notebook.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "PseCCEKqlXY0"
   },
   "source": [
    "# INFORMATION"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "c1XbVKIE-RB4"
   },
   "outputs": [],
   "source": [
    "# NETWORK TYPE       : SIAMESE NETWORK\n",
    "\n",
    "# Dataset            : ...\n",
    "# Iteration          : 760\n",
    "# Training set size  : 75754\n",
    "# Test set size      : 16368\n",
    "# Model Parameters   : 1,240,588\n",
    "\n",
    "\n",
    "# Preprocessing Time : About 5 mins\n",
    "# Training Time      : About 30 mins\n",
    "# GPU Used           : NVIDIA RTX 2080 Ti"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "HbRjqEm56pce"
   },
   "source": [
    "# PRELIMINARIES"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "KjzZZW-C6sLA"
   },
   "outputs": [],
   "source": [
    "'''PATHS'''\n",
    "\n",
    "# Path to directroy contain notebook and dataset. \n",
    "working_directory = \"/content/drive/My Drive/Signature_Verification\"\n",
    "\n",
    "working_directory = \".\"\n",
    "\n",
    "\n",
    "# Path of training set.\n",
    "training_set_path = \"./sign_data/train\"\n",
    "\n",
    "# Path of test set.\n",
    "test_set_path = \"./sign_data/test\"\n",
    "\n",
    "# Extention applied to folders containg real files.\n",
    "real_extention = \"\"\n",
    "\n",
    "# Extention applied to folders containg fake files.\n",
    "fake_extention = \"_forg\"\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "jY-U1pXB8rsv"
   },
   "outputs": [],
   "source": [
    "'''HYPER-PARAMETERS'''\n",
    "\n",
    "# Hyperparameters\n",
    "\n",
    "global_image_height = 400\n",
    "global_image_width = 700\n",
    "\n",
    "triplet_loss_margin = 1\n",
    "learning_rate = 0.001\n",
    "batch_size = 100\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "JaJW7IPs2A6F"
   },
   "outputs": [],
   "source": [
    "'''IMPORTS'''\n",
    "\n",
    "import os\n",
    "import numpy as np\n",
    "from matplotlib import pyplot as plt\n",
    "\n",
    "from PIL import Image\n",
    "import cv2\n",
    "\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torchvision\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "from torchvision import transforms, utils\n",
    "from torch.autograd import Variable\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 170
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 6297,
     "status": "ok",
     "timestamp": 1562564895423,
     "user": {
      "displayName": "Aditya Ahuja",
      "photoUrl": "",
      "userId": "08809216472288903233"
     },
     "user_tz": -330
    },
    "id": "k0rKiZivdtja",
    "outputId": "0510a771-fc01-4f51-b186-ec50ebb468da",
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dataset.zip  import.sh\tnotebook.ipynb\tonstart.log  onstart.sh  sign_data\r\n"
     ]
    }
   ],
   "source": [
    "'''MOUNTING GOOGLE DRIVE'''\n",
    "\n",
    "# Mount Google Drive\n",
    "# from google.colab import drive\n",
    "# drive.mount('/content/drive')\n",
    "\n",
    "# Unzip the Dataset\n",
    "# !unzip dataset.zip\n",
    "\n",
    "# Explore Directory\n",
    "os.chdir(working_directory)\n",
    "!ls"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "-BkJG23L61Bp"
   },
   "source": [
    "\n",
    "\n",
    "# AUXILLARY FUNCTIONS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "4QU2bvgXeS2V"
   },
   "outputs": [],
   "source": [
    "def get_single_perm(is_train, location):\n",
    "    \n",
    "    \"\"\"\n",
    "    Retuns a list containing all [anchor, positive, negative] path permutations \n",
    "    in a given folder.\n",
    "    \n",
    "    Note : `location` must be an integer denoting the folder number.\n",
    "    \"\"\"\n",
    "   \n",
    "    index = []\n",
    "        \n",
    "    if(is_train):\n",
    "        root_path = training_set_path\n",
    "    else:\n",
    "        root_path = test_set_path\n",
    "    \n",
    "    originals = os.listdir(os.path.join(root_path, location + real_extention))\n",
    "    forgeries = os.listdir(os.path.join(root_path, location + fake_extention))\n",
    "\n",
    "    originals.sort()\n",
    "    forgeries.sort()\n",
    "    \n",
    "    for i in range(len(originals)):\n",
    "        for j in range(i+1,len(originals)):\n",
    "            for k in range(len(forgeries)):\n",
    "                \n",
    "                path_anchor = os.path.join(root_path, location + real_extention, originals[i])\n",
    "                path_real = os.path.join(root_path, location + real_extention, originals[j])\n",
    "                path_fake = os.path.join(root_path, location + fake_extention, forgeries[k])\n",
    "                \n",
    "                index.append([path_anchor, path_real, path_fake])\n",
    "  \n",
    "    return index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "XS9Ly_wm7pTV"
   },
   "outputs": [],
   "source": [
    "def build_file_index(is_train=True):\n",
    "    \n",
    "    \"\"\"\n",
    "    Returns a master index file that contains all the \n",
    "    [anchor, real, fake] path permuataions across the whole dataset.    \n",
    "    \"\"\"\n",
    "        \n",
    "    if(is_train):\n",
    "        directory = training_set_path\n",
    "    else:\n",
    "        directory = test_set_path\n",
    "    \n",
    "    folder_list = os.listdir(directory)\n",
    "    folder_list.sort()\n",
    "                           \n",
    "    master_index = []\n",
    "                           \n",
    "    for i in range(len(folder_list)):\n",
    "        if(i%2==0):\n",
    "            master_index = master_index + get_single_perm(is_train, folder_list[i])    \n",
    "    \n",
    "    return master_index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ZJR92LB5sGhT"
   },
   "outputs": [],
   "source": [
    "def get_image_from_path(image_path):\n",
    "    \n",
    "    \"\"\"\n",
    "    Returns a numpy array containing the image at the provided `path`.\n",
    "    \"\"\"\n",
    "    \n",
    "    image = cv2.imread(str(image_path))\n",
    "    image = np.array(image)\n",
    "    return image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 695
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2964,
     "status": "ok",
     "timestamp": 1562564900622,
     "user": {
      "displayName": "Aditya Ahuja",
      "photoUrl": "",
      "userId": "08809216472288903233"
     },
     "user_tz": -330
    },
    "id": "dmtJUYwdu0gS",
    "outputId": "3afbd5e6-fff4-4b06-b448-90649fc22a7d"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def print_img(image):\n",
    "    \n",
    "    \"\"\"\n",
    "    Plots a simgle image. \n",
    "    Accepts raw data, NOT path.\n",
    "    \"\"\"    \n",
    "    \n",
    "    imgplot = plt.imshow(image)\n",
    "    plt.show()\n",
    "\n",
    "    \n",
    "print_img(get_image_from_path('./sign_data/train/001/001_01.PNG'))\n",
    "print_img(get_image_from_path('./sign_data/train/001/001_02.PNG'))\n",
    "print_img(get_image_from_path('./sign_data/train/001_forg/0201001_02.png'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "WxuIevZYVge-"
   },
   "source": [
    "# DATA WRANGLING"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "P75_TcYgVhGJ"
   },
   "outputs": [],
   "source": [
    "'''DATA PREPROCESSING'''\n",
    "\n",
    "\n",
    "def resize_images(new_dim1=global_image_height, new_dim2=global_image_width, is_train=True):\n",
    "\n",
    "    '''\n",
    "    Transforms all images in the dataset into correct sizes.\n",
    "    Then rewrites them at their original path.\n",
    "    '''\n",
    "    \n",
    "    if(is_train):\n",
    "        directory = training_set_path\n",
    "    else:\n",
    "        directory = test_set_path\n",
    "\n",
    "    # Get all Folders    \n",
    "    folder_list = os.listdir(directory)\n",
    "    folder_list.sort()            \n",
    "        \n",
    "    for folder in folder_list:\n",
    "        \n",
    "        # Get all Images in a folder\n",
    "        image_list = os.listdir(os.path.join(directory,folder))\n",
    "        image_list.sort()    \n",
    "                                  \n",
    "        for image_name in image_list:\n",
    "            \n",
    "            # Get image path\n",
    "            image_path = str(os.path.join(directory,folder,image_name))\n",
    "            \n",
    "            # Get image\n",
    "            image = get_image_from_path(image_path)\n",
    "            \n",
    "            # Resize image\n",
    "            resized_image = cv2.resize(image, (new_dim2, new_dim1))\n",
    "            \n",
    "            # Overwrite the Image\n",
    "            cv2.imwrite(image_path, resized_image)            \n",
    "\n",
    "# Applting the tranformations over the training and test set.\n",
    "resize_images(is_train = True)\n",
    "resize_images(is_train = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1697,
     "status": "ok",
     "timestamp": 1562564903539,
     "user": {
      "displayName": "Aditya Ahuja",
      "photoUrl": "",
      "userId": "08809216472288903233"
     },
     "user_tz": -330
    },
    "id": "rRBvdwLG0Q78",
    "outputId": "57d62877-05f2-4b7e-8324-d309975a26fc"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'MODIFIED IMAGES'"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "'''MODIFIED IMAGES'''\n",
    "\n",
    "# print_img(get_image_from_path('./sign_data/train/001/001_01.PNG'))\n",
    "# print_img(get_image_from_path('./sign_data/train/001/001_02.PNG'))\n",
    "# print_img(get_image_from_path('./sign_data/train/001_forg/0201001_02.png'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "qUGdxtpu8YKJ"
   },
   "source": [
    "# CREATING THE DATASET CLASS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2916,
     "status": "ok",
     "timestamp": 1562564905742,
     "user": {
      "displayName": "Aditya Ahuja",
      "photoUrl": "",
      "userId": "08809216472288903233"
     },
     "user_tz": -330
    },
    "id": "fd0aWaWI9doA",
    "outputId": "fe5d06eb-6a16-4b56-a2bc-86fa851b5def"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training set size :  75754\n",
      "Test set size :  16368\n"
     ]
    }
   ],
   "source": [
    "# List of all the path permutations in the train file\n",
    "train_master_index = build_file_index(is_train = True)\n",
    "\n",
    "# List of all the path permutations in the test file\n",
    "test_master_index = build_file_index(is_train = False)\n",
    "\n",
    "print(\"Training set size : \", len(train_master_index))\n",
    "print(\"Test set size : \", len(test_master_index))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ctfrEb2QIwtT"
   },
   "outputs": [],
   "source": [
    "'''CREATING DATASET CLASS'''\n",
    "\n",
    "def transform_image(path):\n",
    "    \n",
    "    \"\"\"\n",
    "    Rotate the image so that the model can process it.\n",
    "    \"\"\"\n",
    "    \n",
    "    image = get_image_from_path(path)\n",
    "    image = np.rollaxis(image, 2, 0)\n",
    "    return image\n",
    "\n",
    "\n",
    "class signature_dataset(Dataset):\n",
    "        \n",
    "    def __init__(self, is_train=True, transform=None):  \n",
    "        \n",
    "        self.is_train = is_train\n",
    "        self.transform = transform\n",
    "    \n",
    "    def __getitem__(self, index):\n",
    "        \n",
    "        if(self.is_train):\n",
    "            master_index = train_master_index\n",
    "        else:\n",
    "            master_index = test_master_index\n",
    "        \n",
    "        # Get paths\n",
    "        anchor_path = master_index[index][0]\n",
    "        real_path = master_index[index][1]\n",
    "        fake_path = master_index[index][2]\n",
    "            \n",
    "        # Transform images\n",
    "        anchor = transform_image(anchor_path)\n",
    "        real = transform_image(real_path)\n",
    "        fake = transform_image(fake_path)\n",
    "        \n",
    "        \n",
    "        if self.transform is not None:\n",
    "            anchor = self.transform(anchor)\n",
    "            real = self.transform(real)\n",
    "            fake = self.transform(fake)\n",
    "        \n",
    "        return anchor, real, fake\n",
    "    \n",
    "    \n",
    "    def __len__(self):\n",
    "        \n",
    "        if(self.is_train):\n",
    "            master_index = train_master_index\n",
    "        else:\n",
    "            master_index = test_master_index\n",
    "        \n",
    "        return len(master_index)\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ioWpL6ZGoE4A"
   },
   "outputs": [],
   "source": [
    "'''LOADING DATASET'''\n",
    "\n",
    "train_dataset = signature_dataset(is_train = True)\n",
    "test_dataset = signature_dataset(is_train = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "WVC50Cizn7nI"
   },
   "outputs": [],
   "source": [
    "'''MAKING DATASET ITERABLE'''\n",
    "\n",
    "train_loader = torch.utils.data.DataLoader(dataset=train_dataset, batch_size=batch_size, shuffle=True)\n",
    "test_loader = torch.utils.data.DataLoader(dataset=test_dataset, batch_size=batch_size, shuffle=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Ap_SJnaLEW8B"
   },
   "source": [
    "# CREATING THE MODEL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "5GfbsMKXnyz9"
   },
   "outputs": [],
   "source": [
    "'''CREATEING MODEL CLASS'''\n",
    "\n",
    "class SiameseModel(nn.Module):\n",
    "    \n",
    "    def __init__(self):\n",
    "        \n",
    "        super(SiameseModel, self).__init__()\n",
    "        \n",
    "        # Convolution 1 [03,400,700 -> 16,77,137]\n",
    "        self.cnn1 = nn.Conv2d(in_channels=3, out_channels=16, kernel_size=20, stride=5, padding=0)\n",
    "        self.relu1 = nn.ReLU() \n",
    "        \n",
    "        # Max pool 1 [16,77,137 -> 16,38,68]\n",
    "        self.maxpool1 = nn.MaxPool2d(kernel_size=2, stride=2)\n",
    "\n",
    "        # Convolution 2 [16,38,68 -> 32,36,66]\n",
    "        self.cnn2 = nn.Conv2d(in_channels=16, out_channels=32, kernel_size=3, stride=1, padding=0)\n",
    "        self.relu2 = nn.ReLU()\n",
    "        \n",
    "        # Max pool 2 [32,36,66 -> 32,18,33]\n",
    "        self.maxpool2 = nn.MaxPool2d(kernel_size=2, stride=2)\n",
    "        \n",
    "        # Convolution 3 [32,18,33 -> 64,16,31]\n",
    "        self.cnn3 = nn.Conv2d(in_channels=32, out_channels=64, kernel_size=3, stride=1, padding=0)\n",
    "        self.relu3 = nn.ReLU()\n",
    "        \n",
    "        # Max pool 3 [64,16,31 -> 64,8,15]\n",
    "        self.maxpool3 = nn.MaxPool2d(kernel_size=[2,3], stride=2)\n",
    "        \n",
    "        # Fully connected (readout) [64*8*15 -> 156]\n",
    "        self.fc1 = nn.Linear(64 * 8 * 15, 156) \n",
    "    \n",
    "    def forward(self, x):\n",
    "        # Convolution 1\n",
    "        out = self.cnn1(x)\n",
    "        out = self.relu1(out)\n",
    "        \n",
    "        # Max pool 1\n",
    "        out = self.maxpool1(out)\n",
    "        \n",
    "        # Convolution 2 \n",
    "        out = self.cnn2(out)\n",
    "        out = self.relu2(out)\n",
    "        \n",
    "        # Max pool 2 \n",
    "        out = self.maxpool2(out)\n",
    "        \n",
    "        # Convolution 3\n",
    "        out = self.cnn3(out)\n",
    "        out = self.relu3(out)\n",
    "        \n",
    "        # Max pool 3 \n",
    "        out = self.maxpool3(out)\n",
    "        \n",
    "        # Resize [100,64,8,15 -> 100,64*8*15]\n",
    "        out = out.view(out.size(0), -1)    \n",
    "        \n",
    "        # Linear function (readout)\n",
    "        out = self.fc1(out)\n",
    "        \n",
    "        return out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "MninVWyfnz-C"
   },
   "outputs": [],
   "source": [
    "'''INSTANTIATING MODEL CLASS'''\n",
    "\n",
    "model = SiameseModel()\n",
    "\n",
    "if torch.cuda.is_available():\n",
    "    model.cuda()\n",
    "    \n",
    "# summary(model, input_size=(3, global_image_height, global_image_width))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "CZNxotzcCwPk"
   },
   "outputs": [],
   "source": [
    "'''INSTANTIATING LOSS CLASS'''\n",
    "\n",
    "triplet_loss = nn.TripletMarginLoss(margin=triplet_loss_margin)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Ja6FZ2ZWCwUL"
   },
   "outputs": [],
   "source": [
    "'''INSTANTIATING OPTIMIZER CLASS'''\n",
    "\n",
    "optimizer = torch.optim.SGD(model.parameters(), lr=learning_rate)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "hIttxOp3CwWx"
   },
   "source": [
    "# TRAINING THE MODEL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "yEadh_rGCwcS"
   },
   "outputs": [],
   "source": [
    "'''INITIALIZING SOME VARIABLES'''\n",
    "\n",
    "iterr = 0\n",
    "\n",
    "# Initializing inspection lists\n",
    "iter_list = []                           # Saves Iterations at which the model has been evaluated\n",
    "train_loss_list = []                     # Saves Train Loss\n",
    "train_acc_list = []                      # Saves Train Accuracy\n",
    "test_loss_list = []                      # Saves Test Loss\n",
    "test_acc_list = []                       # Saves Test Accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 68
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1675,
     "status": "ok",
     "timestamp": 1562564930944,
     "user": {
      "displayName": "Aditya Ahuja",
      "photoUrl": "",
      "userId": "08809216472288903233"
     },
     "user_tz": -330
    },
    "id": "WdtFNjXUCwkU",
    "outputId": "e4ddb3e9-02f7-4b97-e087-733dc512e7fc"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of Iterations     : 760\n",
      "Number of Epochs         : 1\n",
      "Number of Sample-Points  : 50\n"
     ]
    }
   ],
   "source": [
    "'''SETTING CUSTOM PARAMETERS'''\n",
    "\n",
    "# Number of iteration to train the model\n",
    "n_iters = 760\n",
    "\n",
    "# Size at which model is evaluated for later inspection\n",
    "inspect_size = 15                                                 \n",
    "\n",
    "num_epochs = int(n_iters / (len(train_dataset) / batch_size))\n",
    "\n",
    "print(\"Number of Iterations     :\", n_iters)\n",
    "print(\"Number of Epochs         :\", num_epochs)\n",
    "print(\"Number of Sample-Points  :\", int(n_iters/inspect_size))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "EvIPbfEmBw65"
   },
   "outputs": [],
   "source": [
    "'''HELPER FUNCTIONS'''\n",
    "\n",
    "\n",
    "def get_encodings(matrix):\n",
    "\n",
    "    '''\n",
    "    Accepts a Tensor and returns its encoding.\n",
    "    '''\n",
    "    \n",
    "    if torch.cuda.is_available():\n",
    "        matrix = Variable(matrix.cuda())    \n",
    "    else:\n",
    "        matrix = Variable(matrix)\n",
    "    \n",
    "    matrix = matrix.float()\n",
    "    matrix_enc = model(matrix)\n",
    "    \n",
    "    return matrix_enc\n",
    "\n",
    "\n",
    "\n",
    "def return_diff(anchors_enc, positives_enc, negatives_enc):\n",
    " \n",
    "    '''\n",
    "    Accepts the encodings of three tensors.\n",
    "    \n",
    "    Returns d(E1,E2) and d(E1,E3) where d(A,B) is the \n",
    "    Frobenius norm of the vector A-B.\n",
    "    \n",
    "    Returns the result as a pair of numpy arrays.\n",
    "    '''\n",
    "\n",
    "    assert(anchors_enc.shape == positives_enc.shape)\n",
    "    assert(anchors_enc.shape == negatives_enc.shape)\n",
    "    \n",
    "    num = anchors_enc.shape[0]\n",
    "    pos_diff_vec = []\n",
    "    neg_diff_vec = []\n",
    "    \n",
    "    anchors_enc = anchors_enc.cpu().detach().numpy()\n",
    "    positives_enc = positives_enc.cpu().detach().numpy()\n",
    "    negatives_enc = negatives_enc.cpu().detach().numpy()\n",
    "    \n",
    "    for i in range(num):\n",
    "        pos_diff = np.linalg.norm(anchors_enc[i] - positives_enc[i])\n",
    "        neg_diff = np.linalg.norm(anchors_enc[i] - negatives_enc[i])\n",
    "        \n",
    "        pos_diff_vec.append(pos_diff)\n",
    "        neg_diff_vec.append(neg_diff)\n",
    "   \n",
    "    return np.array(pos_diff_vec), np.array(neg_diff_vec)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "LjmF-RxtCwmV",
    "outputId": "0999c292-90f1-455b-c3a6-48e593072089",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iter 1 Done.\t Loss : 0.14557\n",
      "Iter 2 Done.\t Loss : 0.21168\n",
      "Iter 3 Done.\t Loss : 0.11992\n",
      "Iter 4 Done.\t Loss : 0.04521\n",
      "Iter 5 Done.\t Loss : 0.10324\n",
      "Iter 6 Done.\t Loss : 0.11340\n",
      "Iter 7 Done.\t Loss : 0.10455\n",
      "Iter 8 Done.\t Loss : 0.15453\n",
      "Iter 9 Done.\t Loss : 0.08825\n",
      "Iter 10 Done.\t Loss : 0.08118\n",
      "Iter 11 Done.\t Loss : 0.11669\n",
      "Iter 12 Done.\t Loss : 0.21851\n",
      "Iter 13 Done.\t Loss : 0.16140\n",
      "Iter 14 Done.\t Loss : 0.17823\n",
