFaOj60a3KKGCL7WRKPBpIU3RPGgrUVVFaG1uMXFIQQ5+p1zBKbxjQXiza8JUZPj5cyRyB5po8fC9TiI3c5n9aHixLcEFGOYEIjzPNHMPwjJUcyoQkYLYIrkw+yGa7Qr/IRWjwU1KByEDM/FCfmY5/1zumf++V2oBA2YmUftcAzmjut5qa3nY9OCIXPeoxJa4Gn0yMr2DMHQtCPbhUR3rtXEtoJeobyHQc4rGvkx1nNoe67aDVfNhqt6IzB9WNGFij/sH+ODpzc4HRv+6M5VtictsTNUdw12kMhXp4sfsxPbRG6vK27XgU+sDnnqyhGrauCzrz7FGxaf4FBveV39DLXqxUlXageV7qLhaX/IbXfAB7ZP8v+79emc9DV37hygbteoUdGeKkyf3pCuT6g1vjUEA3euVtxdrLCN4xNXDlhWI69e3eVk1WJUYBPqMlcvxM7bAEP6iWoH8Vf5mPMmnGF/JdthcSJIr1EyrpknEPKH1AHdeJSO4gRYz7IeecXqhNaOHA8tx32Lj4puqBicwZjAsh4xOki6xWsGZzntD9DOoJ3CdKCcHFtBNnRMSMHuMcs5zbbwuTORTxiKI1Qs/x4m9CHmgEffZzyfh40+rZkJFYiA95oQNDGCd4lD4BXB6Qlxytct/65mjoGeOQM6bfhBxkel6xjyuXgtY5dSN5GMTuiddWpu83TmHJnug03Iqi5rIYDVEuxmVDUjp8LFcVTKlZSqRvhtSy3r7VIPHJh+SiMgQVrhcGlHW4Jdd+7xPps9uEOQxyIoVK+JJjIaywZwQbNualm8tS/RZqW8LOqzdEGOjuam9xwFWcglMm+ULAI9joAsPrf8AXf9ko/2V/nD9XU6b9mMNYM3jF6z6Rq8V8RENsluulKyoPbesKxGTutt2gRlA1zpngqJvLIzcKgHNJE+GtYJ4m7VwEbVHJiex9o1jXV8YnmEX1TolN+LVhO1nqA0rVFG+AOx1vj24hCC7BCMGMAWpyDn5LOjk20eHc/JluLADQVin0PemkAXK078oqRn5gS/8tn55kikmREYMxFGabQPKaVgCiEmm06oi0T8E4FxqQeWZnIyzawfl1YBE0MZgzl6kREFz+S8yPmfRW7yHNRMfINHYcJnkJt2YUYGO9BWjnUb8E7ja0WsBPZUXV/C0gygVNZzZLccmJ5Ds2WlRjpV0YWKk9DyTH/Ax9dHnHYNm1tL7F2L6aC5rdADaBeLQxBq4SyEWmE6i28i44Hh46Ohqh2NdRyYnt5WvMLe48j0CVmaxs8ntO62O+Dj41X+YHOdj926writsM9UNLcU2kF1GklgRhpz8DX4RhEtDL3FLwxuUfG0M9jK0ztLo31aUxQuXGCa7T63X+EOpA1hZ7OHBJfvvSnu7qPoCCZOPIHkGJiFo21lc182A43xHDUdbzh8mgPTc3M45Cl7yBAsJ7ahc5ZKB46ajlq7QkrrvWVz2uA3mjBKkJZRAhVSIB8yUinoU5n+5/EJVDz73HnOwBzuyMtJXl4vaDkLafOM6fcYk0OQUgXBJ96AV+D0riOQ0z2F55CC2ITalOg/OwUxoQhhQgoi8hkZlQDhMATkx3MWWQyz9XRMSFZ2Cqa0a0IzE7w/33daPbLUQyHStukGzUTd/FkezajNDj8u86mqhHznINyj6EL10OnPB+cQzCZFTI9kQkeQxX8IljHIgJgYysCkGbqz8D6ohRJNRnyU1MSJb7njVtwdl9ztF4xBs+lrRm9wTjN0FTFPFq8oRBklUclpLQiCVpET1xLMIPnPUIGWTVITGJQpF3p/ElQFmpGcjU4LhURySlICRoExU4WBtaA1wVxsymBid8/zS2HmGISJuJn+7/fg1wxjBTQ+5a182M3dh6joYlXy/fl7dmzmXCzMSBNlko/alM1dq0jQu/mw6binlWgihBr6FB2O2DMOpRz/9Fw+z2GGAIS9ebeLpLiSmghKFxj8hdrEfYilGmbfgdJkpjbT/KkkDRcriwqRYGWBiwq0jqXKY34uXazoQ8Xa12yHimEUiN50YDqF6WVD1l6cghw9EiV60j2Su7Yav7UMQXHct9wdlxglzuCAZoiaMZGjuqjYRMtxaLntDrg1HHCnXzJuKtTWYLYKu5XoVQ9ght3IMi/aIUiUmxLuuNoyNprjpuFmd0BtXEIZLxCY3v+oMxt/elKp4gAUmD5HxGrv/RmO1glRTT+68igTaduRg7YXh6AaqbTwkEJU9MGy9RWnY8MYhPjpfJ4zk+Nday9ENRPxVURHiFqVfVpFWWrR6fc53D87zx2gfy/lcN44qagmEmVOHeTUSVBnnaTnafmejxnCj2r3J6ULCKpwUSZkZ3LoVPJSoiKlC0CpdA4Z5ciOQj5VJVldpRTzHsCjN6xHQTzvjosdNLQQHxPXaoxmpwpAIvlZykHFmVMgyIRRwpvTaComJDSj1RpdHINKuZKWz9VVOYjJ5OnMVzqJC+755UON/4O7DyFHutNIiTOgcc6wHaXKoNaOE9fi9IQSTGVnDrA7Eet8IZdBYef/0+ITGSPcCks+2D3Bx7sr/OHJNZ66fSQeZGdg1KhRYTca5dghBOUJ75vIaWdYLxzrgxqglJGYBJljYVQyNKMaUq50uqg5gl6aniObItfa0VeRUIFrFH5pUT5SrZayfiwXxLaGpsY3Gl8/8Mg/sBkiYzqPhB+W0k6T8r3TRigeZ0gsfJOi6i5UDMpyx6247Va4oDn1DUOwWBVYmEHSCalcrTDocw4/3V5XzJZrdr13fJMnO0eNsvkoFSIezSbU3HPLslieemGYz2G7vEiubM9SDzulkHIs8toM0ecx2rdKGcZ0XKTF9CIsk+AEEZDvF2a+oCs5/WF0SBwC8A24ZYW2GqMUqvW4lZFovopURvKN+Xw84qTlUsZPbA65d7wkbC3tLUN7C0wfae5F9DhFglGJoxEshB5Z1C3oQTGOltAYPt4cCZS9aPj0+hatGgm6g5QS/IQ/4Gl/yE13xG+evIqPb4746DNXqZ+qsGtFcyfS3o6oEDFjRDnEKa9kobZ9QtN0xAwKXyt8A2NXEarISWf4PWeorOeg7TmoBi7KVN5MZo5AQQcofu0O+VA2+Wn8ps0SCY4SKhBNRC08uvZUlefKakttPNfbDY+3pyXnq4kEFMduQYiKj26u8LHjI0LQCSZXVJWn0oGFHQWxsT1We9rFwOnK4q3G9Kqcj3KS9w4279cqOSkxbZTTJloiZ70PceQTZuJVkH6fOUcT9MqFpQzy/R3S3hJCIhJ6cQbiqGUvCmpKj6Q0gVRbqFkpZkwUgowYIE6ESimEXCkyKzs9U5KpYKtrnuYAq4X8+kd1v3OsMSpcmjCjNxMymdAOo9P1VnL/Vlpy/EdVx8KMjJXhwEh6NjsbuZS3Ui69N5SqqTkikW2eUq2UZxMafte/kg9tHz+LiD6LPTRCkIkcKntwQZ6T/JYpKIEm1bmjINqUO9aYMzQLMR/FC9qtPVf4ErnCiJJ86bjkeGw56RrGbSV1tp1GDwrtFPZUYMrMvJ2bHhV+oQlYehvZjBJ5rhMRCyTnbHRgxKSFe1oUdkh3aaNptMWYUCKIaASKjVZDlbQHrAFj5DktXv1FedX5+MZzns+kSJ13uhly4ONUGpg3dI/GRzjxLXfHBUOw3BtaOl9Ra89RvaVJE/rAdJAIoGV8k2ORuQjzjT9XDwClrG0e6Y7Rctf7ksY4VYGQYLchMXPHYKZFY5aeCEbt8BPmtl8GO7+eWomPrRP05ZUupXQv1HYQm1lZ4zziy+mRnN+MWhEqBRi0k6qdUKniMBidq3Zi4SVk7YxtqOmcJfQG1WvsFuw6YoaI3QaUy4tegkjTfaxCTHM2zUuj8B7GznIyNLTGCXmXxKFJM20da+76lTiP/ZJ725ZxU7E8VdgNVGuo1gEVIsrLT7S67CPByv0Z0/2gXEQ7QKmUytB0y5rRJsfpgkrbgB0YPOeaz/iKZfNLfzAzB2GOKMyj8IQgKBuw1tPWI1fbLa1xPN6e8qr2bnEEfNRsfcXWrwQdGBr6vhJ4HAQKV5ExaKqoaZD8M0BlPKoKslFm5AJZ81AKFeNOlDsRAveO/by/56f2nIOznzGDFS7o0swrC7K2QEaiC3EwqIL+lpLLnDrISAEU3oBUQcTp+PPzMZZKkRInzTgSeVzCqOm1xaVrP8yQm3ysuTTSR3FkgEIO1VocAaUiTeWotKEyohUAUmUwBktl9gj5zKquZnwAnxC2+RqS17Rclj8qRx8q7g6Lsk4+iD2UDgHMYKjkjRWUwGtGI07B1ssm2wWBMSvlqaKTmnCCnNDMASiQR2H1Tyc4JqbmOmg20XLXL7nVL7m9XbLeNKi1EVRgq9C9EKaqNegx7joEeWI4hV0KUcRbw6YXKGjbVozB0FGlPJEuFQ+DmurfMzFsnLE5hQgXJ4cgb/hGEY2WdIExwh/QWhY7w4V51fveIswFgKZHIqkgNL8mEtJYZxRkE2rGYPnw5jH+8OQa27Hi7vES1xuUiRgbUDrQtiOrZhDSovFFWKMyKe9vhx2BjYwW7ebNerSKhbfg0axDU6A3kBRCo4WFbRCOCkwpjulcs+O4OxaaSKUnYaSc4wN2UyqZ0Rv8uSjC87G5Q2DSrTYvj83e/hz1KPPGi3OQtSxUUGgXGZzh2AnR7q5fUinHLXfA7XHF7WHJybZFbQxmozFbsF1EO4nSVVr4ohXYPhhK5JTvFT0iJLWgGDaWe5sFRkVuuxXX7QqvdUE7TvyCe37BnXHJydCw6WroNSalB5SfHBBJo0k6zdfp9/TdUaW6+sSa14Mcj+mEwOzqQF9bmsoVcvALNeUmJ2D+keWYsnBQhv+hcAOISP56z2LO19uArTxN7Vg1AzfaNSvb85r2Dp9ePwPAJjR0seKj4RofWV/lXtdy93jJeKeZHBQFfWM4NoGhlrl0YIVQZnSUtIGJU/Sbzqfk0oWKtbO57R6w2j35/Jr7DfEMaVVREUMelwtayOZfFefOAFOaIDkDyk/6EARxBFQENc4cvFw+qCCTHOIs3TORLmNxkgtKwHSu0SnG3jAq6EyDMmEHScnHCRROw84YmoiuRZ3KVoIaGR3oWktrHUMwrExPHy2jkfUpO/0wle8XblRJYyduwuy7Mqn6rl/ysf4qH1tfuS8Z8jx7aIcgD36UcIroNVFHRi8RzVZXbFyFi5pGu6Jil38yyz1Aib51ypUUvYJ88ii6KGN8LzQch5ZPuCt8YnPI7ZMV472G5rZBj2A3spCZMVKfxOIQaJc2CislXa6VGl43KKK2rFcNLmiO25bTRmDpPlS0ahD1QgzmDDEksNI9XZCKA6c11gQhEpkEw9YS4VBZcJ5YWdEmqDShUlwkYXqu4jeJAbk9J2sqgTEqYKI4Ayh5rdTOVtxzCzah5v/cvcEnPnYVtTUsPm44OJ7Qj6iFEHanlZsnVCSoFEIjUa1qPfVywJhAW48sKtnUl9VQFL+u1xus8hzZjkPTnXtORgUOzUCT5tFVs2Gl+x3y3zo0nIS25NLzRlzQnORpezT3CinSlEqJLCBy0daHaqpeIKe+qvL9O987dyiNoEux0lJ2pZAUmFP0g+X2uBIdA9Ph0Xx8vMrHNle42y/YnDRU9zR2o2juRZpjn6Jg+Zqo1cRJMNPGrELKBnYSpftaYe8Z1m1LCIpPXD/iit0wGlPG/aY74qn+Cp/oj7hzuqQ/bbCnBruJ2A1oNy3ArtFyX1QKt5DvFpg3Q9llGLDbWFILvjGEWjM0Ff0FOgR62It+kbkt7HwmNMBEVCWDo1LZZwwp+itphxSpGqSSoPKsFj1Hbc+NxSn/1+om1+ya19dP81n1TXxUfMwfcssf8NH+Gn906yr9vRZz17K8LZ8VDKDBLTUbBcNCFoxrzaakmYz1ZV2LehcR1VpN1Q15YOceQTrmspnFeP7r8vjsow055ZD+Hi/ouuTDmKsOxpBIhGESbFJe0OAyd9K5242gw9OBUoLBM48lLaTEKcgOoJreV8413T8qzByOMwe/95g+P1QRX1dCxG8iQxXARraLGlt5NquKpR04sAMbU7M0g2gRINyCVo3iALJb6j1Ek4jkkzooyD5w2x3w+yeP8bFbVx5q/J+fAkvxiORmzsTCEITZ7aJBByEAivJbTAd6th4+m0CgZ9XpsuPgUYwYxmAZnMU7XUqo9EiJSvQoOVMziKekvHyeCrIQaisTRqcFtuTrZuSQuYWEZOTjAUq+vfAg1EQMky+jkMCSVqYIEpVHLgwduJ89aPlcke4lFjbrECzboUJ1BrMVidTqJC3SCdnwjcL3aYOpIKaFyTeyyfhR0QeFMoGxMfS1CNusbY01nmUlVSFWB7ZVzdZOEtFZxKNslnmW6inNoefIB+LITRH3RHqcX6fpnPXs9dOFuJ+64wsxcX7jmeud0aczppii5/xUSdPJRtR7y8II0VKYxIYhGMagiU6jx5Qy87FsEHnOCW8gOQQJycrfWb4rOdLaA6PGjYL69aFi0LakVLpo5bmQBHWSRK7KQjkxn4/MEV8pQq4sMOxwfCTaS6heXnzTfUrahGGClF+o7exfMzg8lmuQBlxHVJbrTfd5QO9G5Oz9nl5rdBDnVw8s9cBK9xyqiFexIFEhitBbJoDarZx/qCSloq2CURNqhfNTuiwr7Kl0nOIQTOWSO+kCpvPbGYDzxvK5xnf+5/kGeEHr2TxlkNMm51Z77B9WnP2Es6+dBJSY1ueSJgA1X5ezn5PVDRMakR3rM77PHGXKxzm/p5ySMkejkryIJoaIS8J1g5M1t/eihuijJig1kblzZUNCB4oy7/z39DdxDISLtRlr3GCe+5rO7OGrDDKjEwQhQEpDsnohIDCINiztwNZXhVFeRQcql17tSsnuk9NAcjQDmhHNSVhw4hfccUvWXS1lTaea6lQgxmotfQTMEKmPHXqYbQIafKUhaLQVp0E7YVw7P2liZ0RjqXtWqQwE5sp3E7lwh9WeJ/F+fS5MDkEhsVy8J2DO2fzOlBqmXHOeQF2sZj0bbIla+2DpvaXrZXztqaK9E1ncckSlCtycI1k5b/mOaGTRjxrcwuCWpuSmvZGAajCyeD1TgW9lwY11gDotujYrhgW0kd/beqStRyoduN5uWFqRu71SbWdEG1XGIKCK2Md+2eIYbEJMJHcH7KQULlL4Zr8SQ6uYkJk4VahELblxk8SJ5jXdPqJdkM05MfXH0XDqGrSKbOqaTWg4dQ1bV9GNViD7TtAy7YAYiVYxrqTef1wqxsNZ3jnD9G561IPc7LpX6LVhNBW3+hXP1AcAHGkhk258w9qLAJIbDGrIPJ6I9jEhSUKg3T6uGVcQ6ohbBtGgGpUgHzmyG3PKTxwZM8jzPsAYFJW9uHJQ083mbtpQQyPRXLQRWo+uJApv21ECxZTXHkeDj1YIfE5hBhm7zBsKEQYnMtEuZOlsKWNeasMYA5vQ8JS7wh9trxFv1zS3DPVdWN5Mqb1GnKhxULiVxlPR1Z4hWGrtaIznYNFzCvRNLeXORDmmvEFn5DrB6jtcB7LTk56bySbL/2eDNQ9gzlu+LnBJGwbZkgoqEKfHjESoJAUdc/omVTnkO1elPSrP6YxC5WPdQQA4+5qJlCivVSGi9sdjZhKQxun+SUFo5on5OqNiMBxp3BJCIwRwt4B+rOhchVWB68pz3a5p9MhVsy7CYGO0jEgq4G6qHMicqUo5jkxHqxxPjQs+MV7hqeGIj9+6gvl481CX5+FTBjlnQyT6NDhO40epE9U60DmL15qNq9naOp3QVAueSyNgUsybYN3dwx/RjAkKPg4L1r5h6CvoNHarqE4jZhDykt0G9BCwJwN68LJhaZ1KT6xcaKskneBkMYpeEfxUKtcocQpaPcwkfPWkhhf1Dvy+Y4ozzlhUajqjXOp1wTbpN0yoQCakhKhL06biQaKKyFIR00DqbF1yCtxgaE4V1Sk09zzNrV4muBbnRiUlPRUjavRCGFOKWIkok19Y3NLImBhKmWW+0XytcK18nsDClFSEwMVSERI1HLeRu62kZD62GoV5XTkO255Kh8LcBbCp90StRSyr0p5D25EFPLKTJNLV4vS1euQw3XjnaSu80OtS+m+QSYxq4soo0UknMdhjcRoTEW/06DE5saMijIatq7DKc+oalnooZMLBWfQgZELTxbIwBasYlwpfw3CkGK7GAjNHLY6x6SQKMluoRyBOG/JYGe52C243Kyrl6aoKHUUsaOsrtq4ijEacgeRsZ7JgqMEtFN31yHgtEKuAOXAYHXCDIfYGnCJag92kVX2dUhhjxGzlphqCEmLWhVwZSIrIySlK6FYbxBmwEdt40WJI80yrmMbYEKNijAm2HuWHANompAypvBpSRJ9rw1vlWKqaDQPr0PDMeMjN7QH1HakGae4EFs84CBG3MvhGQwB7VRAJtxINAq0ijXGYNuCDoqsWBBvRe+TIOedvxxmYR/Q7r1G71RdMr9lPCRRBpgu2MMq9F3NKZj8fn9AbhfCFpMxSNuyCHsU0p4co18Wzu6HvHXh2uIkxvXYm6qR2EYEwW8uy2S5QncoaqHuP7lwil8reExaWcWUJlWLbGYarsvb5VhOMYRwNnbelvPaK2ZQmX0vdl7V6TBy6jwzXgamp3BWz4TF7SqtGxmh4ajjiY9sr+DsNB0+rs4jGs9jDIwRpPGNMuZQgk0VYoBHvhViYPWqf8u8ZIvUljXAWyt03zySVm2HeMRhC8sxzukCPggzoIWDGgBoDuCDEF4sQs+YY2uxmUClCyOTArHo3Z8A/m6RtFquIaeKq5G2q2e/kHx8neOkCYba5PduY7qsYZoJK3gTl2qQGImHG2PUp9RLibkoynZfyeUJMJ6R9kIVSKaJLb5h752kMok6bxyC/m4QwhEryvFFDGMB3hmgjflR4G+mqwLZp0EYkX7MWec4xWx1oKldqvg+rLjUo8UkUxLOyvZRQmp4rdnMGoXqhdp5WQ5jdB5nHsIMwzUzFCCFMMLwHwlSXnu+rMUjNunNmB67PPIFQKVyb0jwLcMuY0Kp0KdLnljSanSBaPSrUoDjta+4MCxZmpEtE4ek8d+e1RFSxpCd8pfBtJC6lFG+xlA22N5ZBQRw1oU4iO25ymlVazIMHnCj0Xdi1SYefOWDlftSAjcUZaOuRZTWUNFZW0pOyt3RdxuzAQBwUQYEbDYMzpXHYUJzuwBgDJ6Hl2LVsxwrTpzTnCHoMKdKMssl7Va6LdxPfxeiAinHGXWIWIk+O97mpDabXTX+bwfIPujY9gvVrmpR70dV5KRCVpJNnaEhJheXNPP3/vM84w6ecnb8iKRnG3ffm1+ykfWffJWt/TA2WAkrl6pkUHIxRnEiDlOGmgHSeCpP7en9dzmuGnThSUcjU8pyU829Cw9o1nI4NakjqoI/CISgQUr5Zi+uUnAKtCV7hNPSjpA82ri4VB721VNGio4iceBQtYylZm3MHSrc65XfEacZoWLuauLVUKV3Q3vPoPlKdOkznJFpdd6hE5KOuwARUbWZ5UwpBTsqDAgdVz43qhEO95TFzyqHudiLFEpvkxYMkCBMqhmBxXpf2p3kDVS7C6MB5GB1aS8mhOuOxXozlSoL8e97wSw6qvGbKOYWo6VMeWkqgKjpviV5TcmZBNiY9etR2kM1faymlVIpQW6hM2nxMWZjsRq6tSpr8MTV2QknkmaOzbFGlCCulJnwld1kwE8rg65yGUPh2WjP2b1xnYKjkOj/TRnyDQPJNJFoh9ZiFQ+vAoh25suiojOew6mmtsOj/3y/wesxrgye5ZlEQ24SatRPIfQx6WrjyohIiunOo7YDdOqqNFcRrFIfbWWk21UfL2tWcbhv6TYXtEwLmSWx+w3Co2D6pcIuIu+qpr3UoFXHOEJyWHge1QTmJlHVqvZsdboLmzjOHdEPFybWWVzT3St30wowFncmE48zRcQtBJsYjiK/oeN0rbrOwIzfaUxrtudkd8Mx2xXa03NaHBFtRGfD3MoonZZPKKfSJ4e7q4URWns3sJh1yJdoPIab1YOGwjePG4Zqr7ZYD23OjOQXgw+vH6JJTojtNdarRPdQnQAAzKFynCE2kr2pOgNp67owrWuW4GxbcC8fcDfCh/nF+9+RJnrl7wNHtyOKZQLX2mLXMvWjkPrCVoCRRKVxn2CS9l4UdWdqBMRieaTxx0NLpz5AqC2JxCFRQkNHcfUt/jyqm181elG8qnRCBc957v7bRz9fi/D6+nxOjkJJB+WXiaiXuRa6mAVXWr/3yc5UCnelEJBWhkcBmTj5UAXRGDWZvyWiB8grdaKIH5SLKBTmRACoGtPIYBdpqqkZI+HqE8VB6UvjW4FK1UW4KZ7B0oQYt6YIuVKl3j+ypIWqCUYxKRJCq8QqV8ny0v8pH1ld5+vSA+q6mvR0eHUIwh0/iLHUAU54nONHcdj4WVMAFiYRGZRmVyJAaNEFp2KtFn1uuQsgbs+ji6+T5iPqa6SKm9+IMbEeU96hhBJe2QJN3/+zipSg1l0MrqROttWOVuAMC70nToyER1eb8gbmFmHT9S1lM3kRn6EAIKB+IIYijkCGuC/awhTcwzfy5yuIcHZhzDnL6ICtMumCkKVFBONK5eJnoatuD91BXxFiJY1DbUmKZN3Q9Bnl9nN0gKjkMe1GL8lKrjlJlIcxM+3K90s0ZKl1q832tE8IQJ7JPQmWCncrb3ELhFvKZbgGhNoQq4pYWZyPHy5rtqsLawHE7sKovRgBnauUci9JYTtFkvsbgkyLkGXQAmTejQ42hbM7Kq9nCIY7dEAxuNMTBSESfxiMahdfgW4VbRfwqYA5Hrh+tRaWzaxgGi7NG6qiNIgxGFlUPpk9RrwW1NnS64d5COilWyVm3KpQ0TT7uvAjnlIGvYXnQ87rD26zMwOP1SSlBtdpzMrSctAvG3hB6U9CLjBCgRLho7C9OuliPKc+rFMrmSDKibaSqRFL48eaUle15oj7BR81T5kgUSWNCKAcZI9PlDUfCVB8EVfGDYTtUpftqFyu6GNmEijvjkrvdAr+xVBvEGdjItSbNaZ0iSzOkjc4JsdAFjc1KoMZN5YdmQul2kPZ55D9DEXbn24RwFlJcIRucs6M8AmfgXDuDDEy5j9JLIm/cOpbKGUoQAdqrtHfNsf/psRAMo2zwslYxORr3CbElhZpQOCvORFaplSV2Wv+1E96D6QOmltfoQRA45yYCtXC6DIaQFH+FxNvFqgSgpQV9sKBJAUZDpTzHbsHx0NANFaaDahMeKvh8CA5BytckxnQep5jOO0M4JNhWqf2uTpNIznk2F1fIx9/Fik3iD9x0R3x8uMKtboXZTsQp03l071H9iBpG6R8/jETnJVXgA5NI9eQMBAPRRoz11FZ05bMyVK4Bnde0e3ZLD+W5Sa86RjU5AnkDDXFqY3veTXVBdh7MPT9Ok0N9ZJMqEpjKy12TzsUltnrvLJSUTCwRPlBQgdjUxEVNrC3DlRp3YArZUBCAQHXiwMsoqtETjZHWXcXTno1J+nyVNgAVpRYfNb0+KhJZTRbknKfWLpZJU4IaLdBc1CpVnsg1t5uJae/b5DAsNW5lcQa2TeCZ6uLhG58hfrK8qWzkQ0hk3AyBJg5BVDKHcB49eEwfsRWoQTF4ySUPqbvnECzBS3RX0KcSJZHK/QQZ0Xrq3R6ClmYxOSpUgqKESmBYcQgiuleYtcYry+lBw7FbiPhJNEVYSVnJvwcjiI7KKoiVEAnbynFoO45sx41KHILsFGkVqRvH2FhCJZFPLqMrKFVKG1yUZZ0XrSPRpXmbIh6tRUfjsOq4YrdcMVs84vz4oPFOY3oKV8NuJ9KZONEKs9G4ytAtKu6NC57RB9ywB9w297gVltwdF2yGqpxT1EpaoieH2K0MrtW4VJGRLXf/A4ozpnXAZ4JgcYrTf+fLTv77/fe46XUP8hzP8TnPx2Y8gHNtvnToGVKQHLVQyQnmklcVJOUk12c6iVKqGNMcd6qkzXRpIb37pVGR+tRQ0nGZ7xS1LpUIkBy60YPXMqdrQ9QKt9AFOfOLiG8juvE0xlHr3SZxOSUIwgnzSPv0hcmqq1Nb5RPfAnBnWLDua4bessoprYfYex4cIShrZFLBKpCKonTS0qQ2nRGTHjMZby7+kk92biKxK4SrTHLbhIa7fslxWPDB7nGe6o64ebqiOlbUx5HmJGDv9ejBSeQ6jOAcseuIo9zxyoqboWYIQb6ooY6smpGDZuBqtU3NYgZqpMmO3yPOzHPu8n9puNIHi/c6pQpUUWVTozgDMYRzL8oFle6e20VyUOB3hHFiqejIE2uMPV2s0Gky9d5Ky9yxRvU65TZTXjPxBKIVgaWwanFXG1xrWL/CMlxVO4tRe1slTgewdajeiTPQ2pkXrgoKQfaoU1yS8+dZYBEQko7VgjRYnW5INTlesFdDnG5kqwgmow3yZ0EgVLlJx2VKTVhDMOm2+H9dzPWBqcSxTx7+NtR0vmI91hJR6qnKIJp07KPMZb0dqU9lIOxWselrtIolxdPP1AkLkhAjvs6aGBCagFp4bOVT62iRhPWjTvlOQEepmV5E1AjVqcJ2Eu03dxW+M2zblqeePCzOcqWkVEpXgVDHhAjIdfWNpHV8GzlqO15RH3PFbnhdfTOhcCKDfada8tTqiHE0DL2oeeayRO1lw9aDKm2bL8JsL/eKQ+2o/CmEf3Kt3vDK+h5XzIZXVPcYosHqT6d3gsYsTxX1Palsqo8F4TKD6Jv4XlApMAy24WPrK/ioODA9V82G2/6Ap7eHnG4aVJ/QT6tS50crMHIjCFcuP8ybU0gdXLNgV2tGjA2MVSBqMwtKlHgP5yEC91t3Zihwrr4489780vnidZFOwYM4KgrZczKSke9paexQEC6dSwVz2WB+777jlIKfnO5SfveEs2NLIj2HOkeX8mA6hTuV6ppoDKGS302nUT5xaRrRn+mPNMMVQSrHw0g8cCyWA0d1x8oOs9bKwiPIOgMgejGNHrlit8VRyCJux8OCPlie3hyyPmkJp1WqNJoFcw9gD83SEZTgPj2wU12syQiBiqldqqAF55G2dNIekA0rzmDuqbLg1LecjC3HQ0vfV1SpBEsPUVrEOi8wtnNEH4g+yP+zrnIekMx0zyVXBqzxwlJXXjQKU8ST6/MnJ2CPkJfUDIWZn7sqqh2IXSZdlBvzEdpzKevNGxsBZ5CagEBWpa1x0FLTnKH4VE0ATNG9TZFcyuW7RfqwdJO5TRJmAozVkroxqmzEMlZRLtHMGS/fEyTlIJ+ZNBwyvqeUOABKcD7lQnEIVC7xnFn0CqV35x1KlfSGcgblJf0QrJqUyi7IQlRnmjb5BPvnpi3Tcc0PPM1d5wsxSaUy2dHntqqSRttRaztTZoU46+m+nLrJqcIVKTa7NyBdey8LZtQIuS+VplrtC/qnICm+7St2yo9VIbVmHVkpScut9MCh6eijyLYaE6QMdQ5Dz2HssDc+L8DyMqQ9hBB3yvKUkuCk0aNUHKlJFDwE0dXPqo5F7yEkeH/MPIzEv3CiG9H5ilPfCOkrNNL4xpuSOiy8mJTmKs5ANVN13LOCzqi4A7qV8TrPznv+Icb0fgJE6rwDfCF2v89Te4/yYnnQaqYqCSEFHFGLoFR53xxliKDT4GXuQabEZIcjO2VSiSJCbHltV4gDEWpAKXwFxsqCppz05ok2idHZdE3rhCpUAVUFrAnp/vC76wSSUp/vPRI4e7L8Ncjek5tjdc4SRuEDqeex7zywQxBsgv997qcu8Ew0kVhF1NJha8dqMfDEwSmNdTzenHKjOaVSnkMjJRSV8qUzkywKfSoBcyxV5F5Q3A1L1qHh//RP8vvbG9wZlvzmU69ke2eBvW1Z3Iw0xwG79Tv1sjE3VwjpSR9kMVXiLUclsHGo5cLG1vP4as0TixOu2zV1UkrcJBnBAZOa7cxSB4np6ZFoTy5ChXOimJi7upk+SPnJMMI4QGhKhJtL7i5qcRs4m1stnfZmkFP+f2DKY2f+QJ5Q66Fm3depnjylDMYg5yGEC7K4Ur6BxkNFd0Nqy2Mj3zMcGWnSM0SqTY3pV6l6QBY400eqjdTY241Cd142soREqHFy5PxBgzuoZaFcGMnXjRG79bIQQ3EIyiYaJE0BoKwWlUhFij51+g65pSxMHIb5gvECrYtVcbwq3A4ECBKJmgT5qrRpSz5z1ufCi7Nr+kC0CrNVbNc1IShOli2HVc/ozZmFrlQZ5OcCRK8Z+opjEwhBSx+QTidmveSuo434VipEssMEiUughEx33LfU2rG0A1Vq7qUz2z0tnCFpVGTxod5bTn07lXoqx1WzLgTjVTVQWc/WzqLa/P58ac3DL3D3M7uN5fOljExSTCGo0uVzKucVDk7nK4beQm8EPRsm9rggVZScvz2VwQ+V4enjA8ZZVH/sWm6tl/hTix1SQ6eVZlwphsNMtmViy+vp2vrkTOZKk5yujDEBtZnUnMdtFg3LIO7C5vLfuDPnCzow23yfS4nwopQKHwhtyN+1v+zFlMNPpSPntkSeay/kS+dS6WhkquYp35WIp22U+V0HqII4kKM408EYkRavBFH1g1RWKSfcIN8qxqXGV0rKfq9IqsBcGVmtOm4crLlebziy23R/+JLa7ah2NG8q5VjqAY/i1LdCBg81N7sDtq7iZNOmTqNp/W/U/ZUVz7EHdghKW+XkGeeJFlMuvl2MHCx6ri62fNrqLgszcrXacN2u04bfl85NuQNirs1tledQRZbacBKi9CvwB/z+9ga/fecV3N22DH9wwOppTXUCi1uO6jSRcEpEGXOCbXbQAZybIsZEVst5TbtwvHp5jyebY27Y46KNsI61yBZHw0nqWCf5GrkwWQ2qCxVrV6c6bI1NCnFmiJjOoQdHHEfiKMeZiXK+lqj6omxfm/88K6S2lMceZpr6fbC4aNi4ms1Q0XdVIrykGt0hETWNIbY1MTkFmYQzHkTiYwO29lw92lAbz1NHR5y0LXpUmE4g3/mmVZ0o2lsKk8hdNqZIS0m0r5Qq/A+/rOivW3ylGFdy/ewGFrdI5MU8B9JljlHKTvtByJxNRa6MiKkaQrkoqZCEROgE3T5Mvu1BrstIqjY4py+5JmJVQKQdYlmE84ZAjEQf0KNU0EQr7YTjxjJExelRzdpJlQIwKavNNoDSyMgrolN4pdlsGikT7gy60yI7bUGZSCQgfQ50yo2mYx2Eha57xXqoqcwCq4V2C0h0XyWn0Ezvywts70VQ6Zo1JQAY6BiNMPgPbC+9CmwoG+GcoJiFvS7KSgVM1JJ+UomT4ieHoDT8Sihh5yvcaFCDQg9xcgjSvqNCTMGSiCtpLw7w5riVlCJSlXEyNqxPW/RagohQK8Yl9FcV/WNJnCpM5591DoDUSCfLv+9Jbu9wmLKzs3ve55XRZYdo5wUX6Bg/lO0jA7NDK/wHEMTLSI5HzY+Z6bW5VXJuowzTe0DeF6MieEVwCSlL8sg7Dk4bsK1Dm0BdO2rrGL2h6yqCN3glZaZmUEVQqiANSmS7RQdE4Q5gvBKIC8+NK2seW6650a55rD5lqUWiPXMDfGK++1mJft5Lx2jZIGjf1lfc7pash4p+U2HWOrUzn9J3D2oPnjKY5WlKIwiNeFylDnze6/ksbyAwdTvMEPZuueGDH/mcw6C02ilfY193YP6e2e8Z7tT3GbF8jF4orM8pBTxnucuXPMf5XNAN92yOANxfm0D6FySnwAvj3TlDCBrrZwpdOe2ho5AEVUon5JpbRWl8tKxGWjvStiPrZZ0IWxqdyoHyQidVApA725X5RFqkrCauWrCa8dDSH4kWvlvmPhARt80yvRHda4EF3e657ouIkDvvpVJKUuYhQplPF2WlOiZVf2Txp/30U4bys7xqJhbKm4McswvoMZSa9OhUaa86TzmU+uv54gkpSpQUQUgvVE704EM+/yT4ImjoLnSvEvimAtLILLHdId1HKV2Y14cS1SY/x4ddVCqn5LIoWdbnV+rs907Hz4XdM6XHySyaVglFCUExpmqQIdV4d6Fi8ELA1CFzX+JERoTJmWN3U8615i7oVNZbyed4NaEfiTsSUk8Q5dkpAcxriyAYaofwnPsRTA4UO/u7pHlnx1neOL3m+SD+OU1wYejA3PYcF/lCJuRiRmCXR3lOMW30sZAOKXNH61ic73yLea+JNsh7Upvl+Xea1tG0IyZpm9TGY7SU7TrAGzu7dsJF0jH1RzD5us6ur42yXupAbVKaenaiueOsQVrPzS0L/AlCbYuy7Has6MeKOBiR/B5V2Yce5to+eMog9wGvKZKXvp0gFK2njbUIp6QyK1SYutcxdTvMBL0Kad1bpauWlb0OTM9R0zEGzcky4JZC3HMLXQhHupLPVVqjjE4lHymiMGeh9AKj5vU2w2/3CT98+ltxZKIuz81z76XOp0zkfLdplBFmPrNcVYHzPok26RRI2mPjG059w+1hxfHYcrdb0G1q4tbQdEkspZ+VDboE648aZRS6NlK+lW60unK8YnXMlWqL1YGPVyOjN2w2DWNnYVTojZnguSClN7bLdHhE68AFxmstmydrXKtYv1LR55RE8vDNqWY8EOLj4hnFQoEZg9RxD+k4rUmOTEoRhIAeHIwKkgqgipFgKmItcydU0o3yIqxLqSevNKMy5TkXTFG/NElp0VZSxherVAVRTfOFcURvRlQEu26wpxoXYTNUbFw9lS2mTbg0LcpIA5MTwajSyoR0ROwRaPsQtEm6/TrgncEvDL5O89WBHSOmU6y3NUpFDlJfeE2kto6htrg64mtdeohoL+/d9DVP94cSIYdFqTLIQUPuGJpzvVmtUa5bruC5uBvGbAUNjEZJGZhCYNZO09dSGXDHrei1aHRsQs29voVeFBkzOhaVrInizcWyuWonv5tOoXqNs4Z1X0sDqrEmdgYzUD4LZNPwK+FRqFELGjtIK2nl5PdxsCgFvZM28y4Y4S+l1Iw4jJTxipl8d44zUJwRxfmQ+sxUVI9m4382O+MMTD8qdXuUhlMBk343RpxLcWpnDlXislXWF8XL3Jl19KakYXyYRIIy8rCsRo6aTuZ5UhPcuJqbOtCPluOg8Z0mDpIKNi2SMggKbRSukYqmnKpWrcc2jtY6au2olDTy80kTpme38928Ai9rx3Sh4ta44t64kGZ/d1eErcXesTR3U0njcE5A9Bz24CmDnL9TEPKNUEleRZm4R4ZQZbMsTRmUTsTCVNd/Tr18SO81KiRGpbBoW+ugsJhF+UzXCmMUGCX5nLThKvIgmGlBfRY7L2Lb/btOZXpnxYqn3vbqfDDgft8fZz+P2PbRgUmwaIrYNk7kZ7dDReyTJn1KFygfS16//KgwEfkyTK8DtfVcqbY8XgsMbFWg85ZbdsW6rhkGiw+CtqggnSdN9p6VQsVUzeACvjFsb4j+ffdkQD3RSYvpdLO6Wtj6ppeURHNP0hhm48pmJE7hHBaKE1HRJ0JqiEAlzoARbQPfXMzGk6O4MZrye3YgM9GwEG+NcAmCiUnwZDrm6IOkbHTmpwjMPY42kUBVik4nQl8uOyzBa9owSJsGIS8auVQrLZpW2rOOqYtlsArtRUUvixUNzjA4u4MQWBOE7W6nqCgqxMnzCudEynztmlSPL4teIRQrUQvdaRLGFBkXUuEFmR5SL5XBoL1GO4X2oAdNGFL6zNdFIW4TavrRosap7W4uPQt22lQzIS0TPCdERzOOhs0oMtMqNaHaac9ugSqgrGxoSmlwSPSfrp/30owtC9mUYCQ5BWRk4jnWGBVnTsCsAuvZgpRndQoeRXBTJu9sTuyhA9kZyPdPlaJ3mBwCk7g6CoSrYjxWBVo7oom4qMtcdqmMXKupbD6XoJrZc2szsnUVRkW2VU1XRRn+RASFmAILScdlYmKoQFfSHju3jbd62lsEPdurwEtreCYSiuNgiyLheqgJG4veiPy33aRSyvuJUT2LPXSVgUCBeeGZLkyeJpmpbhMqMAYL2okgC6kOPsHvOTdnosJnSIdY9OWXZuCw6ul8hW48vrHoRMLRTljs0WrwkWi06A4EyTvnPT7GOUM+/ejpPJ7LNOG+g/pc0suph+cMtnr0HkBGOuYkwmd9fUz9C5xldEai+KJJL3r4ak+WGCBLFutUiSDwtTSIWuqBI9vhGsPa15wMDVstG4DyIvlKlAUwxNwFL6WQtpKEDVZKc/wCwtJzlORu8zFvADdoQq9wxxrXSpOgYAW9mF+2mFTMAHE60vlEY8CAby3jgcXXinGlcRfE7xgTrLZfYTC3DJVbHWQTnufgtRaUK5dphiniFnb8VKkgqbs0ppYpgi25/CS165jVXE/aGRJ1BYwJ1Fai522COWPaZEqVw6hxTr43N5FqraOvnCyMFQQn92BOEzlnOBka7lYLbrmDncqXk9CWDU6uFxOpLp2DfM4F7jq5C6oXRURt5VGNgqL0zrL10oel1WOJHnM1R06FBDU19NKe0nAnb8gxAR0CdqTz26sqkZyzwtdSk65MxHuRgi9CO/fZrEtX0DBLGcTsrTzLEpeOKcPphSTM+d/zbKbiuTVnz8/OSxVMX1SgfkEIQkEF8qZfW0+VKoryfMrkXasCq6qnNSKIdWCHsslmCXHZcNXO5r+yPUe2K5+Zq8vOrP+zYxYdAxngMEsbYCS9rrXo8lgVpio8FQgJSZ+nhPJzMAWhfbB03pamZqrX6E6c/Fwd9nwuyfPgEKRJr+TkMsSUyS6jl0YNISrZHMwgkJZG+gMQMKmX6YhEqPmzNVCrwEoNoOGGPWHT1FgV+IODa5x2FhdhPEibx6ip1galFLozsoACkhTOcqpRmNpMi0tIfc61noknMTUDyqaRrowe6Sp1XtObUna1fxMVzf70/bmmPkx9Di66Uud+tp8O8bNUxxgN67HmtK/puwqzNtgtVJuI6QKmT+NoMqyfZpkPqfFOkChz1ISgOTA91+0pS9NzozrlxEu56GknXbdUcjggNS8y4JZyTe0W7KlG4fGtNOEZDwPLxzZ85o2n0SqmUjvFndWSm+0BQ2/ptkvqu5pqG6lODbozks+eVxxkcSjnhWhYWeJBQ6wMw9Wa9RMG3yj66zAeXMzilhXF7GyByvXFkCJrJVFCZUXBc5ujiUqBNSijUUqJUxmC9O3o5e/jODXQ0TZIfrmJuIVKSoUkciJJm0ChRjD9vogWoCLGRJrKcXXR0TvLujkk1MKgNmPAdgG71dAbxsoyBCEILszI1XaLVpH1ssW31cQXCRIl+63l9nqJ1YEPLR/nJLSlDPnUt6xdPTkEafHMP9Gocg4XZXpIKYPRoJ0lOrCp/XCoDcdDw91xwRg1jXZCvPWCnJWSrkiZv9Fk/ZGZ7HMS1lIBCHsoYhqfqCkEY7eKrA7E8V2nICl6NVWfzJYUYEpZplbwpbdC/slOSMxfOP9+NXNaUhl5am2eUYAHLiWMF5jOeY5bT6W0htYRa72kKq2jtqKvkREAmBxxqzx1QgWu12sRodMjV4x07cyl5ZBTwnqH/3aoOw6TVPeJb9mEBkOgNlcnVAt2xhRyml2qdTKRPSZ0oLbCRai0xyope9fpps0oQLbihCTnT2Tma+70S+5sFmzWLfZEY9cJHRhS4JB5QA9hD4cQJLimtM3c+zIh+otYjzCQU+/6PcIE5PIZga6rvd6BWgVMjEXedGHkIqsqpBxvWiyyTKQioQKzu6ZsyIGpqDp7blCIKHuWv9szvS8L/+y8bu+5+wb/e7Wgc4/9k5mSyzLQ+xaiFvW1IAI1Zk9ciTn8Duy6wXGKfsIEg8vkFijYR102xEiONLNjmR9VyZtHI2VwIUOxdaStRw6tLJQucThcNKwbyWV3dYqMz+lEdsay4BGI1LLRSUBHkCe3jPiD+6NCD2N5rus4qV7uIwQgqFiGNAX2F73zqJTAxjObb+TE2eelhVJgbJgT3PL78AkVmDVA2s8vGy2d9EJUpftiIQdmklygqBsK1O9TS16HNl7UDo2SWCF/h1OMzrAdpR6/1QLXVtolXoXeRbjTPTrfCAvEfRE2r06KcZrzTkiqzotqZ5MkvXMDMzVDCEhDl9MGSmUtfIlHdshpaZxlUd87lHydTCyQ90bFnTEoUDm761bIG36uSkg8hvydz7rGJFT2BdnORfskWCHgxsJbM0n7xqaUQa3F2csOQa2l1XilAgszstRDaXMvHXYndGq35l/4CCvdc6i3s3Jtu4P6xaiK83veWBSicOL0aC3HbJOewD6p/TxUd+4MuCi9DzKxNziFTf1D5kqlMl4PN7wPrlTo1PSOPFG1YGbRK6nDV5ExeFzUDH7Sx5cT0uSEmY9aoJTEJxjTlhtkvEQtUHs6WxXi0RMHp4zesGaBW8mKZzuEwBQi0eoiSqOskZsxxOQQAHoiWmWUQ++L1Zxj2YOcm1EZVYhFV77k+fLFyBHp3gaVIcTnyjZclAXm7Y9Dec6j02ScHUhIm4WbQcS5qUeVpkotYxwrK+WHOQIatbDP0+KZ1SbHaDjuWzabBn9SsXpGSkdzUxmQjby7qjCDIeoW09V01xV+GaAWGFCrSKV9cR6HYKmNZ9Rmxmyfr5RB1BH3PTVjiJUlrBr6x1r8QrN5QrN5VcTX4B8baQ/7Cxl7l1IGQSmCmiKQXDKmywKfFjTjJ1ayUcTKoOoKKukZERMCliPAGKYKA2OEMBoSQqD9tDioQOkcmWWcYbbZQ+kUaXRgaaW7X7SBYI1IEYeYGrfIphlT179GO3SMHFbyoU3j6OsUzc+6KOqNYWtbbnrN79RPcrU5wqYFeusrntkecNo1hMFMteDZF30ETnSsZT5Hq2fCQpQ2090oqp0AVnuGYBkGK1wKJ+MWrNR5jyv5XdI5wtGofJT8P9N5hKAlNeezcEx6Pq2p0abNTQuxs/BCTC5ZJnE8XIosUx+J5DWJsxLL5qRydmw+jjlV8FzjOEMK5PX32VnOiRNekJ33PWr2mIM5LQ6ANcLUX1TSZKs1Y+pomtZoJZ1Nq+S0HpqOA9PRqpGrZlPW96FA9ZNWSC5Dl74Coj1z1y+541bcHA64tV1y0jX06xp7aoQgvaVIWRcuBzOHIDl9dsYfqJSn0klK/pwM7xiNaI0g60fvbakKGwZLHHVJ85ZSUwXueZCjH9ghmNqFqh1nAA9RabxXOK0Z3HTwfRK8mZ9kdgB82qiGpIeeIZtKwaEeGaNn4LQQ+l6zukuMio9GRXdYgdK4NYRK433Eaj2lDKJJfQx8yRzMkYGYyiWfb4XZHDFwMTWnmeUVc5vaEgrkCoNSkiTR90WVuOVclqAxZz2N7BSYVM+aJ7xRk659RDaYvKDp1LVLuSDEvOwApJptYbQnpb9UWuWcXO95eVkfLOuhxm0s5tiy+ERkccczLjTjgZTouKWivyabmG80etR01xVx4aXkZ1aa0+gRqwNbX9FYJ6VgZnbDkaKkpEOAD+KU6SRI1NTEyuAOarY3LG4Jm1coxlf32Mbx5LUTnlyeXMh1ySmDSk+pqDHlAHfSBolDYNLiX6DySqOqSqolrCkEycnxlDSdVlLyiYqMTcAvlEQN3cwhSNohIpyV5kWWxZ2ZTQ6B1eKMhUocFBU5U/YYo5J24dFzaDvZ4OuRTS3Sx3pMJDsnQj0uWoZR80f2KjfrFbX1NMbjgubetqXvLbHXpX/93FedQ7EXYaFNDoFW5Z7NzYp8r+h7y2ac0hiDt/jeUA2TTr60lRYdjlAlx2dMTY+6yfGSL4LgFb2TRbw4dUnHJRrAiFO4U8apEvrl5XV17WmspzVCui5OZd7o9xCkncBj//85m6DOgZb3/n8eRyCqePEKhfezTCTMargmlIqBRTWysgNWe5Z2oEktzkGCt1q70tr+ihGJ+laNHJotFT6lru1OB1g4y+7vQsU9t+TmcMjN7oC7a6nK4tRiT2VDtutItUlrana2o0pOHSgrx90Yn3oXhFRV54oDcsYpCEzttxOhsPOWfrS40UIigZtB9pYkXyDclodkCT44QpDZs/mA45Tbwov3630kpKYtIpGa6rBzFLMfkXI2v51fhwrUeFa6Z4yGhRlorIhCbK3wAKTDHhL95/RBTJtv0QKYoP/cNGZKGdxfgwB22wk/m50HA6s9SHJqh/ncn/d8bb7ZZ5t3OiwVH7O73SA5MCUHPTv+vWPNHSLvJxCzd14hasZcJ+sMjFoU3FKHyqhFax9LUTwMSGe+YCXSVZWQ3Coj+UE9c2AAgcy83s2HZ32BWadJtEmpCBEmCq3FNzpJLit8G7HtSNM4ltXI0l5gshpJj+3PM7P3/xyd75RXaTU5k+dJMsfd96t8jTRCo5lFb4U4mODl3S+fkApJ+QQcyWHO90rY/ZkT/Exa1IJ2VMYnoSO146RpLxtlrBRDXxGjYvQeZz0+CB/COyMNmjJcGGcRzwXfN8HKgUU7TeZCykvOlgsSLeZGVHN56BxYRAOxgliltFAqk5SxjlPk7hUEJdLHJa+fzzGx9wM4n9ZOr1MH2dm5Kya4fMY8Lxv6c+3NzxHNl/lyXpBeCJ+fzPzA+ZYPL99TRkmkbbX0oNEzoh7IGpflfnNTt7zZ56qrkbx+SrA6D/om1NOyCTWnrmbjasZR2ofPybm7ayZJ5CvNEyOlklWSy899fkDWZr23sGZnJCPRubmei9KIzgVN8FMaq5AJyamssw7/c9kDv9xsUkSTvVUNcVRFaMFT4a0heI01gcEalnZgqKwsdDA7eQ0xFDXAKgmV+ATvV0ryrld0T6scS93zmvY2IWq2ruLO4kjqpJu93LMxiSCDLFhpY4hhtgKWxXYmoLS3QmoVqGGXYJgmkMDgolTYR5sWDTWxjwvLN21G3hO9JzqH8p6YYOSHrQ991mszu8NzOWdOE8h4yzEbQvm78DNG4WjYUVTiZtKwOXertAgLkVqrCjkvLQ4ZSk6Q4uQDSZnWM+MBz/QHnN5bUN8yNHcUi2dG6ptb7EGNHquk2mhgBaGG7kAqWMZrnsMrW1bNwLVmw6HtpEIlndPa1dxeL+m2NfZUYzcihWw2I2rTl66XMig1cSkEwvFqi1saumuG9auSyuIrO97wxC0Oqp7HmjUHpn9WR/H52MQhmJNWBR2oo8OZxOPILP0qbVpWnJmoJhnhqQwvidREhdZpp7chdVaTLoWQFqns0LtEdiM5smV1FQfA6EBjXHku8wf0KIqVZoiJRKpEjjctqteqDT5qHlts+NihwxmLClJaB6B7qAeFGzVjbHBVDSZyUqX5NGiUV9gTjdmKBobtY1GyVOHhFNeey8YrVb4w5FbNKsQCvY7pvt6OlURm3pRILHfZ9LUSQuxVD3WQv3ea2EG8LfC9HiNmK96Dq6U6w3shAWov1yV60ErhTw13jpeSej2pUV0qTXTTRqNVLOVqgpZJVYI0xkrnsR/dqzQJYpIcP29A0msKP2z+p3mfh3O9hed9GZ63ZYGhKon7tEbSKAsjYzJvPZ7RgSqR9yrlqNNjpXxBVkPa9HO60yCOwolvOQktp77lQ+vHeGp9xL1ty3hai0xwpybEJ82LqCiKhW4JbhWIbeBg2XOt3bK0A0e258D0VNqnrrm7wZocv3DxsgjR2tVSajgKCTwmmWLpvMkMAReyqls+3Lg+uEOQYcYAWfM9WvGGg09ekFc4E+lGS4iS5/XJs5lvvD7lsLPAzzyqLQUWClYq0MZAFQOP2xNO65Yr9SGqCYRelzrpYEUSODPho0kRr9aZ6chckjZmdcWSZ4o7EZtJNWs6htlzMSEGmi6qCWKae/w5qi7ef2K6Jz16fJjVyD/oyD+cFaiMXXGofbXF7DVLbk1YuPocrXhRgiR5gDO0Q+V67LgTcWS0pAuVNKQaW+LWUp0qqpNIdTxg7q1RIRCNQrcGPcoqFmyUDboJmKOBq4uOZTVwpepYapmAfbDS+MqLxLJPIkq2j9guoHuH6kdyl0k5WU2sLaE2jAeGcaVFU/xaIBw6blxd8xmHz7CyfVk4Ltp8PHvBdarzymmDyngopYcqNYOapcKSzUljPpHxhGQFGOlYKEpps/eE6bGAdBlEm0WeVges8nitJvY0abN0IaWTJG0QoiroQJNKAI6qjqp1jEE0HYKV7xM4M20uWpf24yEJi2VtBNsp0VoojYNkfJ5vGdX9zC0SLOyiyLxmx8fLmhZDGlsVic4yOLvD5I+aUleuFk7O2VgCFkIWSCNpGyTUxukS+Wf/TUWmVsy9YtxYvAbVaemWl8esTGWJfm1Cy6wOKB0JyRnI17NUJMyi/n1nIKbnd56cOQ/lRfvpmvPWrkfpFETI5KvMdQFKEz2r/IxTIUT0bGWdmzkDJiEFJlWPzZFgHxVDCp5CGvRNaLgzrjj1Dbe6FXe3Ld22RiXp74nVL4FesHItQi37om8htgHVelbNwFHVsTCjkOW1IBdjsAWlKOjPLGURosYFM5UaOisy2mPucCrOczAT8TZU4BcPd9M8OIcgj7GCOGZUXhGiNJcJLokCOZH+zG06cwOO0mwnwR4wpQv20waGyTHQCioCS90LGcQ4lAmT8EmZ8BLJCjoQJog1yRpnDsF8XVZ7IUce/AsBjGdd+/btovOh+2mY8jwBo1RBPOYTLHc8nFeAZALSHH4v4kOp8VAp39Op/XDp6ojArNHQRSsNXPoVt7dLzFpTnQrhZupgKL0EolZUm4g7UeikA+7R+N6wHiQ6u1MtaLQjoDgeW4ZgeOr0EHdSobcmEXkCug+owcEoE1QlXkm0qaVuUiH0VS4DCuja01ipZMnOgL6gi1PNBEf2HYJM6NRK0AMbpCaZlGsMhhkikMY9ET2zHgEz3f08l5UWxC5EioxqvqbyAopgUfkeAypDmWlBDVGjk3MhMtOaUJmiWCqHpQrHJ/8cVR3tYpC/LQx+oQshOZOeTKfQCc7M60pW1zMdO6JYwppXRanvomCCIZcu95Hah8IJkLbDoE2kNl6AmChIjPKq8Buy/n+oENW5ZpSq1qhKBcZUtSBoKhHhepDGPvEOTNKdt1uFP7VEJb9nhAdIaRjhaCzsWMRsQhSp3VzumNNFO6kn2Ek7yN6aetbm+ZCiyh2Le78/SFriUVh2UmY/ahbMZRGhSk/zECZuVU4RSLStC2LaEaWnS+YQkBRooyIwEeJPfMutccXaNZz0DdtNg+slMje9zMs8f1GUQNUtSXLrEb10VLVjUY0FxYDcJ2PiG0lwKsfdKIdRgY2vS6vzLrU770dLzOmKGQE3zyuZy1La/TD24AhBly7AKF2dohKIVyX2a0Ryad5oxkE+dgym5DtyieEcFskqYKgJUjVKUaXfc1vKOgaeMCd0tuLJ5gp16+h6Q8gd7DLD3OokopKblXjZGObOgMo3TSzsbhBPsmyWRJ4tRjQqEpJnqWaTs9j+f8/UGZ19zQuxmtxDe+7tUDxg9G4pS0An3QWBzOyMlZuVznRq/qNG0VBQPkoNfz8I8lFZVGUlakwTMgRhwG58w1PdEX9w9xon65b2pmb1lMeuZcOOWqFGj1kP6F6z1FCtLeNSoaJ0fRtjxXGzZFsLcfBOv2T0hjubBcNgGe42tB+Tnt/LpwP1nQGzHVEnG+JmC3WFOlgRrSEsa/zC4huDW4jgkVuAWjmWBz3XFxtuVKc0ejwXtnu+tkiohiws5oxT0Oi8cM30MGqPX0RCJ7wYlKS+VOfBaPQQMEMkjAo1aobBpJrs1EukCviFl5arW4H2cxSqIkUVUiKY/BNp25Gjtueo7lgaqTJo2pHtqkU5xbhUmCTelCtlspaFVoFDvWWlBz6tucPrrl3leNnyMXWFQbWoXlHf1YV5Xd+T95dyvUipiDB9pFpLtCOVDSm1so2E9Sxt8gJt/cq0thxPbP/xQDEcgTuILJY9V9st67HmuGvoRxFF00M69wrQCreM3DjacLXdcq9pOW0a1qoFZbCdQPkuEc76G7Bse5w3nCxb/NZAB/VxFEGZoDCpyVZOTfgaxiOpgFELz+OLNUf1lgM7lAqnMGrMILlkUaWORTtfGOcZlYxTqlIh6GVCemMStiqqgAXxVBe6Vj2nndd6PB2bymlePSFqudLCpnr+pRkKmqhnDkHuIAiktIClSxLBc7S6C1VxDja+IaD4SHeNPzy9xnqsuXXngHi3xvSK+q4S+enU5RbS/bQQp3K4FghtRF0ZeOKxY5bVyJOLE65Wm7LeTqkA+a55L6CFkQDl2LXcGRZ0vuJe33LSNQyDhU5jOpV0RRKSUIFbyfePh4FwZZZvegB7cFLhjKygHHKBciPBWRMPQloowi6bem5Zx/1+lh0Bk4gWgVj6pzfaSYmVyRM4/STCm8qogH+OQXie68pzEg1nf1Ph/vpdIhpxse525jjsP5dJmjA5BuY+hMq5I545GCpLF4cgrP0gDY5iTsmkxSNGSlOYzlds+wrfWcwW7CZgu1m76hClI2GUNsZiFtOBrwQ+9YO4aJu0SIzOsN00+N6ITOeWkm82vUf1HpwTvkZlmZT+dBGliomAGixoK1Fxa6b84gj4mTLYC7GSkiKemTPzsiidSIclH6wyQpCuSSJKKpiIQ+knBk1M1zaTAr2RNFnZEBJgFuVg0odSym+jofRUqFP5k1eayng2iSMkablE4j3HWj3S6oFD03G13gBwq10y1jVEXbQRMh8hR7PazaLpQIJfEzqQxVXirBz2hV8WYIJSfT85JTkFEKpYNhuthCgdUiWRSmJDFLJYpLWOg6qXUjBrpIUzGWGTjogirqSoTEII0rjm1Jt2OeiS4CWPR1kjtKAWSztIhDlPawUlPxnp3wmA5nmhWP4fExtAAqp4LlH4uZyBhxYwehC7XzoiOypqtzpsnj7ISEFGQqf0aSyaAjChdZ6p1FCe16Vx0LxSausr1qN0gg29wfQirmaSI6DnvByV9FMqCG0kLjx14ziqexZWqiFySmNMMsmudJzVOw4BQNBKSl6DyJSPqQGdd6l0Ne+9zBx9iwTpdUTXD5f+fPDmRpUMYoYbz8s97c8LHyTvMQZDHy19qKiU58B01Mpx1Wy4bk6pkfILDxAjY85zK5MQA80V3ePtXV5Z3+Nw0TGOBreQ5jdEIWAZjTgCo0ONTnL21oIVCEM7qTVWARHh0ZFD25UWzQW5yASwRDLxcWq45FG0Sk/M1dkNVzbT1J0upwuUVrIxVZLHFiLdxd1EGSITpmqytLaEUoqYZKKTQ+PjNOllsmmBoNIkN11Eb1LZXvnMKM5W+hblPGoM8p5eYP7b/YpGe546PaS/22JODM3dSH1nQA8ONYyiFJjHJShBCgaD6QPR1IwLzXCo6E8aooW+jWwbgV+rtaLpJf3Q3pYWtPU9J42KUqpI1RWqqoi5XA/Kwmz6iLWi1tc7ibBc0LtQ4yMKiUyh+J9XpjGznCkwSkiFUDgwGf4XVEaVI83SrblJUtARtzCUUjQ3I+VF2cjcMhLaCAeOJw5PeXJxwpPNCdfsmk2oqawvDsNOW+N075tEbFvqgSPdcd2ccmIXPNmc0GjP3eWC/tDiaovbanIzq9yhEtIGGykok57BnwBF8GVvbF6oRZPHYVrbJKqW/O8+53feYndewgykKgQr6503BD/bSNMiLdVQUkmyU5UUwXaCgNheEbax7N0gKMF4IGVr46jpvKUxsmwHJd+7L6d8rs2OP39xRJ3/hrKkySZ8P8fgkZQc7h+PQgiveuoIKnyZsOtwJ8uclrntVydl3lquHgAKYbkPon3ThYpj1zJGw91hIShRXwmJcCtRudlKb5GMfGXuRjTga3EGqtXA4bLnWrthZQZWVkiEPkptYEEn8g+KEFQqM9RsVcXp2HC3WzB4w+m2od9WhMFgO43uU8ogOfg5VRCaSGw9i8XwUMP/wA5BFpHZmXRq9jj34kgBZVS4qOmDDHCjHI0duWrWtGrkCXPCddOlXHYJesj590pFGjSVgseNZ6U33Kxv8eTyFB80z6wWuKUser7WVFoDHtUNxH5AGS3OQGVRUTaDrMcuDkHgul1z3Z7S6qHkjMo5I7WnHs2KoTgEaDDR7joEUCI37cQpKZup0ihr8YnY5lp1YXr5wE6qI9eros4qXk38AU9Ik78Pqe2xN0ngIjG8tx51upVrkYltmReRtfVH2YglQofQGW5thdZ6596K6hlLdaJYPjNibx4nPkJykoCYlFt0J2kIbQ32pBWdgFXFcMUSc7ewRiZ+vfboQdIZtvMoF9HdKNwB55PjVYuYT2VLWZkeExrRSZJ13CriYBhHS4ha+meoAYPFq+fYsJ+HZc30nefOQclU/ielwtBa7lIvfAsSvD4po8lnWuNL9Kl1xFlNP2qcTk5B3qR0iggNsHLYduRw1fG6g9u8qr3LDXvKK+w9TkLLshqFpDgXS9JTzrnSgVY5VrrncXPCdT0wVre41R5wzy5Zu5oxaNZ9zcnWooLBaNFHiFHJtUsOQBZxEd5AcmKy6l6I99/onqflW0PEhdLvVc6l3+fLZs7AnL/kggixDakMNvppg5bvmtAEoyJBC5swKnGA7DZSn+RGWwjnKSFa2hmGK6msehAybZ2k2IMKUv3gZ6Jos2PdOc5n2bulQiinCmZOQir3jIn8esYxeAT+wBmCIxTtAaVlbotC4cQfmFt27OfVPJIO3kVIpYNlw5h0cDK62oeqNLM6dgu2vuJuJ/LAvjdUp5rqVBzbah2TJHhCB4zMjSxTXK0Grh5uubFc84r2mKUeCs9BE+mxhV+XZahdMOX3Lp3bydBw3DU4Z6Qb7VqabIljMhH+M3/ALyK+idil42jZPVTF1IN3O8zjm4Oc2WQr8yT9P6NW97PC2Of8aMwzqzYo71FUiR1qtcfokORdOX9ixhSJnanB2f3vlF96sEGbv25igz7ggGsKnAdc2A21//2ZvZ6rJfZx3qxNUMSDfNIK8GpikefF6TkaMuUqA0kZSSe83luCV9hMwEqtqGMRjoqy0aUwLKb+AkDhLBgFlRaYWo8aP8qiZ9ceMwTZTHqX8uOJ6Hi/Yw0BNQKYiayWfTU16RtkbsgnNWf6bKYEJdiPxFScWPdzH01QAi33hkG65lUpXVEi4rTwm0jVOBbtyKoeObJbDkyX5FwlbXCGG3Of+Zr5N62iOAijMazswLIa8UFzUgVCIzyCYGXB0FEa+CgoWgOiJyIpCqk8iJN+CFzctdn/nHmgE6e9ufRX2IvqVRp7laoRfNSMqRMhc4QAnn1DjgI36zGU/0PWWFFT6sSTykwlFTsG4Y4UJzNSUgZqvr7Mkdszx3CfwdxhXk8v+6QKET2klTRA1Owry577+oSYZgch+1J5XXRBAtmsuhu8glxlMpMJzqmCuR+Vyz+NmUpEc0WXTkTujGHM0wPZSufFVME2eAlcvNPEUUs/mHGWLojz+4cidmWSANLD2MM7BFDygTlXEYwsPMGS4B1pRWlTh6msxlSl3Emp5UcxRL1DqoJ5mjky7tH7chSsZ2jEjgWkxC+V/JV1RE055OezEeeWzAAmVqUSQatJ1CfDqVGxs+Gdaxd4X53nzBgioXhv05gPqd71tl/x9HDIR7urfOz2FYa7DfXNBO+fBvQQiE0l8L5JjlWMUt+/t/EWydZBcbppJefaWbSXsd4+ZoDrO6WfwpSXldd0Hj2IU6B6L6iDC+iNtPyNTUWojaAIW5faFjNrX8zkEKQOk4QAw4gKBtVF+X9liXYFUdrPqtqzageu1B0r3bPSPV2oGM64o4/GpmqbCTJUKpWQGUWoVZLYdWUuaR8mItModf3RSrOU1joqHWjs1EBpbHZvcWMD1soi9cTBKY8vTrler/nc1R/yhD1JSInMbl0iwwkO3UEDkwmpcOSKrhkZeF31DMemZVwZKu25PSwZneFk0TKeVhAtZlDoHsygps04ykIbrJac+iCoXr6vnoV29NCWv9f0k6Kg/C7jPjgp7xpT8yjI6ZqU6nBSOqhHxem2IUQ43bQMmwq1MQIlJ+7NnMVvdCjXpqR/+ojuvNwf6fOjFd6LtQq7lYoeNSi2rpLyVCvRpgs69WGgCHyhEgdFT2M3d0qiipO2gOLZxYbSdZHj3Vs795Hii7AseJXmXkYtYiSt6Rl9Zgdmh4nkuo/0nmeioipoQJ9KyHMXwT4Io//usJQmX11N3Fh0Jw2EqrWQQG038UN8Jton/kCspPlSpQN1UpQUYaS0DwaR/SbAoKZ71KVSdhc1m7FicIZN19DdbcEpzFpjN1JVYlNFjiDkcm3cEtyhhyZwuOy5sTh9qOF/Xg5BQQUsUnOZvBLmXQRTvmzq+b5bU+mJwvZMIkVGxZ1lWALOWGpBQSoQYLYJP5sl/YFdOTee9+TVsyqE3eeTczLnVqRSx+etjfzQx3a+Q+CT8lbOmfk4sWjvuSV3xyV3+wXDcYO9Y6mPFfWpx65TxJ6j+nwe55RQAiWHpp1i7CwbgFEJ6U3BeKgIOZefI1o35d/sxmC3Hj0EbL+VagbnhQcCUFfoyk4OSS59LAMgpYU7KoVePiPGKJ/TD1BVmGVNMLLhaBtF9tQMrHQvKaFHkC54WMsbSDBJX8NLR89c+ilcmJQ79HIzVjqwsGNZJH2U6KK3EulbI/dibbwIUVnH61a3eEV9zI3qhD9Zf4LrpmNMVRZd7KUMcrdON5HcZOyLUhyRRsFS11zB87hZs9Q9XawIUXNoO46PWqwJ4ih0mtBrzKyTYXEIPICkE3KKpCCPF+gQqJEkxywkR0ibfOpm6L2oE/rE04jp+PJx5g1YjzAOlq2KjL2F3qB7Pcl/Z2cqMfj31y7RaAjozk2CZkD0IrVtetHIz9yLwSdimdaMwZTGOjqVOpIze/k7Zw5BnAdRD7kWftKRgRkkNO/HEKMqP2H2mHuEZFGhuWnOqgBmjZZM7JsTCIdgpczPVWzHiqGvRG+gE4jebuXamlGEp3yNqOmpfN8KElcqIfSUssj7YFCh8B1ykyM5LlUaF236mtEZhs6i10aIjFvpaKh8mrtZEyOLILURtfDYxnHUdlyrtw817A/f/ljNEAIzeb/ijUYoOZ4snhGmvMk5qoAwlRyefX4yT8Q/B3z9oLbDvp1ZUVKMZucCEifBiEzSyw7CPmFliqTyDZom4qyfQXnNBdl545c1ucdo6WIlkW80dLEuPQa2vqL3FtJCmBdHPSZnSorkJ0vtg7MwUWmJnFMGiawZk3RrtMIQF77EHuztI66VNICvNW6p0WOkMQrdN5jtiD7epAhLTdF/Hst8PDNZ36iTjtq+zO+87FNNeXCtA42VFFS5pio8UITxyKywqSHzCFRONeU/h1gEdAgTsgBTq2UXZPOHVEGQSIdV0nuvjefA9Fyxm+QMOSoinsg6WtahYeuqXSZzGX95EDlVWUwD4GMouiSGSK08Sz0wGsNR3dF7aSF8d1kX8aVM7svQp06OpHYK7SOxZ3/qXMwwh1mE7qfnmO2ZOY1kdcCbIMFrfo+TF+pe4beGPiJR5EZLH4NhSk1NqYi0iaXfM2lKqhH8NGfnczc5STo1lRqTo9JEQ4guNVabvVyVoFrmD3m9uc/aqfYe53bfjEI8ixY8Ksv56AT8PcgWMAZbUtKQyjD3TkY0B9SOM+CC/PTe0jmp9x+8wTtRjMz9AvSYGmGla6wNBXGZD7OPE2m5D6lDYspk77Y3njgEPogzMIbULG6whN5ge1Uak5memYS4fF7p9lpJZYGtPE3qd/Ew9uAIQVojgxXCREEJculSldCBSpo31LnxhBlTcwlhI9dqShvMPbnszMrCIo8VEa+ipA5iZCRJ8uabqnw/6S7Ymy3PEqVLdWJMeSSFJlDn3U8JpFSRCSCy+dcEvJKFLqBTDX9IGvSzxbJs/lqIjemn6CWkk70op2A+jnmR7pIjsA4NT7sjNqEuHvQYDbeHFbf7Jfe2LebUiJLgaZQc/dalVIFB6ckJUM4Tt1uJvhcLqCtRG/QRPSrMoBiTNKvAV5L/9A24VF4W9ZRzK6Qhr4vH29wVsY/2TmD1EYPqR+EVJLRAUj4SBqmEAJWURuIqSP+C3YggRhG5CVYTak2oFU07cqXecsVKs5NaeUYcg3o0DoEuu83UQXO/D4bKczrl0UOdkA+jU8lnFFJlaqSjRk1w8llZqa1OO1ytPZ23tEbK4prUojg3e/n05havqu6kap+RQ6V5ysNNf8hT7qrMjQRRSudLpAwQuXY+Sq32Wjd0UdFHhycK/4PAVb3BV5pDs2U8MNyo1zzVHPL7KtINFeNgcUPuTIqw5QeNP5YFOGpVINGoLzhlUNIEUqkCaaHP8LsSoZv5Cjmo7AwkhyAKq3t8piLUlnotVTB2A/Wxx2xGzErg07w8jCnCV16J0zMi6YJulJRWZRJxMTmAQY4vGIlOu6GS1rkJivZRTfB/SQEg64shrTWxEARh9vhczkC8PzLwyHsazD/eq9KSJgZNCLEoxOaNFCQw6kIFGnSMO05B+ah0sn2o2Pg6pQmqUvZ3MjYMwbIea+lkOFriRpRWTQ/VSaS5J7osduslsIkWX2X561RtlsT5Ri8l2KdeCIxZYG0XnRCeQpdQifUgqartuiF2Br021PdU6aZYrROakK6vMxJ0hSbiDzzXDrYs65HrzYYj+4gQgjwvcl6qCF7MEINcHlJKoFROBYSpFnQPIRD5093JVQgeCI9g33YW0eezqc7qcedWGPqczcvrXMcaUxlLDAUd2OcziEeuJg7BDHbPBKn7kiGfh83HIztaWXxjTOWeOV8270KY61oFvs8IQUCPfvI8Q4q6Y5QNyXvi6KAJk8OVod5AunkVqNTBTQnxKdqZIwClnI382SBKdOTcmKZtDCbmSCu930xjKSvWfC7I8ajzIq3SeZLC4LY6SNe4pAimMxp0waxCQ8A9Cy/hjFaHioXzQp5Lc9QjRamlq+bscG3qReCjojZOVCqNKAhmjYGs7X5oOo60tIJtlaJS4pJ3oWYTapzXSaZ4iqj3N5OAmgm7TAm+LA/bJsWWa3YNiE7FM+0B1gQ66xmsLcqmMSiCNoRenOYs/XqRjkAZ4hkqUHgBOeKaTavSjdJkgi4lZaCdOGe2U/gAdiPOgN0mUa+k8rlT7gl7UueiBqqcTz0rdrvDyrHmFJGkMnzIkacpSpUT8W96zOjAzsb/EFN7xxnIqZsXw0r0N41f5KwzTXrOZ+8xOd3zNu9FMTdxBmQ9lJz9mBCCnJYZvZn6TiQEVdIEoQi3kUrMVdDT/ZHGOAQ5zlyWCjKfxtwWPTsySbwv6/a4pDUQnRb0dpjQibkDS63K/SZaIUAVaapJETETGR/UHhohKO06Z7AmOhb+wPwSBYQNa9Uk8iDncZb5OKLK5j9FvNOJeDJy8AA53kToyyJFsonM/p6O26hIq2RxrHOlgQKT6MMmpzj28spZ//pM+kNNN6RKEezOpZiphL0QPsOD2uQcSNnnvEY3a3/vlE3K3VaOnxQRFrjeaGgaVFWLvoMPMDgWtxwqWKojRTQWvzACm25F6tV0Aq+Vc0faHLuFzKfxMOIPAr6KdI2kEVyrUbHFdpH2mZHqmdRYaRgnbkE2o6f0RnZS5giBVqkE1YgORJKnXVjZHCvtEgIkWhMPMsUexPLYnrdwTRUqsnAUZfCdeveMiEzvU1HSBVItgfQU8EJEmkdL+ftz6s4qv1Mmm53DLlboGLjlR9bK8QfuMX63eyUf66+yPW1oNypBlMkh07mV63Svh6jYRMsmjqxDZBMqhjSIeTyvm1NaJQJQ28OKTeoYt3WVwKSphn/d1WzDgjBodC/XSnuJkC9yU9K579UQ0X1Mvyen2EktOCQycSX3zC0raTQVU9fOPkgL5HvST6U6jSLR3Ufs2qF6L5tGvqRBmiQNziSmOEmRcCrHlYh+moByvYWAq5y0GB+cpXMeq2v60U4R1HxNKdyB3LAoTvkEIDKrGJgvUvtLbx7z88b9Ua1f+0jG/PlEMJy4A9NPrufPkfdO9VWY/6qKBkHuETCp6opjMIYkwe/15CT67IhP6GSu+NE+EmeOQ+wVfVdxTy+kNBSojaTrVkkJNHdl7MN9goWAEEbzd89+inOXAvJc6qhrqSyQlusJMXkIewiHIEdolGoCyAcUwQa0DWgTyqKWCRIubUiVamj0SMu4gxQIbG922j3uowZpfM45sNnvuTWvUqBNUambR+iQ7ouEZDR6TBHSWEpCioxy8ix91AV+0ipiYjwbRSpKRDffWM895HSDPUqezlRaaOkSJCaQcY+JAjc21mFzWUqJWCgOwNwZUEHCJrVIAgqJT6C2Pe0f3aN5yjBeX6LHhnGpktiKeNLNPY89HXeu1XilprsumgzHfwJc61FVoFkOVNZzfG2Fb2vsFg4rxeHg0b3DbHvo+ik9oBTYmBbRlCqwVlI0meOQnovW4FuRL/YNtNZJQ6PUi7xSnpZH09wIJq7H/ty2KjDu11QnZyBYhS5OjpShCYqjS9ogdx50QRP0pA6auynua7yHqAhKmrh0sUodE0U99AP9K/jfJ6/m5vYA7tbUJxKVZOEVyVeK8291KOpu61hzN4zSxyI25fOyE9BWIz5qrttTlnqQORltgW1PvUC1N7sDPhhvMHQW1ze4k0mvfV7b/0ItR1lZc0N+N9gO3KBETyAqGus4sD219vxhNUn/mm2gOnVlUQpGUa8D1bFDu4g96VFdjxkXafNQ4KEbLc4lgtggTogavaTlgszjaDL8kpxyH5OjAn4w9NqilPBzhtHuqMjm1PQukktKGcQp2p49yBt2x+ehJIsfxTpW4I3ZExHhEShdUAJh5EuUT4BRGwgQlLpvZJwdAhdMUlYVXoZL0utj0mXxXpyC0sMiOwTZQQgTeqkcwp0bwHRy3MNpxcYrhl56D1jjWdUjy2rA6sDSDlgVcBnJ2CfLe1UaauU0VeaTyFvU5BDUkbAI1JWntSO19oxRc+LaR4MQ7LJTd6O9fcs1u9lr8ynX81wa8dJy8nnc8Q/xlvkc0zmd8RyY5HlEyPz+YnvEwvvaJwF2K3LFTGhGrvSolCcgPe8FSts9pvseeyLw5U1Wyg+DqEF2A4wa21RUmxpgcgjGiD0dMaf97scZRdXKCmYGgcaw0FSOw7Znu6wYDy3RKsaVwi8sKIW2iXGvzwnj98mE5/ytsPeTXK9WU+ezPHb3cT0v3B74Rk3s/nJmc2gykKKmKVJ6NsvKbDrGCV5FGr0QpZHL6diwGaspXZD9IzXbdHScOf6Z4S0/pR9EcqSzHoZoFYwcmE4clCTu5aNGq0gfLCemwVqPM6bkSB8JmpacX1FxTM5+Ei3LSqbZgbNa+n2QCdQqv1ecMzNolAHTBUwfBEIefap6mX+npEa81+W7JcpL89ZoQjWV+Mr9JseQXx/D9BmDM3ivprz1Xjqn/F7W7IyyxJ0bXUX1wJyAM697BNflPl88pUWgCN9lC1ETlKj7ofzu3xAHeN8EUXsw3YJ5EBd1GsI54ptMnITUbMop4qjxOjKMomRpEuE+rz21donMqCduHDmtpHbv9eyEpPt+finyKWRHMSAiR1v/qBCC2ZipkIRCZv3S8YqoNN7HApFsnUCDIWpOTSORPxGvFUkt9IFEfaT8cIp6zy07TOOn7rMh7MCvey/JSIBRUQiG5TGUPgB5k81OTa6YsCo3OWKn3ri0Y1YT4e2iexdkq2YOTYWM01L3kj9Wnq66Vxbhle7xUfEJe0RrR6zxM3IoxEoTnXjZam9jjFpBXcm5MCR56Ew0DJjRsbIav7CY3ouCoI+o7SBNkZQSOWGtqe5sMeuR0Fp8s4RYMVw1VDdOeNXBPY6ajjtXF2z6mjuLq7hFQ3UauWoV1U1b0ImycM43zHyj7pQmKrAGt9CMBwq/jFypO67ZTRLjEfns8/pBPF/LjqQj9e5IbOJ9y53YztwLec4Wxq1sFDkqyZ33CCTY3eCSFDVMC2bvLVtTl57rAYVVnhPbsgw9Ac0QRTXtQ9sbfPzkkJN1i10ryYcniFJIjhCbgKqFq5Crh7pY0UXDOtasY42PWlC3dJ9kx3Sle66aNaOWvvP5eyt1wCbUnFQtlfGM1uOTSmKBSS8IHQBKl0XTR8x6RMVIvbREDb42rDcVd7cLWMBjTaTWnlgFfGtQPqVMk4ZGc0c+y65HzHEnTvIwyhz1oUSV2sHQi8BM1YsOgwoRv6xRbsV4bUH/WE3USsS3/CzCj+mYO42Pls4rRmsZtxX1SGmhXDat+fqcnieTI/Yh+XOme1S528GeE7D/2gtM4+wcU67CmH9ncohQiFMUNKOKZTOdAs/ME0iSxEqjs5pguidc0Ge5O6SyQB2wxmNMEEHWJuBbSTsPBwrtLcqD6YTwG60oqQarSikqgD3WhK0m1JautWAi3aLmtBkxJrBqhlIJlPlr3SiVON6JvoTIfKu9lEHq57B/P6iId4Z7fYsZAzfj6qGH/+FJhelAFHuoeFBELxfMe7kgmWEJJGKbo9WJWvgsXtkcUvXknOdEvjpPsjL94f5RYi4323cG9tABoyLj7KMlepyaMZWmGYkPURbyxOQt4kc5VfFs4kQXaDnaNCqSfUKDeKKDMSx1Tz3rx3BoEslMh4m8lqJnbTTKzMalyOFpYUFrjR4dhEB0nni6JnQduu+plMLWlTgLwyivCYHoPcpaVFPLuGw69DBg6prVUUUwUkhrdODV7V2apePges8m1Px/whs5DUfU9zTtvRrde/R2FCcjN7FKBEiB1eUpFebd3eSa+Ea6HfomclD1HJqOlR5EPItAUIqH86kf4NpkkZVnsTzndyrOSlSyO69VmBocCYwp7Xl9Shu4oHcIskEptr5CYxPDWVGpwMY3bHTDGC0nvqWPlk90h5ycLhg3FcutwvahHIvojgBWeibUxpf7Z0jOQBeksiVEjTGBNkl4lfQbmiPdFedenBHJ5xovHd5q6+l0nNKUOfq5oLJjGUP5PO2iSGf7gD2tCEZRLTWqM2z6isY6NJHGOKgCvknkV4Vs9h5ML1U4+niLOllPX6J1Shcga6ZTUlWRJcJHifZCY/CHDeNRRX/FCOG1UxN5LKMZDnQvydbgFcFG6LSQzhzlHpbvnjkGmfT8kMP3rI7Ao7I95EKeS8cSU6tnHQlpzmuV2xWrcyt3AqYQpOd/y/yBsLenWB2I0eO1Tg6BYrQxtRFWuKVicFo4A0YX7mKwqlTCmCE56qnRUagh9JZoIt5pulGjbCAERT3TCQGm5kVeUFNV7vE4cyxlXSupKCiBbgiKdV+jVCzqhg9jD97tkBnaNPfikHroGORF0afmDCmX6YLGKcnV9EGigtzy+Lks7D2+UNvPDEzNMabNPkeHpXXw/PWJX5D/9lyphvNsx7G6QMutM/cty0SbGHd5GzE1ZIlqhx27A0MpNS3CqduhGn2KPBQsWiHVWCNCQk1NXLaCjsQoURKC2sSc79+3tCibJLLhg+SkG+0k3xwrri23nF5ZMURFf6Sx65pKK9TaUuh4aYFXWZAq5Uz3L1FMNbvRRuFU6LGkDEzauO+ni/Gwlkm0OTIvzweTNkepsMl/k9auQRjDpZmQmpyaPZMFQhWEIC9uWYsAzhIap9JH2ISa2+6APlpuDodsfcXT6wPG0xq11dJNMm1KvlJTNZGJaCNExX1tkTnpN/NXgDIHM9E1OwP7UrN5DqtU615g/QvkD0Aa25grOc5xtqJEoSApAxekBLcgaYmsrHyQrp0hVQqkVJqyVoiCCeLN91hM2hGlogEh2ILFLTXDgZo553mDmfV3SETSsg/Mosdz5TPmFVD7j+cOzPytD55KuDDL3zcnPKp0LCSnYEYshOkx29wZzjYx+ncf5fXScdTqgA0BjCgGVplfVYVSau8XijFxCoLdbSwUFam9OJPGxUg+AUn39JqoItErxtSkSs96Z2RnYGo0N6swGCK2D5guEBXYpE7oa4XZKMCIQ+AkpxFHLcJlD2EPziFgQgWUlxMUITM1TTAvE8hVJqmkWXpriVFxMra4YLA6cMMaiQoeIHczzqKC84iG59p5SoFZcz99hPAPRUrSqEmDwBCpCIxKS7VB/m4CdVrExWlwiTAlEs2qCDNlJEJJN7+c787jlBd4FR8pqTBLF/tE7Joz53PzjtOxYTtUJU9cFt0iPCQhhvJh6iDZDbLBH60YbxwkuMxIx7iSF400t3tMl7puaC1jkQc+VzJoSavo3lOfBvq1oR+l1efS9HxW83G0Ctx5xYr/Wfd89N4V7vVXGQ9qFjctV9YDbPoUiYVSFolzQjicf19iyUunSYG9r1YbrhoR5inprAsMhY5nHaxyOioTmdgrQzSpGqCtR7Yrh/cK34heQhzy4hjLj+SvoywYiQQ3pg1sYcQR672VdMXeTmqVdNx7uj/ko+Eqx2PLH9y5RtdXDLdalh+RNtSrjwcWNwd8pemvWXxqbmTqQF072pIykM/fV4k79i13WSbBp91jyN3lch96mFInRgdZJAMSSQ9MjZAuyPKl8a0mNFURBppIYkq04yE5jg5b+6StIRBxrDWqj+he0gNxvSEcH0vacrVEVdWOs6090JtSxmZG2Uj6q7IMb57QbF4pzqw9UdjUvGZxK0g1hItSueMEHYgWUc/rpJIn1ODNtDlNFQcRZUJxRvYrB+ab/tm+Gapoh5yxR+kr7OXO5WAoehU7vBkk8NQ67qQBsqM5MjmaJWWQKhFCcsqNkcAi3xu1doQoypDea7ZR4UdZ60wnjoAZxBnP6SCY1lAVwWxTAG0gbFOzql463UYbGZ3CNRaVEDeAsbfEzqBGTXWipWplHWnviiNQ3xuxd7dEpbDrhtAa6lOLOJSy10Rtd47lYeyhHAKYTjYmgopKHmj0SiLHVIce1aS6pFVkCKYQh7Iu3HOZj88fcd/nEoiXfj50O1U3TGjEPrlxTjrLm+25eV84Pye356A8Smcg27wiYn4+AYGVRy9klhyBnTmVnJtPUsBxrrV+7QC3soRKMRxKtUBuL6wdVKcGYwQpYK4dsCc5DJITkw6GkSHlwivluW42tMrzJ5qb3DtaoFXk/75yRL812I0iVkacirmSYZarzspvmkmDgBR5VRESQpBZ8A/sbD6EldrjshtQyD5zE8JnKgc1AV1JCWY0qUteniupFFQ+aIIQVZjkXEEWNQCvfSmfm39XtnUq/bvTLTg9aQlbS3XPUN8D00XqU4/ZOGgMKtiioVH6lGhf0LIc7WfzTCqZQOlHn5E1UYqb14Tr0qekHHFkypleNEKQGiwFmx3fNBdz0BAoJM1y3DbgbUIIzAy58QE1OuIwEoeBqDSm9dJxcx6IZFSgsNZT1NnIZuFW4A5D2vimICIYcR8l4lSy9qAECMsMdB/Turt3os+CCty3WdF8A97/nE8yYFAsZpSAM+jAXJjoPJPUwO5zmW+QLTsMWTgooGiM9CBo65GxlbUuANFKK+8wIE6rk0i+iFYlMa3cJC4msatCyE3Qfhi1PBeUBPI5qnfSt0I6z6afThwCvRlR6+3k641yX9aHoidTbu/9xwe0B+cQTCOX6lqnk4sJRoxJj0DrBCkaUSustCj6NUkYZd9MYr3nTSsTDiukf4GPkQ2KLmFiCyOlG9gw5TVB2MIuQD8Quw6qGpWEakKlcQuNbyBWkltXKpZa7IF+cgYSdFylnHtZuFL+dR0tm9Bw0x1xszvg9nZJ6EyBd5QHjMDkerFAVRWxqQmVlgVIJaj3gjahnag2Ts9l1cIhCiIzRss61HLswwF3ugXdtpZGM0myWCDTKRKdm6oqUSe0hu2rDzl+ncU3iv4auFVAjwp7KuVww8GCK/GadCTsU1+CEBIRMew5BopQSWfDXea6HP91c8r/1d7EBcPvPPYkvW8xvcavKmFzD04QjFTPXZpahSAL8NwRMwmKtYFl6mFQzxpomXPgxkdleUHKi5FOaJMxQQh1FWVczggT7U2dvEDudk8z5Xsy2zhvci4Y7g0t66Fm3dWE0wqz0dhThekithfnI1R6atm9kNaqdeM4aAaObM+h2Zay3VaPECadkVH9/9n791hrs6wsHH3GmPO9rLX25btUV1VXd9N0S//60NpHCIiGQGhppTEQA0qCKIrRVmM0IaYVNYfQGI1ohCMxKJgoEmN7/KMF4zEcPIlCDD/50XJQLtII9g26uqvqq++2915rvZc55zh/jDnn+6619/fV/qr2V9WX+VR27f2tvdba73rn+8455jOe8QxtvRxyeoCyYdF+IJ2Yt7nSGrN0Vt5pXxFSy2PXEPyiAo8e8/bS5AniNchK81ZVOQy1gKOPhW9UVMZtpdqVtgEvFjpOTRNLXVMdoO4ozdrEBSSmyGrCuNLvw5EAR8rueFfn4wiWsk18ClgkxEsiTF8AIguJrGdSWlc/C5IjZFSv7wQD+2zB/Pfp3M9vi1cjMMgMwZwh02MXpGtZf3UuNRYjsH2t2tQIKQWjqqdpIkOQTIwshWwo5Fstc/RC6OtKKX1PcGPUJwwMs1URoN3o+5OLBTI+BQXQtdJGi/WggSBIE9FSQVnUMFUoaHAhsclXUKF2P0LW23isgIw1KiLUZ1adL51ktkLmm4lL4pGNicAy/aEkpCOJn15Ase2itQG1VT/l1jgsjDonNewuzM23UQxniDJ3YEBgIowIGANjEyqMYrAyA641W3Dr4eso8gF0cegHhPUGYb0Bx5p5gu4EhgOGWxGkUoqGSbII6pC3ufOiTlqCgKjKj6LCVFZ1xx/gJLR4tr+OT50d4WTdgtZKs3KPePPoDlYOl4AI/KrOFDCQbuCrWXjmEXLKn+2YEok2NLrtD/DCeIRT3+LZ9TXcO11gPKvRbJHzxeqwJkrPJC8C6OdBXcHfOIBvLe6/tcK9dzrwasSbnrqLtxzdxt1+iU/cu45tX+Hu9QOQLGC3gsWLA+y9TvUHUWioBxwAGISK4WsV4NDe8QPAF9i7uGnOcMOe4ePP3MBvLa9hM17DcFiBhgA7ei199B4S87bC0NQBYmqCGUJxcq11/A9Mh0Peoo0aAoYaYD1K3e7DMBfBPmzs5juVynjUtcPYWLi2gmsJtovd7nI/g12mK9302WWT1E8jmZa6wNi4OndR652FC4yTTYuhr+A3FvVtA7sm1CdAcxqDQ4EaOcXKjPGA4FYBr1tu8cTiDK+rT3HNbNDSiEPearDPmntnBHSocgliH+9dreaYetDPHeSmL12IH5d+AADGlS6s4woYD6yWDgYB+xCNnwjjoLTyoenQ8IjDtsfpMsB5wrgkVAt1FqSxBlcG7FVLAADUNoAxkymaKK1fB53o7VZTAK4l9NcIbikYbzo8cVO7073oj+C9BYkGL+wot7ilEAVllH5GrsJQgVsMBtImzcfdW1pc43vkc5oCB7244oSpran1oSkw2E8v5H9fVQy9H4TMgkJJKQPWx5UtSC2hGSE3yYvzS3TeTJU1QGQGZKpKABDtvl326qhYLfeZAlwwOLA9rjXbHeMvFxidqzAGxmnXYH3Wwg8GcldPPhOAzZSKpZDOkX6Q4ABbUXQjZ0gTLxJPalrlEjOgJdxm62A2I+h0A3/vPiAB3PdAXcEMI9pGPVbs1ucSb7+sIfVjEhWmLcmOcnXGFqR/5+ZGHHJjkBRx7fd9BpLoTSfjKk54Ju2c9mj/tEDY6LJH+W/HJ6RdrfdA8LENclRJm6j4jLXNFFWluS/2LIUxCXUT1T6xF2nH3cUJbnQG3pnc/SwtoEIApW51AbmdadZcPKYI+0G727T7SiZFozcqmkre7RelDOaIHQV9a+GXFm4FmKMBq2WPLzi8g3ccfAov1EcYgsFJ3+JTB0uMqxiFVwzDrLnMNKb7nRMn1nZy04sDW1HACrroHFUdVs2As0Zd4sTEnfM+o7HT0Gj+M3T8eUqlcLz+PhOQ1M7p2haTdnqzcd3nPzFdz6nNKjBpblLTlCFaVffOqv3taOAHBsbJHpXHfavoGERZ0o5qlfZ4r00yO3K518dFQlz9+1O3zaTbCcI5TXKZOvCrhlh1EQ2VtjsmYQ2Gcztt/UrMSxadWm0HnuYTtgSplJJma1RMCOj9Ynhnh6aOdolKRp4DJLbMRR2wqLRjJdkw9bbPTOwD6P355ZCo5/2nzoKBB2IuQLzo79BuEPCqCw5nxzJPGwDYWawfhHkwsI+U0komXp5U2JzuobTRSnOTE8bG+FzyOAwWjgWhMghWNSjzc5mZriQMjflpFQVrkJD3XjkYmqWw5mDSlBJ2319JEcmVV+wCvHn4OdnHpQOC7IAYbUt3ggAStS1mga18bm60qEbtrMYeK9tjyQNaHqOYKAqJ4kTSEqGNTWVSqWEKDHJ0F6dtA3Vfi2yYQtftHbvg1NBGG+BQpGCR0xo2GvUk29qaAiqaUhUIARvyYFA28lH9gOZHt75Wc5CRYRzt3OCpKQ3rgWh9f+ruhnnE+PiRmjEByF0OO2cRRvXppnizCwOhTuMCpd8JQGVBRPDXljj7ghbDIWH9Ro//2+tfwBPtGb7y+CN4R/ss7jQHeKI6wx23wv+na3B2/xrsGcMMFeza6XvO+zqEkHc6ycOhtR5HVnftQRgjTakDADisOhw3HZ5bevRHFdjV4MHDnmkVA8VqCO2dPmM3op4gROttaz2WuctfwGQOd4U+BLPJJP+MqWT2nL0qNP9fW62B9g3gWoKvpnSBiiehu9F5nENqFJQoT6ZocQyBE4OTocHgDLqhQt9XWg201rauZkvqSLgVbe0ac/a+JbiFxbgiDEfAeCSQlcP1ZoPjaoul6VHBZ00AYloqpadGsTvnMlXlzA3I5poCBmU/EOYAMdGfPUzNla4Kw824kw8Gm42B7QXVhmE3QXvLp1LH2TldVgPMwsM7xnBsQV4bcQWrKTcQYFM30KhtETtVVKkJEnJVAMeJ29cCXwtM43HUdCr4rB3G2iIMKtj1ce7aaTk/u292RIQSdxw7FP/uNT33GXjJy11m3+dBwfxvXhV2Vd/xPgay7bJo2kMNmgiOOC/K2QhvvrkjiWukBgPJmnj6fbzeKPXakcx0jUY7Sq5iJyw1MfPR3E03V0EILy4O8OnmCJ2r8Em5jpFr+C3DjHrfkhNgQGY5pvJBtQHnkTQwNypY1T+mIlG30H+Y6zV4aVG1FuZA01JhoU21/NJi+7pKqw36CtVG82Gh0kD+UXD5lEEV8oHChnwxUPTJZqvfbeXRVC73XV/ZATU7LHnA0gxZQ5AXdxJdhInRUKXyJBGEvZ3GfGKxrANHe5Npttqd7z5TR7yUpzYAxd7wKooKMac5BQMtMSpieHJoSS0gq1hlMGLqo90Hi3GwkMHs7LCFtUsd+4AQWYNgWXcC2TQHV8YSzBeVefSbxI8VuVmXL1Xxjy7WQ88DGdKccTpvZiRNC1UWwozxqMb69YzhWNC8foPfe+PjeH19D1/RfgxvqQJOwxmesXdxzy/x7JPX8H/eXWI8qbB4gbCsOfd4z06DzuexSuVcySq24TGmPDi3cgaAA9PjqO5gFh7jYQ12jOq0guXpuBNjQH5PrxInGTGiNfI8oiENBB/HHnWuDUglq2kX4oXAe+kCPURRUxQb4GvRBj/VTCzmRSVPTrLAkOJHNvGaTvdYYgyCELqhQjdUGDqLsK4033mmC5rZxk6XW4mK/vinaoZrNMc9HgncYUCzGnBcdziyHZY85P7uyrRhFgxMXvEXBVge6veQrl0jSagbmwnNAgKKRN9VBgR8XSf5YWxh1wzX6SJHQeDr8+I8A0FrHOpmROcI44EBBVX4gzhqcCqYTTNpV4DICiKzgnMRYJJTiQWkETS1w1HVIYBQ1x5jJQiV6MR+QTCggbRM1/X8mEmmRXuOPXZyovxnzMjO29CFjz+WYACY6FmZ/Y0QH9jZ/KnfDVFa7M+7DqZKFmDSJKWeBfPAHJhdh5HJrgAsY2OuijyWpoeB6CYiOg+llMOL9RGeqM9w4lpsxgq3wwG8reDOtM+EGbQfB2anOgWHzHGT6ClvatNH9DVArV6X7Ax4ZO0Tc6CalXFl4GvStNN1vUZ4INhO30uvk8cUEOSBml9kiaeIbAHhPHWZ+hLkf7/Mu3pHJR8n0pRnPLe47jW22e2Ip19z29qaPGqEScOQUheUUhdpd3f+ODJ2brL0naIx2Px5kqmgx4GLanCB81UTXghwHEVGaYJJnRgvvogyy1IDi8rFbnlbNORhwKiJUEPdERdmBFdaUy8mpkpmnvwqV5+NTTwnPt60fajQSQUDwQCt40+KdY674WB1d6bMC2tAOC8VySclKbZVtQ+rtLeOfcgi1tcK+xNT+nyZOTlH/+rz53RikhWk5i4A0AeDzltsXYV+tBhHg9AbUM+zHKV66nPudilZ/Q6kQE1LzyQyawljXPw9AhL1pXoVk9uUz4M5QNm+1K/e03Ss++3Q9QNiSiVd8RjVtYMIoWuCBsGCmFKkB/49yx7MAoqCT18BCJrj1w2HGncl8R6JTO81X8wuQqTAE/Ute7TxPgg65vv08rSbf8AJe+Dfp4dPSq/WPZKur7R+JlYiOhTOMfcjCLNzl+e/2JQuPw7trTN4LX+3UZCQ/HIC6a4/sddT75HEJIfY80Qj5lWuHJo2K7+9vI5NX2NLgFtom2swoiW2TAEiZoFBgHpLGEzjQzEAbAAQYXTKKPhaYFoVJrpFErdq0C4VQI0ye/k9HnHcLh8Q2LQdQR40iu2OiZHznsZou1XNM6qY0LJHyyOqmXCjji1n00IVRDAineAHBw0ejK2vsXZ1tHdEzt/rQVH+ImOgTY54otgIYKPOUCs74JrZ4BpvcMgjlmxUyJjqoqF2tiw0SyckijNFnjRFdWkxTRMAQUVFqb3vqBdE8ky/KswXe4/JenlusuQRdmvGRwvestbUxvRFnvxTdJ52OintYgnjoWA89njTwRpvrG/jSXOKlgIYVkcnBlkr26NpR2x7g1CZbOQCayDBgjjeblUszfMaMW/6GreGQ1Tsccsd4ZTHvOu855cYxWj3vsrDLQEeCW5h0FRafpMbMYXoSZDo28qqWn4JVIc9nlhu8KQ5xfGsrjVyFTsGTleFiy2Lp13DXIColLnsCsTm9bcBmJcdIiBbuXa+wtbXcMJ4YXuIs6HB/W2L7Z0FqGdUa0Z1orqRaq3lhaYH2rsepvMxzx0gTBhXHHPsQGgEaGIFhBD6YHHfLfEcH+98Jh8DtyCcDZly7/edwBhZZJhed64kk/TzUwDEvcSC9Yh4/bUTBCE86xj9mmF6gul1dxUqTNqoPFYBNXs01YjQErZHFYQNzKDUrO78DczQqGdCbH/sWxOrBKYFIM1D2YExmg15x9HqXbsaJsU5Ut45jncWFwLZmIhTz4m0RxNMZj4PmWpISFNrhHNBQWYHLjNVXRXbWev1IEIqbBRoKfvefCmxE6EDMDiD3us8svVVZuFsXE/G6OnvRM/v6A1MbC7EJIBrdq679PoFxyZErNVIqeqs5QEVPG6atVbHVVol14nBE9Up/ufBG/CpzTF+lV+P7UkNs2bUd5MVsQbeIB0vFiD1iJk3pJJaMBxLFo124zwo0qDVtxJ1PR5og7rLEs4x54+CywcEZnah5ABsYgcQdzUUT6idCQkr8jn/Phce7e9aHxYI6PP1932wGLxFcvyahHwzP/vUQyDZCM9BWnZoSb3VlzyiIaCKZVqJGTBCuTGLsgN7DEEKwS4Q66QdN7HojgHQm3U+kT8GPMhcJ+Vv0zn02Sd7tsukSHGm05XO7az1ra8BNAGHVR8DqW7qS4HJp6EiD2v9JI7i9P6sqv8g2OlIGCdL51nr432NdWjgwegiW7AJdd5tGhNidKwTMqJoESKxvjoGAt5rLtcwxDJCJWgah1XVx3I5o70y4pbMYBqux4W5ecpuuofzY0SS6eC55o5EJm+ISDGSfmx1eQyMPvZ171yFzVCh7yrQVg1V7Jpg17qAVGv1jbC9wG58tuCFAElUkco0xUhOtaX+7b1YnPrF7meLjED+Hhd6tUyeBdEAwMidROfPA+KkNnMGvOod6vVmgyCEO4slTpYNQFoBpIv3xcGHZdVGjd4DdUBYRFFr9AdwraZZKGqsyMd0TyoDjIE3zTYMwCyoSztVoWgiNF/csUM5J1CcU+YswUs6PF9mvbjsc+iSz70kONlVR9pSKw1ndT+z6giJ5ylVGnieWoB7IaT0eWprnLxXem9RwUejPK/dBoPdYRlq9rC1B4uy2rlrZwwGWh5xyCNaEhwS44gXcPC41/42AGBlejx7dIy7ADxq+E6FuyRAqnJOGzFKTcR4MpwSAqRKA4qJ9Yz3IkhACw9beRALmmbU9Lv1aK0yGPNKisviZRgTTTbFQJqE074q5IuayOTWxyxa729kqjRgxIYumKv6L87kpnTDxCYo9ZMMH7LXc6JSjZm+MmMQ3yz+mPtR5yBFkdMEs2Pxom5WI6ANXEKDM9/idGzgO6VgTVJpz3JF588d8uCye3Q650HILlx7tGuamKdWyErF997COwM7ANxTpImRzVLIhel8RoHe1M9AF4bWjhpMkUNLpOdNgLVYnIQW98YltpsGsrVxFxrLGRnKEiAGG/F17LV16Pa0xSdW13EytjhxrRqFxMWi9xZ3hwU6r4tcNSKbf8xd/HZmRCJ1LawrhNogVEBbOSztiJZcHOcU9BECAvwVeubPx2je/ng3KEhOfhdEiWny46nzHUIMfsPErIyDxdYG3OcWhgPGYHB7vcR2W8OvK1Rr0oBgA1QbHW/baY0zx3JT5DEXgM05SjM4xjBa3OmX6LzFGAzObJPFVvOOmukzz78DcZIC58qHffTe5rLDTJumXfEVYmk1P7yoR5xUguBSnb/kz4uR0Y0Wt8cVerG4Nyyx6WuMo06bYgSIaTSw6j18E/sKQKsWglUxYDCRwZ4KnwDouU2tkMetxf2+1fltMOCBdaF4QNnlVPE1+9p5AkCYMQAz5PbGr3QOuuJbJb/dLG3ySm/HpBdIXQaN7NrOB9HKMUDLCZOL4b1xoQxxvL5bGmPvk15TRXxx4JgqFoDZ6U33UbKZnpNlMegVIwi1xEAAyKJ9wbRLMTIJ+GuHuvbKdjQDDAka67Knwhi7KD4KLt/LIG1yA7IAQgx09xUnMxID5wRD7HbY+bj4CKFhbYurk5/mmU1UdxuoqLCiXQe3xBgEIbTk0cf8TR+sUkTJvMHpQpbEg6gsyFdKRxsz+cHHm4YiLVtRygkFVHFRYzAYib6NOVCoKdIggtPo//7ieIDb3Qp0ZlGd6kRreslBwbnrJEX6Xi1L/fBI4/TIyIFArIhQgZ5WGGxcjdBp3Xl2whoEZtD2rezDrEd7mGh46JhzFdQLgntc44Al1Wiown0MuOeXuOMP8Hx3CHdSw943qM8EZjPoQma02yERqWcAEcgFmF50x/pCjU+FG3iu9vjY4oaWsMYbeV5m5NYV2gHRvlcmY6J8viUHA1JZhEUFtzTwy4CbyzWeaM6wZIeKWiB67fkobPNXNMvtVxDMF8YUACRWwFKIKvuZBmRG2WpAYGIQFRdar9bFpiPtvR4I42iwHSr4wFjfWYDPLKotob2tY23XgvosgL26n/EQ6+5Hn9sAa5ol+rTH4JUH9WHvtxWeOz1EbT1etAdo7QgmydVECzPiWrXJdr8tj7ncEBQQohg3i8D2VqS1r6fPHydICF15QPBUc4oghOebI7y4dPCwCNbE4FRpXeoZ202DT2xuoDUOtzYrDbB8rBwwAmlUCxG8Nr5xC7UWNqzaGdeqGZpv4zwVd9TJBImdwG51ZXcLizunK41pN1YbHHWqUt8pU5wv/pxYPV1EdqquZPYzLvj5MnjA8wlxDbjqaoP4AaZg4NHeeN7oKAjnihtLAeDU52Myxkqag7zpcBZOGFuqcDY2sBx0vQkWDTs8VZ3g2GzgDWPEKbwEeIrzxr6RW9ICAWA/pXZS6jhUMagzSv/DCqTSTqIg1arsG7xSnA/ZBKwWPRaVQ2MdDqsetXGa1opui2tXK5P+CHh0hgCYopWk/uQY0ccSkeTqpAInzr3XH7VpzJTLFxjap1dnVP3+RbvTP4CmaCazBHGQZgzBRVd0kN1cfFpkO7G5ll8HmHZyfMBeOi7XEc/e/Oo3oRdicuTiHNwE0JRuycctUYshWseatAPAVJs9Q2Zt4rn1ogvqRhqchAXWY53tN9khLzYSDYJAc6Md/RvsBNwTaGsQRsLWa8qFTFQEx9w6EYBxlx3aYQb2RZGx/EtYPcRr46NBVrSAjYORWA6z//rHjEvZJhtS90s/u1IlpaA01ywjw7HBwNpVDQODY26cB4AHgRn1i2IQzV5ixcLeViyyBXMnvPQ3+tjbXVsuc3Z3TK1cV1Yd3JgElUyWzXoNcmT4TO7tMG/MldrSZmtaQr5/rhJNtHOrjAdzgN/z60+fOXjCxulmZkiNZ5IiPB2TUYo75GZUsRJAMOWEOXao2188RRmeFISMo26KaKSpRHF2aHNGQOK/czpuxoKm997/W1eNq26A9FA24ILULDCloB7mQ5AbGJFHINpx9EzBQFqjtMEaEFiv2Y2rsLVaYtiZCks2UVDL2fwxpP9m13i2E09r1P5X+lgsmh5Lu/+oBdC5Lm7E4mcjDjBGy3IrE9BYh4o9WjtGcyV1XfRCGNm8pDfDPi5fdphEHVHkoph9WFbnqBC00QoQWzmGqePbGAzCfj7/JcCgnQk6iBqs9M7G3UvqC78fnZFydFZ3pDLLdUmcdJwYDDDoxKARDxPGHTOkUwk4lSar24Mw7oUl7o4r3BlWOO0amE1kB7ZRpe3iAFqGiBpV6L8JvlZxo48lRFcFsxftpmNNzMAQ3Qr7UGHrq3juKOcmtdugLgw0BrCbHNcyRH9vNoTxtMLz3SGe9wfoZRs1Ig4fdU/g/3vvd+G57SE+/umbWDxnUJ0B9YkDdbHzYXwvfT8PMQzuHeyG0TBh8YKF3RoIG4itYgQ9lVqNtf7c3mXUJ4Jqo+mI/WOdzyxijLaYbQiwKiZKpUNhdmd6EYwQdFecMjCzySdNTPN2rfN8pz5Hphx6NMDxjdHx8QHw+rj2jtDrz54wQqflrl2lZYX1HdZUQR/1AkPsNTFKFKpJvHf1elXDk8huMYF7Qb3WoKC5SzCdgVswNj0jlx8bDdjqdoS1AQdtj+7AojUOh1WHa5X6JW69Cg23vsKpa7It7BjZxDTxbcYaZ12jC+PAMZhEFsxdFY7tBgBwUPUqULWxnXkKMmMDIjcY3O0W2Jgap+sW4ex8c2whyc270gIdrKrJlR0Q+BbAEDvDxs/CTsekOtNyz2AZXas15NUpw24odsuLlHJspSt2Ep+BtAMfCakd835VyvycEc6fw3P/3s8tXPCcvc9+lbio5clutJP+PS2Ugt1ggKGiwIbHnCqoYmdXpga17JYjuyC578jojXq0iLZXBoCx0WsjmXGldMO9oPqZtQSchh6dGHxkeBIf3b4Oz3bXcLZt4DoL7mPnQoccpImJqaQmjmklgA0gG8AmeooYbfI1Z0mM0eZfhgPaGAzURo36dqv5JAe9j4LLL0susQI0KeRjBJwoMIKqnVPeYggGQ0i5GYORzCO1P9ZggBFkcpLzYAxerVen1p+yd+Hr5EaGlSmwZidqFiDubqZIrxdNYUCy5gOnwaiyHSbm4FXpfntc4WRsse0qVBsVaaV8bGYIGFBBGOUJIlSqNg4Vcv+Fx4EpXUC5wcwoFptQo3OaxpnTieTj7nGIwYAL2e1qDh4D7IbgW4Pb2yWec8dYmzqLxX55+wX4hRfehHsnS/CzLZbPaRqguj9oK+RoK5zp/HjNUOdgSctyVs8zxpNZOSlFUVt0mXRLRqiA6hRo7gfYTsD97AafzyTpZ8vwMSCgWtMdqcZ4LmRNaYPuisQdlymxVXOi87sbAiYxZgwmyRuwI9AYAJPsTQWVAdxJ6geh5ZjsdBGvzpQNqDaSFyANCHaNfiRW4lAQwAcIEcwQgDVAnhEqhlkQbENwGwswJgEUA8OqQl8FdAdVnqyccNZLDDHN13mLu/0SozcYvDp9AgBzABPQO4NuWyP42OBlxr6ltNtV4IZZw4NxaHtY69VHBVOeN6v2B8ZZ12BDgnFTw5zp3BUqUWZglufNbICoOAxGXRBDIwhtAAUGJ+dBAFqjLmhOpxLaUOvmxW4ItsPkVZDmjZhjFqsBo2eGW8eeEakiZa4STj8/bGF/wONpsb+wAdLs91eKnHun82zB3r/TNJJ+lYy+0qLYsMsslbpjRi1YMJm9DqLlhoivHwNjcJp7Hwbt1Oti0NoYh5UZcjfRe36VRbF3IFhLjY/1r8MnNjdwu1uh31ZAryxdTm1H/Y9E4yHfxLRYFYOBKsBYH7OdAcy7c4jlABuDgtQWwHLsF0Rq2JfmndSc6VFw+WVpflHkFGfKH0l+jLCbO0nI7mkXXH2XoWjDzs+UUwaTCneP8gSmBXn//XPeiGa7aEaVnNQiYz4KY4RBF+qsmh5FWY/BG0juFLgbkEg0m8hlfBRv9ljnnIxGHjdSuaGBCiIT0uKzU29N+ZfnzmNqxkGC2JubsB20B8ScdrvvFjjdtBjXFdoNwW4DbBdUGW9jTS787mBGN0lJDEQfIMSTehrIZVsS1d/BEUw/YzX8BQzBvMPhLDADqXnPfLGe5/48dqsAXk1cmFJLlHDqsOenShql9GNuMu0k47/JARz1IZkB8tNXTjeIXKhlBKY0kgYSSm2STNc2OcpmOaFSbj+MjMFZGBIM3uZysN5bDMGgiy6Zg1Mb5XG0sTJU54xxNOoJ7whmUNqcHbLi/qowLxtO9ez5MycluwAUDXCERS2ene5IiQGBMgP7C6OmOeIik8yErFLCaaNAEpkYr4EaRIM70+l8YaIvRPIJSe8r0VgLsZGc/rx3L8sDVv+X2O0/8HmXfd0V4IF/Km9gZhFAnHbnHQ8T0nqju3lGBbUjtuJjdQvFTqABuQlYfB8fGCEwgo/pLc+ZNVj7GmexM9ZpaHf+5jrUOHGttpUfK4iLm9Y0hnGDg5mIUKJzKmIJP+e06GT/n44NwI7PD7A7Z6Sqgrkg/lF9fy4vKkwMAWHWvwB6YUblOaUuhzGKqdlnR8AkMKrYzUrgtKGRVhqcTyUECIJ4jOLzhjWZSDg/8yDQWTzuekTL2YKmCvQLO0IcieVtW6/lbGvRHGEIPrvjAcBtv8Itd4QhOq8FYdz3C9wdFrjfazMLjvk/kumm9KmJjkEuOwpWO5qJAYYjwK2u7g6b+ixIrm03EHgKMCD1e5Awc87TyUSPKe0saNpQzK+hZDNsCDwENHfVQ+H+yRK/NdzEsdlmLcbH1jcxfHKFxYuMg2cFh7/VgYeYQz5e6uTXDYDzkfqOZW5BWQmWEe2LuiNWSjvpDqIGwKhDXKi0I199OmqKoxsnHUHqpJgqJJijElwtgLkKODBqWQwA44w+9BCMAoxX5FsYZltBbfcziZeyu1oKNOPz06QkQDYDmvcTIE95TMgBZtRFpjrT85auc3LA4nZAfeJVnzFGTYCbdbOcITtoAjmApjHABlG1u4eq6GtCdTYxXRKZr2Ek+EbgpMLposFQa7vz2mhf+bOxQe8t1kONeydLeM8InQH1ZqJRST3gude/V51qsyXyMtGtV4RN0Ht+6ys4ZyCOMyug1s0xsBpJyzZZwGuD6n6sAfcxLWAF8a1254DYJtktgbDyoIVHCBVC1AYAyrhBVNwJAOwseFQGIrEiORhMOqQ66G6yim2yYeFrFZfuCy8fWoKYFqh9XKRDuOjfF73mCkAkQLr+9zVi8+8yGdNJDur0QFIwkDxvzGyXveE6M8Nr32AMBkOwUbvCcN4oezwauN5CHGHr9N/WahXdxtVY2gFbX2FhxrhJZGxDjf91/0k8f3Ko18zawGzU4yKzA/G+8bWuAX4VABtgGtWyVJXPDED6PjVwQmYTSbTSzgQGQ0spLXuMnjHEYCFpCh4Fl2cIUprACHKNXgwGwJgMilhgeGovnI1WZgOUoBUHlHsWzJHEXkmsASQFKWdB0yR4ukDV/yBhIaYLyInWuI9s0MVTkToEjmKxkQbr0KAPVWYITn2LjdMeBjKP/oBJUqGGBRBD0cVMAwNf6+/8QlRVekWIhXNKh8ns3MbzlUo856U24Ej3JiFUrBgB4s5zv3QPukBXm2iYs7W4Oy4BIPtM3O2XqO8x2tuCxW0P++IZEALC8RJ+UWk5o/daYQA3WQvHtsUUAOO08kBbJcffc+oDYcBDjVAzeAgw21Hfc/TTzHeBfbUwRUdDpaYth2znPK8oCCLR8+fqGYKp7WoKBqbZO0Tl3KR4ngffM6bJ7FrqqhATMKw6AXHTZ+ERqM88qtNxanqCOLYxuk47WfU1OH/MqeuliYI76QDTMHjU8Uj++qlNMwV16xsHCyKg98oIBKGYqtIKiLGzwMCgzkSFffqDKX0VjZM22hXwcXh2pAB/CAY+Gt/kuWSmr0EAxGkobXuC3U5sDSSWku7NokLILKBvAGoCbO0w2tj9MBkLJcbGqYmRbrJsXthVP6PWtEk4mJgBrgKs9WqJnoKG2abnUsFTDMQeKgwkZUKAB6QOHgOZRiSzYGAWFFwg0NsPeJIzbuqRkcpgK/JZHD6SyU3e/CxdkHQDIZoeychqiiQEFwjeBpzGMtveWjAECzOiD5oKG4LFnfUS202N0BsNBEaAZkyPrgVJPyBA48E2xPRANPazHkTqpjpvxz4//zKbK5LPQtrIOeimr2ZNJzwKHr3bYQoCCBocxPIINpEdsB61dag45LbHdWIHYiMhzgs8YxDZMWEDkEMAH3sajDJlen28iZ3n2DNaRYVI7nQya2oTAnbK5mZVAPNuh6NYtDTmioMABmKJY2p00Ycqkksx9RE/fxpciWyAnqN47Vo1K0FkCEKt9OF4KJDWX9nNtD9fMkn2hvczNkZ/p4EajDZUoZGy53W+3i6o6acg4MFrQEAEXht8anuMtWtwrdpgYUatKydM9LwonU+D07RQ0ifMG8AQqS+B2S05JZrK7MRQZnp0EYwT6H7Qko53r4eBGI7ULWCMoI1iIy+EUUIOSD3U1Gm4IobAJaMhISR5zyiclfXphk6ptOS17kVp6lwJkio+JOW4AyB67aedte01HZUMi9ilskI/LfbzoAkAMUOCBvI7E+v8ukwBpg9qAmO0MkEk/r05Hbo3Oe90o4tBeAgEjAwa1SXTbGZ6EWDK3wdttmQGQdIoXWXp4XP9MQDgdreC6ypQZ2IJrmi3Oo/MEoTIjmqJZywZs4iphclIJmkcdCGP93wlsI3Doh3heouwNfAOufEMA7nVNPmYMsspR5pMsiLVDCN5AbE2wKXeBmbvA9LLCBD2kcbkNUqhRQXyzmbuImiGcFocx2Bgo14t2eVflKF1MShM12baaHrPep06bUUMJmBgBCM4E2AYLaz1WI81GuvUDMwbOG9UeLq2en33sYtoZJF37pHItKcqqqpSJqO2Ho11udTacsiavPTSpIWo2cNyQM0usvFxrp9pKOYb8MvgEZwK03cBNbqY0SzvUTcOldHGRgfVgMp4HNVbHFeqQl+yCrkqcll1rbvxuFjRxAqM4iN9OwUGQ5wNOqlw1jfouxp20Fyb6QN4cKDR6WLjve46vQc5D4w+G+8gaDCQdj1dqDBwFI3ELlaA7h5XNOCIt7nKAAEq3uAAa7xSdg20HCvlhihRqToZuEXMH1YCVKrKrhYj2nae1b8aeKHsVGiSWlCmbofJPKZmD9N4+EXseV9NTMG0qAY9n0Cm4fm0Q3urgl0btC/U+PVbT+Fw0eENB/fxdHuKzlkIx3wyIS/OfNZBujGmBmIwEP0BYBihqRBavRQpdvETIWUxoOdVf0ngwUPGuJNzU1AgrDQavIeMI5IHASHtshi+JbT1iAPToaUBIxiduOisSBhF0EUR5lVg3lUt91GXSXQ7r8NPxkW9s+hHC+eM7i6SNW0Uz/IYQL0HsTIDVGv1ATvdKSaRLXtBdTKANyqe3NHRZMErYhtWAmwqB0VOH1DcQUIENAhYBCQmvla0hJQRfdqRF8R5QDFnPZIlL3UMs+VZSiDulnN5YwwIBoHtdHx9oym3q8L/On0KAPD8/UPQ3Qp2Q1q1cjpCqAKPJp93PzBICNUaaO+pgyR7NSFyS8oBQZ704xzgG8AvBTePNrix2Kg3xKD0oWsJoVFuzwBACDCdA8UKEt+wps7AkwjNAlQHmMqjbUYs6hEiwLZuEAYAAeDE5CY2ITEAkJde2HPwEaMaYDIw2sO5xklXLTAkOd8HBojXWDRbkhiQxZ39GFRwvuUKTEH7EhDjkJR92e9Doz1TDIaobxljKtp7RhgMuONpQR9T1YjBWAn6SnC2XIKqEF0l9YvXBtVW00LVerIq5kjdhtg4S1JwVwXUtcPhokPFIRsLzfV2Thg92x320MbntmaEpaC9Y1KlQQyCDkz/yJUGly87TAMe0wMpIEgmCSnnYWNJhLYWnugajjbG+6rHABVy5bQAwrlgQBmXqVGKD2oEohNRQCwE3dvZBuT2x2HGEFxw3abmFCbm3zl/T52v9DP0Mc60cWEllrzzzGVArMrRFBCEZSzNqgK41jzRoh2xbB6fM5GJi8u+mBBAPv9MEncXMzoSiOE2zlPwaXHvHYxRinqzrUEkOGsabKsOPvAUGKXGRam/QAoukuDPAEhVICZG4dEQR0Jszbq/64mvT2KsByJItt4FkBe/lDLQ0qF4g+Zzll66S+e/EuQSUEw3slrT8s5j8+f4uFORoFu8vHue7fIpKENAQcBeRVMS/RkoitTYBSCxMYBW3KRzEXQ7SyT6M8d7hGeBA6VFJP4ztVv2ei8J7S0U8oB7azaJJavZFECk6pbktrifAjSj5LbCvr54YXq5OBsaBCEMfaUq8J70nKW5Im/HEhOgx8ij3jPaqU5ywKae18hOs3P3udY6NW6yDmsboh6GVROSq7diUMRBUy925kyZU0fKxBqjHTEtK80MEz0O9s/PjCF4pF1+DAIe+JrXiDB4EOb6gXQfuVhJMLfMnsPnFJ3ei5mCT/feLDDlkcDaHFPHJgYIDoDExlgUX8Od+n7otaJfWQ8SWSR9o7iGQq+jioM2XGOPOi7iea0MwJg2zWn+hsCSVhYkkfQ8GEibv8fHEKRzaiS2CYiBQPxKLY9X1aC1vTFqSbRF6haVOkalaG0QzXGPEsBwGOFxGnwMBoAxCrDuBc3nv+gOcLZpIGsL0+kNaoaYRx5Gdb/zXhcFH932rJ8Z2OikFEQn504qdKECDFBB0LLAS4DOk4Jg9Dl6zLp8HFQ9NnWN02WPsxsN/IJVSVyJLob1JPo5XHWojFe73GqEoYCDqs/WqVeJFNAAQEWpx3zIXgSjTEYVbAKkSmYqmGi5gF39AKbfUZAs4Ft9qoZbLNCvWvzG6xd4/vohTtctsBR0BHBvYLY3YDsPsx71dV6AYQT5AKkspNa2ymoprBdYym/nhT/+rBPUxApMra5nx5i+9nJQYinuMIHDSltxt5RamAIgYIj6gS5aL18Fbner/POOhiAzBLtBgUCbO43RAIf3h0FmzAgA7gkSTFxEtLSW3JSTPtf+eY59AVpaBKMoF3ECSudUfTWg/SCSnsHMqzemoFIicTN6zlbE29FicBbesZYSpjbAScjnpkAgmfHke5aQe4BcFT7x0ScBANU9g/aWtjE2g0BqLbFM5ZTnskfxWGwfUwkyBdV2oxUZwnq9jYcCWXosqwGtcThoBmxWAwYCxkOD/sigNoTqhMAikBDAI2KrdIZvGeOS0N0g+AUw3HR46vopFtWYW8D3zuKsCgh1pLfn966nvPCcW8RnTE7e7M0CTxJd4OavfWiZ4RWlFSg1zousK0Gm1MdOABorQ2IDpET5D95gYO1RkDppaivuqbeGj2kCbQWvnjaJ8g+B8o6fkrtgr43ApsVcUzquNblDYQrc7VbTTuRjhY+XiXkjZY12Pm+syKuMR2OcjuvsPGfh8d4GYj6eyXU3tT7nmYbiUc0AH7nbIcXaSGKBtaqItMZjWY1ojMPCjlrbyx4r22ftwBSxaLthXahYFd0CdBIAOHQScBp4xghoTvdeWOLUL3BrOMS4rmFPDew27jCc0tviPBA8JInUvAGcAzmTtQbsoQFBNEbpQ4WOdeffUiyBJJ0HKnEwWGMUzi0vAeBGvdHXriw2NxutNqg96srDGPWVbq2ei6eWJ1iYEQemx5HtYCjE1MmjRW4PwzRnzWg80ShyROo2Z3ciZWNUrSwVR/fAeCPOGYEUJCTVYgigTQ/qRhx+wqDaNBiXjHvbBvef0nQBVgF+BUAMSCxMZ7G8ZdHcNaDRq/GT95DKaEBg1Ewn1PHYZuWC6SbT8re4i3Uh21TvBC+xtJCykHQm2qsYvtGufa11OOIt2uxDoH+SocHnWmqcht2GPS8Xd7vz76M7kWkC2/kdtAulGw3CyJiXIKdxIReAyLZQtH0WVt98obiTT0GTC/M/vDue2SVyxhKI6HsGDxAhWJ4Wg2S8kwKB6AuR8tfzTVjSCjivk60XwuAshsEiDAbGx13XOA8IohYoBYKij6WdFXtAru6WweFvWJAAdiOoT0P2dEh+Famh085iFCd1Cuo7on3qKTPsttPP46soHD5yqFa6QVrZAdfaLQZvcMKC4ahGdaZ02oKg95boAs5AXHAIwwGhvylwBwHNE1t80bUXsTCjVkjFss7bTUAYWS/7MDEF2jwnVZ7sR5d736GfkZIL4+wxGGUndl67/5wrQmqvLaJW+LvallkUAyhNT1ARYNIBiM7rNgQMwSIQoWGXg4O0uI5ichmsatJUlyap5DTqd9hpn4nqTLLtcA766knXkgOCTr1RsmhUVC/i2lmVELATQKXuiqoJ8NFeGbPU4lQOeVEnQ9ULjFEsPZknpeDnUfDoDEH6POc2kZHKmAkbHgazF+57JFEX4oBxVvwP0QugE3XaQzIkcphZrwYg+B1xYU4bpD+SIuBAuXyxCxUGsfBQu0qI2vFetBkxmJq31KwLftuOGI2WiyzqEYYDVvWAhR2xtAOuVVssjDbFODBdbpIxDzAeB+YNdAKm9rNJ0JZPyUUBZFo8Lsg759bSLsBuNTdnNwx3pp0Ew0IfEwu4hd4w44rBowWPBpZJ2RzWndi88ZQQsuMbAEjaHQqmoCQd334lxBy0t9BmyhW5A+dUqkkxxYIoMrR58nilSI6d58R1mGjO/Y/gfRTe7bVczFQ+T+NCIhDQ5OMAmtI987TPRZgFUjko0F9EFXz0HBBM18G+BfjO+6UPCIin3I7Zpe+R9cj3bqJkc5nfxFLspEnS55rtXq8C9Ym+WRISpulIqwMom4fJrMurZMYkBc+RUh6Q9RvpGMUAqATWhuwvX7NDax36ymGIpWeq39HUmRBFATLnMlltPx0gTUBTa1vxhRm1QuphPvXz83WZTWJa5CV90PT4AwKJVws7Acv5P54zm9jZS+zAzO53RmqU9wC6KTMqssN6pXPDkc2aRPbTyyCYvFHiRgZBYrpudu8Qsjg/pd0tJ+pfsp5ojswmpgPioKJI+Lh51hYB6WWPGggkXDogSH2qNSU8uRIGlrzbCRxyTiZ1WvLRqjhNsjU9/EBH0Y6CSdy1Dg1GsXjOHeO+W+JWdwBzZlSQdCawpyPMZgBte4SuVwovpQyiqGzKg8eoeWT0fYWzocHtcYUAwk1zhpu8zUIOLeFTW+N5B8GKPI7tFkEIB3bAtXoLJ4zWjFiYERUFHNktGnZYmh43zBotj2hJuwMytLHTVQYEZnbRpk1UCgY8COuYbjlzje4qnNXJGbNd+E4kTurlkJT7QbsfSmUhi1pzqAJUJwNMb3D4SUZ9ShhXjO2ThNAKQiXontSgbTxkmG0dbVpr9dOPLXczW5kyFoby5JTn3llvBUqVCvuuh7McOZimLpe5ZBEQK1jYEUfcoaVRhUXgrLMYhXHPL3HLHV7JuJxtmvzhsjc/MC2+MgsW0nlIwcCoLbamXDS05t8yuFHbbhiKO8BZ8JZEgrnttuwEEeeQKg7SzOqjUVR6fmV0bCxn0ax+pOn9KEw7eB4B9AY+ELaB4JwqtsdNrVbEG9amWoOWFNouTIJJN78IJ+ZASI2RrlLtfuPDaqmsHhfKLvXXLIYDg3FFGI91Vy6xrDoJBV1DeVEgH+2gh/2ViOBbQXvY4/rBBk+2p7huN7Ds0RqHZbXEvesH6IcKIMJ4ZHUcYgrG14zumkF/nTAcAeGJAQdHHd507R7eungRDY/43+Ep3B9auJC6vtJOHxUd00hsCM6L80jieKbPl244PFrgdYVB2gPffxbc7IgcY/AisTJgXqKXkPLoyygCYAqx66tFzR4uMGrWyjgigzqucy6QulEKVM8RrwFtRpU2vZgW+BhYcx9gRv19ukdCbNHuK8pW1qEBqNaNZFvr5rE1k+IrCMNBOy8mV89UBaFZPcmN0FxgWAqZCVnyACbJ7Pyj4NIBgTHxQ852OCzTTicriZHckmKb06iI0UBBMo1RXXBvp5gt+QCsQ4PTsEAXKtx3S9z3C5wNjZYrddEueDuCuhHSDxCXVPFehWlp0Ui5SUk7E4J36ll94hawHLAONdZicyRpRDJDAUx9240o5e8t4wgdnmnuAQAa1na6FTksuc9BwBF3SP2021x2mWplH+/dlE1vUtvjUKkpR+CpzjdhnrPPi6k+DhcnDmsQWgthAo8BvB1BvcfCMuzWoD82GA8JjlVQOS49SAi+VQtaHlRVzgOhPtWFJ/vqB8l/U6KNcRK2pePTbcCsbDF5DXCUy8+bJiUfitjlMi2qtXE7DM082OviNXfmr0ZD4PrZ7ZV22yk3Gx/bn9wSyMXfxQknRHMiFaQZEM8W7Vnd+qWWzIBpbOePAXrP+Lg/SeWeJh0endNn7IjvvKrcaSCk4uJRAPEMdAweYqnhkDpsat5+XhmRPgeAyVFRoJUUl/lsl0T18Rf0b60WCEcLhMYi2ArjAcGtoNVBCz8bF8o15CBSOhjKbnCvQsRQcS49DnXAsu1xWPc4tlsc201mF5kCqtUIt7IwHcEtGNybnI4JNWFcEsalmtcsD3s8cbDG6xcneH11F4YEv8034CTOIGmjky6A2TVFiDEzsLvBJpx3mr2Q+blchPDYSxPn7z9nMmKVQWqoB0wsQGIC9v1vDCRb/KbyvirS9JXxCFY9B4LVDc3cBZK8gIeQy3v1QeSUGg9xw0KYAnGZBfapHXYVy0fj304bylQGqUZls1JKb6JHgrJuavrls96gD5MmrYeF5YAW42N0Kkw5pDDb1QC5BlRmH0D9ofX5XhgcTSFSkDA5FYZzi6JWHaS2wxyDA4M+dhic93dWP2gDDhbcNmC31GDAexWgtQ2obSBtrY6BJuY9U3SQ/uaMKk7NbjxNrWk91O2PRS+uA9MBQI48J/MLFUuuuM8eBisatNoCAS1NkeVVTm7pvebsQEq7DDCT0+K5XcL85z0fgvhYRqLpZ+p+iRUC7ALsWiPp5q6B6bQky7fxJhkpN6nh2LJ4Tq/m4Yg7U0IM4FyYhHQvYbObYVgH2RpIUwOV7vrcQiPz1ow7olYgCv1AOQCclwu+EohLK3V6IFH6e5Pb7HueWNOOL2B3PiZo0BRmF1EaO9bgW2xcTB+Uttu/+JIWI5abJh8HGuMTmcFBIIaVPWCK768W08HHe9HpguQbRuih+pRKOx/atTZ4sRuCXYu2Yt4GmC5MXRf37KbVuCfEtAZfWfVHOmcAtOy1Ug2LbwhuAbgWkNbDtF4Fkj4uL/Nqov31j6KuItmSW1EhsR1wYDocchdPvTKPB6sOd48qDM6iu84IqZEX63uMh4A7EPhlwKIesaoGLMyuEHnHBj4t7Em6TtNxJsvu3eOdfWHvOwsuEQPs4LH0NbgsUgoHyKp7G5sbJVF7amaGoKxBQw4NewTjtM+OUTOgwQQE8XDGwFmZ3DitBl2pJ82OSVA0PgMArlgDAk62+ZoydQvVHGiaSAXoJpavN9bhwA5Y8JANs2zUQgTR3T+AvPEWUSOplPY2pN4mhgQsDBvTBwzJPVsui0sHBFUVqRTHCLEhSQgEIk0b+MDwJGorLAyWoPlqJgQijMHmiYgxlfIxBPXcVx5TQ54u6Ncm0t2nY6ud+hK7WRH8qtJObQHgyiq9PTrdPS5ahIMW0lQYl1FYFr2S03BOPQ0M1qGZtA2ikWTWREDQ0og6CjZumjMc8hZP21NUCDs7/oqQjTCqOPFUxKhQP5bWumn5GjEFAyntsolOi11kBxIo5cgiJZrn2rQomAsmXx+rOUKk4mMXMHM2wHpBdWZRn2rwNRwYDIeUy6VAyMpbFWUFpYhTmVmki00MOPJjInlhyEHB9CGwK5Yj7ZlABFm2GG8s4RcGmycZ/ZMedDzgqeY0G2QByKLVMXaE1K6QNnvdvxLQdi+wiDsbmqUM0vf5fCqkYqYcSPnp92paw7FEENPnjrsRYXXaEyPgELCfmTrX1yMFA6PTc+38xMA4D8TzmdgWaSrQqOY5pjMx0CZUGy2jG88IpqOYf6fYMCl1AxXYraC978GDpozs2YAsgPSJvk5BUWT5iBAWVe4ceiVIltyVgV9auIVBf43QP6HeIc31DofLXpstdRW8NfCN1fJHEsiAnUVUAPiaMawIbkUwK4ebiw1e157hKXsfN+0ZbkZh7xPVKe68bomPNwNeODjAfXeAas35+hAL9DcC5NChWg54/eEJnmpPcb3aoCa/c21S+vspK5QurdiCOaUG9gOYbD+f3iSlELJqfmKvHrrYX/F0lrRodNE8OUsdXPQ6E2vzD6seKzPgyHZoecSx2eR0bUtjFqvfrZaoWPuxWPYYggWTqJ8KgNNFBWGG63RzI4xoQ8w5TQCK61BDWXCaUm4pIBuXhOFYz+1wLPAHAai1/HxVjziut3iqPsGh6dRBUSy6UOVOwalB4FQaSRDRJkwJG1fnDZ+JY7k0PV5nTx7p/F9eQxA9lS8cp1nKYPp5umi9aEepRIfMcZH40GPqna5fFB0FeRKO5GhabyRuDMRXumhF69tU2hZqMzXIidFyciqc/iarH8HObiwAwjOhpJZNMgICMY64ww12aGm6swwITJMdc+rRUJFBRVc4oc1giOBFNKcl826HjCGV3Mzu3Py5kzCGMiv60gixez0jNj2SbB/MIaAKArEM8jUAEx3b9NxTiKKbmHPOC+F8oU+HluvBowf/Bf77O14JCbMyOd+kL0CagKqeqMN5qiD1MB/F5p+vxL7Y776HftbZ4j9jDuZaCgKAGTtw7haJXg2Jvk8aAn1tNBy6zPHtiTTzV3L4RAwWkhYjajO4MrlGngJHQyuAo5FOdu5MVFhU5fMI2F5gtgFmiEY8g4sVJGFShaW0RBIIE4Esg686mGb9XGnX52vA1wJp1CxmUemu0lltZJZpXz/dMxpgSZ6PJPqSsInCYzPm9GEFNXTzILyuPsPZokHnLF48XCJ2S5p0IwsP2zq0reaXF2ZA8yDd0R7hlIIEzIP+hzH/dP7nKFd9+ef2ZeIiFf3uE3COMcuzL+3W5qf0jDK46dw5mGjSlsThDRsMrFVuAxsEQ6iMB5mQS0p3GAJLOU0gpB4Zvon3rqSDRPZ38S3UpbaKnQ1tAFce1kR3QvZYmj7rHFJKz7KHDee9exJyqiRW5bnAGMlkH4OKfK7ouywub0yURFECSJzoBIxAugD7QPBMWsIRkpBQJ1eIlkAw5MKd1/yRZB+cFl9V5Ts0rBRPWzncXwhcTxiOCNvBgp2gWhjYrtFa+d6DfCwhag1Cw5obXGo5EDUBtvJYVCOuVRsc2Q6HvM0DkoIZpoA6EfGzhWSExSA4l59JQUCYqT3TrKup8vMDczUFbhoUjHFhNJAYfE1nNvUIr9gjGIK1AYNR6sovgHFFYG9QHTYwdbwskuAs9QqIC21efKI+Q5gBK5P9sACm86gTvT0TK5FP9D+ytmNnp5x+n9mIGZ091w2ITPbFNJXiTSZKEnfOyH/be8ZHN0/gZ6u3n0v3GAQMoh7nAB54Ez4KeN8VSnZ7X+RLZHYOspA/Ne7KgYFMr43nTdICai45cV+woOZyxnQunVctTogMgQTkqg0mUF2rmpkZHPtLgAFbGQgRqtbAbm0WYgnHzWcs1+JRhcDs1NKaejeNWcIsUEkpAwxulw16hZBGxzmVvPqacm8PGK0LX9joVREIzILNstHmWoMegx+m0jQQ0B8xxkP1H1gtBhxXXaZsswYpphSfbu7nufD+EwuMq0rnVadBVHUwYLEYcND2qGMbX2X+qrxZ0p2tts2VKiCAYdxe3BzvwVwbedFlPVvEMP8x6UoQtTyvQnxgZ1q14CcNUEqHSOqlE83ewJMfDpPaEffBgH2F+7xAFXTeHsTm2nz1vTFYmiHX7DMpo80k6LxeG+5Y23NvTYutqcAOGA8IXWdyAAikahRMwUqMytLc4xcCtwoQK6CVQ7sYUVcOTyw3uN6q6PR19hRL7rOOJgRCFSuiLKlpERDnR1L9iuFpYzv3NEmbbBMZkUc6/5d9YnaCCqQiIQDCatsYSGuOmYCRJOY+9AS7oOrkVPs5X6TS4q8/T2AKqETzPoPohNPyiIUZcFAN+NQqYPQUaRntUmYPGaY3YK90NPmo7KxV3TkcEdxK4JbqLd7WIw7rHterDa7bNa6ZDVakN+8Ik8vR5vRyOvHr2fFfBA/JLXVDZAw8fG48FOLvPOTKAoKEudJ2EDNzYQyw5LMhUmU82Ai8FbglsmNata7gexN38UEX5Maeq+nJDpACVUcbmydr8gGmE5gu1sv72YSfhGOVGhLBUA4YkvEQR8aBtsPFi0BaKJzPu1lJO9okLAwh5mQnej44xkfv3cTG1WjNiCeaNWp2ODA9ju0GgHbBe2BJ0iOC3ey440SxExA8gCEAkFmUeVmesgUz90BJuxLZVZGnXWs8VwiYrIlnmMydYqDlA+AcMIyABMjoJqEuAAQBL1qQcwCzej6k85so+LpCtWxiq2qabJBTcyUXQP04NbRKgRxfkKJJ6SsmUA/9+arQxA6nCwvXKuUbLACrO7i20rLiNOH23mG9WmA8ZPhR0yRpQ5e2qcORBgP+wOOZ5RbXqg2WZppTkinbCsCb6xdxbLY4tlsMweKkb7EdK3SjBZFg1Qw4iOXLCzPCQDdV69BkpjUJ4tgG+ErHMrUq3xUQivadCQDm6apEJ2RaAchuWH72PLx6QUFtff6zzjEQWAW0PLumSUBW1PmVoh9OdMh1gWNDLc7OfX2wODNtttBPLOGB0YCt4woNO3gwLHtsfYWlHXBY93CB8eJihXvLJUJgbDYWNPBUnbGP+XlNjf9qj3oxwpiAVTvguO3QGIc3LO/hyHZ4fX0fT9v7eUOajJQsq6NqbbRcVYOAAON1tbSx+VFiVeYMvEFAy+PEOlwSl/chACaVNM4H9XpA6bveILsd3lRc+DBw/MqRnITowz+5HVbGA1VAqBmhAfxCrSSRqFIPCDHYyVTLO2/za2MTJqNNIZJrnSrPZxOOmFyvvr9bTK2FAUQltURtgOTHpudOAcCc8vQPCCauEoYCshYC83pXNZdiDqqmrSS3tnUtq0e+F7WHFuSWuRR33ro7zSzn7geeq9AzExAmMWKuDFC6GYKYu6SoVA8TEzC/yGh3wVMTItqZ+XbyjpIsfJWqpl4lWCdrrSCojcfG1RoQVD2uVfq4ClevSFSYjguzH2j6ZV78sbegz7BbDqpsS6ImH5bVOKcVSLiMNEIife/9+UZR0f2TTByjtJCngAwAjAYLQrRvCaHXQdQokI9/J9lWXxD46XtrJHWV1h0hsmCp5j9E9bfE5kG18WjtOFOiW3DjERrVQvlx91iFgdAIQhPzw3aMrOa4s3GYM6SJAa1Z+8CMgcHOYN7vHgBcMOipwiYEnPkWXjhfpz7wg3f+dMHXwzCn42n2ff46wcXv8/inM8UjMHc+Bk1p/QGU4U36LmWiPTwTmiheb9jlYCIIITBhWdXoFgO8Z/QkCJXJGqz9zox5jSTJvX6M1RSU5YBlNap/jUnXh5sJ0kPcJJ8X2wO6IRVMaZV9R8M0Ll40XexfhhD30gGBi4rp7PUMKDvABATaF0vn8sNRjObOHgCGivBaYjRkEXjEoQzKEMTa/U7UPKgij5PFAp+4cR3btsFwaDAeG8AT7Db5kauSmRxyPk8MMFwTuCMPaj2uHWxxY7HBM4v7eGN9G9d4g6ftKZ42Hl4EG1EvBL5gUAIRBmhrYQ/CnUgvVbF3w3wg9bOJMlyk9qR6bq4WXXzfUSb/AW1/rELIY7NGRVqjempbWArolxZMwKausOmN2qSu1DVQHeRE/eVFqW+ONdd2HbTBjmDX8z0hLXIxiJAQSwuRqOn0Aq9OfElEFgWCKT2x47KXfAZo8ncnbzRVsU95x50nRofqXgezsTisCewMfG0wHlS41x5AGPh0HTvF1TqJgwWmDuBIz+ErXtm4hNTiOu0akBiCPd95gX6G2b+1ImNqhBJs3Pml0sO8WZrKD/eDhwdZOefdPM3m9ngexSszkH/eCwjIRzdQmQfPMtmFDwOoH5RBmP2t3LcijdE+WP39MdeLcMw3+ADZdspcXBG6p5dK8x8bbG8yfAsMNzzam1scLjv8zmvP4c2LFzEGi02o0Qe1YP4kX4cbGe6sAnfTeRUCwqHPVP8XHtzBG5q7OOQtDnmLmrympKTCKBanfoHT0OK+W2A9qpagH202swJqjN7gjGtsXQVDGqQs7aCCuH6J077Buq/htwY0MGicpaTiQn6uCdFLYR5T73tn51+c//dVlR36kJjoWTXZ7P7JG9KQ1h/kUjwT4qaHgqZHE+XOYZYalLjRm3LzqXosiGoLetNjFIOVqTEKY2kHPLnUNt7DrNIt7cZ9LAvUhkkM582OFqI2mp42FHBUdzi0PRrjcL3aRJZCd/FJLzeKzcLHaT6X3KE1IXU9nPdxcLE6wZBgE5pHtmG/dEAQUpogXXD5i5CMV/KB5e9qrAAAi3NtdiYYqAq/IYsRHqtYFjYioCWHTkac8gLBEG7UazxxsMaJdRiW0Q7VM8atBXptq2rXqtAWI7lUyB158GpE1Thca7e40Wxws1rjSXOKa7zFDXY45gajeCC45Fu4s3j7KCeuoXWmQRgnUHpKWYbdphQVtJEzQ/sk1JQuoKsNp9Mam8SDSTBXUezayEO+QFa21zxZ1SEIwRqP/qiCryxCYzQ3PKovOqfmHr3S36k3vRia7IQfACbtzAeC5kWBaccv6oKXXr9rchPP+Ly9MWtqSJkEE6kk3VEm2lviQkPDGAMLBz7ZgpmxZAK7GsFSLv9J5V1CFEvONHAcDwS+vRpqNNRJmDft4NL9sy8sjGQOMLOezZM6U1aNax+BaeHMYcRshz2Jmx6AXJUh2InVRYCgrIBINPjau1Yl9UggzkyCiECGQYMCY0D1OPtT8W/VFTBrcU1E+m9O6UeaMQSSA8Qc0PQ9wtkaV4X+muaB+yMt8fOtAIcONw7XuLnY4IuWz+Ot9QsIUDfTTircG5cYvcF2tLhfLeF7M6N5BO3BgOPVFsdNh6fqEzxt76PlAS2PqKEBQTJaOw0tznyLra8xBIPB6ULj3KQod9Gqto9phCo2jxNoz4vBGQx9pSZWUc+wU62STJWA87v8ByG9Ac9etLMgX/B+V4iptA66ruRf7D8xfSXVPWWG2rLHvMmPJk6nxj8pGKjjfJ1y7UmXoaXHugH1YNys1nlOVw2CIDVNSj0RtqGGF8LaNRhSY6UorrczW+IDO2BlejW4M9vsWQMkyp93fp6nAR4muJx3FNX2zyFXmT2KF8FVlsM/VlzqQxGwk2B9FfAwSuZRG0u82kgX+f5RXrri4GECr0f46KnCYKo0eJlj9xKCsyxgnH8B0wK9v2t/pUjz6SWuxSvaYF0dHuQDe6nXTveqXHHwe2V4yPneZwbT3JPSB0nnpu9z8efbt2a/CoT5AjnH3u758xHyKnz2FAwA56+RhP11arpmPkPvgz2QfMbesQUFBQUFBQWvFj5rGIKCgoKCgoKCx4cSEBQUFBQUFBSUgKCgoKCgoKCgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIASEBQUFBQUFBSgBAQFBQUFBQUFKAFBQUFBQUFBAUpAUFBQUFBQUIDPgoDg4x//OIgI3//9339l7/kzP/MzICL8zM/8zJW95+cbyrh85qKMzWcmyrh85qKMjeKxBAQ/9mM/BiLCL/zCLzyOt3/N8RM/8RN4z3veg2eeeQZN0+CNb3wjvuVbvgW/+qu/+lof2kPxuT4uX/iFXwgiuvDrbW9722t9eA/F5/rY/PiP/zi+9Vu/FW9961uxXC7x9re/He973/tw79691/rQHorP9XH53u/93gvvl7ZtX+tDe0mUsbl62Mf2zp/D+JVf+RVcv34d3/md34knnngCzz33HH70R38UX/EVX4Gf+7mfw+/+3b/7tT7Ez0v84A/+IM7OznYe+8QnPoHv/u7vxtd93de9RkdVAAB//s//eTzzzDP49m//dnzBF3wBfuVXfgU/9EM/hJ/8yZ/EL/7iL2KxWLzWh/h5jR/+4R/GwcFB/rcx5jU8moI5Xs2xKQHBy8D3fM/3nHvsve99L974xjfih3/4h/EjP/Ijr8FRFXzTN33Tucf+zt/5OwCAP/En/sSrfDQFc3zwgx/Eu971rp3HvuzLvgzf8R3fgQ984AN473vf+9ocWAEA4Fu+5VvwxBNPvNaHUXABXs2xec00BMMw4Hu+53vwZV/2ZTg+PsZqtcJXf/VX46d/+qcf+Jp/+A//Id785jdjsVjga77may6k6H/9138d3/It34IbN26gbVt8+Zd/Of79v//3L3k8m80Gv/7rv44XX3zxZX2eJ598Esvl8jOeAn0pfK6Ny7/+1/8ab3nLW/CVX/mVL+v1n0n4bB6b/WAAAL75m78ZAPDhD3/4JV//mYzP5nFJEBGcnJxARC79ms8GlLF5NLxmAcHJyQn+2T/7Z3jXu96Fv//3/z6+93u/F7du3cJ73vMe/I//8T/OPf9f/st/iX/0j/4R/tJf+kv4m3/zb+JXf/VX8bVf+7V4/vnn83P+5//8n/h9v+/34cMf/jD+xt/4G/iBH/gBrFYrfNM3fRN+4id+4qHH86EPfQhf/MVfjB/6oR+69Ge4d+8ebt26hV/5lV/Be9/7XpycnODd7373pV//mYjPhXFJ+O///b/jwx/+MP74H//jj/zaz0R8Lo0NADz33HMA8Fm/M/1cGJe3vvWtOD4+xuHhIb79279951g+m1HG5hEhjwH/4l/8CwEg/+2//bcHPsc5J33f7zx29+5deeqpp+TP/Jk/kx/72Mc+JgBksVjIJz/5yfz4z//8zwsA+St/5a/kx9797nfLO9/5Tum6Lj8WQpCv/MqvlLe97W35sZ/+6Z8WAPLTP/3T5x57//vff+nP+fa3v10ACAA5ODiQ7/7u7xbv/aVf/2rj82VcEt73vvcJAPm1X/u1R37tq43Pt7EREfmzf/bPijFGfuM3fuNlvf7VwOf6uPzgD/6g/OW//JflAx/4gHzwgx+U7/zO7xRrrbztbW+T+/fvv+TrX0uUsbl6vGYBwRzee7l9+7bcunVLvuEbvkG+5Eu+JP8uDdS3fdu3nXvd7/29v1fe/va3i4jI7du3hYjkb//tvy23bt3a+fpbf+tvCYA80BcN1MvBf/2v/1V+6qd+Sv7JP/kn8nt+z++R973vfTIMwyt6z8eJz5dxScf+hje8Qb70S7/0Fb/Xq4HPp7EREfnABz4gAOS7vuu7ruT9Hhc+38ZFZBqb7/u+77uy93wcKGNz9XhNA4If+7Efk3e+851SVVXeaQOQt7zlLfk5aaC+53u+59zr/+Sf/JPSNI2ITJHcw75+8Rd/UUQez0DduXNHnnrqKXnf+953Ze951fh8Gpf//J//swCQ7//+73/F7/Vq4PNpbP7Lf/kv0ratvOc975FxHF/x+z1OfD6NyxxPP/20vPvd777S97xqlLG5erxmVQb/6l/9K/zpP/2n8U3f9E34a3/tr+HJJ5+EMQbf933fh4985COP/H4hBADAX/2rfxXvec97LnzOF33RF72iY34Yrl+/jq/92q/FBz7wgSs1t3i18bkyLh/4wAfAzPi2b/u2K3/v1wqfC2PzS7/0S/jDf/gP43f9rt+FD37wg7D2s7/Q6XNhXPbxpje9CXfu3Hmsf+PVQBmbR8Nrdjd+8IMfxFvf+lb8+I//OIgoP/7+97//wuf/5m/+5rnHfuM3fgNf+IVfCECFFwBQVRX+wB/4A1d/wJfAdrvF/fv3X5O/fVX4XBiXvu/xb//tv8W73vUuPPPMM6/K33w18Nk+Nh/5yEfw9V//9XjyySfxkz/5kzu11Z/N+Gwfl32ICD7+8Y/jS7/0S1/1v33VKGPzaHjNqgySuYLMSil+/ud/Hj/3cz934fP/3b/7d3j22Wfzvz/0oQ/h53/+5/GH/tAfAqBlf+9617vwT//pP8WnP/3pc6+/devWQ4/nUcpBXnjhhXOPffzjH8d/+k//CV/+5V/+kq//TMZn87gk/ORP/iTu3bv3Oec98Nk8Ns899xy+7uu+DsyM//gf/yNe97rXveRrPlvw2TwuF73XD//wD+PWrVv4+q//+pd8/Wc6ytg8Gh4rQ/CjP/qj+Kmf+qlzj3/nd34nvvEbvxE//uM/jm/+5m/GN3zDN+BjH/sYfuRHfgTveMc7zrnNAUrDfNVXfRX+4l/8i+j7Hj/4gz+Imzdv4ru+67vyc/7xP/7H+Kqv+iq8853vxJ/7c38Ob33rW/H888/j537u5/DJT34Sv/RLv/TAY/3Qhz6E3//7fz/e//7343u/93sf+rne+c534t3vfje+5Eu+BNevX8dv/uZv4p//83+OcRzx9/7e37v8CXqN8Lk6Lgkf+MAH0DQN/ugf/aOXev5nEj5Xx+brv/7r8dGPfhTf9V3fhZ/92Z/Fz/7sz+bfPfXUU/iDf/APXuLsvHb4XB2XN7/5zfjWb/1WvPOd70TbtvjZn/1Z/Jt/82/wJV/yJfgLf+EvXP4EvYYoY3OFeBzChCT2eNDXb//2b0sIQf7u3/278uY3v1mappEv/dIvlf/wH/6DfMd3fIe8+c1vzu+VxB7/4B/8A/mBH/gBedOb3iRN08hXf/VXyy/90i+d+9sf+chH5E/9qT8lTz/9tFRVJW94wxvkG7/xG+WDH/xgfs4rLQd5//vfL1/+5V8u169fF2utPPPMM/LH/tgfk1/+5V9+JaftseNzfVxERO7fvy9t28of+SN/5OWeptcEn+tj87DP9jVf8zWv4Mw9Xnyuj8t73/teecc73iGHh4dSVZV80Rd9kfz1v/7X5eTk5JWctlcFZWyuHiTyOWZNVVBQUFBQUPDI+Ixvf1xQUFBQUFDw+FECgoKCgoKCgoISEBQUFBQUFBSUgKCgoKCgoKAAJSAoKCgoKCgoQAkICgoKCgoKCvAIxkRf9pP/DwAAkYBJQCSoOKAyHkyCxjhYCrDsYTmAIbDsUVHYeZ8AQu8tAgidq9B5Cy+Ms6GG8wa9M9huGoRACL0BegYFAg0E8gSaV5qSfgkBYgViBDCANB4wAlMHNO0AYwJuLLe41myxtANe397Hke1gKKDhEQCw8Q36YNEHixO3QB8MhmBxMrTwgdF7i85Z+MBwnuFFbTAlficSmHheaqvnoDIeq2pAzQ6tcTisOhgSWNJzBgD/6Ev/X69sBAF8yV/6f+qxEAGczgcQDCAG8C0QrEAqwNeiz7ECsbHi1AhAks8naFaJGj/fdL4FRACZADL6POYAZn2cOWDmELo79oHgHEMC6/iODAQCdQbUE9gR7JrAA2A7oDoVsAOqdUC1DmAnsGcjePDg7Qg62wDOA95DnIufxahFKRFgTPzOADOE4+PpK4EveAzAT/3K33lF4/LH/68/BwCo2IOh52oURhCGE8bgDZyY6TAgYArgeI+l+wkATBwTL4QgDJ7dVwsz4sh2YBJU5FGRvqblMf+c8OJ4gNvjCkE4H1cAYeur/N4JQQgBBEsBN+o1GnY4tls8WZ2gJocb5gxH3KEihxWNqCjA4+LB78TsvDeA/BmCMO6FJdahgaEARoChgFEshnh+5q/942/70GWH4IF429+N9wzrFwjwrUDqALAA9XR9EwlECNIb0MCAB8yWQR7wS4HcGGBrh//j6Vv46pv/G8dmgzfVt3GNNwAAH/ddBgEGu1XeTAErcmAIWgo4ZIrPPX8ePQSjCDyA+8FgLRZBOJ/zlhyu8YCKgJYIVXy8kwAPYBRgBMELoaWAau9PBABe9HtDQE0EBtCSQUUGnTicBo8Q36uLY1NRQBU/1+9403n3vkfBm3/s7+s5twG28mATcO1gixuLDZZ2wDMLnbuP7QZP2fuoyePQbLGiAUwBNTyY9Dyncz2AMcZjHWP7noueyyQIMp2UNEcbSL5/p3NFGOO578SikwpeGCZe0xV5tDTCQFBRQBvvQ30vPceDMAIonz8Tx21J6bwa8GzPHqDv7UUQELARj0Ekj0e6DgwEPo5zwu/8gsl98WH4vGUIwgMmrgc+Xx78fHnI7/S1/NB/XxkuOoxH+5gXvH73Rthf7F/O2xMV64tXC/MFOv28f/2liePcdQp66HV/WTzoPR52H/jHdY88DnweXM5+ZlfjBecCvxC/Hgc+2+aL/aDvlSLE/y6L+dgE0COtdVdiXcwxusq7m/hvE3c6O5AYeQl2djkPAgnlGy6xA+kt0zVKgF6NRBASIJAyBwKEwCACXNBdmROGB8MLgyEYg56CUQxGMfBgBCG4YOACKyMgygiMniFCcHGHq8eQvrP+2XhwwRAEQM0GgQk2Mg7p3DxqQPIwuGU8BsLEELCyA2IAvxBlCypBaIKeMBsAM9vts55gwu4NKKKfA0I7v2cTYEyYWAHouDIH8P7OQ+L7CIHZQAQgZ+AASCCI1wMOBASrfys4PfYggBhCiNsZMQzhADGUz6CITBdDEMBcfG4pCCT9TuR8dPMYMd9thHyoFy9683vGBRP/vce07bEEADI7kFiBMV4APu5EvDDuuSXuj61ew/FxFxhbV8EHPne/Eglq9ghCWBhl047NFoZ1B++hdNQagHmAx5kH5V3yuc8dF5fEBHjhnSgzT66XmCseBexiEGQFYgFh2Ql+aXYvgwCCxPsA8QaLhxT0Gg6BMXqDTahRkUcXaoxx13oRMzCHB+UdaIDu0kZIXoRH6CKcoM8R1AjwJIAYDDD7bwsmwigBI3QHmWBIUBGwor3AEIJudpxD/PunQdm3+6HCx93rsA4NTn2L+34JAFjygCX3AIC3PfBTXg5cx5209ahrD8MBy2rEQdWjNSMOTI8lDzvM1ygWAzwgBoFYGabZOfegzAykxfKiOy8Fruk5w979md4v78AxMWrpvkt/lyMrMDEEygwk1iVAmZtRCds8ZXkRdPAw0LHT45HI3ig7lJicThgjDIJQPh79HHpMnVT5vnrnJc//KwoIErVJkQa3HDLNaSLlWfEuZekjDemFwHGy84LzgQOwE3nvpAri3EDpfxTpcoqTfAoIPCEEApH+PRcYLjCGYDGy0wsgjnkKCIaYNnCiz3OiQcHoDZzXS8J7DQh0kYvHEieQtMb4wPCB9OaLaZV5KsVd4Q5oTE3jKE5siAtrpEN9GzQ9YAUUKVG2aUEXWBvisU9pjzwEoucuBT5pzJkDajvR2ZZ1UAyfn7h9PO8CYPQGIoTBBPQkmkYIMeggjgu/IIwaBDAEvtJ0gq49BPGaAsgnOwjgPUB88Z0+DxjmA6Yf6FUJDPLifY4tohxE7yMHjYJzQdY+KvIwFJSqjGmwU99iE2q4wOhDhVEYd4clbm0P4ANjO1YYnE4ow2ARPOvYU0r/TOPshLGqBjAJnqxPwBLiPWMxQief/FnjDRqglKrSpNN1kR5PLEDAPmNxfhCvetfFg37XlIFADHbpLhLQbtyS5x9KXwEgT4BjBCPYugpnrgFDsA41VlyhgkNNSk3PA6H9hWm+QVA6WTLF34nBKJxp7URhN+QjdW3yOWKcvwXm9D5D7/GWCEfc7jzPwcOEUQMDEfSii+L90GAtNZ4dr+MXzt6Ce+MCd/oV7nYLAMBBNWBh9Zr7c29/pGE4h6bV96msx6IeYTnguN7iuOqwMAMOTI9D06HhcZZGY3So4rljcFxIE33/wOtpL50GYGehH2BySia9RxpLfY+Q36siPe4aIY9RSx4VARWAlhhMhJYsLAwCBPdDlwOwkP9+YmMkL/zpGghCOf2hG9fp2NI9NIhBAGMQg01osA4NAOCyrZCurLlRXigSM5DyofF7vhlIL0rQFARcNBlm5CCAdhmClNOWuHnN+e845wcAQroDZYIPcXcvJgcH4OlGTKyA5lGn4MHn7/p4CPHL6+dJO+h0O6fF1Bgd4jHoRWcpYGCT2YGHfuZHRKj1vWR+DhgQoxOd1KJ50SqAK138jfU5EKiMj9oQXdBTzjQhMSPpOek1VXxuCnSSvmQfI5n8noSUBwecYRAxvAkQpul4Q2Q3eP5F+bsyQReskBIACfFYX73d/+NCmAVhl8H8mvKRKtTdA+cgt/OVBreB0TuDcbTwnuAGGwNpAXG8P1lAMRLpnYWlgD4YjMEikFOmLV75acI0FOBhZnlMhgdgiHYeA2JgcMFkPc+FPmxSf0WY37Tp3+m7kM4dM8ZDAl3glk8gkfzvdK4fxv6do9qFdxYmLwJDtEPBp1z1/nHvzyEehDFOgAECLx4jgHuhxlpq1IgMEgJGcTuvNcRwMt91A6ehQicWz/lj3HYH+PR4DZ/cXMP9YYF72xZnGw0oTpsRbbX7fi8Xad60ZtJhWdaNZUUeFbsLmeUAZQb0PPF0rh4B82DAx0DLx114WnQ7CfmazMc8C3hr8qjIwUAwkpvpB0LUhTgsiTGKvzAJMH8s6QJG4RgUagDeSaXahVBPwctMI5EChrO4IXgUXDog8JEiN6wHnXaShgIsB7TGoWEHyx4LM8JAH59TO0H0JFvyCMIYg0HNHgOQqehELUuMvGkgUAB4BHjUoIAcdm/oSDUnihzEurgAcGQQAmNrK1TGYwgGB7bNoqqe9RSsXZOFhGdjg95b9M5iPVQIgdENFcbRKD04MuB10oA/f/OPNgCswpiuqWBMQGU9lk2TA6aLdtIvF8PrXD4PWRzIOrETi1JvJsBaj2U9gkjQWofGODAJWqOR+DzdkwRlALL4LbEbAFCz19fHx8xecJdYoCCMPhi4YLKQNIB0N2U9nDfYcMBgKviRERxpOiAAvtdgjryAvQYDvmZQEPAwYwgkQEYHMKsQLMSoMU/otJsimLMDc1xxW49EsTMFVBR2gs90jnJ+neapN7ngeZwfmwLKKW1Qsc8TVWK7ulBloeza1RiCwUnf4t5mAecYw6aGdAbkCdypaFfFp8pjuig2HWtNDfW1heGA2/UZ+pguGMWAEaad0wNOYQiXX9AvmvANhSsNCtI8KaTpA4EyiohsQYCmpwDEa4lAA4OjuJldXMcDpiACM+qYpkVigE5McwFgQht3lum+MbNAlqE7RiAFDik4IoTZfTpAx/tUWqylziLNe16Fmh/vnsCJa3FgejxZn2LJA97RfhLAfVREWFKFCgajeHQS0Anwa+MT+KXNm3F7XOEX77wJL5wcoNvUoOdamB4wW0K70c/uW+AsrSR/6JWNy6pR6qY2HstqgOWAlR2w4AELM+bzW5ODQXhg2jkJUxOS4C/R+em1c+FhEM40+wiDU7/AIAanYYH7bokAwqlvsfUVnBicOT3X87myZg0CmAQr26Min4W4LQ140p7ihjmDgbIGFU0sQz7WOHdu4uK+lhq3/QFGsbjlDnHXrdCFCi/2Bzh1DQZvcTo2ymQHxuBieiRMzO7fvmTO4NIBQXrjIKqGnFPIlgIadliYEQ2POLA9GJIjOn2d7gZG0V2yC4yGKwyxX7XlgMHH3tUxL6c3ni4OZiDQGGm6+DW/t8hprlnzgVFlHwMDHwRjZbA1FUJFOHNN/JseDevfXLsanbdwYrAeawzeYHAWXQwIxsHCd0Z3UaMeGwKB/DydoTsssaw0pGUMjkEmYKwCnOecczeX3PVdBu31Ts9BDNLmeX3DAYt6RMUBjXFY2BGWPZZWbzBDggXrjWcwpXh8vNDz7jKYrBEx0PFu2EVV+LSTSzfoGCz6mLfb+BqjaFDQW92prk0NhuTUCREwGoNhMJo6cIRQ6RiHkeBrABCIJYSKESxPWVOJKQMRjQwpZDZCH5PzuoL94OAxoOFp16SLJU8B0yXFevOAYf5YAMHGbHJi5CrySp+GKmsDXEyDbWJAcNrX6La1BrWnFnajanm70Ws5VATfakCgVSoCGRnbqob3jNo6nLgFPKbPUpHHintU8HGx0jGdpwlUb8CaY03XygV7pPlCmioOgPNphVeKYKf7j8Z4GfgZE0UMWFFNEkXNwBjnI6+pgunep3Ox5PxzzhXuc/YjfTfodZMwG+N5MJDzw6JnY0o/TQr6USzWKYUjBh/tn8Sn+ms4cQ0+fnITZ32Ng2bAMwf3cWh7NDziaXOmeW52MEQxVaC56Y/2T+EX7n0BXtgc4tmPP4HmBYPFmrD6lKDaBthNQHWq17dfGARLr1zADGBVa0BQxTnKcojristVMzU58AOCgf30VGKlJvW/mwUFSu2nYGCAQScV1qHBKAb3/BKdVLjrVrg1HGLrK9zuVzgdGgze4Kxr4CJ76iNjbEzIqba2cqiMx2Hd402re1iYAW9o7uH11V2seMDT9l6uztkN9pQVWEuNLlQ4DQs8P17DJtR4tr+G57ojbFyN588Ose5quNFg3FaA0w0qjfOF8dHO/6UDgjQxUbr4044wigl1cdWBa8jlfGZiCDzFXU7QCQSsC3J6/RwiUQeQFn8fv+K/OTEEsw+d53WK59aTBgmOARF4pzSp8QaDN+jIwsYSMAAYIjswxt8PzmJwBs4ZBM86gboYBIysxxAI7LGTWwRUEAcmhEoQSAMVL6RMCAcwM5ivLiCwMZdPUfQHaCBgWJmIeka/1bE8tIpfOm4ul6ClBX6EQSACi8CB9ecYDEypoSkYuOjmTCVtSUQKzHa9e9+BuEGff81AYXqcvChVu/PHEk0YZVlBgBCmx19DzM/NKIzOa7mYCvsoBtU+n9e00DqZfu9iMMEkaGP+1MbzXsfADEjlUJpv7EOFMRgtmfUWvbcYnIUfGBgZZiBl3lz6ng9YtShV1JCYJJyL6TRhuGDi32GMAAaZppK8k9+bjMwFAcDD4PeEk/tU7StBPlyaXWqJcDKaYpOXukd3Ug7KMmQtUqSbH4a0qDEkl+35LORjbMSiE4tb/ghdqOBn5XNDXPhHMbjvlvlnF/Q5H1/fxAvrA/SjxcnpAqEzOFspG3FQ93jL4hDrNgnt9Lz2EnA/VDiRBp8ejvH85hD31gvYE4PqhFCtBfU6wGwDbOfBvYsBkYCrq7nP0oakNl7nrXh9N6xspGbIpw3VPK00R2ID0nMA5CCgimK/LPQV0VQLNAc/igYGm9CgCxXuuwVe7FfofIVbmxVOt+3EsLm0VqXAXYNpsGDdeDALTpc9RAgLO2Yx8CF3OOQtWuN2NwoxGEiBngYoddYD3RsXuNsvsR0rnG4aPYaBwWsDcspc8aDHIiyPXEd46YDA+Ymy1NiVMzvQGhdrQ7dY8oAD06EmFwMCF6kyVep3UuO+X6CPdObWV3lRkJijFz9FOjxoNM4DYAYAMX1AYfdmDRUp1WcBgHSj6AFAWQJnDTqq4SoPazx6b3VXHC+m3lsM3sAHxrqv4ZyBGw1cZ/VYeobpI2vRx/RFOpaoWciBe2QpgiH4JSEYzfO71uSLJdP7V4CjRWQIZo9lISOFzArU7LGyAyr2WPCgNxmJqnbZ5ZsFmFVdxEUJYab5iMFARVMd77ymHAAChTx5zRf9MS4mLhg4idF1XGwk8BQIZu1IDAC8LljsRP/tQtz9zyaDnEKQy0XGj1lnkM7JlAIAToYF7nSqzjYxTZNSbpbn6TU9j040IE3szM1mjdfVZ9P535sQu1DhzDdwweDeuEDnNQV2e7NC7wzWpy3o1IIHQnXKsFu9hs1WPR9CBfhG0zau1dp8gOBGhreMwWmAUbNDH2xW1Qfh2fWgxzqGSfme5oJ9zP0V0neT6vbJ77ANVxkQuIP4XlFUCIoCwZiKlJj2S/olAFPwEH0LkoaJvM5Z3WhxOrYwCDitFliFAYYCWhrP7VrTTrWlES0FtKSzqgo0CR91N/DceA13/Ar/e/Mk1q7B2djg/tBi9Ab3t60yPcmvJTKW6Xiqu4z6vi4Q19cCHoHtkzV++60NeKWq/XcsnsUNc4ZDPkMrHvcC8D+HZ3DLHeLnb38hPvWJmzBnBscfAQ4+5WD6gPpOBxo9aPTqAQKtZnhJ1eslca3ZAlDq/cAOsJFyPzRdFsym9FRNfodNmp9XAPl6SwFAUvzv0/SnMgVbm9Dgnl9iEzQo2voav7W5jk/eP0Y3VNi+uIQ9MeABWJ3ousQO4FGmayduAt1C14L+cIHfeOIA1AT81o1r+OTxNdxs1mivjfp54LP4ds743PYH2IQGt9whPtHdwOnY4rdOr+PW/QO4wQAvNqhPCWZLaO6J+rf0AtsFzWbXU3XWZXF5hiBwzPHHhVumqItJIoWsKYMV9zAIaDneCDQTEAVgQzXAaZIIGPYXxsgQJBVvZgjiFzuZqpCSlmAmhAtO11zhpDcgiGOEwHBO0I82XkQqjgOQWQHnGeNo4J2BdwyMiRUgDQI8YPoph8jDPI0RxX1GacdQAWACm1g2RFpSpTuQRxqnh6K15yfaKhlEkWRWoJnpPCyHTPlX7PJNsz9pMwVwqHeCAWAy0UnPOVceJufLoFIufF4bG2YVDBmprCt9xXObdiPkoz5gzhLsL+4SYlT22rEEF4kBnTA6Z3VcACBqSZKxV0oHBNnXEBiw6ISWJse0yI5xJ5nymUOwGIOZWK9g0I1WNTADw/QcdxLQ4NYBJn7X4dPyzGABcaSagpCEdhTTEbHcKQYumhIxYETqWEJWQnNcFPdhIuMEAD7u/VQYpqkFnSQfD6SKY8OixlyIgtW0+Kd7NDFT6XPFICALeIEcuIbAGCVVdJgcEHsi7N/wc9q6ItUO+FRZIIxTv8Dz4zFujys8u7mGs6HBaV9jvW3gnIE/qWHOGOwJVQ+Qo52NyeJFQXvXg53AbAN4FFCo0d8wGANwt1d9wZL6XNLYicEtd4gXhiPcWS9hTiyqU0JzL6B5sQcPDnzagUZ3/v67ItSRplqYMebjPRrStSVd83kTshcMzDGfk5LiHwBamgpgDSFXuIWg69OQhHuhwtbXWPsap0Oj530wMKcG9T2C6YHmjsD2GkibYaafId0MjktGsLpjD8YitIKTeoEXa70XUmpC9UNz/Y8KB7ugXxvf4HRsceYarPsa47aCdAbNKaE6Idgt0N4R2D7AdAK71nPoVha+eUwBQVLPU14YEEsMQ87vtOSw4l7zieR2FphRLDwpnZl2oRVNu1iluKPLHSXKLi7q843BDk0X9T7Q5wjt6gso5vhV+aJiQBGgH22siQ9abQB9zAWG9ww3WhW3jQwaNYjhYcZWuMgMOMAMspPWSKV/Kh6OgUGsrU87DDEqXnpcmIvS9helJGRLAs8AAoeYM55Ra2lCG8VEJqeevf+0cM3TDOn9AWRBWxDCNtTovcUojI2rtLTTW2xGrX0fIhvjHWfai1wSksYd7CAwo8D0AZx2KD6xBPPAIIoN0/eLsF+C+Jjgwu5yFkBozYjjtkPFHgdVH1mbPjsNbr2eN4ZggyqPlY8B+dZX2IQaDTmwCage8LeBpO9R2jUHvlVQPwqoRiPrNGpM3hVE2aAyLXYIQPB6HHPGJzEVSx7UmY12Kd2E+c5t/lpGUBYgAINMQQLvTfIXTfqvBNzFKqHk2ElplQeyMBd6bjQwkFwFA9KKGICyy6GmLS9/fOo9oJiqCVTZfyIN/nf/FH7t9PW41R3go59+AuGsAm8Z1Smh8oTFBjAd4gZFdJOUUmpBg4H63qCblBhMm76C6OqDilX3UZPHqVjAO/zm8Az+693fgRc2h7j36SMcPkeozgTt7RH2pNPgOgScc/UM4crupZwKoMl1M21YlHV2OU2QGKOcEsB50SCAzAqk4CvdMx6R2JBUoWFycJ3o+Y2rcNI1GDYV0Bu0a4JdA7YT1Os0Hwl4CJM/DgFiGSRAsATf6CYSAPxg0DtN4Y2zTVOI+pCkYeikwn2/wqlvcdctcTK2WI+1rl0j60Z5Vnmnf5eido3ycTwqXlZAkGvWOaA1IxZmxJIHHJotDrnDNd7MAgIf7R01avbMalE6E6a5qPjPQYERiJdM5wnrRSwP0K0kQble/MospGoe8ogKYYIQQzyh5wrOaSlcKnNxzsB71lxpZ6IGgbP6OilrKeiNqDlXgemjrsFLzsGGWNPsK/1duihSgIAY6FwV9hf9eTlSWlBC9GJQalZ3mzkNwOoPkLQd2FOpn7gWa9fs5HMXZtSFK964c9vPlL9OAcHaR5Fmvvd2EQAAqn5JREFUsNg6LXvrnEU3WnjPGPoKfjCQUdMyPBBMr+ecR4HtBHYr4DHAbEZwN4K6US2LfYi6ASgTwDEYSBTmg9iB/UAiPXaF6MP52+u46nBY9ViYEU/VJzg0Xd5BezBeHA9x1y0za+bizmU76jR2Wjc4cw1GNsrAcdwpyZS7T+yNpYDAhMpoTbdxgr72GNsAMoQwMLxXIS4k6mEw7X4zO+MBeGXZUkCQtCFN3AjcNGc4NEr37kzQWSw1lSXuL+ypFDGVTu0HA48D9kwnExVRhkk3Mb9cKNKPkS2QoL4ZaZ5himWyKQU4wzyN9iAE4cgeqHZgIwa3wxK33BF++eQN+PCtp3B2d4mDX6/R3BXUp4LFrSGnzCjod+5GDZBnKTM6WUNOz9TKe7UEKovq6QUAAxitLDrkLVoacS+0uCOMD63fiv/fx78Acq/G4ccMrv8vB7vxaJ69D9y5DzIMtA3ArBUYVp3DaHS7qbtXMi7xnNkcDHi05NDSgJr8Tllfwpwp2B8DAKgwGQO1RKiiwcQoYRaM2bwz72Pa7U6/VGbmbAE6qWC2hPpe2o1roGQ6D/JBz3+ULwGAGIbpKxVAG4vxUAWpbmuw6SssKq1QmJsnAcoa3PEHWMeUxYlb4M6wxN1ugW60GPoKNMTqu6Srk2nzLEYrsfQY4ib0EVKjL8uHYGIJkhth2KEw58HARNvoTiBFdvPoKC1gFN+TKH1AiTfqNEnNa+2BPe3ZAz43AUi1xfBQwyLwTkQfvAYLkgQiQZ+borDUR2GeHtj52U8pA4busJhFX0cAeyDEARSDiZp8DAggsEw7jxwUZBFlgAsmlz6OwaBij1EMjIRzDEHK98+jGJPGXiQLDtPfAhCpUxWdpWBg9AajV68HHygaPEWTp0A5D5pZnllFCfm4C/JBJ0MXqwrkaiaix8EWXFRJkEo0Fzzg2G5wyN3OTvmMW1TkMcRbMzljpqNLojEbKfXdcd2FpvM4GoQFBKPe8DCa+gsGYBvnsWhDkFm3GUOg1z1BArKwMOX8sysbj+fSAvNFcVfYtTtmnOcFyU5vV6kXuAjkNeIhEz8bBETJe4DitbU3yST2kmg2Fz0aK5C+8znqc3LUG8Rg42r0XQV0DHsG1KeC+r5Hda/Liw+JAM6D+iHn87UiQiDrNcJmA7IW1NRaLy5x7mHZKQnvQoVOKtwbFwhri+qMYTcC03mY3se0GwHGQKzRHiFMGmyTQDxdtZEkAMzKoENeZxILAEyBADAFA/vX2zlWCgQDragwKf+Tzn+8j3wSzQaDMbLG7HQO39ExBbl4Y3GRpkKAB8W4WlKqP6d0xZyd7fw0dwZPk14tXqLJwyXYxBhEHVet+rVHwaUDgmwYEdXrlfGoY6nhwoyRHdjiiDsc8rAXwZE25SDdyXAUHamHgYcVRm28mt1YD2OjQK1m+BogI3rTxntIbEwFyG7KOlGeu54EUyoBLgoPe4OQGpewntUkZEx6AYpVCjxGpiHS2GnHxLnqQeLPkv1FKEicSwhmQFTF67/nZjtXhbNhovPnzZYAvam2ptKdfCzlYVI72n1HybnPQKL1+2Bxb1ig97uXSmMcWjPuKOPnuoDBWwwxz9zHC9oLoR+nBlFDX2lOujPAyKAhitx6zYtVGwGPgvrMozqNTY1OuzwBinOaNgBAaZKi9PlpYgdeyo3wMeVD5wxBSt+k9MCx3eCt9Qu4ac6yoGkUqzsEtwJDMqPiY8oAADpf4dS1cKIMgYFEMVStFQXB5nHQci31BamN0yobEtxhgXOMsa7hOwMaCH6tzIxOeMgTmBkAQH0KAhkMg8V6bLIzZcO603ydOcG1aF97UYOjZK86BQbThD0KY0MNOlRx8g8P3PVdFVK5IHkBD2p0lehXYUAqFSOnFCCggUPSt+TpTSgGSxJNzEw2EgrCMLEUVJsL7ZYdps92k3swxfsmCdu6BcZ1BXtmcjBQn4zgzQCMTgNiHyn8yJSRYcDGa84YUF2D6gpyuIKsWvTXGP7YYXHc4YlmjRoeAwx+rX8Dnu2v47/feiMWn6zQ3AXq04BQM0bLcKsbIH89aqN0s2M7D+482AWwD8B4NcZEzUxDkMTOKrzUdFRiCNI5TNhnYzQlQ9kWuhOt5PAkqBCbR8V4oEKIG1kziaNBWnHmDMQxzEggR/n6CJYwHFnwaPScxNMeon5Md+qqwxmOCMN1ga8BPhxxfbXFjXaTG+upyVATPQ9abEITUxZL3B20quD+2QJuMJC1hV1z3lj6FvAN4Fa8s4ECkN1qHwWXDwjSDRyDAUOipSHsspf1EXdYco8VOVQ068D0gNUvlZFUpKUlNVsMRgMCEUKoAkKji7OP0W2i3aedI5D/BOmuR8V80Jxfoj8DciQrYCBevzn4j5EXAsUqBg06kpCQUxDgNT2ApBuYDQI7fbdZwKcBgZnldOjqA4IkkpTIBMyXNoK6fqVxG4PJmo06l/i4HbEgMJVhDt7gLNbdzlEbj8ZWORhIplJpMRq9RtdBSO1xg1aZJFbAjQZ+ZLUhHhncM7gnmC6lZzRNYEaB3XjYs0HVzestpOs0Kp+zA2Z2fPPonB7w8z4eQ0CQPBYSQ5P0Nk9Uak7yhdUdPG08OhGsA6MXgxf4cKr0iKkVPTw99j6akIzBYGXU+GQUs+P1kHbvTcy9ejAObYcgnMtPB29w0rTouwpuMBhNBe4ZZgDMZm4Gpp+Fh0ixD6mMUVMYK+5xZDrc4A6vM2q3mjqwJXhJhNtMe6B3IQwBo3jcC+OOWGw/Pzz3ar8SxAMknzf8Ubui6Ulfp50XosZg9tr9SyUGF6kkc87YpF4PiRVNFs8p4KnIYRBGQyEqzJW23gwVaGtg14T6TIMBc9aDNp0Gw8MIjMPeYdS5QyPYKDvQtvAHC/iDGsMBoToccONggxvVGhV5rEONj25fh984eRIv3jrCjU8LFrcDSCRWbhGGA4ZrMeWrBWhOGO0dggwBvDHnd8kvE2mRbHnMvgP687BTRbDPNCW/gTnSGIyxwgB0cYugpFeoUy+CeO25wGpX7+LmMK0ZsaR8jBuOUGlFjvrfUO4yGyp9zK0E41GA1AGHBx2eWKxxs1ljxT2YArrQ4CS0eUOwiWWGp7GyZK5hMGuG7abAxLeatvI1dnRpErUvL1k6u4eXpyFALJmKC/quA9Tkpz0310gq5H2h0E4pWzTsYQ5gQ6oliDdjsHoSiCTSfXpC0tw/TyvMS4nm14hqC3SBFxZtnAQgCacoex/MaJkZdY3ZDZGYwv3vc9Yii3xEGQ4KaWK52nRBEkZml8fZNTClFdOOS7TZ0uwYkjiQZ+ORggEvupA7v5ti8dEO2gPZsnheMTDG/g8+EFws50ztjyGEEHPSCBp5zw2edtMyc2ou7ojmk0+QmV6AALpEpJVcC2fiwv0y1qvEXOORfM5TgD151nNeEMboXT8Ekx1Ck3X0XKeR0zrZEyBVKExlipZDNDDS1xzaHpuqRs1aVZPHq7YI8TplCw2YaW9MYkrtUTogJgX7vB2roTR5C5KgCzhvVTxnBy6a8F8R6AE/PwRZc7iXtsTO5UiZDfGxaiIxHkxh5rM/nwdTuo1zDntwFtxT1irtNU3JFP70GIOs0bw+ESgESLCAtZDGwC8MfEOoaodFNeZAspMKa9dgM9Y6thXgWkKwrKLTuMP1C0w7UFH2FlTBDAHkW9ir0hBE5ikb20VHwnyuZsHA+XSA5PMeRyO+V2QKhOAlNoR6AFLju6S3CinVPGOPUmk76piFbghuCRVm25lQt5Jp0V54UKVmRckUjjFZ+of4N1PKAojpwaA9dDDyZIwVg3SJQuBggJD61cRD0w8ujyxev3RAkBrZZIObSEUmd8IV95Ha8WgJ0Y9bEETgz6mLJeeqddLS96vZo7EObT1iNAEiOi+JJ3irDW/I60U7Ufe7d3NqVDJX/wLQhUcire9mFOEFPRLyzt8ntgBThDhjBtjLzsKVFv9kD6SP68/MgMRJVljwCHPqS6Lrom52r4RPEPNwsSuhMSEv7JWJAs70nKjdSDdk8rv3QmrbvMcQOG8wmFSHPAWMylLo792MDUg6ARmj14CPFFwATMfZmtoM07lO9b3kou+A85BIkZ7DnBWYOxOmHF9UnpKk9MHsmkwpgytmCZJAKpsP8aSzAYDTUMNA0EWqUJXFS9wbF7g/LrAeamz7WgMu1vHRcTG5IsEg5Fzj3EbVQhmCVKK4ZO28d91ucK3aoA8VXmgOcL9f4Gys8YIQXG/grQUJg6JhkV7nUSUtAoxTQKDs0NxNcApwdhiCGAyMcaLuZ/oZJskU+37J67xKSbuTXh0klRqmuQKR2cvaJWDSDOw+JgFQ2/8YMUU20XvOzdHGYGNFlcOSVM3fyoiOqqkEG4hzpqAmtSDWsr9DbE5atHcZ1Wm6B4Kef2bAQtMDoZ5y+0wQYyCVTulUWdBQQZYt+hsN+msGwzXgqcM13ri6h6XpcS8KGJ/dHOPWyQEQgO1ThP4aYTgWuBsOVAdcu3GGp1cb7YoZK4Nu3znAyfMNTGdw+IklVs81583CXgYOjXqqHJgOh2YLjuXrFVLzrl2nwYSpQ2GVLaITG5OafQWQ9iIQdbdIWoK8EINmVVXa+Gsco8BcdLh9i5wicEtd9N0ywB97wAaYxqOqfGz4p8e3MAFNdC18y9Ed/I7VLRyYDisecqfPVO6YtHUuaN+RzllsuwrmxMBsCXZDqM70uh0qFa77VYB9You2jYZlsWld2mA/Ci6fMkiRW1w0kqnKZJASUEf638QWj+nLkJyj2SZKcCoRsayq0hR8+EoXFPG6yAgx4OP87klvXjOV2wA6aGLiP+fsQKT1MD0NcwOcOR1G2QNhLzWRFv/EAgjiQpL+hux8F0msAOXgQOIccpWeON7NFuuddS6WocTdpW6Mp0ZDJlJeNAsI5kYxqWNjKsecY/axc1ABIKcsnDNwjiGB4R1HNgBA/E4uMgMx2KJIy+2mYeLKklwH4+Kems7kBkYXsAIiAkoeBC+12ItkLcLjQNJlVLNSKQDRbtZrzXH82oQ6szPO63nX8dKgQMms1IDLoKcqsgnRdS4t1qBYyaO7wUOzndqzUsj2xil9dL/VznXOMYLV3dQUTM+C79iZMgUFF004ARc08cEUSAzzpV0mavdBjYwS3XuVyORgWvgxsYxpQ5g2FHnaiIEBYUpbzj9HStmloGzeIW8uuAyInfTAmVGtIrOz8Q3OfAPpGbbT8jZys5I2w8qtpACYCKgsQKSW6Tbe01JBRCBNBbdgjAuCXwgO6h7HlV4LqcTtbGgw9BYQwrgSUAu4J0c89fQ9HDY9fue1T+Nti+fRhwp33Ap9sPjlxRvwEfsEhnWFbl3Bdnw+lfIykFIG50TqNFkNz22HAWRtyly0Oe8AOIpRliFmhVMb4fklml8TTfSCaLog5KZWkZmz+jpfAeOhwDcCOXQ4uL5BbR2O2h6HtWpp0rWQNlyWPJ5qTvCEPcOS+5m7qB6rNiWaDspFljV4o1VXKZ3aS2bMxQqkCjhcdThqezTGoTaqB0up4EdppPfIVQa79GfY+WNpJ+AlPHTFS6YjqawEABr2CEZbEre2Umo7aEDgOVLhFG8EYUg2KaJseZsZvNkNnDGn9NNjMwGGLvCpogC7TEEMHEj2gg8g0nUSa0Dj7/IEQ9kbPYlNJnXyJU72JSEXNFjCbNc+R2oH7RAXTdLeB0Kxm6HxeYEX0l0Rk8Bav8M+zG2S94OB5DgZvNHd1MDaqMoTKLo9Jl0GZNJpJFFm0oL4Jvai6C14qEGjAY8OUYEFuAuETBKmACGlGZh3F/xzJkazlMQVIgmkkj23IXXvux+d0LpQZWOhTTQp+cT2Jj61PsZ6rHG2aeC6CqldNZNgayusbY2BTXY3nNONmc5nxN1OBRNpQ9XrqMU4U8DK9BitVpAs6lHjIsdKB0PTdNmbgM9ftymYNxTiAqj93YdoPjX1Zyd0UVCXBJQ73RFnk/dFuNJUQcS8kiJVC1DquJnu08TkJZZgxhZQqvnm6fcJKRjwM01AWnDGaLVkosAaAHrRV52EFvf9Aqdjm6uQggHcyoCkBoUKFDRwC5azyC9UhGAp5rJp8mMRgWsJm6cJ44FA3tThi4+ewxc0d7DkPuerXdx0URXgb6r16o0nTvFF117EUdXhd7S38LS9j1EMltyjkwonhy16b3F3ucD6/jFMfzX8zbwqLSGdS1CYNGnpZ0wBZRJyPgw+sgIAgBgYbGJgdBoWOPVaYr1xFUZn4EeTNy4QvRe80TSAO/RAE7A46vD00Ska43Ct3uKoUpZjjKLqBMseN6s1js0aLY9YkmqAkmiSOWAUi56qeHiaqg2eYGdrUULWohlBbT0WdkRrRrQzTdij4vIagvQ9l4LErn171I1OBHqyUwrez05KyqUZcKYyO1RY2T6mEDSvMgSj+c9Ic3fRfjgEQjBK44gnyKgr7DwHvc/HZxYAyAt9YgUo2X3ONAK7QsHpNfPgIF9TunoCItmcT6JKKaSKB57ZKtP0+6uC9CmXiJwzotRxMQUGkRVIO31mgqfdRkjWeDBxZgwshzyyfkbznnMWzL8gZXREWQs/sNJtnQH3ygjYDeVqjdwHIiLfdJENHVYE9gQKBkALHgNsAKiutPa56ycfgrTgm8jpBo4B2iUW+n2B4hUhdTtM/SISvf/p4Th6NTyVF/Fk6fyJ0+u4c7pSt8yTGtQzYAW+DvBGsGaBYUFtNRdZsTYUGqLnQxYwCmPtmyxQvBGPSfsgBPXCtzYH5fcXCzS2hghh2xsEwwgbyZbgeeGbIfVVqKJSaIT2bt/ExT/3awfjNIqmfMyTe6GZH32khfdSBkmQB2BHaHgliMG/iABx4VdRGPQetZKFWTl9wND7K+wZE3EKEiaNRwrG0ufTDoi71RN1HJu1WHQiuOWO8OnuCLe7Fcjpwh4qQn9souud+pkIA25Bkb4GQgMEI/ALgT9ygBVw7WGs0tdPHZ3huO7w9qPn8YeOfhk3zAYfH2/gt8ebuDuu1OzGMarVgDfcvI/jusM7jj6Nrzj4KJbU45rZ4JBGjGCchhoejGequ/ji1afxwnCE/7f8Lpw2h1cyLHO773lQEKAMxAgLkIOfla2n6o05s3RRgJkC5zFOOj30GrsXFrjljnAaWtweV1r33y/RbyuErVXLeqeXv2+gNvRLQXNzi8Nlj9cfnuD/fvxs9OLRVEdq5JcC4KTredrex+vsiYoYETuE8lRy2kmNKtTa7C9Enc/I4JFyxRuAyIRrZ1KqAw7qAUd1h5UZsLLKUPTBZubwsnhFPgQ7JxvnW3s+DPNUQaKHHBk4CqhNdDc02ncg5U5FADgzVQ8I4gIQ8/ZxQckCsf2Fa54WSIFDmH5HYfpOe/+ev3b3ZCCzAskAYtphTAyBftF5GvIqIHs/z9jE+SQukQdNKYP0gimlsHtQSSio6aL5DSZZK6DvO+WUfX4GJv416jFS+WayyM2q3bTzlOm7EEGsNofyFcEks43KgLyJ7AZpMDaXFGTBoAYF0+MvsZg8hiqDySnN5zzmxtWan0wtiWM5ZwDlDmpDbxFGNR/hQRkwiZqI4LVk07B2MpwzA7t+E5obne/GU3tiD1LXN3aoxKoJjPGogrbJhplc+fIu5IJrdnIl3D23uz3lp8l6nifVCXN6n0o8ku364/Yg2MHsM2Wh8YzFO3efZsbgwUH9XFiYOoZO512fk5r0AOrZESDaejdouW76W2LiImQ0GHBLDdDcIorVGAhtHK+lx/LaFtYEHLQ9DqoBjXV44/IertkNvrB9EU+ZM6xYjbASPZ7+lrUBN1tVwL++vo+nzX205LBkhwqCGiEyT4yb5gyAXtvHiw5ny+WVDMdFhlS6tqQ5yM9Eg4p0nV0Gaps+vU7vEZuN8/p4/n3g7ElD881QYoWsoK4dVvWAo6rDDbvGknscmg4r7nNAMIiO9CAWBiFW4Wl1yJT6iJtqctrJcXb9y/6aNUdkq5h1Y16zy/b0gAZA7hHzOJcOCC6iINKE00UzjU4qFQJJyDmdUSaxRqJzavIYoLuVUSwMCw5iCVXKt45GG/KkFrlM2oPAVR49aUviwJwXHiKo0CfewXq4MjEA+s9ZNQDtVgjMFvudwMCn4OC8ViFPGKwLWGaq4ySabmJhjfRDVKXO85ZXgszdS9zNCMgo/Q8gd1bcrcCTcz/PhSiGQxYYzmmv9HNyrJPZ48le1weGMwGB5dw6u6/JAACYWVAQJ0FiPWAKAj9odYIxar8KhlrPjg7kI4tx0YI+DwLmv59XJqR/y171whXgyCp12LLWVI9i8LHhJp5dX8NmrHDnZAU3GhArIyNCcOsKtDEwDjCd9s/wLSAVKWWNabySjkCDicl/IDWjSgK3ht3O7kk1PXExCmn3IlkwbFqHwAa+YW10xECwkk1t8uWGiSHkmAs3ceHwJPAxGEgucGHnGMJEBafAYJYT3sdVWxcnRXZOieTrb8YM5CfH7wG48MYN0DSc49w0bRNqZQgYqEJSy8c0CSSaMQkGAU5DCwA48y02rtbNTxMwHOvzt0/q8bhlAI4cyAa0iwHLeoRhQWsdDAdcbzZ44/IeGna4bjc4thtU5HHNrNHSiGtmg5Z071zH1NGx2eINx/fBHPDEcoP/4+AFPFGd4U31bRxzjzpS82NePHWhW9GAlb2DFfd46/GL2AwPM9G+PCZjod2UAWDhkYyJdsdg7u8wx4OulxQIpLTWENmcPmhA1vkKY+C8luTgEMhlqGJjxUClAcFBDASSuH5fn5DwOnOKazzE406baAfPPUYxMX3gsvvrvKIlrR2aHorVDHWArTyO6g7Xqm30BOry3wuPuNC8Ig1BEC2LGGmKsDpYmDC1lpyX4KRoiRFQEzCSxzKeCACoqFIL40itante7TegTXq0fjpR3yMZLWsKpNMRpcVcXdV2YpjMCsyDgt3HdoWF2Ck3zP+evacQ6SIs+e30M9ukG0id45AH8LEGBAQ1WoqqdGPkXA+KB9L9iILlGAwkcQoweRMkaHObaGOLaWfqA2MgAx8EozMYI72aGYt0nvfyYclAChRTpwRQ3J1SILhIlYlhmG7mcDlo6oD8Xh3LfnCQqYxZgOB3fy8XVS68Qky2xA6HZot1aHBvWOLT947QbyvQCw3sVhXLodKFqOpUPEQes2ZDsazJ7HpMJPGaC5xNoAC9Rx303tz6WgMCTCwBI8BHNXbuZxEX9aZyqGsHRwLXWu12mOj0WUoKiDt7xMWOBAa6jlYUwCB0MqNzMZWDJVfGTKfGC2RynrvKeoKLkdN7kQ2ZJtx51Dy7sQUgxKBsvnkQRHZSNU+9s7AUdGGJKYNkS10jltCR9qDQczJ1Fb3vFliPmtOnxmM8YkgtMEcDqsrj6aMzfPH153BgejxZn+K6XQMA6phquWHO8AZ7Dy15HLPHIZtzVV6qb9DqjRX3CJbw1gPVCry+vY/fuXwW18waT5sTXIsbgo1oJcwoHLUgjCPqccOMOAwDfufBpx+Zmn4QUopr3lJdF3AzC6j2GKkLNCgcz/NFNtn///b+bMmyJNkSw5basPcZfIiIHKtuV3Xd27hoNJqYHkARPPG7+D/8C75QhEIRPEAoAJvdYA+4XXVryCkGD3c/5+zBzJQPqmrb9nGPyIhMz2pAJFTE0z08j59hD2aqS5eu5ZpkYNYWlkm1n3LEmIPoD7R/apuy09ZvYOzijKtuwGUYatIlPj55aRM1r+/BuHYznjmHDMZdyZhhYl0TZnjcaets4+bHeQDa1ipRWheuz+j7GddxwPN4RE8JOy8tg9Zg60Pjo8/iebUo8wWki0sQZIBShciAx8keriEkAbJQRZdQigi4oEjvNRSBqbyyoT2LKREz4MiZCRvqXMh5PAbz/0jU5ED/3siEKySh+f3q5Wj5Qm0VNF/UbIBPFa3qomOpOH2p6pItGvCgYqdFV8K+gn1plmpfwHL+Ezkk51dQddaphewKxuDhggiBcHAoQZIDq8hWUK1vqjQL690Cq5aLEalENpUAan4+j/bDfgxH4InaB99P0lcNrmDn9jiWDq9PS28yqmdDKdLyIgjBsvJW2mtXkR/R6+AHKIERmEwLQoSnBH4cdQRuKPHRqQBTqPTOqw6IXEMLaW6N4LTXg0UlgxHgmCvzO1ICdF2osjDkAC4qZ051E2jjXMToqYObBR7NZ3t0sWjaa3Je7CaX39akgJvKk2lVnS2ujoKIFFqQkQJrpwhUPWcpbuABhIK+n7HrZ7zYHPF1f4udm/A8HHDlTqvxu70bsaOEjgo6IkR4FBKOTQavGPZmvZwd4Xk8AgCex2OtcG1aDJDzabwxsBSC4qEhyd/OTbjUvvXPjbq/EGrf3UKuqaeZZli9Fky2eL2GkDNiNS+JYkMiXUnfs5yzzO7R67m9P94VlqxXyWbofW7XqUNtnMiayXCe1dW2VMTu58SH2x9b9aELjhGhxhIQSViajgqyIzheq0jlhl9g6EAGSRbEQiI0otNICQ6M2flKXBxzQIquLkJRmfDZF5CXiQY+G70zwmCt8M/aAgDqQteeo3adI9v0DR2w9kF+bNpgqTBsgytByHHsZUyl9PyLJARhIxUCuSI3qmPEkNHHhyx86/23SEHwchFuQsI+TuhcwsaLLLVZW1srx8IkWu1nQCCwIUekIlLUXrkfJ2IUL+SYmR3ctIwbApLx2ly4BSWSfb5ouyXIuSiB4IIIp7jgl/FCO/8r2L98/Ob+hFyC//sf/7k+Jakok8f0aoN44xFnmSd2E1A6gtvI9eMn+R2A5jqS0SLqBR7sQhJvApZWwSlFvB02SNlhnAPmKcD5gsNFh+t+wLQJ+CweMAdf+TpmXWz6CJfqvJiLwzFIZTsFrpKs1cEzFNEK8fNq4XNAtfC1lsTOzYicVT62YOBYN79IQMfSD5ZqKtUq7xz+Ne1//8hC+1Oj9IqYmduhfojaL24LAkOyTLPfSMzt4EoBShJVztGFajudydXxbEFmCCBX0YMMV21wX04XuB16jHMQgmJXEHYJ//TFG/xqe4vfbl/jv9v9Xvr/TUVsyMqeJknoIBv/iIQMxlG1IbJmPZkJG0r4TXyFmT0+C/cYSoeNm/CZO4hrLSU4PQ+RgIJSC8BJRbA61Zr5dXyDu+3mo+Hpx6JyAYqQnc9t1AstSNd7E0Vqptn0OEM/QyFDrJZJmFYkD9ACKRZkACU55EkT8KhEvrBs3ADkmm7e+zJSuPbzmBmYiZHZEjQ7xlLsdubq6NJiJOgYJTJysgJb95KtoAPbbq5W9gAW/s7ZlMOHxAcnBOtllkBKXkrFY3bCH9jwjKHwgwv2sRCOgSUH8uwOUqEa56CFPLqcMancbvRF7YtZio1zcOBBJd/824L4rCTFahF48Bx1woCXSYP2T2uFsbQF2Gu7wCvMo9rStW3wRBGicQUEFSBAoF+/XkAZS2JnSR2wJFjRZ2z8jOAKtn7WyQ9WTfF28efVuWntlE8uI7EQ3ubsMWUR95hVgyH3y1iUq7Btc9/b8aZFSbIiCFkSruIJzhHYOZBzgH1X1y99U/KdyLhIHx5PNH5498OF/GAiTJnQvyXEO9lQ/AC4zMhlGUl1E+BmrJAUIzFRkPPraUEIEkvfepiDJBxjRD6Jo919WFCf+9zDk+gS7NwkxENbDKFoAhh9SDLZUxZyIbBU0uTFE6Nz+YF+vHAITDcA8MiIKJjhhFRZFvZ31g0+NiemIopnN8cvQTKsHALPFWFDXngaIEMCsCABECRgIRyf+R8UWiS6G/S0JRDK55NWAQAdORW+gYjhyHkEAQgFsUv4m91b/