      "Iter 15 Done.\t Loss : 0.28434\n",
      "Iteration : 15/760 [ 2%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.28434 | Acc: 90%\n",
      "[Test ]\t Loss: 0.07691 | Acc: 97%\n",
      "=========================================================\n",
      "Iter 16 Done.\t Loss : 0.06771\n",
      "Iter 17 Done.\t Loss : 0.09421\n",
      "Iter 18 Done.\t Loss : 0.14141\n",
      "Iter 19 Done.\t Loss : 0.05604\n",
      "Iter 20 Done.\t Loss : 0.12675\n",
      "Iter 21 Done.\t Loss : 0.08704\n",
      "Iter 22 Done.\t Loss : 0.12934\n",
      "Iter 23 Done.\t Loss : 0.10722\n",
      "Iter 24 Done.\t Loss : 0.12542\n",
      "Iter 25 Done.\t Loss : 0.04365\n",
      "Iter 26 Done.\t Loss : 0.12651\n",
      "Iter 27 Done.\t Loss : 0.12307\n",
      "Iter 28 Done.\t Loss : 0.08618\n",
      "Iter 29 Done.\t Loss : 0.05183\n",
      "Iter 30 Done.\t Loss : 0.03513\n",
      "Iteration : 30/760 [ 4%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.03513 | Acc: 100%\n",
      "[Test ]\t Loss: 0.05231 | Acc: 98%\n",
      "=========================================================\n",
      "Iter 31 Done.\t Loss : 0.07757\n",
      "Iter 32 Done.\t Loss : 0.05548\n",
      "Iter 33 Done.\t Loss : 0.13307\n",
      "Iter 34 Done.\t Loss : 0.06423\n",
      "Iter 35 Done.\t Loss : 0.09676\n",
      "Iter 36 Done.\t Loss : 0.03183\n",
      "Iter 37 Done.\t Loss : 0.08352\n",
      "Iter 38 Done.\t Loss : 0.05199\n",
      "Iter 39 Done.\t Loss : 0.04702\n",
      "Iter 40 Done.\t Loss : 0.06836\n",
      "Iter 41 Done.\t Loss : 0.08562\n",
      "Iter 42 Done.\t Loss : 0.04121\n",
      "Iter 43 Done.\t Loss : 0.02137\n",
      "Iter 44 Done.\t Loss : 0.02705\n",
      "Iter 45 Done.\t Loss : 0.04771\n",
      "Iteration : 45/760 [ 6%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.04771 | Acc: 99%\n",
      "[Test ]\t Loss: 0.05005 | Acc: 98%\n",
      "=========================================================\n",
      "Iter 46 Done.\t Loss : 0.06888\n",
      "Iter 47 Done.\t Loss : 0.05880\n",
      "Iter 48 Done.\t Loss : 0.02857\n",
      "Iter 49 Done.\t Loss : 0.03597\n",
      "Iter 50 Done.\t Loss : 0.06454\n",
      "Iter 51 Done.\t Loss : 0.08365\n",
      "Iter 52 Done.\t Loss : 0.03410\n",
      "Iter 53 Done.\t Loss : 0.05433\n",
      "Iter 54 Done.\t Loss : 0.05341\n",
      "Iter 55 Done.\t Loss : 0.10030\n",
      "Iter 56 Done.\t Loss : 0.05245\n",
      "Iter 57 Done.\t Loss : 0.08076\n",
      "Iter 58 Done.\t Loss : 0.10843\n",
      "Iter 59 Done.\t Loss : 0.11835\n",
      "Iter 60 Done.\t Loss : 0.03176\n",
      "Iteration : 60/760 [ 8%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.03176 | Acc: 99%\n",
      "[Test ]\t Loss: 0.12153 | Acc: 95%\n",
      "=========================================================\n",
      "Iter 61 Done.\t Loss : 0.03595\n",
      "Iter 62 Done.\t Loss : 0.01089\n",
      "Iter 63 Done.\t Loss : 0.02019\n",
      "Iter 64 Done.\t Loss : 0.04376\n",
      "Iter 65 Done.\t Loss : 0.04817\n",
      "Iter 66 Done.\t Loss : 0.00405\n",
      "Iter 67 Done.\t Loss : 0.03448\n",
      "Iter 68 Done.\t Loss : 0.03800\n",
      "Iter 69 Done.\t Loss : 0.03190\n",
      "Iter 70 Done.\t Loss : 0.11027\n",
      "Iter 71 Done.\t Loss : 0.04101\n",
      "Iter 72 Done.\t Loss : 0.02629\n",
      "Iter 73 Done.\t Loss : 0.06654\n",
      "Iter 74 Done.\t Loss : 0.01953\n",
      "Iter 75 Done.\t Loss : 0.03884\n",
      "Iteration : 75/760 [10%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.03884 | Acc: 99%\n",
      "[Test ]\t Loss: 0.03197 | Acc: 99%\n",
      "=========================================================\n",
      "Iter 76 Done.\t Loss : 0.06406\n",
      "Iter 77 Done.\t Loss : 0.09432\n",
      "Iter 78 Done.\t Loss : 0.03979\n",
      "Iter 79 Done.\t Loss : 0.05243\n",
      "Iter 80 Done.\t Loss : 0.02018\n",
      "Iter 81 Done.\t Loss : 0.00979\n",
      "Iter 82 Done.\t Loss : 0.02009\n",
      "Iter 83 Done.\t Loss : 0.01721\n",
      "Iter 84 Done.\t Loss : 0.00950\n",
      "Iter 85 Done.\t Loss : 0.01423\n",
      "Iter 86 Done.\t Loss : 0.06109\n",
      "Iter 87 Done.\t Loss : 0.00898\n",
      "Iter 88 Done.\t Loss : 0.00554\n",
      "Iter 89 Done.\t Loss : 0.02641\n",
      "Iter 90 Done.\t Loss : 0.03472\n",
      "Iteration : 90/760 [12%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.03472 | Acc: 99%\n",
      "[Test ]\t Loss: 0.05173 | Acc: 98%\n",
      "=========================================================\n",
      "Iter 91 Done.\t Loss : 0.05283\n",
      "Iter 92 Done.\t Loss : 0.02987\n",
      "Iter 93 Done.\t Loss : 0.00817\n",
      "Iter 94 Done.\t Loss : 0.00932\n",
      "Iter 95 Done.\t Loss : 0.03743\n",
      "Iter 96 Done.\t Loss : 0.02809\n",
      "Iter 97 Done.\t Loss : 0.00650\n",
      "Iter 98 Done.\t Loss : 0.02005\n",
      "Iter 99 Done.\t Loss : 0.02829\n",
      "Iter 100 Done.\t Loss : 0.01798\n",
      "Iter 101 Done.\t Loss : 0.05279\n",
      "Iter 102 Done.\t Loss : 0.05580\n",
      "Iter 103 Done.\t Loss : 0.06461\n",
      "Iter 104 Done.\t Loss : 0.01913\n",
      "Iter 105 Done.\t Loss : 0.00561\n",
      "Iteration : 105/760 [14%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00561 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00000 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 106 Done.\t Loss : 0.03624\n",
      "Iter 107 Done.\t Loss : 0.02400\n",
      "Iter 108 Done.\t Loss : 0.01897\n",
      "Iter 109 Done.\t Loss : 0.03626\n",
      "Iter 110 Done.\t Loss : 0.01286\n",
      "Iter 111 Done.\t Loss : 0.01209\n",
      "Iter 112 Done.\t Loss : 0.01007\n",
      "Iter 113 Done.\t Loss : 0.03286\n",
      "Iter 114 Done.\t Loss : 0.00000\n",
      "Iter 115 Done.\t Loss : 0.02470\n",
      "Iter 116 Done.\t Loss : 0.04050\n",
      "Iter 117 Done.\t Loss : 0.01131\n",
      "Iter 118 Done.\t Loss : 0.02468\n",
      "Iter 119 Done.\t Loss : 0.01268\n",
      "Iter 120 Done.\t Loss : 0.00000\n",
      "Iteration : 120/760 [16%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.01101 | Acc: 99%\n",
      "=========================================================\n",
      "Iter 121 Done.\t Loss : 0.03332\n",
      "Iter 122 Done.\t Loss : 0.01218\n",
      "Iter 123 Done.\t Loss : 0.00320\n",
      "Iter 124 Done.\t Loss : 0.03479\n",
      "Iter 125 Done.\t Loss : 0.00229\n",
      "Iter 126 Done.\t Loss : 0.00832\n",
      "Iter 127 Done.\t Loss : 0.03550\n",
      "Iter 128 Done.\t Loss : 0.01428\n",
      "Iter 129 Done.\t Loss : 0.03841\n",
      "Iter 130 Done.\t Loss : 0.01029\n",
      "Iter 131 Done.\t Loss : 0.06962\n",
      "Iter 132 Done.\t Loss : 0.00760\n",
      "Iter 133 Done.\t Loss : 0.05042\n",
      "Iter 134 Done.\t Loss : 0.06471\n",
      "Iter 135 Done.\t Loss : 0.00365\n",
      "Iteration : 135/760 [18%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00365 | Acc: 100%\n",
      "[Test ]\t Loss: 0.02676 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 136 Done.\t Loss : 0.00676\n",
      "Iter 137 Done.\t Loss : 0.01500\n",
      "Iter 138 Done.\t Loss : 0.02154\n",
      "Iter 139 Done.\t Loss : 0.06095\n",
      "Iter 140 Done.\t Loss : 0.02328\n",
      "Iter 141 Done.\t Loss : 0.08534\n",
      "Iter 142 Done.\t Loss : 0.00895\n",
      "Iter 143 Done.\t Loss : 0.01225\n",
      "Iter 144 Done.\t Loss : 0.01281\n",
      "Iter 145 Done.\t Loss : 0.01640\n",
      "Iter 146 Done.\t Loss : 0.01244\n",
      "Iter 147 Done.\t Loss : 0.02426\n",
      "Iter 148 Done.\t Loss : 0.00125\n",
      "Iter 149 Done.\t Loss : 0.00000\n",
      "Iter 150 Done.\t Loss : 0.00000\n",
      "Iteration : 150/760 [20%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.01472 | Acc: 99%\n",
      "=========================================================\n",
      "Iter 151 Done.\t Loss : 0.01323\n",
      "Iter 152 Done.\t Loss : 0.02014\n",
      "Iter 153 Done.\t Loss : 0.04379\n",
      "Iter 154 Done.\t Loss : 0.05184\n",
      "Iter 155 Done.\t Loss : 0.03519\n",
      "Iter 156 Done.\t Loss : 0.05188\n",
      "Iter 157 Done.\t Loss : 0.02880\n",
      "Iter 158 Done.\t Loss : 0.05233\n",
      "Iter 159 Done.\t Loss : 0.00000\n",
      "Iter 160 Done.\t Loss : 0.00189\n",
      "Iter 161 Done.\t Loss : 0.04018\n",
      "Iter 162 Done.\t Loss : 0.00448\n",
      "Iter 163 Done.\t Loss : 0.01547\n",
      "Iter 164 Done.\t Loss : 0.02518\n",
      "Iter 165 Done.\t Loss : 0.00000\n",
      "Iteration : 165/760 [22%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.03242 | Acc: 99%\n",
      "=========================================================\n",
      "Iter 166 Done.\t Loss : 0.01447\n",
      "Iter 167 Done.\t Loss : 0.00000\n",
      "Iter 168 Done.\t Loss : 0.00224\n",
      "Iter 169 Done.\t Loss : 0.02638\n",
      "Iter 170 Done.\t Loss : 0.01025\n",
      "Iter 171 Done.\t Loss : 0.00472\n",
      "Iter 172 Done.\t Loss : 0.01778\n",
      "Iter 173 Done.\t Loss : 0.05179\n",
      "Iter 174 Done.\t Loss : 0.01549\n",
      "Iter 175 Done.\t Loss : 0.00296\n",
      "Iter 176 Done.\t Loss : 0.01939\n",
      "Iter 177 Done.\t Loss : 0.01740\n",
      "Iter 178 Done.\t Loss : 0.02227\n",
      "Iter 179 Done.\t Loss : 0.00829\n",
      "Iter 180 Done.\t Loss : 0.02454\n",
      "Iteration : 180/760 [24%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.02454 | Acc: 99%\n",
      "[Test ]\t Loss: 0.01427 | Acc: 99%\n",
      "=========================================================\n",
      "Iter 181 Done.\t Loss : 0.02462\n",
      "Iter 182 Done.\t Loss : 0.01769\n",
      "Iter 183 Done.\t Loss : 0.02716\n",
      "Iter 184 Done.\t Loss : 0.01013\n",
      "Iter 185 Done.\t Loss : 0.06035\n",
      "Iter 186 Done.\t Loss : 0.01727\n",
      "Iter 187 Done.\t Loss : 0.00557\n",
      "Iter 188 Done.\t Loss : 0.04214\n",
      "Iter 189 Done.\t Loss : 0.01211\n",
      "Iter 190 Done.\t Loss : 0.04358\n",
      "Iter 191 Done.\t Loss : 0.01934\n",
      "Iter 192 Done.\t Loss : 0.00224\n",
      "Iter 193 Done.\t Loss : 0.00000\n",
      "Iter 194 Done.\t Loss : 0.00633\n",
      "Iter 195 Done.\t Loss : 0.01450\n",
      "Iteration : 195/760 [26%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.01450 | Acc: 99%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[Test ]\t Loss: 0.01223 | Acc: 99%\n",
      "=========================================================\n",
      "Iter 196 Done.\t Loss : 0.00000\n",
      "Iter 197 Done.\t Loss : 0.03378\n",
      "Iter 198 Done.\t Loss : 0.00535\n",
      "Iter 199 Done.\t Loss : 0.00378\n",
      "Iter 200 Done.\t Loss : 0.01261\n",
      "Iter 201 Done.\t Loss : 0.01187\n",
      "Iter 202 Done.\t Loss : 0.02785\n",
      "Iter 203 Done.\t Loss : 0.04122\n",
      "Iter 204 Done.\t Loss : 0.00399\n",
      "Iter 205 Done.\t Loss : 0.00705\n",
      "Iter 206 Done.\t Loss : 0.01873\n",
      "Iter 207 Done.\t Loss : 0.00104\n",
      "Iter 208 Done.\t Loss : 0.03819\n",
      "Iter 209 Done.\t Loss : 0.00609\n",
      "Iter 210 Done.\t Loss : 0.00831\n",
      "Iteration : 210/760 [28%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00831 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00521 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 211 Done.\t Loss : 0.01715\n",
      "Iter 212 Done.\t Loss : 0.00000\n",
      "Iter 213 Done.\t Loss : 0.02605\n",
      "Iter 214 Done.\t Loss : 0.01251\n",
      "Iter 215 Done.\t Loss : 0.00158\n",
      "Iter 216 Done.\t Loss : 0.01230\n",
      "Iter 217 Done.\t Loss : 0.00266\n",
      "Iter 218 Done.\t Loss : 0.01261\n",
      "Iter 219 Done.\t Loss : 0.00000\n",
      "Iter 220 Done.\t Loss : 0.02017\n",
      "Iter 221 Done.\t Loss : 0.00841\n",
      "Iter 222 Done.\t Loss : 0.00171\n",
      "Iter 223 Done.\t Loss : 0.00420\n",
      "Iter 224 Done.\t Loss : 0.02727\n",
      "Iter 225 Done.\t Loss : 0.00000\n",
      "Iteration : 225/760 [30%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00763 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 226 Done.\t Loss : 0.00706\n",
      "Iter 227 Done.\t Loss : 0.00771\n",
      "Iter 228 Done.\t Loss : 0.00562\n",
      "Iter 229 Done.\t Loss : 0.00979\n",
      "Iter 230 Done.\t Loss : 0.01699\n",
      "Iter 231 Done.\t Loss : 0.00000\n",
      "Iter 232 Done.\t Loss : 0.01890\n",
      "Iter 233 Done.\t Loss : 0.00376\n",
      "Iter 234 Done.\t Loss : 0.00079\n",
      "Iter 235 Done.\t Loss : 0.01014\n",
      "Iter 236 Done.\t Loss : 0.02639\n",
      "Iter 237 Done.\t Loss : 0.01799\n",
      "Iter 238 Done.\t Loss : 0.00395\n",
      "Iter 239 Done.\t Loss : 0.00679\n",
      "Iter 240 Done.\t Loss : 0.00000\n",
      "Iteration : 240/760 [32%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.02232 | Acc: 99%\n",
      "=========================================================\n",
      "Iter 241 Done.\t Loss : 0.01362\n",
      "Iter 242 Done.\t Loss : 0.00000\n",
      "Iter 243 Done.\t Loss : 0.01116\n",
      "Iter 244 Done.\t Loss : 0.01784\n",
      "Iter 245 Done.\t Loss : 0.00000\n",
      "Iter 246 Done.\t Loss : 0.04178\n",
      "Iter 247 Done.\t Loss : 0.00000\n",
      "Iter 248 Done.\t Loss : 0.00839\n",
      "Iter 249 Done.\t Loss : 0.00661\n",
      "Iter 250 Done.\t Loss : 0.01715\n",
      "Iter 251 Done.\t Loss : 0.01412\n",
      "Iter 252 Done.\t Loss : 0.00103\n",
      "Iter 253 Done.\t Loss : 0.00390\n",
      "Iter 254 Done.\t Loss : 0.00000\n",
      "Iter 255 Done.\t Loss : 0.02406\n",
      "Iteration : 255/760 [34%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.02406 | Acc: 99%\n",
      "[Test ]\t Loss: 0.00398 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 256 Done.\t Loss : 0.00000\n",
      "Iter 257 Done.\t Loss : 0.00586\n",
      "Iter 258 Done.\t Loss : 0.00414\n",
      "Iter 259 Done.\t Loss : 0.00170\n",
      "Iter 260 Done.\t Loss : 0.00000\n",
      "Iter 261 Done.\t Loss : 0.01619\n",
      "Iter 262 Done.\t Loss : 0.01772\n",
      "Iter 263 Done.\t Loss : 0.00000\n",
      "Iter 264 Done.\t Loss : 0.00000\n",
      "Iter 265 Done.\t Loss : 0.00600\n",
      "Iter 266 Done.\t Loss : 0.01287\n",
      "Iter 267 Done.\t Loss : 0.00000\n",
      "Iter 268 Done.\t Loss : 0.00000\n",
      "Iter 269 Done.\t Loss : 0.00714\n",
      "Iter 270 Done.\t Loss : 0.00993\n",
      "Iteration : 270/760 [36%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00993 | Acc: 100%\n",
      "[Test ]\t Loss: 0.01818 | Acc: 99%\n",
      "=========================================================\n",
      "Iter 271 Done.\t Loss : 0.00000\n",
      "Iter 272 Done.\t Loss : 0.02285\n",
      "Iter 273 Done.\t Loss : 0.00122\n",
      "Iter 274 Done.\t Loss : 0.01376\n",
      "Iter 275 Done.\t Loss : 0.01596\n",
      "Iter 276 Done.\t Loss : 0.00746\n",
      "Iter 277 Done.\t Loss : 0.00000\n",
      "Iter 278 Done.\t Loss : 0.00659\n",
      "Iter 279 Done.\t Loss : 0.02471\n",
      "Iter 280 Done.\t Loss : 0.01525\n",
      "Iter 281 Done.\t Loss : 0.00810\n",
      "Iter 282 Done.\t Loss : 0.00000\n",
      "Iter 283 Done.\t Loss : 0.00000\n",
      "Iter 284 Done.\t Loss : 0.01654\n",
      "Iter 285 Done.\t Loss : 0.01271\n",
      "Iteration : 285/760 [38%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.01271 | Acc: 100%\n",
      "[Test ]\t Loss: 0.01432 | Acc: 99%\n",
      "=========================================================\n",
      "Iter 286 Done.\t Loss : 0.00000\n",
      "Iter 287 Done.\t Loss : 0.02837\n",
      "Iter 288 Done.\t Loss : 0.00000\n",
      "Iter 289 Done.\t Loss : 0.02238\n",
      "Iter 290 Done.\t Loss : 0.00000\n",
      "Iter 291 Done.\t Loss : 0.01761\n",
      "Iter 292 Done.\t Loss : 0.00174\n",
      "Iter 293 Done.\t Loss : 0.01702\n",
      "Iter 294 Done.\t Loss : 0.01042\n",
      "Iter 295 Done.\t Loss : 0.00000\n",
      "Iter 296 Done.\t Loss : 0.02261\n",
      "Iter 297 Done.\t Loss : 0.01427\n",
      "Iter 298 Done.\t Loss : 0.00000\n",
      "Iter 299 Done.\t Loss : 0.02016\n",
      "Iter 300 Done.\t Loss : 0.00000\n",
      "Iteration : 300/760 [39%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00000 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 301 Done.\t Loss : 0.00000\n",
      "Iter 302 Done.\t Loss : 0.00392\n",
      "Iter 303 Done.\t Loss : 0.00802\n",
      "Iter 304 Done.\t Loss : 0.00076\n",
      "Iter 305 Done.\t Loss : 0.00862\n",
      "Iter 306 Done.\t Loss : 0.03334\n",
      "Iter 307 Done.\t Loss : 0.01139\n",
      "Iter 308 Done.\t Loss : 0.00833\n",
      "Iter 309 Done.\t Loss : 0.01100\n",
      "Iter 310 Done.\t Loss : 0.01255\n",
      "Iter 311 Done.\t Loss : 0.01046\n",
      "Iter 312 Done.\t Loss : 0.00000\n",
      "Iter 313 Done.\t Loss : 0.00000\n",
      "Iter 314 Done.\t Loss : 0.00000\n",
      "Iter 315 Done.\t Loss : 0.00000\n",
      "Iteration : 315/760 [41%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00000 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 316 Done.\t Loss : 0.00000\n",
      "Iter 317 Done.\t Loss : 0.00528\n",
      "Iter 318 Done.\t Loss : 0.02014\n",
      "Iter 319 Done.\t Loss : 0.00681\n",
      "Iter 320 Done.\t Loss : 0.01204\n",
      "Iter 321 Done.\t Loss : 0.00655\n",
      "Iter 322 Done.\t Loss : 0.00585\n",
      "Iter 323 Done.\t Loss : 0.00454\n",
      "Iter 324 Done.\t Loss : 0.02109\n",
      "Iter 325 Done.\t Loss : 0.03771\n",
      "Iter 326 Done.\t Loss : 0.00115\n",
      "Iter 327 Done.\t Loss : 0.01136\n",
      "Iter 328 Done.\t Loss : 0.00000\n",
      "Iter 329 Done.\t Loss : 0.00007\n",
      "Iter 330 Done.\t Loss : 0.00000\n",
      "Iteration : 330/760 [43%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00051 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 331 Done.\t Loss : 0.00688\n",
      "Iter 332 Done.\t Loss : 0.00527\n",
      "Iter 333 Done.\t Loss : 0.00044\n",
      "Iter 334 Done.\t Loss : 0.00697\n",
      "Iter 335 Done.\t Loss : 0.00607\n",
      "Iter 336 Done.\t Loss : 0.00653\n",
      "Iter 337 Done.\t Loss : 0.00000\n",
      "Iter 338 Done.\t Loss : 0.00000\n",
      "Iter 339 Done.\t Loss : 0.00208\n",
      "Iter 340 Done.\t Loss : 0.00355\n",
      "Iter 341 Done.\t Loss : 0.01135\n",
      "Iter 342 Done.\t Loss : 0.00000\n",
      "Iter 343 Done.\t Loss : 0.00000\n",
      "Iter 344 Done.\t Loss : 0.01045\n",
      "Iter 345 Done.\t Loss : 0.00000\n",
      "Iteration : 345/760 [45%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00057 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 346 Done.\t Loss : 0.00244\n",
      "Iter 347 Done.\t Loss : 0.00235\n",
      "Iter 348 Done.\t Loss : 0.00300\n",
      "Iter 349 Done.\t Loss : 0.00000\n",
      "Iter 350 Done.\t Loss : 0.02471\n",
      "Iter 351 Done.\t Loss : 0.02722\n",
      "Iter 352 Done.\t Loss : 0.00396\n",
      "Iter 353 Done.\t Loss : 0.00000\n",
      "Iter 354 Done.\t Loss : 0.00000\n",
      "Iter 355 Done.\t Loss : 0.00978\n",
      "Iter 356 Done.\t Loss : 0.00000\n",
      "Iter 357 Done.\t Loss : 0.00067\n",
      "Iter 358 Done.\t Loss : 0.01214\n",
      "Iter 359 Done.\t Loss : 0.00234\n",
      "Iter 360 Done.\t Loss : 0.00000\n",
      "Iteration : 360/760 [47%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00000 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 361 Done.\t Loss : 0.00000\n",
      "Iter 362 Done.\t Loss : 0.00199\n",
      "Iter 363 Done.\t Loss : 0.00000\n",
      "Iter 364 Done.\t Loss : 0.00535\n",
      "Iter 365 Done.\t Loss : 0.00817\n",
      "Iter 366 Done.\t Loss : 0.00000\n",
      "Iter 367 Done.\t Loss : 0.00000\n",
      "Iter 368 Done.\t Loss : 0.00000\n",
      "Iter 369 Done.\t Loss : 0.00000\n",
      "Iter 370 Done.\t Loss : 0.00000\n",
      "Iter 371 Done.\t Loss : 0.00000\n",
      "Iter 372 Done.\t Loss : 0.00132\n",
      "Iter 373 Done.\t Loss : 0.00000\n",
      "Iter 374 Done.\t Loss : 0.00998\n",
      "Iter 375 Done.\t Loss : 0.01437\n",
      "Iteration : 375/760 [49%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.01437 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00000 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 376 Done.\t Loss : 0.00699\n",
      "Iter 377 Done.\t Loss : 0.00335\n",
      "Iter 378 Done.\t Loss : 0.00430\n",
      "Iter 379 Done.\t Loss : 0.00000\n",
      "Iter 380 Done.\t Loss : 0.00985\n",
      "Iter 381 Done.\t Loss : 0.02798\n",
      "Iter 382 Done.\t Loss : 0.00421\n",
      "Iter 383 Done.\t Loss : 0.01579\n",
      "Iter 384 Done.\t Loss : 0.00000\n",
      "Iter 385 Done.\t Loss : 0.00546\n",
      "Iter 386 Done.\t Loss : 0.01711\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iter 387 Done.\t Loss : 0.00000\n",
      "Iter 388 Done.\t Loss : 0.00452\n",
      "Iter 389 Done.\t Loss : 0.01316\n",
      "Iter 390 Done.\t Loss : 0.00000\n",
      "Iteration : 390/760 [51%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00214 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 391 Done.\t Loss : 0.00000\n",
      "Iter 392 Done.\t Loss : 0.01824\n",
      "Iter 393 Done.\t Loss : 0.00217\n",
      "Iter 394 Done.\t Loss : 0.00557\n",
      "Iter 395 Done.\t Loss : 0.00000\n",
      "Iter 396 Done.\t Loss : 0.00000\n",
      "Iter 397 Done.\t Loss : 0.00000\n",
      "Iter 398 Done.\t Loss : 0.00245\n",
      "Iter 399 Done.\t Loss : 0.01419\n",
      "Iter 400 Done.\t Loss : 0.00000\n",
      "Iter 401 Done.\t Loss : 0.01299\n",
      "Iter 402 Done.\t Loss : 0.00752\n",
      "Iter 403 Done.\t Loss : 0.00000\n",
      "Iter 404 Done.\t Loss : 0.03586\n",
      "Iter 405 Done.\t Loss : 0.00000\n",
      "Iteration : 405/760 [53%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.01417 | Acc: 99%\n",
      "=========================================================\n",
      "Iter 406 Done.\t Loss : 0.00000\n",
      "Iter 407 Done.\t Loss : 0.01228\n",
      "Iter 408 Done.\t Loss : 0.00919\n",
      "Iter 409 Done.\t Loss : 0.00000\n",
      "Iter 410 Done.\t Loss : 0.00000\n",
      "Iter 411 Done.\t Loss : 0.00000\n",
      "Iter 412 Done.\t Loss : 0.00769\n",
      "Iter 413 Done.\t Loss : 0.00000\n",
      "Iter 414 Done.\t Loss : 0.00000\n",
      "Iter 415 Done.\t Loss : 0.00332\n",
      "Iter 416 Done.\t Loss : 0.00367\n",
      "Iter 417 Done.\t Loss : 0.01834\n",
      "Iter 418 Done.\t Loss : 0.00000\n",
      "Iter 419 Done.\t Loss : 0.01011\n",
      "Iter 420 Done.\t Loss : 0.01996\n",
      "Iteration : 420/760 [55%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.01996 | Acc: 99%\n",
      "[Test ]\t Loss: 0.00000 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 421 Done.\t Loss : 0.00000\n",
      "Iter 422 Done.\t Loss : 0.00717\n",
      "Iter 423 Done.\t Loss : 0.00642\n",
      "Iter 424 Done.\t Loss : 0.00000\n",
      "Iter 425 Done.\t Loss : 0.00245\n",
      "Iter 426 Done.\t Loss : 0.00000\n",
      "Iter 427 Done.\t Loss : 0.00000\n",
      "Iter 428 Done.\t Loss : 0.00000\n",
      "Iter 429 Done.\t Loss : 0.00000\n",
      "Iter 430 Done.\t Loss : 0.01025\n",
      "Iter 431 Done.\t Loss : 0.00054\n",
      "Iter 432 Done.\t Loss : 0.00000\n",
      "Iter 433 Done.\t Loss : 0.00388\n",
      "Iter 434 Done.\t Loss : 0.00000\n",
      "Iter 435 Done.\t Loss : 0.01201\n",
      "Iteration : 435/760 [57%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.01201 | Acc: 99%\n",
      "[Test ]\t Loss: 0.00267 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 436 Done.\t Loss : 0.00000\n",
      "Iter 437 Done.\t Loss : 0.00395\n",
      "Iter 438 Done.\t Loss : 0.00132\n",
      "Iter 439 Done.\t Loss : 0.00000\n",
      "Iter 440 Done.\t Loss : 0.00670\n",
      "Iter 441 Done.\t Loss : 0.01087\n",
      "Iter 442 Done.\t Loss : 0.00354\n",
      "Iter 443 Done.\t Loss : 0.00000\n",
      "Iter 444 Done.\t Loss : 0.01001\n",
      "Iter 445 Done.\t Loss : 0.00000\n",
      "Iter 446 Done.\t Loss : 0.01139\n",
      "Iter 447 Done.\t Loss : 0.00000\n",
      "Iter 448 Done.\t Loss : 0.00168\n",
      "Iter 449 Done.\t Loss : 0.00909\n",
      "Iter 450 Done.\t Loss : 0.00782\n",
      "Iteration : 450/760 [59%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00782 | Acc: 100%\n",
      "[Test ]\t Loss: 0.01378 | Acc: 99%\n",
      "=========================================================\n",
      "Iter 451 Done.\t Loss : 0.00235\n",
      "Iter 452 Done.\t Loss : 0.00000\n",
      "Iter 453 Done.\t Loss : 0.00000\n",
      "Iter 454 Done.\t Loss : 0.00681\n",
      "Iter 455 Done.\t Loss : 0.00000\n",
      "Iter 456 Done.\t Loss : 0.00550\n",
      "Iter 457 Done.\t Loss : 0.00490\n",
      "Iter 458 Done.\t Loss : 0.00000\n",
      "Iter 459 Done.\t Loss : 0.00617\n",
      "Iter 460 Done.\t Loss : 0.00767\n",
      "Iter 461 Done.\t Loss : 0.00106\n",
      "Iter 462 Done.\t Loss : 0.00096\n",
      "Iter 463 Done.\t Loss : 0.00000\n",
      "Iter 464 Done.\t Loss : 0.00000\n",
      "Iter 465 Done.\t Loss : 0.01276\n",
      "Iteration : 465/760 [61%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.01276 | Acc: 99%\n",
      "[Test ]\t Loss: 0.00303 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 466 Done.\t Loss : 0.00202\n",
      "Iter 467 Done.\t Loss : 0.00034\n",
      "Iter 468 Done.\t Loss : 0.01257\n",
      "Iter 469 Done.\t Loss : 0.00000\n",
      "Iter 470 Done.\t Loss : 0.00000\n",
      "Iter 471 Done.\t Loss : 0.00000\n",
      "Iter 472 Done.\t Loss : 0.00000\n",
      "Iter 473 Done.\t Loss : 0.00786\n",
      "Iter 474 Done.\t Loss : 0.00555\n",
      "Iter 475 Done.\t Loss : 0.00185\n",
      "Iter 476 Done.\t Loss : 0.00898\n",
      "Iter 477 Done.\t Loss : 0.01247\n",
      "Iter 478 Done.\t Loss : 0.00000\n",
      "Iter 479 Done.\t Loss : 0.00000\n",
      "Iter 480 Done.\t Loss : 0.00000\n",
      "Iteration : 480/760 [63%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00385 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 481 Done.\t Loss : 0.00000\n",
      "Iter 482 Done.\t Loss : 0.00166\n",
      "Iter 483 Done.\t Loss : 0.00000\n",
      "Iter 484 Done.\t Loss : 0.00000\n",
      "Iter 485 Done.\t Loss : 0.00000\n",
      "Iter 486 Done.\t Loss : 0.00834\n",
      "Iter 487 Done.\t Loss : 0.00000\n",
      "Iter 488 Done.\t Loss : 0.00354\n",
      "Iter 489 Done.\t Loss : 0.01305\n",
      "Iter 490 Done.\t Loss : 0.00084\n",
      "Iter 491 Done.\t Loss : 0.00000\n",
      "Iter 492 Done.\t Loss : 0.00000\n",
      "Iter 493 Done.\t Loss : 0.00330\n",
      "Iter 494 Done.\t Loss : 0.00191\n",
      "Iter 495 Done.\t Loss : 0.00000\n",
      "Iteration : 495/760 [65%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00000 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 496 Done.\t Loss : 0.00037\n",
      "Iter 497 Done.\t Loss : 0.00000\n",
      "Iter 498 Done.\t Loss : 0.00641\n",
      "Iter 499 Done.\t Loss : 0.00000\n",
      "Iter 500 Done.\t Loss : 0.00273\n",
      "Iter 501 Done.\t Loss : 0.00591\n",
      "Iter 502 Done.\t Loss : 0.00000\n",
      "Iter 503 Done.\t Loss : 0.00000\n",
      "Iter 504 Done.\t Loss : 0.00000\n",
      "Iter 505 Done.\t Loss : 0.00537\n",
      "Iter 506 Done.\t Loss : 0.00000\n",
      "Iter 507 Done.\t Loss : 0.01488\n",
      "Iter 508 Done.\t Loss : 0.00000\n",
      "Iter 509 Done.\t Loss : 0.00152\n",
      "Iter 510 Done.\t Loss : 0.00000\n",
      "Iteration : 510/760 [67%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00004 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 511 Done.\t Loss : 0.00000\n",
      "Iter 512 Done.\t Loss : 0.01157\n",
      "Iter 513 Done.\t Loss : 0.00000\n",
      "Iter 514 Done.\t Loss : 0.00679\n",
      "Iter 515 Done.\t Loss : 0.00000\n",
      "Iter 516 Done.\t Loss : 0.00000\n",
      "Iter 517 Done.\t Loss : 0.01079\n",
      "Iter 518 Done.\t Loss : 0.00000\n",
      "Iter 519 Done.\t Loss : 0.00134\n",
      "Iter 520 Done.\t Loss : 0.00045\n",
      "Iter 521 Done.\t Loss : 0.00378\n",
      "Iter 522 Done.\t Loss : 0.00031\n",
      "Iter 523 Done.\t Loss : 0.00259\n",
      "Iter 524 Done.\t Loss : 0.00000\n",
      "Iter 525 Done.\t Loss : 0.00000\n",
      "Iteration : 525/760 [69%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00198 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 526 Done.\t Loss : 0.02605\n",
      "Iter 527 Done.\t Loss : 0.00000\n",
      "Iter 528 Done.\t Loss : 0.00000\n",
      "Iter 529 Done.\t Loss : 0.00000\n",
      "Iter 530 Done.\t Loss : 0.00293\n",
      "Iter 531 Done.\t Loss : 0.00496\n",
      "Iter 532 Done.\t Loss : 0.00027\n",
      "Iter 533 Done.\t Loss : 0.00000\n",
      "Iter 534 Done.\t Loss : 0.00815\n",
      "Iter 535 Done.\t Loss : 0.00208\n",
      "Iter 536 Done.\t Loss : 0.00000\n",
      "Iter 537 Done.\t Loss : 0.00000\n",
      "Iter 538 Done.\t Loss : 0.00000\n",
      "Iter 539 Done.\t Loss : 0.00130\n",
      "Iter 540 Done.\t Loss : 0.00000\n",
      "Iteration : 540/760 [71%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00640 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 541 Done.\t Loss : 0.00000\n",
      "Iter 542 Done.\t Loss : 0.00876\n",
      "Iter 543 Done.\t Loss : 0.00000\n",
      "Iter 544 Done.\t Loss : 0.00000\n",
      "Iter 545 Done.\t Loss : 0.00738\n",
      "Iter 546 Done.\t Loss : 0.00000\n",
      "Iter 547 Done.\t Loss : 0.00000\n",
      "Iter 548 Done.\t Loss : 0.00000\n",
      "Iter 549 Done.\t Loss : 0.00000\n",
      "Iter 550 Done.\t Loss : 0.00000\n",
      "Iter 551 Done.\t Loss : 0.00000\n",
      "Iter 552 Done.\t Loss : 0.00000\n",
      "Iter 553 Done.\t Loss : 0.01406\n",
      "Iter 554 Done.\t Loss : 0.00000\n",
      "Iter 555 Done.\t Loss : 0.00000\n",
      "Iteration : 555/760 [73%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.02473 | Acc: 99%\n",
      "=========================================================\n",
      "Iter 556 Done.\t Loss : 0.00000\n",
      "Iter 557 Done.\t Loss : 0.00000\n",
      "Iter 558 Done.\t Loss : 0.00022\n",
      "Iter 559 Done.\t Loss : 0.00748\n",
      "Iter 560 Done.\t Loss : 0.00000\n",
      "Iter 561 Done.\t Loss : 0.00000\n",
      "Iter 562 Done.\t Loss : 0.00381\n",
      "Iter 563 Done.\t Loss : 0.00000\n",
      "Iter 564 Done.\t Loss : 0.00000\n",
      "Iter 565 Done.\t Loss : 0.00380\n",
      "Iter 566 Done.\t Loss : 0.00000\n",
      "Iter 567 Done.\t Loss : 0.00000\n",
      "Iter 568 Done.\t Loss : 0.00000\n",
      "Iter 569 Done.\t Loss : 0.00000\n",
      "Iter 570 Done.\t Loss : 0.00000\n",
      "Iteration : 570/760 [75%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00000 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 571 Done.\t Loss : 0.00000\n",
      "Iter 572 Done.\t Loss : 0.00668\n",
      "Iter 573 Done.\t Loss : 0.00000\n",
      "Iter 574 Done.\t Loss : 0.00000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iter 575 Done.\t Loss : 0.00300\n",
      "Iter 576 Done.\t Loss : 0.00229\n",
      "Iter 577 Done.\t Loss : 0.00000\n",
      "Iter 578 Done.\t Loss : 0.00000\n",
      "Iter 579 Done.\t Loss : 0.00000\n",
      "Iter 580 Done.\t Loss : 0.01718\n",
      "Iter 581 Done.\t Loss : 0.00261\n",
      "Iter 582 Done.\t Loss : 0.00423\n",
      "Iter 583 Done.\t Loss : 0.00000\n",
      "Iter 584 Done.\t Loss : 0.00280\n",
      "Iter 585 Done.\t Loss : 0.00000\n",
      "Iteration : 585/760 [77%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00000 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 586 Done.\t Loss : 0.00000\n",
      "Iter 587 Done.\t Loss : 0.00423\n",
      "Iter 588 Done.\t Loss : 0.00607\n",
      "Iter 589 Done.\t Loss : 0.00196\n",
      "Iter 590 Done.\t Loss : 0.00000\n",
      "Iter 591 Done.\t Loss : 0.00000\n",
      "Iter 592 Done.\t Loss : 0.00000\n",
      "Iter 593 Done.\t Loss : 0.00000\n",
      "Iter 594 Done.\t Loss : 0.00000\n",
      "Iter 595 Done.\t Loss : 0.00109\n",
      "Iter 596 Done.\t Loss : 0.00078\n",
      "Iter 597 Done.\t Loss : 0.00000\n",
      "Iter 598 Done.\t Loss : 0.00247\n",
      "Iter 599 Done.\t Loss : 0.00439\n",
      "Iter 600 Done.\t Loss : 0.00000\n",
      "Iteration : 600/760 [79%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00000 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 601 Done.\t Loss : 0.00000\n",
      "Iter 602 Done.\t Loss : 0.00000\n",
      "Iter 603 Done.\t Loss : 0.00000\n",
      "Iter 604 Done.\t Loss : 0.00223\n",
      "Iter 605 Done.\t Loss : 0.00000\n",
      "Iter 606 Done.\t Loss : 0.00215\n",
      "Iter 607 Done.\t Loss : 0.00146\n",
      "Iter 608 Done.\t Loss : 0.00000\n",
      "Iter 609 Done.\t Loss : 0.00077\n",
      "Iter 610 Done.\t Loss : 0.00000\n",
      "Iter 611 Done.\t Loss : 0.00130\n",
      "Iter 612 Done.\t Loss : 0.00000\n",
      "Iter 613 Done.\t Loss : 0.00592\n",
      "Iter 614 Done.\t Loss : 0.00198\n",
      "Iter 615 Done.\t Loss : 0.00000\n",
      "Iteration : 615/760 [81%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00303 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 616 Done.\t Loss : 0.00000\n",
      "Iter 617 Done.\t Loss : 0.00560\n",
      "Iter 618 Done.\t Loss : 0.00387\n",
      "Iter 619 Done.\t Loss : 0.00000\n",
      "Iter 620 Done.\t Loss : 0.00000\n",
      "Iter 621 Done.\t Loss : 0.00000\n",
      "Iter 622 Done.\t Loss : 0.00000\n",
      "Iter 623 Done.\t Loss : 0.00000\n",
      "Iter 624 Done.\t Loss : 0.00153\n",
      "Iter 625 Done.\t Loss : 0.00273\n",
      "Iter 626 Done.\t Loss : 0.00000\n",
      "Iter 627 Done.\t Loss : 0.00000\n",
      "Iter 628 Done.\t Loss : 0.00000\n",
      "Iter 629 Done.\t Loss : 0.00000\n",
      "Iter 630 Done.\t Loss : 0.00000\n",
      "Iteration : 630/760 [83%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00000 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 631 Done.\t Loss : 0.00668\n",
      "Iter 632 Done.\t Loss : 0.00000\n",
      "Iter 633 Done.\t Loss : 0.00000\n",
      "Iter 634 Done.\t Loss : 0.00000\n",
      "Iter 635 Done.\t Loss : 0.00000\n",
      "Iter 636 Done.\t Loss : 0.00000\n",
      "Iter 637 Done.\t Loss : 0.00040\n",
      "Iter 638 Done.\t Loss : 0.00875\n",
      "Iter 639 Done.\t Loss : 0.00000\n",
      "Iter 640 Done.\t Loss : 0.00000\n",
      "Iter 641 Done.\t Loss : 0.00000\n",
      "Iter 642 Done.\t Loss : 0.00000\n",
      "Iter 643 Done.\t Loss : 0.00000\n",
      "Iter 644 Done.\t Loss : 0.00000\n",
      "Iter 645 Done.\t Loss : 0.00000\n",
      "Iteration : 645/760 [85%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00000 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 646 Done.\t Loss : 0.00000\n",
      "Iter 647 Done.\t Loss : 0.00072\n",
      "Iter 648 Done.\t Loss : 0.00000\n",
      "Iter 649 Done.\t Loss : 0.00000\n",
      "Iter 650 Done.\t Loss : 0.00000\n",
      "Iter 651 Done.\t Loss : 0.00000\n",
      "Iter 652 Done.\t Loss : 0.00228\n",
      "Iter 653 Done.\t Loss : 0.00037\n",
      "Iter 654 Done.\t Loss : 0.00117\n",
      "Iter 655 Done.\t Loss : 0.00000\n",
      "Iter 656 Done.\t Loss : 0.00465\n",
      "Iter 657 Done.\t Loss : 0.00000\n",
      "Iter 658 Done.\t Loss : 0.00000\n",
      "Iter 659 Done.\t Loss : 0.00000\n",
      "Iter 660 Done.\t Loss : 0.00000\n",
      "Iteration : 660/760 [87%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.02409 | Acc: 99%\n",
      "=========================================================\n",
      "Iter 661 Done.\t Loss : 0.00000\n",
      "Iter 662 Done.\t Loss : 0.00000\n",
      "Iter 663 Done.\t Loss : 0.00000\n",
      "Iter 664 Done.\t Loss : 0.00280\n",
      "Iter 665 Done.\t Loss : 0.00000\n",
      "Iter 666 Done.\t Loss : 0.00000\n",
      "Iter 667 Done.\t Loss : 0.00000\n",
      "Iter 668 Done.\t Loss : 0.00000\n",
      "Iter 669 Done.\t Loss : 0.00000\n",
      "Iter 670 Done.\t Loss : 0.01611\n",
      "Iter 671 Done.\t Loss : 0.00000\n",
      "Iter 672 Done.\t Loss : 0.00000\n",
      "Iter 673 Done.\t Loss : 0.00410\n",
      "Iter 674 Done.\t Loss : 0.00000\n",
      "Iter 675 Done.\t Loss : 0.00000\n",
      "Iteration : 675/760 [89%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00387 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 676 Done.\t Loss : 0.00000\n",
      "Iter 677 Done.\t Loss : 0.00000\n",
      "Iter 678 Done.\t Loss : 0.00411\n",
      "Iter 679 Done.\t Loss : 0.00789\n",
      "Iter 680 Done.\t Loss : 0.00000\n",
      "Iter 681 Done.\t Loss : 0.00000\n",
      "Iter 682 Done.\t Loss : 0.00000\n",
      "Iter 683 Done.\t Loss : 0.00000\n",
      "Iter 684 Done.\t Loss : 0.00000\n",
      "Iter 685 Done.\t Loss : 0.00289\n",
      "Iter 686 Done.\t Loss : 0.00000\n",
      "Iter 687 Done.\t Loss : 0.00000\n",
      "Iter 688 Done.\t Loss : 0.00000\n",
      "Iter 689 Done.\t Loss : 0.01624\n",
      "Iter 690 Done.\t Loss : 0.00000\n",
      "Iteration : 690/760 [91%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00000 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 691 Done.\t Loss : 0.00000\n",
      "Iter 692 Done.\t Loss : 0.00000\n",
      "Iter 693 Done.\t Loss : 0.00000\n",
      "Iter 694 Done.\t Loss : 0.00423\n",
      "Iter 695 Done.\t Loss : 0.00000\n",
      "Iter 696 Done.\t Loss : 0.00000\n",
      "Iter 697 Done.\t Loss : 0.00000\n",
      "Iter 698 Done.\t Loss : 0.00000\n",
      "Iter 699 Done.\t Loss : 0.00230\n",
      "Iter 700 Done.\t Loss : 0.00000\n",
      "Iter 701 Done.\t Loss : 0.00000\n",
      "Iter 702 Done.\t Loss : 0.00004\n",
      "Iter 703 Done.\t Loss : 0.00000\n",
      "Iter 704 Done.\t Loss : 0.00053\n",
      "Iter 705 Done.\t Loss : 0.00000\n",
      "Iteration : 705/760 [93%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00000 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 706 Done.\t Loss : 0.00502\n",
      "Iter 707 Done.\t Loss : 0.00000\n",
      "Iter 708 Done.\t Loss : 0.00000\n",
      "Iter 709 Done.\t Loss : 0.00103\n",
      "Iter 710 Done.\t Loss : 0.00000\n",
      "Iter 711 Done.\t Loss : 0.00000\n",
      "Iter 712 Done.\t Loss : 0.00000\n",
      "Iter 713 Done.\t Loss : 0.00695\n",
      "Iter 714 Done.\t Loss : 0.00000\n",
      "Iter 715 Done.\t Loss : 0.00000\n",
      "Iter 716 Done.\t Loss : 0.00000\n",
      "Iter 717 Done.\t Loss : 0.00000\n",
      "Iter 718 Done.\t Loss : 0.00463\n",
      "Iter 719 Done.\t Loss : 0.00000\n",
      "Iter 720 Done.\t Loss : 0.00000\n",
      "Iteration : 720/760 [95%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00853 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 721 Done.\t Loss : 0.00000\n",
      "Iter 722 Done.\t Loss : 0.00953\n",
      "Iter 723 Done.\t Loss : 0.00000\n",
      "Iter 724 Done.\t Loss : 0.00000\n",
      "Iter 725 Done.\t Loss : 0.00214\n",
      "Iter 726 Done.\t Loss : 0.00000\n",
      "Iter 727 Done.\t Loss : 0.00000\n",
      "Iter 728 Done.\t Loss : 0.00000\n",
      "Iter 729 Done.\t Loss : 0.00000\n",
      "Iter 730 Done.\t Loss : 0.00187\n",
      "Iter 731 Done.\t Loss : 0.02375\n",
      "Iter 732 Done.\t Loss : 0.00072\n",
      "Iter 733 Done.\t Loss : 0.00000\n",
      "Iter 734 Done.\t Loss : 0.00176\n",
      "Iter 735 Done.\t Loss : 0.00000\n",
      "Iteration : 735/760 [97%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00000 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00000 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 736 Done.\t Loss : 0.00119\n",
      "Iter 737 Done.\t Loss : 0.00000\n",
      "Iter 738 Done.\t Loss : 0.00000\n",
      "Iter 739 Done.\t Loss : 0.00883\n",
      "Iter 740 Done.\t Loss : 0.00000\n",
      "Iter 741 Done.\t Loss : 0.01717\n",
      "Iter 742 Done.\t Loss : 0.00000\n",
      "Iter 743 Done.\t Loss : 0.00000\n",
      "Iter 744 Done.\t Loss : 0.00000\n",
      "Iter 745 Done.\t Loss : 0.00275\n",
      "Iter 746 Done.\t Loss : 0.00000\n",
      "Iter 747 Done.\t Loss : 0.00000\n",
      "Iter 748 Done.\t Loss : 0.00347\n",
      "Iter 749 Done.\t Loss : 0.00000\n",
      "Iter 750 Done.\t Loss : 0.00615\n",
      "Iteration : 750/760 [99%] \n",
      "---------------------------\n",
      "[Train]\t Loss: 0.00615 | Acc: 100%\n",
      "[Test ]\t Loss: 0.00000 | Acc: 100%\n",
      "=========================================================\n",
      "Iter 751 Done.\t Loss : 0.00000\n",
      "Iter 752 Done.\t Loss : 0.01257\n",
      "Iter 753 Done.\t Loss : 0.00234\n",
      "Iter 754 Done.\t Loss : 0.00000\n",
      "Iter 755 Done.\t Loss : 0.00000\n",
      "Iter 756 Done.\t Loss : 0.00000\n",
      "Iter 757 Done.\t Loss : 0.00000\n",
      "Iter 758 Done.\t Loss : 0.00000\n",
      "\n",
      "Training Done.\n"
     ]
    }
   ],
   "source": [
    "'''TRAINING THE MODEL'''\n",
    "\n",
    "import time\n",
    "time_begin = time.asctime()   # Time when training started\n",
    "\n",
    "init_iters = iterr\n",
    "\n",
    "for epoch in range(num_epochs):\n",
    "    for i, (anchors, positives, negatives) in enumerate(train_loader):\n",
    "        \n",
    "#         if(i == 5): break\n",
    "        \n",
    "        \n",
    "        # Get encodings by forward propogation\n",
    "        anchors_enc = get_encodings(anchors)\n",
    "        positives_enc = get_encodings(positives) \n",
    "        negatives_enc = get_encodings(negatives)\n",
    "           \n",
    "        # Clearing the previous gradients\n",
    "        optimizer.zero_grad()                                        \n",
    "\n",
    "         # Calculating the Train loss\n",
    "        loss = triplet_loss(anchors_enc, positives_enc, negatives_enc)\n",
    "        \n",
    "        # Backward propogation\n",
    "        loss.backward()\n",
    "       \n",
    "        # Optimizing the parameters\n",
    "        optimizer.step()                                            \n",
    "        \n",
    "        iterr += 1\n",
    "        \n",
    "        print(\"Iter {:.0f} Done.\\t Loss : {:.5f}\".format(iterr - init_iters, loss.item()))\n",
    "        \n",
    "        \n",
    "        # -----------------------------------------------------------------------------------------------------------------------------------------------------------\n",
    "        ### Inspecting the performance of the model ###\n",
    "        \n",
    "        if (iterr == 0 or iterr % inspect_size == 0):\n",
    "        \n",
    "            iter_list.append(iterr)\n",
    "            print(\"Iteration : {:.0f}/{:.0f} [{:2.0f}%] \".format(iterr - init_iters, n_iters, 100*(iterr - init_iters)/n_iters))\n",