NPtK/xt/wN+G94gUtHJhLiqos8h6gzGzIyBCcPZxhqpYEcJmal6UABLezASaq/ebIMnZr1urPUjVsKX7oTP493POh/Le7bCURHns3Ov+7IgBY94Gjz6nOzqNXkejxmDFSVZ+yA7Uw4FRe2MTTnXeD/1b7T95alghvBy5HxkRQ/a9gc/WI7s/bmm1VCf38T2bALGMUoHUFfQhSw+QJosZ70eVoTRj4gPTgjOleqKIww5YPQBLheMJeJAPeDE09lu9FaT3DL8rDeHgxy8VoPMxDoKuPZi7CK17+2mxjYBULBIDmdakdcebN4GRZMy1dvNvzQteV6eQ5yueKVQuDoeXtEBx2BP0gv3WCchtWXAT5sQBFMMZB0dNK+BszEus65lXfiUBOgNiiZrFZTKVnXE2PmpXnCuyUItIci6uczF1/9n8/IAEGMWDwMmMY1hAJmq4WNlsmOpukgX2brY1pYBADVYqXEuPPQYObAyPv96DPb+L1IFiieGfCY3L8hI6ZT1rBKxAKpZu/UqbSEgL8hPC9dnFuGiIQUMpw5p9uBjgD84sAeO6mqYi8N3m0udI58EXVGI8dIPFY2b2ePtvBEyU/M5mHQhdIImGUTZLsQzxMBsxrrC73SyoUPGTKIYFtXrpI3Vov6OBOAXky4+Q/oe7C/2O9tTmlbBuuhY2mfAuk1glDvjUAEqI45F9t0EwMx5FVm4UofU4Zg73OQdXpVdlR22anoD0ZUQgpxTPofcExlABAN1w6D62rLRt5vaWmLaokV+IiVBbGjtVyHV+C/P/fiYWLQuGFYRGGLQho27eyrV6M1TkIIT0HZSO4YqVbvXdbJqCajjp4lQGUIBiIKiR1HCfdLfuareuX7fC3JBAJwXhACM2oYuHSN0Cft+wmU3onepijotjo2uKph+aHxwQmAGOrkQssreHn0nBCZ2eJ32cpF6XwkVkRI2mCtBxG6Qjd5RmXQsiUxSUrxYJxfgWNTUTNSlzcZYCWxVDjc7mByuSd6K0M2iT7CMT3Htx8hzyYH2UOMjO3aaHNhz+YnhRwYxw01if1mDTGeAqgc1WP3VLXlRkmHN9vwjm9ZPjH0vGLN3YnNLxFWkAkCtBk0b2wiAi2GVLPJmVtW7jK2b6uTHdThWt7FFplZGaQDUCZKZPXZ+qkSxVByGEKvXwuQK5kTIQUWKkh0brouuTYWACCgMR9IuKEFh2UDaNmgucl7UCx8lELah/eu/Rnz1P2krJzH8KJnm8CJgunTIPTA+kwmCutgA4HmZ+y8dowQgbwpclxFiVqMqef+nWeb5b48bzG96uNGhf+vQvZEk4jT2OF4HDPsOnc+42ezwrDti7IKaLN3jwg8oTDh6ga3v5g2+URXS2gbzAHdFRq16aRdsm5ZBhsPAHh5ptaFsqCBq1XNUNCBSRscPk8v1HPnDBazgcYvnnxrV9IWAB1bkei0yGbOj/UNDC2lVPLC2Fe2utiTbQUZ3Z4QGRnZ1fcxMsjYyC9o6e5TJi8R3JqTZ43be4IfpEo5YN5/F1KijLNMDmDHD4wiGZ0amjA0VeAI2BOxJ+QMtLA7Z7Hty8CAMnPG2+q0syG0EIRIBEIKsqQRmTQDN0v6c7PdLRJtQtejIYyiSJUzSjvboIHuOjYMCcg122kbrKWEfRgxR9DIEgTakTJ6TPQNRkDprq5r2wJ7E5G+jLbmhmbI5cFff06CIzLlFs0xz+DqRYwWeDxnTrogSZdT2YVfw4uqIr/e3eN6d8DwecO1PgjbpWnwqnYhcfUR8dMtAxv2kuhS7T1fNVexr4kWNywQlNjTXTHh5UoF9DB7KKDiHIeWkPb55smbltpGsxgWbL2hLoXYJdKGrv7eFjxY0oeUPVKRALY4pc3U2tKNT4JTXpkZGjPpVg0wA5XH46qeGjQeapHQgaa10PmvFooIpTqpFEFC01whgRWLxTQImojVpMRmhRVu+cFPpIEivlCUjdoXVvVJe396X9w6zZ2H1A8tF5XTxZfGlr8kU0XocsSWA1UP/CDLwYxv+Xykp6G4kIXNzBo0Z8IS085h3kA0nyKZfg+VtsSWwJgikHDMRjFo9XOSlswPN4hbpRiCcZAHzI5BHhxLFNfGUIjZexqvsntrQhEICLzouqnzYGBhpdQRnKIVcX4+RE6dHNmwP2W9bBOCBZwEVhUGKehYsG+d/8ji/T/nsu/38yP38QJ+BXYOYLiiBRWJBBgwdIDUAO6WIQ+pw43bqTbGpCI1JU8MtCIrBzYYSbBwhguTaae4Xk5j3IETyyGA45ApnG4GtPl6fO9NaofZ/D/G+a8aIfq1y5nmYsVabyD2mtwOgir21e1W7twnCbf1+4VwMRRIC73jh2GlroLUFX30mvd+dk7ZWmwy4LmMbZzVKmrBRDYOZPEbFQDJTdVz90PjghCApaYt9AZIckDGEWq3cpU090BuaMVNW0aGkFx3jQY+LGA4Ez4xZoSjPXNnHwMKilr60U/UmNc7JSyZNs4i4UIao4p2NthVgxfJvK3RWRI0LVZc3+R/a604MNzP8UOTnqYjZSBMuOJTowIEqI5Xd0rowZILlrnpahCAuCIH1di17NSg/sYMrjOSc9EYd6jhnoKLGRlkgMxKbXksGNjRh76aVuUgrBjIpQjCpZvhMASMHHEKP4DIOsa+vNfde9Amyq2I7FZIrBOgxY+ursyUGDZa7SrIeWYk/pD2w8j/IMqnwxNLFaW+3l1oYO+D0wmO6Fn2KvFEosOGUsIeYYOnjCVBuJFVfg14lvgcE5EIo2Qkypq0JPwmi4geC74EUHYY5oPMR0fXYh0UYZe+magJT2OGQehzGDuMYgcmJYp4DqMsIvUCU+zBhq2ZJMzxQgJkmRCp1nAuA6BE0x6NDXl03hhoa6ficO2DjifbzUwa1ct+GztCCVAENh8Aeoxu/AViP7YnmOLok2EvboK2grTI1DsEx93gzbME3HfzgEA8EPwBp6/EPw9f4j7vP4TQha18rdglfXt3jqh/wrDvit9s32LkJ/6R7hb+Jb7ChGV/ghEtanwtAEAMPIBJXGN0TUJjr/89AVfYbzj6vtRJmXkYsf24Y3N6Kecl7WVodAKqoXeuQuf5s6za09fWF2Lq0D2zDtuvPHA4TO6QkPgJQN1aw7DMMj8lFfHe8rOTRqCP0OzdWfZ27vKnn9y5Lpf5VvMVX8QYdZXzhb9HRXJO3qJ/fdEuedUectlF4dVmk4ENM6LuEbTfjNxdv8NvtG1z6AV+E22oXLQh9j8LPcT/3H3X8PzghMMlbZkJxQugbUqhynfepa3zYTVwh6QiLJQeLiImcKBUuIoG5IjLm5uQWLMQcM3qY1fRlnj3K7PRkSTLgJsDlxTFv2Yw1GbD9xHoyFsYazgvEX5GGLKiAHwvCMYMKww0zaM72JuXh0cNFMReRhMBJi8DcFGHJAK+JTE8Ql50IUQQdNfPE2PoZnZON45Q7IZ8hICnpr03/zagquFInCiQRSDqvPOHSnaqRVT0/etMtalxeRKdUrOoqnDCWiPs4IrGDd0UcxJyvplU27UCaEPDkRWBKuQIoC6rzo3EuNvShSIAlIZYUuKepTOcLeZ4SCFknJaZnhOlSevJ5wyjd+nrg4MAjLdek9aazECkBYONnJCN0auvMJarypm4Wi3A/AuFEKNFhHCNOPiN6qTYt9m4Usp/CjMcUcTr2yJMHzct7cLFgu5lx1Y24Cidc6JjZofSizFYrIqrTJxMcIhVkNhSqIDeLrwf96HltpX6fcvyQ5iUZtfYBKQcIQCW0Ag1y2BQKD57PHtiEtUrFBjnXkUNLqgUp6TGWiNu0wdvjFt1rj3ACNq8Y3S0j94TT24DcB0n6Jqz8VvIW+NPXl8j7DH894/ef3+CiG/EvrvaYdwGX/iQwv5se1M4FOkHQaBU03VTYdpxZHjvxMi1m0wyAnHNLKn9ubBpzI0sO6rhhE49NGbwLUZKNvpPEoQBzg1BlHZdf2qriJirIG6EoGZiSNo9GQXAKAl7f7XGaRLcAEO+SrZ+kbVo8btIOYxbnxKPec/cX0pq78AOeuSOcm3QEVNZV4yBc+hNedEdMJaD3qQq+7bsJ192AXZjwn++/xz/pxJfiM3+vf7vVZEla+ce5Oz8c740PbxmY2QYgxh4scKVlxKl4lAo3OuRHKjePBZYGpPp3xLWPdj5iYo+xrK1gMc8Rq1/UXj+M+X/WLqDS3NC2MT/YYBZ42toHFhX2lxcGivAHKkJgWYbdSc6cDVl1+PHoAlKf/Amitgw06TJnrUhZSJtU4PRYA+sWzGPQn+lDWOtg7UvfbLBkv5OyyZP2Q5UY2vI/jKhmgkkyAuTXkL/B09Q8ucVjUG39O9JrsvmfZoBEDupV+95Y/e0TRVbmXO5UsdIDaSMjQ8VLu0DIegwY6z0zOMh7ryNw7fnSxM36184VFaPCqr1yjnK1YZCo3I9yrm2yZypBuDlJJFvtaZwXRnP0WRQNtRIBlmqsNEnBedR1AW5dvZ1Vck+NBLwrWk34RZGOV4nAowhAs0Z8SI5qLPYZi+hOgVslONZqTcnBTdLq8QMQDwUuE9K9adkz/NiubYw0EnJPSJNM8nzrL9H3W7GSdwnP4wFXboDHLSIVbMgG8CQyC1IrY4UNett8dlMpfNdYYWvb+3NjaUMuiWBmv+KbnEeuUNoaGXhwHeq+5Or1uqjoVjVdLGJfUrQI36kasUGI7EwO0yD3ineM78JlHdfuXcLMDjfTDmMSKeQhBSFoB+FmAdApEVq1jSxRFBMyRXmZMHYBqThcdiOuOlEk3LkJG5rRnSHvFnPxorj4EfHRCAGx2eiamI2YqgxZPnB0ucJHs1uevk0GYsXkSrVZ9aybEC8QW9bxtVRkoTKUwBjr4kNOS8/eXPTmFiEQJ8LSYT0uWLkEmpwEqaoKZEqAg0C3tZVaGC4V0JxBpwk0zVix22MAdRE8Z/jgAASwJ7jEsrgmAuXGg0Erv6cIEwWJLlcxoZ2bKsHPo2B2HkjA4AJS8ZqILc9hm0Rwy4Jv7QLbNOx31jKIOlIjozZBkgLI4tImAsEJc7cEQhck23XOIRkYY1UwOeTHNjRW/oadP5br0BIwEyhaUwvescG3iIFyDrg1QHpC18PhM7ln0hZIe7nG0kVB2WbAi8qYqUmSVqZzCCjBS+N9cFLJeoCU2bwJCddxQFLNB0c9Tl3CsBFGbN6I492KJ9NcZ53LuAxS5V/6E/ZuxJAj7vMGL+cL3AxblPtYfRTAUkFv+hkvtke86A+q2z6JUZJCtuURxvQMh0E9FYbG9c8it+NYP9LrFJnjp4t4p4inQ536KR1EDtaS0jY3aLgUNjZr/iltXs9N8WKb5LF0MM+WWvUqIS4z4Zg7HFKP6RRx/RaId4z9dzP67w4ofUC83yD3BD8ywinJa6YClwpKdJj/3CH3hHnnMD7bI3eEf/3ra/x/vv4bbC9G/F9++7/hv97/EV/Ht/jn8fsqX6x1FO4K46hVs60JNWEAVZVC+7dFATAzcJs3eDk/jbmRGT4Z618E7JbXtpG68+vNN/uGfA56nE8AmfVvzYdepwu8nC9xyhFvpw0Oc4fD2CEfImh0CAeHcE+r/aN4QrrbgAPwarvDy/0V4E0ynlFmBz56uMnV9Z498PLrC7z6fI/PNgdc+EFaDJgrer6hVKv93/avcOkHDCXidrNBYYerIATCDSV8FW/wzB1rwTY3FgLH0uF23uD29EuRCq3nxprJK4Qxa19zKgGhFMxFDCI8rcdQvCYDNtMqzyU3XmGuzNU2TPFrZleZ8azIgBlPGEKwQgXUSrduIrR4GlS4wJKBhtjMTgtK+2oahcRyE1Iq4rQ3KaCmkrekRDnKHi4V8FzgZlc1++vFVOjJq1Ezs2hd9XZ+qvr5s5M+b3C5qdrfvQBbf65TIqFloGbCUVnGIBklZQ8YB4SKJiH2eCEgGVchmnqivpYhTHZMshNOiPxP1O8PRrzOw+B+fR4iJSi2cZYMLJMJT58MAIAKFSLtGPOVoAC8S4ibBHKMGHN1mjQjKWahsXBy4KQtE8ea9zCiz9iHEal47MKEXBxiyDhFRklcdc7lmKyPlSPUv9/6ubbzANRN6TQH0OjgR1p2BRJdi8tuwFUYJdk0zXaNVuRlOcQqDwuHCQvZuI1zU5q/lg2yyb2zg/J+IGuAV7TQphDaS8jJuSAor8IqI3scW41gNvC2OSmZ0C2EP2AhHY5FYGUePcKB0R0KutcD3Mu3cF0EzQXce7ghgY4jUAponIGU4J1D7CLYO/CmR77qUXqPm/seh8MGw7MO/+vVV9j7ERkOfxdfYnOGwgy6vjqwjIliGT+0tu6gapMG1ZvNfQEwsFw7TxEr6eIVSuDqmiPvz61QiQfXnk5BtQRCR4yoaKHxHjIcjkWv/Rwx5ChVffKgycGplHgYUM3ZqIhyYTiSmk8R0j4uk2wOCAnobqmKjIGE13P0W/ylu8IpRby52uMYe+EIKFIQqWAP8agwV8TChCGK5sS1P+KZlyTg0p3qPTg3Ew0TB/FmSAHT9Au5HXI7mkNAcRAr4uLAXFYSthYt7GPECVf/n+znMy+wdXuyzSchqflKNsGPQuCyZF2tOEi7cbRyw4zl/7WPfbC+NLBrHU08ayHIweC6+Zj9LhGBFKamMYOCAyVexiCLfmB1VHtKkrvNnwYnNrJO2ce9KkFGJ5alsSxV+3m0s9PLuXr/KFEdl1FYD3hY6WVNHusiqAp7zKW2f0pR9cvCj7Zs6vmCLsLMWL2ttWsTVmOIFo8gA/LzWesHT9c+qKNKEeBYgMDwnYjNOLUw9q59XUKKClU6lv4lORlzChkhCNISSUYPd0oOvN90uN3sUAhII2Ee5XikjY4u9gXbLqFXy+QLP1YDlGPpcSw97nOPu9RjnKNIJ9thcQCrbvt1HCqyYMmE9cMtWplygaKXNeGcIPixUwRPySGwpMmUINspltp6cVhahcC6HWOP15zA3lkpDrmonLQq+kWXYazRlRQvu7qZ3qde+FANEboiX8zgote9yXV7QSHhCOxd5b24uQBE2LwV9MBPDn/45jMMKeCb6+tqv/vMDXhmdqNo5+Ufr6ofHD8QZgADe7zNW7yensjtUD/8Y62j1bWl15QlW+fXkkkR215izzmUWMfdTdjpmDscsriQnpJspPPsQaNIi7sR8AMv/iIZALFa20PaOZkWvw8ISt2/YYRRUUgdn54vHA4XWzATXs4XuO03cCi4VC4BoIkLMq6c3KMzB2xY1vidGig5mIGgJEaidhlwk3d4k/Z4m7Y4pSiaFh9x/D88fUgGscmFyY6RvUMOhDnbCOJiVGJhvTJDBhyAjuxk8irrA5QRqqS0Ud3ahhyqGtucvWwgqj3w6Khh++8MgFBJOFatM2MZc9PPxQpBs5f9o3iBhig3WtaAfP5SgJTA0yybyjQDUwBCgAteNi5P8KNfLpqk1cPS8nqSMC1yg/QdFVw6cd+yiz5SFqb/IyhMW7dZu6B3c53PfSwsabDJkHKG8LRmNoAkfV61DjwxZufgs0dhVHU3AEjqx4CzpMDO6ZL02SJNy/fHJgXOM6/zNsEZSvCU6E3eSosobRm0y3CxYLcfcLURBcg+iNwoK6vZrKaDL8jFYfSMkgkhZuy3I7ogrmbWgzTjr+AKjmOH0xgxhh4lBoBJxpQ8A5cJn+2PeLE54Nfbt/ht/6rOsn+fL/GX+Tn+dHyGl6cLDKdOpnWSQugdA33B1/tb/O3uJT4Pd8JoprmaypwnBZNdUdpCmuBqn1ZaSg/7vO9qGdgGEZGfVLFwvtDrwCZ+6iavyUForsH2krDHeOjmLC6V1p7JyWEOHoOSyRZBL1H5M9TMiHj3eYNvTld4edzDHx383GicBC/rTinSOhIDFwAO3MUlQQheCLEAaJzhxxkXxxkXvyfMVz3evO3x9rOv8P/6zQvE/ybjn+1+wH+3+z2+6F4iEmEGq6ohMEDaAyulvMeOHzvcFeCm9PjD6XP84e75k5yXc9QJQNWpWDgo0g7I/JCTAkhRY8JNVfVWUeuRYm07DEVMjV7Ne/wwXGDMAbdDj9PYYT522Nw6hKO0cPobhsuAnwrOpmZRIiF30qaTNYoRToztt0e4u8V5kPsAPz/D29zj+CLg33/2Jf6mf4M5BDzzuoZDeB7CwbvDM3dcraWdannY9TOzx13Z4tv5GsfS43Xa44fpAnfzBm9PG+TpzKTpR+LDE4IGPmSrvHmBfI2M8UDlqrJXzR51PdvqsJ7FzdofspOf2Cnz09Vq0qC5GkoyfEA8M7QAzaZibYP3HCV7ulUlADTwIMvGUlhaBqWowJGQCSllcCqCEFj7whISXhKSpwpbdCwhiDodIBeOq/3Lc1KhkTrtu2zya8OpFro7D0su5kf/b/O4+nxq80vSPmJXJCljQU1Mu7se+3qusCA8BesKH5Dxzsde+MdgmF8wGQAkoQQABIaLBc4LKa8PSdQgg8gFFxBIXTxn59GFjJQZOTpkcogxV4nSTkdBq8iNI+z9hJ2KU82bgLIVsqZtdD4WbMKMXZhx4Ud1ZEsCLxYxPTqmDkMKKIngC6TFRvIcFETKuvY8aa5ckunskFk1JnKt+UF12c5ct6OGj6EFMpXQzu4/3flZpIuXn+UXaFAAxjvZg3VRoYVfgAXxSjqKm4nrlEErsFbYoZAKyKSIYQ7Kh+KG8KgvXABShnyV4pZeqxQwzsFIATIpxaDjKE6gwx67FwGUHeaLgG+OV9j7CYdNX82LPElRMKteySIf/PjxrudUx/hOOeI0xfc6qX5oPIYe2Z4gPy+vYZvkeVug8gg0ibAJgqKQnQNhLjIaPSs/zSbYjKOGRKAZOrUD+Jl15Fw1aMpibFdmgps1IcuyB4RDgn91B75dJJ3dZoPu7grh4JA2pkDZY3ADZvbolCmspTc2lGWU9+zzWSE2cETWcW9B+jrc5x6H1OOYOiRTvPyI+PCEwN6UVtysv2svgkXYRghqokEgs+sdlaqM1YYQU6THaGzVQfs7UwmYsszQ5xWH4Ec+5FkLobYMbBpBF7wVhwC68ZvfdWkV8mhRIVTt/DriVoq0DXKpyQFNMyh4uCnKRWSKiYZWlB/fqz4mLhT+jWqM4VBw5QcRg+KCO9qgkKAEwWUZU2vOk8kVVyIgWJ8r67nLdRPYVatbrsSkCjfCP8ohiFTQ+QRXHIonZHLwzsFlX5OROSsHw6Eurm0SZwTRVevnPBqUgADwO8YOuUUXfoHpAot8oYnGNiN2CTEmXPYTrrsTHLFqRnC1LrZqP6saYVC+RR8SXmyP6FzCi+6o1qiltoqGPuKL7R6HOIOIceekaoVOdlzsB3y+OeDz/h6fxzt85u+RQfjjdIHv5mv8aXiO7+4ucX/YgA8BbpYbvAQGbwv6zYzreMILf49n/ohn7oRIBQN7RGsVgQD2eh8vv4tAlVKdOShBtRldfYz4ZRl4U6G/C6n6qVEFodo2ALBoERhSZa2+Nmj5zoRKCLX1sJhEuD7QCJjCCpfEbWKPIUe8nC/w+rDD4X6D7kRwSbROKJV6HaMUMFzTLhN0i1KWazhyJdeyE0siygUYJzgi7L7ZIBwiShfx7379Fb67voSjgi/9HS7dgBduwt6ROFUSo2O5Z6OSHme4VdK/6JBIdfrN6Qpvb3d4inztyMJFqPwL7YubRbS0AKhJBs4cEfU92OSGXXstnwMQZPSYO0nIctTCU48dybk3H5oSZGJIUGId72FaTK/cck0IuV2SAsyKIOs5BIBwyAjHgLAlvD5u8afxOQoIfxPfAE6O/14Yq4gyC70qmOuUniY6R+5xl7d4OV/gPvd4OV7g1bDHKUXMU0Blbn9gfHBCYEIe7CBJAaOSoKoQBy0WjKb6ZIS0SMDmjMhWwHW2dVZ25MCipDYWaRNY9pZ08yhVAP8dH9Tu84oG6P1uZEOPOp642hBqz1C9CKL0i0rQzxrFYc9nqQqEtAZwLuCcUSXmcgYPA8g70NTBzSJq5JL0mQrJNMN7OH0fHRVu0s3boVQbU1dKFcoQje5WldBGFItqD5SqTtg1OhImxxlRsHelIj2AXAqmbdaxvLZrkguxUhYznKB3UHIeXoWKChOgehZsKEFL6FolcagjnQ828vZcOlfJnh8UlkScjy7+3LicAQK6PuFiO6ILCc/6Ez7rj4q+yGvdpw5TDoAT0p98HJGfDq7gIoz4vL9H7xKuwwm9m4U46sZ6f40XAYfU4zIOeL3Z13HdXBxebI/47fY1Po/3+HV8gxf+HgNHvE4X+P3pM/zp8Axv3+yB+4BwkNE3JoA7RtjPuNyN+DLe4bNwj8/cAV94mWk/csaxBK1WlikCIwq2m/ispEIxMzNi2kO4t402MZARsafLorlrrjE75ZaMEsRMyrFMrFbEse1jNc/V/FyyIgTaBouUcWnIihPRmcION3mHY+nxcrrA/f0G/LZDOAJ+ZOEBJC0yACBlkOmW6LVPcwLGCfDSnmTvAO+B4MAgQSmPJ2AYEOeEGAOofInx+Qa3zzv8j+F3+N3mFT4Pt/hvN3/CpW6xEQWFCBvK6AkoxBibcaR2VPJQetzkHX44XIBfd+9ekz8ibvOaFW/kwFYe/VzZz52hTtZiGM1Rkhcjr7nIZEEqDqfSCUKgyEC9DEwZMMgIMAdC7gDODMBp21wE6wxR4sa3xulYOs8zeBgBlj2CUkI4JHR3MhVye7fDP15Kq+Wm38FBpKGBRiCKytlnhX5OxqTn4K5s8N14hbvU49Wwx5vjVtCOIYDSLzR2uMr+KpT7sFpvLSF9szm8L8yKNZ9lc1WtUL+vXsuypgrdQXt673//lgxQ0b8t9ODmFi6BjCuyI/lZehvqYOjEFMkrUtDsPZX8Yzdu+9F/oWLUpGCtojdVSKvybZP+sVjrEywXYp0soKXt45v2ybug3IoW6GRDYRl3dLygE9Cq5LzV1AY9uPbwix3LGo8ZJH1kuCBZaQjiQRCdICXR5RUcey4DTJpIbfyMzmfslQgYXa7JgFmDG1KwcxOKJ+xCxBBl+mAqHqzqlWZ+skyJyLk5pA6nOYJnmXWvrS3tlftgdsupehFEyPmfmWsV86540DJgh0i6cKMRunnPCc3N/PiT8QjqMX+k+l/1GGlpX53HB14iC/K2vHdjuY85qFQx1kRZb+uLW5Av5SUtc9++thWE+c46BilrD+v6w9MMKgV+LPAnIPSEw9Dhddqjd7OOFIr/YkcFE7t6ny/jh4IH2uab4TTJE0a+oUo/N85bBQAaDQuqG35BQ9BsxldbvkH7HNaae1fYGuRdQXZOkkHzFDGnWiMskxDZ2yTArhGXGCU6uCAOrBSDFIxFjfTs4Syk/DFr24K9qH7a+8HSOgCAx+6wWdE4c8o8pk4IkeowW6fwPiI+PiF4UJhJ39cMcqzvXDckZHQoiAAiOWRwVcbKDIzWIlB0wIgSYwlVkMhex7kC50jmPAGZHw6qMxCXPj17QWgc1+tEpg4yyX00yQktXtf9BjIEUG1nSw/kDcBeZrvTNsA7gtt2oFxA3sNNsyQOuYDTLJMGVq3W17aDhQozPiWr0NioxhvwKNjTjF41zjduxsCdtAyoYCZ+Z4JQSaC0CGSYoImgPASzRXXqhz5wQUSpvW2TnwbWmu1FNzpHDNcgBHaOkzvTD2/bBe0oaWY5/tYvtSSsnf54T6W/QgIeaxs8QTIAANudkP/2/YTnmxM2Xpj6l0HOl5k+HXKHqfhqPGV6Hl9v7/AsHLHzE56HQz23xuuQNk7C3o34PN5XV8rossy3pw6peDzvj7jUFtJe0R4jU/3l/hqv73dwdx7xfjGcknKRsd+OuO4HXKtmwY4SLl2Ag0N0CRvOGDiL6RncMm2iqJ8kHqF+QbkLwLo3DCp1PvyxMPGZpxLAWe4/Xv5hZFbVfCCn7P7zm/WxhNQgY10Pg1t08S0G7jBnqXa/m69wM+/watiDJw+XdE0KhNw55H0H4ALsHDguLcpKONx0oJwlAVBiM6UMGiegMHictGjxIE0swmHG/psO8Z7w5sUe/+8Xv8Fv9hf4++5b/F14CxCwA6Nv2ruZhRR6LP1CHqWCY+nxQ7rC9/MVDocN4q17koTgUJbxRZO1rtC/Vv2Li6Eg0oUKWlG79hpaphYkIXBOvg+IcEWtvH3GJsyILC2DLmSk5DBdRpTgUDpCidb2VZEi4OFeWBh+cAgjI+4c/PEF3KaX1sE4gbzDtA3IUfYeTg73U4e7JJM++zIhe6rJmCBj65hZJjsG9niVpOX35/EZ/nK4xt3Y4Tj0GI9RSPfKa/iY8/JxQ4oWj5Dy6px5kwys1e4ITrPMUslohAnCHxg4VtanCUacq1JZf0dg5SUhIJJN28SF6vhQMwYC1g3FQcgixo5UK132vIwZqlAJBaAoe7REUQQDPEIXwH2UTDwGWQumSU7Ce3oBC1mRnwJdq7Hof6fqRtaTuJ0VSBuh0839QxAbAFVzoEPWzR41qasJARxmylLMgB9ceOe21aaYCEJV2ytwCFRQHFVfDAA1GUCdIuElGTBCT4PELH+nyUDhx4kaNhqq/VYAqHLFT+xlsO1Ep/yyH3HVDdj6Wcf2BmQWKenab1ejMNtAOpfxIh7wq+4GG5rr7PGk7GlfuTnC77jwg4yZKjt8Zo/eZZxyxGWQ/yda56mej0PqcTv0GE8R4UTwp+WiZAIQCvbdjIs4aiKRsHcFPfWI5OGZEDnDm0veWWR2mBEqeVBQIHWgA1cjNN9UzzZzfh5P7XaIpu9fE0BNBlpewWoM9hydepAUcP2bRQ1yrUg4csTAAbdpg5t5i8PcVSMjAEpyJORtgBEWObhagdaEgBnEETRn+FJkp8gJfByAkivhsHKdiECnGdtXGWFwOLz2+Ob2CgBw92wL4C08aMXxyhCvA/MpAYCJMjoWlb27vMWbeYc8eGyPP/eESIyNJ4KtUy0XwLgEgCDHP1b5A9Z3L0v9zQufKhUvZnA+o+h6kYrD2AeMu4zsIUkZ0TK19sgGa2T1cALyAJTg0T3bIBDBjTPo6MHBI/dOLdwBzoRhksp+btoiFYFVrwkLac8uuhB3ZYvXaY/X0x5vTxschw7zEMDH0OjufNzx/2kJAVBhNUHOF59208E3prugA2vqUIZU+CI6ZC6JoSGCrPUMgEVv3ywpCQ7kC9hpL983CoNeEgQTPqpv2arNAiBJUrAaOTIUXOeLOShTnCGe9VEuitIH8TIoDIpRLhYoCuG9QHneaYsBK+vlFST5RGEtA2N+OxKBEU8QeB7vnhT4kKitAkUHIi2XqQia6OOwbh+ct3wAbUto0eV0x3+Xq9jKV6L5qmhA+6WcgZoM1OfgZSLkrxw7JfldxgGfdwf0bsZVGKpg1LF0OhIlxFlmglNVx01o/NXdYj3t4WBzTwuitgiweBI/CseM4klbCgtyNHPAHUfclg0OuUPObqUxwgqPlo4RNwlX/YDrOFRkAQBGnjFzxsgJBy6yF4EgpmRCQDufJPCKADymQW9xLju7GmN9ynZBG8TNpvkj1widoQXWSjgHEBgrtcJBbd2BpYVisrQX3Qh3MSMTY5gD3OyFczR50YMgXYOUwWsSusJ6B9zM6N52cFOGP81wb4OQDccJPDdUwJxBOcMPWZwwTx6HU4e3/QY3eYcDy4YZye5ZVHtjALX1aJ/hNm/wMl3g9bwHDb4KPf3ceFcLyXgApi9QpzW08nvs2qit66YIkqJGkBUTcdt6mbZJZRFoiiGDuqIkUUnYjZNWz3l73vX37JR8Dkbaebg5ClkwZcCr+Z25mBLq2mejhfJ93ZY1kykj4A8ccCg93qYdbuYt7uce4xyQZjH7o0SPv8cPiA9PCNon1pXZaavAa7/T9JUvvCwge5pqpbrIT0oyMDHjyBF3ZavZ5gZ3ZYNj7qv+gJ14IyjGkOGKnaSCRB6JJNMqeTkIrheUwE9Uq0liAJnhmBAUPZA2gzw2b4DcziA7HajYASUR5olAycH3DMoRMRDcKcAbwWcYxXyGHKjrwF0ERy/KcXoRmJCFOR4+VWxqQiDQ/kLiJMy6ENtiDOABlGniMY/ZggK20Ysvek9R0BUNR05d29bPafPBktEv46hR8bbMBJnKJgz53QuxnDtWlADVR4JSEbJVykDK4KQ4npE7gQUlKCx9PADvNS76WDLij8Sv92/hiPGrzVv8/fY7bGjSCY2C2yLMYBkREjEUBnDRjXjRH/AsnvBlvMXX8aY+X+vsNnPAt/M17sqmckYAEamy8zx6gVd3fhQJaiq4LRvclg1eJ2Ejj0MHnszJVFQV02UB9wW/eXGLf3H1LX7VvcXX/i0u9dz9kBMKgLsSccsiSGPz0RaWCCwEwwTnSp0fr5+p8kcCSpO0yvazNtP6MXnjj4rGXMwWZT7vudrPSiKTaQNZZNiZ7wFjBfcxRFeiOIw5wIHxNu0wKnoTVQzoeTjiOpywDyM6n3FKEa9/vcXd323AxYFcgXNSAMUg0yipiANfzoT5toc7ePjRoX/dw5+A/i1j/+0WfswIr05wN3dCck4ZKBl0f0KMHv4YsXkVcPPDFt8mj3/31df4l/2fsaGEF35GhIwSzyxCRZEyLt0JgPoWwOOb+Tn+1dtf45vDFeJrh83r8iQ1zlzkeilU6jozl1CTq5l9ncZZTuCax2KthOjSSl0RQHWL7N2MmLMmzdI6mItwaiadFx6vAubkMU0B6cLDVG7JimHPK000LsB8F+EP0n4Lp4DSEcIxoHMOTMC8c8i9trh9yylZkHFAkoBo7VmSkdCZgbvS4dt8jZu8wx+GF/jD/Qu8Pu1wut0Ag4ObxAYdAPKGwd3HJdEfjxCcJQaWFNhBjS5XhnodNyTpNwsRST6sjatNChvOKrdosInNjtY3qvKOgBO5VxI2ryuu4RMQ2ETBXFOZ2zEx6DkRoFwdki5AFTTSlxCCToBcMGCZMogAMaF0hDw5IHv4KAs5pQyEIC0D7dmxtSNoyQgNiajjTU8QsWa/5iYpmbCrX8sc92M+9gB+FI4V4yISUlHbFuGP05urZDoSBjN40Sd4N1KwLt4eoAQ25QE8ngzY44ClPfBXiMsoZdN1OOHLcFuRHJsdrjarmjSxJsBbTa5tUqSd1a/wMwQhOOZeWkG+kZU2GWkn42GbZlTUWnN3ZYMxBUmky3KNsge4L3CKDnwe73Htj9i5GRsiDMwYWHTkb7mvCYl3J+19ckUIDOK06CjrpEB4dHPPH3UlPU3Y4i7/wAqVenB/Wj6gPzPWRLE2Wm2WmUU63CYMHJU6OgoAaSejb8/7I97utgCAbZixCeJYehEmBMo45Q6H1GHIAb/fvcDt3Q7zIMa54SSLWTx4hEDwxwjca9WRtIWQEmgQB4AwAP7kkDYBd0mSRLgBmWchfWoyUBSpEQldQX8yi9zv22mLw9iptO/TJATvipZgaN8L1kjBuW5CmwyY1XbVWkFBcU5bawkze1nnNUGYglgMeycI5mxTUJqoEQFBEzV71VIcjuqzk7JD3gBpcCrytSTdFSGwvzPSZN33HtaLWVG4CTJZcCiqNzB3OE1iVe4mVx1KQc1+9hHx01oGS1tK+vkQpz1z2DN2+0PYx+AR+YDy4d2qXTCrkca7qlVAqwpAT4zyCRxrUiA9uCLrDpy3I9O8/dKgf07zBZs3bhcHXSALDEaV/5969ShwBD92oFnYvgT5HW97QQf6oAgBGl6DYd8/6cg/fjwqNGbQ/gLvtx/c9Mkzi7Sqaa47fT/v68VlVSMsxA8u1mJfTJrcrb8swVsMjwpS8ZX8JteOjuE9dmxofWGbVsxq/LBNBICaDMjvzrLkJimoXAIlI8q/33OwPyK23sYDJ0mQkaV5RlIpn0qHu7nHKUVMSc5HYWp4OKm2gSb2cCiY4WtCUT+OVU9YQ611BIuoSra+zXu81d7vlPU5PIuqYgDSVUa8HNH3Cc+6U33vd6WDxwgzu7FzahLXbSzs70XO2hwV20TAkQkP/fXbOXQM+hZ5hRZYIsAQVv/KM+XH3iabL8dyLzkwNm5GTzqpURG8VNdKB8aoUweHvoejggs/YuenVaU7lIj7LPa5uzDj9eUOx7nDd1eXmMaA8U2H+cLDDx77a4/9PsKNGeHVPeg4SKFSBF3zIyMcHHLv8f1wgW/TNebgcekmFGQcOOBYRNVvQxmXmDHC47ZsMHDEn07P8adXzzDddbg6yqjdU5AK6xrUXCc/xhGw49wSCK1N0B5rAKvWQu9mRM61DTFrK9TIzmMfMIaAMSaMUa6X4AUR98TYxrkmIeaQ+F0hnBjIOSD3HiXqntRsLW3bgbQYMo0FMSciUY6E8FBmBu44YGCPH9IV/jI/w5t5jx9OF7g5bTCcusWMzD4ePXJtf0B8XEJA659lumBxs1sc8uZ6EpbxCYmi6EABquHJzKKWZsZIogPuKgvbwlTunCnceUIJGZSdiKgUlnZxXFKjklBdEKuwjULUTuEYFMAl0dKH08rVhCk6UQ7LPSFtARfEsIKdQ4gEStq/Cw7Oaxa7lXZB7h1KVF6DtQvsgDxhQrAgBNIqsJajoDLrx7boy6Suh1S4QWAeRtbEQto9ZsTRjgexZq+uOm7ZpEhS5ryFU+Kp94y5eDgQOp+Q2CHkUpO8x97Kg+kM5QdwKQ9JhOVs0sCSgoZISP4stdHxLnoicuF1OMGBsXMjIiRJnlgGvCYOuJ03eDttcT91GKYIIkmqjIdjhMEZviHmCeN7MgKStQeUjNUujABq2+ZYejiKeJku8JfhmfoWyO1PgVEukyQjlyO+fHaPi27ErzZvcelP6CjjpuyqMlrrSWBOmG10WBQK6+N5vRibre1/qgjqdlgCo/TcJOuo1xhbkXAuTGRhv27XXPNa0XAkCWHv5ro+ehRVi5RN6otwu3paDxbjGhWeMob9wLGy8MteJ1RKjz999QL3uce/v/sS//brLzEMAePzDabLHvHAuHKE8NoBWSYReE6Ix4Lu1qEEwreHK/zp+jMM3OGZO2JPMw7azo2UsPf3eOYKboqMut2VLf7x8BzpLzt094R4Kwp+TxGL9sTDcdZ3tYxqAtBc+wufSl0m9VhWy25y8Fx0I9bWhI5FhyJIt4mGpSIongNX1CZSwT6MVQ22MGHUdsf3xDjyFmnrkXrxQ4CZrVnrUw+XeZksfJOIY1HnYJbE41gCXun998f5Bf7j6QvcTFt8e3eJ+9stePAIJwc3Q69jRfwCg2LzYh8QH88hIKxuGmP9G7u59qnP0sVyRvBrdaqX3y0KX4sE6gfsnJUkgHpAahai/36Uj6SZ2ooswqhyuuevwY4WWNVUrCKB2KFkD4qCFAh3wKFEJ+MldoKeMAloY0EIfpyacG5Cdf79nWNfkDZE/TcXeHIoBg03s8LGCK/nUjkK7uzzy7WymB8RLT05s6WW40YVOWChcD98g+8yLwIW9MDiHTDvU0crKV2lhuFqRTIVjzEFJJVLJRLyUhsm4V3/DVVrwzsWy0cWUoF6JfkadGZ5SHFJAB2DVKin6zL2ccIuTJWMaM8/IK5eqyM88DH4P0qYHj2BQDpmbJm0zfOvXIseayGc8w3aztRZEtwqd54nRvKypfk3KwdrFla5JoTyZHL8OycI0iU3jn25x8vLPe5ij/vLHvOlTCekrYfbdHBTkvYBqwzvDLhZ+Q4VpQ3IJmtdpX65EtxMP+E0R/hRnADN/Y9+QdVPOzbnUuv1GDcnY5FK/7DEvr0/I2UUIhQiIeeScNWKis1YOy9QrtoglnSb+qj3RcTV2gKGGy7bI2jT0i7QBIWBmZyOGepYfhGpY9McSNmBZydTKkqWt5F5+2p5Ch8SH+52aNCDwhDki7iveVGh633Czk/YKHfAhG3aCnVGxqwQiE0YCEwSKlmkuhyyeBjkIn4Gc3GiJlVhubUoEpG1DXSzVpQg21SBA0hnfVGUqKaIOhUoY1d6f8I7cIIsJMgIB3RjD0DekApWKCyYPPzo4bdKiOkk8047h7QllF4IjIY6PCU6ACy5j00CAKijgcCyicwlCPKi6Muc5SKeyWs2vLRuJlUEszGXGVlHxbKMIxGhMCvbHDjqRXvU/tZQIpIy6HOT5JmYDlh1CIjRuYTkHDoX4H0RnQkdK4UzQqZc9BxEMZLCqgn3OCLQ/tuQAUMBLClw+rznI4hPEDv1PhWBpoAMIfUdS4/v5mt8e7jCq/sd5imIqhgBb/cbvJ52KEx4lS7qSOniRmfnJ6wqm0ERAqtC7Wcb4zuWDpkdvh8v8efDNSa9l0KXEEIRJUWf8fn2gN9dvMLOTfgn3Wvs3VhRidZYBhCFTONFWBhbuk30f86Ey/lzP1X4we4TgI+CFOYNo3Sa+CuxGECFYetoMmMZ68qNwVqmSnDOxamuhNx30YujgdMqcUCsI5fyPtRyHMKzmDlg0r79gbtq1GPJWKSEPYnuxN/E15g54NIN+KK7w9u0xf/Y/Q7fXV/DvY1g6rC/DOhvZnR/mkFzgh8KwoERtoTTFHHMHY6+w4E7OC51emUxlxIJ4z9On+Gb6Ro/vL1A/4oQdNxwunyae+Yc9pcD7+BUO8OjAA6NHseijHse1qZsrZLPryHRTRFxr1KPba7JcC1S9bm2fq6Iw4UfECljLBHH0sGjW41O25iiy4CbsugUTAw/suwhWbgGUxY0deCA27zBjdvW58gs8sRmXvSH4QX+dHiGu7HH6b4HHUQUyvapEljQ8gDQJqPbnDkx/Uh8OEJgCYGTZIC8QPdeSYVmuhIp1RNULW8h0wXGGs0wIyRXxYeKLSQN4z3p5pWLfUkScF5FydEXmB96MxeQjAWGRXPaQfYIl7C2c9aMzSVBBthJHxBFWw32MOUb5A46XsIAhE/gO4LvDYZEFTMqvT4+cB095I+AcD4kor7B1jjKKYfA/p0rEuBUVlWOPQNCjlTnPFONtEXfRmFmJRfNLJoE0l1hzMiY2FW4ayiL9HSb2BVr/Nf3t2T7kUTUKqqGv/NFpEONGKoTGlQkMXBBSJsLnGDy0WfH9bxNoD/TOcpgScG5g+LPDPMaMHeyDNm4xaJ0h7enDU6HHjw50OjBjnEYO9zPPTwx3uZdhZbb6YJWytXaCDaVUxzJgtkskKbbPpaAV6PMLFs7rusytt2M3169wXUc8Hl/j7/tf8DGzbh0ImY0weOubOrrmlfFzo31NVqUwGbFC358VPCxNtVjnIKnTAYAwKtPfesxMhdCgt7nLNfd6q2QIlcrNBGyTjDkjzKJD0qzrgHGl5KWW8HCJgcvm1vUpBu08D9Mo8WmNQBJsKqRGTL2fkIG4Wv/Fv9F/xccSo8LP+J/3v0T/PntNY5vX0AqoojuzwTkAjdmxBMjnQhj8jrZJVVoRF4KgprEO8zs8DJd4JvxGtN9h8sbhh8FKZ13T3N+2lHBxSmyLIixkz65qX1aUvAYGpAVHi4omBCq0NH5NWl2w+YqFZFgDrHn/IW2/XPpTvDEONDiv2DTD2yFp01GzaWqRbpZ4f2s+5mun3MJGLjDXZGEwNqCt2WL7+ZrHEuHH4YLvDrsMEwRfAwIRycFrZKDQVJ8cizwXcam+zHruXV8RMtgSQiEP6C8O5s0QHnA8qwwNKAmJ6zEwnX12sZSTcpXNqOQBhWoMLcmCAszmOo3IlRSIKi2myuiSsvD5eMZkgBSBTwlCdabnyoMw96ej1AiqxbRsuEJiVCmEkrLHbA2Cy+2y08R1iZYiITQZMBVOPJjiVttVWcOaBmMTNyc5yKtoHeE6em3/67Pr9cMyFURlwfTBg30tdo3Ghhu9bV68bL++bzqN+EW/XmFLjwR9GmL2KH0dTF6ma5UTGSHYYzgwYuz2ij+GMOpw5txh8IOrzd74R9Qxl43X0u4bLG2e2V5TapaHk77k8bnEE6HqyZhRmjqQsLGy1y8sN+Xz28toGVSRQiRCwzOWoVZr24d5/bm9Xmfirn5U0M/ok0VrbkpDbRrj1U00VqK735euVizFjRzkQQqIsnkTzvuywtLPusG9FjIMVy3GawN1IajgghBMT6Pd/jV9hanFPH7y2eYDw7zHQHBawLM1WclZ+mRm0+ArcmLoJLDyAV3ZY8fpkv8MFyICp4iIyVCRdueLlYCak3P1zGvkgGL8ympAt1z7PlQ6lRCZle5BIuT5tIycBAFVtuaWmXEFRekmViwMFdeMz8ylICaNWohF65bty06C6CS7c3F8JQ7HFOHKYnmAKXlNVbXsyJZIWT08RdCCMgsQh3DhQLnCrqQtWWQ6oV97qE963H1YBSy+VbJNldVaMsqbZIBMwlJWX4uxYmYCgNFRzy40GqGWDZsNc4xPwEisy+XvfssoaSiLQPCSu2whv4deyBrFkbFOATSk3TJetyoOgd5o0mBHWm27ORDj/yPR2zaBJEEmPRECPCIKmyzKXOdgX5fyPhoWiUQWeFCGVuTKidirWoGKMnMFriK9iwtCueXa8RGgByEwBNcRqCs8tSC9rC1gZwdU/WScOY/zzJymIUopU/++CiYJQU5yzkClg2sTR6esA96p0Ytr9Mef8ZzDCXim+EKL4cL/HC4wPRqg3jr4SaBsJmAIfT4E57j5VZ0A262O+zDiK/iLSJlHEuHocSKsNlCkopoOiT2cEU2aRupus89Xk97TMXjOHc1ud7EhD7IeOFX/S2exyMu/FAdM2VcrpMRRmJsMKGQw0blavZurCiBbSZZrxNPC/S94hi01d67+Cq6SLbtxl+KpyDKgHpbqghZVbNuEwJAkABgnaiWpriQ/4ALYUoBx9ShcxljEeg9+ryQ2tBwpBqCbms3bomUkbSBZXMaVFzKOAcOBVduwDM/4Bkm/Pfbf8Dvupf4t7tf4f/2txe4310AHPDs30T44wA3JvQ3CbmPmMcgfeks11ZXW06C9t6UHjPP+P+Nv8L/8urX+OHmAvFGhZMATM8I43N+klaorVFt66sYutzwWR6LVrwIQOW8nHMEjGgYVSrbo6DTyj7ykpTZuWi1M6qBnCXEVHAo/TJVVdRHYCa4hGqbjGJS6+v3zJAkwv5eUBqR8rvLWwwc8HK+xB9Pz3FMHb6/v8DxvgePHuHoausLwFKs9gXUZ1ztB3y9v8PHxEcjBOSsihNCobUGTEAIWBTrrAcNJZ6Buc5T1q8HicCyWMjJpdoyKMWhFNEfMFiuJgJtC4CaTcQQAsuijHihZ6MWo7onkCEvZvP8WJXqNT8ozfOo5CiAqixm1pnsH7Lm31tlfGS03gKWDMh3p1LDj/fYLNpkbkV4gm3uBM+SFBTKyMyQw8vNczz+/O8ihbakILFbtutIrqv2mK++nFV1tqnzesKgOCko3pcUmDnML5QIWMyswksl4pBEcOvlcIE3wxb3px7+5OBPMrXiR/m8/uiQdgEnAG/GHTqfMbOrFrqWEKw+lrXc9FgLwUmPJzNS8ZiK2O3WCgai/tmHtBIV29C8QpQW5ThBczyyTplw5QkBkujbJmfXQtZRyY9TqvgrBS3fa2tKCwY+TwT059X/0/YB2fpQUUr5yup4mHiR222nQmzMzGPxu18SofYelIzDAfr30n4aiqhUembduOVvbdLoa3/EM+V/fH11hz8MEWkfxBeBSEYPhww/BXASEaVKKtT3acnGoMTFt3mHt8ct5kOHjZIJAWmJposnrHDaz67H2j7/2urYPfi5dTZcSRxzO5a4XNPLGrQmIpqXi+1jLb9geZ4FZbDk3JJtsE22QY7Tu0jtgAhZNQiN8EVEkbAlER7mDuMcwJMTVHHpcsheRxAEPxS4ULAJCVfd6aOO+UdPGVSVJlog3mVDWcMnFVJ8z9MaqaPlHNgClxruQFZkoBSqcqtchJgBJi3bdTe3m5O4VuzFc9XRL7ofkPLOiJvNXtsKSvSsiMIDxMAebp0Cd75YtA9CTUbI6d71tAgbgDWR0D3yAgbvtgnA+95GVg6CzLJLlp71eIjGuczKFiXtFJhXOa3aBUuip7PADTxnY3ZB+2/BZzgXVJYai5BHVXpUlOBsw2ddnExVmfndojH6APn+mLnRE8VfxmsAwNt5ix9OFxhTwJvDFsOxQzkGbO6cWN4OQDjqtR8IVCLyNuDP7hr3Y49n2xNS8cJudrnprZ636NrJHIdJxz5POWLKoSq8meDK1/s7fLG5x1U44VfdDZ75YyUNFgYO3D9o8wB2/hJmqCyvhrVFVu/liZKBp+YQFLtOHMDREAKuBQC3k0oaRipkWuDkpd1gq3vzI9t0zTIXL39SsKEJHS2tGADYq+ZDKwHeEjSXzYJUrlqgdA8HMHDgDq+LmJHZeYuU8NnmgLdXG7y63mL8YqNEXZmCYidZ0HnryazPMzvc5D0OpcMfhs9wvN3A3QVQAnIv9+V8xaDn05Ocl5YQW69vLisOhUXb37fN39plduyApTXTihcVLIiGKeECC7nTg7Fx06ooAiA8jyaps/dakwRredLCgSqexJaaF4E79hA+HiQxNw0fQx0sgZzZI/Fi0ZyzTRXYHqfXsHLUSsfwfUbXCfL3LJ4erBPviw9vGZgZiGbFzi0KhY549WE+1G63FTUxIxCnF7MlA3MSlnpKDtnEW2bVX8/6BYH5radlm7iQHHS993IRUJFNjq2/Y8hBcyNb1kXnicAZ0mytLZYPsLQVGlTBepE2csJF0YEnXN9cs/Gdw/j1MVgyX1ugpNBeJ3fnIcnAAqsMcOjqgRYJamkBhao4aSOHSUmFyyhpu+DkOuLTQ2DCqQREJRe6oE6WRacMolQkRYWnuMGTRYnQ+o2Ex4Yva4LQogKGFjyxqZHFHw/PUZjw/f0Fbm72YjF8F+CPhG4gbF4DfmCEE6N/W0DMiCeP6Y1MqBynPV5ebfDmaof5c49dnPCiP+JFd6hEqKUHvSRfM8sGMequN5WAQW1Wo8/4bHfAxif8n67/gv988w02bsZn/h4bmnFTdvh2foaZfZUSB5ZFL1LGzo3IcHUmPkJY1x1Ecf2cEd9Ga2/+MXE+fvlzw/aXEhmlq7eu/D9bN0KzyZMlBAt6Zf+LG+tfarlOZze5Kf55lEcrTdOdAB4ibqZDMGgLxwyHTBLew+GuyDnxVLCnCTs3oqOM/2z/A7Z+xv/zqz3u/maLeecQBoY/5ZoYtegsIOf00g04lh5/mZ/jH8cX+Ldvv4R7FdG9EVLcvBfSdPp8wu++ev0zzsYSy+dfHFPNlqj6W1CpEsdGyJNCZBG0s3/LsVn2KTj5bPJ8DjbEu9jGp0rwNEKvHZeWN2NunhmEu6K8AuU2GGq0kKJlFB0AckfInSQF8Mu627uk5mNZkbdSP9uYA8YcMKSAnNTW2JTajXTdSZLLm4LtdsK2m/HV5g6/2XzceflopUICHkjMtspQrVCJ9IjXvcDH4n2VjkEwRiDkAlUPQ/0itn83L9SQLKi+cVTUoEJ9uoG3xIwH0SYMtHxfve0GZrSHrAiJzNKD1MzxKVuildhHUGhMSTG8kP5+rFL7mCwygxDBKwnq1t8+ny3e6xG0h/PDRswJpCpgbsmwVzeW8Qgs6VkJiTfvXwmDLUqwQgv+StLFpxTBLI5mfPKgWVoE4UDwM+AmmQX3E8NP0mMMnavQtT8RSueQu4DjHEHEmKJX6HRNQPMkolzy8ddtOJsqKUxinOQTdmHCdTjimT+KLbKOsLWV6axERDlmCr0+QuT6sWvr51T3T40MWLRJfUUFmmKiihRZNOvF8jtbCNDAiRLn+h7n4c/WSuDd45ntMaiCT48c86xTCV6Jdxv9kDs34TIMiDEjb8S11WXAze6dqpyVm0AFx9LhNm1wnCOc9saJlSMVAb9JuO6eyN3oR6KF79dmWOtrvv13hsGy8jcOi/dHYUI5c6i1wnaRAs+r/2cGbe/l7zd7iySTCzRta5m13SuhGutpiUquB1UisLSqGwhar1VBtBjwIjLXeZEC2FD6YD0G4CcgBCYe45zo4geFoE2hsGvVClGqcp70tWX8cKPjNebkhiI636MLOJVODoYxdbMSCZNHUREGzE4OSpKLs2665++5qfrXsL/qUisD1IhEpBt1+3fWTiBGzcqoeX43699kVJJNXWyCSDRLwiHCHZV0+KEH/gPiphgED3iD5XUDvCuMV/kCN3mHY+kwKnRsuvkA6sUIaMategRZe5aRsrhWKrQmfUXCQXkFr/IeN2WHu7zF2ySvc8qxXtCt7oC1LKJb1MPsPczssQ0z5uzRdQmnXseEIlUHS9OagBfYk0JQv/czVEB3RzsXpPoDpkOwirZt8IRowcv7PZgJpzdb9N9G+BGId0C8F6e6MBS4BMS7hO7VACoF8TagdB7zRUDxEePgME0RN9sdhm1AcAWXYURgJ5oNWmFs3YTOORySVO2JHYYUpYedPU6zCBFddwN+d/EKF37E33U/4HdxXUE4LHPbb9IOr6c9HHFVZfs8Zlz6Qe/dCXs1TWpbhXa+z70MPjRaEtemFhc/z7HzPPJmuT5KXLJ4A8M4yOJ6blxkiGStMore39piMBSBAczZY9KkKlKsLpcg3Zio1J59BqGjjEHbp5FzJfdZRWprpnlaZKIVtD1wxCFvUOCE/MbCN7j0A+bocbEdcdxfws2aVDOQNgQXi2jJOPHPMA+NSAWlOPx5fIZ/9/ZLvL7dV85L2gHzJSPvCn794hb/5dU3H1VU/FgY2VIQL5nRNwj9nDx4PmUjx2xpX7XXpiQAC6O/wCFyru2Xlfia7OSr95WpwPPCj2t5M4IONMegJgMAeycodZBEigNAQYj5fRB0YOPmFen0mEVs6m7ucZgihimiTL7yBwy1Em2cAo6MuJvwYn/EZRzxIh5w7Q8fddw/PCGw7+YfUGHmM9lis2wlcTnsNBHwIIW2CwoYHksCUchVyCWQ2dg2vIHsJBnQhMCNDiiASwR6jDh/BunVaPv8LFmVMzUyKXVrVd8mVRX2t180rQU3Q4WNeHkv+sASADA11xQtGd0TFj53in8KwUgvKEVL7nSu9S5vccydQvgL8aYd9RPdbF9vthkiBNKhYKfZjlWQA3scS8QEj9eacNyVDW7TBmOJ2r92zWtkhIbdW3Xcjf3LGbP32IUJuTjchR6nWARYieIYSQkNaqAbu3eA96BWV2ClWuiEgGhJwGM6BERrN8QnSgqO97I5+5uA7Q/SHujfFnS3WVpIyjqOdxP861tRkYMk0OFqj7R7AT/JBz4+l+mA+37CqAJY2TsEhZ/tmptcwCF1SMXhlGJNAMckf7MJM37bv8aLcI/fxZf4tZeBtrvCmJRMZpMLh9zj1bhHcML16H0Shz43YkdjvdfbHreFa9pMHxMFIlQmx2HxwGjtd58iSrckBG1rgJmUmGUJAS/IYyJpTQLLGqPIAFdUgRVtIMxFRGdOOcKjYHBRK9Sim7kkPiIJTZgp1Lbp3o0AY1WpLgtMFiQATvkIAjVP2VcBKVmDhQOwcyNyIFz0I273LEI2heGSoAXeyxpusPXOjSIwp8f7+/ES399eYLrrsB9Fw2F6xkifzQi7hL9/9gP+690fn+S8rCp7PS02o28E3fkDryt/dr0UJmTSvcOxJme5jvp1gKIFMn2QyeEBG5CXq7Dld7ROpPWhBDCRjqBr8hJILcYZ3ov+SueSeF0ob8IrAjGWgPsko4bDFDGNETw5+IR6TbJTblVfgK5gv53wxfYel2HEi3DA3k0/qgXSxk8yN2qzoJ+bFb6Pb7DSHVBNcTKxIPUnOJ+ks/6f/UxNciBvHhWyF9ZwCyFgSQ7svm9aAfUprIWeNREoqGOH9lQWxfaZDLBmdWB6QFj6OTEboUZ/NsEfAFUsaDEakvnoNs5h/HPI39FCcqqfi8V5y26mmUNFFmZ2q572u66Rc4lRD7lBzFIbOn5YiVuWDJBWONWuEljYmu8JonUyYFChK3pnPS2XoCR5/pBl86+cFUAnI7DIvXovv0sJnDJomBCOGbEnhEHQsJLco6Jc5/CxI4EN+6AjXJ4QvVyIl2HEhVb4d2WDv+Qipin5EkPp8EO6xDfzMxxzh9t5U61gAZsGyRXZ62C8obxapFsBmPPF25NI81iF7M9vBF7QBlsbfomRQ+uEoCUIWtuwVj/63TVIQWuAhkce7xY01aK1Gxf4WmzTAeVGKJHToxlnazaY1hei5WZYy6E9Pk6JSsZR6LBwP8zEDEZ2U1t2ci1sLa85s8cRELvsaYNxVEe9GaDEkhipKuOrcY9/GL/8qI3nXfHYKKoJ2bX2x+dj7nbNtcT083gXR8rOX7Ue1oNkCMLqvSgaZsnAouYqyIWZW1WkWaOiw+/63FjWy3YEcSrCH0jJizNpphVfbvlwAHlG8EXllacVT+VD48MRAuIH3+3nc0GiDwkHJWkoBGObglWviZ1YSCaPkrRNMDqZ7TzJQXEzlpaBIQKElZGQ9QLZL/0cG9GgAsBrsp9Zx0QISFggRGspNAREl6AIBcNP8hivIh9ycOT5i0oIsl8qdutZ/gRe1Tvj23wFYD2rbYvIXd7iZbrE27TFzbzDkCPm7FfGUdQsWK26mj1nREGv12Bm+T7B4SaL4cbrfIE3aY9j7nCXNnIBsygUAhB0QA2wKuMai6iNwZ8777H3E1Jw2HYzbvssNLXAghBk5RBYzyx4QY20bcCsfbqzxIAscXCCJsBbH0G/l6x67E50DZ4qTh7EYm5i15e+IVmINSko0YEut6BUQK/fgu/ugHnC9k8bdG97sNvh8GuP7BhzWhzZLFpvCgexTwZk83fa5tm6CcEV/G3/A/6L/i8o7PCvht/g98NnuJl3+P39C+kRN89L+jrbMAO9KC/u3IRnTuyQhRQs7G2xMo+L4IuOGxqBq2q0Aypt7h4mA5C2V8cL09xIXvKcT5cY8LY5z3bbJrcUhI5BjnVxt5aSQrWgZczLEgpCHfkirfy8JmbGIC8glbgVeL8mPnrMNySQvalAmkmUmUqZNDgAreLnB74IHSWARCPimTuJHK4Tlcnocl0X80YWuPkC6PsZ+zBip9VkYYdXvMFd3uIP0+f4xzfPwd9v0N0RulshwaYdYb4NKKPDv3G/wh9vngEA/q//8uedlxZlsuRgLKKRYKjVrE6pnUsqO5wEpdbKXv723ZuvTRzMJaCQyDQ7Zpl2QwHcYi8OLImYkQ6tlWPv9Vj6Kh405oCSXZ14o8J17JDAS9F5dikvyV7EkCPu8wYvJ3E0fH3YYbrvgNHBH7zsf9wgzSQmRiFmXPQjftW/xYUfcelPD6TFfyx+mv1xEz8XIXjsJm/nco1EWB0LEwnkpeiAm5Y+v1X/BYD2KQDIBgL5sUL2lfDBtJwchrgp6ua9Sq4UOTCot77+zErQEY1qAFX7gBjIkYRL4hXxq+/r6cLY3m3YzXRbtrjLG+EPVOeuZRbd4n3n0ROjI4K4Gkq0UsXH3OOYO5yycBQmlSpuEwtLGmtG35KpSPLwSCJQ1LmM6ET8qnhX/cMrAUz+SDJu76qFMZWCdypAtq0CcuuWAcvvqeC9WfzHhpu0kktYOZxVToq9vCOUPoCc8KnLOIKY4d/ew08zui828JMXhIDPiX20Ylg7kgUSkA08OHHbexEO6N2Mv4mv8Zk74cAB381X+F9vv8btuMF3r6+QRg8XCnzMcI6x7Wdsuhmx8j9Yx9ES9s1CM70ju5Vk7/FpA6/jOY8lBUbiahffJ4/YTKXoqDITgyo0CNhElfBzCcSLDDpDqSc6tmejiMZFemwsGzBtCq7tFEcFG8xNpaubDnS8VB9nNvGmYrcDHj02IrZDELfZgqwoTkepIgRMOnbZCXwdvbQL2nvyUHq8yhd4OV9gOHXwR0I4EvzI+iXaGVQc8k2Hm9PTLGprszuqvIFW+MemZwLlSlivRWnVEfhxNcyiG4a9Bmgtklftw7Egnptm3TD0YCgm2R6q/DtKc8+fIQWmQXO+BhviYM95TB2OcxRX0snBjYLQOJvwDITWUsD5gt4nXPgR1+FYJzY+Jn5WQmBwfu05q8mKZVYRojEAKJucLfOiam40NSdbnmMx3inFgbMDkgPNkgiIQ9fSv/cTKkLQVt6sfIECgJhQrKoE6gYjJ4vV34AELWjaBS2fgIrA/oYQWELgZ2g/ThWpgKozwETiE05irUyWfDySIf6cOD6SEBjsbwS/Uxb4aVbDqLaqtAgNF0SY/ypLy4SJCwqAGXbujOQTVhv/Iju9bjkATVIAIxc+JKPZPPAKkTJkp/l6ZwHgSHgDq981ZkZNy8B+9/SAtAQ/kxtyoAiw3MzxzqO7d3CJEQ9FJg2yAzyBIsN3SmvKGXw6gXKGm3Wm2TOiz7gIk1RHbYXJpSIeNoFQUR8QBg7IhfDn+QVepQu8zhf4f3z79/jzX14Ao0P32mMzEfKWkS4KODDylSy8dk7M8OVc2hhQZE/h13MtkvVsbhNU4Gk9v76Iyfy4D8LPitGECLAgBNoOYG1VGfQvOlYM9k7Y4g0iWQ89EbjwSkWbiBFINtudXzTwfU2OSx1BtH/bMRgg/JzCMkrYan3Y8Vnkq2XDGkq3kgpf+D4Rd3krTHVFS7MXHYH5gvFlP+E6nHDhh0pgvMl7/MPpC/zx9Bz5LqK/I3R3QHdXEI5ZjgM5lCBuiqV/moTgvMVU2yMkY7bRabLo8oNrEUBFrNo4v45s9NWe+1i69yIKdt1Hl/DCHwC/JG4AlHcxYfYy1uuIkQ2EVCTZzXqOk7aXM6FguUdtD7TE4jZtcTf1OIwdpjFKMjDRqi1dAoOj8GF8KIhRkFiTO95Q/1ETBsBPSAi4yW5s1nwuHqNZNCqsNZNHIYcZ5kNtPTTCoKYNd2Vb/2auWusydzkXh3n2VZXJqdWmm0XIRdABriYlK4Qgk1oTQ7J7q8r1Z2FmsvKB5AYvBDgl/bXXU00GFA2gIhmaJACAH4smJ2JcAUCFc+Tvcy8v7LKK+jxxMgAAr9PFg99ZgiUjQ1sclJwyZ18NpdqQ0VHGzk9VFU9mc0UDfOTl3M0KX4rpiq/J4MyLM9j5Tdl6Xljf81y73SBoV5MB+T03SYAgBdYCaP54NVZIza9pSQLo/RXDU8evv34DALh91uPwYoMye/g3AfGtgx8J2x8I8SiJpB8dKDGCJgScEvKbtyBHCKdfgSkAoWDfzfiiu4MDYyyhQQZEBvrhCC9hLh5HyHz6P5y+wPfjJX44XeC7/+UrfPbvgTAw+psMPxWcPgs4/Moj98DpbwhHL0hNdDJdsHcjom5L0kKSSZMJXufi1fug5YfoxMF5bGh+gBBmIhWh+YWQAQ1/31wLtqZFri6H5Bik1ZdAvYSs64bkOIYsYKn4HAHJgZV4beNf1/6ESz9Iq6TpebdyuvI7Y69rgVRksz+UvvIMTEUw84IYtE6UrUMhIInaXd7gdRLp6hIZude59Z5RLhJ+vX+LX8Ub8bFQfsefp+f4/978Gt/fX6D7wWP7PaO/Y+y+OcHdT+hfR+y+jWBHyL1D6Z4GWTvqlBmwbkVby2XvRxS3tj9+LCmQE9cUo02yJToG4u8xs8fNvMXttEVih9txgyEFWaucrGLPNie86A/Y+hm/2bxR0uaEZz5V8uall7HL3if4UDA7rsZGbma4e9mowrCBnxySFu/mjmib+Nu0xd28wdt5g1eHHQ53G5RDRHcguGlRPzTtgbxh8KZgs52w7ycEKrjPPYq2OD5Wu+PD7Y+ZHugPAEtFaJvC1MBaE+WqDV30JE2a6ZqLnjm2tTOXNmHALZHQIJjcZllQ44hl0yheT4L1+QIeEDyEwIDaIgAA5GUTr2TE+uFRkYIKBVl/SAljLrGwxgFrvq7RAHsOfvqK9FzKFsCahMOLn4BJQddoCaJ1Qy4VIQAaqWkW0ZtFknp5rQ+NdsQRWKrJx4hx8v7k22r6q325M30BKqUSRReLY7ckA/p9pUvwCyUKzzYiG+pIxlzn2WOcBO0CEXIHHQGDwucQXgQ5gDNQsqBNWVESJwuIoTiSjGFFIrUwBGhmj5kWd73X2pd8eb9HvCVsX2f4oaC7meCmjNztMF7LzUQzyYRPnRZZE4CN7PW+6srCV5RgHQ8qGHZojXxWn+kJmbh0rllinUP7Na3qAlknjC9w/vbsUmU1LePFmr3VaGmj9SwAmvugQUiWvvKSDFistT8We9/z5zVVw7EEue/1OiqRUfoC12fsw1QnRuy171OP26nHaYzwIy1aGWMGTTOIGSEVwBN850US+QlirrwjWXBb90PAPWw9/ozJE3MMnUrAKUVMxeN+FHlg6UaWStDdhan5m7V2gXgjSMslaKvzQWFpsFGzD7Sfw/bQVKSwmooSCZNb2uU2BQfU4heeq8CRvN9Snyt/4L3ZxkeTCgFUJmVmwpQ9Bgq4z33NwjodoTiUHndugQQ9WCwd9eI+lA7H0iuktcGpdDikDmP2olCorQKahZjlB20TDFqdT0vfvhV8SEWtiQstSk6F6pngwEA4GxFhgO2GmfSIA81EAavnNKphhUBBRTkFDMr6nDri1vaNV1D3+yDvnxBv0m71b+spJ82C304bGV1JEcMcGqc7eXxm01xvdMA1YXNwiujIAnTgDoUdBtXUtznq8zaBvY8fmzCw6FRAw8ZYHTGcY+3TcjU5avkEdcrgkb7/KhlYXvT8QT9yZH9e/LfP/gQAuJl3+H5/gWPq8B/9CxzDFvno4SaP0gniFI/inFn2Pdx+Jy0DJThydODAlaxmSdvGzfBc6ohoYcLNLBXGVDzejDscpqXiKky4ud1hftPDHzxe/Imx/W4EZV3o2a5pacW5iZBnh3HW8afcC4GKAwrEY2HSTdqjVF2J2j74kagucmh04x9h1v8SkS7OREXa0Gu2vXUBSIGh8sXyP6g+XEbMpEfJDFFZzR5DDnibtqvPce6at8EEkNOJjq6OXorNdKkKdi2iJpW/2OQaGdFGEE2pcNZ2w++Hz/CH4wvcDdJaLB4olxm7F0dc7Qb87fYlvg43mDngrmxFpvj4Aj+8vkK+i7i+BbpDgR9YNjYl5VLOQIa0RaenuZcWW2hJANoqfzGcWda4mT1sXBpYJwhLUVFQ9HHnxYu1qEUN0GOcA8ZRyLVBuTRj9piyoAbiXhlEuwBulaQ6EnG14AsQCnIPpI2Dnxj5sgcKY945pC2QtozYJWzDjI1PD8awHyXqW3Fkcu6BwbHAqYX55sy19NINeOaPH3X8fxaHYM7SLhhyrIIo4h6lC5bOx56HwZwDx+ruZH3uIQuJYkpeINZJEgKnJBY3AWFQEt/ECEPRVsAiXgNA9KPBAt8HRRKAagBBcbm4AIAz6TwyLUYnQEUGnCETZhuaFvtQl4uM4qQiZDflDTzQyP8FkgFAqr7zmDXbnIrH3bzBmET6cs6+ihIJXCU3RHQNWqOjUO0iNuuojcCTVD3al3HGx3kJFpUzQAtU2sKmrcd7rapckUTUEB0jFno9386J/sC7lActMTvb+B/1OCAHuCJI9RMpGf4PF/8eAHCT9/huvq7uh39wz3Ha9piGDUrUsUKSSYq8jej2O3BKoJTAuaBE6dW6wCvTmd7Not/AMpJUIOf6m+MVhhTw6uYC8zHKzqaeH90rj6vvpVVx9Y8jwndv5VoNHnBOWl+zLDZuBtLokTq5x4+lw6H0OHAHYMLAATMH0QlAxoZEeGeihRzWIkkGdQNShRo8nQElwYn1LCC1YHlHq+EpgvbajCVpD4CBoi51dY23VoC1sBxroSHXYG1ptbc5C6qZi8NcHIYccZu2dc0DZKO69AN6N2NDSe/DhIE77We7OiosgjVvqz6BoXaH0mHgDh0lXPpTVZu8dFbJSrJ2V7b48/AMf76/xuHUyfoTGPFiwj/7/BW+6O/xd/33+Nrf41XZ4tt0jZu8w7eHK5SXPbp7mSyI9xl+bLQ6mEGz3Ls0P91ZqtNPDnA6NeCbTdL4MTM85rKgYBYtr6b6Fuh6l7HwCWxs0BCCMXtMyWMaI/LohXPGBOcLpijrqMulMSDyqlWwLoSCy+LHEgtyB+QNkCaHtIsgZqQtiQNuD2xCxsbPdUqi9fM5j4pe60RL8QyODEThD2yCJBeG9uzchEv/V0wIqrCbcQl0AxopYChRxlfIPSDMtEzKsZmPT7z0tk1/4BxekZ9ljMMQPEsIiVkWPRK/AtK0nRqSD6NJCHTSwPbtalxiI0QNdNhW9o8eC/VVJmO+k2YTROv+Nz3y/E8Qp/ywZWAtgqkEjEk4GbmQEDUZCn06EOXaoqkEUYUiC5xC2XKsqsWtJQxN6+B9bYPHLvJ2ztrc9VrN8YVHIAxubo+hHd8PDOa1idN7jY+ecPTQFu+6mJOIkHQhYQoBc2+oE8F16tUQnYxRNu9bRJiWvrQlV3OJC0FKN5ypeAwpYJgi0uyBaVH1pAz4IyHeM+KJ4cYMShlwDhzDivvy4LAQ14RNGPDCYG+jwtjNPV7/X7NotsQ4ABXibmVp6zFser9PKUxkYcqr0i7QRF77eosi+Ar7hfFXbKKlOpc6+//6OXlR6pTkVx/WHEtgIZS9zTu8nC/ER2Le4JQjtn7Gcdth5yZd5E8Qa+qgan6LR4J8yTEabAqIIyatfLk4WedAMqIWxqpAaQ6KN3mHl/MlDmMHfzLelhQ8yFzXNwIAQ0R/IXOwj4laiNDaQfKxtaf1zTGFwVVYDw+6N9j66FzVB/BORkdRsPJwsTAfHHO8lRF0qkqFJbAkDrbeabvW61rYIqiWkNY96B37R0vetmvuY++Zn5QQGOScC2HOkp3dzf2KqGEiM48hBMAiwGDfTZFpzFLFpuSqMpgRKVZfBtNrciBMX9mYKMuCb3A/jGOg1SXFZbTKKZN4ZL2oHKEEEYmhIuREhwamZkUijNgRjfmph55ElYoDVWerHJefa4X7hKOHPwwLqbCObJaFMzBMYnubs0Oafc18nWPk4hC9LiI54Jg7ODAObqpGKQUzHBdxt1NzlYG7mtCNOr1QSYVYbsalB8hqmbtMGNg0Q1UrVEZ27xI6l+F9QfalnjcOrMZG/GHHz+yOH4nzJGEVT9RKsHHQQ+nrsdr4hGfbAd4xXn3mMe8C8r0HO4fSE+Yrj363Ac1JmOspo0QCdwVdP2Pj59VGMjTnYCoBb8YdXt/uJRm4iYgHB0pAOMl9tPuOcfmPE/yYEd4cRR0xErgTyWRxwFsvPuQY1/GEr+Itvg43+HU4YUOEQ0kY2GGGw13pMJReFkf4Wtk9Nla8Qg1gbpkPl6JICRGNYMwTtg941ucKDRQNwMyLuOEYrDkrqH1b+2irT0gA6f1kXB27rjdurq0CQU8TjqXHy3SJY+7wb25/hf/w6nNMk8d828MdPcqmoHs+YNPP+PLyHv/86nvsVYXu2h/hULBzI565AREFG90UvuUOf5w/w5+mF3h5usDh1Euhtc1gAj6/PODvL77HtT/BU8Fd6fCX9Bz/09vf4YfTBW6+ucLVN4R4YGxuMvwpabuAwD4AUwLNrEhBWpKDnxnBPYYerq+hc97Kul1ZUODr5I2s6UBUfsR5IiYcgaybMECu1GKRAUDX0cPcoYDwetrV8zmzR6+W5NJO6zBYcUaspD/CXAA/y/U9XRPmy4JykbHvZrUen2pbqHdJ2g6uLC3TdyUBWhAbX8WSz+iWFsQvJkz0WDAL4cUpNGZvClCiRSkILq9+34ax0a3tkIrXcUOSfr5W+Ocogf279uiLVOdgZXbyQkKsf0tYxok8I4RSHRsZwOwDSiiAZtGimw/ZjKwlodBiZbsrdF2CFQeCEHAgFC9yu8XT0vNpIW//cKH8qXGcl2NvF8ecXc1s5yTHtGRJCsCQfFT7dHN28E7QhLrB2ygoSCv4hahirYLWaewxDoFFO264SLGWFQzqV8ScXEcQSRGdeuzqVwPtPhIrd0NfLd3kolA54/oYEbsAgKpp8BRhUGY7Q937hH0UWPe46zD5gpk75EHu+Nw5cBelCkvS1y+BANU971xjQKTHPxVXp32GFDCPATx6hJNDOMpkTjgI52bzJqP7/iDEsHESNCQGcHDg6FcukhZEIna0cyOu3IBLcti5KLMFXDAw4w6oyUDllaDgcW3xJUzwxfgqFovAkcicC3nuCcOEwkoDBrbtPKZaZOg4kmkToeoTVDbzEsxoxhXl/wVXGhhXKvyNCjsN3OEub3CXNvj2cInD93vQ6LB96RBvgbRzGE473G4LpjngMg647gZEyrj2R3jiKhUfwdioXohV+6/THqdZ0SIGqCtwnnHVD/gy3lbuwcwBN3mHb45XeH3YIdx69DeCJIVjhpuSIJ5Oiiaadb3NRZLK9HECOD8W5+TjetpaMvM79hQj1hlZOGBJLKxP3yYGQcmD1JxXQ5iZGEWTgjEFHFOHQ+gxs0ekLGOCLIWU6RAwC4LCgVE6QkmEtJFLJfdA2RZQLx4GvTuXcufKoTIbZQArxHr1sVkMj9pjYeur/fwx8eEJwRn0Ur0GNAses0BYycl4SJXCfCRD8c1zZd1IBN5eK+hVxzsv0IvzkE3WEHKGaosbXIS6CZcgUA17+851Iw6+VGOYLmQUBuaYpQ/LwjngrEzcmURoJEsS7LIKJnkZb/QkKEXJAGUhJZYo9rypJ7kAItVRnzr2+IQIgaE07blp2wDMZ22Ys3ANAr+I3Cz2qlLFFUxakdpM9MiCDCT2K/2Bx8iEDzUIFpXCd8WDG/5917baGPOHwJdnOgRygh1+VPr4I6NVmLNKPlDBxs8oTNh2UrWkMVSOQAkAgpOkxcuYDHtUWdLepzr/vBovMxJeceBG2tTaLDWxbj+ic0AXwVHQgdx75F4WL3HFK3CbhL5LtcKVz1WAMuPIYnI1GGTabuj0OFz5aJVPYirTGtIsUycFHzs69dFhlT4DK8dUkE4XyM/L9WetwEeQJJ0oKUWM2cYccEh93Yh6DugoobCrks93aYObaYvb4wbhVpToNi8Z21eMeUtws0PuHcbTHv+6/Ap9P+Mvl9f4Zn+NfRjxw/YSL/x9NZbLIPyH4Wv8aXyOl+Meo6pb+iBqdjFkfLW5w2f+XqdVAr7Pl/jz+Bzfvr3E6W6DzVtCf5fhTwVuyKA5g70HgtOix4FiUBMoB3D38Fj8wvHYdM151MKjmfaIlAU14AAUVIO+ZRPGonoKmQ46uYjZO9yFjewbLiMVj6DfT1n0A4YUa1GrgHJtG7DTUfjAcLEsDsFna2XbOq/CeQUtp1KQ8ETgLJLmU5Zx/VMWgn4GYVPmvw5C0PZV5uSRqiCKwMKmm/5YmHQtAJX11Dn3FGrfu4446ngMEQn8wnJQUwZcMMa5WynAsVe/afOI7rRn0wHcMdAVYXd2M7wr6PW9Wm89eUaenJIUdePIVLW/XYaMi5k40UyKVrAq3aFqhOcOmC81OemBtOHKVH5K8rTJ2baReeELCOIimaRVLQSu9psWwgXxSEUy32ORhSyTJD4GUYsTV1/lRM2vOzVTCo8mBQ0hzsYa15vaYtNb2On7xaMI0So5OPcgUPtjedGzKYPHxg4t1NjkqcL0Ie6ymD4Vdogu43l3xMbPSMVh6CK+yw7jvbredYTSqX5fiiAi5EhwnVQUVqkDQO82kqDpZ0msCXpyoOTqPibHpEXMtIKNQaq+bYf5MiJvHKZLh+laEth8lXF1OeDziwM+j3diuAPgpjh0VHDggGOJlWRq48atrW+dreeFkX2eFBRjx+vKZ+2DDMLcyHo+qVARn30HJBkw86JiiN/CCXjnRIJrfia5z1JyIBKXydfTrrbTImXM5OGUC3IsHb49XeL1sMfp1RZXfyHEO8bz/zCg+8MrcN8hfbZH3gQcv4q4/80FUg/8+6+u8L89/wIhJnx29RtcxKmqIxYm3Axb3A89SiHMkxzPfjPjN89ucNUN+K8u/4S/777DzB7/evwbfDdf41/d/BrDP16if+Nw+UfG7k9HuCmDpgTMCdRFpI1yTeLC5xfBpl84adNYjR1yiySUig7Y/zcEwESMPBUxbdKJluImzOSF1OezeCQoV4cLoSRBVWZFV8mJ++Pt1CO6gpdxj07/LmtB++a4xTioKyGWJCBtZQ/JW4bfJfS9TBesiuOzospErqgQXF6M/Ii1QB0FzWTn5VyzGJA9iyfc5x5zCbjz2486vh8xdrj8bBu2jR56ACl7EDGKK+/N2hZ2O8CcUdxCSrSKdnlR2TjJRs5IK35v7QHp31coD5qNVXieqocB15YB4L2Y50RXEL04uRlqUIoIkECnEzjI+yiZKtqMIgfEek22eFhCwEHfR5RkwNytOGj26bHIKT9BPHa8uV5Q1Dg8Pg4Ht0lBNoQA8t2BAQ4wQxpDD1atAqzn0d83bgg0iYFWkecz1q2wUZslv+cp9bEfdkzXGgRPwxd4LEwfYiyhkmY9iSNlooKNT2AmhJAxeBmtLGrrzL6IP0ORnj4BVRffE1dFOmMm279zI5jzrmAitYyWaq90fsV5KQHiBhgL+ijsZfMVAIC5Tgn5qqZnRNTzZMC+z2fn+MfifOzwF1Et1FNfJWSbJIHqNbfAZ48mB0xYbjBanqRB6abiaztOrm1eiJXKcp+LA80OflCY/u2I8uoNqO8QSoHvO3C4wLyPwlTvHSYfkaPHawLuY79UuACOQ1cTAeM5B1dw1Q140R3xzB9x6WYcC1dF0yGp5sAo7SWaC5B0jtI5ndhyQnz1VB1GS/d4q+mXjnOEoCWun6t1LtW4ogFGKmS3kPrO4fkCoBCYHIpjEMskAhEjK8I8q0x7bjReTLq4PldtL0sx6JQf0CYv0jJbxzlCACxtcqaFT8eFkLMU53P2OGVBa46l++txCFgrUEBqKoP6iRay1mNrEgHKrpTv0dlJlMzaTIBADHiZsyxEyL3aFc/KUUjL4oUmIZCMTLOxXqp0DkDpBR2QSitjE4S4Zo5wc/bIhYRoN3sU52VhdmqkpFU+CuA2RnQUxADGZTCugp78EoC0k/ddOkbZaMPSNQvLXyFskoKI4bTg8r4gBCHuRS/HoVNyDYA6fjiyeKvbzK+NRRnbdmYnbQO9Eex7mxC0kwMrYRvd6YybMPHilpiKCOJUchcDaLghZLrUFu9yKmw5BI8dmDasZfBErod/Gp8DQIXy2v5nKg6DoioAgMAoHSNvHdJlhJs8AgAaPUokITsRI7HHXRbYckMJMRzwJu0xlUscVIkSeal02TNKJKQdUDrC8QsP0JWaOel90hGG56I2N10B04uCsmFcPj/id9ev8Xl/wAt/L3a6ILzOonthmiIWj7rvacLnPhD6Ny0CB1RY3XgGTxqNBnyVx46lOqFypnrN/Wg0689qIEHvgaq+Wjk5ghR0lGX00M/ofFd7zrJmOTgi6c3f3sM5h94BV7hA7hw2bxymC4/SBYzPOgw9KxKp0stWdARGfz1g28/48uIe/9n+B7wIB3wRbnU8nLEhmWB41p/whxcJJQjJdd5d1SIHpIjnlaCv9lmZFH0NT3x+gNp6fizOnTQf00ApTVK5mB4pt6UJR6WiK7U4Yr2HVByvJEkMT5kwhQjnC06d6BUY8loKYdSxRajTqbUM0EEQ7wB4t/a3MFRghhQPoyaQRRGr6uxrQnwZIA+UYdn0xk2HnBx+UIS+cwnX3VCNzj40PkKp8OHvSiGU4h+tQKvE8dnfOBITBiIghIzocxXJISVwALJWu1Aq6zmzEv2iwHKUAddjkcy08yhtV3UbVGnHAFHl6jJCzNjGGfsoevAbLwlB6uR1p+wl2/JKCIle4SPp2aCYcx1V6KZWr5YQOCxoxqYIgTAwqJPMgX7hhKDWK6viZUnUiMSL25KBzotGuBFAAbkZ7YbLIHjmJgmQnplpHdjIqaEEJmttCcC5b0ErHrIoXJrlp8hX5+Kk1aHZdiWRamvmnftDec/CVFp49/ETwK0Y/c+Mfzw8b156QU9scWrbYxQLOBPSFpguvUgZZ4bzSVpUuojMxeOubLChhN7N6CjhqC5rx9Rhnj1oFldQhianJO50VOR+mS/CWt0zAvOltOfmCwaeTQh9wu+ev8G/vPwG1+GIz4IkBIfS41XZiTiLQq+G9DyWDCwKhe8nzbR/12nbJpNa2uLhgv+zQ5+uHT+mwEDIgqwlt3AKCq03ffu9BSsyYLUMYQVnTcUjGEpQAqKX6RqTvd34hN4noCtNEaPE1zmB7w/gwvCnE7Z3z8DeYR+DkED7iOGLDdLWIW0J06VMOU1X4lOQ9wXdFwlfXtzjtxdv8C+2f8Fn/h5f+DtEAiILwfHSD/hqe4v950cM+w7Hq4DhS4W9O9Vf6DN2VwO6kBF8QRcE8r7qB1zE8WnPz3tCHBllkk24P/L71egzO+Rm0sCMjN4V54gmMxarYUb9W06E5BlwjLkLVcNC/oZUat+BJv07fXs2YYbAlczeqhTO7OEqSdjLPpS8vgdd+xKqSm+dmFMV37kPmGeHOyZ85xhdSLibN+JU+hHxE1sGzXfb+LXCX/5tD14/T6le9AznCJkEsmFleBbjD4CWapohaEFQMkXkCledQ8m2EVtCYPC9TRdYu8BGO2wTDE5+75mkdRAyAA8ORatUp/c9i8yxjiVWOUmmZZFVW0r20ARGEwIvCAG594y8/YQ4374WEuHDx7btgfY9tPOwi5ywVGZiWlNqL9jCoP3zlsVjttiGPFRjl0cIZ1XcCIuENfNautp64B+NIDfmP3+tGPNye1WRlNonXCZCAKAKa1hCqSJb7F3dsAwilWrCwymiMTekTiLIvaJQdrEheyeIiqECYCwiJxGY9yJ0wtuMbpPQdQkXYcSlH7BzEyKW6YZ2c67tgXecEOMOtEiCtQN+rOo3TsEvQiq0EfqqGbIkBgBkvSiNU+rqrdKC21o0rYPz++7c3Ks9fpGy8ELCBNdlEbKZCfNFRLi+Ak0zyr0DzbMgXarDwlm2P/Z+uTdg14zypnoGdwX7fsJ1f8KzeMTebJPBmFmuxw3pBEkYcL07wbmCMUTMvVy/LopIWNdlvNgfsQ0z+pCw8TMCFTzrTtj7XyYhOEcJfm7rKIMQ8XiC+Zg0f4tMyhMsUuOcnKz1bUFcCb1Uf1VzFofqc3H+Wuax8JgPzIP3ybwUSnbus7y3kgmTkUjp8UmN98VPahlUMR+tcFghNi5Nz8OShQcQAaMEJ/yA4lCiIgROeitVN93J5o0eQogLLDBeJpReYZTUuD/VSritfLhmt26fsNnM6BUd2IQZnUsVUpnChMKEzklPaAoeSSWUDQ7iQuvNloGS3UOEhJZxOe/1AnAFIdho29PCa6khFS7mU83PZdl8bJRlsWd9N6QvF6dAapm9+IfrmKhxCPJZMmCb1oIOSM/cYRHVOdemyDqHbgJVYxYxpTR7lFmdLqfF6VKyZAaVIqMf2uOkopBvYaAo/6SOHp4tANJDeXgwS3mylsHLe1GQtGqKsBzrdlEgAAvRZUlwa882kPQd9SIfSsRMHm/zFnPxuEk7QQeKFz7CLq8HJtr7sK12e1FU875gt50QfcY2JjzbnLDxM/7Ly2/wd/332NCMKyejbhs3Y8MzsraSzr3hW3jf9AhajkirQfBYK6COoYJFFpmVtENPixLQYKQvHWkFpKUY5fOQtiylBmg3fjmePFnpB0EOmYDA4CJ/J2ujnITOJdnwlUCdlZ+TQdi4Gf90+wovugPefLXD7+fPMRwCQD0uXvwNwqmgfznCHycwc1X0NcZ/2kcMLzzmPWG8JgxfsCQCLyY8e3bA9XbA//D5f8Q/23yPZ/6IL/wtNpQwweEHlT7+0t/hS3+Hr+MNftW9xbF01V8GWHgckcTgylRozQxp7yZs3Nq18qeGFSNSZBiM81A8qPb8WZDHhIfSxDZ6aO3OSFn4LrT46Kw0UxQ5dapIafuI6N7Ys1IVwEPC4oVj/1fbDC5haR/r/2PVr5AJN/ndXDwC+YpgFJDqx7xjK2crjFgcekmSWn9yKIlRKODoewy+YOw9TuGhaN374qO9DNqbg5l0hJskW7LFxuA0JdvVxxPX+W8QV0IvuQLvl0yrVpgq8CHkQLkZODthVip5qtjNCKwqdADS14pFRky6hD7O2MQkanHOZkCXm3bjHQIVJHaIriBzwpQCCi9VN4C6CTILmaMei/r+l8/hmozQP3EiYFHK8n7kOx4iN1gjAjZhsPJtbxi81nfzmrVaGLEQWCC6B86JtHYiW0YMl8mC89FTWyBnbUFkS8IywSVh2FZjKxWkkrGTs4P+zoPEgBO9AYHpzx7P5f3thp8Qp1FuxjYRtAkch8XpDEBNMNsJCvZUW2Z2HRl6MpeI27TBWEKVpk5F0DbfL5saNa+xnHPhkFxuB1x1I6LPuIoDep9qkty7hN92L/GFv1V1QknuPOs4IUHH3JJuGG6VFACoegQfG9bb9qyrHz+9hDFNliCb3oh5ZdCC4tX7uMIJAICSsbQHihQmpNV2RXnsdYilJVfJn2tp8EgJX8ZbXPsTXl5e4P6LHvcXPQ7DHiU4hKPDRSDEQ4QfM/xRWNRC5HNIW4/pgjBfEKZrRvpshttkfPXZLf7+2Q940R3xf774B/wuvJJjqu/spvS4K1tESvja3+PaZXzBJ/wm3AhbnTL2riAzcGSPoWkfAuK40CvBVVxKnyZaGL3lIrVIgQM/uB6crlOPJgUaKx4BPyS5Om2rCl9nUZkF60YP/dlJRs0ZD9cR45M1XDjWvJYdJMkkXt379b1RqSOHjxLz7VesoFaWwSn2UiiBCRwd0uiFkMxA6T7uzPzksUMYMqCVcyVgoDl4Z71eAgnZRT8cZ0JxJDQbktS3blJu+Q6g9pO5FOQgCQFn+ZI3tWx8Ff7zDFIBoq5LkgxYz7zpbQNA7zKSCkJs2dVxyOgeVottP5jt+9lJ5MdO6C8U70oEgIcAjaEXq2QAC8u2ZspUVvCcoATK+G6EPlARBlm4g06ZBJLj7MAr7/JISXrL72kZ2BjP4na5ZOlUjFyjRistPvuI7PB7FQkfP5gfPK3wY1HMFD1om0qnWjrXVKEQci1pmwkK47MjJRPKhjXPHvdTj7dhgx/8JQDgVk2MUnHwrqADsOunej06Z+JOomFgyETU9thnmwOedyexNg7DSt2sd/PKrGgiwLMw0lvnwvdt+JY4rI/vw8cL4VBlsLX3a9LY5oz65ByC5n3Jx5E1jGcn5yAUmMS5c5bQ6ft1hMRanKhGLRcSouJ5nqnXc/GL7r6IOXUYeEZhh44S4IBn8Ygv9vfYxhl/+jLiGDr4gZC2HuHk4WZGEJddmaTyMs52+pqR9gXlKuHFF7fYdzN+e/kGf7t7JdbL7qRS061bqZqXsYmNyfNKws7olNjZDk600tKT/t6BEVGevNh535QSoDwCpfAHV5CLTBYUXtj6589xvp5Fl4TTQSIpXpjQhYw5eEGvvYpTa6utPp0hbO94i6zra/HawqucMtmPrD3dIrJL603W30DSpsmeq9wxabEg47DaEq/TC6hrR/tGSvm4feij7I/bF5LNuUEGVGYY2levb96qV0LNlhAgSAGUM+BYWJgocDoWaLrt7UJmcLSJIclYz7s3YedKZXTuu6kSCfd+wtbPcFTQu0VhK7qMzIR9mDAX35wsOUHR+AaUV5msfFe3QFbnQK10hxTl3yocYX/zsbaU7wtuEIJ6jWrCVhNYg6f1YnXEiL7ZpHyqG3f7BTw0qCmN5Gcqvm5wwDJy2rmM3osM59ZNOj8/VXa1jR4uWbtVlU5G9ExgJwv8Zu0Cb4ZSiYFSBCl4V5gewfs2eOdg2gNPlQhYJJ1Fjl3CvpsQXcZFHLHxqY7aFibcxR7kuepTyIitttYgN/t07PDaMaYsOvcAcJojMhM2IeF5f0SIBZfdgLQzC1kV33IZz7qTomIJOydSqZ+HO3wW7hX2Haul8sCCbESIaE2mAlPSEllywy2xaAs8knZ5epj4RVonEe0iba9legYmjrU81xOeH3u7TbLJ7MBJ25PNmtP3qSJ81tace19n/NMpqHkUZKHTjydFAjDkWBd5W2+OpavyvDs3YocR/3z3Lb7s7jCUiO9eXOFm2uKYOvxw2mOYA6YUcK+jhNZi3XQz/pvnr/B5f8BX3S3++eYbbNyMZ+6ISzcgUqkqhkf2uOMOM4cqmgUCjqrNL0p+Rb0RgE5VD4ElGRg4VpLxEXKNdVhc+n6peAxAb82OrOJuJw8szDBtbbvOAAmpfR9GXHcDupxxSnLtD77gmAllFvyDM68LXMv9zt6WoQqr0XInaDV7RuxTdSbsFKG29wfI3mLTb12XkXpp/6WdTNW5jpCVJ1i5RipoZq9ha35hEguAj4iPRgjWzl7LF51Vc3VEjDWhIqCmVYXA+h2FJTtfQe5y8wVfpMpUqPUcTpL38xAeao0tLBMTm8lZyYS5Lpb2uGrCBAeonrT0u0tNHKzKld7Z+kpYXP+oSv+OJcCBJTGguBC/mOCeMiFo0YAVWtAe+6UiNYLaql1AC0ogx6HhE/DarEagx2WGVo6lq/8uTOL6pUmUeVpElx7OqJ+NIRpKkFt2Ny/kmTpp0DBbWTkEj5oSFf4xgvvyOEA5BE+z8bR6+IHkWu6ckMjsOkksAkLGTDcqAYGWaRUHYTfPHkMIiD4KzyV5FG0T2MbfLsxGKtr6GV90dyv53EgZX/hbPPNHeLBsGlQwsMddkTnm1hHu58ZqMz/bPNr/177WaqyRAf8OX5SfHdZyzFr6ERbejSFmWlx4V2QmncQHhFksopmcoqTra8eQw/MvWy/s83sqeOYP6hgZ8CIccNyIxfgPu8vqBntMoqkfSO6r6zjgX+y/wefhFl+GO/zT8AYdFfS0XPY6GQ2wnNNJxx8LnKpDCnIQqaj58jKM04ahOOZQKe9djpf/KyKi1hKwtcOQAkCS2NrSbN7TY2REu+4iyVoVqKBzGV1ISNkJwlxKJZ7KfYlVUrAKWn+vl7Im+8Ypi07WxXi2D5mmSHDq+OoKXFBEPMo+WcC1DV/J8wEwefdzhOCDxmab+MnmRnXzsYXbxiOyzuZXeEPTgNqPkZsMhRRKcYAX4R+QtQ6Wfus2ivzixifdyCXTBh5CQnYBpEcWsM6JcYQjXm3ubVLQq/BKr89jRjuRMnZ+xM5NjT+5yscagUpH5wpcNbM5lg7fu8va57X3OGf/o0zSnxJyeHm5COiMK+CLejiI4MwuzlK1hhH7MGGvrHKzZe0VMjZBlejF/7w0fTMxCAmVi2HR+4Stm+BIrF4v/IANiUf3xk3ad5QZ9YEi5saIY5HtxGrCgIq2CAr0632Vvy7ypspiKoXkmt+59eOzJhZPhBR4HTPtQq6s7Ms44Fk8IRWH27TFWARt4QJJqHVPYigsrOJaFLhyaszCetYR2RhkVKm4jOgYnUuKauXKYr8Ox0oEk8oxVT39RU1QNCDs3+/9bGer4Yf8zUcfv6Y18dTCRKXTdSYtfilQXZFSAI4qSOM9CicsimQOhdHIgqNyDmrLrlmXDOk0xMysbgFU4a+BIzyX6poHQE2QEnYuVORmLEF1QKhZlyZ8Hm5x5eWczjrnlsGIYBQAk66Hg97HgFXIqZ7HCW5FEHZ6DxSocZUiAzOH1ZRITeCe6NSfFwjvOu9tQdZW2DM8HKR1YFpJi0+AeRjI2m1r/85NuAon9C5hyKEWpKk4pOgxOdaEykbP3/HmtS2w9Cx083PSKiDPuNiOuOoHbPxckWqzKnZU6tghAFxuRikciDEnQs4Emh2y9ncsAWDP4F64ctRnxD5VBOljddc+PiHQZKByB5TcJ8xvqvOS1PSzAbkxWNfkAvEHKAavQPgErAQfgog3dD5jHyYEl3EZR2z9jEAZF36U7ApLZQ/I87aGOxlu7Y2Ahk16DmW6JUuzC2znx7p4fhbu8cxJNbVzc+2b1RuKXVVwuylbHLjDTd4jUsZd3tQNM7HDROHRpOWnxgPCJ7FUmIa46IXhbXbYSQvlujuh8xkvuiO2fsKFH3HtT3W+faMOYbZQLZm3LGQG+Rsi0oZB05EyrsMy7iRVkDyvMMkXm1bXJASFl2trnRQ0icGjx0LRJwCkgkTkvYxsmUKfCRXVpjA1f/t00fXyOTfqarbxMz6Lh4XJDQckUZjj7CrsvMCOVGFBUt0IAEhZOBbzHJCTw+iFCJvYYQvGhR+roU6vG79p3e/ciL0bFRWYm4R2Df37R4hbFq3YkEV5tKb8eSGV8y+ECmhCwLQQlN0EFRojcO/AVFCCEod90cmBpW0pOizS8pRes5aQViHqxdS5VAmbtpHbZjpjccVsY0MTNjoB9ZUSxYcSMahnwIammtBJu0fu18KEURn3A+RemhTpEfGvRV7azljWJKI0SV3L3TDTKmAZIxXeh1a1T83v0HiX2+H5/wcvyoQWbdGyTghS5cnY9XvpBxSQ+o04bHzCMcjUxJQ97n2PE+keNbsqOvQgiIG4jJe7oNhFIwL3YnvEF/29SgwfxdZa95k2+endjNvtFo4Yp5hw6xg5O5TkkDOtEk8z6yMdDd1280LM/0hux89yO3zX6mnJQPtedIuSvjaa9by2HewuOle5K9XzIDT2uAZBrxICvehH9YkvWLTQ21nmxwhtgBmyaC9Nn9vQgIiMjWbU1pPz2i70ACIVzFwwwVVFN7E4lYvR0I0Kxf9CN9GDqG2ChUy4EFq4tlAW3kCCR6k3kBwvVfdiqyxUdlhZ4I4YEWtXy5agds5LaM+ZZ4JNqj/oEZ9Dc7wUX++9zt2yGLwzRX5P6vyoz8FPiBBMS71o/zjrtZsQya+sn+uUQX0TqH7qLMCGEG3xkCsDLIluaJEwSohOkB6vbTHrxT/W8/2pFf6HJAPtiGF1a/wFULIPjgbeZaerk9fz4LCWKcZyzLP+XEWzVs/JaP/osfN0XvE+rpy3TCStBJsoA5jqzx3leu/Ie2tae4/cH+/ScyhaiRtKZI+1JP9951eqeH7nXvCx0ep11PeHdysWAlhxkezn82mD976mrvtR22bS5vRVxTb6jMkXFHLIhSEsQ1o+c72WJBlwQRYq48JJIZYFmQ02xSNoXXRJORup3qNmg9z5hD4kZCZEHc8vxMhnbqzOMUKU5+9CWo04f2z8vISgHgdqNnYsyUCzyJE2Srm9eM6SAj5DFIKzvmuqFWzvEi79UDcvY0IDy0jcSKnKhI4k6bXZ9QJrjgGwJAK2efWaeW9orpXtlRtwqb+/dISelDCmCceMjJklKSiY4FguzJ0bkUG4z71ceExoWxVPcvwVriRihTZlQSNY5YsmGWD1cciikKaV/IUfsfNj9eVuZ8wBSaImePROtbodatLlHD9ACCzzjZSxdyMu3UmOJ00VLSnalHsMFqwtqbPrqCYFH6soSNYyoJZU8fjjnig+2x/hiPG8P+Lz/h5bP+NFOOBFuEfMGd+5KyCr7PcsWvbEqLbaaQuVHWZEnZIBtFNSHJwrYA90Qdo+z7oTXsQDvoq3FR7ttCIym1sA6mJJSmpbQ/0fwheoOvzNRvExLYYFWXD/yZICUulipiJIjLYMKo1gk+uEkiQAQFFUAICgBfIEzbVpvR5xoAOkvfNg9I3Wsrt2DAQ5WJPjWp05aUsqMRkJuQgHYKaw3K8OqoKYEElIvjWJZ1oTeWulT8j8cF7dHmdoQoG89/OCSrgFT2jfevb6NlFwnhQ8NMlaC6e1j18mWMyfBU3BIxMHKMDWT/W1nLYNvE5OiT5NqKPmj/XnQ0y1LbuJqXrmbMOM6DP+ye4Gv+5vEBsEtaOEvZv03BB8KDi6Hrf9FoGKTJsQY9bpq5TXn9urMqFXrtw+jpVT1KrPfkg8SULQBtUFG5UMhjaZ0nyA7Of2HJ8lBbVaV1LfzssB3CkjOuqBBNazzx5FSTuh9nNcYcyw3v1iotQqrZ17ZW/cjB2NCrXO2BAjAtiRR08RDgSvZLqZsyQFlDGwnIRBs0CReZWZ3dIQ+J70uBOWpACQ1g7pXG1DIPQ6itYpqcX6mhs3VylVR6VWH4AgBAbdzqr451lQlLZ90IYR2ORiH2v/2ljPmQkzOXRs2gSPV62txsSjOdRH2B0/OGD1R+kXw5WnNDvEF9t7AMB1HHAVBpWIPeHKDVKV6fE1uNlUL5lELz53hNIDpWNso1QLqTigODCxblaMoK21vR9xHY6ScDQcF69QaVUaVCjYrHKBdycC56qCjwkMncf7pgFaieOJ8YtsJB8U2iIkx6DI62JERdHInO90XcpJJgvs4z1AANrEVceiS1l0B1oJXbdKkJZjbPdRu7Etx9ytE4la2Wub1M2InFAg5zbCNEXefa7sWqgmY817qfd/s3E6blwr9bE/RWviXbHwiNpE6iGX4OdwsEywqE2BqiUySQEaFNUt7KqU8JyF75Oya0a97bssKV1MKu2ccaGTRRs/YxdE2+OL7g5fxbeN3DdrYSv7mCQGghR81t3DUcHBS0tpMiOls8I2UKmk+12YcBmkpW7TXR8TfyXc+lP87zGeOin5FO+PDz7e71nrlkmR90CoNZH7dH4/NJ5aPfSJJ1gfhHGknnIz/hS/fLy39fGR9+s5yvwU6znxUw9ff4pP8Sk+xaf4FJ/i/3DxKb38FJ/iU3yKT/EpPsWnhOBTfIpP8Sk+xaf4FJ8Sgk/xKT7Fp/gUn+JT4FNC8Ck+xaf4FJ/iU3wKfEoIPsWn+BSf4lN8ik+BTwnBp/gUn+JTfIpP8SnwKSH4FJ/iU3yKT/EpPgU+JQSf4lN8ik/xKT7Fp8CnhOBTfIpP8Sk+xaf4FAD+/0HZlKtWgnAXAAAAAElFTkSuQmCC\n"
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "First 10 labels: [1 5 2 8 3 3 7 2 5 5]\n"
          ]
        }
      ],
      "source": [
        "# Use X_train to visualize the first 10 images. And Y_train to print the first 10 labels.\n",
        "for i in range (10):\n",
        "    plt.subplot(2,5,i+1)\n",
        "    plt.imshow(X_train[i])\n",
        "    plt.title(f\"Label: {y_train[i]}\")\n",
        "    plt.axis('off')\n",
        "plt.show()\n",
        "\n",
        "print('First 10 labels:', y_train[:10])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kzoyeXHOy80N"
      },
      "source": [
        "## **Data preparation**\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "NqndzQXng9rL",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "100d647c-1ff9-4431-f8aa-399d48ba9451"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Shape of the first image: (32, 32)\n",
            "Pixel array of the first image:\n",
            " [[ 66.0302  79.0289  86.3271 ...  27.4839  29.7117  38.7817]\n",
            " [ 65.0303  75.0293  82.3275 ...  27.071   30.0707  42.3684]\n",
            " [ 66.3183  68.3289  75.3282 ...  26.843   32.0705  47.4388]\n",
            " ...\n",
            " [ 77.4528  73.8553  75.9691 ... 105.2265 106.5854 106.5854]\n",
            " [ 77.0507  71.7523  74.8552 ... 104.5147 104.6888 103.9878]\n",
            " [ 74.7628  70.8664  73.8553 ... 102.5149 103.6889 103.39  ]]\n"
          ]
        }
      ],
      "source": [
        "# Print the shape of the first image\n",
        "print(\"Shape of the first image:\",X_train[0].shape)\n",
        "\n",
        "# Print the array of pixels for the first image\n",
        "print(\"Pixel array of the first image:\\n\",X_train[0])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f4CQkKtQ0XII"
      },
      "source": [
        "### **Normalize the train and the test data**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "q_yUUTp_mUzB"
      },
      "outputs": [],
      "source": [
        "# Normalize the datasets\n",
        "X_train=X_train/255.0\n",
        "X_test=X_test/255.0"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "t7FSqOpamWkH",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6b546b0a-5278-4f99-9869-46d4d729c7be"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "New shape of the training dataset: (33600, 32, 32)\n",
            "New shape of the testing dataset: (18000, 32, 32)\n"
          ]
        }
      ],
      "source": [
        "# Print the new shapes of the train and test datasets\n",
        "print(\"New shape of the training dataset:\",X_train.shape)\n",
        "print(\"New shape of the testing dataset:\",X_test.shape)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0uLxXBpz81vk"
      },
      "source": [
        "### **One-hot encode output**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "zL0lYER4sqWw",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e48a7bb8-17fd-4994-a578-717fd7691e55"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Shape of the one-hot encoded y_train: (33600, 10)\n",
            "Shape of the one-hot encoded y_test: (18000, 10)\n"
          ]
        }
      ],
      "source": [
        "# Encode the target variable\n",
        "from keras.utils.np_utils import to_categorical\n",
        "y_train_encoded=to_categorical(y_train)\n",
        "y_test_encoded=to_categorical(y_test)\n",
        "\n",
        "# Print the shape of the newly encoded target variables\n",
        "print(\"Shape of the one-hot encoded y_train:\", y_train_encoded.shape)\n",
        "print(\"Shape of the one-hot encoded y_test:\", y_test_encoded.shape)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yH-gVrzuByNA"
      },
      "source": [
        "## **First ANN Model Building**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "BcKRwrGn0XIL"
      },
      "outputs": [],
      "source": [
        "# Set the seed for numpy\n",
        "np.random.seed(0)\n",
        "\n",
        "# Set the seed for TensorFlow\n",
        "tf.random.set_seed(0)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "A48z6ucF0XIP"
      },
      "source": [
        "### **Build and train an ANN model as per the above mentioned architecture.**\n",
        " - First hidden layer with **64 nodes and the relu activation** and the **input shape = (1024, )**\n",
        " - Second hidden layer with **32 nodes and the relu activation**\n",
        " - Output layer with **activation as 'softmax' and number of nodes equal to the number of classes, i.e., 10**\n",
        " - Compile the model with the **loss equal to categorical_crossentropy, optimizer equal to Adam(learning_rate = 0.001), and metric equal to 'accuracy'**. Do not fit the model here, just return the compiled model.\n",
        "- Call the nn_model_1 function and store the model in a new variable.\n",
        "- Print the summary of the model.\n",
        "- Fit on the train data with a **validation split of 0.2, batch size = 128, verbose = 1, and epochs = 20**. Store the model building history to use later for visualization."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Cmi81Gr5sqW-",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d8a5ee82-2ae7-4c3a-818e-eb363286acfe"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense (Dense)               (None, 64)                65600     \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 32)                2080      \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 10)                330       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 68,010\n",
            "Trainable params: 68,010\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "Epoch 1/20\n",
            "210/210 [==============================] - 9s 19ms/step - loss: 2.2862 - accuracy: 0.1324 - val_loss: 2.2137 - val_accuracy: 0.1688\n",
            "Epoch 2/20\n",
            "210/210 [==============================] - 2s 10ms/step - loss: 2.1184 - accuracy: 0.2308 - val_loss: 1.9948 - val_accuracy: 0.2845\n",
            "Epoch 3/20\n",
            "210/210 [==============================] - 2s 10ms/step - loss: 1.9204 - accuracy: 0.3253 - val_loss: 1.8370 - val_accuracy: 0.3433\n",
            "Epoch 4/20\n",
            "210/210 [==============================] - 2s 9ms/step - loss: 1.7816 - accuracy: 0.3769 - val_loss: 1.6861 - val_accuracy: 0.4234\n",
            "Epoch 5/20\n",
            "210/210 [==============================] - 2s 7ms/step - loss: 1.6816 - accuracy: 0.4191 - val_loss: 1.6809 - val_accuracy: 0.4167\n",
            "Epoch 6/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.6013 - accuracy: 0.4557 - val_loss: 1.6123 - val_accuracy: 0.4463\n",
            "Epoch 7/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.5486 - accuracy: 0.4741 - val_loss: 1.4943 - val_accuracy: 0.4945\n",
            "Epoch 8/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.4944 - accuracy: 0.4900 - val_loss: 1.4180 - val_accuracy: 0.5302\n",
            "Epoch 9/20\n",
            "210/210 [==============================] - 1s 3ms/step - loss: 1.4533 - accuracy: 0.5097 - val_loss: 1.4289 - val_accuracy: 0.5109\n",
            "Epoch 10/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.4236 - accuracy: 0.5224 - val_loss: 1.4046 - val_accuracy: 0.5299\n",
            "Epoch 11/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.3846 - accuracy: 0.5390 - val_loss: 1.3999 - val_accuracy: 0.5385\n",
            "Epoch 12/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.3550 - accuracy: 0.5544 - val_loss: 1.3439 - val_accuracy: 0.5561\n",
            "Epoch 13/20\n",
            "210/210 [==============================] - 1s 3ms/step - loss: 1.3261 - accuracy: 0.5672 - val_loss: 1.3253 - val_accuracy: 0.5621\n",
            "Epoch 14/20\n",
            "210/210 [==============================] - 1s 6ms/step - loss: 1.3056 - accuracy: 0.5751 - val_loss: 1.3046 - val_accuracy: 0.5814\n",
            "Epoch 15/20\n",
            "210/210 [==============================] - 1s 6ms/step - loss: 1.2777 - accuracy: 0.5877 - val_loss: 1.3446 - val_accuracy: 0.5656\n",
            "Epoch 16/20\n",
            "210/210 [==============================] - 1s 6ms/step - loss: 1.2573 - accuracy: 0.5939 - val_loss: 1.2569 - val_accuracy: 0.5935\n",
            "Epoch 17/20\n",
            "210/210 [==============================] - 1s 6ms/step - loss: 1.2374 - accuracy: 0.6032 - val_loss: 1.2718 - val_accuracy: 0.6065\n",
            "Epoch 18/20\n",
            "210/210 [==============================] - 1s 5ms/step - loss: 1.2230 - accuracy: 0.6077 - val_loss: 1.2205 - val_accuracy: 0.6079\n",
            "Epoch 19/20\n",
            "210/210 [==============================] - 1s 5ms/step - loss: 1.2053 - accuracy: 0.6125 - val_loss: 1.2568 - val_accuracy: 0.6001\n",
            "Epoch 20/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.1939 - accuracy: 0.6186 - val_loss: 1.2194 - val_accuracy: 0.6040\n"
          ]
        }
      ],
      "source": [
        "# Import necessary libraries\n",
        "from tensorflow.keras import Sequential\n",
        "from tensorflow.keras.layers import Dense\n",
        "from tensorflow.keras.optimizers import Adam\n",
        "\n",
        "# Set seed for reproducibility\n",
        "np.random.seed(0)\n",
        "tf.random.set_seed(0)\n",
        "\n",
        "# Define the model building function\n",
        "def ann_model():\n",
        "  model=Sequential([\n",
        "      Dense(64, activation='relu',input_shape=(1024,)),\n",
        "      Dense(32, activation='relu'),\n",
        "      Dense(10, activation='softmax'),\n",
        "  ])\n",
        "\n",
        "  model.compile(loss='categorical_crossentropy',\n",
        "                metrics=['accuracy'])\n",
        "  return model\n",
        "\n",
        "# Call the function and store the model\n",
        "model=ann_model()\n",
        "\n",
        "# Print the model summary\n",
        "model.summary()\n",
        "\n",
        "# Fit model on the train data\n",
        "X_train_flat = X_train.reshape(X_train.shape[0], -1)\n",
        "X_test_flat = X_test.reshape(X_test.shape[0], -1)\n",
        "history = model.fit(X_train_flat, y_train_encoded, validation_split=0.2, batch_size=128, verbose=1, epochs=20)\n",
        "\n",
        "# Extract the training and validation accuracy\n",
        "train_acc = history.history['accuracy']\n",
        "val_acc = history.history['val_accuracy']\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MeF8XSWz0XIU"
      },
      "source": [
        "### **Plot the Training and Validation Accuracies and write down your Observations.**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "lt77zgGMP4yw",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "outputId": "13941d8f-3755-470e-efec-ed3d686b4981"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "# Create an epochs range\n",
        "epochs_range = range(1, len(train_acc) + 1)\n",
        "\n",
        "# Plot training and validation accuracies\n",
        "\n",
        "plt.title('Training and Validation Accuracies')\n",
        "plt.plot(epochs_range, train_acc, label='Training Accuracy')\n",
        "plt.plot(epochs_range, val_acc, label='Validation Accuracy')\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pGBbQpLONX7k"
      },
      "source": [
        "**Observations:**\n",
        "\n",
        "1. The training accuracy generally increases, showing that the model is learning from the training data.\n",
        "2. A small gap indicates good generalization, while a large gap might indicate overfitting.\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "I_ih3wEU9wIk"
      },
      "outputs": [],
      "source": [
        "# Clear the previous model's history from the Keras backend. And fix the seed again after clearing the backend\n",
        "tf.keras.backend.clear_session()\n",
        "np.random.seed(0)\n",
        "tf.random.set_seed(0)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lT6o3TIKuCtk"
      },
      "source": [
        "### **Second Model Architecture**\n",
        "- Write a function that returns a sequential model with the following architecture:\n",
        " - First hidden layer with **256 nodes and the relu activation** and the **input shape = (1024, )**\n",
        " - Second hidden layer with **128 nodes and the relu activation**\n",
        " - Add the **Dropout layer with the rate equal to 0.2**\n",
        " - Third hidden layer with **64 nodes and the relu activation**\n",
        " - Fourth hidden layer with **64 nodes and the relu activation**\n",
        " - Fifth hidden layer with **32 nodes and the relu activation**\n",
        " - Add the **BatchNormalization layer**\n",
        " - Output layer with **activation as 'softmax' and number of nodes equal to the number of classes, i.e., 10**\n",
        " -Compile the model with the **loss equal to categorical_crossentropy, optimizer equal to Adam(learning_rate = 0.0005), and metric equal to 'accuracy'**. Do not fit the model here, just return the compiled model.\n",
        "- Call the nn_model_2 function and store the model in a new variable.\n",
        "- Print the summary of the model.\n",
        "- Fit on the train data with a **validation split of 0.2, batch size = 128, verbose = 1, and epochs = 30**. Store the model building history to use later for visualization."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f-ZjNBmH0XIV"