    "            print('---------------------------')\n",
    "            \n",
    "        # -----------------------------------------------------------------------------------------------------------------------------------------------------------\n",
    "        ### Calculating train accuracy and loss ###                                          \n",
    "\n",
    "            # NOTE : Using encoding obtained in current training iteration.\n",
    "        \n",
    "            # Append train loss\n",
    "            train_loss_list.append(loss.item())\n",
    "            \n",
    "            # Use encoding to obtain vector difference\n",
    "            pos_diff, neg_diff = return_diff(anchors_enc, positives_enc, negatives_enc)\n",
    "            \n",
    "            # Append train accuracy\n",
    "            num_sample = anchors.shape[0]\n",
    "            tot_correct = np.sum(pos_diff < neg_diff)\n",
    "            train_acc = tot_correct/num_sample * 100\n",
    "            train_acc_list.append(train_acc)\n",
    "            \n",
    "            print('[Train]\\t Loss: {:.5f} | Acc: {:2.0f}%'.format(loss.item(), train_acc)) \n",
    "        \n",
    "        # -----------------------------------------------------------------------------------------------------------------------------------------------------------\n",
    "        ### Calculating test accuracy and loss ###\n",
    "            \n",
    "            # Use 100 samples for inspection from test set\n",
    "            anchors, positives, negatives = next(iter(test_loader))\n",
    "\n",
    "            # Get encodings by forward propogation\n",
    "            anchors_enc = get_encodings(anchors)\n",
    "            positives_enc = get_encodings(positives) \n",
    "            negatives_enc = get_encodings(negatives)\n",
    "\n",
    "            # Append test loss\n",
    "            loss = triplet_loss(anchors_enc, positives_enc, negatives_enc)        \n",
    "            test_loss_list.append(loss.item())                  \n",
    "                  \n",
    "            # Use encoding to obtain vector difference\n",
    "            pos_diff, neg_diff = return_diff(anchors_enc, positives_enc, negatives_enc)\n",
    "            \n",
    "            # Append test accuracy \n",
    "            num_sample = anchors.shape[0]\n",
    "            tot_correct = np.sum(pos_diff < neg_diff)\n",
    "            test_acc = tot_correct/num_sample * 100\n",
    "            test_acc_list.append(test_acc)\n",
    "            \n",
    "            print('[Test ]\\t Loss: {:.5f} | Acc: {:2.0f}%'.format(loss.item(), test_acc))        \n",
    "        \n",
    "        # -----------------------------------------------------------------------------------------------------------------------------------------------------------\n",
    "            print('=========================================================')\n",
    "            \n",
    "            \n",
    "print(\"\\nTraining Done.\")\n",
    "time_end = time.asctime()    # Time when training ended"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "OHsDltFoB1er"
   },
   "source": [
    "# BACKUPING UP"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "itbzqjgXGxjN"
   },
   "outputs": [],
   "source": [
    "'''SAVING THE MODEL'''\n",
    "\n",
    "torch.save(model.state_dict(), 'state_dict_margin=1_2.pt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 221
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2173,
     "status": "ok",
     "timestamp": 1562561270748,
     "user": {
      "displayName": "Aditya Ahuja",
      "photoUrl": "",
      "userId": "08809216472288903233"
     },
     "user_tz": -330
    },
    "id": "GKHgD87vGxlu",
    "outputId": "c04efc3c-1ec2-4782-b2be-4f688204052d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SiameseModel(\n",
       "  (cnn1): Conv2d(3, 16, kernel_size=(20, 20), stride=(5, 5))\n",
       "  (relu1): ReLU()\n",
       "  (maxpool1): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "  (cnn2): Conv2d(16, 32, kernel_size=(3, 3), stride=(1, 1))\n",
       "  (relu2): ReLU()\n",
       "  (maxpool2): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "  (cnn3): Conv2d(32, 64, kernel_size=(3, 3), stride=(1, 1))\n",
       "  (relu3): ReLU()\n",
       "  (maxpool3): MaxPool2d(kernel_size=[2, 3], stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "  (fc1): Linear(in_features=7680, out_features=156, bias=True)\n",
       ")"
      ]
     },
     "execution_count": 57,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "'''LOADING THE MODEL'''\n",
    "\n",
    "# model.load_state_dict(torch.load('state_dict_margin=10_1.pt'))\n",
    "# model.eval()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "PULOGKYJB9vB"
   },
   "source": [
    "# ANALYSIS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 221
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2456357,
     "status": "ok",
     "timestamp": 1562560448913,
     "user": {
      "displayName": "Aditya Ahuja",
      "photoUrl": "",
      "userId": "08809216472288903233"
     },
     "user_tz": -330
    },
    "id": "Y4OQKFpiCwe7",
    "outputId": "8bdb2dbc-b6c3-4bcb-8179-fc6826f064e4"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total Iterations     : 758\n",
      "Total Epochs         : 1\n",
      "Total Sample-Points  : 51\n",
      "-------------------------------\n",
      "Loss - Train     : 0.00\n",
      "Loss - Test      : 0.00\n",
      "Acc - Train     : 100.00\n",
      "Acc - Test      : 99.90\n",
      "-------------------------------\n",
      "Start Time        : 06:49:27\n",
      "End Time          : 07:25:57\n",
      "Total Train-time  :  0:36:30\n"
     ]
    }
   ],
   "source": [
    "'''POST TRAINING RESULTS'''\n",
    "\n",
    "# Formatting the date-time data\n",
    "from datetime import datetime\n",
    "FMT = '%H:%M:%S'\n",
    "td = (datetime.strptime(time_end[11:19], FMT) - datetime.strptime(time_begin[11:19], FMT)).seconds\n",
    "hr = (td//3600)\n",
    "min = (td - 3600*hr)//60\n",
    "sec = (td - 3600*hr - 60*min)\n",
    "\n",
    "print(\"Total Iterations     : {:.0f}\".format(iterr))\n",
    "print(\"Total Epochs         : {:.0f}\".format(iterr*100/60000))\n",
    "print(\"Total Sample-Points  : {:.0f}\".format(iterr/inspect_size))\n",
    "print(\"-------------------------------\")\n",
    "print(\"Loss - Train     : {:.2f}\".format(np.mean(train_loss_list[-10:])))\n",
    "print(\"Loss - Test      : {:.2f}\".format(np.mean(test_loss_list[-10:])))\n",
    "print(\"Acc - Train     : {:.2f}\".format(np.mean(train_acc_list[-10:])))\n",
    "print(\"Acc - Test      : {:.2f}\".format(np.mean(test_acc_list[-10:])))\n",
    "print(\"-------------------------------\")\n",
    "print(\"Start Time        : {}\".format(time_begin[11:19]))\n",
    "print(\"End Time          : {}\".format(time_end[11:19]))\n",
    "print(\"Total Train-time  : {:2.0f}:{:2.0f}:{:2.0f}\".format(hr,min,sec))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "mUDtxgT9I9A1"
   },
   "outputs": [],
   "source": [
    "'''GENERATING AVERAGE INSPECTION LISTS'''\n",
    "\n",
    "# Defining roll function\n",
    "def make_roll(input_list, roll_size=5):\n",
    "    output_list = []\n",
    "    \n",
    "    for i in range(len(input_list)):\n",
    "        if i==0:\n",
    "            output_list.append(input_list[0])\n",
    "        elif i<roll_size:\n",
    "            output_list.append(np.mean(input_list[:i+1]))\n",
    "        else:\n",
    "            output_list.append(np.mean(input_list[i-roll_size:i]))\n",
    "    return output_list\n",
    "\n",
    "# Generating roll lists\n",
    "train_roll_loss_list = make_roll(train_loss_list, roll_size=10) \n",
    "train_roll_acc_list = make_roll(train_acc_list, roll_size=10)  \n",
    "test_roll_loss_list = make_roll(test_loss_list, roll_size=10)   \n",
    "test_roll_acc_list = make_roll(test_acc_list, roll_size=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 515
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2526,
     "status": "ok",
     "timestamp": 1562564643326,
     "user": {
      "displayName": "Aditya Ahuja",
      "photoUrl": "",
      "userId": "08809216472288903233"
     },
     "user_tz": -330
    },
    "id": "aK_BkGFksKaw",
    "outputId": "b5061321-12b0-48ae-8347-836f03ef651b"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAuMAAAHyCAYAAABBH8AHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAAEnQAABJ0BfDRroQAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOzde5RcVZ3//feuqlN9yYUkhHAzoExgoslviBKUQNDOhTDkxyA8+DhGghhvzIgorDU6RlEeUS5GRmUeBxxRwUQkuhBQxhGRQACJGXkQmCGCmMVPo0kIhJB0J+m67+ePququ7q6+nTrf9OV8XmuxUqmqPnXq3Rln9+59djnvPSIiIiIicuglRvoERERERETiSoNxEREREZERosG4iIiIiMgI0WBcRKTCObfROXfFAI8vdc495pzb75zb55z7uXPuLb2e8zbn3MPOudecc3udc//tnHv/UB+34pzzzrl5zrlvVs5/v3Mu45wr1vx9v3PuOOfc7c65XK/79zvnZlSOtdE5l63ct8c594hzbn6d13TOua3Oue3OuWTN/Q/UHDPnnCvU/H1v5Tm/cs59rOZr0s65LzjnXnTOdTrn/uyc+xfn3ISa5yytvM/1vc7jG865b9f8fUHlnKvfg2eccxdHW1xEZGg0GBcRGQLn3HnAvcBa4Gjg9cCjwKPVgahzbhJwP/BDYAZwBPBB4OWhPH4oeO//wXs/0Xs/EfgH4H+qf6/8t63y1Jt73T/Re197nv9cOcZRwH8Bd9d5uTbgOGAycE7NOSyrOYc1wL01rzGln1P/IfBO4F3ARGAZcCrwc+dcquZ5ncC5vX9IqnLOTQF+Dqyj/D2YAXwYeKWf1xURMaXBuIjIIJxzDrgJuMF7f6v3vsN7/5r3/nrKg8QbK0/9a2AC8C3vfb7y3xPe+/8c4uO9X/cm59x3e933z865X1Run1WZWe9wzu1yzt0S/bsfmPc+B3wPmOmcO6LXwx8E/gP4ceV2KM65pZQH8+/03v/We1/03j9HeXD+v4D31Dz9IPCvwA39HG420AR8p9I/573/jff+/rDnJyLSCA3GRUQGdxLlmfAf1HnsB8BC51wL8AKwD1jvnHunc+6oXs8d7PHe1gEXVo5ddTHl2XkoD4K/4r2fBJxQef4hVTm3DwK7gddq7p8CXAjcXjnPc51zR4Z8mbOBTd77P9Xe6b1/jfJvGpb1ev6XgVOcc0vqHOt5ygP2O51z5zVwTiIikdBgXERkcNMrf+6o89gOIAlM8963AwuAPcBXgR3Ouf+qLpkY7PHevPf/H/AXyjPAOOfeTHnZxz2Vp+SBWc65I7z3B7z3mxp/q13+sbKeuvrf73s9fn1lffcB4L3A/+W9L9Q8/l6gg/KSkI2UO70v5LlMp357Kvf3mJH33u8DrgNuqPxWo/axvcBpQDvwdWCnc26zc25eyHMTEWmIBuMiIoPbXfnzmDqPHQMUKQ+w8d5vrazL/ivgdcBW4KfVQeFgj9exju5B7PuAH3vvD1b+fgEwF/i9c+4p59y7G3qXPd3ivZ9S899f93p8dWV990xgO/A3vR7/IPCDylIQX3kfHwh5Lrup357K/fXWe/8bcCTlNeY9eO//4L3/iPf+hMr5/xH4SchzExFpiAbjIiKDewH4E7CizmMrgMe99529H/De76C8dvlYYNpwH6+4A1jinDu28lpdS1Eq66cvpDxz/EXgB4d62YX3fjvlCyC/7Jw7BqAyy/wW4P3OuZeccy8BlwGznXNnhHiZXwKnO+eOq73TOTcV+NvK473PKwN8HvgSkOr9eK/z/zJwnHPusBDnJiLSEA3GRUR6Sjnnmmv+a6rM7F4JrHbOfdA5N9E5N8U598+ULx78FIBzbnblAsvXO+cSlXXTHwNe8N6/Otjj9U7Ge/9n4FfAd4Ac8HDltdLOuYudc1O99yVgb+VLCvWOY8l7/1vKS1E+U7nrg8BTlC+WnFf5743ABkJcyOm9fwB4ALjXlbdnTDrn/prycp3ngPX9fOlaykt5umbHnXNznHOfrHwPXGVA/zHgd5XlLSIih5QG4yIiPX2F8vZ41f9+D+C9v4fyBYmrgJeAbcBiYJH3/r8qX9sBvBl4jPKa5N9TXs/8d0N8vD9rKV/E+P3KDwZV7wW2Ouc6gP8XeG9/g/oQPur67jP+5gGefy3wIefcTOAi4Cbv/Uu1/1Feo/3uyhaPw/V/Az+jPAA/QHlg/xSwzHufr/cFlR9SVgOH19zdDsyn/APOfsqD+cOorMsXETnUXM//XRcRERERkUOl33V0I62ydu8Eyr+WFREREREZD9LAi9WlcaN2MA6ccO+99/521qxZDR+os7N8XVVLS8sgz5ThUFcb6mpHbW2oqw11taGudtR2aLZu3cr555//FspL7Ub1YDw3a9Ys5syZ0/CB8vnycsIgCBo+lnRTVxvqakdtbairDXW1oa521HZYulZ+jObBeGSy2SygfxxRU1cb6mpHbW2oqw11taGudtQ2nFgMxkul0kifwrikrjbU1Y7a2lBXG+pqQ13tqG04sRiMNzU1jfQpjEvqakNd7aitDXW1oa424tg1n89TKNh/BEF1h77q2nEpS6VSA/62QPuMi4iIiIxT+Xye9vb2Q/JaqVSKVCoW87zD0t7e3rWevp5YFKuuYYrjT8OW1NWGutpRWxvqakNdbcSta6FQoLW19ZDscFIsFgFIJpPmrzXWFAqFfmfHYzEY1z8KG+pqQ13tqK0NdbWhrjbUVUabWCxTSafTpNPpkT6NcUddbairHbW1oa421NWGutrJZDIsWbKEtrY2Jk2aRFtbG21tbezZs2fAr7v22mvZtm3boMffunUr73nPe6I63VEjFjPjuVx5K0dttRMtdbWhrnbU1oa62lBXG+pqp7m5mQ0bNpBMJpk/fz4bN27s8XipVCKR6DsP/NnPfvYQneHoFIuZ8WKx2LWOSaKjrjbU1Y7a2lBXG+pqQ10PrQcffJDzzjuP888/n3Xr1nH99dfT1tbGKaecwoYNGwBYuXIlzz//PA8++CDLly/nggsu4OSTT+a5554b9PjPPPMMCxcu5PTTT2fNmjUAfOMb3+Btb3sbixYt4qc//SkvvPACCxYsYNGiRXz0ox81fb9hxGJmPC4XaRxq6mpDXe2orQ11taGuNuLc1ZeK0L4vmoNNPgyX6Ln+3jlX96n79+9nw4YNOOc4ePAgq1ev5qWXXmLFihUsWbKkx3OLxSL33HMP9913H7fddlvXALs/q1ev5rvf/S4nnngiS5cuZcWKFdx1111s2LCBiRMnUiqVuPXWW1m1ahUf+chHRuVe6LEYjIuIiIjEXvs+CjddG8mhUp/4LEyZNqTnzp8/v2ugfvvtt3PnnXeSTCbZtWtXn+fOmzcPgJkzZ/Laa68NeuxXXnmFk046CYA3v/nNvPjii9xwww1cfvnlAHzmM59hxYoVXHPNNVx00UWcc845rFy5ckjnfajEYjAet22MDhV1taGudtTWhrraUFcbse46+bDyIDqiY/VW/dCf3mrXid98883893//N7t27aKtra3Pc2tn1/s7Xq3p06fzwgsvcOKJJ/LUU0/xiU98gunTp3Pbbbfx6KOP8pWvfIWbbrqJG2+8Ee89c+bM4aKLLup3Fn8kxGIwXu9iAWmcutpQVztqa0NdbairjTh3dYnkkGezrZx22mksXLiQBQsWMGHChGF//caNG1m6dCkAZ599Ntdddx3vf//7ATjvvPOYOXMmK1eu5M9//jO5XI4bbriBe+65h1tuuQWAc889d1QNxAHcUH7qGAnOuTnPPvvss3PmzGn4WNVPPdKV09FSVxvqakdtbairDXW1Ebeu1Y+mPxQf+lNdjx3nH3jq6f092LJlC3Pnzp3rvd8CMZkZj9v/4R0q6mpDXe2orQ11taGuNtTVzmid4B3tYjEYLxQKI30K45K62lBXO2prQ11tqKsNdZXRJhaD8VhepHEIqKsNdbWjtjbU1Ya62lBXO6NtLfZYEYvB+LBt21b+r1CAOlf6ioiIiIhEIRaD8WFvY/S5z8HatfgjjsC9/LLhmY1tsd4eypC62lFbG+pqQ11tqKsdrRkPJxaD8WH/2uTww8t/trdHfzLjiH4dZUNd7aitDXW1oa421FVGm1gMxpubm4f3BdPKe3C6bBZyOUinDc5q7Bt2VxkSdbWjtjbU1Ya62lBXO/pBJ5xYbARZKBSGd/X0tJoN8ffsif6Exolhd5UhUVc7amtDXW2oqw11tdPZ2cnixYtpa2tj0qRJtLW10dbWxp5BxlLXXnst27ZtG/T4W7du5T3veU9Up9t1zBkzZnDttdcCsHLlSt761rdy5pln8u53v3vAfysLFy4kk8nw7W9/m29+85sALF68mNNOO21Y5xCLmfHqnqJD3vC+ukwFyoPxo44yOKuxb9hdZUjU1Y7a2lBXG+pqI9Zdr7gCnn66/8ery3MnTx78WPPmwde/3uOu5uZmNmzYQDKZZP78+WzcuLHH46VSqe4HAn32s58d/PUMLV68uMc5rF27ltmzZ/P+97+fDRs2cPbZZw/5WA888AALFy4c1uvHYjA+7Is0amfGX3012pMZR3Txiw11taO2NtTVhrraiHXXp5+GRx4xO3y9ZSoPPvgg//qv/0oikeCCCy5gx44d/OIXv6Cjo4M1a9awZMkSVq5cyVVXXcVf/vIXvvrVr9LU1MSLL77I+vXreeMb3zjgaz7zzDNcdtlllEolzj//fD71qU/xjW98g3Xr1tHa2sqVV17J7NmzueSSS2hubuaNb3wjN99886DvxXvPvn37ui5KveOOO7rexxe/+EWWLl0aLlIdsRiMD5uWqYiIiMh4M2/eiBxr//79bNiwAeccBw8eZPXq1bz00kusWLGCJUuW9HhusVjknnvu4b777uO2225jzZo1Ax579erVfPe73+XEE09k6dKlrFixgrvuuosNGzYwceJESqUSt956K6tWreIjH/kIpVJp0PN93/vex6uvvsqsWbNYunQp+XyeG2+8kc2bN9PZ2clZZ52lwfhwDXsbo97LVKQubQ9lQ13tqK0NdbWhrjZi3bXXspKo9be14fz587tmzW+//XbuvPNOkskku3bt6vPceZVB/syZM3nttdcGfc1XXnmFk046CYA3v/nNvPjii9xwww1cfvnlAHzmM59hxYoVXHPNNVx00UWcc845rFy5csBjrl27luOOO46//du/Ze/evWQyGV7/+tfT1NREU1MTiUQi0m0cY3EB57BpmYqIiIhIJGrXid9888088sgj3HnnnXVnqWuXugxlwDt9+nReeOEFvPc89dRTnHDCCZx88sncdtttrFq1iq985SsEQcCNN97I97//fa677rohHbe1tZV//Md/5Ktf/SpHHnkkL774Itlslr1791IqlSLdOSYWM+PDvkhj0iR8MokrFjUzPoBYXvxyCKirHbW1oa421NWGutoZygD1tNNOY+HChSxYsIAJEyYM+zU2btzYtUTk7LPP5rrrruP9738/AOeddx4zZ85k5cqV/PnPfyaXy3HDDTdwzz33cMsttwBw7rnnDnkgfeGFF3Ldddfx+c9/nn/6p3/izDPPJJFIdO28EhU3Wj8tyTk359lnn312zpw5DR+rs7MTGN7/Afoph+H2tcOll0JluxrpKUxXGZy62lFbG+pqQ11txK3roXy/1ZnuejumjFZ/+tOfWL58Oe9973sj2dVlyZIlTJ48mXvuuafrvt7fgy1btjB37ty53vstEJOZ8VDbGE2aBPvatUxlALHeHsqQutpRWxvqakNdbairndE6wTuQ448/ni1btkR2vA0bNgz7a2IxGE+H+QTNyZPKf2qZSr9CdZVBqasdtbWhrjbU1Ya62tEncIYTi8F4qF+XHFbZ8F4z4/0aS7+GGkvU1Y7a2lBXG+pqQ11ltInFYDyTyQDD+2nYH3YYDjQzPoAwXWVw6mpHbW2oqw11taGudsbiMpXRQD8e9mfK1PKfGoyLiIiIDKqzs5PFixfT1tbGpEmTaGtro62tjT1DGEvdfffd7N69u8/9K1eu5Pnnn7c43VEjFjPjzc3Nw/+iKVPKfx44ANksxPHDAQYRqqsMSl3tqK0NdbWhrjbU1U5raysPP/wwiUSC+fPns3HjxiF/7d13382b3vQmpk+fbneCo1QsBuND+ejTPqZO6b69Zw8cfXR0JzROhOoqg1JXO2prQ11tqKuNOHctFQocrPOpl2G0HnkkidTgw0jvPZdddhnPPfccqVSK733ve6TTaS688EISiQRTp05lzZo1/PKXv+SFF15gyZIlXH/99QMe84orruDJJ58knU5z22230dra2uN4d999N6tXr+aRRx6hqamJNWvWcOqpp0byvi3EYjCey+WAYf40XPspnBqM1xWqqwxKXe2orQ11taGuNuLc9eCuXfx02bJIjnXeAw8w8dhje9xXb834T37yE4466ihuvvlmHn/8cdasWcM555zDGWecwXXXXUepVCKRSHDWWWdx1VVXMXv27AFfd/PmzezevZvHHnuMhx9+mC996UtceOGFPY4H8NBDD7Fp0yaSyeSo/wEsFoPxIAiG/0WH1wzGtaNKXaG6yqDU1Y7a2lBXG+pqI85dW488kvMeeCCyY/VWb2vD3/3ud9x111089NBDlEolTjjhBBYvXsymTZu46KKLmD9/PldeeeWQX3fr1q1ds9ynnnoq11xzTd3jXX311axatYqWlhauueYajqxzvqNFLAbjqSH8GqWPw2vWLOkizrpCdZVBqasdtbWhrjbU1UacuyZSqT6z2dZmz57NihUrWL16NVD+0KVCocAXvvAFABYvXszf//3fEwQBxWJx0OPNmjWL+++/H4AnnniCE088se7xFi1axPLly1m7di3f+c53+MxnPmP0DhsXi3+R1Y8hHdZPw9OP6L6twXhdobrKoNTVjtraUFcb6mpDXe3UW6ZywQUX8PGPf5xFixYBcMkll3D88cfzuc99jmQyyXHHHcdRRx3F8uXLufzyy1m2bBmf/vSnexzjAx/4AK2trQDccccdTJ06lYULFxIEAbfffjubN2/uc7xly5ZRKBQoFAp861vfsn/zDXCjdU9I59ycZ5999tk5c+Y0fKz29nYAJk+ePOSv8X94HnfSG8t/WbMGPvnJhs9jvAnTVQanrnbU1oa62lBXG3HrWv3ho6Wlxfy1qjPbyWTS/LXGkt7fgy1btjB37ty53vstEJOZ8aYw2xIeNgWfSOBKJc2M9yNUVxmUutpRWxvqakNdbairnXprxmVw+tCf/qTT0FK50lqDcRERERExEIvBeDabJZvNDu+LUkH3YFy7qdQVqqsMSl3tqK0NdbWhrjbU1Y73vu66cRlYLJaphLpIIwjwLc040Mx4P3Txiw11taO2NtTVhrraUFc7WqYSTiwG46G2MQoCaK1c7KCZ8brivD2UJXW1o7Y21NWGutpQVxltYvEvMpPJAMP7adglkvhWrRkfSJiuMjh1taO2NtTVhrraUFc7WqISTizWjIddw+QnTCjf0GC8Lq0Ns6GudtTWhrraUFcb6mqns7OTxYsX09bWxqRJk2hra6OtrY09QxhH3X333ezevbvP/StXruT555+P9DxXrlzJW9/6VrZv387WrVuZMWMGbW1tnHrqqTwwwCeUPvjgg117oM+fPx+AdevWcfzxx3d9EFEYsZgZD72N0cTKYPzgQchkoLk5upMaB7Q9lA11taO2NtTVhrraiHPXK664gqeffrrfx4ezB/u8efP4+te/3uO+1tZWHn74YRKJBPPnz2fjxo1DPre7776bN73pTUyfPn3wJ0dg7dq1HHvssWzdupXFixezfv16/vjHP7JixQqWLVs25ONcfPHF/P73v2/oXGIxGA9t0qTu23v2wDHHjNy5iIiIiDTg6aef5pFHHjmkr+m957LLLuO5554jlUrxve99j3Q6zYUXXkgikWDq1KmsWbOGX/7yl7zwwgssWbKE66+/fsBjXnHFFTz55JOk02luu+02Wltbexzv7rvvZvXq1TzyyCM0NTWxZs0aTj311EHPde/evV2/Ndm3bx8rV66kvb2dY489lrVr10bSo55YDMarWxgN96dhr8H4gMJ2lYGpqx21taGuNtTVRpy7zps3z/RY9Zb//OQnP+Goo47i5ptv5vHHH2fNmjWcc845nHHGGVx33XWUSiUSiQRnnXUWV111FbNnzx7wdTdv3szu3bt57LHHePjhh/nSl77EhRde2ON4AA899BCbNm0imUx23defhx56iIULF/LMM89w7733AnDLLbfwzne+kw996ENcffXV/OhHP2LGjBlDzTMssRiMh75yenLNYFw7qvShK9JtqKsdtbWhrjbU1Uacu/ZeVnIo/O53v+Ouu+7ioYceolQqccIJJ7B48WI2bdrERRddxPz587nyyiuHfLytW7d2zXKfeuqpXHPNNXWPd/XVV7Nq1SpaWlq45pprOPLII/s9ZnWZyrp169i4cSNLlixh69atnHPOOV2v8+STT5oNxmNxAWcQBOGump7ca2ZcegjdVQakrnbU1oa62lBXG+pqxznXZ6/x2bNns2LFCjZu3Mijjz7KrbfeSqFQ4Atf+AJ33HEH9913Hzt27CAIAorF4qCvMWvWLJ544gkAnnjiCU488cS6x1u0aBFr167ljDPO4Dvf+c6Qzn/lypX853/+J3v37mXWrFn85je/6fE6VmLx42H1V1LD/j++KVO6b2sw3kforjIgdbWjtjbU1Ya62lBXO/WWqVxwwQV8/OMfZ9GiRQBccsklHH/88Xzuc58jmUxy3HHHcdRRR7F8+XIuv/xyli1b1rVjSdUHPvABWltbAbjjjjuYOnUqCxcuJAgCbr/9djZv3tzneMuWLaNQKFAoFPjWt741pPN3znHJJZfw7W9/m0svvZSVK1fy/e9/n6OPPpqrrrrKbL29G63b+zjn5jz77LPPzpkzp+FjDefq4FqF2/6N1Ac+Vv7Ll78Mn/pUw+cynoTtKgNTVztqa0Ndbairjbh17ezsBKClpcX8taoz28lk0vy1GnHllVfym9/8hh/96Ecce+yxDR1r3bp1/Mu//As33XQT73jHO+o+p/f3YMuWLcydO3eu934LxGRmPPzWhpPwqSSuUNTMeB1xvPjlUFBXO2prQ11tqKsNdbXTe4nKaPW1r30tsmNdfPHFXHzxxQ0dIxZrxkNLp2FiZd24BuMiIiIyxiQSiUF3ExFb1R1j+hOLmfGw2xi5VAATW2HvXu2mUkect4eypK521NaGutpQVxtx65pOp2lvbx9wMBiV6qD/ULzWWHLw4MEBP8woFoPx0GuXUgFMKF8woJnxvkb7mrCxSl3tqK0NdbWhrjbi1tU5x9SpU8nn8+avVf1B51CsTx9Lpk6dOuASnlgMxtPpdLgvDAL8hFYcaGa8jtBdZUDqakdtbairDXW1EceuqVTqkOyvXn0N7VQzPLEYjOdyOSDEP45UANWf7jQz3kforjIgdbWjtjbU1Ya62lBXO2obTiwG40PZRL6uIMC3NJdvazDeR+iuMiB1taO2NtTVhrraUFc7ahtOLAbjoS/SCAKoDsY7O8v/aR1Ul7hc/HKoqasdtbWhrjbU1Ya62lHbcHS56wBcKoCmmrVlmh0XERERkQjFYjCezWa7rvAdliDAN2sw3p/QXWVA6mpHbW2oqw11taGudtQ2nFgsUwm932UqgHTNRQjaUaUH7SNqQ13tqK0NdbWhrjbU1Y7ahhPpYNw592XgdOCPwAe89/nK/X8HXAXkgSe995+o3N8BPFn58su99/8T5flUNbJm3NdeEayZ8R60NsyGutpRWxvqakNdbairHbUNJ7IfYZxzJwPHeu/PBJ4H3lXz8DPAGd77hcAM59z8yv2/9963Vf4zGYgD5PP5cJvdBwG0Nnf/XYPxHkJ3lQGpqx21taGuNtTVhrraUdtwopwZPx14oHL7fmAVcCeA935bzfNyQKly+6+cc48CW4ArvfeZmudN2rlzJzNnzuz6SSubzRIEAalUikwmg/e+x2OpVIogCMhms5RKpa7H9u3bRyKRIAgCcrkcxWKxx9clEgmamprI5/MUCoWux3LZHM0t3YPxzPbtuMpaqNpPmSoUCuTzedLpNIlEgkym/Daam5splUrkcrmu8+7s7ARo+D1ls1mSySTpdHpY7ymbzeKco7m5ueu8ax8bznuqnkOhUBg372k0fJ8OHjxIPp+nWCyOm/c0Wr5PnZ2dBEFAqVQaN+9pNHyfOjs7yeVyFAqFcfOeRsP3ad++fTjnCIJg3Lyn0fB9ymQypFIpisXiuHlPo+X7tG/fPqC8z/h4eU8W36cDBw5QK8rFPVOB9srtfcC03k9wzp0KzPDe/7Zy1yzv/duBncBlvZ5+8fr16yM5sSAIwn3iViqAIIWvfK177bVIzme8SKfTsfwkM2tNTU36wAQj+jdrI51O69+sgdD/v0sGpP8dsBMEgf63IATnvY/mQM59FNjvvV/rnDsFWOW9/1jN468Dfghc4L1/udfXvgn4tPf+fTX3nbZp06ZfL1iwoOFzq/5UM9y1TH7vHgo3XUvqm3fgXtoFH/oQ3Hprw+czXoTtKgNTVztqa0NdbairDXW1o7ZDs2XLFubOnTvXe78Fop0Z3wQsrdw+G3i8+oBzbhKwHri0OhB3zk1wziUrTzkT2NrreB2TJ0+O5MQa2doQgMOmlP/Ubio9aAsjG+pqR21tqKsNdbWhrnbUNpzIBuPe+6eBXc65x4A5wI+dc/9eefgK4A3AN5xzG51z7wBOBJ6orBlfDtwU1bn05pzDOTf8Lwwqv8aaclj5T13A2UPorjIgdbWjtjbU1Ya62lBXO2obTqRbG3rvP9nrrksr938R+GKdL3lLlK/fn+bm5sGfVE+qOjNeGYxrZryH0F1lQOpqR21tqKsNdbWhrnbUNpxY7M5eKBS6dvwYDpdIQCIJkyeV79DMeA9hu8rA1NWO2tpQVxvqakNd7ahtOLEYjDe072UQaDDeD+0nakNd7aitDXW1oa421NWO2oYT6TKV0aqhq3pTAX7SJBxAJgMHD0Jra1SnNqbpamkb6mpHbW2oqw11taGudtQ2nFjMjAXmEgIAACAASURBVDckCGDixO6/a3ZcRERERCISi8F4Q1vtBAF+Qs1MuAbjXbSFkQ11taO2NtTVhrraUFc7ahtOLJapNMKlAphYk0k7qoiIiIhIRGIxGG9paQn/xUEAyWT33zUz3qWhrtIvdbWjtjbU1Ya62lBXO2obTiyWqTS01U4qgNp/XBqMd9EWRjbU1Y7a2lBXG+pqQ13tqG04sRiMN7y1YWvNYFzLVLpoCyMb6mpHbW2oqw11taGudtQ2nFgsU0mn0+G/OAjwCQdNTZDNama8RkNdpV/qakdtbairDXW1oa521DacWMyMJxIJEomQbzUVQLEAhx9e/rsG410a6ir9Ulc7amtDXW2oqw11taO24cSiWCaTIZPJhPpalwogn4dp08p3aJlKl0a6Sv/U1Y7a2lBXG+pqQ13tqG04sRiMNyQIoFAzGNfMuIiIiIhEJBZrxpubm8N/cXVmvLpMRTPjXRrqKv1SVztqa0NdbairDXW1o7bhxGIwXiqVwn9xkIZ8TjPjdTTUVfqlrnbU1oa62lBXG+pqR23DicUylVwuRy6XC/fFQYDvvUzF++hObgxrqKv0S13tqK0NdbWhrjbU1Y7ahhOLmfEgCMJ/ce9lKtksHDwIEyZEc3JjWENdpV/qakdtbairDXW1oa521DacWAzGU6nwb9N1XcB5dPede/ZoME5jXaV/6mpHbW2oqw11taGudtQ2nFhU6+zsBEL+xNZ7a0MoD8Znzozo7MauhrpKv9TVjtraUFcb6mpDXe2obTixWDPekCCAQs2H/oB2VBERERGRSMRiZrypqSn8FwcB5HP4qVNx1fu0owrQYFfpl7raUVsb6mpDXW2oqx21DUcz44NJVX7Vctjk7vs0GBcRERGRCMRiMJ7NZslms6G+1lXXPU2uGYxrmQrQWFfpn7raUVsb6mpDXW2oqx21DScWy1Qa3toQIEhBczNkMpoZr9AFGjbU1Y7a2lBXG+pqQ13tqG04sRiMN7TVTvUfVnVHlR07NBiv0BZGNtTVjtraUFcb6mpDXe2obTixqJbJZIAGtjaE8l7jhx9eHoxrmQrQYFfpl7raUVsb6mpDXW2oqx21DScWg3HfyMfXV/5B+XweV91rXDPjQINdpV/qakdtbairDXW1oa521DacWAzGG97aEMrLVKp7jWtmHNAWRlbU1Y7a2lBXG+pqQ13tqG04sdhNpRHOJSCZLC9T0cy4iIiIiEQoFoPxhrfaSQXdF3BCeTCuX8VoCyMj6mpHbW2oqw11taGudtQ2nFgsU2n46t4g6L6AEyCXgwMHYOLExk9uDNNV0zbU1Y7a2lBXG+pqQ13tqG04sajW8FW9QRryue6ZcSjPjsd8MK6rpW2oqx21taGuNtTVhrraUdtwtExlKFIBvnaZCmjdOPp1lBV1taO2NtTVhrraUFc7ahtOLGbGS6VSQ1/vei9TAe2oQuNdpT51taO2NtTVhrraUFc7ahtOLAbjDW+1U72A8/Ajuu/TzLi2MDKirnbU1oa62lBXG+pqR23DicUylYYFqZ5bG4IG4yIiIiLSsFgMxqPZ2rDQczCuZSpaG2ZEXe2orQ11taGuNtTVjtqGE4tlKslksrEDBGl8PgctLeX/Ojs1M04EXaUudbWjtjbU1Ya62lBXO2obTiwG4+l0urEDpCoXcEJ5dnz7ds2ME0FXqUtd7aitDXW1oa421NWO2oYTi2UquVyOXC4X+uu7dlOB7h1VNDPecFepT13tqK0NdbWhrjbU1Y7ahhOLmfFisdjYAaq7qUD3unENxhvvKnWpqx21taGuNtTVhrraUdtwYjEYb3irnaDXMhXQMhW0hZEVdbWjtjbU1Ya62lBXO2obTiyWqTQsqJkZ1zIVEREREYlILAbjUWxt6OstU/G+8ZMbw7SFkQ11taO2NtTVhrraUFc7ahtOLJapJBKN/cxR9wLOfB7274dJkxo8u7Gr0a5Sn7raUVsb6mpDXW2oqx21DScWg/GG1zDVu4ATyrPjMR6Ma22YDXW1o7Y21NWGutpQVztqG04sfoTJ5/Pkq4PpMOpdwAmxXzfecFepS13tqK0NdbWhrjbU1Y7ahhOLmfFCodDYAVJ1LuCE2O+o0nBXqUtd7aitDXW1oa421NWO2oYTi8F441sbpqGQx3uP08x4F/06yoa62lFbG+pqQ11tqKsdtQ0nFstUGhYE5T8LBS1TEREREZHIxGIw3uhWOy5VHYznew7GY75MRVsY2VBXO2prQ11tqKsNdbWjtuHEYjDunMM5F/4A1ZnxfB6am6G1tfz3mM+MN9xV6lJXO2prQ11tqKsNdbWjtuHEYs14c3NzYwdIVTLV7qhy8GDsZ8Yb7ip1qasdtbWhrjbU1Ya62lHbcGIxM14oFBq7wjdIl//svaNKzGfGG+4qdamrHbW1oa421NWGutpR23BiMRiPZJ9xwOdz5b9X143HfDCu/URtqKsdtbWhrjbU1Ya62lHbcGKxTKXxT+Css0wFYr9MRVsY2VBXO2prQ11tqKsNdbWjtuHEYma8Uc4lIJnSMhURERERiVQsBuORbLUTBH1nxvfsAe8bO+4Ypi2MbKirHbW1oa421NWGutpR23BiMRiPRCronhmvDsYLBejoGLlzEhEREZExLRZrxltaWho/SBD0XaYC5dnxyZMbP/4YFElX6UNd7aitDXW1oa421NWO2oYTi5nxSLbaCQJ872UqEOt149rCyIa62lFbG+pqQ11tqKsdtQ0nFoPxKLbacal+ZsZjvKOKtjCyoa521NaGutpQVxvqakdtw4nFMpV0Ot34QepdwAmxnhmPpKv0oa521NaGutpQVxvqakdtw4nFYDyRiOAXAPUu4IRYD8Yj6Sp9qKsdtbWhrjbU1Ya62lHbcCKt5pz7snPuMefcOudcUHP/3znn/ss59yvn3E0193/COfe4c+6nzjmzqyAzmQyZTKaxg/Q3Mx7jZSqRdJU+1NWO2tpQVxvqakNd7ahtOJENxp1zJwPHeu/PBJ4H3lXz8DPAGd77hcAM59x859x04DxgIfBD4LKozsVEEODzufLtpiaYMKF8O8Yz4yIiIiLSmCiXqZwOPFC5fT+wCrgTwHu/reZ5OaAEnAo84r33zrn7ge/1Ot6knTt3MnPmzK6PV81mswRBQCqVIpPJ4L3v8VgqlSIIArLZLKVSqeuxfD5PIpEgn8+Ty+UoFos9vi6RSNDU1EQ+n6dQKPR4zDlHc3MzJRKUOjspVjazD6ZOJXHgAKVXXiHb2Uk+nyedTpNIJLp+KmxubqZUKpHL5brOu7OzE6Dh95TNZkkmk6TT6dDvqVAokM/nezwG5a2Jqo8N9J4Akskk7e3t4+Y9jYbvUyKRIJvN0tHRMW7e02j5PpVKJYIgYP/+/ePmPY2W71M2m6W9vX1cvaeR/j7V/v+u8fKeRsP3yXtPKpWio6Nj3Lyn0fJ9ql68Wf2a8fCeLL5PBw4coFaUy1SmAu2V2/uAab2f4Jw7FZjhvf/tEJ5/8fr16yM5Me89vtFPygwCXKH7CmE/dWr5RoxnxkulEqVSaaRPY9wplUqN/3uVurz3+jdrQP9mbejfqw39/y47kYy3YshFFc0591Fgv/d+rXPuFGCV9/5jNY+/jvJylAu89y8755YD87331zjnDge+570/t+b5p23atOnXCxYsaPjcqjO3kxv4cJ7iL+/Dv/wSqYs+XL5jyRJ46CE4/XR4/PGGz3EsiqKr9KWudtTWhrraUFcb6mpHbYdmy5YtzJ07d673fgtEOzO+CVhauX020DVCdc5NAtYDl3rvX67c/QTw9nrPr+iI6psZBAFBEAz+xIHU7qYC3RdxxvgCzki6Sh/qakdtbairDXW1oa521DacyNaMe++fds7tcs49BmwDbnTO/bv3/lLgCuANwDeccwBXe+8fcc79zDn3OPAacFFU59JbKhXB2wzS3bupQPcH/8R4mUokXaUPdbWjtjbU1Ya62lBXO2obTqTVvPef7HXXpZX7vwh8sc7zvwZ8LcpzqKe6iL6hn9Zqd1OB7pnxPXvAeyj/kBErkXSVPtTVjtraUFcb6mpDXe2obTjanX2IXH/LVIpFaG+v/0UiIiIiIgOIxe8TqtvKNKT2Q3+ge5kKlGfHDzus8dcYYyLpKn2oqx21taGuNtTVhrraUdtwNDM+VP3NjEOs142LiIiISHixGIxns9muTdtD6z0zXjsYj+mOKpF0lT7U1Y7a2lBXG+pqQ13tqG04sVimEsmFBEEAhQLel3Au0XeZSgzpAg0b6mpHbW2oqw11taGudtQ2nFgMxiPZaidV+QdWKJS3OdQyFW1hZERd7aitDXW1oa421NWO2oYTi2qZTAZo7Cc2V/3afL7vYDymy1Si6Cp9qasdtbWhrjbU1Ya62lHbcGKxZtx7j/e+sYN0zYxX1o2n0zBxYvl2TGfGI+kqfairHbW1oa421NWGutpR23BiMTMe2daG0HdHlf37Yzszri2MbKirHbW1oa421NWGutpR23BiMTMeiSBd/rPeXuMxnRkXERERkcbEYjAeyVY7lWUqvt5e4zEdjGsLIxvqakdtbairDXW1oa521DacWCxTiWY3lcox6g3GY7pMRVdN21BXO2prQ11tqKsNdbWjtuHEoloUV/U658oDci1T6aKrpW2oqx21taGuNtTVhrraUdtwYjEYr/7KpOF/JKmg/sz4nj1QKkEiFqt+ukTWVXpQVztqa0NdbairDXW1o7bhxGIwXiqVojlQEPScGa8OxkslaG+HKVOieZ0xIrKu0oO62lFbG+pqQ11tqKsdtQ0nFoPxyLbaCdI9Z8ary1SgPDses8G4tjCyoa521NaGutpQVxvqakdtw4nXuopGpQJ8Ptf999pP4YzpunERERERCS8Wg/Gottpx/S1TgVjuqKItjGyoqx21taGuNtTVhrraUdtwYrFMJZlMRnOg3hdw9l6mEjORdZUe1NWO2tpQVxvqakNd7ahtOLEYjKfT6WgONNDMeAwH45F1lR7U1Y7a2lBXG+pqQ13tqG04sRiM53Lldd4Nb7UT9JoZnzq1+3YMl6lE1lV6UFc7amtDXW2oqw11taO24cRiMF4sFqM5UCrA1w7G02mYNAk6OmI5Mx5ZV+lBXe2orQ11taGuNtTVjtqGE4vBeGRb7aQCqN1NBcpLVTo6Yjkzri2MbKirHbW1oa421NWGutpR23BisZtKVPrspgLdF3HGcGZcRERERBoTi8F4ZFvt9N5NBbov4ozhYFxbGNlQVztqa0NdbairDXW1o7bhxGKZSiIR0c8c9WbGq4PxGC5Tiayr9KCudtTWhrraUFcb6mpHbcOJxWA8sjVMvXdTgVgvU9HaMBvqakdtbairDXW1oa521DacWPwIk8/nyfceRIfRezcV6J4Zf+01KJUaf40xJLKu0oO62lFbG+pqQ11tqKsdtQ0nFjPjhUIhkuPUvYCzOhgvlWDfvp57j49zUXWVntTVjtraUFcb6mpDXe2obTixGIxHu7VhP8tUoLxUJUaDcf06yoa62lFbG+pqQ11tqKsdtQ0nFstUIjPQzDjEct24iIiIiIQXi8H4IdnaEGK3o4q2MLKhrnbU1oa62lBXG+pqR23DicUyFedcNAcK0lAs4H0J5yo/x/RephIjkXWVHtTVjtraUFcb6mpDXe2obTixGIw3NzdHc6AgKP+Zz0O6si4qxstUIusqPairHbW1oa421NWGutpR23BisUylUChEcoWvS1UG47Xrxmsv2IzZMpWoukpP6mpHbW2oqw11taGudtQ2nFgMxiPb97J2Zrz2vsmTy7djNjOu/URtqKsdtbWhrjbU1Ya62lHbcGKxTCXSrQ0Bev/UN20atLfHbmZcWxjZUFc7amtDXW2oqw11taO24cRiZjwy9WbGoXvdeMxmxkVERESkMbEYjEe21U5lMO7zuZ73V3dUidlgXFsY2VBXO2prQ11tqKsNdbWjtuHEYjAemWQKcP1/8E/MlqmIiIiISGNisWa8paUlkuM45yCV6rtMJaYz41F1lZ7U1Y7a2lBXG+pqQ13tqG04sZgZj3SrnSDof2b8tdegVIrmdcYAbWFkQ13tqK0NdbWhrjbU1Y7ahhOLwXikW+2kgv4v4PQe9u6N5nXGAG1hZENd7aitDXW1oa421NWO2oYTi2Uq6XQ6uoPVmxmvLlOB8lKV2k/lHMci7Spd1NWO2tpQVxvqakNd7ahtOLEYjCcSEf4CIAjw/c2MQ6zWjUfaVbqoqx21taGuNtTVhrraUdtwYjEYz2QyQDQ/sbmBlqlArHZUibKrdFNXO2prQ11tqKsNdbWjtuHoR5jhGsoyFRERERGRIYjFzHhzc3N0BxtsZjxGg/FIu0oXdbWjtjbU1Ya62lBXO2obTiwG46UotxsM0n1nxqdO7b4do2UqkXaVLupqR21tqKsNdbWhrnbUNpxYLFPJ5XLkcrnBnzgU9S7gTKXgsMPKt2M0Mx5pV+mirnbU1oa62lBXG+pqR23DicXMeBAE0R0sFUAu0/f+adNg375YzYxH2lW6qKsdtbWhrjbU1Ya62lHbcGIxGE+lonubLhXgD3T0fWDaNPg//ydWM+NRdpVu6mpHbW2oqw11taGudtQ2nFhU6+zsBCL6iS1I9V0zDt07qmzb1vhrjBGRdpUu6mpHbW2oqw11taGudtQ2nFisGY9UKoB8oe/9RxxR/nPXLvD+0J6TiIiIiIxJsRiMNzU10dTUFM3B6u2mAnDKKeU/X321vFwlBiLtKl3U1Y7a2lBXG+pqQ13tqG04sRiMRyoI8Pk6VwqfeWb37UcfPXTnIyIiIiJjViwG49lslmw2G8mxXL0P/QGYNw8mTizffuyxSF5rtIuyq3RTVztqa0NdbairDXW1o7bhxGIwHgRBdBcTBEH9ZSqpFJx+evl2TAbjkXaVLupqR21tqKsNdbWhrnbUNpxYDMZTqVR02+30NzMO3UtV/vAHeOmlaF5vFIu0q3RRVztqa0NdbairDXW1o7bhxGIwnslkyGTqfFBPGP3NjEPPdeO/+lU0rzeKRdpVuqirHbW1oa421NWGutpR23BiMRj33uOj2m4wCKBYxJdKfR9761vLj0MsLuKMtKt0UVc7amtDXW2oqw11taO24cTidwmRbrOTqgy2C3lI9zpuS0t5QP7447FYN67ti2yoqx21taGuNtTVhrraUdtwYjEzHiVXnfkebN34M8/Avn2H5qREREREZEyKxWA80q12amfG66kOxr2HTZuiec1RSlsY2VBXO2prQ11tqKsNdbWjtuHEYjAe6dW9g82Mn346OFe+Pc7XjeuqaRvqakdtbairDXW1oa521DacSAfjzrkvO+cec86tc84FNfef5Jx72jmXcc5NrLm/wzm3sfLf/4ryXGpFu894uvxnfzPjU6bAySeXb4/zdePaT9SGutpRWxvqakNdbairHbUNJ7LBuHPuZOBY7/2ZwPPAu2oe/gvwDmBzry/7vfe+rfLf/0R1Lr1ZLFPx/c2MQ/dSlSeegHG8xY9+HWVDXe2orQ11taGuNtTVjtqGE+XvEk4HHqjcvh9YBdwJ4L0/COCqyze6/ZVz7lFgC3Cl97525Dpp586dzJw5s+vq3Gw2SxAEpFIpMpkM3vsej6VSKYIgIJvNUiqVuh7r6OggkUjQ1NRELpejWCz2+LrqY/l8nkKh0OMx5xzNzc0UCgXy+TxN6TQOx8F9eykd1k5LS0vXY+l0mkQiQeGUU2gFyOXI/upXZN/61q7z7uzsBGj4PWWzWZLJJOl0uvH3VPMYUPc9VfcNbW5uplQqkcvlus6hvb193Lyn0fB9ymQy5PN5vPfj5j2Nlu9TZ2cnQRCwf//+cfOeRsP3KZvNksvlKJVK4+Y9jYbvU0dHB865cfWeRsP3KZPJkEql6OjoGDfvabR8nzo6OrrObby8J4vv04EDB6gV5TKVqUB75fY+YNoQvmaW9/7twE7gsl6PXbx+/fpITiwIAtLpdCTHwjkIUlAo9PuU4umnd91OjOMP/0mn09F1lS5NTU36NZ8R/Zu1kU6n9W/WQKT/v0u66H8H7GiZSjiuv83ZnXOfBe733j/pnGsD7gXywIXe+z5XJjrnPgrs996vdc6dAqzy3n+s13M2Aud67/f3uv9NwKe99++rue+0TZs2/XrBggUNvUHo/qkmqv0v82s+R/J/v4vEnJP7f9JJJ8Ef/gDLlsEvfhHJ6442UXeVMnW1o7Y21NWGutpQVztqOzRbtmxh7ty5c733W2DgmfF/AP5QuX0N8HngU8CN/Tx/E7C0cvts4PGBTsQ5N8E5l6z89Uxga6+ndEyePHmgQwxZ5GuYgqD/CzirquvGN20acBZ9LNPaMBvqakdtbairDXW1oa521DacgQbjh3nv251zk4C/Af7Ne38bcFK9J3vvnwZ2OeceA+YAP3bO/TuAc26qc+5B4GTgPufcOcCJwBOVNePLgZsie1e9JJNJksnk4E8cqiA98AWc0D0Y37+//AFA41DkXQVQV0tqa0NdbairDXW1o7bhDHQB50vOuTOANwGPe++LlW0JS/19gff+k73uurRy/2t0z5rXesswzzeUyNeGpQLI5wZ+ztvf3n370UfhlFOiPYdRQGvubKirHbW1oa421NWGutpR23AGmhn/AvAw8DW6l6YsBcbcNG8ulyOXG2TwPAxuKMtU3vAGOOaY8u1xut941F2lTF3tqK0NdbWhrjbU1Y7ahtPvYNx7fwcwBZjhvX+4cvcmYMWhOLEoFYtFisVidAdMBf1/AmeVc91LVX71K+jnQtmxLPKuAqirJbW1oa421NWGutpR23D6HYw752YCae/9Qedck3PuSuDvgd2H7Owi0tTUFO2VvUOZGYfuwfgrr8Dvfx/d648SkXcVQF0tqa0NdbWhrjbU1Y7ahjPQMpW7geMqt28APkD5g3y+Zn1So14whJlx6LtuXERERESkxkCD8VlA9SPq/x44FziLnh9zPyZEvtVOKsAPZWZ8zhyYOrV8exyuG9cWRjbU1Y7a2lBXG+pqQ13tqG04Aw3GPZB2zv0N0O69/xOwB5h4SM4sQolEgkQiug8bdUNZM15+YTjjjPLtcTgYj7qrlKmrHbW1oa421NWGutpR23AG2tpwI/Aj4HDgnsp9s4CXjc8pcpGvXwoCONAxtOeeeSb8x3/An/4E27bBcccN/jVjhNaF2VBXO2prQ11tqKsNdbWjtuEM9OPLB4FngQeBL1XuOwn4V+uTilo+nyc/lJnsoRrqzDj0XDc+zmbHI+8qgLpaUlsb6mpDXW2oqx21DaffmfHKB/V8ttd9PzM/IwOFqD+Ofqi7qQC85S3Q0gKdneXB+EUXRXsuIyjyrgKoqyW1taGuNtTVhrraUdtwBtraMOGc+7Rz7nnn3P7Kn592zo25zzk12dpwqD/5pdNw2mnl2+NsZlxbGNlQVztqa0NdbairDXW1o7bhDLRM5RpgJeVP4lwC/D/Aeyv3x9tQd1Opqu43/rvfwauv2pyTiIiIiIw5Aw3GVwLneO/v9N7/l/d+PeXtDVcemlOLTtRb7bjhzIxDz3Xjv/pVZOcx0rSFkQ11taO2NtTVhrraUFc7ahvOQIPxVvp+2ubuyv1jinMO51x0BxzOBZxQXqaSqizPH0cf/hN5VwHU1ZLa2lBXG+pqQ13tqG04Aw3GHwa+5ZybDuCcOwK4hfKWh2NKc3Mzzc3N0R1wOBdwAkyYUL6QE8bVuvHIuwqgrpbU1oa62lBXG+pqR23DGWgwfjlwDPCyc+4A8BJwLHDZoTixKBUKhWiv8A3Sw5sZh+5147/9LezfH925jKDIuwqgrpbU1oa62lBXG+pqR23D6Xcw7r1/2Xu/BHgdsAiYCfxvYPMhOrfImOwzXiriS8Whf0113XixCJvHXMK6tJ+oDXW1o7Y21NWGutpQVztqG85An8AJgPd+B7ADwDnXBLze+JwiZ/IJnFCeHW8a4k6PZ5zRffvRR2Hp0mjPaQRo+yIb6mpHbW2oqw11taGudtQ2nIGWqfTHR34WY4xLVQbjw1k3fvjhMGdO+fY4WjcuIiIiIuGFGYyPOZFvtVM7Mz4c1XXjmzdDLhfd+YwQbWFkQ13tqK0NdbWhrjbU1Y7ahtNnmYpz7uPDeX4shZkZh/K68W9+EzIZePJJWLAg+nMTERERkTGj3uD6gkG+ZsxtlN3S0hLtARudGYfyuvExPhiPvKsA6mpJbW2oqw11taGudtQ2nD6Dce/9opE4EUvVbXaC6iC6UZXj+EKeYW1t/7rXwetfD3/8Y3nd+D//czTnM0Ii7yqAulpSWxvqakNdbairHbUNJxZrxiPfaieRBOeGPzMO3bPjjz8OpVJ05zQCtIWRDXW1o7Y21NWGutpQVztqG04sBuPpdJp0Oh3Z8Zxz5XXjYf7BVfcb37sXnn02snMaCVF3lTJ1taO2NtTVhrraUFc7ahtOLAbjiUSCRCLitxoEw7+AE/quGx/DTLqKuhpSWxvqakNdbairHbUNJxbFMpkMmUwm2oMG6XAz4yedBDNmlG+P8f3GTbqKuhpSWxvqakNdbairHbUNJxaDcROpFD7MzLhz3bPjjz0GPvafoSQiIiISW7EYjDc3N9Pc3BztQYOQa8ahezC+cye8+GJ053SImXQVdTWktjbU1Ya62lBXO2obTiwG46VSiVLEO5e4sBdwwrhZN27RVdTVktraUFcb6mpDXe2obTixGIzncjlyUX/8fNgLOAFOPhkmTSrfHsPrxk26iroaUlsb6mpDXW2oqx21DScWg/EgCKLfgL6RmfFkEs44o3x7DA/GTbqKuhpSWxvqakNdbairHbUNJxaD8VQqRSrV58NGGxOkw8+MQ/dSla1by2vHxyCTrqKuhtTWhrraUFcb6mpHbcOJxWC8s7OTzs7OaA8aBOF2U6mqXTc+RmfHTbqKuhpSWxvqakNdbairHbUNJxaDcQvlCMU9lQAAIABJREFUCzgbWBd16qlQ/ZSqMToYFxEREZHGxGIw3tTURFNTU7QHbWTNOEBzM7ztbeXbd98NY/DqY5Ouoq6G1NaGutpQVxvqakdtw4nFYNxEI7upVL3rXeU/d+yAjRuH9CW5jg4e/fjHKejXQCIiIiJjXiwG49lslmw2G+1BG50ZB1i5Eqo/QX7720P6kr0vvMBfNmygY9u2xl47AiZdRV0Nqa0NdbWhrjbU1Y7ahhOLwbjJVjtBgC8UGjvGtGlw4YXl2z/+Mbz66qBfcmDHDgA6X365sdeOgLYwsqGudtTWhrraUFcb6mpHbcOJxWDcZmvDCGbGAT784fKfuRysWzfo0w9s3w5A5+7djb92g7SFkQ11taO2NtTVhrraUFc7ahtOLAbjmUyGTCYT6TEb3k2l6h3vgBNPLN++9VbwfsCnj6aZcYuuoq6W1NaGutpQVxvqakdtw4nFYNx7jx9kkDtsUVzACeAcfOhD5du/+x38+tcDPn1/dWb8lVcaf+0GmXQVdTWktjbU1Ya62lBXO2obTiwG46Nya8Nal1wC1V/r3HrrgE89sH07yebmUTEY1xZGNtTVjtraUFcb6mpDXe2obTixGIybiGpmHODII+Gd7yzf/uEPYd++uk8rFYsc3LmTaXPmjIrBuIiIiIg0JhaDcZOtdoIASiV8qRjN8aoXcnZ2wg9+UPcpmVdeoVQocMS8eaNizbi2MLKhrnbU1oa62lBXG+pqR23DicVg3OLqXpeqbN0T1VKVs86C448v3+5nqUp1vfjhJ59M5pVXRnxdlq6atqGudtTWhrraUFcb6mpHbcOJxWDcap9xILrBeCIBH/xg+fZTT8GTT/Z5yoEdO2iePp2Jr3sdpUKB7N690bx2SNpP1Ia62lFbG+pqQ11tqKsdtQ0nFoNxs0/ghOjWjQOsWlUelEPd2fED27cz4dhjaTniCKC8bGUk6ddRNtTVjtraUFcb6mpDXe2obTixGIyXSiVKpVK0B416Zhzgda+D5cvLt3/wA9i/v8fDB3bsYOIxx9A0ZQqJVIqDI7xu3KSrqKshtbWhrjbU1Ya62lHbcGIxGDfZaidIl/+McmYcui/k7OiAH/2ox0P7KzPjLpGgefr0Ed9RRVsY2VBXO2prQ11tqKsNdbWjtuHEYjBuorJMxUc5Mw7lmfGjjy7f7rVUpbpMBaBlxowRX6YiIiIiIo2JxWDcYg2TSybBJSCfi/S4pFLlteMAmzfDs88C3XuMTzjmGKA8GB/pZSpaG2ZDXe2orQ11taGuNtTVjtqGE4vBeDKZJJlMRn/gKD/4p1Z1VxXomh2v7jE+sTozPn36iM+Mm3WNOXW1o7Y21NWGutpQVztqG04sBuPpdJp0Oh39gVNBtBdwVp1wAixdWr69bh1kMl17jLdWlrC0zJjBwREejJt1jTl1taO2NtTVhrraUFc7ahtOLAbjuVyOXC7i5SRgNzMO3RdyvvYa/PjHXXuMp5qbAWg54ogRnxk36xpz6mpHbW2oqw11taGudtQ2nFgMxovFIsViRB9bXysI8PlC9McFeOc7Yfr08u1bb+1x8SaUB+OdI/wpnGZdY05d7aitDXW1oa421NWO2oYTi8G42VY7KcOZ8aYmuOSS8u1HHuHAc88xsXLxJpQH46V8nty+fTavPwTawsiGutpRWxvqakNdbairHbUNJxaDcSsuCKLfTaXWhz7UdXP/U0/1nBmfMQOAzhHeUUVEREREwovFYNxsqx2rCzirZs+GM88E4MDLLzPhyCO7HmqaMgWXSo3oB/9oCyMb6mpHbW2oqw11taGudtQ2nFgMxhOJBImEwVu1vICz6sMfpgQcTCSYsG1b190ukaBl+vQRnRk36xpz6mpHbW2oqw11taGudtQ2nFgUM1vDFBjPjAO8611kpk2j5BwTH3igx0MtRxxB5+7dtq8/AK0Ns6GudtTWhrraUFcb6mpHbcOJxWA8n8+Ttxg0pwK89cx4Swv7ly8HoPWRR+CPf+x+aMaMEZ0ZN+sac+pqR21tqKsNdbWhrnbUNpxYDMYLhQKFQvRbEDrL3VRqHDjtNJrzeVKlEnznO133txxxxIgOxq26xp262lFbG+pqQ11tqKsdtQ0nFoPxMb1MBTiQTDKhev7f/S5U/qFrmcr4pK521NaGutpQVxvqakdtw4nFYNyM9W4qFQd27GDirFnlv+zYAT//OTDyy1REREREpDGRDsadc192zj3mnFvnnAtq7j/JOfe0cy7jnJtYc/8nnHOPO+d+6pybHOW51DLbaidIH5JlKvu3b2fCaafBhAnlO269FRj5T+HUFkY21NWO2tpQVxvqakNd7ahtOJENxp1zJwPHeu/PBJ4H3lXz8F+AdwCba54/HTgPWAj8ELgsqnOpc24456I/cBDgD8XM+PbtTHjDG2DFivIdP/sZbN8+4p/CadY15tTVjtraUFcb6mpDXe2obTipCI91OlDde+9+YBVwJ4D3/iDQ+xt0KvCI99475+4HvtfreJN27tzJzJkzu9YfZbNZgiAglUqRyWTw3vd4LJVKEQQB2WyWUqnU9Vg+nyeRSJDP58nlchSLxR5fl0gkaGpqIp/PUygUejzmnKO5uZlCoUA+n+/xWLJQIJ3P0dnZST6fJ51Ok0gkyGQyADQ3N1Mqlcjlcl3n3dnZCTDk95R0jgM7d5KcNo39730vE7/9bSiVyNx8M/4f/gGAPdu2MaHyHhp9TwAtLS1djw30npxzJJNJ2tvbh/We+vs+ZbNZkskk6XQ60u/TcN5T2O9TlO8pkUiQzWbp6OgYN+9ptHyfvPcEQcD+/fvHzXsaLd+nbDZLe3v7uHpPI/19yufzOOe6dqgYD+9pNHyfAFKpFB0dHePmPY2W71N1J5Xq14yH92TxfTpw4AC1olymMhVor9zeB0xr8PkXr1+/PpITKxaLFIvFSI7VwyFYM57ZvRtfKNB69NGU3vIWSpW1401f+xpNO3bgUikyI/QpnLpq2kahULD59yoUi0X9mzVg9r+xMaeuNvS/A3b0bzYcF9V6Y+fcR4H93vu1zrlTgFXe+4/1es5G4Fzv/X7n3HJgvvf+Gufc4cD3vPfn1jz3tE2bNv16wYIFDZ9bdeZ28uRol6WXnn2K4s/vJfjkFyI9bq2Xn3ySB9/3Pt795JOkmpvh4YdhyRLwHubM4d4ZM/ibj3+cE84/3+wc+mPVNe7U1Y7a2lBXG+pqQ13tqO3QbNmyhblz58713m+BaGfGNwFLK7fPBh4f5PlPAG8f4PkdUX0zTbc2NL6A88COHTRPn14eiAMsWgSf/Wz59pYttBw8SOcIzYxrCyMb6mpHbW2oqw11taGudtQ2nMgG4977p4FdzrnHgDnAj51z/w7gnJvqnHsQOBm4zzl3jvf+FeBnzrnHgfcCN0d1LodMkDZfpnJg+3YmHHtszzuvvhrOOAOAlu3b6fz1r03PQURERERsRHkBJ977T/a669LK/a/RPWte+/yvAV+L8hzqqS6+j/yntVQAvoQvFnHJZLTHrjiwYwcTjzmm1+um4Ac/gHnzaCkU6Ny4Ef74R3j9603OoT9mXWNOXe2orQ11taGuNtTVjtqGow/9aYALKlupGy5V2V9vZhzguOPgu98tD8a9L297eAi2WRQRERGR6MRiMN7S0kJLS0v0B05VBuP5XPTHrqi7TKXq/PNpOfNMOlMp2Ly5vHzlEDLrGnPqakdtbairDXW1oa521DacWAzGzbbgq86MG81Il4pFDu7cyYTey1RqtFx2GZ3pNB7ghhvgl780OZd6tLWhDXW1o7Y21NWGutpQVztqG04sBuPVD0yIXMp2mUrmlVcoFQpM7G9mHGg55hhKQG7ixPJ2hxdfDLt2mZxPb2ZdY05d7aitDXW1oa421NWO2oYTi8F4Op0mnU5Hf2DjmfH927cD0Hr00f0+p2XGDAA6P//58h27dsH73gelksk51TLrGnPqakdtbairDXW1oa521DacWAzGE4kEiYTBW60Mxr3RzHifPcbraJo6FZdK0Xn66XDRReU7H3gAbrzR5JxqmXWNOXW1o7Y21NWGutpQVztqG04simUyGTKZTOTHdYkkJBJmM+MDXrzZdQ6J/5+98w5vsmof8P1mNWmhZbQFWQW62DJENrJXGUIBAQEBBURR0Y8PxYE/ceL4cKGoyN5776kigoAM2zLKLrNltNCd5P39kbZQaEuT5rS0Ofd19SJ5x3lP7r6kT06e8xxM3t62hX9+/BH8/W073nkH9u4V0q90RHl1daRXcUi3YpBexSC9ikF6FYd06xguEYwLRacXF4xnVWM8C4zpwXjx4rBwYdrKoGbo1w9u3RLSN4lEIpFIJBJJ3nGJYNxoNGLMIdUjT+j1wiZwZltj/D7cfX1twTjAE0/ApEm2x2fPwogRtomdAhDq1YWRXsUh3YpBehWD9CoG6VUc0q1juEQwbrVasYqa0Kg3FGiaCoDRx+duMA4wZgyEhNgeL1kC06YJ6Z9Qry6M9CoO6VYM0qsYpFcxSK/ikG4dwyWC8ZSUFFJSBC3Mo9MLmcCZmxrj6Zh8fEi8du3uBkWBGTMg/dxXX4WwMKf3UahXF0Z6FYd0KwbpVQzSqxikV3FIt47hEsG4Xq9Hn16G0PmNCxkZz02N8XQypamk4+MDc+faAvOkJHjmGUhMdGofhXp1YaRXcUi3YpBexSC9ikF6FYd06xguEYzrdDp0Op2QthWdTkgwnpsa4+mkp6mo9+eGt24N775rexwWZlsQyIl9FenVlZFexSHdikF6FYP0KgbpVRzSrWO4RDCemJhIopNHhTMQNIEzNzXG03H38cGakkJKbOyDOydMgLp1bY+XLbNVWHHSV0hCvbow0qs4pFsxSK9ikF7FIL2KQ7p1DJcIxoUiqLRhbidvgm1kHCAxJubBnTodbN9uq7ICsHw59O3rtIBcIpFIJBKJROI4LhGMu7m54ebmJqZxvUHYyHhuaowDGEuVQtFqM0/ivJeSJWHLFmjUyPZ81SoIDYXk5Dz1UahXF0Z6FYd0KwbpVQzSqxikV3FIt47hEsG4UPRiqqnktsY42FbhzFj4JztKlIBNm6BJE9vztWuhZ0/b5E6JRCKRSCQSSYHgEsF4cnIyyXkcBc4O5RFIU4EsyhtmhZeXLSBv3tz2fMMG6NHD4SorIr26MtKrOKRbMUivYpBexSC9ikO6dQyXCMaFltrR6SHVufnX9tQYT8fk45N1zvj9FC9uC8JbtrQ937wZuneHhAS7+ylLGIlBehWHdCsG6VUM0qsYpFdxSLeO4RLBuNBSOwKqqdhTYzwdk6/vw0fG0ylWDNavt5U+BNi6Fbp2hfh4u/opSxiJQXoVh3QrBulVDNKrGKRXcUi3juESwXhSUhJJonKjBSz6Y0+N8XRM3t4k5ZQzfj8eHra88XbtbM937IAuXeDOnVw3IdSrCyO9ikO6FYP0KgbpVQzSqzikW8dwiWBcVdUHF8RxFjrnT+C0p8Z4OiZfXxLsCcYB3N1h9Wro0MH2/LffoFMnuH07V6cL9erCSK/ikG7FIL2KQXoVg/QqDunWMVwiGBdb2tD5I+P2Tt6Eu2kqdv8nMJlspQ47d7Y9370bOnaEuLiHnipLGIlBehWHdCsG6VUM0qsYpFdxSLeO4RLBuEgUnR7MZqe2aU+N8XRMaatwpuYiiH4AoxFWrLDljQPs2WMbLb91y/62JBKJRCKRSCS5xiWCcaGldvTOr6ZiT43xdExpq3DanaqSjpsbLFtmK3UIsHcvtG8PN29me4osYSQG6VUc0q0YpFcxSK9ikF7FId06hksE40Jn9wqoM+5Imkr6Kpx2TeK8H4MBFi+GXr1sz/fvh6efznalTjlrWgzSqzikWzFIr2KQXsUgvYpDunUMlwjGHal7GR0dTWpuguy00obOmrDgSI1xuLsKZ0Juyxtmh8EACxdCq1a257/9BsOHQxavT9YTFYP0Kg7pVgzSqxikVzFIr+KQbh3DJYJxe7822bJlC8GBgQwfNuzhB+v1tkDVaslDD+/iSI3xdOwub5gder1tYaAmTWzP58yBjz564DD5dZQYpFdxSLdikF7FIL2KQXoVh3TrGC4RjFutVqxWa66PfXPcOG7GxjJr7ly2b9+e4/GKLu0ToJNSVRypMZ6OQ+UNs8NotFVZqVLF9nzCBFiwINMh9niV5B7pVRzSrRikVzFIr2KQXsUh3TqGSwTj9pTa0Wg0zF+wAGPa1ywDBw4kJqdl5vXODcYfVmM8ISGBrVu3ZvnJ0+Tj45yR8XR8fGDdOvDysj0fMsRW+jANWcJIDNKrOKRbMUivYpBexSC9ikO6dQyXCMbtpVq1anw8diwAly9fZujQodnnhKePjDtp4Z+cJm+qqkqHDh1o3749zZs358aNG5n2m3x8SMxrzvj9VK9uq7Ki00FKim1C56lTzr2GRCKRSCQSiYviEsG4IzlMYz78kGa+vgCsXbuW77//PusD9Qbbv04cGc+uxvjatWvZnTYyvX//ftq2bUv0PSPhJh8fEnMaxXeUtm1h6lTb45gYCAmBmzdlbpggpFdxSLdikF7FIL2KQXoVh3TrGC4RjGu1WrRarV3naLRaPn31VXzd3QEYO3Yshw8ffvDAtDQV1Ukj49nVGE9OTuaNN97ItO3QoUO0atWKK1euAHlYhTM3PP88vPmm7fHx49CrF1qLxW6vkofjyP0qyR3SrRikVzFIr2KQXsUh3TqGSwTjBoMBg8Fg93m1n3mGUb6+aLVaUlJS6NevH/Hx8ZkPcvIEzuzSVL799lsiIyMB+Oqrrxg0aBAA4eHhPPXUU0RFRWHy9cWSnEzq7dtO6csDfPIJ9O5te7xzJ6YxYzDIEkZOx9H7VfJwpFsxSK9ikF7FIL2KQ7p1DJcIxlNSUkhJsX+VzBIBATRq0IAXO3QA4NixY4wZMybTMYpGAxqtU3LGs6sxfuXKFT788EMAqlevziuvvMLMmTN54YUXADhx4gQtW7bkWkICgPPzxtPRaGD2bHjySdvT2bNRP/lEzLVcGEfvV8nDkW7FIL2KQXoVg/QqDunWMVwiGLdYLFgsjtUBr9KtG62TkmiVtgjOtGnTWLx4ceaD9HpIzfvNl12N8XfeeYfbaaPdkydPRq/Xo9Fo+Omnnxg9ejQAZ86coUPPnlwxm0l0ZkWV+zGZYPVq8PMDwPDBB7ZVOyVOIy/3qyRnpFsxSK9ikF7FIL2KQ7p1DJcIxvNSasevSxdSoqOZ/MYblCpVCoARI0Zw9uzZuwfp9E5JU8mqxvj+/fuZMWMGAN26daNjx44Z+zQaDd9++y3//e9/Abhw4QIfnT3L0YMH89yXHClTBtatQ/X0tD0fPBj++kvsNV0IWRpKHNKtGKRXMUivYpBexSHdOoZLBON5weTjQ9mmTUnauzcjKI6NjWXAgAGkpgfger1T0lTurzGuqiqvvvoqqqqi1+v56quvHjhHURQmTZrEe++9B8DNlBT6jR/PkSNH8tyfHKlZk9T581G1WkhOhu7d4cwZsdeUSCQSiUQiKWK4RDCe11I7Vbp14/zmzXRp3z4jLWTPnj188MEHtgP0elQnjIzfP3lzwYIF7NmzB4AxY8YQGBiY5XmKojBx4kQ+/vhjAG7cvk3r1q05cOBAnvuUE0ktWpCU/gEhOtpW8vDWLaHXdAVkaShxSLdikF7FIL2KQXoVh3TrGC4RjGs0GjQax19qhbZtQVWJ2rGDL774gtq1awPwySefsGPHDhSd80bG02uMx8fHM27cOAB8fX159913H3r+22+/zWtpk01v3LhB27ZtM4J5EWg0GizPPw9pCyQREWGrtuKkyjKuSl7vV0n2SLdikF7FIL2KQXoVh3TrGC5hLK85TDqTiYrt23N2zRqMRiMLFy7EZDKhqioDBw4kJikZUs157ue9NcYnTZrExbQc8k8//RTP9Pzsh/B8jx6MbtgQsKXTdOjQgd9++y3PfcuKDK+TJkHPnraN27ZB//621TolDiFz7sQh3YpBehWD9CoG6VUc0q1juEQwnpqaeje/20GqdO/O5d27SYyJoUaNGnzzzTcAXLp0ieEzF6Cm5P1rmfQ0lbNnz/LFF18A0KBBA4YMGZLrNkw+PrQvVYpp06ahKAp37tyhU6dObN26Nc/9u58MrxoNzJ0LNWrYdixbZgvOExOdfk1XwBn3qyRrpFsxSK9ikF7FIL2KQ7p1DJcIxs1mM2Zz3kaufRs2xOjtzbkNGwB44YUX6J22AM7af47ww4rVeWr/3hrj48aNIykpCbAt9mPPVz4mHx8So6MZNmwYs2fPRqPRkJiYSJcuXXjttddYv379gwsXOUgmr+7usHs3NGlie75+vS2H/M4dp1zLlXDG/SrJGulWDNKrGKRXMUiv4pBuHcMlgnFnfG2i0WqpHBLC2TVrANukyZ9//plKlSoBMG7aTA4fPuxw++k1xg9FRbFkyRIABgwYQNOmTe1qx+TriyUpidTbtxk4cCCLFi1Cq9WSmprKt99+S0hICKVKlaJt27Z8/vnnHD58GFVVHerzA15LlIDNm6F1a9vzHTugQwc5qdNO5Nd84pBuxSC9ikF6FYP0Kg7p1jFcIhh3FlW6d+dGWBixp04BULJkSebPn49Go5Ccmkq/fv0cHnW+c/EiVlVl/KefAuDu7s6kSZPsbsfk4wOQsfBP79692bRpEy1btkSn0wG2FbK2b9/Om2++Sd26dXnssccYPHgw8+bN41peV+8sVgzWrYMuXWzP9+yBNm0gJiZv7UokEolEIpEUQVwiGHdWqZ0SgYGUrFaNM2mj4wDNmjVjQv++ABw7dozWrVsTERFhd9vxly7xR2pqRn3wt956iwoVKtjdjlupUigaTaZVONu2bcuuXbu4fv06q1at4qWXXsLf3z9j/9WrV5kzZw4DBw6kTJkyNGjQgPHjx/PHH3/kOGqerVeTCVasgNBQ2/N//oGnnoLLl+1+Pa6ILA0lDulWDNKrGKRXMUiv4pBuHcMlgnFFUVAUxSltVe7WjbNr16JarRnbxj/7DC2rBwHw999/U6tWLUaMGMGlS5dy3e6lEydYcP48AJUqVWJserlAO9FotRi9vUnMYoTb09OT7t27M2XKFCIjI4mMjOSHH36gR48eFC9ePOO4gwcP8tlnn9GiRQtGjhyJ9Z7Xei85ejUYYOFCGDjQ9jw8HFq2hLTXKMkeZ96vksxIt2KQXsUgvYpBehWHdOsYLhGMG41GjGmrWuaVyiEhJF69yrX9+zO2ad2MLB09gqFDh6LRaLBarfzyyy8EBgby3nvvERcX99B2v1+yhLi0coBffvklJpPJ4T6mT+J8GP7+/owaNYqVK1dy/fp1du3axdtvv02DBg0yjvnll18YPXp0liPkD/Wq08GsWTBypO15ZCS0aGH7V5ItzrxfJZmRbsUgvYpBehWD9CoO6dYxXCIYd+bsXpOPD2WbNMmUqoLeQEk3PdOnT+fw4cOEhIQAkJCQwEcffYS/vz/fffcdKdnU3o6IiGDJvn0APPXUUxlVWvLSx6xGxnNCr9fTsmVLPv74Y/bv38/58+cJDg4G4Mcff+T1119/ICDPlVeNBn78EV5/3fb8/HlbQB4WZlf/XAk5G10c0q0YpFcxSK9ikF7FId06hksE486ue1m5e3cubN6MOa38IHo9alr7tWrVYu3atezcuZOGaYvvxMTE8Oqrr1KjRg0WLVqUKe1DVVVef/11LKqKRlH4+uuv8/wVj8nHh8Q8TpisWLEi27dvJyAgAIBvvvmGcePGZQrIc+1VUeCrr+C992zPr1yx5ZAfPJinPhZVZJ1WcUi3YpBexSC9ikF6FYd06xguEYw7u9ROxTZtUC0WLu7YYdug04M588331FNPsXfvXhYvXpwR0J46dYp+/frRqFEjdqSdu379ejZt2gTAgO7dqVu3bp77Z/L1tXtkPCvKlSvH9u3bqVKlCmBLn3n33XczAnK7vCoKTJwIn31me379uq3Kyp49ee5nUUOWhhKHdCsG6VUM0qsYpFdxSLeO4RLBuLPRubtTsX37jFQVRa+HLD4JKopCnz59CAsL4/vvv8cnrezg/v37adOmDV26dGHMmDEAuGs0fPDuu07pn8nbO1c547khfYS8YsWKAHzyySd8+OGHjjf45pvw3Xe2x7Gx0L69rR65RCKRSCQSiQviEsG4iFI7Vbp35/Iff5B0/bptZDw163xwAIPBwMsvv8ypU6eYMGEC7u7uAGzYsIHItMmMvXx8qFSrllP6ZvL1JTE62uHFfO6ncuXK7Nixg3LlygHw/vvv89lnnznudfRo+PVX22h5fDx06gTTp4OT+lvYkaWhxCHdikF6FYP0KgbpVRzSrWO4RDAuAt8nn8RYujTnNmwAvR5yMWGhePHifPDBB0RGRjJy5Ei0Wi0A5X196RYYiM5JM5BNPj62VTiduBS9v78/27dvp0yZMgCMHz+e77//3vEGhw2D+fNtEzxTUuD556FXL3DSiL5EIpFIJBJJYcAlgnGTyZSnUoFZodFqqRwSwpnVq23BeGpqrkeiH3vsMaZOnUpYWBiTJ09m9uuv45WWBuIMTL6+AE7JG7+X4OBgtm/fnpFu88477zBjxgzHG+zXDzZtgscesz1fuRJq17at4OnCiLhfJTakWzFIr2KQXsUgvYpDunUMlwjGRZXaqdK9OzfCwoi9dBlQwWKx6/zg4GDGjBmDR0ICxdJSQJxBVqtwOosaNWqwdetWSpUqBcCYMWP46aefHG+wXTs4ehTSyzlevQpdu8KoUbYUFhdEloYSh3QrBulVDNKrGKRXcUi3juESwbioUjslgoIoERzMuR27bBvMjl3jzsWLeJQv77R+abRajKVLCwnGAerUqcOWLVvw8vIC4MUXX2T69OmON1i6NCxeDLNng6enbdvUqVCvHqTVX3clZGkocUi3YpBexSC9ikF6FYd06xguEYwbDAYMBoPlJnCCAAAgAElEQVSQtqt0787Z7TtsKSoO3oDxTg7GwXnlDbOjfv36rF27Fs+04PmFF15g7ty5jjeoKDBoEBw5Ai1b2radPAlNm8IHH+QqJ7+oIPJ+dXWkWzFIr2KQXsUgvYpDunUMlwjGNRoNGo2Yl1o5JITE6Giio2/kWFElO6wWCwmXL+PhxDQVAKMTyxtmx5NPPsnq1aspVqwYqqry3HPPsXjx4rw16ucH27fD55/bcvEtFvi//4NmzWzBuQsg8n51daRbMUivYpBexSC9ikO6dQyXMJaUlERS+mqZTsbk40OZJxty7twlh9JUkqKjsZrNFHPyyLh7WnlDkSQlJVGvXj3WrVuHu7s7VquV/v37M2nSpLx9TaXVwn//C3//DTVr2rbt2wd168JPPxX5Eogi71dXR7oVg/QqBulVDNKrOKRbx3CJYFw0lbt2JSrqqkOlBO9cvAiAe3pFESdh9PERHoyn07JlS9asWYPBYMBqtfLWW29Rr149fv/997w1/PjjsH8/vPGG7XlCArz4InTrBleu5L3jEolEIpFIJAWMSwTjRqMRo5NqeGdFxQ4dUFWVS3/stvvc+EuXMHp7O63GeDruPj5Cc8Yhs9c2bdqwc+dOqlevDkBYWBgtW7Zk6NChROflQ4HRCF99Bdu2QYUKtm3r1kG1ajBrFliteX0Zjxyi71dXRroVg/QqBulVDNKrOKRbx3CJYNxqtWIVGLTpPYpRoUJZzm7bbve5IiZvQtrIeEyM01bhzIr7vTZp0oRDhw7xySefZNQZnTlzJtWqVWPatGl5+x20aWOb3DlggO15bCwMGQItWsChQ3l4FY8eou9XV0a6FYP0KgbpVQzSqzikW8dwiWA8JSWFlBT7J1fag5+/H5cP/sPpVatIsCOFIv7SJafWGE/H3dcXS2KiU1fhvJ+svBoMBsaPH09YWBghISEA3Lhxg+HDh9O8eXOOHDni+AVLloR582xlENOd/fknNGgAr7wCt2453vYjRH7cr66KdCsG6VUM0qsYpFdxSLeO4dRgXFGUSYqi/K4oyhxFUfT3bNcqijI9bd/X92y/rSjKzrSf2s7sy73o9Xr0ev3DD8wDvhXLUbllcw5+9hkr27ZldefO7H3/fc6uW5dj7raza4ynY0xbJdPRvPH4S5fY/8knOY6s5+S1SpUqrFmzhuXLl1MhLb1kz5491K9fn7Fjx3InLx8S+vSBY8dg7FjQ6WypKt9/D0FBMGNGoU9dyY/71VWRbsUgvYpBehWD9CoO6dYxnBaMK4ryOFBeVdUWwDGg9z27uwKX0vZ5KIrSJG37cVVVW6X9HHVWX+5Hp9Oh0+lENQ+A4uZGo+eHELp7N52XLiWoXz+SYmL4+8MPWdGqFWu7dePviRM5v2kTSTduZJwnLE0lfRVOB/PGw6ZN48S8eTkG8w/zqigKPXv2JCIigrFjx6LVarFYLHz11VdUr16d5cuXO55GU7w4fPEFHD4MrVvbtkVHw7Bh0Lw5HDzoWLuPAPlxv7oq0q0YpFcxSK9ikF7FId06hjONNQU2pz3eCAwFFtyzb909+5oBewB/RVF+A8KA11VVvbceTvHLly9TsWJF3NzcAEhOTkav16PT6UhKSkJV1Uz7dDoder2e5ORkrFZrxr4bN26g0WgoVaoUKSkpWCyWTOdpNBrc3NxITU3FbDZn2qcoCkajEbPZTGpqaqZ9ACaTCbPZjEajwZqYgGI2oy1fnnI9e1K1f38sqalE//svNw8d4vqBA5xeswZLQgKeAQH4PPEE8Zcvo/f2JjU11a7XlJycjFarxWAwZP+aSpcm9uJF3OPi7HpNmsRETq9cCUB0WBiPlSiRUarIaDRitVpJSUnJ6EP6f7ycfk8fffQRzzzzDKNHj2bv3r1ERUURGhpKhw4d+Oabb6hSpUruXtP9vyd/f5JXrEC/YgXGd95BuXQJ9uxBbdgQ6/DhxI8fDyVLZvyeUlNTMRgMaDSaLF9Ter8TExMf+pqc9nu67zXdvn2b1NRUihUrlqt7rzC8Jnv/P4l6TYmJiej1+ozKP0XhNT0Kv6f4+HhSUlLw8PAoMq/pUfg93bhxA0VRKF26dJF5TY/C7ykpKSmj70XlNT0qv6cbaYON3t7eReY1ifg9xcfHcy/OTFMpCcSlPY4FSuViX4Cqqi2By8DL97U3aOHChU7snmB0+ixXiVS0WkpUr07Q4MG0mDKFjps20WzaNCp17szts2dxK1kST39/IV0yenuTFBNj93mRixdj9PamWOXKxJ065bT+1K5dm40bN/LDDz9QqpTtFti8eTO1a9cmJCSEKVOmcOzYMftHyxUFc2go5qNHSX39dVSdDsVqRfvTTxRr0AD9nDmFPnVFIpFIJBJJ0URxVrUNRVFeAu6oqjpbUZQGwFBVVUen7fscWKuq6m+KooQCVVRV/fKec2sAb6mqOviebY3//PPPPU2aNCGvpH+qSf90IgLzjCkoVQPRPtVB2DXsZedLL1G8UiUavPVWrs8xJyWxql07ao0aRcyhQ2jd3Gj80UdZHpsXrzExMYwbN44ZM2Y8sK9SpUp06tSJTp060bZtWzw9Pe1rPCICRo+2reSZTuPGMHEilCqV/XlZ4ecH3t72nZNH8uN+dVWkWzFIr2KQXsUgvYpDus0dYWFh1KpVq5aqqmHg3DSVP4E3gNlAR2D3ffvaAb+l7ZuhKIoHkKSqqgVoAUTe195tu4OwgkSvg9RHawaxKa28oT2cXb0a1WKh6tNPY46P58K2bUL65u3tzfTp0xk+fDizZs1i48aNnDt3DoDz58/z888/8/PPP6PT6WjWrFlGcP7444+jKErOjVevDlu3wpIltgWDLl6Ev/6CDg58UDIYbCUVg4MdeJUSiUQikUgkOeO0NBVVVQ8BVxVF+R2oCSxTFOWntN1rgUpp+5JUVd0DBAJ/p+WMdwG+cVZf7ic5OTnj05ow9AZIfTBNJSfUWzcwz/ge9bqYlTJNvr52TeBUrVYiZs0i4Jln0Ht44BUQQOypU6jZpHg4w2uTJk2YOnUqZ86cISIigsmTJ9OxY8eMT9Vms5ldu3Yxfvx46tWrR7ly5RgyZAiLFy8mISEh+4YVBfr2tVVdGTcOtFrHOpiSArNnO3aug+TL/eqiSLdikF7FIL2KQXoVh3TrGE6d8qqq6n/v2zQybbsZGHLfsYeA+s68fnbkS5kdnR7VnJrrw1WrFcuK+ajnz2DZsgZdv2FO75LJx8eu0oYXd+4kPiqK4GefBaBEUBCWxETuREVRvFKlB453pldFUahWrRrVqlVjzJgxJCQksGvXLjZu3MjGjRs5ceIEAFeuXGHWrFnMmjWLYsWKERoaysCBA2ndujXarALuYsVg0iR4+WVb5RV7eOcdOHoUVq6Ejz92wqvMHbIslDikWzFIr2KQXsUgvYpDunUMl6g/ky9ldnR6sCMYt/6xHfXqZbTPDMWyaCbWM5FoqgQ4tUvpwbiqqg9P7QAiZs6kcrdumNJqlHuUK4fOZCI2MjLLYFykV3d3dzp37kznzp0BOH36NJs2bWLjxo1s3ryZpKQk7ty5kxGYlytXjgEDBjBw4EDq1Knz4OutVMn2Yw///msLxsPD4cQJWx3zfECWhRKHdCsG6VUM0qsYpFdxSLeO4RIrcCYlJWWUnxGFotdDau6CcevF81h3bkLbtTeaarVQHm+AZfNqVNW5FT9MPj5YEhMx31dCJytijhwh+sABqg0ZkrFN0WjwCgjg1smTWZ6TH17TqVq1KqNGjWLVqlXcunWLFStWEBoaisFgAODSpUt8+eWX1K1blzp16jBp0iQuXLiQt4s+/fTdx2mlHvOD/PTqaki3YpBexSC9ikF6FYd06xguEYyrqur44jK5JZcj42pKMpZlc1Hq1EdTqx4A2jZdIOYa6uEDTu2SydcXgIRc5I0fmzWLx1q0oERA5tF5r8DAbIPxfPGaBW5ubjz99NMsXbqUK1eu8PPPP9OyZcuM/f/++y9vvfUWfn5+tGnThunTpxMbG2v/hapXvztxc8UKJ/X+4RSUV1dAuhWD9CoG6VUM0qs4pFvHcInvE/KlxI5en6tqKpYNK0FV0XbumbFN8fRC07QVlu3rUWrUQTE4p7/GUqVAUUiKjsaratVsj7sTFcWFzZtp/csvD+wrERDAqeXLszzvUShdVLJkSYYPH87w4cM5d+4c8+bNY86cORn1ynfs2MGOHTt46aWXaNasWcZIem6pZDQyBqj+119w+TI89piYF3IPj4LXoop0KwbpVQzSqxikV3FIt47hEsF4vpCLNBVr+GHUw/vRDn0Jxc2YaZ+mWWusB/di3bPLabXKNTodxtKlSXjIJM5jc+ZQIiiIMo0aPbCvRFAQcWfPYklJQWtnIJvf+Pn58fbbbzN+/Hj++ecf5s6dy/z587l69SrJyclsv7fuuB38AjwDvDd1KjU++MCpfZZIJBKJROLauEQwni9F6B9STUWNu4VlzRI0LduhqVjlgf2KwQ1tm05YNqxEU78RSnEvp3TL5OtLUg7BeEpsLKeXLaPh//1flpM8vQICUM1mbp87R4nAwEz7HtXi/oqiUL9+ferXr8/nn3/Otm3b+Oabbzh37hzFihXLdTuxsbEcP34cFVgILJo4kb7Hj/Pee+9Rs2ZNYf1/VL0WBaRbMUivYpBexSC9ikO6dQyXCMbzY3ZvThM4VdWKZcUCFG9fNC3bZd/G4w1h7x9Ytm9E1+MZp/TL5OOTY874ycWLMXh54dexY5b7jd7euJUowa2TJx8IxgvDrGmdTkfHjh3pmM3rexgRERF81L07CyIjUYFFixaxePFi+vTpw3vvvUetWrWc22EKh9fCinQrBulVDNKrGKRXcUi3juESEzj1er342pc5TOC0/rkL9dIFtL2eRdFkv/iMotGg7dAN9dDfqFcuOqVbJh+fbEfGLSkpnJg3j+BBg9Bk40dRFLwCA4nNYhJnvngtYKpXr868KVMIB54FNIqCqqosXryY2rVr07dvX/7991+nXtMVvBYU0q0YpFcxSK9ikF7FId06hksE4/mzAmfWI+Pq5Sis2zeg7dILpWTphzajqRqEElQdy+Y1TpmRnNPCP+fWrSM1IQH/0NAc28iuvKHLrLTVqhXVvLyYC4R37MjAgQPRaGz/dZYsWULt2rXp06cPR48edcrlXMZrASDdikF6FYP0KgbpVRzSrWO4RDButVqxZrOku9PQPRiMq6kpmJfPQ6leC6VOg1w3pW3fFfXsKdSTEXnulsnXN8s0FVVViZg5k4A+fTAUL55jGyWCgoiNjHxge754fRQwGCAkBIDgP/5gzi+/EBERwaBBgzKC8qVLl1KnTh1CQ0M5dOhQni7nMl4LAOlWDNKrGKRXMUiv4pBuHcMlgnE3Nzfxkwn0BjCnZhrNtm5eAynJaEN652oFzHQU7zJonmiCZcsaVIslT90yeXuTlLYK571c/uMP4s6eJXjgwIe2USIggDsXLmBOSMi0PV+8PiqkLwB05w5s305QUBCzZ8/m2LFjDB48OCMoX758OfXq1aN169YsW7YMs9ls96Vcyms+I92KQXoVg/QqBulVHNKtY7hEMJ4vpOdIWWzBl/V4GNYDe9D2HIBicre7Oc1THeB2HNaDf+WpW0YfH8yJiSR++CbmVQtRo68AtkV+/Dp1wiMXdbO90hYCupXF6LjL0KkTpL/B3LMAUGBgILNmzeLYsWM899xzGUH5zp076d27N1WrVuXTTz8lJiamIHotkUgkEonkEcclgvH8yGFSdGnBeGoq6p04LKsXoWnaGk3lgJxPzK49j2JoWrTDunMTalKiQ22oSYkY9thqayfXqAs3YjD/8AXRkz/lyp49BD/3XK7aMXh64l627AOpKi6VG1a8OLRLq4SzejXc941FYGAgM2fOJDIykv/85z+UKFECgAsXLvD2229ToUIFhg4dysGDBx96KZfyms9It2KQXsUgvYpBehWHdOsYLhGMa7VatNrsq5g4hfSR8dQULCsXoniVRNPasXJ66WgaNQeDG9bft9l9rhp9BfO0b3BLigdFIaWiP7qho9EOG82JvQfx9S2N156tWI8dRVUfnt/lFRj4wCTOfPH6KJGeqnLtGuzZk+UhVapU4csvvyQqKoqffvopo/RhcnIyM2fOpEGDBjRr1oyFCxeSmk0pTJfzmo9It2KQXsUgvYpBehWHdOsYLhGMGwwGu5dBt5u02prW3TtQz5+xlTHU5q3epqLTo20XgnXvb6g3r+f6PGv4Ycy/fIPi7YthxOsYS5XKqKiSqDNyPuIE1f8zFnzKYlk6B/OUz7Ee+CvHRYtKBAQ8MDKeL14fJbp3h/Tc/5UrczzUw8ODESNGcOTIEXbs2EGvXr0yUlj+/PNP+vfvj5+fHxMnTuTq1auZznU5r/mIdCsG6VUM0qsYpFdxSLeO4RLBeEpKCikpKWIvorfdfNZ9f6Dt2APF29cpzSo1Hkd5rAKWbesfeqxqtWLZuhbL0jlomrdB+8wQFKMJk69vRjB+Yt48PP38KBfSDV33vuheexdN9TpYtqzB/PXHWH7fhpqY8EDbWdUazxevjxK+vtCsme3xihWQi9KTiqLQqlUrli1bxpkzZ3jrrbcoXdpW4vLy5cu8//77lC9fns6dO7N3715UVXU9r/mIdCsG6VUM0qsYpFdxSLeO4RLBuMViwZLHqiQPJS1NRQmuiVK/kdOaVRQFTcceqGGHsF44m+1xakI8lnk/Yz3wF9r+z6Nt2R5Fsf1604Px1Dt3iFyyhGpDh2ZUd1GKe6Jt2wXd6++hadYa6/7dmL/+CMumVaixNzPaLxEYSGJ0NMm3bmVsyxevjxo9e9r+PX0a7Fzsp1KlSnz66adcuHCBGTNmUL9+fcDmcePGjTRu3Jgnn3ySOXPmEB8f7+yeS3DRezYfkF7FIL2KQXoVh3TrGC4RjOdLqR2dHk3HHmi7P2NXGcPcoKngh1KzLtbNq7NcCEi9HIX558mod+6gGz4GTWD1TPtN3t4kXrtG5NKlaI1GKqfVzL4Xxc2ItslT6F59G22XXlhPn8A89SvUFNtEDM+qVVE0mkx54y5ZwqhHj7uP76mqYg8mk4khQ4awf/9+du/eTb9+/TKWEN6/fz+jRo2iZs2avP3225w/f94ZvZak4ZL3bD4gvYpBehWD9CoO6dYxXCIYzw8URUHbuCWKu4eQ9rXtQlAvX0QNy7ygjPXwfszTv0OpWBnd86+glPJ+4FyTry/xly9zfM4cgp99Fm0O+VyKVofm8SfQvfAamM2op04AoDMaKVap0gOpKi6Hvz/Urm17/JC88YehKApNmzZlwYIFnDt3jvfff5+yZcsCEBMTw6effkqVKlUIDQ1lx44dTlmRVSKRSCQSyaOFSwTjRaHUjlKiFJpGLbBsW49qTkW1WLCsX45l1SI0bbrYJowasv40avLxIebQIZJjYwno2zd319MbUPyDsJ4Iy9hWIiAgU63xouDVIdJTVf75B86dc0qT5cqV4//+7/84d+4cv/76K40bNwZsq5ktX76cNm3aULt2baZOncqdO3ecck1XxGXvWcFIr2KQXsUgvYpDunUMlwjGNRpNRiWLwoymRVtISca6YyOW2T9iDTuMdtBItE2eyjE1xuTjA4B/r164pdW/ztX1gmuhnghHTVva9v5JnEXFa3ac37yZ4/PmPbgjvcQh5Hl0/H4MBgN9+/Zly5YtHDhwgGHDhmE0GgEICwtj1KhRlC9fnjFjxnD27FmnXtsVKOr3bEEhvYpBehWD9CoO6dYxXMJYUclhUowmNK06Yv1zJ5jN6EaMQVPl4YsKFa9UCY3BQLVBg+y7XlB1SEhAjToL2CZx3jp5MiNdoqh4zQpzQgJ/f/gh/3zxBQn3lR6kbl3w87M9dnIwDne91q9fn19//ZWoqCgmTZqEX9o14+Li+Oabb6hVqxYRERFOv35RpijfswWJ9CoG6VUM0qs4pFvHcIlgPDU1NdsFVgobmgaN0fYehHboyyheJXN1jldAAKF//EGxihXtupbiURyloh/qsbCMdlJv3yYxLTgtSl7v58SCBSiKQnE/P8KnT8+8U1Hujo7/9hs4ean7+72WLl2acePGcerUKVatWkW7tJVA4+PjCQ0NlWkrdlCU79mCRHoVg/QqBulVHNKtY7hEMG42mzGbzQXdDaegaLRoatZF0entOk/v4djEUiW4Jtbj/6KqKsX9/NDo9RkVVYqS13tJjY8nYsYMarzwArVfeolTS5Zk1GnPID1v3GqFtWudev3svGq1Wrp3786WLVuYMGECABEREYwYMUJO7swlRfWeLWikVzFIr2KQXsUh3TqGSwTj8msTx9EE14IbMXD9GhqdDs+qVTNW4iyqXk/Mn49GpyOgb18qtm9PsQoViJgxI/NBzZpB2uI9jpY4zI7ceJ0wYQLt27cHYMGCBfz4449O7UNRpajeswWN9CoG6VUM0qs4pFvHcIlgXOI4ircvlPbBmpaqUiIwkFsnThRwr8Rx76i4zmhE0WioOXIkJxctIun69bsH6nTQrZvt8ebNkM+L9Gi1WubNm0f58uUBGDNmDPv27cvXPkgkEolEIsk7LhGMy1I7eUMTXBP1xN288fTyhkXR64l589C6uRHQp0/GtkqdOuFetiwRM2dmPjg9VSUpCTZtclofcuvVx8eHJUuWoNPpSE1NpU+fPly/9wOD5AGK4j37KCC9ikF6FYP0Kg7p1jFcIhhXFMXpq2K6EkpwLdQL51Dv3KZEUBBxp05htViKnNeU27czRsW193zNptFqqTVyJCcXLCDp5s27J7RvD+7utsdOrKpij9cmTZrw5ZdfAnD+/HkGDRqENa0UpeRBito9+6ggvYpBehWD9CoO6dYxXCIYNxqNGbWaJfajVPADdw/Uk+GUCAzEkpzMnaioAvMad+4cZ9etc3q7x+fORWsyEdC79wP7/Lp0wejtzfHZs+9uNJmgUyfb4zVrwEkzyO31+uqrr9InbSR/w4YNfPzxx07pR1FEvheIQXoVg/QqBulVHNKtY7hEMC5n9+YNRaNBCaqO9XgY7o89hs7dndiTJwvEq6qq7Bk/nj/HjePizp1OazclLo5js2dTa8SITKPi6Wh0OmqOHMnxefNIvnXr7o70Eoe3btnKHDoBe70qisK0adMICgoC4P3332fr1q1O6UtRQ74XiEF6FYP0KgbpVRzSrWO4RDAu617mHU1wTdRTJ8CcilfaJM6C8Hpu/XpuRkRQpUcP9rz9NglXrjil3eNz56J3d6dqr17ZHlOla1fcSpTg+Ny5dzd27Qpare2xk1JVHPHq6enJsmXLMJlMqKpK//79iYqKckp/ihLyvUAM0qsYpFcxSK/ikG4dwyWCcVlqJ+8oVYMAFfX0CUoEBBAbGZnvXs2JiRz63/+o9txzNPrgAzyrVGH3f/+LNY+fwtNHxWuOHInWYMj2OI1eT83hwzk+dy4pcXG2jSVLQqtWtscrV6LGxaLG3cq2jdzgqNdatWrx008/ARATE0Pfvn1JSUnJU1+KGvK9QAzSqxgKs1dLSgoRM2diTkoq6K48QGH2+qgj3TqGSwTjkryjGNxQqgZhPRZGiaCgjIV/8pNjs2ZhTU2l5vDhaPR6mn3xBbGRkRydMiVv7c6ejaF4caqmp5zkQJUePdB7eHB83ry7G9PPi4rCMvkzzD9PRr1ZMFVNBg0axMiRIwHYs2cPb775Zr73wfLXb6g3nLsqqUQiKVycXr6cf774ggubNxd0V4o8V/76i5OLFhV0NyR5wCWCcVlqxznYShyG4+nvz+1z50i4fTvfvCZcu0b4tGnUefXVjNVEPcqVo/HHHxM+bRpX9uxxqN3kW7dso+IjRuQ4Kp6O1mCwjY7Pnk1q+jL0PXpk7Fe27QCTB+a5P6MmOLZMfV7v16+//poGDRpkPF6yZInDbdmLGnMN66ZVWFbMR1Ufvaou8r1ADNKrGAqrV0tKCmHTpqEvXpwza9YUdHceoLB6zY4j333HwUmTMlf7KiCKmtv8wiWCcYlzUIJqQkI8XsWMqBYL8efO5du1j3z7LcUqVaJqem3vNCq0aUPQgAH8+eabDy5ZnwuOzZqFW8mSuRoVT6dqr15oTSZOLFhg21CxIjzxBACak2fQPf8KGAxY5v+Kmpr/aSJGo5ElS5ZQsmRJAIYNG8bx48cdasti55uq9fB+KFka9eplrPsd+4AkkUgKN2fXrCElLo4mn3zC1b/+IuHatYLuUpEl7tw5Yg4dQmMwEClHxwstLhGMm0wmTCZTQXej0KMUK45SoRJuVy/iVqoUSRcu5IvXG+HhnF65kgZvvokmfbLkPdT9z39wL1uWP996C6vFkut2k2/d4vjcudQaORKNXp/r87QGAzWef55jM2eSmrbyppoWzCtXrqGcv4BuwHDUO7exLJuLas19n8A592uVKlWYM2cOAHfu3CE0NJR4O1cJTbl9mzVdujy42FE2qKoV65EDaBo2Q9OmM9at61DjYu3tulDke4EYpFcxFEavVrOZsF9+IfjZZynfqhXuZctybv36gu5WJgqj1+w4s2oVXoGB1HrxRU4sWIClgOcJFSW3+YlLBOOy1I7zUIJrYj0eRonAQG6llTcUiaqqHJw0iQqtW1OmUaMsj9EaDDT78kuuHz1K+LRpuW47YuZMjKVLUyV9WXs78A8NRdHpOLlwoW1D8yZ3d65ciVLcE93A4ajnz2BdtxxVVXPdtrPu15CQEN5++20AwsLCGDVqlF39OPLddyTGxBA+bRrmhISHHq+eiYTbcWhq10fzZHMUb18sG1Y43H8RyPcCMUivYiiMXs9t2EBiTAzBgwahaDRU7tqVM6tXF3S3MlEYvWaFarVyZvVqqvbogX9oKOaEhAL/4FNU3OY3LhGMy1I7zkMTXBNiruFVoQK3Tp4U7jVq61ZiDh2i7tixOR5XvFIlGk2cyNEpU7h24MBD2026eXd7bEYAACAASURBVJMT8+bZPSqejs5opMbzz9uqBSQkYE2IQy3ja9u5whaAKt5l0PYfhvXIfqy/577utzPv1w8++IDWrVsDMGfOHCZNmkRceiWYHLgRHs7JBQto9sUXoCicXLz4oedYD+9HCaxm+wZFo0HbrQ/q8TCsx47m+XU4C/leIAbpVQyFzavVYiHs558JfOYZjKVKAVC5a1duHT/OrRMnCrh3dylsXrPj2t9/k3j1KpW7dsVQvDj+vXpxbPZsuwZdnE1RcZvfuEQwbjAYMORicp4kF3iXgVLeeLobuHPmjFCvlpQU/vnqK4KefRZPP7+HHu/XqRP+oaHs/u9/HzqR5diMGRi9vanswKh4OgF9+qAoCicXLcIadgi1Q3vbjr/+gqAgaNcOzfsfor1yG3XyV1inToHISHhIqS9n3q86nY4FCxbw2GOPATB+/HhKly5NixYt+PDDD9m7dy+W+1J7VKuVvydOpEKbNlTq0IHqzz1HxIwZOeaPq8lJqBFH0Tz+RMY2pWx5NE1aYlm/AjX50ShvJt8LxOAMr9bjYVh27yjQQOJRo7Ddrxe2bCH+4kWqDxmSsc3L359SNWs+UhM5C5vX7Di9ahVlmzXD5OMDQPDAgdw6cYJr+/YVWJ+Kitv8xiWCcY1Gg0bjEi9VOIqioAmuiacliYRLl7AkJgq71ol580i9fZtaL76Y63Pqv/kmbl5e/PXOO6jWrKt5JF2/zvH586n14otodDqH+6czmag+dCgR06ZhuX0bZeQ9/Tx5ErZtgxkz0Hz/I7qVm9CMGg2BgWAyQdmy0KgR9OkD//kPzJsHaTndzr5fy5Qpw9KlS/H09ARsXyP+8ccfTJgwgcaNG+Pt7U3v3r35+eefOXPmDKeWLiU2MpL6b70FQGD//lhTUji1bFm211DDj4BWa5vkew+apzqAVot126ORMyrfC8SQV6+q1YJlwwqsW9diWbkA1SK/5obCdb+qVithP/2Ef+/eGcFhOpW7dePsunXZvifnN4XJa3akxsdzYfNmqt5TzatYxYpUbNuWY7NnF1i/ioLbgsAljCUlJZH0CC48UFhRgmvimXQbgJiICCHXSLpxg3+nTqX2yy9jSAsic4POaKTZV19xdd++bN+QImbMwL1MGSqHhOS5nwF9+6KmJHP6dgpKs+awZo0tuO7dG5580hZ0Z8XVq7BvHyxdCv/7HwwcCGXKwJAhpG7ZQlIucrTtoWnTply6dIn169czZswYatSokbHv1q1bLFu2jJEjR1K1alVaPPccKzw92bpvH3Fxceg9PAgeNIjwX3/NdnKQ9fB+NLXqodz34UYxuKENCcX6959Yo7KvvpN6506+jIjK9wIx5NWrGnEU4u+gHTgC9dRxWyWiR+TblIKkMN2vF3fuJO70aaoPHfrAPr/OnUmKieFqAY7Y3kth8podF7ZuRdHpKJ+WhphO8ODBXNy1i7h8rHZ2L0XBbUHgEsG4xLkoFSujL+6JqXQpbp8+LeQaR77/HpOvLwF9+9p9rlfVqjR8910OTZ5MzJEjmfYlxsRwYsECao0aladR8XR0Oi1BgX4c//sfWxpH167w5ZewZAns3QuXL0NiIpw8ibplC5bRI7C0fwp1QH9o2xYCAiC9H/HxMGsWHl27Uuzxx2HCBFtai5Pw8PCgc+fOTJ48mbCwMKKiopgxYwb9+/fH556RrCtJSSzcsYOnn36akiVL0qBBA3YkJnLtxo0sJ2KpN6+jnjuFck+Kyr1oAqqh1K6HZc1i1Cyq3dy5cIEVrVqxvmdPIpcswSzw2xbJo4eqqlj37ELz+BNo/IPRPf8q6q0bmGf+gHr74fMbJAWPqqr8O3UqVXv2xCMtJe5eTN7elG3alLNr1xZA74omZ1atolLHjuiMxkzbferXp1SNGhxPq6YlKRy4RDBuNBox3nfDShxH0WhRgmrgVbIECQI+fd86eZJTS5ZQf9w4hwPmqk8/TeUuXdg9duzdpeuBiOnT8ShXDr/OnZ3SVzX8CP41ArBYrJxavjzrg4xGCAhAadcOzdc/oA4egLlxbdRVK23pLAkJsGoV9OoFaZNJNefPw4cf2tJamjeHX36BWOeWCSxfvjxDhgxh/vz5XLlyhW3z5tHP15fmTzyRkfNntVo5ePAgr40dy4uHD9NjxAi+/fprLl68mNGO9cgBKO2DUr5SttfSduwOt+Ow/rkz03ZVVdk3cSIlgoMp17w5//zvf6xs25ZDkycTf/myU18vyPcCUeTFq3rhLOrFC2iaPAWAUrI0umGvoGi1mH/9FjXmqjO7WqgoLPfr5T/+4OaxY9R44YVsj6nStSvnN29+JD5sFxav2RF/6RJX9+3LlKKSjqIoVBs8mNMrV5Li5L8ZuaGwuy0oXCIYt1qtWB+RXLWigia4Jl4GDbFOniGvqioHP/+csk2bUq5Fizy19cS776I1GNg7YQKqqpIYHc3JRYtsueJZ1Ct3BOuRA7jVbUi1wYMJnzbtoTVeFa0WbZ/B4GbEsuBX1JRkWwDevTssWwaXL5P6v/9hrV//7km7d8OIEbaUlwEDYNMmsKOeeu5eiJWkJUt4Zdgwfv/7b27evMmGDRt49dVXqVChAmD73YTfvMlrr79OhQoVaNasGZMnT+bczq1oHm+IoijZv26P4mg7dMP622bUGzEZ28+tW8e1fftoNHEi9caOpee2bTz+2mtEbd/O6o4d+eONN7h24EBGCktycjJRUVEcOnSInTt3cuzYMbtWe5PvBWLIi1frnl0owTVQSt/9dkbxKIb2uVEoZR7DPP17rBfOOKurhYrCcL+mj4pX6daNYmnvFVlRoW1bUFUu7tyZf53LhsLgNSfOrFlDsYoV8a5XL8v9lTp2xODpSeTSpfncs8LvtqBQHtWZ64qi1Pz333//rVmz5sMPfgjppdw87cg9luSMmpJM5KihHA4/Q28Hl6LPiou//cZvo0fTZflyvAIC8tzezePH2dSvH/XHjeP2uXNc3r2bLitXOiUYV2NvYv76Y7TDXsZcwptV7dtT7403cpVao96Jw/zrdyg+ZdD2G4qiudufjPs1KgpmzYK5c+HSpcwNlCtnm/zZrh20bAl5vLePzZrF0R9/pOvatZi8vTPts1qt7Nu3j6VLlzJ/+nQuZ1GpptETT9D7mWfo3bs3lStXzvo1qyqW2VOxqCrJvQZy49Il1g4ciG+7dpTr2pXr168TExNDdHQ00dHRXAgP53x4ONHR0SRoNNy2WIjPIhdRURQqVqxIQEAA/v7+BAQEZDz29/enWLFiGcfK9wIxOOpVvRGD+bvP0D43Ck1l/wf3Wy1Y1i1HPbIfbehANNVqO6W/hYXCcL9e+esvdgwfTsiaNXhm838/nT3jx5McG0urH37In85lQ2Hwmh2qqrI2JITK3bpRe9SobI8L++UXTi5cSPeNGx0q3+sohdltfhIWFkatWrVqqaoaBi4SjCemfS0mV4VyLle//IhtMxbQ67ffMJYunef2rKmprO/VizJPPknD995zQg9tnFy8mAOffAKKQpNPP8WvUyentGv5fRvWf/aie2U8iqJwZMoUzqxcSbf163P15qdej8b867co1Wqj7dYnY2T5gfvVYoGtW22B+YoVD5ZG1Gptk0XbtrUF540bg5tbrl9HwtWrrO3albqvv07QgAE5H3vtGl+3aMH5OnXYcvAgp7OYM9CgQQPKlStHfHw8d+7cyfRvfHx8xuvLL8qWLZsRpPv5+REcHEzdunUJDAxEn49/pIoyjr7HWtYvR406h3b4mGy/WVFVFetvW7Hu2oSmcy+0DZvmub+FhcLwt2vrkCGYfHxsaxI8hCt79rBj5Eh67tyZUYe8ICgMXrMj+tAhtjz7LN03b6ZY+fLZHpd86xYr27Wj0cSJVO7SJd/6V5jd5if3B+N5n8FWCNA5YaKe5EG8mjRDmbmQW5EnKeuEYPzk4sUkRkdTe/RoJ/TuLgF9+nDt77+JO3OGSh06OKVNVVVtS7/XaZARRFQbOJBjs2Zx5LvvqNqzJ8UrV845daO0D9oBL2CZ9SNWrRZNx+4oOv2D96tWCx072n5iY22TQ3/4AY4csQXqFgvs2WP7+egjW+nEFi1swXnbtlC3rq2NbDj4xRcUr1yZgGeeeejrdvf1pcOQIVzdu5cpYf9yYNxoViSoLN31O5Fpk00PHDjAgVwsvJQTRqMRHx8ffHx88Pb2xsfHh1JeXmivXSM5PBxjfDx+TZpQvFUrzl68SGRkJKdOnSIyMpIbN25kauvKlStcuXKF3bt3Z9qu0+kICgqiRo0amX6CgoJws+PDTGFCjbmG9cxJtA2bObXd++/ZlNhYkmNjKV4p+3kEamIC1kN/Z/ogmhWKoqB9qj2KpyeWNUsh7haaNp1zPKeo8Kj/7bp24ADX9u+ny8qVuTre98knMZYuzbkNGwh+9lnBvcueR91rTpxZtQrfhg1zDMQB3EqUoGr37hyfPRu/zvn3/6Uwuy1IXMJa+ic1OQrmXFKqBFGsmDu39v9N2UaN89ZWbCxHp0yh1osvYixZ0kk9tKEoCk0//xxrSgqKs+qfXo6CmKto6jTI2GTw8qL+2LEc+f57wn/9FbeSJfGuWxefunXxrleP0rVqob0vyNNU8IMBL2BZPg/ruVPoeg4g0cP29V6W96uXF7zwgu3n9m34/XfbqPm2bbbgHGzVWzZvtv0AlCwJrVvbRs27doWKFe++jD//5PzGjXSYPz/XqTs1hg4lctEiLs6fQ73HytDwP+/zid7A0aNHWbJkCcuWLcNiseDj44OHhwfFihXL9K/JYCBq3lxKl/GlzuhX8fDwwMPDg9KlS2cE3x4eHtleX7VaufT77+z/+GMeu3qVYRMnZtp/8+bNjMA8/d/0x5fvmRRqNpsJDw8nPDw80/larZaAgICM4LxmzZp06NCB0k74wFmQqFYrluXzUC9HoZQsjSagmtPavvc91pKSwvbhw4m/fJmua9bgVqJEludYD+wBkwmlxuO5uoamXiMo5ollyWzU27Fou/VFcSDdLOyXXzCWKoV/aKjd5+Y3j/rfrn+nTqViu3aUyGVKoUarpXJICGfXrCnQYPxR95odluRkzm3cSINx43J1fPDgwawNCSHm0CF8sskvdzaF1W1B4xJpKjKHSQxxcXH8M6g/br5laPzL9Dy1deCzz7i4axchq1ahLQSrd1k2rkS9dAHdsFce2KeqKvFRUUT/8w/Rhw4R888/3Dp5Eo1WS8maNfGpWxefevXwrls3Y3EMNSEey7plqMf+JbXJU5ifaIZnNkFMtly7Bjt22ALzbdsgq7KTigKhofDaa1gaNmR9z56UadSIJ99/365L7fvgA67/toN2r4xE/3R/u849+sMPnJg7l06tnsD0/Mto/IPtOh9AVa1c2bKZHW+Mpf3cufjUrZur8y5evMiJEyc4f/484eHhhIWFER4ezpkzOU8Q1Ov1hISEMGjQIEJCQgrlyLll7+9Yt29ACaqOGnUe3Uv/RdE75//ave+xByZN4uzatRg8PfGpV4/GH330wPGqxYz5m4/RNG6JtmnrB/bnhPXieSzzf0V5rDzaPoNR3HJfueH0ihXsnTABgJZTplC+ZctcnafeiAF3DxRj/n71/ij/7Yo5coTN/fvTeelSSlavnuvzbh47xobQULquXYtnlSoCe5g9j7LXnDi/aRN73nmHXrt2oc9hwOJedr70ElqDgRZffy24dzYKq9v8xiVzxtOrLRTGP6CPMsnJyYS/+xbXDv5Dx207HW4n7uxZ1vXoQfP//Y+Kbds6r4OCUC0WzJMnomnVEe0TuctfTYmLI+bIEWLSAvTrhw9jTkzMmBHv36sXvk88gXr0IOb1y8HbF32vZ1FKeT+88ew4c+ZuYL59uy1Yv4d/69ThuF5P1/XrcfP1tavp28ePsTa0Ny3ef48KfR6e3pJO7OnTbOjVi0YffUQlJRnrqePoRo3NVVComs2oZyNRj/2L9XgY3Ilj34WbxF6/SafFi3OVp5/de0F8fDzHjx/PFKCHh4dnpN7cS8mSJenbty+DBg2iadOmhSJdQo2LxTxlEpo2ndE8/gTmKZ+jqdsQbVvn5JKme43evZvfX3uNVlOnonVzY+uQIbSdPp0yTz6Z6Xjr4f1Y1i9H9/p7DgW46o0YzPN+ATcjuv7PoxR/+B/+mCNH2Prcc9T7z39sq/DOm0eH+fNzHNVVb8di2b4R9fDfKH5V0Q5+EUXJvyJkj/Lfrl0vv4yqqg5NxlzfsycV2rShzisPDmbkB4+y15zY+dJLGLy8aPrpp7k+J32CbbcNG3KsduMsCqvb/Ob+YNwlShtKxOHVoCGxMTewXo92uI1/vvwS3/r1qdCmjRN7Jg711HFISkRTM3ejsQAGT0/KNW9OnVdeoe2vv9L7r7/otHQpwYMGYY6PZ/uwYYRPm4ZSuz7q86+CTo956ldY9u9xfGXKKlVs6SwLFsCVK7YVP597DgwG7uj1hKWkUC8yErc6deCDD2yrguYS9+hL+AVW+f/2zjs+ijr//8/37qaRSiCEUEKv0qQXlSJNwVMUCyf201NPT/Tq7+6+Zz+76N1ZEPUQ9cCGcmJFkGboVcCEQAIJLb2Xze7O5/fHbMImJCEkOxDk83w8JjM7n9mZz7z2k5n3fOb9eb/Z8+lnDa6fUorNjz1Gm2HD6DxtGrZLLweXC2P18rq/U16GsXs77o/fxf3c3/EsehuVn4PtkknYJlzOgLgISo8da3L659DQUAYPHszs2bN56qmnWLp0KcnJyZSWlrJixQpuu+02wsPDAdMNZt68eVx00UV069aNv//97yQnJzfp+Fbj+fozpHUbbMPGIMEh2KdehZHwPSrTf7HcS48dY8Nf/0rfX/2KuDFjaDN0KN2uvppNjz5qJsTyopTCs341tgtHNLqnWaJbm7HIbTbc/3oKzzdLUYX5dW5flpXF2jlz6HTZZfS88UYG3H8/bUeNYvVvfkN5LdGBlLMcz/df4/7X06hjh7FfeQPq6GGM9asbVd+fG3k//cSRVavod/fdjfp+5+nTSV227Ixk3f25UJadzbF162qNLV4fsSNGENm9O0nvv29RzTT+4Lwwxp1O52nFItY0DKfTSfAF/XG7PRSvX9eofRxZs4Yjq1Yx+I9/PCd6GMGMLS49+iIhLRq9D5vDQXSfPvS68UYu+ec/GfXMM+yeN4+1c+ZQ7DYon3Ejtksvx/jmMzMeeXETMxGKwLBhsGAB6tAhtlxyCdEeD10KCkwj/JFHID4ebr0Vtm+vd1dKKYydW+g763pyf9xNxsaNDapCyqefkrNrF8P+7/8QEdMovHwGRsIqVMaJ0I2qsADP5gTc783D/dzDeD7/CAD7Fdfi+MOjOGb/Gvuw0dgumkBIn370HzaQH199leLDh09Zh9O9FoSEhDBhwgTefvttjh8/zqJFi5g2bRp2r69yamoqjz/+OD179mTkyJG88sorZGdnn2KvZxZj315U4o/Yp8+sGjMhfQcg3XrhWfYxSjU9JnBZSQkb/vxnIrt1o/9vflO1/sLf/Q5XURF73nijap1KTYbMY9hGXNSkY0poGPbb7sM+fSZGyj7cL/8D99LFJyUJ8lRUsO7BBwmJiWH4ww+bbc9mY9Q//kFAaCjrHnywKj+AMjwYW9fj/tfTGNs3Yb/8ahx3PYht4FDsl12FsfKram3VaprrvWv3vHm0HT2a1gMGNOr7nadNo+TIEbJPca2xiuaqa30cXLaMkJiYk94ynYrKJEAHPvkEV3GxRbU7wbmobXPgvDDGAwIC9GACCwgICCCqSxdsDgf5G08/1nhRWhrr//Qn+tx662n5HJ5NVHkZKmk3toFDTr3xadD58suZsmgRBcnJ/HDHHZQdOYJ9xMU47noIiotwv/ocxt5dfjnWkb17OZaZydBPPkEWLoTKBEMVFWb4xMGDzdjln3wCbvdJ31dH0yE7g8hJlxF/2WXsfv31Ux6zPCeH7c89R7977qkWYUN690d69sHzvw/xrF2B+82Xcc99DGPNciSqFfZZt+P4w2M4Zt6Erd+F1XpSRQT7L66nS8dYWraLY/MTT5yypy3A7SbA6YTi4tOeWhgGN0yfzrLFizmanMzLzz7LUJ/kTBs3buS+++4jLi6OK6ZN4/HHH+fDDz9k586dZzycYyWqwonnyyXYhl+MxJ14RS0i2C+/GnXsCGr7piYfZ//bb1Oclsbo556rljU3MDKSwX/+M3vffJMCr8uPsWEN0mcA0rLpA2LFbsc2YAiOu3+P/fpbIScL9yvP4f5gAcZhMzvwtqefpigtjUtefrnaAOqA0FDG/vvfFKaksOXJJ/Ek7cH92gt4vv0c24iLcNz/Z2yDhp14gBk4DOnZF/eS91FuV5Pr3hCa470rf/9+0pcvb3SvOECLtm2JHT6c1M8/92PNGk5z1PVUpC5dSpdf/KJRQQg6TZuGIzi47izRfuRc1LY5cF74jLtc5oVTNxD/Uqnrd9dcTYdQB/3eeg9p0bBBJa6SEr795S8JadOGca+91ui092caY/tGPMuX4fjdw4jd/3WuKCoi4c9/JnPTJkY+8QTxU6agPB6MNcsx1q5ABgzGPvWqRr/ed5eW8sWVV9Jx0iQGV47IV8rM8vnyy7BkCfhmT+vUCe65x3RvadsW8MaGzjiG47bfULB/P19ceSUTFy6kzZC6H1AS/vQn8vftq9W3WxXm437teQiLwNa7H9K7H9KuQ4N9c42k3eS9MpflqzYz+vbbiY+LM5Mk1TZZ0DP0E/Ced0qrYxsRIT4+nl69etG7d+9q83bt2ln2Vsiz/HOM3dtx3PvHWgc6etatxPhhJY77/oSEhjfqGEfXrWPV3Xcz5qWX6DRx4knlSilW3X037pISLn3hOTyvP4/9jt+akYQswEhLwVi3EpX8EykFTrZ9t4YJb79N7NChtW6fteJbVjz4OwYM6EWvWbOwjZ2MhNWuhSotxv3a89j6XYh9yum5CzSG5njv+uEPf6AsK4uJCxY0aT8pn37KtmefZcbq1Wd80H5z1LU+8n76ia9mzmzSoNcfX32VlKVLzTwYfspAXRvnmrZni/NyAGdRURFAlc+nxj9U6vrjk09g7NnJ6CeewDaw9hueL8owWDtnDvnJyUxdvJjAyEirq+o33O+8irSOxT7NurBohQUFJL/zDvvmz6f3LbcwcM4cbA4HxuFDeD79L3g82K+6AVvn089QuuOll0hdupTpy5bVPhr/0CF45RWYPx/yfXxw7XaYPh112224927CPuUX2AaPAGDtnDm4SkqYMH9+rcc8lpDA93fdxeT336f1wNrD2CmPp/YwdUqZ9UhPh8OHT8wrp0ojOz+fnTExHIiKYvqBAwSehXTMBrAWeBHYAGTWv3kVYWFh9OrVi549e9KtWze6dOlC165d6dKlCx06dKhyiTldVMZR3PPmYr/2Zmx9as9cqTwe3G+8iLRtj2NG/QmfaqMsK4svr76a9lOmcMEDD9R5jS0+fJgvrrqKQVMvpVvXDrVGIfI3mSuWs/LBhxg0qA/dxozEftEEs0fe27OoCvLwrPwKtWsbh5yw+fNvGPvqq7S7+OJ692vsT8Tz/pvYb7oLW9eelp5Dc7t3FR48yBdXXMH4+fNpO7Jp4WxdxcUsueQSRj/7LB1reYizkuam66nY+vTTZO/axZT//rfR+yjPyeGziRMZ8+yzdJw0yY+1q44/tHWVlGAPCjpnOukaw3mZ9Ke5PnCc61TqGtWzFwe3bMZI2tMgY/zH117j+Pr1TF606JwyxFV+LurgAWwTLM5mJkKPW28lbvBgEv74R3L37GHM888T3KET8uuHMJYvw/PO66hxk7FdMqnBvaoFBw6Q+J//MOqZZ+oOi9WpEzz7LDz8MLz7rrmcmmomFlq6FFm6FEd4GOS7IbI1dOtGv1//mq9mziR7166TfEjdZWVsfuwxelx/fZ2GOJmZyPbtcPDgyQZ3ejqUljbo/C7IzuZQZCQ727Rh2PHjZhbSdu2qTeWRkRAQQHBww8PhNRQbMNY7kZ1N8TffsG/HDpKAJCDRO08CfJ1WiouL60yU5HA46NSpUzUD3XceHR1d6++vlIFn2cdIj95I73511lnsduzTZ+J5+xWMQcOwdenR4PM1PB4S/vhHwtq3p/c999R7nQ3r0IH+d/6KXa+9RodX/2n5jac0M5N1jz1Bl6tm0HPOb1Eb1uD5bBGs+BLb6PGQn2u6y8TGYb/lHrp37kZx63b88PvfM3nRIiK7dq1z37buvVHDRuP5bDFyz++bNHbkVDS3e9fe+fNpNWAAsSNGNHlfAWFhdJgwgYOff37GjfHmpmt9GC4XB7/4ggFNTIYX3KoVXa64gsSFCy01xpuqbf7+/ay84w7CO3Vi3KuvEhAW5qeaNW/Oi55xHWrHGip1zdm0iTX338+MmVMI/PMTiKPu11Ppy5ez7qGHuPjll8+Z6CmVeNZ+h7FjM477/mzpYFPf9lp8+DBr58zBmZfHRXPnVhm7RuJuPJ+8i23wSGxTr6zTpcNTUUH2jh1VyX3COnZk/BtvNLz+SsHGjfDmm7B4MZSUVC8fPx7uuIPVa9eC3c7YV16pVrxj7lxSly5l2uefExgebmYQ3bIFNm8+MaWnn55AIhAXB1FR0KYN9OsHcXEYIUEcXf0da/anMfm112l90UXmtj6c8WtBVpaZlKkyCdPRoxjAYU4Y54nA5tBQDgFZpaWndTMLCgqiVatWREZGVpsiKsqJyD5O1NhLiYqNq1YWFRVFmzZtiImJqXqV7P78I9ShAzju/j3SwN6oH199lcSFC7ns448J8MbLr09X14ovWf7oP4gYOpyLXnyxwed4ungqKvju1lsBmLhgQZULhCopwti4DmPzDxAUjP3Sy5F+g6r+d5RhsOb++ylISWHKokV1JisCUK4K3G/MRWLbYb9m9sn/T0pBdraZgKsJWNZeRUy3s9NwJSg+fJjPL7+cuG+/qgAAIABJREFUsa+8csq3Bw3lyOrVrH3gAa5evfqMdsycSzbB4e+/Z91DD3H1qlVN1ih//36+vPJKpnzwAa361f2Q3hSaom3u3r18f+edtL7wQgpTUwkIC2P8vHn1/i+eq5yXbirn0j/euUSlru7cXJZOnMiUaWNpeef92HrUPhgzf98+vv3lL+lzxx30v+eeM1nVJqOUwv3qs6av6NjJlh6rZnt1l5ez5fHHOfjFFwz