      },
      "source": [
        "### **Build and train the new ANN model as per the above mentioned architecture**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "EEPYLFIPnSDP",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "866d8dec-cf71-4dae-f5b4-9db50834f4e8"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense (Dense)               (None, 256)               262400    \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 128)               32896     \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 128)               0         \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 64)                8256      \n",
            "                                                                 \n",
            " dense_3 (Dense)             (None, 64)                4160      \n",
            "                                                                 \n",
            " dense_4 (Dense)             (None, 32)                2080      \n",
            "                                                                 \n",
            " batch_normalization (BatchN  (None, 32)               128       \n",
            " ormalization)                                                   \n",
            "                                                                 \n",
            " dense_5 (Dense)             (None, 10)                330       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 310,250\n",
            "Trainable params: 310,186\n",
            "Non-trainable params: 64\n",
            "_________________________________________________________________\n",
            "Epoch 1/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.1763 - accuracy: 0.6234 - val_loss: 1.1574 - val_accuracy: 0.6329\n",
            "Epoch 2/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.1664 - accuracy: 0.6277 - val_loss: 1.2052 - val_accuracy: 0.6317\n",
            "Epoch 3/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.1517 - accuracy: 0.6342 - val_loss: 1.1989 - val_accuracy: 0.6249\n",
            "Epoch 4/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.1419 - accuracy: 0.6376 - val_loss: 1.1499 - val_accuracy: 0.6415\n",
            "Epoch 5/20\n",
            "210/210 [==============================] - 1s 3ms/step - loss: 1.1315 - accuracy: 0.6393 - val_loss: 1.1866 - val_accuracy: 0.6167\n",
            "Epoch 6/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.1163 - accuracy: 0.6471 - val_loss: 1.1740 - val_accuracy: 0.6344\n",
            "Epoch 7/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.1098 - accuracy: 0.6490 - val_loss: 1.1180 - val_accuracy: 0.6600\n",
            "Epoch 8/20\n",
            "210/210 [==============================] - 1s 3ms/step - loss: 1.1027 - accuracy: 0.6491 - val_loss: 1.1368 - val_accuracy: 0.6427\n",
            "Epoch 9/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.0934 - accuracy: 0.6537 - val_loss: 1.0634 - val_accuracy: 0.6759\n",
            "Epoch 10/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.0883 - accuracy: 0.6567 - val_loss: 1.0846 - val_accuracy: 0.6662\n",
            "Epoch 11/20\n",
            "210/210 [==============================] - 1s 5ms/step - loss: 1.0773 - accuracy: 0.6611 - val_loss: 1.0796 - val_accuracy: 0.6682\n",
            "Epoch 12/20\n",
            "210/210 [==============================] - 1s 6ms/step - loss: 1.0718 - accuracy: 0.6610 - val_loss: 1.0816 - val_accuracy: 0.6647\n",
            "Epoch 13/20\n",
            "210/210 [==============================] - 1s 6ms/step - loss: 1.0621 - accuracy: 0.6688 - val_loss: 1.0629 - val_accuracy: 0.6735\n",
            "Epoch 14/20\n",
            "210/210 [==============================] - 1s 6ms/step - loss: 1.0592 - accuracy: 0.6667 - val_loss: 1.1243 - val_accuracy: 0.6463\n",
            "Epoch 15/20\n",
            "210/210 [==============================] - 1s 6ms/step - loss: 1.0498 - accuracy: 0.6710 - val_loss: 1.2092 - val_accuracy: 0.6220\n",
            "Epoch 16/20\n",
            "210/210 [==============================] - 1s 5ms/step - loss: 1.0446 - accuracy: 0.6739 - val_loss: 1.1246 - val_accuracy: 0.6499\n",
            "Epoch 17/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.0377 - accuracy: 0.6757 - val_loss: 1.1386 - val_accuracy: 0.6490\n",
            "Epoch 18/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.0293 - accuracy: 0.6765 - val_loss: 1.0711 - val_accuracy: 0.6729\n",
            "Epoch 19/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.0221 - accuracy: 0.6821 - val_loss: 1.1842 - val_accuracy: 0.6369\n",
            "Epoch 20/20\n",
            "210/210 [==============================] - 1s 4ms/step - loss: 1.0265 - accuracy: 0.6791 - val_loss: 1.1229 - val_accuracy: 0.6515\n"
          ]
        }
      ],
      "source": [
        "# Import new library\n",
        "from tensorflow.keras.layers import Dense, Dropout, BatchNormalization\n",
        "\n",
        "# Define the model building function\n",
        "from threading import activeCount\n",
        "def ann_model_2():\n",
        "  model=Sequential([\n",
        "      Dense(256, activation='relu', input_shape=(1024,)),\n",
        "      Dense(128, activation='relu'),\n",
        "      Dropout(0.2),\n",
        "      Dense(64, activation='relu'),\n",
        "      Dense(64, activation='relu'),\n",
        "      Dense(32, activation='relu'), BatchNormalization(),Dense(10,activation='softmax')\n",
        "  ])\n",
        "  model.compile(loss='categorical_crossentropy',\n",
        "              optimizer=Adam(learning_rate=0.0005),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "  return model\n",
        "\n",
        "# Call the function and store the model\n",
        "model_2=ann_model_2()\n",
        "\n",
        "# Print the model smmary\n",
        "model_2.summary()\n",
        "\n",
        "# Fit the model on the train data\n",
        "X_train_flat = X_train.reshape(X_train.shape[0], -1)\n",
        "X_test_flat = X_test.reshape(X_test.shape[0], -1)\n",
        "history_2= model.fit(X_train_flat, y_train_encoded, validation_split=0.2, batch_size=128, verbose=1, epochs=20)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZJYsvjmw0XIX"
      },
      "source": [
        "### **Plot the Training and Validation Accuracies and write down your Observations.**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "01ig6BrF1KVy",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "outputId": "8c854571-a622-49e5-c7a5-4c583fa58acc"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "# Extract training and validation\n",
        "train_acc = history_2.history['accuracy']\n",
        "val_acc = history_2.history['val_accuracy']\n",
        "\n",
        "# Create an epochs range\n",
        "epochs_range = range(1, len(train_acc) + 1)\n",
        "\n",
        "\n",
        "plt.plot(epochs_range, train_acc, label='Training Accuracy')\n",
        "plt.plot(epochs_range, val_acc, label='Validation Accuracy')\n",
        "plt.title('Training and Validation Accuracies for Model 2')\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VPW1LlD61RDn"
      },
      "source": [
        "**Observations:**<br>\n",
        "1. Both the training and validation accuracies are increasing over the epochs.\n",
        "2. The training accuracy is generally higher than the validation accuracy.\n",
        "3. Both the training and validation accuracies seem to be converging as the number of epochs increases.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8kuXx9Bvu00f"
      },
      "source": [
        "## **Predictions on the test data**\n",
        "\n",
        "- Make predictions on the test set using the second model.\n",
        "- Print the obtained results using the classification report and the confusion matrix.\n",
        "- Final observations on the obtained results."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "xbWMEtTj5Ad0"
      },
      "outputs": [],
      "source": [
        "# Import new library\n",
        "from sklearn.metrics import classification_report, confusion_matrix\n",
        "\n",
        "# Make predictions on the test set\n",
        "y_prob=model_2.predict(X_test_flat)\n",
        "\n",
        "# Convert probabilities to class labels\n",
        "y_pred = np.argmax(y_prob, axis=1)\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "NByu7uAQ8x9P"
      },
      "outputs": [],
      "source": [
        "# Convert one-hot encoded y_test_encoded to class labels\n",
        "y_true = np.argmax(y_test_encoded, axis=1)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1_SIoopr0XIg"
      },
      "source": [
        "### **Print the classification report and the confusion matrix for the test predictions. Write your observations on the final results.**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "xRddeJ-3EHT1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f5a6d200-9d8b-42b8-fcf5-282db5bcee5a"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Classification:               precision    recall  f1-score   support\n",
            "\n",
            "           0       0.00      0.00      0.00      1814\n",
            "           1       0.00      0.00      0.00      1828\n",
            "           2       0.10      0.91      0.18      1803\n",
            "           3       0.12      0.12      0.12      1719\n",
            "           4       0.50      0.00      0.00      1812\n",
            "           5       0.00      0.00      0.00      1768\n",
            "           6       0.00      0.00      0.00      1832\n",
            "           7       0.00      0.00      0.00      1808\n",
            "           8       0.00      0.00      0.00      1812\n",
            "           9       0.00      0.00      0.00      1804\n",
            "\n",
            "    accuracy                           0.10     18000\n",
            "   macro avg       0.07      0.10      0.03     18000\n",
            "weighted avg       0.07      0.10      0.03     18000\n",
            "\n"
          ]
        }
      ],
      "source": [
        "# Print classification report\n",
        "print(\"Classification:\",classification_report(y_true, y_pred))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xkR4JioMsuIV"
      },
      "source": [
        "## **Using Convolutional Neural Networks**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YN2YgkGL_6xQ"
      },
      "source": [
        "### **Load the dataset again and split the data into the train and the test dataset.**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "mqM204HbAjP2",
        "scrolled": true,
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6fb65d99-1b89-4ee0-afe0-3c0efb2252fa"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Keys in the .h5 file: ['X_test', 'X_train', 'X_val', 'y_test', 'y_train', 'y_val']\n"
          ]
        }
      ],
      "source": [
        "# Open the HDF5 file\n",
        "df='/content/drive/MyDrive/04-Projects/0.MIT PE/01. Deep learning/SVHN_single_grey1.h5'\n",
        "with h5py.File(df, 'r') as file:\n",
        " keys = list(file.keys())\n",
        " print(\"Keys in the .h5 file:\", keys)\n",
        "\n",
        "# Load data into the train and the test dataset\n",
        "with h5py.File(df, 'r') as file:\n",
        "  X_train = file['X_train'][:]\n",
        "  y_train = file['y_train'][:]\n",
        "  X_test = file['X_test'][:]\n",
        "  y_test = file['y_test'][:]\n",
        "  X_val = file['X_val'][:]\n",
        "  y_val = file['y_val'][:]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "gTLJZWjPAjQB",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "78fb6cb8-9796-42aa-edfd-40d8802d679f"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Number of images in the training dataset: 42000\n",
            "Number of images in the testing dataset: 18000\n"
          ]
        }
      ],
      "source": [
        "# Check the number of images in the training and the testing dataset\n",
        "num_images_train = X_train.shape[0]\n",
        "num_images_test = X_test.shape[0]\n",
        "print(\"Number of images in the training dataset:\", num_images_train)\n",
        "print(\"Number of images in the testing dataset:\", num_images_test)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OJndFfEVAjQG"
      },
      "source": [
        "## **Data preparation**\n",
        "\n",
        "- Print the shape and the array of pixels for the first image in the training dataset.\n",
        "- Reshape the train and the test dataset because we always have to give a 4D array as input to CNNs.\n",
        "- Normalize the train and the test dataset by dividing by 255.\n",
        "- Print the new shapes of the train and the test dataset.\n",
        "- One-hot encode the target variable."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "W4uXqKz1AjQG",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ff298b1a-19b9-453a-e415-176eb9efbe96"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Shape of the first image: (32, 32)\n",
            "Pixel array of the first image:\n",
            " [[ 33.0704  30.2601  26.852  ...  71.4471  58.2204  42.9939]\n",
            " [ 25.2283  25.5533  29.9765 ... 113.0209 103.3639  84.2949]\n",
            " [ 26.2775  22.6137  40.4763 ... 113.3028 121.775  115.4228]\n",
            " ...\n",
            " [ 28.5502  36.212   45.0801 ...  24.1359  25.0927  26.0603]\n",
            " [ 38.4352  26.4733  23.2717 ...  28.1094  29.4683  30.0661]\n",
            " [ 50.2984  26.0773  24.0389 ...  49.6682  50.853   53.0377]]\n"
          ]
        }
      ],
      "source": [
        "# Print the shape of the first image\n",
        "print(\"Shape of the first image:\", X_train[0].shape)\n",
        "\n",
        "# Print the array of pixels for the first image\n",
        "print(\"Pixel array of the first image:\\n\", X_train[0])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "D9YPwf9ysqWU",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "07cc401d-bc55-4b45-f75d-a5820ac09b40"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "New shape of X_train: (42000, 32, 32, 1)\n",
            "New shape of X_test: (18000, 32, 32, 1)\n"
          ]
        }
      ],
      "source": [
        "# Reshape the datasets to include the channel dimension\n",
        "X_train_reshaped = X_train.reshape(X_train.shape[0], X_train.shape[1], X_train.shape[2], 1)\n",
        "X_test_reshaped = X_test.reshape(X_test.shape[0], X_test.shape[1], X_test.shape[2], 1)\n",
        "\n",
        "# Print the new shapes\n",
        "print(\"New shape of X_train:\", X_train_reshaped.shape)\n",
        "print(\"New shape of X_test:\", X_test_reshaped.shape)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "eOGLAn40AjQG"
      },
      "outputs": [],
      "source": [
        "# Normalize inputs from 0-255 to 0-1\n",
        "X_train_normalized = X_train_reshaped / 255.0\n",
        "X_test_normalized = X_test_reshaped / 255.0"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "5qf8S5NQAjQG",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5e202b67-4bc7-4bcc-da9d-497799d0c3fa"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "New shape of normalized X_train: (42000, 32, 32, 1)\n",
            "New shape of normalized X_test: (18000, 32, 32, 1)\n"
          ]
        }
      ],
      "source": [
        "# Print New shape of Training and Test\n",
        "print(\"New shape of normalized X_train:\", X_train_normalized.shape)\n",
        "print(\"New shape of normalized X_test:\", X_test_normalized.shape)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "10QaOV-xR7Jn"
      },
      "source": [
        "### **One-hot encode the labels in the target variable y_train and y_test.**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "3KHWFWKMAjQH",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "1f022fc5-7bd9-43e0-b1c6-1d9f71b0628e"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "One-hot encoded y_train: (42000, 10)\n",
            "One-hot encoded y_test: (18000, 10)\n"
          ]
        }
      ],
      "source": [
        "y_train_encoded = to_categorical(y_train)\n",
        "y_test_encoded = to_categorical(y_test)\n",
        "\n",
        "# Print the shapes of the one-hot encoded target variables\n",
        "print(\"One-hot encoded y_train:\", y_train_encoded.shape)\n",
        "print(\"One-hot encoded y_test:\", y_test_encoded.shape)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H-8jYVQTAjQH"
      },
      "source": [
        "**Observation:** From the length of the one-hot encoded vectors (10), we can infer that there are 10 unique classes in the target variable, each represented by a unique position in the vector being set to 1, with the rest set to 0.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Vjx_LI4_AjQH"
      },
      "source": [
        "## **First CNN Model Building**\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ZY5pyF4-KDNt"
      },
      "outputs": [],
      "source": [
        "# Set the seed for numpy\n",
        "np.random.seed(0)\n",
        "\n",
        "# Set the seed for TensorFlow\n",
        "tf.random.set_seed(0)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1JUAczhzAjQH"
      },
      "source": [
        "### **Model Architecture**\n",
        "- **Write a function** that returns a sequential model with the following architecture:\n",
        " - First Convolutional layer with **16 filters and the kernel size of 3x3**. Use the **'same' padding** and provide the **input shape = (32, 32, 1)**\n",
        " - Add a **LeakyRelu layer** with the **slope equal to 0.1**\n",
        " - Second Convolutional layer with **32 filters and the kernel size of 3x3 with 'same' padding**\n",
        " - Another **LeakyRelu** with the **slope equal to 0.1**\n",
        " - A **max-pooling layer** with a **pool size of 2x2**\n",
        " - **Flatten** the output from the previous layer\n",
        " - Add a **dense layer with 32 nodes**\n",
        " - Add a **LeakyRelu layer with the slope equal to 0.1**\n",
        " - Add the final **output layer with nodes equal to the number of classes, i.e., 10** and **'softmax' as the activation function**\n",
        " - Compile the model with the **loss equal to categorical_crossentropy, optimizer equal to Adam(learning_rate = 0.001), and metric equal to 'accuracy'**. Do not fit the model here, just return the compiled model.\n",
        "- Call the function cnn_model_1 and store the output in a new variable.\n",
        "- Print the summary of the model.\n",
        "- Fit the model on the training data with a **validation split of 0.2, batch size = 32, verbose = 1, and epochs = 20**. Store the model building history to use later for visualization."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JWsAd45JKDNu"
      },
      "source": [
        "### **Build and train a CNN model as per the above mentioned architecture.**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "L1jOYANWAjQH",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "47adf6a7-9d56-4940-b71f-6f813b0694e2"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d (Conv2D)             (None, 32, 32, 16)        160       \n",
            "                                                                 \n",
            " leaky_re_lu (LeakyReLU)     (None, 32, 32, 16)        0         \n",
            "                                                                 \n",
            " conv2d_1 (Conv2D)           (None, 32, 32, 32)        4640      \n",
            "                                                                 \n",
            " leaky_re_lu_1 (LeakyReLU)   (None, 32, 32, 32)        0         \n",
            "                                                                 \n",
            " max_pooling2d (MaxPooling2D  (None, 16, 16, 32)       0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " flatten (Flatten)           (None, 8192)              0         \n",
            "                                                                 \n",
            " dense_6 (Dense)             (None, 32)                262176    \n",
            "                                                                 \n",
            " leaky_re_lu_2 (LeakyReLU)   (None, 32)                0         \n",
            "                                                                 \n",
            " dense_7 (Dense)             (None, 10)                330       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 267,306\n",
            "Trainable params: 267,306\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "Epoch 1/20\n",
            "1050/1050 [==============================] - 11s 5ms/step - loss: 1.0365 - accuracy: 0.6701 - val_loss: 0.5947 - val_accuracy: 0.8335\n",
            "Epoch 2/20\n",
            "1050/1050 [==============================] - 5s 4ms/step - loss: 0.5370 - accuracy: 0.8439 - val_loss: 0.5383 - val_accuracy: 0.8505\n",
            "Epoch 3/20\n",
            "1050/1050 [==============================] - 6s 6ms/step - loss: 0.4549 - accuracy: 0.8679 - val_loss: 0.5068 - val_accuracy: 0.8592\n",
            "Epoch 4/20\n",
            "1050/1050 [==============================] - 5s 5ms/step - loss: 0.3974 - accuracy: 0.8824 - val_loss: 0.4818 - val_accuracy: 0.8630\n",
            "Epoch 5/20\n",
            "1050/1050 [==============================] - 4s 4ms/step - loss: 0.3531 - accuracy: 0.8959 - val_loss: 0.4604 - val_accuracy: 0.8748\n",
            "Epoch 6/20\n",
            "1050/1050 [==============================] - 6s 5ms/step - loss: 0.3166 - accuracy: 0.9052 - val_loss: 0.4536 - val_accuracy: 0.8714\n",
            "Epoch 7/20\n",
            "1050/1050 [==============================] - 5s 5ms/step - loss: 0.2804 - accuracy: 0.9147 - val_loss: 0.4508 - val_accuracy: 0.8749\n",
            "Epoch 8/20\n",
            "1050/1050 [==============================] - 5s 5ms/step - loss: 0.2520 - accuracy: 0.9230 - val_loss: 0.4923 - val_accuracy: 0.8704\n",
            "Epoch 9/20\n",
            "1050/1050 [==============================] - 5s 5ms/step - loss: 0.2270 - accuracy: 0.9315 - val_loss: 0.4766 - val_accuracy: 0.8773\n",
            "Epoch 10/20\n",
            "1050/1050 [==============================] - 5s 5ms/step - loss: 0.2044 - accuracy: 0.9364 - val_loss: 0.4917 - val_accuracy: 0.8756\n",
            "Epoch 11/20\n",
            "1050/1050 [==============================] - 5s 4ms/step - loss: 0.1835 - accuracy: 0.9421 - val_loss: 0.5160 - val_accuracy: 0.8740\n",
            "Epoch 12/20\n",
            "1050/1050 [==============================] - 5s 5ms/step - loss: 0.1596 - accuracy: 0.9503 - val_loss: 0.5501 - val_accuracy: 0.8725\n",
            "Epoch 13/20\n",
            "1050/1050 [==============================] - 6s 6ms/step - loss: 0.1489 - accuracy: 0.9532 - val_loss: 0.6126 - val_accuracy: 0.8645\n",
            "Epoch 14/20\n",
            "1050/1050 [==============================] - 5s 4ms/step - loss: 0.1342 - accuracy: 0.9567 - val_loss: 0.6377 - val_accuracy: 0.8601\n",
            "Epoch 15/20\n",
            "1050/1050 [==============================] - 4s 4ms/step - loss: 0.1227 - accuracy: 0.9603 - val_loss: 0.6777 - val_accuracy: 0.8623\n",
            "Epoch 16/20\n",
            "1050/1050 [==============================] - 6s 6ms/step - loss: 0.1066 - accuracy: 0.9651 - val_loss: 0.6680 - val_accuracy: 0.8725\n",
            "Epoch 17/20\n",
            "1050/1050 [==============================] - 4s 4ms/step - loss: 0.1015 - accuracy: 0.9666 - val_loss: 0.7063 - val_accuracy: 0.8721\n",
            "Epoch 18/20\n",
            "1050/1050 [==============================] - 4s 4ms/step - loss: 0.0951 - accuracy: 0.9693 - val_loss: 0.7131 - val_accuracy: 0.8662\n",
            "Epoch 19/20\n",
            "1050/1050 [==============================] - 6s 5ms/step - loss: 0.0807 - accuracy: 0.9733 - val_loss: 0.8084 - val_accuracy: 0.8615\n",
            "Epoch 20/20\n",
            "1050/1050 [==============================] - 5s 5ms/step - loss: 0.0713 - accuracy: 0.9768 - val_loss: 0.8486 - val_accuracy: 0.8568\n"
          ]
        }
      ],
      "source": [
        "# Import necessary library\n",
        "from tensorflow.keras.layers import Conv2D, LeakyReLU, MaxPooling2D, Flatten, Dense\n",
        "\n",
        "def cnn_model_1():\n",
        "    model = Sequential()\n",
        "    # First Convolutional layer\n",
        "    model.add(Conv2D(16, (3, 3), padding='same', input_shape=(32, 32, 1)))\n",
        "    model.add(LeakyReLU(alpha=0.1))\n",
        "    # Second Convolutional layer\n",
        "    model.add(Conv2D(32, (3, 3), padding='same'))\n",
        "    model.add(LeakyReLU(alpha=0.1))\n",
        "    # Max-pooling layer\n",
        "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "    # Flatten layer\n",
        "    model.add(Flatten())\n",
        "    # Dense layer\n",
        "    model.add(Dense(32))\n",
        "    model.add(LeakyReLU(alpha=0.1))\n",
        "    # Output layer\n",
        "    model.add(Dense(10, activation='softmax'))\n",
        "\n",
        "    # Compile the model\n",
        "    model.compile(loss='categorical_crossentropy',\n",
        "                  optimizer=Adam(learning_rate=0.001),\n",
        "                  metrics=['accuracy'])\n",
        "\n",
        "    return model\n",
        "\n",
        "# Call the function and store the model\n",
        "model_cnn_1 = cnn_model_1()\n",
        "model_cnn_1.summary()\n",
        "history_cnn_1 = model_cnn_1.fit(X_train_normalized, y_train_encoded, validation_split=0.2, batch_size=32, verbose=1, epochs=20)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JPzfIf9kKDNw"
      },
      "source": [
        "### **Plot the Training and Validation Accuracies and Write your observations.**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "07oUCr1kAjQH",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 449
        },
        "outputId": "b21f6a29-435a-4191-ea0b-51e323f2be5d"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "plt.plot(history_cnn_1.history['accuracy'], label='Training Accuracy')\n",
        "plt.plot(history_cnn_1.history['val_accuracy'], label='Validation Accuracy')\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "P6zTLyp9AjQH"
      },
      "source": [
        "**Observations:**\n",
        "\n",
        "1. There are gap between training and validation accuracy, which might be overfitting\n",
        "2. The training accuracy increasing over the epochs and validation accuracy decrease.\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "HbKi93HTolGW"
      },
      "outputs": [],
      "source": [
        "# Clear the previous model's history from the Keras backend. And fix the seed again after clearing the backend\n",
        "tf.keras.backend.clear_session()\n",
        "np.random.seed(0)\n",
        "tf.random.set_seed(0)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ep19Jd8HAjQH"
      },
      "source": [
        "### **Second Model Architecture**\n",
        "\n",
        "- Write a function that returns a sequential model with the following architecture:\n",
        " - First Convolutional layer with **16 filters and the kernel size of 3x3**. Use the **'same' padding** and provide the **input shape = (32, 32, 1)**\n",
        " - Add a **LeakyRelu layer** with the **slope equal to 0.1**\n",
        " - Second Convolutional layer with **32 filters and the kernel size of 3x3 with 'same' padding**\n",
        " - Add **LeakyRelu** with the **slope equal to 0.1**\n",
        " - Add a **max-pooling layer** with a **pool size of 2x2**\n",
        " - Add a **BatchNormalization layer**\n",
        " - Third Convolutional layer with **32 filters and the kernel size of 3x3 with 'same' padding**\n",
        " - Add a **LeakyRelu layer with the slope equal to 0.1**\n",
        " - Fourth Convolutional layer **64 filters and the kernel size of 3x3 with 'same' padding**\n",
        " - Add a **LeakyRelu layer with the slope equal to 0.1**\n",
        " - Add a **max-pooling layer** with a **pool size of 2x2**\n",
        " - Add a **BatchNormalization layer**\n",
        " - **Flatten** the output from the previous layer\n",
        " - Add a **dense layer with 32 nodes**\n",
        " - Add a **LeakyRelu layer with the slope equal to 0.1**\n",
        " - Add a **dropout layer with the rate equal to 0.5**\n",
        " - Add the final **output layer with nodes equal to the number of classes, i.e., 10** and **'softmax' as the activation function**\n",
        " - Compile the model with the **categorical_crossentropy loss, adam optimizers (learning_rate = 0.001), and metric equal to 'accuracy'**. Do not fit the model here, just return the compiled model.\n",
        "- Call the function cnn_model_2 and store the model in a new variable.\n",
        "- Print the summary of the model.\n",
        "- Fit the model on the train data with a **validation split of 0.2, batch size = 128, verbose = 1, and epochs = 30**. Store the model building history to use later for visualization."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "y5IBLS1eKDNy"
      },
      "source": [
        "### **Build and train the second CNN model as per the above mentioned architecture.**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "wk9sl2UEAjQH",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "3763eccd-567d-418d-bfaa-36eaf0df85b1"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d (Conv2D)             (None, 32, 32, 16)        160       \n",
            "                                                                 \n",
            " leaky_re_lu (LeakyReLU)     (None, 32, 32, 16)        0         \n",