9y1/odq2ZPtw4eADPoreQ3v2w/+J6xG5HKUXBgQMcT0jgWEICmVu2YDidRPfvT9yoUfSYNYuQ1o2LXa6OH8Vz9+3Y0zKQ7TuqlWXHxPBtTAyXPfkkLa+6CoC8nTv5evZsxkyYQHxOjml479tX/0EqDe0OHcxMobXN4+LqNCI8q5eT8OzzFDmCmfrxxyf5p5/Va4FSsGfPCcN89WooL6+2SQWm33mKzUZqTAypkZGk2O2kOp2k5OSQW1Dg1yq1bNnSNMxbtaJ1YS6xXbsRO2gIbdq0MddHRtKmpIQOIkT4vE3JSE5m5auvMubWW4kfOJAKbzSSQF/f3+hoGDQI7HaUy4X7pSfI69idFU8+e1oJd04HpRSbHn6YI6tXM/Wjj2hRS/x85XaDTRDbye4/rpISls+eTVDLloyfN6/e2PXqaDrul5/C3ncINmWH/fshOfnE3Pu6vtlit0OPHtC3L/Tpc2Lq3RtanNzbv+mRR8jdu5cpH3zgt84Iw+Xi0wkTGHj//XS/7jq/7LMhnEs2wdo5cxCbzW8PsCvvvJOgli3p/7e/cfToUSoqKnC5XFVz3+Xa1gUEBBAdHU2rVq2qzSvfNjZW26zt21l19920Hz+ekU88gTM/n5W/+hUiwvj58xt932qunJfGuM4IZQ2VuoaHh/PxyJEMnTSW+JHDsU+fedK2zvx8vr7+eqL79OGiF1/0X1r6M4RxNB3P/Jdw/PYvfokCUR+1tVelFPs//JCt//gHnX/xC4b97W/Yg4JQxw5TPP+fZJYbZBoOjm/cSFlmJmEdOxI3ejRtR48mdvhwAv3Q9j0b1mAkrMIx529IYiK89RYsXGj2AAIr4+MJMAwubt0aw+NheXExwW43l6SnU+utu2NHM9zisGEwdCj07Fmvod0QlOGh+N/P8fWCD7ngvvvoe/vt1cqb1bWgrAzWrTMN808/hQMHTvmVQiC1XTtS4uJIDQ/nsMNBgQgFhkFBRQUFZWUUHD1MQXExBRVuymsY+02hLzAaGCaCMz6eEeXlDD9VbPqoKJgwAaNnNzyeEhxPvMi2F17k8MqVXP7ZZw3OIthQkhcvZutTT3HpO+80OCtrTYqPHOGbG24gftIkht13n5ksKyPDdNc61wzuxtK5c5VxXtyuHXv27SNl/XoufuYZOowb57fDlJeX8+Ujj7B/925aX3staenpHM/MxDhNY19ECAoKIigoiODg4Kp5Xcsej4fQ0FAGDRrUPK4FdeDMz+fTceO4+OWXaT92bJP25XK52LBhA0veeotlH33EgbIyv7rrhISE0KpVK6Kioqoe8iuN9R49ejB58mQ61JF06PiGDay57z46X3GFGfrWax848/P5/q67cJeWMuHNN2nhDSLgD1RpiTfS1Bhs8XW7pVnFeWmMl3p9TlvU8rSvaTy+un57443Edu5I3wgH9hm/xNazb9V2htvN97/+Nc7cXCa9957fb8BnAs9Xn6GOpZ+RgWf1tdfsnTtZ++CDhLRuTezw4RxLSCA/KYmAwABiO8cTd+0NxF1yiSWZ1lzzXsDWrRf2idNPrKyogP/9D958k8y1a/muUyemHThARmgoO2JjmbZ/P6FuN8TEnDC8K43v2Fi/1xHM7Iz7fv9bdmz/ielffEFou3ZVZc36WlBYCImJ8NNPsHfviXlKitmr3gCUCIQEQ8d4pGNHKmJiKGzZkoLwcArCwigIDia3uJispCQyU1LIPHqUzOxsMsvLyQSygBygIUeLBkZhGuijgWHAKf+zu3TBNW4cS/bsIWjkSCKuvJL09HTS09NJS0sjICCAIUOGMHToUPr27YujLpeZSjeQw4fNCDk2G5kHDrDi+ecZdvPNdJ8wAWy22icR8zsZGScM7czMasuZ2dmsDAtj8PHj9KwlKVCtVbLZkE6doHt3s8e5WzfzYaQJWNaD63KZ7aqynaWmVo+iBBQFBLCndWtSo6JoWV5O/6ws2jcwGpHCfMuTj+mSlQYc8s7TfD43dJCzlcTFxdUa5Sg+Pr7Rg6f9xb5Fi9j92mtctXLlaQ9mVEqRmJjI8uXLWb58OatWraL4DMQZr4++ffsyefJkpkyZwtixYwkJCTFdlebMoccNN9Sab6SiqIjV995LaUYGl771FmEdOza5Hsa+vXg+/xBahOKYcSPStt2pv+RnzktjXIfasQZfXTc9+ijO3FxGXzYeY9M6pM8AMwRfeARbn36a1P/9j6kffOCXf6QzjfJ4cL/4GLbxU7EPHWX58U7VXsuys9n0yCO4iopoO2oUbUePpmWHdhj/fRMJDMJ+451IqH8HvaiMo7hff8EMkRdThxGdlsZ3t9yCPSODLBEG9O1L7+uvN43v+PiT/LetxLN1A9899HuCuvdk7JtvVV3gz8lrQVmZ6d5T00hPSjIH1voZlwg5HTuS1aMHmbGxZERFkVRaSsLu3STs2kWp1yWlJnabjYHduzO6f39G9+tHR6eTw2vWkLZ9O+llZaRD1dSQfL3BQUFc2KULQ9u2ZWh4OEOBXoWF2I8cMY1wn17/UoeDr7t2pUNhIcOPH/eDCpASGcnGdu0Yl5ZGXOVYCZvNHORcaXB3746Ki8W9aTW2K67GPn5Kg/btdrspLS2lrKzspLlSChHBZrPh8Xiw2WwEBgZWrauc+y7XRV33d6VUNfeDiuJiXAcPUpGSQn5iIqlJSRwvKSGoooI2RUWEOJ24gBKgGCjyzn2nmutOzlDQMCKBkx6VRSAw0HxrFhhoTt7/6YqKCpRSKKUoLy+nvLzcLz2+wcHB9OjRo8o479WrF507d677AbEObDYboaGh1aaQkJAGufp8c8MNxAwefCIM7SnIzMzku+++qzLAjxw5Uut2vePj6eFyce2TT9IiIoLAwMCq+OCVy7WtCwgIwOl0kpubS25uLjk5OSctZ2dnk5ubS15eXtVnTy3XqaCgIIZfcAHxx44x45ZbmPHkk3W2ZXdpKWseeICCAweY8OabJw2wLi8vJyMj45TJhlSFE8+6lai9O7FdOBzbiIsRRwA9evQ440kHz0tjvPJpMOw8GZV7pvDVNen999n33/9yxRdfmC4dn38EeTkcCm7J5rcXMv6NN5ocCqsSpQwwDEvifNeGsW8vng8X4PjdI5ZGTqikse1VlRThfm8+uF04Zt+FRLb0W5083/wPlZaC48459W53fMMGVt5xB9EXXMDkRYssjWdbH0opcl95nm/nvcOY554jfuplwM/sWuBymYNfMzLg+HGM1StQe3/EFt8dyc6B48dPTLXdpEJCTP/gGj7DnqOpGNs2mrHHfdp71vbtfHvzzbS86y4OORwkJCSQkJBASkqKX04nEigH6rudhgKDgaGYPfFDgc7Aii5dsCvFhIMHaVSLCwoy39S0aWNO3uXNKSls3LuXdtOmUR4TQ3l4OGVuN2VlZZSXl1fNS9MPUZqcSFm7TpRUVFBSVERmVhYem42AoKCTDO7Kh8LzjWCHg04tWxIfFUV8ZKQ5j4rCk5eHkZPDTddcQ3B5OWzaZA4cL6wj87CI2W5Hj4ZRo2DkSPNNRGCgOb7HbbpoOZ3OKgO9ctnpdJKXl0deXh5paWkkJSWRmJhIUlISBX4ek1EXIkKLFi1OMtIrp+DgYNwlJRxdt464MWPMwe/1oJRi37597Ny5s9bydu3aMWnSJCZNmsTEiRNpFRnJ0ksvpfett9LX65ftL2peYz0eD1u3buWbb77hm2++YcOGDbUa5+3bt6/qNZ84cSLR0dHk5uZy/Phxjh07xtH0dNbPn8+RQ4cIvPBCcktLOXbsGMePHyffNwxvIzAMQxvjdaF9xps/vrpmbNrEyjvu4NrNm3EEB6MMD1mL3mXl0y8w4JKR9Hn4MaRNXJOOp0qLMbZvxti6HoqLkH6DsA0dja2dtb3t7o8XgqFwXHeLpceppCntVZWX4Vn8NiovF8fNdyOtYppcH2V4cL/4OLZLJmIfXn8Kc6UUu197jfjJk4nsfvpx0P2JKi1m+x23cPDgEaZ/u5zA8PCf7bVAZWXgfv0F7DNmYet3YY1CZUaxqTTMKypM//z4eLOnt+a+3G7crz+PxHfF8QtzUJ0zP5+vrrmGuDFjGPHYY9W2z8jIYMWKFWzcuJGtW7eyZcuWk3qowkJDie/UiY4dO9KxXTs6GgaxR49yJCmJUYWFjM3PpwXgwYwwsxnY4p12YBrpdRFgsxEbFES/Pn3o3r49Xdq0oWtsLF1at6ZLTAwRwcGmC0bl5PFAaOgJ4zs2lkKlSNq3r5phlpiYSHJysk7tXYPQ0FDCw8MJCwurNtVcFx4eTnh4OO3btyc+Pp74+Hhat25dq9FTnJ7O/6ZOZcrixbTq742J7/GYb4HWr4eEBHOelFR3xWw2cxxK166mYd61a/Xlli2retPrGpeTkZFxUhtISkoiNTX1nAqHGBoayrhx46oM8D59+pyke9L777Pt6aeJ6NqVXjfeSOcrrsAR0rhkcr6c6hqbn5/Pe489xmcLFpDocHAkq/b3ZHa7vVaj3Qq0MV4POppK88dX1/K8PJZcdBFTP/6Y6D59KM3M5JvrrqPtsKEMHdQLUvZhGz3ejIt9Gi4CSilU+kGMLQmovTshNAzb4FFIdCuMbRtRB/cj7TpiGzoK6XchEuDfTG6qvAz3849gn3kTtnriNfuTprZX5XLh+egd1JF0HLPvrJYGvTEY+/bi+WCBmXW0gRlWmwsVe3by1W130G78eIY989zP8lqglMLzzmvgsGO/8S6/3FSM1P14Fr6O/bZ7kY5dWHP//RSnpzNl8eJab9i+ujqdTrZv305eXh5tk3bSMTCAVvc8VGu9khcvZvvzzzPtzjsJ3boV8vLMiDmV0XM6dsTVti17c3PZsmMHW7ZsYcuWLezcubPBvcutWrWiS5cu1WK0l5eXVzO4jh071iid7DYbQQEBBAcFERwcRIjbTYvwcFrEtiWkRQvsLheezEzcWVmERkQQ27cvbfv1IywighYtWhASEkKLFi2qlkNCQrDZbCilMAwDp9OJUgq73Y5hGFXrK+eVU32/eV1lFdnZHF2+nNytW2nduze9Zs4kdtAggoKCqlwTarostGjRol63mKawfPZsWvbty9C//KXujXJzYcMG0zBfv97sPW+oH3RkZJVh7u7UCdW5MwEdOlR7E0JYWDV3OsPl4vjGjSQvW8aWr74iMz+fsA4daDN0KG2GDiUk5tSdHZUuSSUlJVVTcXFxtc++U1ZWFhUVFbizsgiKjGxQOMOysjIiIyO58sormTRpEiNGjKge2agOSo4eZd+iRRz4+GMQofvMmfSYNYvQuMZ3nJ3qGvvTggXseOEFhj/2GF2vuork5OSqXvPvv/++alxPXYSEhBAdFESoy0XPESPo1KcPcXFxxMbGnjQWSOXn4tmSAE4ntsEjsNVxL5w1a5Y2xutCG+PNn5q6Lhk7lgt/9zvip0414/wqxcQFC7AFBqJ+2oXnq88gIAD7tGuwdetV776Vsxxj1zaMLQmQeRzp3ss0uHv0qRaSTGVnYGxZj7FzCwC2gUOxDRlVt1/zaWJs24jnu2WmIXqG3GL80V6Vx4Nn6SLUvp+wz7oDW6fGjxZ3f7QQlIHjulsbvY+zyeG5z7LmrYVMXvAfwvubcdp/TtcCY8cmPF98guOePyDR/gn/5SopIePVuWT9uIcst5CXlMTUDz6o821HbW1WFRbgfvkJ7NffVm1Aty/KMFh+000ERkYy9pVX6r0hVhQUcHTdOo6sXk3a6tWk5OSQ2aoVpfHxZLlcpKamkpKS0uRBaiJCly5dqg3m69G9O23btCE0IqLKaA4ODj5p7EFd2TlLMzJIXryY/R99hOFy0fXqq+k5axbh8fH11sWf9y5PRQV5iYnk7NpFxsaNHP7+e2KHDaPfvfcSO2xYk/ffVJI//JBd//oXM1aurDekZDU8Hti9G7ZvN6MRpaScmNfR41ovISEYbdqQ2aoVhwICOFxWRoXHQ1yHDsQPGkRk794cPnyYtIQEig4dIvqCC+h02WXET5lSbaB4UzmWkMCqe+5hxsqVBLeyNnoXmD7ZqcuWse+99yg8eJAOl15Kr9mziRk8+LSN1LrabOVb093z5jH66afpdNlltX43ISGBlStX4nK5aNu2LXFxcdXm4eHhoBRbn3qKA0uWcPHLL9PuoupvbJXhwfjhe4xV3yK9LsA+/RqkRfNyTbTUGBeRZzAH1h8EbldKubzr7cB8oAewVSk1x7v+AeA6zAH8s5VShT770m4qzZyauq781a9o2bs3zvx8jq1bx5QPP6wW51eVl2Gs+BJjy3qk/4XYp/wCCa3uC6eOH8XYkoDx4zZwOMxBFkNGnTKcoHJVoHbvMHvQj6YjnbthGzoa6d2vSUa0e8ErSExb7NOuafQ+Thd/tVelDIwvP8XYsQn7dbfWGf+93n2UleJ+4REzpXqvM/NmwN8ol4t1M6+iqLiU0R+ascd/LtcCVVqM+9/PYBs5Fvsljc9iWJ6TQ9a2bWRu3Urm1q3kJyYiNhstW0YQM3gwnW6/s94kIbW1Wc93X2Ak/ojjN3+sMykVQH5yMl/NnMmYZ58lfsqJAZBKKQpTUzm6ejVHVq8ma9s2HCEhxF10Ee3HjiXu4osJbll9XIRSipycHFJSUkhNTa0y0CvnaWlpuN3m0MLw8PCTImj07t2b7t27NylDq+fLJRiJu3HUkp3T43Ry6MsvSXrvPfKSkmg/diy9brqJ2BEjajV6GnstUEpRcvQo2Tt3krNzJ9k//kje3r0YLhfhnTrRauBAus+cSZshQxp9nv7Gn2H8ADPsZEpKdQPdu6xSUxEfFwgDyGrRgrSICNIjInDa7cSWlNCpsJAOhYUE1Yg0o4KCyOvWjbSWLTlUXk5JWRmtOnem0+TJxF97LS2aaJgn/OlPuIqLT0qiZjVKKTI2bCDx3Xc5umYNLfv0odeNN9Lp8surkmedirpcgHa88AJJ773HRXPn0mH8eL/UdedLL5H4zjuMeeEFOl56qbk+JwvPZ4tQWRnYL78a6X/6DxRnAsuMcREZCPxBKTVbRP4KpCilFnnLrgSGKaX+JiLzgbeBZOADYCLwSyBeKfWUz/78ZoyXeEfCh56DIfWaMzV13fr006R89hme8nImLlx4Umr0Soz0g3iWfQSFBdgnXYH0H4zau9M0pA8fQuK7mr3gfQY0OBugL+poOp4t61G7t0NgkNegH4lERZ/efvJzcb/8JPY7foutQ6fTrkdj8Wd7VUphfP81xg8rsV81C1v/waf1fc+WBIzvv8Hx0N+RsxzmqymU7t7JF7Nvptc1M+j20B8ara1SCo/TiauoCFdxcdVUUVSEq6Skar27tBRlGOakFCiFMgxzXsdnsdkICAsjIDycwIgIAr1z388B4eE4WrSourm4ly5GHU7DcfdDDX7orDTUsryGd9bWrRSmpuIICaH1oEHEDBlCmyFDaNW/P7bEXXi+XGJG0anngbhmm1UVTtxzH8c+cRq2IaeOQLTz5Zc5sGQJl3/6Kfn79nFk1SqOrF5NcVoaYfHxtB83jvZjxxIzeHCDjYLacLvdHDlyhMDAQNq2bWvJTfpEds447NfcVOsxlFJkbdtG0nvvcXjFCiI6d6bn7Nl0mT4dh8+r9oZeC1wlJeT8+CM5u3aRvWsXObt2UZ6TQ0BEBK3796fVwIG0HjCAVv37N+tshmseeAB7QABjnn/e0uOUFBTAsWOUbt/OoVWrSN++nfLSUtpER9MpNJQOLhfB2dknQl7W47usgNzgYA5FRpIWEUFpQAAxQHxMDPH9+hHSo0etYzPqwlVRwZKFCxk1fjzx3bo1/WQbSVFBAft27+ZAYiIOh4PuffvSo29fQk7RFmv2jCul2LJ2LSn79jF26lTa+jns7u6tW/lxyxZGTphAp0A7Kmk3REVj6z+44QEX7rvvjEb7AmuN8XuAEqXUQhEZAtymlLrPW/YM8IVSao2IXAN0AfZgGuiPiUgr4B2l1HSf/Y1cvnz5+uHDh1f9qE6nk4CAABwOR1X4It8yh8NRFXrHMIyqspKSkqrwQhUVFXg8nmrfs9lsBAUF4XK5cLvd1cpExBzZ7HbjcrmqlYHpv1RZFhgYiM1mq0q0ERwcjGEYVFRUVNW7zJseuann5HQ6sdvtBAYGnrVzEhHsdntVT9PRr75i22OPMfCvf6XrjBn1n5PNhvuH77GvX2X+4A47nj4DUAOHE9C+o3/OqbwM947NOHZtxZabjYqJRYmYN0cBZZhtX2wCiqqQYiKCYRhQXmZmtbzrIZw+v6HVv5Pb7a5qF35re1vXo1Z+iYqJPZH+u+b5csK/tKqsIA9334G4xk5pVm2vMf9P+59+gh8/+YyI6NOLMqOUwl3hwlVRgavCZRrRvogQEBhAQGAgAYEBOAIDcQQ4EJsNqWxf3rRHYhNzSXw/S7XjVDiduJwVuJwVpr6u6kHixCYEBAYRGBSIQ4DoVhAUXFUv398Q7+9bZfCL4MzLoywjg8DISFpfeCHRgwYR1b8/Mf36YfO24arfyeVC/jsfW0E+hIZVDWKr9GuubCcKhTK8bcYmZvSW8jI8d/8BR0jIKX8nd1kZa26+mdKjRwGIHjiQ9uPG0XrkSEI6dKiW3e9caHvOQykELnoTWrUBr1Z1/a+VFBaRsjuR1D37UErRIjyssuGZ8d597tFV92sFyvxjPlwVFiEiRLaOJrptDNFtWhPdtg0R0VFV32vwdY9argONKEPAJjYMZYCqUVZLGwI4vD+Vjd+sJrKV/yJB1UV5aRnlJaW0bhdLhx5d6NCtM0EtQk4+J6WgtAwKi6CgAHtWLiozE8nIxpadA9k5iNs01hWQHRJCWkQEaRERlDkcRDmd5j4aiNtmw+lwMGPfPuzNwI3YZbOREhVFUsuWlAYGEnmaicTcNhtlDgfj0tJo4/2f8TeJ0dFsj40lqhFJzvpnZRGTlYXTa9OcqWvE9u3bGTFiRJUx7k8n2JZA5SiYAsycEL5lhTXKalvny02LFy9m+PDhTa6Yy+WybNDJ+YzL5ao2eCh+6lRCYmJoOXToqb9st+MZcTHuHn0JzM2Ebr1wKfybZCE4BM+FI/AMGk5I9nGMw2kYhge73YGI4HGbA8AcjgCzx9PjxmazY7fb8HgHh9k7dz/jT8y+aYf9xoiLqYiIwpaTid3bg1r9fN2AOrnMYcfT/fTdW5ojHe5/EEJa4C4oqDIgK68LhmEgCDa7DcNjVCtTShEQHExQWCj2wEDsQUEEhYYSEBKMLTAAe2AQgYEBeDxGVfsCU0MwY5r7lokI7nraXuXAxMr9uJzleJwuPM5yyoqKcZWW4nFWUFFahhNMNyrvQ7FSqioOstvtxmazVUUlMAzDfGgMDSW8b18iunQhOCSkynCt1U9XhIrLZxJ0IBG89QSwBQRg+JyTy+3C43Z7bzTmOVW0bY+jgW3YHhzM6LlzKUhOJmrwYFpER1d7uD3XULHtqJgxm4Bc02/Z9/c1vL9T5e8b5HHTd/ho+hoe0rZsx1lYhM1uA7ztRGwEBgZgPk8Z3nChYoZ4FcHhcBDaJoaIDu0I8vYEVv5OUkvba9B173SuEbWck8fjRsSGzWHHcHtQyqhRJtgcjmptCKBN994MiIhGeTwoQ2Eoo9r/qFk3e7UyQfAYpjFst9lRmINbbWJDbFIVkcN3P8pQBIQE07ZfX0KiIk/rnMRhx+17Th4PxvHj2I4cxXHsOC3TDxOdfphBR4+RfSyP3Ea8yWnpdDYLQxwgwDDolZtLz9xcjoWFUdCI84krKSHKwqhEvXNziXA6KWjE2IqwZhJm1J894/cCxXX0jD8LLKvRM74XGFpfz3hCQsL6UaOanmRF+4xbg9bVGrSu1qG1tQatqzVoXa3hjOnq8UATY2CfaxQVFQHmmIxzhujos+6m4s+e8QTgIWAhMAX4oUbZRGCNt+w/wH7v9tSyPUCRv/5Rfk6RE5oTWldr0Lpah9bWGrSu1qB1tYYzpqvdDmcgGkpzIrAyoZpuu6eF33w3lFI7gAwRWQtcAHwiIvO8xcuAeG9ZuVJqvVIqC/hCRH7AHMD5qr/qotFoNBqNRqPRnAv4NXCyUuoPNVb92rveDdxay/Zzgbn+rENt6Djj1qB1tQatq3Voba1B62oNWldr0Lpah9a2cZyZLCZnGT140xq0rtagdbUOra01aF2tQetqDVpX69DaNo7zwhjXT2jWoHW1Bq2rdWhtrUHrag1aV2vQulqH1rZxnBePMC6XqypcmMZ/aF2tQetqHVpba9C6WoPW1Rq0rtahtW0c50XPeGVSGo1/0bpag9bVOrS21qB1tQatqzVoXa1Da9s4zgtjXL82sQatqzVoXa1Da2sNWldr0Lpag9bVOrS2jeO8cFPRaDQajUaj0WiaI+eFMe50OqvC7Wj8h9bVGrSu1qG1tQatqzVoXa1B62odWtvGcV64qcgZTnN6vqB1tQatq3Voba1B62oNWldr0Lpah9a2cZwXxnhwcPDZrsLPEq2rNWhdrUNraw1aV2vQulqD1tU6tLaN47xwU3G73XqErwVoXa1B62odWltr0Lpag9bVGrSu1qG1bRznhTGu415ag9bVGrSu1qG1tQatqzVoXa1B62odWtvGcV64qehQO9agdbUGrat1aG2tQetqDVpXa9C6WofWtnGcFz3jGo1Go9FoNBpNc+S8MMZ1qB1r0Lpag9bVOrS21qB1tQatqzVoXa1Da9s4mrObSuD+/fv9sqPi4mIAwsJpOnntAAAKI0lEQVTC/LI/jYnW1Rq0rtahtbUGras1aF2tQetqHVrbhuG1bwMrP4tS6uzVph5EJBLoClT4YXfhwE3Au0CRH/anMdG6WoPW1Tq0ttagdbUGras1aF2tQ2vbMAKBFKVUATTjnnFvBbf7Y18iEgVkAYlKqXx/7FOjdbUKrat1aG2tQetqDVpXa9C6WofWtnE0255xjUaj0Wg0Go3m5855MYBTo9FoNBqNRqNpjpwXxriIPCMia0XkXREJONv1OdcQkUgR2SQixSLSz7vuWhFJEJEVItLBu663iKzxrr/07Na6+SMiw0VkvVezRSISoHVtOiIS69VqtYisFJE4EbnIu26diPT3btdWRL4VkR9EZPbZrve5gojMEpEs77Jur35ARDqLSJaIrPJOMVpb/yAi47wafi8iM/S1oOmIyCiftrpPROZqXZuIUupnPQEDgfe8y38FZp3tOp1rExAAxAALgH6YYw3WYw5AGAPM8263BOgBRAA/nO16N/cJiANCvMtPATO1rn7R1Q7YvMu3An8DVgMtgXjgS2/ZXGCiT3sOPtt1b+6TV9slwDZ9HfCrrp2Bj30+a239o2sI8DkQ6LNOXwv8q/ECYKzWtWnT+dAzPhr41rv8NeaFTXMaKKVcSqksn1U9gJ+UUhVKqR+AAd717ZRSyUqpQiBXRFqf8cqeQyiljimlyrwfK4BeaF2bjFLKo5QyvB/DgQOARymVp5RKA6K9ZcOBlUopN7AF80FTUz+zgI8AA30d8DdjvG9w/4HW1l+MAsqAz0XkUxGJQ18L/IaIBGJqtwWta5M4H4zxlkChd7mAE41E03h8NQWztwyqtyetdQMRkU7AZGAdWle/ICKDRGQjcB+QQHVd3d6bSICP0a51PQUiYgeuAz7wrtLXAf9xDOgOXAK0Aa5Ga+sPYjF1vQKYDzyKvhb4k4nACiASrWuTOB+M8XzMV3pgNpjcs1iXnwu+mgJ4vHPDZ53WugGISARmPNZbMcNBaV39gFJqh1JqBPB/mO5pvro6lFIVgEtEKq+BWtdTMxv40Ofmqq8DfkIp5VRKlSjz3f4STPdKrW3Tycd056nANBovRF8L/Mm1mG/Kal4LtK6nyflgjCdgPr0BTAF+OIt1+bmQDPQRkUARGQ3s8q4/JiLdRCQciFZKZZ+9KjZ/RMQBLAYeVUoloXX1C94emUoKgGLAISJRItKREzeEzcA47+8wBNhzZmt6ztEXuFlEvsZ0o7gf3V79glerSi4GvkBr6w82Y+oowCBgL/pa4BfEDIYxDFinlCpF69okzos44yLyHDASSANu8z6xaU4DEfkS82J2CJiH6Yf3AFAO3KKUSheRvt4yO/CwUmr52arvuYCI3AS8BPzoXfWad651bQIiMhx4HrM3sRy4HdN4fBpQwL1KqZ1e/9GFQCjwulJq4Vmq8jmHiGxRSg0VkevR7bXJiMhlwBNAKZCK2WavQWvbZETkN8D1mP/7twPt0deCJuNts5cppX7r/XwJWtdGc14Y4xqNRqPRaDQaTXPkfHBT0Wg0Go1Go9FomiXaGNdoNBqNRqPRaM4S2hjXaDQajUaj0WjOEtoY12g0mrOIiDwiIp+d5TqMFJG9IlIkIr89m3WpDxF5XUSeOdv10Gg0Gn+ijXGNRqPxIiKrRMQjIgN81kWJiBKRzmevZpbzBLBIKRWulPpnzcKaDwxePQZZWaHaHlKUUncrpf5k5XE1Go3mTKONcY1Go6lOHvDU2a5EYxAT+6m3PIkunAixaTneGMUajUajQRvjGo1GU5NXgTHeuLknISILROQln8/Ves695W+JyMciUiwie0Skn4j8WkQOi0iWiNxbY7cO73cKRSRZRGb47F9E5Lcikigi+d7e+z4+5QdF5P+JyAbMONV9a6lzuIi8ISLHvNPrIhLqLTuOaYwv8ta3Z33iiMgm72KCd/u/eNd3E5HPved3SET+Vpl9T0RuFZEdIvKo93iLRSRMRJaKSKaIFIjIGhEZ6N3+KuAvwHTvMYrr0H6oiPzg1WWviMzyKXvEW59/e8vTvHHRK8snicgur2tOhohUxvnXaDSaM4o2xjUajaY6ucAzmAksGsu1wFwgCjML3VKgG9AVuAGYKyKxPttPBTYB0cBDmIZxN2/ZPcAdwBVAa8x06Z9L9UyjtwK3AGFAUi31eRnoDvQD+gO9vfVDKdUWMyHaLKVUmFJqX30nppQa7l0c7d3+HyLSAjPd+ArMpCoXe8/zNp+v9gPcQDxwE+b957+YDwKxwHbgQxERpdRnwD+AZd5jhNWsh4hEAV9jZrGN8eo0X0TG+Gw2BVgDtAL+BrwpJ7JdvgM8p5QKx/xd3q3vvDUajcYqtDGu0Wg0J/MS0MnbQ9sYvlBK/aCUcgMfAp0xMyZWKKVWAAWYRnEl+5RS85RSbqXU58D3QGUv72+Avyulkr3l/wRCgBE+339NKZWklPLUzDDs7Z2+Efh/Sqkcbwr1v2CmtvfXPWAakKeUesl7jmmYDwC/9NmmAHjSW16qlCpUSn2glCpRSpUDDwM9gXanccwspdS/lFIupdRqTOP+Fp9ttimlPlRKeTCN7UDvMQBcQHcRifHWIaGxJ6/RaDRNQRvjGo1GUwOlVBnwKGbvbGN8sDN8lkuBIu8+fdf59vYeqvH9Q5g9zGAa8u95XS3yRSQfaAl08Nk+rZ66xGAaoQd91qUAQZg97f6gM9CvRh1fANr6bHNEKWVUfhCREBF51etmU+hTv4bWqQPVzwnM8/LV5XjlgjLTTZcBlT3jMzB765NEZLuIXNfA42o0Go1f0ca4RqPR1M5bmNfIW2qsLwZa+HyO88OxOtX4HA8c8S6nA9cqpaJ8phZKqUU+2xvUTRZQgWkwV9IZcALZjayvqvE5Hdhao44RSqkL6qnj74AhwEVKqQif+kkd29fkMNXPCe/nw6euPiiltimlrsE0/h8H/lvDdUij0WjOCNoY12g0mlrwujb8FdOlw5dtwBQRifP6Hz/sh8P1FJE7RcQhItOACcAH3rJXgMdEpBeAiESIyJU+vs+nOg8D033jSRGJFpFWmD3+7/r2VJ8mGZg+8JUsA2JF5F4RCRYRu4j0EpFx9ewjAigH8kQkzFunmsfoJCKOOr7/JdDGe0yHiFyM6Y6z8FSVF5FAEblJRFp6Ncj3FrlP9V2NRqPxN9oY12g0mjpQSn0C7K+x+j1gNZAI7AC+8MOhvgZGYg4efRmYrZRK9pb9G1gALPG6c/xEdV/shvAApkvHXmAP5jk91IT6/h/wTxHJE5E/K6WKgYnApd7j5GA+ALStexe8CHgwje7dwPoa5R8BhUCW1+2lGkqpPOAyYLb3eG8A9yil1jXwHH4J7BeRIuBfwC+VUjkN/K5Go9H4DTHd6DQajUaj0Wg0Gs2ZRveMazQajUaj0Wg0ZwltjGs0Go1Go9FoNGcJbYxrNBqNRqPRaDRnCW2MazQajUaj0Wg0Z4n/DzMH7T6YMsrvAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "'''PLOTTTING THE LOSS GRAPH'''\n",
    "\n",
    "plt.figure(figsize=[30,20], dpi=30)\n",
    "\n",
    "plt.plot(iter_list, train_loss_list, '-', lw=3, c='salmon', label='Train Loss')\n",
    "plt.plot(iter_list, train_roll_loss_list, '-|r', lw=6, label='Train Loss [Roll]')\n",
    "\n",
    "plt.plot(iter_list, test_loss_list, '-', lw=3, c='brown', label='Test Loss')\n",
    "plt.plot(iter_list, test_roll_loss_list, '-|k', lw=6, label='Test Loss [Roll]')\n",
    "\n",
    "plt.title('LOSS vs ITERATIONS', size='30')\n",
    "plt.xlabel('Number of Iterations', size='30')\n",
    "plt.ylabel('Loss', size='30')\n",
    "\n",
    "# plt.yticks(range(0.0,2.0,0.1), rotation='horizontal')\n",
    "# plt.xticks(range(0,len(iter_list)+1,1), rotation='horizontal')\n",
    "\n",
    "\n",
    "plt.grid(True, linestyle='-.')\n",
    "plt.tick_params(labelcolor='k', labelsize='20', width=3)\n",
    "\n",
    "plt.legend(fontsize='20')\n",
    "fig1 = plt.gcf()\n",
    "plt.show()\n",
    "\n",
    "fig1.savefig('loss_vs_iterations[margin=1].png', dpi=75)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 860
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2452408,
     "status": "ok",
     "timestamp": 1562560451829,
     "user": {
      "displayName": "Aditya Ahuja",
      "photoUrl": "",
      "userId": "08809216472288903233"
     },
     "user_tz": -330
    },
    "id": "Zy_q2knxsKd-",
    "outputId": "a3ca26fd-478b-4d81-ba3a-303b8a0d098a"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAuAAAAHyCAYAAACqKHsEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAAEnQAABJ0BfDRroQAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOy9eXwcxZn//64Z3bIlW4d1jXzKkmzL+Mb3JRnbmMsyG8iGJeHIkt/ibA6SbAiBsNkASSDZDTlI+O4GQggJECIZmwSDNZJv4wtfsi3JtzW6Zdm6r5mu3x89I49GM9JImtZFv1+vednqqq6ufqa65+nqpz6PkFKio6Ojo6Ojo6OjozMwGAa7Azo6Ojo6Ojo6OjqfJXQHXEdHR0dHR0dHR2cA0R1wHR0dHR0dHR0dnQFEd8B1dHSGNUKI14QQUggxzU1ZsBDiOSHEOSFEoxDCIoR4Twgxz6nOeCHEH4QQpUKIeiHEeSHEr4QQcfbyy0KIjS7tTrQfc4z97z8IIdqEEA1CiBtCiCNCiHUe+psrhGgWQox1UxYphPilEOKKva3L9raThRA/tLfr53J+BUKIb/Tdgt3jfK5CiA/t/Wqwn6/V6e8Ge/2dQohW5+1CiGqn9i7bz79BCFEthPi7ECLJzXFDhRB1QoiDLtsLndptd7J7gxDihL2ORQhxp9M+YUKIl+3bm+3f8dMutvyy/Tx/4nK8D4QQTzv9facQ4rAQolYIcV0IccjTd62jo6PjCd0B19HRGbYIIUYD9wE1wKMuZf7Ax8Aq4H5gDJACZAGZ9jrjgcNAO7AECAOWAmXAyl525xUp5SggEvgj8J4QItylT5Pt/WkCHnApCwf2AxOAdfa+zAYOArcDzwEC+J7Tbs/b+/pyL/vaJ6SUt0spR9nP8wXgA8ff9m0Ovuu8XUoZ5dLUP9vrTwRuAK+5Odx9gA1YIIRIc+pDitPx3gF+6XScWa6NCCECgB1AGpAOjAb+Gfg88KZL9evAV4UQ8e7OXwiRDLwN/BcwFogH/gNocFdfR0dHxxO6A66jozOcuR9oBL4LPGh3uh18AZgG3Cml/FRK2S6lbJRS/llK6ZjR/CFwSkr5r1LKy1KlQkr5gpTy7b50SEppA34PjAKSXYofAY4Dv8LlgQH4BqAA90opC6SUipTyhpTyt1LKl6WU7cAXge8IIW4RQiyzt/ewdCNnJYR4XwjxA5dtvxVCvGr//wP2NwP1QogSIcQzfTnf/iClbAD+DMxzU/wo8Dqwm6626g1fBCYBG6WURVJKq5TyMLAJ+Ce7HR1cBLYC/+mhrblAiZRym/37aZZS7pRS7utH/3R0dD6D6A64jo7OcOZR4C3UWclQ4C6nsnXAh1LKG93svw74iy87ZH8I+ArQBlxx2m4EHgL+gDpDPksIMdelL+9JKa2e2pZSngZ+ZN//deDbUsrLHqq/CfyL0/EDUGeV/yiECLX341Ep5WhgBrC9F6fpE+yz/g8CRS7bU1DfRPwBeAP4F3v/+8I61Jn6eueNUsoi4Aiw1qX+0/bjpbhp6wgwQQjxGyHEOndhRDo6OjreoDvgOjo6wxIhxHRgEfCGfSY1m84zpdFASQ/NeFPHW/5NCHEDaEGdWf+8lLLSqXwdMA74s5TyIrCvD/0F+Dl2515K+X/d1NsGRAkhFtn/vgO47jRb2w5ME0KE2WfaD3txbG/5sT0W3vHZ4VL+lhCiFjX8ZA5ODwp2HgWOSylPAu8BIcA9fexLFFDqoawU1e4d2L+b36OG9+BSdh5YBoSjhs1UCSE+EkJM6mPfdHR0PqPoDriOjs5w5VHghJTyhP3vN4B1QogE+9/VQILbPW/iTZ12wN9lm79TmYPfSinHoDp0e1Bjyl37+w8ppWNB4hvAF4QQQb3oC1JKBTgD5PdQrxV4FzUEA/u/b9rLGlHfFtwDFAsh9gohVvd07F7wPSnlGKfPbS7lD0gpw4FUwA+Y4iiwL4z8Iqp9sM9cuz5c9YZq1Fhtd8QDVW62/whYK4S41bVASnlESvkvUsoE1BAnf0dfdXR0dLxFd8B1dHSGHfYwjweBZCFEuRCiHDUUxRHmAfARsN51IaQLH6HGkXfHFdQYYmemANV2R7YTUsoa4MuoM+Jz7P2NRnV4M5z6+xPUhaH3OvXlXmdlDh/wJnC/ECIWdSFnx6JDKaVZSrkBdYb4r8AWIcSA/iZIKQuBbwO/FUIE2zffCcQAzzjZ6m7gNiFEYh8OswO4UwjhvEgUu/LKfHu5a78qgf9G/Y666/851Hj+mX3ol46OzmcY3QHX0dEZjtyNqhIyF1UpZDYwC3Xm8hEhhEB1yAuAbUKIOUIIP6HK9t0nhPiRvZ1ngdlCiN8JVY5QCCGihRDfFUI4HPM/AZvtbQghxATURXpveeqclLIUNX7ZcZwvoiq1pDr1N81exzGz+z+oDxDvClV20CCECBdC/KsQ4ut9MZI93OS6/ThH7OEVCCFihBCZdhUZK1Bn/3cw2AJcAzbb/34UdSHkDG7aKhk4Dzzch/bfAK4C2UKIqUIIo1BlKLOAbCnlHg/7/Rz1O3KE8CCEWCmE+P/ETYnKONSHrf196JeOjs5nGN0B19HRGY48CvzFrhZS7vgAv0QNK1htVw1ZhxoO8ldUJ/Mc6kLEbAAp5RVgARCEKvdXBxxADQXZZT/WG8DPUB3uWiAPVZnjqR76+CI3wxgeRQ1RKXHp78+BVUKIKVLKWtSwlRIgB6gHTqIuRvx7303Fm3Y7/NFpmwH4OlBsP6fNwD/Zw1t8wU9FZx3wBiFEpLuKdgWXHwP/YQ8fuh34b2c72W31K+Bh+8OV19hDcTJQH8Z2oUoGvosaW/5AN/vVoz5AOff7OrAB+FQI0Yi6KLOSvj0Y6OjofIYRbtSrdHR0dHR0dHR0dHQ0wpexhv3GHqs5GXWFv46Ojo6Ojo6Ojs5IIQC4KKWsHVIOODB5y5YtnyYldclK3Guam5sBCA4O7qGmTm/Q7aodum21QberNuh21Qbdrtqh21YbdLt6z/nz59m4ceNc4NhQc8DbkpKSmDFjRr8bam9X1cH8/V3Vw3T6g25X7dBtqw26XbVBt6s26HbVDt222qDbtde0wRALQfElra2tgD4gfI1uV+3QbasNul21QberNuh21Q7dttqg27VvjFgHXFF8tZhfxxndrtqh21YbdLtqg25XbdDtqh26bbVBt2vfGLEOeGBg4GB3YUSi21U7dNtqg25XbdDtqg26XbVDt6026HbtG7oOuI6Ojo6Ojo6Ojs4AMmId8NbW1o64JB3fodtVO3TbaoNuV23Q7aoNul21Q7etNuh27RsjNgTFaDQOdhdGJLpdtUO3rTbodtUG3a7aoNtVO3TbaoNu174xYh3wgICAwe7CiES3q3bottUG3a7aoNtVG3S7aoduW23Q7do3RqwD3tamJtPUZXF8i25X7dBtqw26XbVBt6s26HbVDt222qDbtW+MWAfcZrMNdhdGJLpdtUO3rTbodtUG3a7aoNtVO3TbaoNu174xYh1wXRZHG3S7aoduW23Q7aoNul21Qberdui21Qbdrn1jxKqg6Ojo6Ojo6Ojo6AxFRqwDrsviaINuV+3QbasNul21QberNuh21Q7dttqg27VvjNgQFINhxD5bDCq6XbVDt6026HbVBt2u2qDbVTt022qDbte+MWIdcD0mSRt0u2qHbltt0O2qDbpdtUG3q3bottUG3a59Y8Q+trS3t9Pe3j7Y3Rhx6HbVDt222qDbVRt0u2qDblft0G2rDbpd+8aInQG3Wq2D3YURiW5X7dBtqw26XbVBt6s26HbVDt222qDbtW/4dAZcCBEuhDgkhGgQQqTZt31OCLFfCGEWQpjs21KFELvt2zN82QcHgYGB+msRDdDtqh26bbVBt6s26HbVBt2u2qHbVht0u/YNX8+ANwF3AC8BCCH8gCeAlcAC4BngK8ALwKNABfAhYPZxP3R0dHR0hjKKAgcOQGEhzJ0Ls2aBEIPdKx2dviMl5OfDrl3wGVIFMTpmwP2GUVDFrFmwZs2gdsGn1pJStgNV4uZNdCpwVkrZBuwTQvzMvj1eSnkOQAhRI4SIklJWA6PLyspITEzseJpqbW3F398fPz8/WlpakFJ2KvPz88Pf35/W1lYURekoq6mpwWAwEBERQVtbGzabrdN+BoOBwMBA2tvbsVqtncqEEAQFBWG1Wmlvb+9UBhAcHNxRFlByBYPRSEusCYCgoCAURaGtra2j383NzQD9PqfW1laMRiMBAQHanlNAAAaDgZaWli7n5OiDY7/AwECKfvQspikTCAkPwyAEVntWLKPBCIBNsWEQBgwGgc2mIJGdyoQQGA0GbIqClC5lCIxGA9Z2GwW795K8dBH+gYHYFPUYfkYjipQoitJpP9cyg8GAEKIjY5fRaEQ6lXXqd8go5Op1tLZbB/R7qqmpASAqKsrt99R05jSXXvk1ilTtJIRAAIqUABiEQKrXYaeysXGxTF18683vwvV83XxP7XGJWKfP8vqcbuTnY7txg5hVq7wee+1FZzAW5vf9e3LudzdlV47nU33pCkIIJBKD/f6kSIlAIIT6u9mbMld7d1fm7ffkKOs4fndlfex3b87J3z+AtLWrMfj7ub2erFYrik1B8fe7+T0Z/SB9A35BQe7ve1Liv28fyt/+ht/f/46tqorj48ZhEwI5ejSMH4/NZEJGRWGw/5grNhvCYMBgMKAoClJRCImJYcbmzbS1tXl337teg8w/hnHp6o4y6P6+Z62s5Oxrr6FYrR3Ht9kdDYPR2KVvNpsNWpox1lT167uw2RT72DcMibGXOHMGcVOndFxP/b2XK4pEkZ3v15ePnuC6pUTzc5KKVOsZhO+uJ4C6OqisRlRXI5rUcd9bjFIyu6ICf3vftaDNYOD4uHEon/EH3dFtbaTeey+Nt97qM9/IW3+voaGhox9aP66MBeqc/jba/3UOfakFIoBq4MG3336bW2+9VeNu+Q5x4gjU3YDPPzrYXRkQhBA4PWBhLS/lVPY2rCuWMD1jBQiBUBTUW4i9nqKAkIABpKLeyTqVOf3fXZkC1y5f4WzeXkLHhDNx9i2gKGotx51RUTrv16XM/ndPZRI4sh+SUiE2wae28wbRzY3x6h9eo6KwiKgpkzrMJBFI+w1bvalKJxMKWpuaOPVRLhNnz8Q/MLDT+Xr8nhoa8Ss8jTUlzet+F/7f/9Fw6RJrV6zweh///XnqdxAd0/vvybXfHs5J2myc3pHHqMgIQsPDkB12osNuCNWGvSkTgLQ79FJ23c+5zJvvqXOZvR17Gb3sm6/O6cqZU0SY4jHNnOb2uxCKVK9nx3UqJcYzR1EmJ0Oq09hpbMSwYwdB2dn4b9+OqK3tKLocHs7l8HAS6uuhoQHOnMFw5gwEBSFNJuT48ShRUQijEWE0othsKDYbha+9hmntWoInTMAbxMkj+O3LRZkxC8ZEeLXPhXffperQIcbOmoVBqo6bYn/