            "                                                                 \n",
            " conv2d_1 (Conv2D)           (None, 32, 32, 32)        4640      \n",
            "                                                                 \n",
            " leaky_re_lu_1 (LeakyReLU)   (None, 32, 32, 32)        0         \n",
            "                                                                 \n",
            " max_pooling2d (MaxPooling2D  (None, 16, 16, 32)       0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " batch_normalization (BatchN  (None, 16, 16, 32)       128       \n",
            " ormalization)                                                   \n",
            "                                                                 \n",
            " conv2d_2 (Conv2D)           (None, 16, 16, 32)        9248      \n",
            "                                                                 \n",
            " leaky_re_lu_2 (LeakyReLU)   (None, 16, 16, 32)        0         \n",
            "                                                                 \n",
            " conv2d_3 (Conv2D)           (None, 16, 16, 64)        18496     \n",
            "                                                                 \n",
            " leaky_re_lu_3 (LeakyReLU)   (None, 16, 16, 64)        0         \n",
            "                                                                 \n",
            " max_pooling2d_1 (MaxPooling  (None, 8, 8, 64)         0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " batch_normalization_1 (Batc  (None, 8, 8, 64)         256       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " flatten (Flatten)           (None, 4096)              0         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 32)                131104    \n",
            "                                                                 \n",
            " leaky_re_lu_4 (LeakyReLU)   (None, 32)                0         \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 32)                0         \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 10)                330       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 164,362\n",
            "Trainable params: 164,170\n",
            "Non-trainable params: 192\n",
            "_________________________________________________________________\n",
            "Epoch 1/30\n",
            "263/263 [==============================] - 4s 11ms/step - loss: 1.3620 - accuracy: 0.5360 - val_loss: 2.1147 - val_accuracy: 0.2839\n",
            "Epoch 2/30\n",
            "263/263 [==============================] - 2s 9ms/step - loss: 0.6643 - accuracy: 0.7946 - val_loss: 0.5542 - val_accuracy: 0.8299\n",
            "Epoch 3/30\n",
            "263/263 [==============================] - 3s 12ms/step - loss: 0.5462 - accuracy: 0.8331 - val_loss: 0.4469 - val_accuracy: 0.8705\n",
            "Epoch 4/30\n",
            "263/263 [==============================] - 3s 11ms/step - loss: 0.4814 - accuracy: 0.8562 - val_loss: 0.3921 - val_accuracy: 0.8906\n",
            "Epoch 5/30\n",
            "263/263 [==============================] - 3s 10ms/step - loss: 0.4350 - accuracy: 0.8667 - val_loss: 0.4372 - val_accuracy: 0.8817\n",
            "Epoch 6/30\n",
            "263/263 [==============================] - 3s 10ms/step - loss: 0.4074 - accuracy: 0.8775 - val_loss: 0.5143 - val_accuracy: 0.8518\n",
            "Epoch 7/30\n",
            "263/263 [==============================] - 3s 10ms/step - loss: 0.3703 - accuracy: 0.8858 - val_loss: 0.4467 - val_accuracy: 0.8695\n",
            "Epoch 8/30\n",
            "263/263 [==============================] - 3s 10ms/step - loss: 0.3466 - accuracy: 0.8927 - val_loss: 0.3669 - val_accuracy: 0.9015\n",
            "Epoch 9/30\n",
            "263/263 [==============================] - 3s 11ms/step - loss: 0.3261 - accuracy: 0.8989 - val_loss: 0.3830 - val_accuracy: 0.8982\n",
            "Epoch 10/30\n",
            "263/263 [==============================] - 3s 12ms/step - loss: 0.3087 - accuracy: 0.9063 - val_loss: 0.3368 - val_accuracy: 0.9113\n",
            "Epoch 11/30\n",
            "263/263 [==============================] - 3s 10ms/step - loss: 0.2867 - accuracy: 0.9102 - val_loss: 0.4193 - val_accuracy: 0.8967\n",
            "Epoch 12/30\n",
            "263/263 [==============================] - 3s 10ms/step - loss: 0.2711 - accuracy: 0.9144 - val_loss: 0.4381 - val_accuracy: 0.8887\n",
            "Epoch 13/30\n",
            "263/263 [==============================] - 2s 9ms/step - loss: 0.2633 - accuracy: 0.9179 - val_loss: 0.3969 - val_accuracy: 0.8977\n",
            "Epoch 14/30\n",
            "263/263 [==============================] - 3s 9ms/step - loss: 0.2543 - accuracy: 0.9207 - val_loss: 0.3792 - val_accuracy: 0.8967\n",
            "Epoch 15/30\n",
            "263/263 [==============================] - 4s 14ms/step - loss: 0.2408 - accuracy: 0.9221 - val_loss: 0.3407 - val_accuracy: 0.9121\n",
            "Epoch 16/30\n",
            "263/263 [==============================] - 3s 11ms/step - loss: 0.2316 - accuracy: 0.9257 - val_loss: 0.3763 - val_accuracy: 0.9031\n",
            "Epoch 17/30\n",
            "263/263 [==============================] - 3s 10ms/step - loss: 0.2218 - accuracy: 0.9304 - val_loss: 0.3626 - val_accuracy: 0.9130\n",
            "Epoch 18/30\n",
            "263/263 [==============================] - 3s 12ms/step - loss: 0.2072 - accuracy: 0.9311 - val_loss: 0.4024 - val_accuracy: 0.9099\n",
            "Epoch 19/30\n",
            "263/263 [==============================] - 3s 11ms/step - loss: 0.2036 - accuracy: 0.9343 - val_loss: 0.3914 - val_accuracy: 0.9051\n",
            "Epoch 20/30\n",
            "263/263 [==============================] - 3s 10ms/step - loss: 0.1983 - accuracy: 0.9349 - val_loss: 0.4622 - val_accuracy: 0.8882\n",
            "Epoch 21/30\n",
            "263/263 [==============================] - 3s 11ms/step - loss: 0.1924 - accuracy: 0.9364 - val_loss: 0.3839 - val_accuracy: 0.9158\n",
            "Epoch 22/30\n",
            "263/263 [==============================] - 3s 11ms/step - loss: 0.1789 - accuracy: 0.9407 - val_loss: 0.3737 - val_accuracy: 0.9113\n",
            "Epoch 23/30\n",
            "263/263 [==============================] - 3s 10ms/step - loss: 0.1778 - accuracy: 0.9404 - val_loss: 0.4041 - val_accuracy: 0.9121\n",
            "Epoch 24/30\n",
            "263/263 [==============================] - 3s 10ms/step - loss: 0.1751 - accuracy: 0.9424 - val_loss: 0.4179 - val_accuracy: 0.9107\n",
            "Epoch 25/30\n",
            "263/263 [==============================] - 3s 9ms/step - loss: 0.1719 - accuracy: 0.9423 - val_loss: 0.4457 - val_accuracy: 0.9075\n",
            "Epoch 26/30\n",
            "263/263 [==============================] - 3s 10ms/step - loss: 0.1610 - accuracy: 0.9459 - val_loss: 0.3993 - val_accuracy: 0.9148\n",
            "Epoch 27/30\n",
            "263/263 [==============================] - 3s 12ms/step - loss: 0.1538 - accuracy: 0.9487 - val_loss: 0.5745 - val_accuracy: 0.8925\n",
            "Epoch 28/30\n",
            "263/263 [==============================] - 3s 11ms/step - loss: 0.1508 - accuracy: 0.9509 - val_loss: 0.4062 - val_accuracy: 0.9187\n",
            "Epoch 29/30\n",
            "263/263 [==============================] - 3s 10ms/step - loss: 0.1509 - accuracy: 0.9482 - val_loss: 0.4452 - val_accuracy: 0.9089\n",
            "Epoch 30/30\n",
            "263/263 [==============================] - 3s 10ms/step - loss: 0.1448 - accuracy: 0.9512 - val_loss: 0.4451 - val_accuracy: 0.9083\n"
          ]
        }
      ],
      "source": [
        "from tensorflow.keras.layers import Conv2D, LeakyReLU, MaxPooling2D, BatchNormalization, Flatten, Dense, Dropout\n",
        "\n",
        "def cnn_model_2():\n",
        "    model = Sequential()\n",
        "\n",
        "    # First Convolutional layer\n",
        "    model.add(Conv2D(16, (3, 3), padding='same', input_shape=(32, 32, 1)))\n",
        "    model.add(LeakyReLU(alpha=0.1))\n",
        "\n",
        "    # Second Convolutional layer\n",
        "    model.add(Conv2D(32,(3,3),padding='same'))\n",
        "    model.add(LeakyReLU(alpha=0.1))\n",
        "\n",
        "    # Max-pooling layer\n",
        "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "\n",
        "    # BatchNormalization layer\n",
        "    model.add(BatchNormalization())\n",
        "\n",
        "    # Third Convolutional layer\n",
        "    model.add(Conv2D(32, (3, 3), padding='same'))\n",
        "    model.add(LeakyReLU(alpha=0.1))\n",
        "\n",
        "    # Fourth Convolutional layer\n",
        "    model.add(Conv2D(64, (3, 3), padding='same'))\n",
        "    model.add(LeakyReLU(alpha=0.1))\n",
        "\n",
        "    # Max-pooling layer and BatchNormalization layer\n",
        "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "    model.add(BatchNormalization())\n",
        "\n",
        "    # Flatten layer\n",
        "    model.add(Flatten())\n",
        "\n",
        "    # Dense layer\n",
        "    model.add(Dense(32))\n",
        "    model.add(LeakyReLU(alpha=0.1))\n",
        "\n",
        "    # Dropout layer\n",
        "    model.add(Dropout(0.5))\n",
        "\n",
        "    # Output layer\n",
        "    model.add(Dense(10, activation='softmax'))\n",
        "\n",
        "    # Compile the model\n",
        "    model.compile(loss='categorical_crossentropy',\n",
        "                  optimizer=Adam(learning_rate=0.001),\n",
        "                  metrics=['accuracy'])\n",
        "\n",
        "    return model\n",
        "\n",
        "# Call the function and store\n",
        "model_cnn_2 = cnn_model_2()\n",
        "model_cnn_2.summary()\n",
        "\n",
        "#Fit the model on the training data\n",
        "history_cnn_2 = model_cnn_2.fit(X_train_normalized, y_train_encoded, validation_split=0.2, batch_size=128, verbose=1, epochs=30)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PyhUtMy3KDN1"
      },
      "source": [
        "### **Plot the Training and Validation accuracies and write your observations.**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "YVQu7uWiAjQH",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "outputId": "c27568b4-6540-40a9-ff9f-62cd530cae5b"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjcAAAHHCAYAAABDUnkqAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAAB/YElEQVR4nO3dd1hT1x8G8DcJJBD2kiWCIq6qaFGpWlfF4qy7bnH/6uqwtta6ba2tXVZttcPR1jrqtkOtYrV1Vavi1rpRlgzZECC5vz8uCUSGgBmA7+cxD8nJvbknl4v55pzvOUciCIIAIiIiompCau4KEBERERkSgxsiIiKqVhjcEBERUbXC4IaIiIiqFQY3REREVK0wuCEiIqJqhcENERERVSsMboiIiKhaYXBDRERE1QqDm6fIqFGj4OfnV6F958+fD4lEYtgKVTJ37tyBRCLBunXrTH5siUSC+fPn6x6vW7cOEokEd+7ceey+fn5+GDVqlEHr8yTXCj05c12L169fx4svvggHBwdIJBLs3LnTpMcnfY/+v1BW5vy/rLJgcFMJSCSSMt0OHTpk7qo+9V599VVIJBLcuHGjxG1mzZoFiUSC8+fPm7Bm5RcdHY358+cjIiLC3FUp1pUrVyCRSGBlZYXk5GRzV+epEBYWhgsXLmDRokX48ccf0aJFC6MfMzU1FQsWLEBgYCBsbW1hbW2Nxo0bY8aMGYiOjtZtN2rUKEgkEjRt2hTFrRokkUgwZcoU3WPtB7xEIsG2bduKbK/9wpaQkFBq/bRfNCQSCY4cOVLkeUEQ4OPjA4lEgp49e5bnrZvd1atX8fbbb6NZs2aws7ODp6cnevTogX///dfcVXtiFuauAAE//vij3uMffvgB+/fvL1LesGHDJzrOt99+C41GU6F9Z8+ejXfeeeeJjl8dDBs2DMuXL8eGDRswd+7cYrfZuHEjmjRpgqZNm1b4OCNGjMDgwYOhUCgq/BqPEx0djQULFsDPzw/NmjXTe+5JrhVDWb9+PTw8PPDw4UNs3boV48aNM2t9TMnX1xdZWVmwtLQ02TGzsrJw/PhxzJo1Sy9IMKZbt24hJCQEkZGRGDhwICZMmAC5XI7z589j9erV2LFjB/777z+9fS5cuIDt27ejf//+ZT7OwoUL0a9fvydqfbayssKGDRvw/PPP65UfPnwY9+/fN+rfqrF89913WL16Nfr3749JkyYhJSUFX3/9NZ577jns3bsXISEh5q5ihTG4qQSGDx+u9/jEiRPYv39/kfJHZWZmQqlUlvk4T/IfpYWFBSwseLkEBwejbt262LhxY7HBzfHjx3H79m18+OGHT3QcmUwGmUz2RK/xJEz5oVocQRCwYcMGDB06FLdv38ZPP/1UaYObjIwM2NjYGPQ1tS1WphQfHw8AcHR0NNhrlnZu8vLy0K9fP8TFxeHQoUNFgoZFixbho48+0iuztraGj49PuYKVZs2aISIiAjt27EC/fv0q/F66d++OLVu2YNmyZXr/F27YsAFBQUGPbQGqjIYMGYL58+fD1tZWVzZmzBg0bNgQ8+fPr9LBDbulqoiOHTuicePGOH36NNq3bw+lUol3330XALBr1y706NEDXl5eUCgU8Pf3x3vvvQe1Wq33Go/mUWibbT/55BN888038Pf3h0KhQMuWLXHq1Cm9fYvLudE2A+/cuRONGzeGQqHAM888g7179xap/6FDh9CiRQtYWVnB398fX3/9dZnzeP7++28MHDgQtWrVgkKhgI+PD9544w1kZWUVeX+2traIiopCnz59YGtrCzc3N0yfPr3IuUhOTsaoUaPg4OAAR0dHhIWFlbnrY9iwYbh69SrOnDlT5LkNGzZAIpFgyJAhyMnJwdy5cxEUFAQHBwfY2NigXbt2+PPPPx97jOJybgRBwPvvv4+aNWtCqVSiU6dOuHTpUpF9k5KSMH36dDRp0gS2trawt7dHt27dcO7cOd02hw4dQsuWLQEAo0eP1jW7a/voi8u5ycjIwJtvvgkfHx8oFArUr18fn3zySZEugvJcFyU5evQo7ty5g8GDB2Pw4MH466+/cP/+/SLbaTQafPHFF2jSpAmsrKzg5uaGrl27FmlWX79+PVq1agWlUgknJye0b98ef/zxh16di8tteDSfSft7OXz4MCZNmoQaNWqgZs2aAIC7d+9i0qRJqF+/PqytreHi4oKBAwcWmzeVnJyMN954A35+flAoFKhZsyZGjhyp+4AsKWfi6tWrGDBgAJydnWFlZYUWLVpg9+7detvk5uZiwYIFCAgIgJWVFVxcXPD8889j//79JZ7v+fPnw9fXFwDw1ltvQSKR6P3+z549i27dusHe3h62trbo3LkzTpw4ofcapZ2b4mzbtg3nzp3DrFmzigQ2AGBvb49FixbplUmlUsyePRvnz5/Hjh07SnztwgYPHox69eph4cKFxXZnldWQIUOQmJiodx5zcnKwdetWDB06tNh9yvo3o1Kp8MYbb8DNzQ12dnZ46aWXir3eASAqKgpjxoyBu7u77m9rzZo1FXpPQUFBeoENALi4uKBdu3a4cuVKhV6zsuBX8SokMTER3bp1w+DBgzF8+HC4u7sDEP9TsbW1xbRp02Bra4uDBw9i7ty5SE1Nxccff/zY192wYQPS0tLwv//9DxKJBEuWLEG/fv1w69atx36DP3LkCLZv345JkybBzs4Oy5YtQ//+/REZGQkXFxcA4n+MXbt2haenJxYsWAC1Wo2FCxfCzc2tTO97y5YtyMzMxMSJE+Hi4oKTJ09i+fLluH//PrZs2aK3rVqtRmhoKIKDg/HJJ5/gwIED+PTTT+Hv74+JEycCEIOE3r1748iRI3jllVfQsGFD7NixA2FhYWWqz7Bhw7BgwQJs2LABzz77rN6xf/75Z7Rr1w61atVCQkICvvvuOwwZMgTjx49HWloaVq9ejdDQUJw8ebJIV9DjzJ07F++//z66d++O7t2748yZM3jxxReRk5Ojt92tW7ewc+dODBw4ELVr10ZcXBy+/vprdOjQAZcvX4aXlxcaNmyIhQsXYu7cuZgwYQLatWsHAGjTpk2xxxYEAS+99BL+/PNPjB07Fs2aNcO+ffvw1ltvISoqCp9//rne9mW5Lkrz008/wd/fHy1btkTjxo2hVCqxceNGvPXWW3rbjR07FuvWrUO3bt0wbtw45OXl4e+//8aJEyd0+SILFizA/Pnz0aZNGyxcuBByuRz//PMPDh48iBdffLHM57+wSZMmwc3NDXPnzkVGRgYA4NSpUzh27BgGDx6MmjVr4s6dO1i5ciU6duyIy5cv61pZ09PTdR8eY8aMwbPPPouEhATs3r0b9+/fh6ura7HHvHTpEtq2bQtvb2+88847sLGxwc8//4w+ffpg27Zt6Nu3LwAxUFm8eDHGjRuHVq1aITU1Ff/++y/OnDmDLl26FPva/fr1g6OjI9544w0MGTIE3bt3133oXbp0Ce3atYO9vT3efvttWFpa4uuvv0bHjh1x+PBhBAcHP/bcFEcblI0YMaIcZx4YOnQo3nvvPSxcuBB9+/Z97BckmUyG2bNnY+TIkU/UeuPn54fWrVtj48aN6NatGwBgz549SElJweDBg7Fs2TK97cvzNzNu3DisX78eQ4cORZs2bXDw4EH06NGjSB3i4uLw3HPP6b5AuLm5Yc+ePRg7dixSU1Px+uuvV+i9PSo2NrbE67DKEKjSmTx5svDor6ZDhw4CAGHVqlVFts/MzCxS9r///U9QKpVCdna2riwsLEzw9fXVPb59+7YAQHBxcRGSkpJ05bt27RIACL/88ouubN68eUXqBECQy+XCjRs3dGXnzp0TAAjLly/XlfXq1UtQKpVCVFSUruz69euChYVFkdcsTnHvb/HixYJEIhHu3r2r9/4ACAsXLtTbtnnz5kJQUJDu8c6dOwUAwpIlS3RleXl5Qrt27QQAwtq1ax9bp5YtWwo1a9YU1Gq1rmzv3r0CAOHrr7/WvaZKpdLb7+HDh4K7u7swZswYvXIAwrx583SP165dKwAQbt++LQiCIDx48ECQy+VCjx49BI1Go9vu3XffFQAIYWFhurLs7Gy9egmC+LtWKBR65+bUqVMlvt9HrxXtOXv//ff1thswYIAgkUj0roGyXhclycnJEVxcXIRZs2bpyoYOHSoEBgbqbXfw4EEBgPDqq68WeQ3tObp+/boglUqFvn37Fjknhc/jo+dfy9fXV+/can8vzz//vJCXl6e3bXHX6fHjxwUAwg8//KArmzt3rgBA2L59e4n11v5tFv7ddO7cWWjSpIne37RGoxHatGkjBAQE6MoCAwOFHj16FHntx9Ee8+OPP9Yr79OnjyCXy4WbN2/qyqKjowU7Ozuhffv2urLSzk1xmjdvLjg4OJS5fmFhYYKNjY0gCILw/fffFzmHAITJkycX+37y8vKEgIAAITAwUHeOtf+nxcfHl3pc7fs6deqUsGLFCsHOzk73ux44cKDQqVMnQRDEa6XweS/r30xERIQAQJg0aZLedkOHDi1yXY4dO1bw9PQUEhIS9LYdPHiw4ODgoKtXcddPWf3111+CRCIR5syZU+59KxN2S1UhCoUCo0ePLlJubW2tu5+WloaEhAS0a9cOmZmZuHr16mNfd9CgQXByctI91n6Lv3Xr1mP3DQkJgb+/v+5x06ZNYW9vr9tXrVbjwIED6NOnD7y8vHTb1a1bV/ft53EKv7+MjAwkJCSgTZs2EAQBZ8+eLbL9K6+8ove4Xbt2eu/l999/h4WFha4lBxC/3U2dOrVM9QHEPKn79+/jr7/+0pVt2LABcrkcAwcO1L2mXC4HIHafJCUlIS8vDy1atCi2S6s0Bw4cQE5ODqZOnar3TbW4b2oKhQJSqfinrVarkZiYCFtbW9SvX7/cx9X6/fffIZPJ8Oqrr+qVv/nmmxAEAXv27NErf9x1UZo9e/YgMTERQ4YM0ZUNGTIE586d0+uG27ZtGyQSCebNm1fkNbTnaOfOndBoNJg7d67unDy6TUWMHz++SE5U4es0NzcXiYmJqFu3LhwdHfXO+7Zt2xAYGKhraSlLnZKSknDw4EG8/PLLur/xhIQEJCYmIjQ0FNevX0dUVBQAMWfm0qVLuH79eoXfn5ZarcYff/yBPn36oE6dOrpyT09PDB06FEeOHEFqaqrePsWdm+KkpqbCzs6uQvUaNmwYAgICytzVpG29OXfu3BMNb3/55ZeRlZWFX3/9FWlpafj1119L7JIq69/M77//DgBFtnv0b1sQBGzbtg29evWCIAi6ayAhIQGhoaFISUmp8N+31oMHDzB06FDUrl0bb7/99hO9lrkxuKlCvL29dR+WhV26dAl9+/aFg4MD7O3t4ebmpktGTklJeezr1qpVS++xNtB5+PBhuffV7q/d98GDB8jKykLdunWLbFdcWXEiIyMxatQoODs76/JoOnToAKDo+9PmXZRUH0DMjfD09CzS11y/fv0y1QcQ+/FlMhk2bNgAAMjOzsaOHTvQrVs3vUDx+++/R9OmTXW5D25ubvjtt9/K9Hsp7O7duwCAgIAAvXI3Nze94wFiIPX5558jICAACoUCrq6ucHNzw/nz58t93MLH9/LyKvJhpB3Bp62f1uOui9KsX78etWvXhkKhwI0bN3Djxg34+/tDqVTip59+0m138+ZNeHl5wdnZucTXunnzJqRSKRo1avTY45ZH7dq1i5RlZWVh7ty5uvwK7XlPTk7WO+83b95E48aNy3W8GzduQBAEzJkzB25ubno3bXD34MEDAOLIoOTkZNSrVw9NmjTBW2+9VeFpCeLj45GZmVns30bDhg2h0Whw7949vfLizk1x7O3tkZaWVqF6aYOViIiIMgcrw4YNQ926dZ8o98bNzQ0hISHYsGEDtm/fDrVajQEDBhS7bVn/Zu7evQupVKr3ZQAo+v9RfHw8kpOT8c033xS5BrRferXXQEVkZGSgZ8+eSEtLw65du4r8/1jVMOemCin8zVArOTkZHTp0gL29PRYuXAh/f39YWVnhzJkzmDFjRpmG85b0Laus34gqum9ZqNVqdOnSBUlJSZgxYwYaNGgAGxsbREVFYdSoUUXen6lGGNWoUQNdunTBtm3b8OWXX+KXX35BWloahg0bpttm/fr1GDVqFPr06YO33noLNWrUgEwmw+LFi3Hz5k2j1e2DDz7AnDlzMGbMGLz33ntwdnaGVCrF66+/brLh3RW9LlJTU/HLL78gOzu7SCAHiK1jixYtMtmEko8momsV97c4depUrF27Fq+//jpat26tmwhv8ODBT3zetftPnz4doaGhxW6j/bLQvn173Lx5E7t27cIff/yB7777Dp9//jlWrVplkhFnxZ2b4jRo0ABnz57FvXv34OPjU+7jDBs2TJd706dPn8durw2IRo0ahV27dpX7eFpDhw7F+PHjERsbi27duhl0dFlptNfA8OHDS8wPrOj0Ezk5OejXrx/Onz+Pffv2lTv4rowY3FRxhw4dQmJiIrZv34727dvrym/fvm3GWhWoUaMGrKysip30rrSJ8LQuXLiA//77D99//z1GjhypKy9t5Mfj+Pr6Ijw8HOnp6XrfTq5du1au1xk2bBj27t2LPXv2YMOGDbC3t0evXr10z2/duhV16tTB9u3b9T6Mi+tGKUudAXEG2cLdA/Hx8UVaQ7Zu3YpOnTph9erVeuXJycl6SYLlCRB8fX1x4MABpKWl6X0T1XZ7auv3pLZv347s7GysXLmySELjtWvXMHv2bBw9ehTPP/88/P39sW/fPiQlJZXYeuPv7w+NRoPLly+XmsDt5ORUZLRcTk4OYmJiylz3rVu3IiwsDJ9++qmuLDs7u8jr+vv74+LFi2V+XQC637mlpWWZhuc6Oztj9OjRGD16NNLT09G+fXvMnz+/3MGNm5sblEplsX8bV69ehVQqrVBgAgC9evXCxo0bsX79esycObPc+1ckWBk+fDjef/99LFiwAC+99FK5jwkAffv2xf/+9z+cOHECmzdvLnG7sv7N+Pr6QqPR4ObNm3qtNY+ec+1IKrVabdAh2hqNBiNHjkR4eDh+/vlnXat4VcduqSpO+w258DfinJwcfPXVV+aqkh6ZTIaQkBDs3LlTb7bRGzduFMnTKGl/QP/9CYKAL774osJ16t69O/Ly8rBy5UpdmVqtxvLly8v1On369IFSqcRXX32FPXv2oF+/fnpzkxRX93/++QfHjx8vd51DQkJgaWmJ5cuX673e0qVLi2wrk8mKtJBs2bJFl5OhpZ1/pCxD4Lt37w61Wo0VK1bolX/++eeQSCRlzp96nPXr16NOnTp45ZVXMGDAAL3b9OnTYWtrq+ua6t+/PwRBwIIFC4q8jvb99+nTB1KpFAsXLizSelL4HPn7++vlTwHAN998U2LLTXGKO+/Lly8v8hr9+/fHuXPnih3KXFLLVo0aNdCxY0d8/fXXxQZc2jlqAHFUZWG2traoW7cuVCpVmd+Llkwmw4svvohdu3bpDWmPi4vTTWhnb29f7tcFgAEDBqBJkyZYtGhRsX8TaWlpmDVrVqmvMXz4cNStW7fYa6A4hbuzHh1CX1a2trZYuXIl5s+fr/dl5lFl/ZvR/nx0tNWjf9symQz9+/fHtm3big2OC18D5TF16lRs3rwZX3311RPNA1TZsOWmimvTpg2cnJwQFhamWxrgxx9/NFi3kCHMnz8ff/zxB9q2bYuJEyfq/uAbN2782Kn/GzRoAH9/f0yfPh1RUVGwt7fHtm3bypS7UZJevXqhbdu2eOedd3Dnzh00atQI27dvL3c+iq2tLfr06aPLuyncJQUAPXv2xPbt29G3b1/06NEDt2/fxqpVq9CoUSOkp6eX61ja+XoWL16Mnj17onv37jh79iz27NlTpIWjZ8+eWLhwIUaPHo02bdrgwoUL+Omnn/RafADxA93R0RGrVq2CnZ0dbGxsEBwcXGzORK9evdCpUyfMmjULd+7cQWBgIP744w/s2rULr7/+epF8gYqIjo7Gn3/+WSSxUkuhUCA0NFQ3kVqnTp0wYsQILFu2DNevX0fXrl2h0Wjw999/o1OnTpgyZQrq1q2LWbNm4b333kO7du3Qr18/KBQKnDp1Cl5eXli8eDEAcSjuK6+8gv79+6NLly44d+4c9u3bV67hsD179sSPP/4IBwcHNGrUCMePH8eBAweKDH1/6623sHXrVgwcOBBjxoxBUFAQkpKSsHv3bqxatQqBgYHFvv6XX36J559/Hk2aNMH48eNRp04dxMXF4fjx47h//75uHqNGjRqhY8eOCAoKgrOzM/79919s3bq1wrMOv//++9i/fz+ef/55TJo0CRYWFvj666+hUqmwZMmSCr0mILZCbd++HSEhIWjfvj1efvlltG3bFpaWlrh06RI2bNgAJyenInPdFCaTyTBr1qxiB1qURNud9STLjpRl2oiy/s00a9YMQ4YMwVdffYWUlBS0adMG4eHhxbZsf/jhh/jzzz8RHByM8ePHo1GjRkhKSsKZM2dw4MABJCUllet9LF26FF999RVat24NpVKJ9evX6z3ft29fg09QaTImHJlFZVTSUPBnnnmm2O2PHj0qPPfcc4K1tbXg5eUlvP3228K+ffsEAMKff/6p266koeCPDv0UhKJDY0saCl546KXWo8NnBUEQwsPDhebNmwtyuVzw9/cXvvvuO+HNN98UrKysSjgLBS5fviyEhIQItra2gqurqzB+/Hjd0OLCQx0LDxUtrLi6JyYmCiNGjBDs7e0FBwcHYcSIEcLZs2fLPXzyt99+EwAInp6exQ41/uCDDwRfX19BoVAIzZs3F3799dcivwdBePxQcEEQBLVaLSxYsEDw9PQUrK2thY4dOwoXL14scr6zs7OFN998U7dd27ZthePHjwsdOnQQOnTooHfcXbt2CY0aNdINy9e+9+LqmJaWJrzxxhuCl5eXYGlpKQQEBAgff/yx3pBq7Xsp63VR2KeffioAEMLDw0vcZt26dQIAYdeuXYIgiMPtP/74Y6FBgwaCXC4X3NzchG7dugmnT5/W22/NmjVC8+bNBYVCITg5OQkdOnQQ9u/fr3terVYLM2bMEFxdXQWlUimEhoYKN27cKHEo+KlTp4rU7eHDh8Lo0aMFV1dXwdbWVggNDRWuXr1a7PtOTEwUpkyZInh7ewtyuVyoWbOmEBYWphviW9JQ3ps3bwojR44UPDw8BEtLS8Hb21vo2bOnsHXrVt0277//vtCqVSvB0dFRsLa2Fho0aCAsWrRIyMnJKfG8Fj5mcf8fnDlzRggNDRVsbW0FpVIpdOrUSTh27JjeNqWdm9I8fPhQmDt3rtCkSRNBqVQKVlZWQuPGjYWZM2cKMTExuu1K+vvOzc0V/P39Sx0K/ihtXVHOoeCleXQouCCU/W8mKytLePXVVwUXFxfBxsZG6NWrl3Dv3r1ipyiIi4sTJk+eLPj4+AiWlpaCh4eH0LlzZ+Gbb74p8t4f93+ZdvqMkm6F//+paiSCUIm+4tNTpU+fPgYbskpERKTFnBsyiUeXSrh+/Tp+//13dOzY0TwVIiKiaostN2QSnp6eGDVqFOrUqYO7d+9i5cqVUKlUOHv2bLFDfomIiCqKCcVkEl27dsXGjRsRGxsLhUKB1q1b44MPPmBgQ0REBseWGyIiIqpWmHNDRERE1QqDGyIiIqpWnrqcG41Gg+joaNjZ2ZlsfRoiIiJ6MoIgIC0tDV5eXpBKS2+beeqCm+jo6AqvhUJERETmde/ePdSsWbPUbZ664Ea7gNm9e/cqvCYKERERmVZqaip8fHz0FiItyVMX3Gi7ouzt7RncEBERVTFlSSlhQjERERFVKwxuiIiIqFphcENERETVCoMbIiIiqlYY3BAREVG1wuCGiIiIqhUGN0RERFStMLghIiKiaoXBDREREVUrDG6IiIioWmFwQ0RERNUKgxsiIiKqVhjcEBERkUHkqTWITcnGvaRMs9bjqVsVnIiIqCrJU2uQnJWL5MxcJGfmIDkzFw8zc5CSJf4Uy3ORnJWDhxm5SMnKRWpWLuysLOBiq4CrrRwutgq42Mrhlv/T1VYBFxsFXO3kcFbKYSErva1DladGfJoKD9JUeJCqQnxaNuJSVXiQlq0re5CmQmKGCoIAtPF3wYbxz5noDBXF4IaIiMiEBEFASlYuEtJViE/LQUK6CgnpKiSmF9yPT89BUoYKyZm5SMvOq9Bx0lR5iE7Jfux2EgngpJTDxSY/6LGVQy6TIj5dDFri0rKRnJlb5uPKpBJoBKFCdTYUBjdEREQGoNEISMzIQXRylnhLyUZ8mkoXsCSkq5CQloPEDBVy1eX/8LezsoCTUg5HpSUclXI4WlvCSWkJB6UcTkpLvXJ7a0ukZ+fpgqb4QsFTYkbB/aSMHGgEICkjB0kZObj+IL3E48tlUrjZKVDDXoEadgrUsLOCu734061QmbONHDKp5ElO5RNjcENERFQGGao8xKRkISo5GzH5AUxUcnZ+IJOFmJRs5ORpyvx6dlYWcLNVwNVW7B5y1d7XdSXJdcGKg7XlY7uOKkKtEfAwM0ev1SgxPQeqPI0YrOQHLzXsFHBUWkIiMW/QUlYMboiIqMpR5al1uSUphW+ZuUjJykNKlpiDkqHKgyCIXS8SSMSf+ffz/0EikeT/1H8MCZCalYvo5GxEp2SVqWtGKgFq2FnBy9EKno7WqGEnBiuFc11c7RRwsZHDylJm3JNUBjKpRBdQ1YeduatjMAxuiIjIILJz1bj/MBN3EzNxJzETkYkZuJuUiQepKkgk4gepVCKBVO++RLwvlUAmgVgmlUAmkUAqFR9n56r1A5isXGTnlr2FxJDsrCzg5WANL0creDlaw8vRGt75P70creBubwVLI7SwUPkwuCEiojJLzc5FZKI2gMkQ7yeJP2NSs2HKPFKJBLC3ErtsCt/sC923s7KARAIIAiAAgCBAQP7jwvchPobusQCNANgoLFAzP3jxdLSCvZWl6d4gVRiDGyKip0hOngaZOXlIV+UhM0eNdFUeMlR5yFCpkaHKy39OvJ+RU/BcVHIWIpMykZSRU+rr2yosUMtZCV8XJXxdbODrooSHgxUkADSCAI0GUAsCNBoxeCi4L0Ct+5m/bX6ZlaWs2ADGTmEBqZkTV6lyYnBDRFSJCIKAh5m5iE3JRlxaNh6kZiM2RZxPJDtXgzyNBnlqAblqDfI0+T/VAvI0GuTm/yz8vPa5nDwNMlRq5KifvDvH1VaeH8DY5AcxStRyFu+72MirTNIpVV8MboiITEAQBGTkqBGXml3opkJsSjYepBXcj09TGSQAeRy5hRS2Cgso5TLdTxuFRf59C9gqZFDqHsvgYW+FWvmtMbYKfnRQ5cYrlIioBNm5atxNzMSt+HTcSsjA3cQMZOVqkJOnRm5+60hOnga5arHVRHs/R61fpn1cnnwUFxs5athbwcNeAXd7cSiuUmEBC6kEljIpLGQSWEqlkEkl4n2ZVO85C6kUljIJLPLLFRZS2CgsYCO3gFIhY9IrVWsMbojoqSYIAmJTs3ErPgO34tNxMz4DtxMycCshHfcfZhk8QdZWYQH3/ICl4KbQu1/DzgpyCwYfRBXF4IaIqj2NRkBChgoxydm4k5ghBjIJYjBzOyEDmTnqEve1s7JAHTdb+LvawM9V7JKRW0ghl0lhaSG2lFjKpAVlMrHFRK/MQiyzkVvAhl06REbHvzIiqvLSVXn5s8VmIabQjLHiNPjZiE3JLjWPRSaVoJazEnVcbVDHzQZ13Gzz79vC1ZYJskRVDYMbIqoUBEFAVq4aGSo1MnPyCn7mqJGpyv+Zk4fUrFzEpGQjJiVbF9CUZWFBiQRwt7NCTSfrIgFMLWclu4GIqhEGN0RkUBrtWjUZOUhIUyEh/2diRsGigalZ4hwqmTnauVXUyMjJe6L8FnsrC91ssZ7a2WMdOHMs0dOIwQ0RlUlWjlrX1ROTkq1bYK/wz4T0HCRlqKB5wiRcG7k4DNlGLoNSbgEbRcFPG7kFPBwKpr73chDX8OHwZKJ8D64Chz4AXOoCDXoCXs3FpsunCP83ICLkqTV4kKbKz1URu3tiCq14HJOShYdlWDSwMEelJVxtxQUCXe0UcLURFw10sVXAwdpSDFT05lkRgxcrCxlnna1s4v8DVGmAiz9g7Wju2lBpboQDW0YBqlTx8d+fAvY1gQY9gIY9gVptAFn1/+iv/u+QiJCn1iAmJRuRSZmITBLXBYrKD2Cik7MQl6aCugzNLTZyGbwcreHhYAU3O/2Vjl1sFXDNv+9sI2cXUHUgCMChxcDhjwrKlC5ii4CzvxjsuPjnP64DyG3MV9fKIiMBiDwBRB4HHt4BAoeIQYUpnPoO+P1tQFADPsGAnQdw/QCQeh84+bV4s3YC6nUT6+T/AmBpbZq6mZhEEEy5zJn5paamwsHBASkpKbC3tzd3dYgMJl2Vh8jEzPwAJkMXxNxLysT9h1nIe0zwYiGVFHT3FO72yc9f8XSwhr2VhXlHDqXGALcOATUaAB6BgNQEAZRGLX5QXfkFuP03IFcCtu6Anaf44aG92XqIZUrn6tEFoM4DfnsDOPOD+NimBpDxoPR97LwKAh7n/KDHxR9wqAlo8oC8HECtAvJUgDo3/37hspxHfuY/L7MELJVi8CS3zf/5yM3SxjTXQ2GCADy8LQYzd4+JPxOvF90uaDQQ+oF47RiDRg3smwX8s1J8HDgE6PUFYKEAcrPEv5krvwLXfgeykgr2s1QCdTsDDXoB9UIrfatceT6/GdwQmdqDq8Cpb4FLOwHn2kDb14H63R/7H7O29eX+wyzce5iJ+0mZuJvfEhOZmInExyxoKJdJUdPZGr7OStRyVqKmk1IveHG1VUBWWbuDUmOAo0uBf9eKH3iAGGAEdAECQgH/ToDCznDHy1MBt/8CruwGrv4OZCaUfV+ZPD/4KRz05N/svQHXAPFnZQ6AcjKBraOB//YCEinQ41OgxRixayrpFpB4U7wl3QQSb4j3C39omosuAMoPgiyV4nVh6w7Ye+YHpJ75970A2xqAVFb211fnAXEXxWA38rgYzKTHFd3OrSHg2xqABPh3tVjmWh8YsBrwaGKQt6qjSgO2jRN/VwDwwhyg3ZvFX1/qPLHeV38Frv4GpNwreE5qAfi1E1t06vcQz1Elw+CmFAxuyCw0auDaHuDkN8Dtw0Wfd2sATdvX8aBWT9xPzcW9h5m4l5SF+/k/7z3MRExK9mO7jpyUlqjlrEQtFxtdEFPLRfzpYW9V9XJZ0mKBI0uB02uBvGyxzK0BkHwPyM0o2E5qCfi2Eb99BoQCrnXLf6ycDODGAbGF5r99BTkLAGDlKAag9buKH/ZpsQW3dO39GCAzsWzHsrQRWzRc64nBjmsA4BIgtnQY69t9WWUkAhteBqL+BSysgAFrxHyNx8lMyg988oOdxBv5wc9NICe9YDuZHJApxNYYC4X42EIhllnIi/kpF1t9cjJKuKUDqODHmERa0Apn75UfhBa+7yW2Vt3ND2bun9J/L9r34/UsUOs5oFZrwKeV2HqndfNPYMf/xCBIJge6vAcE/88wwW3KfWDDIDHgsrAC+qwEGvcr276CAMREiEHOlV+B+Cv6z7vWE7uspJbi70pqId5klvllFvrP6crzHzvWAlqOffL3WEiVCm6+/PJLfPzxx4iNjUVgYCCWL1+OVq1aFbttbm4uFi9ejO+//x5RUVGoX78+PvroI3Tt2rXMx2NwQyaVmSQ2659aDaREAgAEiRQpPiE44dgTithTCI7fBqWQCQC4p3HD1+qe2KLuABXkRV5OLpPC28kaNZ2sUdNJuxpzQRBjb2Vp0rdnNMUFNT7BQMeZQJ2OYrfF3aPAf38A1/eJH6qFOdcB6nUFAl4EfNuKH5LFyXooBjJXfhEDG+2xALHFpWFPoGEv8TVkZTi3eTnih1h6nBjsPBoEPbwrdmNoSpmXx8GnINjRBj6u9cQPXWO39jy8A6zvLwYm1k7AkM1AreAne01BEAMCWX6gYuj3IAji700b6BQOenIygewU8dynxuT/TmLE++lxYm5KeSkcxHOiDWa8ngUsrUrfJyMB2DW5oHUl4EWg91eArVv5j68VdRrYOER8HzY1gCEbgZotKv56iTfFv4Orv4pB3JOq2QoYt//JX6eQKhPcbN68GSNHjsSqVasQHByMpUuXYsuWLbh27Rpq1KhRZPsZM2Zg/fr1+Pbbb9GgQQPs27cP06ZNw7Fjx9C8efMyHZPBDZlEzDmxlebCVt0HZpaFA/6wCsUXKe1wK9dFt6kdMjFcdgBjLH6Hm0RsLUiWOuGo68u45z8E7jXcUNNJCR8nJWrYKape60t5pMUCR78A/l1TEGjUbAV0mgnU6VTyB2PCDTHI+W+fmPugKTSyS24rBkT1QsUPFUiAa7/l59D8pR9oOPoCjV4CGr4EeLcwTg6HOlcMIhKuAwn/iTka2vtZD0veT24HPNMH6PSu2LJgaDHngJ8Gih+WDj7A8G2AW33DH6ey0KiBjHggNVo/6NG7Hy2e91rPiTffNmKXU0WuC0EATn4L/DFb7Fq1qQH0XSXmvJTXpZ3AjleAvCygxjPA0E1iS4mhpMYA8VfFvw11rvhTkyt2a2lyC8rUuSU8zgMcfYDnJhquTqhCwU1wcDBatmyJFStWAAA0Gg18fHwwdepUvPPOO0W29/LywqxZszB58mRdWf/+/WFtbY3169eX6ZgMbkhPbhYAyeO/eZWFOhe4vAt5J76BRdQ/uuJLGj+sU7+I3eo2utYYJ6UlWvg5o6GHHWo6K1HTyRq17CTwvLUVsuMrdK08UDgArcYBwROf7FteZZcWlx/UrC5fUFOc7FQxgfL6PuD6/mJyIiTQ68ao0UhsnWnYC3BvbN5cmIzE/GDnv/yA57r4OOl2QSuDhTXQehLQ9jXAysEwx735J7B5BJCTJp6DYVsrZc5FtRB7Edg2VgweAKD1FKDzvJJbFwsTBODIZ0D4QvFxwItA/9WA1dPxWVYlgpucnBwolUps3boVffr00ZWHhYUhOTkZu3btKrKPi4sLlixZgrFjC/rxhg8fjiNHjuDOnTtlOi6DG9KJ2AD89iaQmykOb7X3EvvY7R+5actK+A8kPiYSSYe/hueNjbDPE3MucgUZ9mhaYV1eKM4IAfB0sEar2s7izc8Z/m62JbfAqHPFFp8jnwMJ18QyC2vg2ZFAm6niN6LqotigpqXY/eT/wpMHGhoNEHuuoPsq6gwAAfAOEoOZBr0qlp9jank5Yg5M+EIx9wMQr9kOM8SROGX5YCzJ+Z+BnZPEb9x+7YDBPxkuaKLi5WaJo5u0ycaegWKQ4hpQ8j55OcAvrwHnNoiPg18BXlz0VMxZo1Ulgpvo6Gh4e3vj2LFjaN26ta787bffxuHDh/HPP/8U2Wfo0KE4d+4cdu7cCX9/f4SHh6N3795Qq9VQqVTFHkelUuk9l5qaCh8fHwY3T7sLW4Ht4wGh5MUUi5DbQWPngTR5DcTBGXdyHCBJjkSHvKOQS8Rv1Q8ER2xQv4AjDr1Qt05dtKrtjJZ+zqjpZF3+IdQajTh08+9PgegzYpnUAmg6SBxh5VavfK9nKHk5Yk6G1EJs8bKwLvhZ1v9o0+KAY8vEXKS8LLHMu4XYUuPf2XitJxmJYguIbdFu7ypBEMRrYv+8giHHTrWBkHlAoz7lO2+CABxbDuyfIz5+pp/YTWKhMHi1qQRXfxNzcbIeiiO7ui0Bmg8v+nvMTAI2DxfzzCRScbtW481TZzOqtsFNfHw8xo8fj19++QUSiQT+/v4ICQnBmjVrkJWVVexx5s+fjwULFhQpZ3DzFLvyC/BzmPghFzQa6Dy3oN89NUq8nxqNvOQo5Dy8D1l6DBR5aaW+5GWLhrjiMwg2zfohqI4H3OwM+AEhCGI3y5HPxBwRAIBEHL3ToDtQu71h+9uLkxYLXP9DzGm5dajoiBEtqYU4asPCShxpYWFVEPhYKPJHX1iI3SCmDGqqG3UecPYH4M/FBXPPeLcAXnxPzAt5HI0G+GMWcOIr8fFzk4EX3zf9PDEk/n+zfQJw52/x8TN9gZ5LC+acSbgBbBgoJs3L7YCB64CAEDNV1ryqRHBTkW4prezsbCQmJsLLywvvvPMOfv31V1y6dKnYbdlyQ3r++wPYNFRsgg8cIo5YkEqRlp2LS9GpuBiVggv5t9sJGbqFHK2RDQ/JQ3hIktBQmYrGdpmoa5UKZztrOLYJg63fE4xSKI/7/wJ/fyYmxBbm5CcGObU7iD+ftGVCoxFbi/7bJ3bnxJzTf15hLwYieSr9EUbl4d1C7H6qy6CmwlTpYuvLseUFQ+Pr9wBC5pfcspenEpNRL20XH7/4vtjdSeajUYvds38uEhNzHXyAft+K/09tHgFkJwMOtYChmwH3RuaurdlUieAGEBOKW7VqheXLlwMQE4pr1aqFKVOmFJtQ/Kjc3Fw0bNgQL7/8Mj744IMyHZM5N0+xW4eAn14G1Cok1+6J3f4LcPp+Gi7cT8GthIxid/FysEJjbwc08XZA45riT1fbStBs/+CK2LV2+y9xSOijQ1rdGuYHO+0Bv7bisN7HyUoGbh4UW2iu7y86cZ3XswUjjjybFXzL12jE0R+5WWKgo/uZLbbOaH/mFdrGNUAMxBjUGEZaLHDoQ3HaAUENSGRijlbHmYCde8F22SnApmFiK4HUEujzFdD0ZfPVm/TdPw1sGyOOppNIxZsmT8xDG7yh6nanGkiVCW42b96MsLAwfP3112jVqhWWLl2Kn3/+GVevXoW7uztGjhwJb29vLF68GADwzz//ICoqCs2aNUNUVBTmz5+P27dv48yZM3B0dCzTMRncFCPmvDi3gVt9MYmzLB+EVUhSRg5u/rsfgYfHQK7JxgGhBV5RvYq8R5ZW83a0RmNvezGQyQ9oXCpDIPM4qjRxkrHbh8VgJ/YC9Cc1k4gJi3XyW3VqtRZncBUEIP5a/hDqP4B7J/SHRSvsxZl/A0LFmYCf8v9Yq4T4a8CBBQUte5Y2YqtMm6nipITrBwAPLonD4wetF3+/VLlkpwK/vwWc3yQ+fqafGIRW0zWgyqPKBDcAsGLFCt0kfs2aNcOyZcsQHCxOGtWxY0f4+flh3bp1AIDDhw9j4sSJuHXrFmxtbdG9e3d8+OGH8PIq+5wPDG4Kyc0G/loiTpam/eYvkYmTpQV0Eb+huz9Tpb5dqzUC/otLw+m7D3Em8iHORibDPvEc1ssXw06ShUPqQEzInQZrayWa13LEs7WcEOjjiMZe9lUjkCmLzCTxm/ntv8Rbwn/6z0stAe9nxRyj5Ej951zrib/3eqFiEFSWieuo8rlzVEwUjjotPrapIf4uU6PE+8O3igEvVV7X9opLWjQdzFyofFUquDE1Bjf57p0Us/S1H3y1O4jzgWjnXtCy8yoIdOp0MOz6PQaQmp2LM3cfirfIZETcS0a6qqD1oZHkDjbK34eDJBM3bJ7FuXbfILCOB+q4ljIUu7pJjSkIdG4f1l9PRqYA/J4v6G5yrm2+epJhCQJweafYkvPwtljmUlecnM/Jz5w1I6oQBjeleOqDm5wMIPw94J9VAATxW1yPT8VZWQFxevgb+8Wci1uHC0a0AAXr9wS8KN5cA0zeqpOr1uBsZDKO3EjAkevxOHc/pch6SzZyGZrXckIX10QMuTIJctVDwOc58T91ha1J61vpCILYnx95XOx+rN1e7KKi6isvBzjzvfjFpeO7gI3L4/chqoQY3JTiqQ5ubv4J/PJqQVdEs2HiSInCi7wVlpsN3D0iBjr/7Sv49qfl6FsQ6NRuZ5Q+YUEQcONBOv6+noAjNxLwz61EZOToJ8/6uigR5OuEZ2s5IcjXCfXc7SB7eAtY201sjfJqDozcxYnJiIiqMAY3pXgqg5usZHE9k7M/io8dfIBeS4G65ZwrIfFm/kiaP4A7R8TFC7XsvYEROwyyFs2D1GyxZeZGAo7eSEBcqv4Ejc42crSt64p2dV3RNsAV3o6PBFUP74qBTWqUOJV82C8lB3BERFQlMLgpxVMX3Fz9Hfhtmpg8CgAtx4uzmT5p7kxOhpjDcf0P8RjpsYDSVWwh8WhcrpfKUOXh5O2k/NaZePwXpz9BnMJCila1nfF8XVc8H+CKhh72JefLpESJgU3yXcC1PjDqt+q9JhMR0VOCwU0pnprgJiMB2PM2cHGb+NjZH+i9omyzl5ZXZhLwYx9xojdrJ7EFx6v0VdqzctQ4cCUOuyKi8dd/8chRFyyDIJEAjb0cxNaZAFcE+TrBylL2+HqkxQHruotLAzjVBkbv4eJ/RETVRHk+v5+eFbeeFoIgBjR73gYyE8VJoNq8CnR8x3jzJCidgZG7gfX9xcX9vu8tDjX1aaW3Wa5agyPXE7D7XDT2XYpFZqHcmZpO1mgX4Iq2dV3Rxt8VzjblXAgwIxH4obcY2DjUEruiGNgQET2VGNxUJ6nRwK9vAP/tFR+7NwZeWi7OaWJs1o7AyJ3iDMCRx4Af+wJDf4amVhucjnyIXRFR+O18DB5m5up2qelkjd7NvPBSoDfquduWf2FJrayHYstR/BXAzhMI21W9Vs4mIqJyYXBTXZz9Cdj7jjgLqdQS6PC2uHK0RTlbQJ6Ewg4YvhXCxiGQ3D6M3B/6YrpsJnalFaxx42orR8+mXnipmRea+zhWPKDRUqWJs67Gngds3MQWJOc6T/hGiIioKmNwUx1c2gHsmiTe924h5tbUaGjyakQmZmL3uWjsSZiCN9UpeAERWKJehFzFm7B+pjt6N/NCG38XWMgMNNvmvZPAb2+KgY21k5jMXNJigURE9NRgcFPVJdwAduWv6Bv8ChD6ASAtQ/KtgcSlZuO38zHYfS4aEfeSdeVTLd7ETzar0CzjKL6UfgpJk8ZAPQNN954SBRyYB1zYIj62cgRG7BSXiiAioqceg5uqLDcL+HkkkJMG+LYFXlxkksAmIV2FPRdj8cu5aJy6kwTteDupBGhb1xUvBXohtLEH7C27AdsnQHJpO/BzGNDvG6DJgIofOCcTOLZMXAsrLwuABGg+HHhhjv7Kx0RE9FRjcFOV/T5dXOHXxg0YsAaQGe/XmZyZg70XY/Hr+Rgcu5mAwisePFvLEb0CvdCjqSdq2Fnp79j/O8BCAZzbCGwfL07812xo+Q6uHQG2fx6Qel8sq9Ua6Poh4NXsid4XERFVPwxuqqqzPwFn14tDvfuvBuw8DH6I1Oxc/HEpDr+ej8aR6wnIKxTRNK3pgJ5NPdGjqVfRGYILk8qA3l8BMrm4vs3OiUCeCmgxumyViDoN7J0J3PtHfOzgA3RZCDzTt0qtVk5ERKbD4KYqirskJtIC4kJ4dToY7KUzVHk4cCUOv5yLKTK5XkNPe/Rs6omeTT3h61KOxRalUqDXF4CFFXDya+DX18UA57lXSt4nNQYIXwic2yA+tlQCz08D2kwx3nw9RERULTC4qWpUaWKeTV4W4N8ZaPfmE79knlqD/Zfj8Mv5aBy8+gDZuQUBTd0atujV1As9Az3h7/YEK2pLJEC3j8QuqmPLgL0zgLxs4PnX9bfLzQaOrwD+/gzIzRDLmg4Wl4yw96r48YmI6KnB4KYqEQRg96viLLz23kC/b8VWkSdw/2Empm48i7ORyboyPxcleuYHNPXd7Z58LhotiUTsUrK0Bg5/JI54ylOJc/IAwOVdwP45BauW12wp5tXUbGGY4xMR0VOBwU1Vcuo74NJ2QGoBDFwH2Lg80cvtvxyH6VvOISUrF3ZWFhgaXAu9mnrhGS97wwU0j5JIgE7vii044QuBQx8AGQ+AB1eBu0fEbey8gC4LgMYDnjh4IyKipw+Dm6pCm1gLiK0fj6zbVB45eRp8tPcqVh+5DQAIrOmAFUOfhY+z0hA1LZt2b4o5OPveFYM2QHzc5lWxq0pejpweIiKiQhjcVAVZD4EtowBNLtCgJ/DcpAq/1L2kTEzZeBbn8ifcG/t8bczo2gByCzO0kLSeLAY0++cBAV3E1hrHWqavBxERVSsMbio7jQbYMVHMQ3HyA3p/WeEh0PsuxeKtLeeQmp0HeysLfDIwEC8+Y/gh5OXScizQYgyHdRMRkcEwuKnsji0D/tsDyBTAwO/F1bfLSZWnxuLfr2LdsTsAgOa1HLF8SHPUdDJhN1RpGNgQEZEBMbipzO4eE5NuAaBbxWbjjUzMxOQNZ3AhKgUAMKF9HbwVWh+Whlq8koiIqJJhcFNZpccDW8cAghpo8jIQVMYZfQv5/UIMZmw9jzRVHhyVlvjs5UC80IBrMBERUfXG4KYy0qiBbWOBtBjAtT7Q8/Nydd1k56rxwe9X8MPxuwCAIF8nLB/SHF6lLZNARERUTTC4qYwOLwFuHxaXHHj5B0BR9pmB7yRkYPKGM7gUnQoAeKWDP958sR67oYiI6KnB4KayuREuzt4LAD2XAjUalHnXX85FY+b2C0hX5cFJaYnPBjVDp/o1jFNPIiKiSorBTWWSEgVsHw9AAIJGAYGDyrSbRiNg/i+XdN1QLf2csGxIc3g6sBuKiIiePgxuKgt1rphAnJkIeDQBun5U5l2XHbyOH47fhUQCTOrojzdC6sGC3VBERPSUYnBTWRxbDtw7ASjsxTwbS6sy7fbHpVgsPXAdAPBR/6Z4uYWPMWtJRERU6fHrfWUReUL82WEG4FynTLvceJCGaT+fAwCMauPHwIaIiAgMbioPVZr408G7TJunZudiwg+nka7KQ3BtZ8zq0dCIlSMiIqo6GNxUFipx6DYU9o/dVKMR8PqmCNxKyIC3ozW+GvYsh3oTERHl4ydiZVGO4ObzA//h4NUHUFhI8fWIILjYKoxcOSIioqqDwU1loe2Wsio9uNl7MQbLD94AAHzYvwkaezsYu2ZERERVCoObykAQgGxty41diZtdiy1IIB77fG30bV7TFLUjIiKqUhjcVAa5WeICmUCJ3VIpmbmY8OO/yMxRo42/C2Z2K/vMxURERE8TBjeVgTbfRiIF5DZFnlZrBEzddBZ3EzPh7WiNFUOf5SR9REREJeAnZGWgzbdR2BW7+vcnf1zDX//Fw8pSim9GBsHZRm7iChIREVUdDG4qg+ySR0r9ej4aKw/dBAAsGRCIZ7yYQExERFQaBjeVQQnDwC9Hp+KtLecBAP9rXwcvBXqZumZERERVjtmDmy+//BJ+fn6wsrJCcHAwTp48Wer2S5cuRf369WFtbQ0fHx+88cYbyM7ONlFtjURVdKTUw4wc/G/9v8jKVaNdgCve7soEYiIiorIwa3CzefNmTJs2DfPmzcOZM2cQGBiI0NBQPHjwoNjtN2zYgHfeeQfz5s3DlStXsHr1amzevBnvvvuuiWtuYIVzbgDkqTWYuvEs7iVloZazEsuHNIdMWjQXh4iIiIoya3Dz2WefYfz48Rg9ejQaNWqEVatWQalUYs2aNcVuf+zYMbRt2xZDhw6Fn58fXnzxRQwZMuSxrT2VnjbnJn8CvyX7ruHIjQRYW8rwzcggOCqZQExERFRWZgtucnJycPr0aYSEhBRURipFSEgIjh8/Xuw+bdq0wenTp3XBzK1bt/D777+je/fuJR5HpVIhNTVV71bpFGq52RURhW/+ugUA+GRgIBp4PH45BiIiIipgYa4DJyQkQK1Ww93dXa/c3d0dV69eLXafoUOHIiEhAc8//zwEQUBeXh5eeeWVUrulFi9ejAULFhi07gaXn3MTn6vA21vFBOJJHf3Ro6mnOWtFRERUJZk9obg8Dh06hA8++ABfffUVzpw5g+3bt+O3337De++9V+I+M2fOREpKiu527949E9a4jPKDmx2X06DK06BjfTe8+WJ9M1eKiIioajJby42rqytkMhni4uL0yuPi4uDh4VHsPnPmzMGIESMwbtw4AECTJk2QkZGBCRMmYNasWZBKi8ZqCoUCCkUlXzU7P+cmKssSfi5KfDGICcREREQVZbaWG7lcjqCgIISHh+vKNBoNwsPD0bp162L3yczMLBLAyGQyAIAgCMarrLHl59ykC9Z4o0s9OCgtzVwhIiKiqstsLTcAMG3aNISFhaFFixZo1aoVli5dioyMDIwePRoAMHLkSHh7e2Px4sUAgF69euGzzz5D8+bNERwcjBs3bmDOnDno1auXLsipkvK7pdJgjVrOSjNXhoiIqGoza3AzaNAgxMfHY+7cuYiNjUWzZs2wd+9eXZJxZGSkXkvN7NmzIZFIMHv2bERFRcHNzQ29evXCokWLzPUWDELIToMEQBqU8HayNnd1iIiIqjSJUKX7c8ovNTUVDg4OSElJgb195RhmnfdxA1hkxKBv3gfYtnASpMy3ISIi0lOez+8qNVqq2soRc26Udo4MbIiIiJ4Qgxtz06hhkZsOALB3dDFzZYiIiKo+BjfmlpOuu+vo6GzGihAREVUPDG7MLX+OG5VgAXcXR/PWhYiIqBpgcGNu2jluYA0vR46UIiIielIMbsxNO8eNoERNBjdERERPjMGNmQn53VLpsOYcN0RERAbA4MbM0lOTAIgtNx4OVmauDRERUdXH4MbMUpLF4CbHwgYKiyq8hAQREVElweDGzNJTxOBGI7czc02IiIiqBwY3ZpaV9hAAILWqHEtBEBERVXUMbswsJzMFAGChdDBzTYiIiKoHBjdmps4SgxsrW0fzVoSIiKiaYHBjbirtoplceoGIiMgQGNyYmUX+iuB2jk5mrgkREVH1wODGjDJz8qDQZAIAnJy4IjgREZEhMLgxo+jkbNhBDG6Udmy5ISIiMgQGN2YUlZwFO0kWAEDCoeBEREQGweDGjKIeZulabqBgcENERGQIDG7MKDYpFVaSXPGBgjMUExERGQKDGzNKSkooeMCWGyIiIoNgcGNGyfmLZubJrAGZhZlrQ0REVD0wuDEjLppJRERkeAxuzEStEaDKSAbARTOJiIgMicGNmcSlZsNGEEdKyay5aCYREZGhMLgxk6jkLNhCO8cNu6WIiIgMhcGNmUQnZ8FOwjluiIiIDI3BjZncf5gFu/yWGwY3REREhsPgxkz0Wm6YUExERGQwDG7MpHDODWcnJiIiMhwGN2bCnBsiIiLjYHBjBoIgIOohW26IiIiMgcGNGaRk5SIjRw07SX5ww5wbIiIig2FwYwZRyWJQ4yjNFgvYckNERGQwDG7MIOqhNrjRdktxhmIiIiJDYXBjBtH5LTc2zLkhIiIyOAY3ZiB2SwmwFjLEAubcEBERGQyDGzOITs6GFXIgE9RiAVtuiIiIDIbBjRncTy609AIkgNzWrPUhIiKqTipFcPPll1/Cz88PVlZWCA4OxsmTJ0vctmPHjpBIJEVuPXr0MGGNn0zUw0cm8JNIzFshIiKiasTswc3mzZsxbdo0zJs3D2fOnEFgYCBCQ0Px4MGDYrffvn07YmJidLeLFy9CJpNh4MCBJq55xWTnqpGQruIEfkREREZi9uDms88+w/jx4zF69Gg0atQIq1atglKpxJo1a4rd3tnZGR4eHrrb/v37oVQqq0xwE5Mizm3jYpE/xw2TiYmIiAzKrMFNTk4OTp8+jZCQEF2ZVCpFSEgIjh8/XqbXWL16NQYPHgwbG5tin1epVEhNTdW7mZN2GHgtGyYTExERGYNZg5uEhASo1Wq4u7vrlbu7uyM2Nvax+588eRIXL17EuHHjStxm8eLFcHBw0N18fHyeuN5PQjuBX01lrljARTOJiIgMyuzdUk9i9erVaNKkCVq1alXiNjNnzkRKSorudu/ePRPWsCjt0gseVtrghi03REREhmRhzoO7urpCJpMhLi5OrzwuLg4eHh6l7puRkYFNmzZh4cKFpW6nUCigUCieuK6Gog1ualiqxALm3BARERmUWVtu5HI5goKCEB4erivTaDQIDw9H69atS913y5YtUKlUGD58uLGraVDanBsni/zghi03REREBmXWlhsAmDZtGsLCwtCiRQu0atUKS5cuRUZGBkaPHg0AGDlyJLy9vbF48WK9/VavXo0+ffrAxcXFHNWuMN2K4BIumklERGQMZg9uBg0ahPj4eMydOxexsbFo1qwZ9u7dq0syjoyMhFSq38B07do1HDlyBH/88Yc5qlxhGo2AmGRxCLithPPcEBERGYPZgxsAmDJlCqZMmVLsc4cOHSpSVr9+fQiCYORaGV5Cugo5ag2kEsBKw0UziYiIjKFKj5aqau5rR0rZW0GqShML2XJDRERkUAxuTEibTOzlaA2o8icT5Dw3REREBsXgxoS0E/h5O1kDbLkhIiIyCgY3JqTXcpOd33JjxdFSREREhsTgxoS0w8C9HRRADltuiIiIjIHBjQndf6hdNFNTUMicGyIiIoMqd3Dj5+eHhQsXIjIy0hj1qda03VI1bfLEAqklYFF5loYgIiKqDsod3Lz++uvYvn076tSpgy5dumDTpk1QqVTGqFu1kpadi9RsMajxUOSIhVb2gERixloRERFVPxUKbiIiInDy5Ek0bNgQU6dOhaenJ6ZMmYIzZ84Yo47VQnT+zMQO1pZQajLFQubbEBERGVyFc26effZZLFu2DNHR0Zg3bx6+++47tGzZEs2aNcOaNWuq5AzCxhSVLAY03pzjhoiIyKgqvPxCbm4uduzYgbVr12L//v147rnnMHbsWNy/fx/vvvsuDhw4gA0bNhiyrlVaVH7LjTiB3wOxkMENERGRwZU7uDlz5gzWrl2LjRs3QiqVYuTIkfj888/RoEED3TZ9+/ZFy5YtDVrRqk47gV/NwhP4cV0pIiIigyt3cNOyZUt06dIFK1euRJ8+fWBpaVlkm9q1a2Pw4MEGqWB1EaWbwM+qYAI/5twQEREZXLmDm1u3bsHX17fUbWxsbLB27doKV6o60g4D93ZUAgnaCfzYckNERGRo5U4ofvDgAf75558i5f/88w/+/fdfg1SqOtJfV4otN0RERMZS7uBm8uTJuHfvXpHyqKgoTJ482SCVqm5y1RrEpWkTiq24aCYREZERlTu4uXz5Mp599tki5c2bN8fly5cNUqnqJjYlG4IAyC2kcLVRANkp4hNMKCYiIjK4cgc3CoUCcXFxRcpjYmJgYVHhkeXVmi6Z2MEKUqmkUMsNgxsiIiJDK3dw8+KLL2LmzJlISUnRlSUnJ+Pdd99Fly5dDFq56kIv3wbgJH5ERERGVO6mlk8++QTt27eHr68vmjdvDgCIiIiAu7s7fvzxR4NXsDooaLnRBjfMuSEiIjKWcgc33t7eOH/+PH766SecO3cO1tbWGD16NIYMGVLsnDdUaBi4tuVGO88Nc26IiIgMrkJJMjY2NpgwYYKh61JtRenmuGHLDRERkbFVOAP48uXLiIyMRE5Ojl75Sy+99MSVqm70ght1LpAnPmbODRERkeFVaIbivn374sKFC5BIJLrVvyUSCQBArVYbtoZVnCAI+t1S2lYbgC03RERERlDu0VKvvfYaateujQcPHkCpVOLSpUv466+/0KJFCxw6dMgIVazakjJykJ2rAQB4OFgVzHFjqQRkzFEiIiIytHK33Bw/fhwHDx6Eq6srpFIppFIpnn/+eSxevBivvvoqzp49a4x6VlnaLqkadgooLGTMtyEiIjKycrfcqNVq2NmJH8yurq6Ijo4GAPj6+uLatWuGrV01oJ3jxsuRc9wQERGZQrlbbho3boxz586hdu3aCA4OxpIlSyCXy/HNN9+gTp06xqhjlRb16DBwttwQEREZVbmDm9mzZyMjIwMAsHDhQvTs2RPt2rWDi4sLNm/ebPAKVnXa4KamI+e4ISIiMoVyBzehoaG6+3Xr1sXVq1eRlJQEJycn3YgpKqAdKVW0W4otN0RERMZQrpyb3NxcWFhY4OLFi3rlzs7ODGxKUHQCP21w42CmGhEREVVv5QpuLC0tUatWLc5lUw7RydkACrfcMOeGiIjImMo9WmrWrFl49913kZSUZIz6VCuZOXlIyhBncOa6UkRERKZR7pybFStW4MaNG/Dy8oKvry9sbGz0nj9z5ozBKlfVafNtbBUWsLfKP9VsuSEiIjKqcgc3ffr0MUI1qqeo/C4pb0frgpwkznNDRERkVOUObubNm2eMelRL2gn8dF1SAFtuiIiIjKzcOTdUdgXDwK0KCplzQ0REZFTlbrmRSqWlDvvmSKoCBcPAlQWF7JYiIiIyqnIHNzt27NB7nJubi7Nnz+L777/HggULDFax6iCquJYbXbcUgxsiIiJjKHe3VO/evfVuAwYMwKJFi7BkyRLs3r273BX48ssv4efnBysrKwQHB+PkyZOlbp+cnIzJkyfD09MTCoUC9erVw++//17u45qCNuempjbnRhA4QzEREZGRGSzn5rnnnkN4eHi59tm8eTOmTZuGefPm4cyZMwgMDERoaCgePHhQ7PY5OTno0qUL7ty5g61bt+LatWv49ttv4e3tbYi3YFBqjYDY1Ecm8MvLBjR54n0GN0REREZR7m6p4mRlZWHZsmXlDjI+++wzjB8/HqNHjwYArFq1Cr/99hvWrFmDd955p8j2a9asQVJSEo4dOwZLS0sAgJ+f3xPX3xjiUrOh1giwkEpQwy6/W0qbTAwJILc1W92IiIiqs3IHN48ukCkIAtLS0qBUKrF+/foyv05OTg5Onz6NmTNn6sqkUilCQkJw/PjxYvfZvXs3WrdujcmTJ2PXrl1wc3PD0KFDMWPGDMhksvK+FaPS5tt4OlpBJtXOcVNoGLiUA9WIiIiModzBzeeff64X3EilUri5uSE4OBhOTk5lfp2EhASo1Wq4u7vrlbu7u+Pq1avF7nPr1i0cPHgQw4YNw++//44bN25g0qRJyM3NLXH+HZVKBZVKpXucmppa7HaGphsG7lB4jpsU8SeTiYmIiIym3MHNqFGjjFCNstFoNKhRowa++eYbyGQyBAUFISoqCh9//HGJwc3ixYvNMorrPifwIyIiMoty942sXbsWW7ZsKVK+ZcsWfP/992V+HVdXV8hkMsTFxemVx8XFwcPDo9h9PD09Ua9ePb0uqIYNGyI2NhY5OTnF7jNz5kykpKTobvfu3StzHZ9EtG6Om0LBDSfwIyIiMrpyBzeLFy+Gq6trkfIaNWrggw8+KPPryOVyBAUF6Y2w0mg0CA8PR+vWrYvdp23btrhx4wY0Go2u7L///oOnpyfkcnmx+ygUCtjb2+vdTCGquOCGLTdERERGV+7gJjIyErVr1y5S7uvri8jIyHK91rRp0/Dtt9/i+++/x5UrVzBx4kRkZGToRk+NHDlSL+F44sSJSEpKwmuvvYb//vsPv/32Gz744ANMnjy5vG/D6AqWXigc3HB2YiIiImMrd85NjRo1cP78+SJDsM+dOwcXF5dyvdagQYMQHx+PuXPnIjY2Fs2aNcPevXt1ScaRkZGQFhpV5OPjg3379uGNN95A06ZN4e3tjddeew0zZswo79swKkEQuGgmERGRmZQ7uBkyZAheffVV2NnZoX379gCAw4cP47XXXsPgwYPLXYEpU6ZgypQpxT536NChImWtW7fGiRMnyn0cU0rJykVGjrjGln7OTf5oKebcEBERGU25g5v33nsPd+7cQefOnWFhIe6u0WgwcuTIcuXcVGfafBsXGzmsLAvNv8N1pYiIiIyu3MGNXC7H5s2b8f777yMiIgLW1tZo0qQJfH19jVG/KqnYLimAOTdEREQmUOHlFwICAhAQEGDIulQbxU7gBzDnhoiIyATKPVqqf//++Oijj4qUL1myBAMHDjRIpao63TDwR1tuOM8NERGR0ZU7uPnrr7/QvXv3IuXdunXDX3/9ZZBKVXXRyY+sBq7FlhsiIiKjK3dwk56eXuyEeZaWliZbt6myu1/cBH4Ac26IiIhMoNzBTZMmTbB58+Yi5Zs2bUKjRo0MUqmqTpdQXGLLDYMbIiIiYyl3QvGcOXPQr18/3Lx5Ey+88AIAIDw8HBs2bMDWrVsNXsGqJjtXjYR0cRVyvZwbjaYguGHODRERkdGUO7jp1asXdu7ciQ8++ABbt26FtbU1AgMDcfDgQTg7OxujjlVKTIqYb2NtKYOT0rLgiZx0AIJ4nzk3RERERlOhoeA9evRAjx49AACpqanYuHEjpk+fjtOnT0OtVhu0glVNwZpSVpBIJAVPaPNtpJaAhZUZakZERPR0KHfOjdZff/2FsLAweHl54dNPP8ULL7xQ6ZdFMIWCCfyU+k8UHilVOOghIiIigypXy01sbCzWrVuH1atXIzU1FS+//DJUKhV27tzJZOJ8ujluHB9pneEcN0RERCZR5pabXr16oX79+jh//jyWLl2K6OhoLF++3Jh1q5KiShwGzjluiIiITKHMLTd79uzBq6++iokTJ3LZhVJou6WKTuCXvyK4wsHENSIiInq6lLnl5siRI0hLS0NQUBCCg4OxYsUKJCQkGLNuVVJ0CltuiIiIzKnMwc1zzz2Hb7/9FjExMfjf//6HTZs2wcvLCxqNBvv370daWpox61klaDQCYvKXXii6IjiDGyIiIlMo92gpGxsbjBkzBkeOHMGFCxfw5ptv4sMPP0SNGjXw0ksvGaOOVUZCugo5ag2kEsDdngnFRERE5lDhoeAAUL9+fSxZsgT379/Hxo0bDVWnKku7ppSHvRUsZY+cWrbcEBERmcQTBTdaMpkMffr0we7duw3xclVWwQR+1kWf5KKZREREJmGQ4IZEBRP4lRbcsOWGiIjImBjcGFBUaS03upwbDgUnIiIyJgY3BhRd0gR+AHNuiIiITITBjQHdL1O3FHNuiIiIjInBjQGx5YaIiMj8GNwYSFp2LlKz8wA8LueGLTdERETGxODGQKLzZyZ2sLaEreKRJbvUuUCe2KrDbikiIiLjYnBjIInpKlhIJaV3SQHsliIiIjKyMq8KTqVrU9cV197vhvT8rik92mRiC2tAZmnaihERET1l2HJjQDKpBA7KYoIX5tsQERGZDIMbU+BIKSIiIpNhcGMKnOOGiIjIZBjcmAJbboiIiEyGwY0pZKeIP5lzQ0REZHQMbkxB13LD4IaIiMjYGNyYAnNuiIiITIbBjSkw54aIiMhkGNyYAue5ISIiMhkGN6bAlhsiIiKTYXBjCsy5ISIiMhkGN6bA4IaIiMhkKkVw8+WXX8LPzw9WVlYIDg7GyZMnS9x23bp1kEgkejcrKysT1rYCtDk37JYiIiIyOrMHN5s3b8a0adMwb948nDlzBoGBgQgNDcWDBw9K3Mfe3h4xMTG62927d01Y4wrQ5twwoZiIiMjozB7cfPbZZxg/fjxGjx6NRo0aYdWqVVAqlVizZk2J+0gkEnh4eOhu7u7uJqxxOQkCE4qJiIhMyKzBTU5ODk6fPo2QkBBdmVQqRUhICI4fP17ifunp6fD19YWPjw969+6NS5culbitSqVCamqq3s2k8rIBTa54nzk3RERERmfW4CYhIQFqtbpIy4u7uztiY2OL3ad+/fpYs2YNdu3ahfXr10Oj0aBNmza4f/9+sdsvXrwYDg4OupuPj4/B30eptK02kAByW9Mem4iI6Clk9m6p8mrdujVGjhyJZs2aoUOHDti+fTvc3Nzw9ddfF7v9zJkzkZKSorvdu3fPtBUunEwsrXKnm4iIqMqxMOfBXV1dIZPJEBcXp1ceFxcHDw+PMr2GpaUlmjdvjhs3bhT7vEKhgEKheOK6VpiKI6WIiIhMyaxNCXK5HEFBQQgPD9eVaTQahIeHo3Xr1mV6DbVajQsXLsDT09NY1XwynOOGiIjIpMzacgMA06ZNQ1hYGFq0aIFWrVph6dKlyMjIwOjRowEAI0eOhLe3NxYvXgwAWLhwIZ577jnUrVsXycnJ+Pjjj3H37l2MGzfOnG+jZBwpRUREZFJmD24GDRqE+Ph4zJ07F7GxsWjWrBn27t2rSzKOjIyEtFCuysOHDzF+/HjExsbCyckJQUFBOHbsGBo1amSut1A6LppJRERkUhJBEARzV8KUUlNT4eDggJSUFNjbmyDgOLEK2DsDeKYvMHCd8Y9HRERUDZXn85vDd4yNOTdEREQmxeDG2DhaioiIyKQY3BibLufGwbz1ICIiekowuDE2jpYiIiIyKQY3xsacGyIiIpNicGNsbLkhIiIyKQY3xsZ5boiIiEyKwY2xseWGiIjIpBjcGJsu54ajpYiIiEyBwY0xaTRsuSEiIjIxBjfGlJMOIH91C+bcEBERmQSDG2PSttpILQALK/PWhYiI6CnB4MaYCs9xI5GYty5ERERPCQY3xsR8GyIiIpNjcGNM2ZydmIiIyNQY3BiTihP4ERERmRqDG2PS5dywW4qIiMhUGNwYky7nhi03REREpsLgxpiYUExERGRyDG6MiYtmEhERmRyDG2Niyw0REZHJMbgxJlWK+JM5N0RERCbD4MaYmFBMRERkcgxujIk5N0RERCbH4MaYmHNDRERkcgxujEnF5ReIiIhMjcGNMbHlhoiIyOQY3BiLOhfIzRTvWzmYty5ERERPEQY3xqJttQHYckNERGRCDG6MRZtvY2ENyCzNWxciIqKnCIMbY2G+DRERkVkwuDEWznFDRERkFgxujIUtN0RERGbB4MZYOMcNERGRWTC4MRZdcMOWGyIiIlNicGMsupwbznFDRERkSgxujIU5N0RERGbB4MZYmHNDRERkFgxujIUtN0RERGZRKYKbL7/8En5+frCyskJwcDBOnjxZpv02bdoEiUSCPn36GLeCFZHNhGIiIiJzMHtws3nzZkybNg3z5s3DmTNnEBgYiNDQUDx48KDU/e7cuYPp06ejXbt2JqppOWlbbjiJHxERkUmZPbj57LPPMH78eIwePRqNGjXCqlWroFQqsWbNmhL3UavVGDZsGBYsWIA6deqYsLbloEoRfzLnhoiIyKTMGtzk5OTg9OnTCAkJ0ZVJpVKEhITg+PHjJe63cOFC1KhRA2PHjn3sMVQqFVJTU/VuJqHLuWFwQ0REZEpmDW4SEhKgVqvh7u6uV+7u7o7Y2Nhi9zly5AhWr16Nb7/9tkzHWLx4MRwcHHQ3Hx+fJ653mTDnhoiIyCzM3i1VHmlpaRgxYgS+/fZbuLq6lmmfmTNnIiUlRXe7d++ekWsJQBCYc0NERGQmFuY8uKurK2QyGeLi4vTK4+Li4OHhUWT7mzdv4s6dO+jVq5euTKPRAAAsLCxw7do1+Pv76+2jUCigUCiMUPtS5KkATW5+BdhyQ0REZEpmbbmRy+UICgpCeHi4rkyj0SA8PBytW7cusn2DBg1w4cIFRERE6G4vvfQSOnXqhIiICNN1OT2OqlBej5zBDRERkSmZteUGAKZNm4awsDC0aNECrVq1wtKlS5GRkYHRo0cDAEaOHAlvb28sXrwYVlZWaNy4sd7+jo6OAFCk3Ky0XVJyO0BapXr+iIiIqjyzBzeDBg1CfHw85s6di9jYWDRr1gx79+7VJRlHRkZCWtUChOz8YeDMtyEiIjI5iSAIgrkrYUqpqalwcHBASkoK7O2NFHzcOgz88BLg1gCY/I9xjkFERPQUKc/ndxVrEqkiuGgmERGR2TC4MQYumklERGQ2DG6MQTuBH3NuiIiITI7BjTGw5YaIiMhsGNwYAxfNJCIiMhsGN8bARTOJiIjMhsGNMTDnhoiIyGwY3BgDc26IiIjMhsGNMXCeGyIiIrNhcGMMbLkhIiIyGwY3xsCcGyIiIrNhcGMMHC1FRERkNgxuDE2jYc4NERGRGTG4MbTcDAD5C60z54aIiMjkGNwYmjbfRiIDLK3NWxciIqKnkIW5K1DtaPNtrOwBicS8dSGiakej0SAnJ8fc1SAyCrlcDqn0ydtdGNwYmi7fhl1SRGRYOTk5uH37NjQajbmrQmQUUqkUtWvXhlwuf6LXYXBjaLrgxsG89SCiakUQBMTExEAmk8HHx8cg326JKhONRoPo6GjExMSgVq1akDxB7weDG0PLZssNERleXl4eMjMz4eXlBaVSae7qEBmFm5sboqOjkZeXB0tLywq/DkN/Qyucc0NEZCBqtRoAnri5nqgy017f2uu9ohjcGBpzbojIiJ6kqZ6osjPU9c3gxtA4OzERkVH5+flh6dKlZd7+0KFDkEgkSE5ONlqdqHJhcGNoXDSTiAiA+C28tNv8+fMr9LqnTp3ChAkTyrx9mzZtEBMTAwcH0w30aNCgARQKBWJjY012TCrA4MbQuGgmEREAICYmRndbunQp7O3t9cqmT5+u21YQBOTl5ZXpdd3c3MqVVC2Xy+Hh4WGyLr0jR44gKysLAwYMwPfff2+SY5YmNzfX3FUwOQY3hsZ1pYiIAAAeHh66m4ODAyQSie7x1atXYWdnhz179iAoKAgKhQJHjhzBzZs30bt3b7i7u8PW1hYtW7bEgQMH9F730W4piUSC7777Dn379oVSqURAQAB2796te/7Rbql169bB0dER+/btQ8OGDWFra4uuXbsiJiZGt09eXh5effVVODo6wsXFBTNmzEBYWBj69Onz2Pe9evVqDB06FCNGjMCaNWuKPH///n0MGTIEzs7OsLGxQYsWLfDPP//onv/ll1/QsmVLWFlZwdXVFX379tV7rzt37tR7PUdHR6xbtw4AcOfOHUgkEmzevBkdOnSAlZUVfvrpJyQmJmLIkCHw9vaGUqlEkyZNsHHjRr3X0Wg0WLJkCerWrQuFQoFatWph0aJFAIAXXngBU6ZM0ds+Pj4ecrkc4eHhjz0npsbgxtAY3BCRCQiCgMycPLPcBEEw2Pt455138OGHH+LKlSto2rQp0tPT0b17d4SHh+Ps2bPo2rUrevXqhcjIyFJfZ8GCBXj55Zdx/vx5dO/eHcOGDUNSUlKJ22dmZuKTTz7Bjz/+iL/++guRkZF6LUkfffQRfvrpJ6xduxZHjx5FampqkaCiOGlpadiyZQuGDx+OLl26ICUlBX///bfu+fT0dHTo0AFRUVHYvXs3zp07h7fffls3MeNvv/2Gvn37onv37jh79izCw8PRqlWrxx73Ue+88w5ee+01XLlyBaGhocjOzkZQUBB+++03XLx4ERMmTMCIESNw8uRJ3T4zZ87Ehx9+iDlz5uDy5cvYsGED3N3dAQDjxo3Dhg0boFKpdNuvX78e3t7eeOGFF8pdP2PjPDeGxpwbIjKBrFw1Gs3dZ5ZjX14YCqXcMB8fCxcuRJcuXXSPnZ2dERgYqHv83nvvYceOHdi9e3eRloPCRo0ahSFDhgAAPvjgAyxbtgwnT55E165di90+NzcXq1atgr+/PwBgypQpWLhwoe755cuXY+bMmbpWkxUrVuD3339/7PvZtGkTAgIC8MwzzwAABg8ejNWrV6Ndu3YAgA0bNiA+Ph6nTp2Cs7MzAKBu3bq6/RctWoTBgwdjwYIFurLC56OsXn/9dfTr10+vrHDwNnXqVOzbtw8///wzWrVqhbS0NHzxxRdYsWIFwsLCAAD+/v54/vnnAQD9+vXDlClTsGvXLrz88ssAxBawUaNGVcoRfGy5MTTm3BARlVmLFi30Hqenp2P69Olo2LAhHB0dYWtriytXrjy25aZp06a6+zY2NrC3t8eDBw9K3F6pVOoCGwDw9PTUbZ+SkoK4uDi9FhOZTIagoKDHvp81a9Zg+PDhusfDhw/Hli1bkJYmfvGNiIhA8+bNdYHNoyIiItC5c+fHHudxHj2varUa7733Hpo0aQJnZ2fY2tpi3759uvN65coVqFSqEo9tZWWl18125swZXLx4EaNGjXriuhoDW24MjS03RGQC1pYyXF4YarZjG4qNjY3e4+nTp2P//v345JNPULduXVhbW2PAgAGPXSz00dlsJRJJqWtwFbf9k3a3Xb58GSdOnMDJkycxY8YMXblarcamTZswfvx4WFtbl/oaj3u+uHoWlzD86Hn9+OOP8cUXX2Dp0qVo0qQJbGxs8Prrr+vO6+OOC4hdU82aNcP9+/exdu1avPDCC/D19X3sfubAlhtDY84NEZmARCKBUm5hlpsxuyGOHj2KUaNGoW/fvmjSpAk8PDxw584dox2vOA4ODnB3d8epU6d0ZWq1GmfOnCl1v9WrV6N9+/Y4d+4cIiIidLdp06Zh9erVAMQWpoiIiBLzgZo2bVpqgq6bm5te4vP169eRmZn52Pd09OhR9O7dG8OHD0dgYCDq1KmD//77T/d8QEAArK2tSz12kyZN0KJFC3z77bfYsGEDxowZ89jjmguDG0NS5wG5+RcZgxsionILCAjA9u3bERERgXPnzmHo0KFmWQV96tSpWLx4MXbt2oVr167htddew8OHD0sM7HJzc/Hjjz9iyJAhaNy4sd5t3Lhx+Oeff3Dp0iUMGTIEHh4e6NOnD44ePYpbt25h27ZtOH78OABg3rx52LhxI+bNm4crV67gwoUL+Oijj3THeeGFF7BixQqcPXsW//77L1555ZUyrcEUEBCA/fv349ixY7hy5Qr+97//IS4uTve8lZUVZsyYgbfffhs//PADbt68iRMnTuiCMq1x48bhww8/hCAIeqO4KhsGN4akbbUB2C1FRFQBn332GZycnNCmTRv06tULoaGhePbZZ01ejxkzZmDIkCEYOXIkWrduDVtbW4SGhsLKyqrY7Xfv3o3ExMRiP/AbNmyIhg0bYvXq1ZDL5fjjjz9Qo0YNdO/eHU2aNMGHH34ImUzs6uvYsSO2bNmC3bt3o1mzZnjhhRf0RjR9+umn8PHxQbt27TB06FBMnz69THP+zJ49G88++yxCQ0PRsWNHXYBV2Jw5c/Dmm29i7ty5aNiwIQYNGlQkb2nIkCGwsLDAkCFDSjwXlYFEMOSYviogNTUVDg4OSElJgb29gVtXHt4FvmgKWFgBs+Mevz0RURllZ2fj9u3bqF27dqX+UKmuNBoNGjZsiJdffhnvvfeeuatjNnfu3IG/vz9OnTpllKCztOu8PJ/fTCg2JObbEBFVC3fv3sUff/yBDh06QKVSYcWKFbh9+zaGDh1q7qqZRW5uLhITEzF79mw899xzZmlNKw92SxkSR0oREVULUqkU69atQ8uWLdG2bVtcuHABBw4cQMOGDc1dNbM4evQoPD09cerUKaxatcrc1XksttwYEue4ISKqFnx8fHD06FFzV6PS6Nixo0FnpjY2ttwYEltuiIiIzI7BjSGpUsSfzLkhIiIyGwY3hqRruWFwQ0REZC4MbgxJm3PDbikiIiKzqRTBzZdffgk/Pz9YWVkhODhYb8KiR23fvh0tWrSAo6MjbGxs0KxZM/z4448mrG0ptC03TCgmIiIyG7MHN5s3b8a0adMwb948nDlzBoGBgQgNDS1xNVdnZ2fMmjULx48fx/nz5zF69GiMHj0a+/btM3HNi6Fiyw0REZG5mT24+eyzzzB+/HiMHj0ajRo1wqpVq6BUKnXLqj+qY8eO6Nu3Lxo2bAh/f3+89tpraNq0KY4cOWLimheDOTdERAbXsWNHvP7667rHfn5+WLp0aan7SCQS7Ny584mPbajXIdMya3CTk5OD06dPIyQkRFcmlUoREhKiW0SsNIIgIDw8HNeuXUP79u2L3UalUiE1NVXvZjTMuSEi0unVqxe6du1a7HN///03JBIJzp8/X+7XPXXqFCZMmPCk1dMzf/58NGvWrEh5TEwMunXrZtBjlSQrKwvOzs5wdXWFSqUyyTGrK7MGNwkJCVCr1XB3d9crd3d3R2xsbIn7paSkwNbWFnK5HD169MDy5cvRpUuXYrddvHgxHBwcdDcfHx+Dvgc92m4pKwfjHYOIqIoYO3Ys9u/fj/v37xd5bu3atWjRogWaNm1a7td1c3Mr02KRhuDh4QGFQmGSY23btg3PPPMMGjRoYPbWIkEQkJeXZ9Y6PAmzd0tVhJ2dHSIiInDq1CksWrQI06ZNw6FDh4rddubMmUhJSdHd7t27Z7yKMeeGiEinZ8+ecHNzw7p16/TK09PTsWXLFowdOxaJiYkYMmQIvL29oVQq0aRJE2zcuLHU1320W+r69eto3749rKys0KhRI+zfv7/IPjNmzEC9evWgVCpRp04dzJkzB7m5uQCAdevWYcGCBTh37hwkEgkkEomuzo92S124cAEvvPACrK2t4eLiggkTJiA9PV33/KhRo9CnTx988skn8PT0hIuLCyZPnqw7VmlWr16N4cOHY/jw4Vi9enWR5y9duoSePXvC3t4ednZ2aNeuHW7evKl7fs2aNXjmmWegUCjg6emJKVOmABAXu5RIJIiIiNBtm5ycDIlEovvsPHToECQSCfbs2YOgoCAoFAocOXIEN2/eRO/eveHu7g5bW1u0bNkSBw4c0KuXSqXCjBkz4OPjA4VCgbp162L16tUQBAF169bFJ598ord9REQEJBIJbty48dhzUlFmXX7B1dUVMpkMcXH6K2jHxcXBw8OjxP2kUinq1q0LAGjWrBmuXLmCxYsXo2PHjkW2VSgUJou6mXNDRCYjCEBupnmObakEJJLHbmZhYYGRI0di3bp1mDVrFiT5+2zZsgVqtRpDhgxBeno6goKCMGPGDNjb2+O3337DiBEj4O/vj1atWj32GBqNBv369YO7uzv++ecfpKSk6OXnaNnZ2WHdunXw8vLChQsXMH78eNjZ2eHtt9/GoEGDcPHiRezdu1f3we3gULQFPiMjA6GhoWjdujVOnTqFBw8eYNy4cZgyZYpeAPfnn3/C09MTf/75J27cuIFBgwahWbNmGD9+fInv4+bNmzh+/Di2b98OQRDwxhtv4O7du/D19QUAREVFoX379ujYsSMOHjwIe3t7HD16VNe6snLlSkybNg0ffvghunXrhpSUlAotH/HOO+/gk08+QZ06deDk5IR79+6he/fuWLRoERQKBX744Qf06tUL165dQ61atQAAI0eOxPHjx7Fs2TIEBgbi9u3bSEhIgEQiwZgxY7B27VpMnz5dd4y1a9eiffv2us9xYzBrcCOXyxEUFITw8HD06dMHgHihhoeH6yLOstBoNObvnxQE5twQkenkZgIfeJnn2O9GA3KbMm06ZswYfPzxxzh8+LDuC+jatWvRv39/XbpA4Q++qVOnYt++ffj555/LFNwcOHAAV69exb59++DlJZ6PDz74oEiezOzZs3X3/fz8MH36dGzatAlvv/02rK2tYWtrCwsLi1K/WG/YsAHZ2dn44YcfYGMjvv8VK1agV69e+Oijj3QpFk5OTlixYgVkMhkaNGiAHj16IDw8vNTgZs2aNejWrRucnJwAAKGhoVi7di3mz58PQJwyxcHBAZs2bYKlpSUAoF69err933//fbz55pt47bXXdGUtW7Z87Pl71MKFC/XSPJydnREYGKh7/N5772HHjh3YvXs3pkyZgv/++w8///wz9u/fr8ufrVOnjm77UaNGYe7cuTh58iRatWqF3NxcbNiwoUhrjqGZvVtq2rRp+Pbbb/H999/jypUrmDhxIjIyMjB69GgAYkQ4c+ZM3faLFy/G/v37cevWLVy5cgWffvopfvzxRwwfPtxcb0GUpwI0+c2OnOeGiAgA0KBBA7Rp00Y3AvbGjRv4+++/MXbsWACAWq3Ge++9hyZNmsDZ2Rm2trbYt28fIiMjy/T6V65cgY+Pjy6wAYDWrVsX2W7z5s1o27YtPDw8YGtri9mzZ5f5GIWPFRgYqAtsAKBt27bQaDS4du2aruyZZ56BTCbTPfb09CxxehNAPAfff/+93ufY8OHDsW7dOmg0GgBiV067du10gU1hDx48QHR0NDp37lyu91OcFi1a6D1OT0/H9OnT0bBhQzg6OsLW1hZXrlzRnbuIiAjIZDJ06NCh2Nfz8vJCjx49dL//X375BSqVCgMHDnziupbG7KuCDxo0CPHx8Zg7dy5iY2PRrFkz7N27VxcBR0ZGQiotiMEyMjIwadIk3L9/H9bW1mjQoAHWr1+PQYMGmestiLRdUgAgtzVfPYjo6WCpFFtQzHXschg7diymTp2KL7/8EmvXroW/v7/uw/Djjz/GF198gaVLl6JJkyawsbHB66+/jpycHINV9/jx4xg2bBgWLFiA0NBQXQvIp59+arBjFPZoACKRSHRBSnH27duHqKioIp9jarUa4eHh6NKlC6ytrUvcv7TnAOg+Qwuv6l1SDlDhwA0Apk+fjv379+OTTz5B3bp1YW1tjQEDBuh+P487NgCMGzcOI0aMwOeff461a9di0KBBRk8IN3twAwBTpkwpsRvq0UTh999/H++//74JalVO2mRiuR0glZW+LRHRk5JIytw1ZG4vv/wyXnvtNWzYsAE//PADJk6cqMu/OXr0KHr37q1rtdBoNPjvv//QqFGjMr12w4YNce/ePcTExMDT0xMAcOLECb1tjh07Bl9fX8yaNUtXdvfuXb1t5HI51Gr1Y4+1bt06ZGRk6IKAo0ePQiqVon79+mWqb3FWr16NwYMH69UPABYtWoTVq1ejS5cuaNq0Kb7//nvk5uYWCZ7s7Ozg5+eH8PBwdOrUqcjru7m5ARCHtTdv3hwA9JKLS3P06FGMGjUKffv2BSC25Ny5c0f3fJMmTaDRaHD48GG9aV0K6969O2xsbLBy5Urs3bsXf/31V5mO/STM3i1VbXCkFBFRsWxtbTFo0CDMnDkTMTExGDVqlO65gIAA7N+/H8eOHcOVK1fwv//9r8ggk9KEhISgXr16CAsLw7lz5/D3338XCRICAgIQGRmJTZs24ebNm1i2bBl27Niht42fnx9u376NiIgIJCQkFJvHOWzYMFhZWSEsLAwXL17En3/+ialTp2LEiBFFpjQpq/j4ePzyyy8ICwtD48aN9W4jR47Ezp07kZSUhClTpiA1NRWDBw/Gv//+i+vXr+PHH3/UdYfNnz8fn376KZYtW4br16/jzJkzWL58OQCxdeW5557Dhx9+iCtXruDw4cN6OUilCQgIwPbt2xEREYFz585h6NCheq1Qfn5+CAsLw5gxY7Bz507cvn0bhw4dws8//6zbRiaTYdSoUZg5cyYCAgKK7TY0NAY3hpKnElttOMcNEVERY8eOxcOHDxEaGqqXHzN79mw8++yzCA0NRceOHeHh4aEbYFIWUqkUO3bsQFZWFlq1aoVx48Zh0aJFetu89NJLeOONNzBlyhQ0a9YMx44dw5w5c/S26d+/P7p27YpOnTrBzc2t2OHoSqUS+/btQ1JSElq2bIkBAwagc+fOWLFiRflORiHa5OTi8mU6d+4Ma2trrF+/Hi4uLjh48CDS09PRoUMHBAUF4dtvv9W14oSFhWHp0qX46quv8Mwzz6Bnz564fv267rXWrFmDvLw8BAUF4fXXXy9zD8hnn30GJycntGnTBr169UJoaCieffZZvW1WrlyJAQMGYNKkSWjQoAHGjx+PjIwMvW3Gjh2LnJwcXT6tsUmEwp1wT4HU1FQ4ODggJSUF9vZGSPwVhDINkSQiKo/s7Gzcvn0btWvXhpWVlbmrQ1Quf//9Nzp37ox79+6V2spV2nVens/vSpFzU60wsCEiIgIgTvAXHx+P+fPnY+DAgRXuvisvdksRERGRUWzcuBG+vr5ITk7GkiVLTHZcBjdERERkFKNGjYJarcbp06fh7e1tsuMyuCEiIqJqhcENERERVSsMboiIqpCnbIArPWUMdX0zuCEiqgK0axUZclkCospGe30XXpurIjgUnIioCrCwsIBSqUR8fDwsLS311twjqg40Gg3i4+OhVCphYfFk4QmDGyKiKkAikcDT0xO3b98usi4SUXUhlUpRq1Yt3dpjFcXghoioipDL5QgICGDXFFVbcrncIK2SDG6IiKoQqVTK5ReIHoOdtkRERFStMLghIiKiaoXBDREREVUrT13OjXaCoNTUVDPXhIiIiMpK+7ldlon+nrrgJi0tDQDg4+Nj5poQERFReaWlpcHBwaHUbSTCUzaXt0ajQXR0NOzs7IqMo09NTYWPjw/u3bsHe3t7M9Ww6uF5qxiet4rheSs/nrOK4XmrGGOdN0EQkJaWBi8vr8cOF3/qWm6kUilq1qxZ6jb29va8kCuA561ieN4qhuet/HjOKobnrWKMcd4e12KjxYRiIiIiqlYY3BAREVG1wuCmEIVCgXnz5kGhUJi7KlUKz1vF8LxVDM9b+fGcVQzPW8VUhvP21CUUExERUfXGlhsiIiKqVhjcEBERUbXC4IaIiIiqFQY3REREVK0wuCnkyy+/hJ+fH6ysrBAcHIyTJ0+au0qV2vz58yGRSPRuDRo0MHe1Kp2//voLvXr1gpeXFyQSCXbu3Kn3vCAImDt3Ljw9PWFtbY2QkBBcv37dPJWtJB53zkaNGlXk2uvatat5KltJLF68GC1btoSdnR1q1KiBPn364Nq1a3rbZGdnY/LkyXBxcYGtrS369++PuLg4M9W4cijLeevYsWOR6+2VV14xU40rh5UrV6Jp06a6ifpat26NPXv26J4397XG4Cbf5s2bMW3aNMybNw9nzpxBYGAgQkND8eDBA3NXrVJ75plnEBMTo7sdOXLE3FWqdDIyMhAYGIgvv/yy2OeXLFmCZcuWYdWqVfjnn39gY2OD0NBQZGdnm7imlcfjzhkAdO3aVe/a27hxowlrWPkcPnwYkydPxokTJ7B//37k5ubixRdfREZGhm6bN954A7/88gu2bNmCw4cPIzo6Gv369TNjrc2vLOcNAMaPH693vS1ZssRMNa4catasiQ8//BCnT5/Gv//+ixdeeAG9e/fGpUuXAFSCa00gQRAEoVWrVsLkyZN1j9VqteDl5SUsXrzYjLWq3ObNmycEBgaauxpVCgBhx44duscajUbw8PAQPv74Y11ZcnKyoFAohI0bN5qhhpXPo+dMEAQhLCxM6N27t1nqU1U8ePBAACAcPnxYEATxurK0tBS2bNmi2+bKlSsCAOH48ePmqmal8+h5EwRB6NChg/Daa6+Zr1JVhJOTk/Ddd99VimuNLTcAcnJycPr0aYSEhOjKpFIpQkJCcPz4cTPWrPK7fv06vLy8UKdOHQwbNgyRkZHmrlKVcvv2bcTGxupdew4ODggODua19xiHDh1CjRo1UL9+fUycOBGJiYnmrlKlkpKSAgBwdnYGAJw+fRq5ubl611qDBg1Qq1YtXmuFPHretH766Se4urqicePGmDlzJjIzM81RvUpJrVZj06ZNyMjIQOvWrSvFtfbULZxZnISEBKjVari7u+uVu7u74+rVq2aqVeUXHByMdevWoX79+oiJicGCBQvQrl07XLx4EXZ2duauXpUQGxsLAMVee9rnqKiuXbuiX79+qF27Nm7evIl3330X3bp1w/HjxyGTycxdPbPTaDR4/fXX0bZtWzRu3BiAeK3J5XI4OjrqbctrrUBx5w0Ahg4dCl9fX3h5eeH8+fOYMWMGrl27hu3bt5uxtuZ34cIFtG7dGtnZ2bC1tcWOHTvQqFEjREREmP1aY3BDFdatWzfd/aZNmyI4OBi+vr74+eefMXbsWDPWjKq7wYMH6+43adIETZs2hb+/Pw4dOoTOnTubsWaVw+TJk3Hx4kXmwJVTSedtwoQJuvtNmjSBp6cnOnfujJs3b8Lf39/U1aw06tevj4iICKSkpGDr1q0ICwvD4cOHzV0tAEwoBgC4urpCJpMVyeSOi4uDh4eHmWpV9Tg6OqJevXq4ceOGuatSZWivL157T6ZOnTpwdXXltQdgypQp+PXXX/Hnn3+iZs2aunIPDw/k5OQgOTlZb3tea6KSzltxgoODAeCpv97kcjnq1q2LoKAgLF68GIGBgfjiiy8qxbXG4AbiLygoKAjh4eG6Mo1Gg/DwcLRu3dqMNata0tPTcfPmTXh6epq7KlVG7dq14eHhoXftpaam4p9//uG1Vw73799HYmLiU33tCYKAKVOmYMeOHTh48CBq166t93xQUBAsLS31rrVr164hMjLyqb7WHnfeihMREQEAT/X1VhyNRgOVSlU5rjWTpC1XAZs2bRIUCoWwbt064fLly8KECRMER0dHITY21txVq7TefPNN4dChQ8Lt27eFo0ePCiEhIYKrq6vw4MEDc1etUklLSxPOnj0rnD17VgAgfPbZZ8LZs2eFu3fvCoIgCB9++KHg6Ogo7Nq1Szh//rzQu3dvoXbt2kJWVpaZa24+pZ2ztLQ0Yfr06cLx48eF27dvCwcOHBCeffZZISAgQMjOzjZ31c1m4sSJgoODg3Do0CEhJiZGd8vMzNRt88orrwi1atUSDh48KPz7779C69athdatW5ux1ub3uPN248YNYeHChcK///4r3L59W9i1a5dQp04doX379mauuXm98847wuHDh4Xbt28L58+fF9555x1BIpEIf/zxhyAI5r/WGNwUsnz5cqFWrVqCXC4XWrVqJZw4ccLcVarUBg0aJHh6egpyuVzw9vYWBg0aJNy4ccPc1ap0/vzzTwFAkVtYWJggCOJw8Dlz5gju7u6CQqEQOnfuLFy7ds28lTaz0s5ZZmam8OKLLwpubm6CpaWl4OvrK4wfP/6p/yJS3PkCIKxdu1a3TVZWljBp0iTByclJUCqVQt++fYWYmBjzVboSeNx5i4yMFNq3by84OzsLCoVCqFu3rvDWW28JKSkp5q24mY0ZM0bw9fUV5HK54ObmJnTu3FkX2AiC+a81iSAIgmnaiIiIiIiMjzk3REREVK0wuCEiIqJqhcENERERVSsMboiIiKhaYXBDRERE1QqDGyIiIqpWGNwQERFRtcLghoieehKJBDt37jR3NYjIQBjcEJFZjRo1ChKJpMita9eu5q4aEVVRFuauABFR165dsXbtWr0yhUJhptoQUVXHlhsiMjuFQgEPDw+9m5OTEwCxy2jlypXo1q0brK2tUadOHWzdulVv/wsXLuCFF16AtbU1XFxcMGHCBKSnp+tts2bNGjzzzDNQKBTw9PTElClT9J5PSEhA3759oVQqERAQgN27dxv3TROR0TC4IaJKb86cOejfvz/OnTuHYcOGYfDgwbhy5QoAICMjA6GhoXBycsKpU6ewZcsWHDhwQC94WblyJSZPnowJEybgwoUL2L17N+rWrat3jAULFuDll1/G+fPn0b17dwwbNgxJSUkmfZ9EZCAmW6KTiKgYYWFhgkwmE2xsbPRuixYtEgRBXLX5lVde0dsnODhYmDhxoiAIgvDNN98ITk5OQnp6uu753377TZBKpbqVwr28vIRZs2aVWAcAwuzZs3WP09PTBQDCnj17DPY+ich0mHNDRGbXqVMnrFy5Uq/M2dlZd79169Z6z7Vu3RoREREAgCtXriAwMBA2Nja659u2bQuNRoNr165BIpEgOjoanTt3LrUOTZs21d23sbGBvb09Hjx4UNG3RERmxOCGiMzOxsamSDeRoVhbW5dpO0tLS73HEokEGo3GGFUiIiNjzg0RVXonTpwo8rhhw4YAgIYNG+LcuXPIyMjQPX/06FFIpVLUr18fdnZ28PPzQ3h4uEnrTETmw5YbIjI7lUqF2NhYvTILCwu4uroCALZs2YIWLVrg+eefx08//YSTJ09i9erVAIBhw4Zh3rx5CAsLw/z58xEfH4+pU6dixIgRcHd3BwDMnz8fr7zyCmrUqIFu3bohLS0NR48exdSpU037RonIJBjcEJHZ7d27F56ennpl9evXx9WrVwGII5k2bdqESZMmwdPTExs3bkSjRo0AAEqlEvv27cNrr72Gli1bQqlUon///vjss890rxUWFobs7Gx8/vnnmD59OlxdXTFgwADTvUEiMimJIAiCuStBRFQSiUSCHTt2oE+fPuauChFVEcy5ISIiomqFwQ0RERFVK8y5IaJKjT3nRFRebLkhIiKiaoXBDREREVUrDG6IiIioWmFwQ0RERNUKgxsiIiKqVhjcEBERUbXC4IaIiIiqFQY3REREVK0wuCEiIqJq5f8jd2kQpfNunQAAAABJRU5ErkJggg==\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "# Extract training and validation accuracies from history_cnn_2 object\n",
        "train_acc = history_cnn_2.history['accuracy']\n",
        "val_acc = history_cnn_2.history['val_accuracy']\n",
        "\n",
        "# Create an epochs range\n",
        "epochs_range = range(1, len(train_acc) + 1)\n",
        "\n",
        "# Plot training and validation accuracies\n",
        "plt.plot(epochs_range, train_acc, label='Training Accuracy')\n",
        "plt.plot(epochs_range, val_acc, label='Validation Accuracy')\n",
        "plt.title('Training and Validation Accuracies for CNN Model 2')\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qrrt0Ac3AjQH"
      },
      "source": [
        "**Observations:**\n",
        "\n",
        "1.   Both training and validation accuracy are increase over epochs.\n",
        "2.   The accuracy raises quickly and remain stable, which proves the model architecture might be appropriate.\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kja4SnOdAjQI"
      },
      "source": [
        "## **Predictions on the test data**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eHCRwRbgKDN2"
      },
      "source": [
        "### **Make predictions on the test data using the second model.**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "f1d-VvaLAjQI",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "778ddba8-5e2b-4ff5-dd64-a8137993a572"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Classification Report:\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.93      0.94      0.93      1814\n",
            "           1       0.93      0.87      0.90      1828\n",
            "           2       0.88      0.94      0.91      1803\n",
            "           3       0.85      0.90      0.87      1719\n",
            "           4       0.90      0.94      0.92      1812\n",
            "           5       0.91      0.90      0.90      1768\n",
            "           6       0.88      0.90      0.89      1832\n",
            "           7       0.95      0.91      0.93      1808\n",
            "           8       0.92      0.86      0.89      1812\n",
            "           9       0.91      0.90      0.91      1804\n",
            "\n",
            "    accuracy                           0.91     18000\n",
            "   macro avg       0.91      0.91      0.91     18000\n",
            "weighted avg       0.91      0.91      0.91     18000\n",
            "\n",
            "Confusion Matrix:\n",
            "[[1698   15   20   10    8    3   28    3   10   19]\n",
            " [  22 1593   24   38   81    6   17   31   11    5]\n",
            " [  14    9 1701   13   14    2    8   18    6   18]\n",
            " [   2   10   29 1554   14   51   16    6   17   20]\n",
            " [   5   14   24   17 1696    5   19    4    8   20]\n",
            " [   3    5   10   66    9 1591   51    2   10   21]\n",
            " [  20    7   13   19   16   45 1652    4   48    8]\n",
            " [  14   41   47   20   18   10    3 1637    5   13]\n",
            " [  18    5   25   68   12   18   73    2 1562   29]\n",
            " [  39   10   31   30   15   20   12    9   18 1620]]\n"
          ]
        }
      ],
      "source": [
        "# Make predictions on the test set\n",
        "y_pred_probs_2 = model_cnn_2.predict(X_test_normalized)\n",
        "\n",
        "# Convert probabilities to class labels\n",
        "y_pred_2 = np.argmax(y_pred_probs_2, axis=1)\n",
        "\n",
        "\n",
        "from sklearn.metrics import classification_report, confusion_matrix\n",
        "y_true = np.argmax(y_test_encoded, axis=1)\n",
        "\n",
        "# Print classification report\n",
        "print(\"Classification Report:\")\n",
        "print(classification_report(y_true, y_pred_2))\n",
        "\n",
        "# Print confusion matrix\n",
        "print(\"Confusion Matrix:\")\n",
        "print(confusion_matrix(y_true, y_pred_2))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DrV1tOG0AjQI"
      },
      "source": [
        "**Note:** Earlier, we noticed that each entry of the target variable is a one-hot encoded vector, but to print the classification report and confusion matrix, we must convert each entry of y_test to a single label."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "dUSHU9W0AjQI"
      },
      "outputs": [],
      "source": [
        "# Convert each entry of y_test to a single label, and print the classification report and confusion matrix\n",
        "y_true = np.argmax(y_test_encoded, axis=1)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TNN9v713AjQJ"
      },
      "source": [
        "## **Insights of the 4 models**\n",
        "- As both the second ANN and CNN models are more complex including more layers and number of parameters, they have higher accuracies and better overall performances.\n",
        "- The second CNN model achieves the highest accuracy (91.27%), while the possible overfitting appearing in later epoches.\n",
        "\n",
        "## **Conclusion**\n",
        "After compared the 4 models, I would choose the second CNN model since it has higher computational efficiency concerning the relatively small model size. This decision is based on the theratical decision based on the given datasets. There would be more concerns on the specific cases in the real world, such as inference time, memory consumption, and deployment.\n",
        "\n"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "provenance": [],
      "gpuType": "T4",
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "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": "3.11.1"
    },
    "vscode": {
      "interpreter": {
        "hash": "ff4b1fca65a764b45acb559e482afe389d289dd599b9f8c5fd12ff5c2ea46a65"
      }
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}