Ac1ylis3WUSZtNkRVOUpTE4SE9v27UGzATadyMMdeXWUVZ6triJsyyX7C/b+Xu5bZ2to4td1MZGICgaNCND0naf+ukIb+XTOKAjdq4VoNVF+Dtja1vr2vfeFqWBjRTU1MrKvruXIfKQ4L44rjevsM09fvyNdo7YDfAMKc/rbZ/1WctoUDNfb/v/noo48+HhZ2cxfnuCJ/f/9OjXdXFhUV1bG9N/u5lgUHB3dbZpUKsszCaCQiLLyjPCgoqE/9Hoiyns7JQUBAQKd2goKCOlY6O45R/eFWAMqrrjPnnx5Uy/BMX8vKf/xj9d/aZpL/6V+6qdl3HMe3vvEK4kIhgcnTbpYNwPfkOmadvwt/f3/KTpxk6vq1zPjh816fk621lb8tXcq1KBPj163rVObJ3rK5CevPnmVUTSWGsSk99lu0tlJ99CjSaqX54kWibrnF4/k6zknW1WItL8H46NcwmLxzonrqt7uyuhMnaG1o5LasbMISBv6Bajiz/8knKa9rZup9X3Jb3t7ejj+dx4P1L7/HcLEQP9NE/D/4ALKy4KOPoNn9rGBxfDyJ0dEsam6Gq1c7F168qP4bHg533gmZmbB+PYSG8o9Nm6jYvZtbNm/uqN7dtcb5AgD8LhZhXLyyx/uelJLSnTtJffBBUr/k/vxdkYoN689/iHHNHRjmLPRqH3e43mMHk6pPP2XHF7+IbdV6gu33Jwd9vZe7Ur1nD4qisOLPb+M/alSf+ukt/bJtczN8/LE6prdtg+vX3dczmWDjRvXfXrBvzx6KJ09m4sqVve+blxSbzYwPDmbhkiU+bdf5jeWwYc4cnH3N/vpGznWdcW1nlNMY19oBPwdME0IEAPOBk/btZUKIKUAlEGEPPwGodzZIf3CsytX8JmY/jlKYj3HBUm2PNQRwtWvd0SMIg4HrZ8/SWFpKaHy8z48ppcRiNjNuwQLK9u/H2tyMn9PF4GtE6kyUA7swrL+n2xlpX9PdmK09eIDGugYS7/1cr9o0BgYSt3y5euN1ccA9IYJDEBOTkAX5MCWlx/qlu3fjFxTEqMRELLm5nRxwTyiFp2FUGCIh0as+9RWL2czYtDT8I7yb9Rzy3LgBe/d6dGh9SWJwMPu3b8f65z/j52ZMKvZZP5x+jAwnCiDrb8hLDyI8KSMsWACZmVhvv53yRx5h6bPPwurVcOyY6txkZ8OZMzfr19bCW2+pn6AgWLcO07hxWD7+uJMD7gl5/RrnzpymJHg0c48dJnJxzw7OjcJCGktKMGV4rxEgiy9DcxMieYbX+7hjwH67vCBy1iyCIiIoycsj6XO9u/d4i8VsJmbRIs2db+rrUXJyoKmp05jtjhuNjew9dozmgwfV8ekY867Ex8Ott6qfKVP6tJahuq6OorffpvRf/gWDl9+9v78/ixYtIjY2tse67Y2NlL/9Nsv/539g1ape96872uz3o+A+/i5LKTlz5gxnz569+dZBY6ZERTF3QI7kGZ874EKIfwCzgRTgVeAXwE6gBXBMJXwf+ANqSMqzvu4D3HzS7euA8BqrVX09VXAKPgMOuLNdZWM9dZcuEZmSQnNtLcVmM6kPPujzY14/e5amsjJWv/oq2++7j7J9+0jUcPGEIWUGyvYtUF4Ccb2bxegP3Y3Z4i1ZhEWMIeyW2b1u15SRwZHnnsPW1obRyx8ekZqGsicHw4ZMhOheLKnYbCZ+xQrGJCdzaetWZn/jGz22LwtOYUhN67Ht/iClpNhsJmHDBtrb27W/F2hFRQW8/77qnObmwgDp7cYJASkplH/lK5ic4hYduNM8cPttGgywcqU6g71xIySqD11lO3aAwUDskiWqwzJ3rvp57jl1YWZ2tvo5dOhmWy0t8P77JAYGkj9lCg2rVzPqc59T23V6+JdScuzYMbKzs8l660+cuXQZAH+jgYwtH7Ppc5/jnnvuYdy4cW7PvdhsZkxKCqN6MYspz+Yjxk9ChPbPkRyw3y4vMBiNJKxeTbHZrIkDrthsWHJzueWrX/V52wBUV8PWreq1k5NDoBcLI8uB94FsVHUIrwT2Skthyxb1018eeKBX1YUQLFmyhMzMTDIzM5k8ebLbemV792IwGoldvLj/fXShL2NWURQOHTqkXqNZWZw/f97n/eqOf/u3f+OVV14Z0GO64nMHXEq5wc3md1zqnAGW+/rYzgyYJI61HTFpKvLyeWRzEyI4ZGCOO0g421UWnqGuqYWwubcQGRKCRSMH3GI2M3baNMKnTCFu6VIsZrOmDrgYEwGxCSgF+RgH0AHvbsxaDh7GtGB+n9pNWL6cT5qbqTx8mLil3j0kGlLSUP6RhSy1IBLGe6xna22lbM8eFj3/POFTp3LiF7+g9uJFwj38CADIlmbk5fMYlqzq7an0irqLF6m/fJnxa9YMP4msS5duOqD79tljZgcWPymJbWzEMnq0Wwe8O6S/P2LdOti0Ce66C1zCFwCKc3OJW7oUP6fXtx2kpMCTT6ofi0V1bLKyYPdusNkY09pKaFsblpMnSd25EzZvxrZwIfvnzCGrpYXsvDyuXLnSpdl2m8L2HTvYvmMHX/nKV1i2bFmH4zJx4sSOehazuXez31KiFOZjuHWZ1/t4YqiNVVNGBnu+9jXaGxp8Pkt97eRJWmpqSFi92neNFhffvHZ273ZaM+KZi6gOdzawHxj4q63vSCnZt28f+/bt49vf/jazZs1i06ZNZGZmkpaW1vEWt9hsJm75cowajC9vx2x7ezu7d+8mOzub7OxsSktLfd6X4cQw0owZmkirFcOUFGRpMfLcWcQt8wa7SwOGUnCK+pZ2Jk+eTMSMGRT9+c+0XL9O0NixPj1OsdnM+LVrATClp/PpT3+K0t7u9Wu6vmBInYly5jjG1es1O4a31J89zY2qayzcuLFP+weEhxNz663qDdhLB1yEhSMSxiPPnoJuHPDyAwdQbDbili3DPzSUsEmTsJjN3TvgRWfAPwAxcUpvT6VXWMxmwiZPZrSTYzVkkRJOn1adhqwsOH7cfT1/f8jIgEmTBqRbpvJyjl+4gHLPPR3qDw7cxn36+6NEj8XmZ8Pvu//lMYRLaW+nZOdO5n/ve150wgRf/ar6uXYNtm1DZGVhOnqUC6NGcammhmzg/YMHqTx40G0TMVFRTEtLY9/ePbRb1X5LKdmzZw979uzhiSeeYM6cOWzatIl1ixdzvaCAxS+80HPfHFSUwo0aDKneL1weLsQuWoTB35/SPXuYcPvtPm3bYjYTNWsWwdHR/WuooODmtXPkiPs6fn4oK1Ygk5IwGAzk19SQfekSWRcvcuLaNbe7BBgMpJtMTJo8GSZOBI3fStRevMi1EyeYtHGjV+GPFRUVfPTRRzQ2NnZsO3HiBCdOnODZZ58lKSmJzMxM7rnrLiy7drHwmWe07L5bmpub2bFjB1lZWWzbtq1D7cuVhIQEbrvttgF787OiF4IBWjFiHXBnmTxNsbYjAgMRydNRCvIxjHAH3GHXACTK+SLqr9UQNnky0XPnEhAWRunOnUzOzPTZ8eqvXKH23DkSX3wRgIRVqzj4gx9QefQosYsW+ew4rhhS01B2bkfWVCMius7eaYGnMWt576+EjAph7LK+L84xpadz+tVXkU8/jTB4F/YhUtJQTh7BuOYOj3UsubnELl6Mf2Mj1NRgysjAkpvLjH/9V4/7KIX5iORpCKO2t59i+yzmgN0LeouiwOHDN+Oez51zXy80FDZsUEM4NmxQFyUOEAk1NRxauZKqRx4hZsGCTmWNdrUGx7qd1tZWrl69SmLYKIy/+5nqlHpQEqo8ehRrYyPxvf0RjIyk4Z/+ie2jRvHn5ma25+biKRp+EpAJbBw9isUPfAG/O+/iuuUKH7z5OlsVf/7xySc02eXEAI4dO8axY8d4BogPDubkCy+wfvVqQqZN6zGm13b0E2RTO35F54H+vUp3OFOhoaH9aseXXJ86lY/+9Cdm+XAthZSSPe+9R+KaNRz08ODUzc5qmNLOnern8mX39QIDYdEiNeZ56VIqW1rIycnhH//4h8eQh9DQUDZs2MCmTZvYsGEDvlqX5g0t166RvWoVGV/+MuPmeedLtNjPKSsri61bt3LN6WHi/PnzvPTSS7z00kuM8fPjvi1b2Gg0EuHjNTHuxuyFCxfIzs7mww8/7PSA4ExycjKZmZls2rSJ+fPnY/Dyt2nE4JDlGQofYEZ+fr70BbW1tbK2ttYnbXVH23//l7QdPyxtp0/ItueflEpbm+bHHEwcdrXlH5c1T/67fGv6dFl35YqUUsoDTz0ld331qz493pnXX5fvr1snFUXp2Jbz0EPy8PPP+/Q4riiKIttefl5a9+VqehxnPI3Zj9fdJg899mi/2m4sL5dvTZ8uq06c8Hofpapctv3nE1KpKndbbjtzRr43d648N2eOlHaZ3KrUVPnW9Omy8aOPpHT6zjrabG+Tbc8/KW2nj/f5XLyhsaxMvjV9uqw+eXLA7gVe0dYmZU6OlJs3Sxkf32G3Lp+ICCkfekjKrVulbGoa1C7v+NKX5JEXXuiyvba2VlosFvnOO+/I+++/X4aEhEhAjh49Wt6/YI58+wdPyfr6erdtHn7uOZnzyCNe96G6ulq+/vrr8u6775YBAQESu+Kb6ydt7Fj5g5AQeeymSJ7HTxPI90E+BDLCQ3v657PxiYyMlA8//LDctm2bbG5u7vO14gs+fvBBefSnP+3Tvu3t7TI3N1f++7//uzSZTINuV3efuXPnyh/96Efy9OnTnX7XPyvk5+dLYIaUcuTOgA/YAharFfz8EEkpIBXkxSJESv9WwQ9lHHZVCk/RMGoMBn9/Qu3ybqaMDPZ9+9tYm5rwC/FNLLwlJwdTRkan13GmNWs4+9przPve9zRTKRFCYEhNU5VAlvgwPrEb3I3ZllILVZYS0v7t/+tX2yExMUTOnIklJ8crlRIAERUDkdFqLPwyu1b38eMdM7bVly7ROmECpqKijn0iCwoInjqVks9/nqmjRqmztps2wbJlYDQiL55TNdyTUvt1Pj1hyc0lOCaGiBkzOpLzDBrNzbBjx035Mg+vYElIUO2VmQkrVgyZrHKm9HQK33yTuU8+iRCC6upqtm7dyt/+9jfMZnPHGwYH9fX1vHP4GO8cPkbgT3/O2rVryczM5O677yYyMlJVNcrNZdojj3R73JKSErZs2UJWVha7du3qCHlxZXpMDF964gkyMzOZOnUq2Gywbx/yvfeQf3wDUeteVzkYuNv+sQK7gXeF4K9GIzWeFFx0Rgwmk6kj/n/58uX4DaHrregvf2HOd77T6983Pz8/Vq9ezerVq3n55Zc5cuQIWVlZvPnyy5QMgHKSO4QQLFu2jE2bNrFx48ZOay0+6wyNEacBAydD2K464AGBiMnJKAWnMIxgB9xqtYLNil/RWeqNIYyeOLEjK1zskiVgMFC2f79PFkk2V1dTdfw4s775zU7bTatXc/SFF7h+5gwRM7SztSpHuBvZUI8YNVqz4zhwN2Ytf32XgIAAxt3uOQzEW0wZGVx6/31mP/GE1/sYpk5Hbn0fsv6uOpBOC9ssMTFENzcTZLNBWBiEhCDKyzHV11M8ejRTr16FX/5S/URFwd13IxOiEFMmIQK0DQex5OZiWr0aYTBgtTuIAyrrVlsLf7fbbPt28LK31uMAACAASURBVPAKluTkmw8p8+eriiFDDFNGBjuee46ffP/7fHzgALt370bxsLAtKSmp06v91tZWtm3bxrZt2zAajaxYsYJ1ixYRUlzMbenpXfYvKirqUEU45Kx+4oTRaGTVqlVs2rSJBVFRXPjxj7l382b8Ha+/jUZYsQIZNQZbTAh+t21CbNsGJ06AoiDrapGV5YgpyR0Ojh+QDkyoryfj2jVM4eG8X1HB6bIyZFOTe8MYjRAdrWbwNEh1EsYHDFVN5WunTmEMDGRMcrJP2qv69FMCx44lzHU9g9UKVVVQVgaVleoDlTtCQiA2Vv2MGeOV9J+iKMyYMYP77ruP+fPnD6jMrLckZmRw7KWXuFFUxNiUvo8pIQQLFixgUmAgM7duJfV3v+PD3bs5dOiQx4fZvuJuzIaGhrJmzRruvvtuYmJifHq8kcKIdcAHVIbQT/1hN6SmYdvxAVKxIQxD6+bpK9rb2zFcuYBfexv1raLTYju/oCDily6l2EcqJSV5eQRFRBA1u7P0Xmh8PBEzZlCck6OtA26aAKGhyMLTiHnaxZs7cDdmLTt3Ej89FaMPHFZTRoaqUnLhAuFTulkA2damyt1lZWHIzkZUV3epIoHiMWNITkqC3/1O1XH284ODBzH9v//HzoMHaTMYCHA4atXV8NprGAEZEgIf7lQdz6VLVUfGh7TV11Nx6BDTN26EsjKsjqxvozV+iGppUWe6s7PBbPYsFzh37k2n24v4Yldu3LjRke5YS65du8a2bdvIysriyLlzYE+G5Yxjxm3Tpk3cc889xMXFUVpaqs5cv/zf7LpwueMNhM1mIy8vj7y8PAD+tHEjmZmZLFy4kNzcXLKzszl9+rTbvgQFBbFu3To2bdrEnXfe2RHDam1p4dKLL1K2d28XnXul4BRi6jSEQ5/ZQXMTtp89i/ELX0a46NwXP/44iZGRLPrRj1gMdLs41maD8nIoL0f6+SEmT1W/1zVrVL3yPlJvH6+jtR6vveTcBx9w6o03yHz1Va/XkXiisaKC9x94gPWvvEJEcjK0tqr3nOxsNabbk1zgLbeo101mJsyc2etrp85l3cJQZFRiImOSk1X1r3444A4sOTlE3nILc5cvZ+5ybcTnhoNdhyIj1gF3zVSkBVJRQLF1vCoWyTNg21+RxZcRE7RVeBgsAgICEJeKEJOSqPtoLzHOP2yoTt7RH//YJyolltxcElat6phh73Sc9HSufPghs77+9X4dozuEwaAuRCzMxzAADrjrmG2vvUH5hUss+bb3M9bdET55MmGTJ2PJze3qgDc0qDO1WVmqbq59xrbLz9ukSZCZyY1bb6Xxv/6LxN/+tkPXGYDFixk3bx5+K1ZQ+utfM7GyUv1RPXGio4poaoJ33lE/GlASHo5fbCzj7roLgEF3Y4SA5ctv6mD34RXs2bNnycrKIjs7m6NHj/q+j70gODiY9evXc9ddd7F+/Xri4uI6lcfHx/P444/zlSkmrhUVsD08luzsbD766KNODw6HDh3yOMsN6o/5XXfdRWZmJuvXr3e7KNEvKIi4Zcuw5OZ2csClzYYsOoNxw6Yu+6iJpqZ0STTV3thI+YEDLPvv/3aqLCAtTf0884wqD+mQRHSShxRWK3z4ofrpJ4M+Xj1gMho5nJxM9dSpRPfzAdAydiwhUVGM9SYhzJIlN0Ozups48IKB8At8gWnNGixmMzMff7zfbVlyc5l0zz0+6JVnhotdhxpD732njzAYDNqvqLXZ4wTtM+AidBRi/CTk2XxtjzuIGASIc2cRKTOpu3Spy+vDhJUraW9spLKfTkJ7QwPlBw541OI1ZWRQe/48dW60fn2JITUNebEI2drSc+X+HstlzJZmvYcA4jZ2dSL6iik9neKcHPWPa9fgD3+Ae+6B6Gj43OfgL3/pEi4hJ4xH2XiHmgnuwgX4+c+xXLvGmORkRiV2zWRpDAggfsUKiouK4Nln1VnDCxdQ/vUhZNLkPmWJ6w2W0aNJqK8f3Jubvz/cfjv87/+qs6S7dsE3vuG18y2l5PDhwzz11FOkpqYyffp0nn766UFzvsPDwlgWHs5br75KdXU1WVlZPPDAA0RGRnrcR6SmMbamkgfvzSQ7O5uqqire/O1vWRoeTpiH2d2YmBgee+wxtm/fTlVVFX/605+49957u1UEMa1ZQ8muXdicshTKy+ehvQ0xdZr7vtkfrKW8GUrjVaKSSZPgm9+EPXugrAzl+0+ipExBDoGslVoTbLMR1dyMxQcz85awMEx1dV0f8EGd0LrtNvjtb9UEN/v2wbe/3W/nGwbIL/ABienpXC8ooMFi6Vc7tRcuUHfpkqZ5M2D42HWoMWJnwFvs8lKaPpnZY3aF0+INkToT5eAeDOvuHpLxZf2l9fJFghrqaY9NoPX6dcJc9J4DwsMZt2ABFrO5XzKBpXv3YvD399hGeFISo8aPx2I2M72HxVz9QUyaCkY/5PkCxIzeZ6HsDa5j1pKTQ+zUKfiH+U52zjRzJmf+7/9oWrWKkL17PcdXLlzY8apXGhRsf/4/RGpKx5juKVFJYkYGn3z/+9haWzEGBiInTcI2IwnDww9hnJSiZnZ8/301yYoPsUpJqaKw2GAAu3Nos4fBGAfiB2LaNHWWuw9ygVarlb1795KVlcWWLVsoLi52W2/MmDEkunnw8TV+fn4sXLiQTZs2sXLlSrbffTdTAwIIsS+w7ukeKxInQkgosvAMYu5CQkNDmevvz3eXLmWtfWFlVlYWBQUFzJs3j8zMTBYvXtzr2OeORFNHjhC3ZAkAsiAfMXGKx8RohtQ0lA+zOyWasuTmErdsmfvEQO6IiUGZkQS3Po/fyvVqzP+WLaomdT8Y0PHaS0yKwvmQEGaPG9fn37dWKalUFNJGjwbnTKRTp6rXzp13go9zSTgYEL/AB4xJTSU0IQFLbi6pX/xin9ux5OYSNnly1zh7HzNc7DrUGLEO+IBgtcd4+t2c/TCkpqF89H63GrjDGeP5ApTYBOqrVK3RMDczeqb0dFWl5Kmn+nyTtuTkEN9N1i4hBIkZGdo74H5+iKmp6uJajR1wZ2wtLZSeKWDuIw/1v7Gioo4Y1shDhwieOhXLyZMkOzvfRqOqlbtpkzojnnBz7AqbFfwDkEVnELfMo8Fi4XpBAYuee87jIeOWLUOx2Sg/cICEVaugouxmopKxkfCVr6gfH1Oelwff+hbxe/eqi7Toqlc9lOhOw9eZuLi4DsWGlStXDuyCUjum9HQsZjPTHnrIq/pqCNd09dqZuxCwa7OnpxMYGMjatWtZa0+w1R8ciaYsOTnELVmClApK4WkMy7ou8uzoW9gYRHxiR6IpW1sbJbt2Mf/pp70+rmxtQV4ownjfl9SHrS98Qf30k6E8Xk1XrnB8wwZqs7MZk5TUpzZKtmzB/6c/JXrPniGj9DPUEEJ0XG/9ccCL7SpiOkOTofeI7SOCgoII6sdCGK9wSFU5z4CPiYDYeJSCkRmG4nexCJGaRt2lS4TExbmVGzSlp9NUXk6NhwVVPWFra6Nk9+4ebxymjAyqT5yguaqqT8fxFkNKGrLoLNKTNNnDD6sr8e+6C15/XV1w2Aecx2zFhx9gbbeS8Ln7e9+QlGq4yDPPqK9tHWm9Dx1CAKb6evU1clCQ6mz/4Q+q2kBODjz+eCfnG0AY/RDJ01AK1TFtyc0lND6eMamepQT9Q0OJXbwYS24uoC6IIyYOMdZzyIIvsJjNxC5e3GlcDsi9oBfU1dXx9ttvc//99xMREcFdd93F66+/3sX5TkpK4jvf+Q4HDhzAYrHwm9/8hjVr1gyK8w3q9VZ17Bgt9n56Y1dD6kzkxXPI1haaq6q4duKEJg6BKT0dS14eUlGQpRaor+0xK6VIndkxpiuPHMHa3ExCLxapyQuFYDCoiy99yFAbr86ETZhA+JQpWMzmPrfRsbZnEJzvoWxbV0wZGVR9+iktnmRLe6CpvJya/HxMbtSGfM1wsutQYsQ64IqieJTK8hkdM+CdbySGlDTV4RhhyKoKRE0VStI0ai9e9PhaKzQujogZM/p8k644dAiltbXHLHlRs2YRFBFBiV1VQSvE1Glgtapxpa4UFqoObEUFfPABPPIIxMRAejr86lfgIYzAHc5j1vLhP4gen0BQbFwPe9mx2dS41CeeUONU586F556Dixc71wsLw7RgARVhYbRdvKi+Mv/Sl6CHzGiG1JnIcwXI9nZV4s9Fm90diRkZWPLyUGw2dSFr6kzvzqWPKFYrJXl5mFziHQfkXtADVVVV/P73v+eOO+4gOjqaf/7nf+bdd9/tomYya9YsfvjDH3Ly5EmKiop48cUXWbRo0ZCIr4yaPZvAsWOx2K83b+wqJk8FgwF5vhBLXh5BUVFe69D3BlN6Os2VlVzLz1fDT+ITEWFjut3HkJoG1ZXI6gosZjMxt95KQC/ChpSCfMTUVISfbx+IhsJ47Q6T/c1jX7A2N1O2b1+Xa3SgGOq2dSZ6zhwCwsMp2bWrT/tb8vIIjokhMq37B1FfMJzsOpQY/Lu6RrS1tdHmtChHE6ydF2E6MEybCRVlyOvuXycPV5SCfJSIKNpGh6sLMF3iv51JtK/i7gsWs5mYhQsJ6GGxjzAY1EWF/ZiN8QYRFIyYlKSqJriSnd11m6JAXh587WswfrwqgfaTn6jOejc4xqyiKJQcP0nC0qXdd6ytTVUueewxiI9Xk7f8z/900uoG1DjLxx5TFRqqqhj33nv4hYZScuRID2d+E5GUCopC8/EjVB096tUsZsLq1bTduEHVnl1QXtrjjGR/qTp2jLa6OhJWruy0fUDuBW64evUqL7/8MqtWrSI2NpYvf/nL/OMf/+jUFyEES5Ys4Wc/+xkXLlzg+PHj/OAHP2DmzJlDbg2JwWjEtHp1x1sNb+wq/PzVEK7CU2pSrfT0fkvYuaMj0ZTZrMoPevGwJ6LVRFO2MyfVh8pezBRKmxVZdAZDiu/H9GCNV28xZWRQc/o0jWVlvd63bP9+gI5Y/YFmqNvWGYOfHwmrVmFxLJrvJRazuSMXgtYMJ7sOJUasA+7v76/5q1rpYQaccXEwJmLEhaHIglOQPB1/f3/qLl4kvJuFHab0dHUF9uXLvTuGonTMsHqDKSODik8+ob2hoVfH6S2O19XOqgmAOoPs4DvfURcRuWrPHz4M3/sepKbCjBnw9NPw6acdEmYOHGO2ZlcezY3NJH7uvq4daWiA996DBx5QlUscShuVlZ3rTZx4U62htBRefRXWr4eAAIwBASSsXNmrByQREIiYkkzJtq0EhIURPWdOj/sERUQQNWcOxVu2wJgIiIn3+nh9wWI2Ez13LkEus/kDcS9wcPbsWV544QXmz5/PhAkT+MY3vsGuXbs6zQ75+fmxbt06fve731FaWsq+ffv41re+xeRuHmiHCqaMDMr376e9sdFruxpSZ9KWf5KKgwc1jUc1ZWRg+fhjqK70+mHPkJrGtd07aa6s7J0DfvmCqrKSPL2v3fXIQI7XvhAxYwYhsbEdD2K9wWI2E7d0KX4DlanahaFuW1dMGRmU2a+33tBWW0vF4cMDFv893Ow6VBixKyAGJK2s1QoIcEm60ymN+eKV7vcdZsi6G8jSYgxr70babDSWlHQ7Ax42ZQqjJ05UF0k++qjXx6k+eZKW6mqvfwxjFi7EEBhIye7dTNywwevj9BZDygyUv/8NabmqqjsAlJTAwYPq/8PC1JCPgABVxu+jj9TZ8W3b1KyIDs6cUT/PP6+Gqowa1VHkiKArlAoRRj9Gbbiza0dKStSEL+5IS7uZ4GXWrG7l/kzp6Rz4/vextrR4rfpgSEnD8sZfiO9F/GZiRgaFr/6O2eu+q+mMrpQSi9lMyoMPdinT8l4gpeTIkSNkZ2eTnZ1NgQcFjJCQENavX8+mTZu44447GDOm+/CIoUrsokUY/P0p27uXOC+vUTF1GmVXSzAGBHTJG+BLHImm6pbMJjLau8x7InUmJb/9HZHTpxPSi2x9qspKEiLI947kUEmJ7gnnBYIpDzzg9X6K1UrJzp3M/Y//0LB33TPUbetK7OLFGIxGyvbtY3wvFiyX7N6NX3AwMQsWaNi7mww3uw4VRqzVHLGVmj6VtbeDv79bx0KkzkT5ZA+ysR4ROlRTK3iPUpAPo8NpHhNJXVERUlG6lTbquEnn5vbKAbeYzUTOmkVwdLRX9R2a05bcXE0dcDE6DGEar74FcDjgW7ferHDHHarzDRAaqjrBmzapYSI7d6qJO7ZsUWPFHVRUdPrbMYpKJk9mwvXr3i3mdJILZKr3i8Hili1D2lVKTKtXe7WPNXESFSUVLH3Y++yjCYsX8emLL1IbNAotl1/eKCigsbTU7YyPr+8F3soFjh07tiORzNq1azvk+4YzxsBA4pcvpzgnhzF2veye7CqCgim90Uhc8hSMGsqUhU+ezOix4ZQ1tHk91kRCIiVl1Uy8zfuYZFVlJR/D8tv61tEeGJDfrn5iysjg3Dvv0HrjBoFePkxWHT1Ke0ODqoo0SAwH2zrTKdFULxxwi9lMwsqV/U6G5y3Dza5DhREbgjIgWK0eZZScNXBHArIwH0PKDBCChitX8B89mqCoqG73SVyzhurjx71WKXHMYib2ctW2KT2d0t27OyXi0AKRmoZSkI90hI44x39nZrrfKSAA1q5V07WXlMDevfCtb6kx226oCwigNigIkyN9uisGA2RkwK9/rWpof/IJ/Md/9Mr5BieVkl6EoZQfO44wGBgX4r0TFdp4g/Cx4ZSc6T7+vb8U5+YyNjWVUQnaSH+2tLTwwQcf8OijjxIbG8vq1av51a9+1cX5joyM5PHHH2fHjh1UVFTwxhtvsHHjxhHhfDtwXG9Ke7tX9W2trZRdukr8mFFdQ7h8iKyvJSEmAsu5iz1XtlN36TL1tfUkjPWc6KfLcUqKob5OvR9+Rhk3bx7+oaG9WiBYbDYzbt48rx12HRVTRgYlu3Z5fb1ZW1oo3btXlx8cBoxYBzwwMJBADxrSPsPa7tkBd2jgFg7/OHDZ3IS8fAExbSaBgYG0lJQQNmlSjyEFkTNnEhQV1aGa0BN1Fy5Qf+VKr1fIJ6xYgdLaSoUjHEQjDKkzoaYaqivgxg11oSVAYKAaX90TRiMsXQo/+5nqPNfVqe3YP60VFVz+2mZGhY0irKysU1nHp6lJlQvcvLmLXGBvSVyzhpK8PBRP8oouWHJzibslDcPFwpsPIT2gFOSTMGcWJX2IF+0Nlm70bvt6L6ivr+edd97h/vvvJzo6mrvuuovXXnvNo1zg/v37qaysHHS5QK2JX7kSW0sLtSdPemXX8oMHUWwKsWNDVedVI5SC08QnTaamsIgm5zdN3WAxmxmdkMDo+hpki3fp1WVBPiJhPMKHCbKcGZDfrn5i8PdXFwh6+QDvmFwZLPUTB8PBtq4krFiBtamJSi8XzZcfOIC02YjraRG/DxmOdh0KjNgQlAHBau2igOKMIXUmtnffQLa2IAKHr0amPHcWAgIRE6aA1Ur95cteZdZyqJRYzGam3udmQaELHVm7vEzX7cB/1ChiFi3CYjYT3wsd394iIqMhOgblbD5GS+VNFZw1a6C36ZmF6LpPayulx45jmj8Xg0aZ4JxJWLWKQ88+S/Xx44ybP7/bukp7OyW7djHv61+DwiNqYp3YeJqamvj444/ZunUrl10X3EoFeeUS1rAx1Jw7z0+WLvWYWKk3mEwm7r77btavX8+oUaNoKC7mRlERi3/yk363XVVVxdatW8nOzmbHjh0eV/bPmjWLTZs2kZmZSVpa2pBTLNGSgNGjGXfrrZTu3EmsF2oWDm32gImT1RAu0wRN+iULTxG1dBnBh05RkpfH1M9/vue+5eZiWr8O2mqQ584iZs7tcR+lMB/DrO6vl88CpowM9n/3u1ibm3tcVHn97Fmaysu9DnfTuUlAeDgxCxZQbL+OesKSm0vckiX4h3r/VkdncBixDnhrayuApk9lspsZcHDSwL1QiJg+S7N+aI1SkI9Ino4wGmltbKTu4kUmeRlvbcrIYPfmzbTV1/coK1ick0NiH2dIEjMyOPnrXzP/mWcw9DKVdW8wpM5EFubDx/tubty40Sdt1xYVUlNWwdzvfMcn7fVEUEQE0XPnUpyT06MDXnnkCNamJky3b+BG6QW2/+plthScZ/v27TQ1NfVwpKvqP3YJMl/w5ptvEhQUxNq1a1kYHU3iuHGMSU52W7ene8HVq1c7FlHu2bPHrZ6tQy7QkY1yOCiWaEnimjWcfOUVpn/zm93eYxWbjZK8PG752tcQsREoxw9hXONmcXE/kS3NyEvnMX7hy5hWr6bYbO7RAW+qqODayZPM+973EOdPqhk7e3DAZXWFXWVFO037gfjt8gWOGday/ftJ7CHcoTgnh4gZMwj1EHo3UAwX27piysjg9P/+L/OfeqpbWUFHLoTZTzwxgL0bvnYdbEZsCMqAyOL0MAPeoYE7jJPyyPZ25PmCDr1bP6ORhitXulVAcSbm1lsxBgVRumdPt/Uay8qoOX26z1m7ElavpqWmhmsnT/Zpf28RqWnIK5eQH35o3yDUDJg+oPrvWwkMDiRqjTaLu9zhWCjbU0jJsS1bODR6NHfffz8J3/0hX3zhRbKysrxwvrWjpaWFrVu38v3f/56Hd+9mzZo1/OY3v6GkpKRTPXf3Am/kAo1GI2vXru2QC9y7d++wkQvUmoTVq2mtrqahqKjbetdOnKClpkbNfJiaBteqkFXehYf0BnnuLPgHICZOUaVJDx2izZ7S3RMleXkER0cTmZaGIcWeaMrafZytUpAPUeMQUeN82PvODBdJN7/gYOKWLPEqDKW3OutaMVxs64pp9WqaKyp6zC5dffw4bbW1A77QdbjadbAZsTPgAyVDKHo4jiF1Jra//w1psyKMw8/c8mIRKAoiKQWA9mvXsLW0eBWCAk4qJWZztyolltxcgmNiiOhj1q7g6GiiZs1StaC90KjuKyLOhCivRjh0WZcuVeUEfUDZ3n0kpM3AMIDjxJSRwacvvsiNggLGTpvWqezSpUtkZ2eTlZXF/n37cOeiGwwGli9fTmZmJitWrOi47qTNhvWNVzAuy8CQPJ1rp09z8JlnyHj9dQJ7kW3QFZvNxr59+8jOzmbnzp3YbDZ1u6KQm5tLbm4uX/3qV1m4cGHHbPWkSZM65AKzsrK8kgvMzMzkzjvvHLZygVoTMm4cETNnUr57NwndSAsWm81Ez55NsGPBdtQ4lIJ8jF7KBHqLUnAKkTwNYfRj3IIF+AUHU7J7N5Pu9DzbXmw230wMlJQCioK8eK5bbW9ZkK9J8h1nhpOkmyk9nU9fegnFavUoTVp/5Qq1585heumlAe5dV4aTbZ0JiY0lIi2NYrOZyJme374U5+S4zYWgNcPVroPNiLVai10rWdOnsh5CUMCexry9TV3EOCVFu75ohFJwCjElGRGgvlqqLipCGI2MMpm8bsOUkcHBZ57B1tbmUYbMMUPSn1haU0YG5999l9nf+pZmMblCCAzFTjN4Pgo/aa2qovJKMQvvu98n7XnLKJOJMSkpFOfmMiY1lfz8/I5wjOPHj7vdJyAggDVJk9h4z91s/OZ3iHYjGalcLMIWEY7fhrsQwSEo06dz7de/ZmxlJVOWLetXn2fPns3mzZu5du0ar/3nf/LXP/+ZU01NHdc8wMGDBzl48CBPPvkkqampNDQ0YLFY3LY3EuUCB4KY5cu5+sEHzPVwvTkW3jmHghhSZ6px4Mt9p9Agre3IcwUY71GP4/zQ78kBb6uro+LQIaY99BAAIjAIMXmqGobiwQGXdbXIkqsYbvegeOQjBuS3y0ckrF7NwR/8gKqjR4lZuNBtHUtuLqPGjyc8KWmAe9eV4WRbVxLXrOHS1q3M/sY33JZLKbHk5vZKm91XDGe7DiYjNgRFSum1UkOf6SEEBXpIYz7EkYoNWXi6U1a5hsuXCU1M7JW+aPzy5Sjt7ZR/8onb8tYbN6g8fLjP8d8OTBkZNBQXU3vhQr/a6RabDXHk2M2/feSAl2b9FaPRyNi1Xqip+BBFUaiZPJn/+sUvSE5O5pZbbuHZZ5/t4nwH+/lx33338fbbb6sLFX/+Ux5OnuDW+QaQZ08hJk5BBKvOrMFoJGH16l7JHvZEZGQkC2w2frV5M9XV1bz33nt84QtfICwsrFO9goKCLs53XFzciJcL1JqYFStovHqVuovuZf9qz52jobi4U+iBSE1DlhYj6274rB/y4rlOb+lAXRNStmcPVg9JqxyJSsY5JSpR13ecQbpZAwDq4ktGhyHivZ986AsD8tvlIwLHjCF63jyKu1E5Ks7JITEjY0gsVB5OtnXFlJ5O3cWL1F265Lb8RkEBjSUlgyI/OJztOpiMWAd8sGUInREpM+360dpp4GqBvHoZWpoRyTf1bpstFq/DTxz0pDlduns3fqGhjJs3rz/dJWzCBMKTkrDk5PSrnW7Zvx9RUwOAnJoEU6b4pFmL2UxcShIhA6B+0t7eTk5ODps3byYxMZEHf/5z/nb+POfPn+9ULyoqikceeYQfLFjA3pdf7pDkCwsLQ0xLQ169hGxs6NK+mqjkNMJloZopPb1PaZU9nkdjo5pIKD2d0NBQ7r33Xt566y2qqqrYvn07jz32GDFO4UEOucADBw5gsVhGvFyg1kQmJzN60iSP13Wx2cyY5GRGjx/fsU3Em2B0uBpL7SOUglOIyVM7KU3FLVuGYrNR4eGh32I2E79iRac3ciJlOjQ1Iosvu93HEX4ihLY/m8NN0i1xzRosZrNbB6y5qorqEyeGjCb1cLOtM+FTphDW3fWWm8uYlJRevZ32FcPZroPJiHXABwLZTSIeZwwpM6ChTlMNXC2QBfmI8ZMQoTfTpddfusToXsoEgup8leTlodhjdp0p9mHWLseiQs3YsqXjv/IWz7GivcHa1ERZ4TniVmkn0dXU1MSWLVv40pe+RExMDLfddhuvvPIKyn5eAgAAIABJREFUpaWlneqZTCa+9rWvkZeXR1lZGf/z9NOkNDYyZd26TvVE4iQIDkEWdV0UJEstUF/bJVbWOa2yLyjbuxeD0dhFmisgIIB169bx6quvUlJSwieffMKJEycoKirixRdfZNGiRRi6URLQ8Z74Vaso9uAQWOwx1s4IYcCQMkNVEvIBUlGQhWc6vaWDmw/97vpmbWmhbM+eLsodInQ0YvxENUTG9TjNTcjL5xGp2sZ/D0dM6ek0lZVx/ezZLmUleXkERUQQNWv4qoANJUzp6Z6vt36oiOkMDiP2V6i1tbVDGkczbFYw9uw0irBwRMJ4tzf2oYqUUp1ZcpnFrLt0ieA+PGF3qJScONFpu7W5mbK9e3uUsfKWxDVrqDl9mkYXx9InSNkp+6UtJgLZ3v/sm+XvZ6MokrG33+nzMbtnzx7uvfdeoqOjyczM5I9//CPXr1/vVCclJYWHV63i1xs2cPXqVV5++WVWrVqFn58fFrOZsEmTCHeZ6RcGAyJ5htuZTE+JSjrSKvsoDKXYbCZu+fJutcWNRiPTpk1j4sSJQ+IV+EiitbWV6GXLqMnPp6m8vFNZY2kp18+edZt4RUybibx8AdncfwUdWXwZmhoRbrJSJmZkuH3or/jkExSbjTg3axFE6szO2W4dx+lQWdE+jnlAfrt8SGh8PGOnT6fYzZvHYsfaniHywDvcbOuKac0arp08SVNlZaftjlwIg6U0M9ztOlgMjatCA/z8/LRfmWtt71EFxYEjjfmwoaIUaq93mllqq6uj9dq1Ls6YNwRHRRE9e3aX2enyAwdASp9l7Ro7fTohsbFeZ9/sFadOgT3+Tk4YD3HRyAvdy7B5g+Xj7YybNJ7gceN8Omb37NnDypUr3coFzps3j+eff54zZ85QUFDAT156iYirV2m1h9c4KO5GPsyQmoa8UIRs63zjVQryER6UInqbVtkTtrY2Snfv9urV9oDcCz6D+Pn5EZmWRnBMTJfr2pKbqzpmqald9hMTpkBAoOrU9hP1Ld1ERGjXHAMJq1bRVltLtct6BktuLrGLF7tNVGJITYMbNVBZ1mm7UngzF4LWDMfx6u7NY3tDAxWffDJkwk9geNrWmci0NIKjoylx+X2z5OYSmpDAmJTBEXoY7nYdLEasAz4UdMCdMaTO1EwDVwuUgnyIjUeMuSlnVGfPdDi2j6vZTWvWUJyT02l2yWI2E7tkCX4+WvwmhMCUkaFNHLjT7LfYmIlI6r/Gu629nZKTp0lcscKnY7a1tZXHHnusw9YGg4GVK1fyi1/8gsuXL3PkyBGeeuopptmlB6PmzCEgPJySnTs72miqrOTaiRMe00eLyclqoqnzNyX9ZHUlVFdgmObeAU9YsQJrczMVhw/36/wqDx/G2txMwooVPdbVNWq1wd/fn4DAQLevxYtzcjyqGgmjEZE8vd/Xjqe3dA6CIiOJmjOn08ysYrNhycvz+MZNjI2EmDiUszf7JttVlRUtk+84MxzHa+KaNdSeO0f9lSsd20r37MEQEOBRHWUwGI62dcaRXdr1bUOx2YxpEBe6Dne7DhYj1gEfkFciXi7CBNTEDXYN3OGAUnCqSwxv3cWLBEVHo/TxQjOlp6sqJfbFforVSsnOnT6fITGlp1N59CitN3yntAB0iv9m40Z1BrjoDFLpGtfuLdU5H9Pa0krC5+736Zj96U9/2qF1HRcXR3FxMTt37uTrX/86EyZ0TQXuUClxdqScE5W4Q/j72x9Cbo5ppSAfIqMRUe51nh1plfsbhmIxm4m59VYCXBRP3KG/HtUGh10TMzKoPHyYttpadfuNG1QdPdrtdW1ISUOeL0T2501IZRncqOkS/+2Ma6Kp6uPHabtxo9tEJYaUNFXxxI681FVlRUuG43gNT0piVGJip1lwdwtdB5vhaFtXXBNNtVy7RvWxYz4L4+wLI8Gug8GIdcAVRXGbUtqn9GIGHJzSmA9x5PVrUFGGYZpL/PfFi4SOH99nu44eP54xyckdT+9Vn35KW12dz7N2jZs/H//Q0E6zuf3m0iVwvMqOjIRly1R1mJZm5FX3slDeYPlgG1HxsYRMmOizMVtYWMjzzz/f8fcvf/lL4r1IAZ2YkUH5gQMdKiXFOTkk9BC/2fEQYrMCIAtO9ThTaMrIwJKX51HurSekomDJzfX6B2dA7gWfQRx2HTd/Pn6hoZTs2gWoD24BYWFEz/Wc1l0kpYBU1ERffT3+2VMQE6fOWnsgMSODRouFG4WFgDqmo+bMISjS8z6GaTOhvFS9D9qP45wLQWuG43gVQpBof8MJaohYye7dg+oUumM42vb/Z++9w+Oqrv3998yMRs2WXOUmmxoM2JhubExxoYXgBJu0m1BC4AbITYgTE7gk+AcJCS0EQk0CfEmA3ISbggktEJCNAdMMhOIGXOyAG66SJdvSFM35/TFFXR6dOUuSR5/3eXiYOTNnz553DmLNmrXXbk16o6n1L7wAwLrnnydcXs4Qww3odkc+eO0J8jYA7542hNl1QUnjHDged90nuLXbDSeVO4mVS2HAIKgY0eJ47erVlO+7b05em9cKrq2qYuiRR1Lkc+u9QCjEqKlT/e2G8ve/N92eORNCIZySUpy99sVd4e1LVSKRYO1b/2LUsZMAf65Z13W5+OKLiUaTi0PPOOMMzjrrrKzOzXQpeemlzEYlu/sfaPONpty65EYlu+sUUTl9OvUbN7J1qTdvW5cupX7zZkZNy65rjFpk2ZD2GigoYNSJJ2Z+1Vi7YEFy6/lO/jY64UKcfQ/IqQwl8f7ud6XsN3o0Aw44IJMFz2pL9GEjoXwgifeXJbusfLCs0yy73+yp12vl9Olseecd6jdvZuNrr5GIRBh5/PE9Pa0W7Klum5PZaCr1/7c1VVWMmjaNQDesT+iIfPDaE+RtAN4duF0oQYFmPXB7eRbcXbmUwIHj29ST1a5eTf8u9gBvTeVJJ1G9fDk716/P1K1ZUHnSSWxYvJh4fb0/AzYvP5nVtBOec+AhJN5v2zUhG7YveZWdtTsYfdaX/JghAA888ADPpzL/paWl3HXXXVnXBQYLCxlx/PGsqapK9mZvtVFJezjFJTh7JzeaSqxcBv3KcEaN7vSckmHDGHzIIZ6/IK2tqmLwhAmUDPN3O3PhndEzZrB+8WIaqqvZsHhxVv9dJ38RXOaphMut2Qafrm/zK117pNee1Lz/PjvXrt39l0rHyezY6a5ZDfW7WuyFINpn8KGHUjRoEOsWLkyWiE2aREG/frs/UXSZ0TNmsP7FF2morubTV17pdb80iOzI2wC8e2rA4zhdKEHJ9MDtxe0I3Z11uJ+sbrOwKRGLUbdmDUWjRuXkdeCBB1IyYgTv3XUXuzZsMGubNOLYYwHY8PLLuQ+2eTO8+GLydkkJnHxy5qHA2HGwvRo+XdflYdc88ghlgwZQNuEwIPdrdvPmzcydOzdz/9prr2VMs01QsqFyxgzWv/ACnzzzTNb1m86ByZpZd+W7qS9uu/+zUjljhuc68Pb6S3eG6hNtaO51xJQpkEjw9s03QyDA8NR/f53hHHBwqoTr311+7cTKpVA+MJmt3g2jp0+n5v33Wfnggww44AD6je78CyKkfq38ZDWJt15rsxeCNXvq9ZpZR/Lcc10qEetO9lS3rRlx3HEk4nHevvlmAsEgwyZN6tH55IvX7iZvA/BgMEjQ+ieZLmbAIfWH3aceuBa47y+HklKc0Xu3OF73ySe48Tjl++6bk9d0l5JVjz7KwAMPpN+oUTnOuH1CxcWMOPZYf3pOP/EEpOvbTj0VioszDzkDBsHwUZ4W1659/XUqjz4qcz/Xa3bu3LlsS7URPOKII/jud7/b5TFGHX888fr6Lv0PNDB2PNTV4q76MOuNStLbKm/vYBvzjtie2oq5K7+cdMvfgj5Ic6+hkhKGT57MqkcfZcSUKYSKinZzNjil/XDG7IPr4b+djn6la48BBx5I6ciRrP7737O+bpwxe0NRMe67b3bYUtOKPfl6rZwxgw2LF9OwbVvWJWLdyZ7stjnpjaZWPfooI447Lqv/3izJF6/dTd42bgx3x8rrLi7CBHD23h8KwjQ+9yTOUNuf0Dd/+H/0Hz6Mov5te+R2hPveWzhjD26z8K529WpCJSWUVVbm3Opo9IwZfPCHP5j3h62cMYO3bryRlQ8+mNtAv/sdDEq1Y6yshFbjJdZvxV3yR5yXXsl6yEQkQs2mrRxz5pmZY7lcs8899xwPPfQQkGw5eO+993rqyxouL2fYxIlseuONdjcqaY/MRlNbNuHsnV2P+PS2yu/deSdDDjss6/ltefttyvbdl/J99836nG75W9AHae218qSTWPf8813KfDoHHkLi5edhQBfWgSQSuJ+sIjD11N0+FZq+9L//0ENZz80JBHHGjsN9e0m31n/Dnn29Dp80iVBJCQMOOIDioUN7ejpt2JPdtmb0jBmsX7Sowzax3Uk+ee1O8jYATy9CM+1N2cVFmJDsgRuYPJXEindx131iNLEkrz/0V0bsPZrDju9CH1bHIXh42+fXrl5N2T77EEu1DcvF69AjjmD0SSexz8yZnsfIhspp0/job39jVfP67a6SSMD69TBgQPL+xx/D2rUtn9PYCDXbcD/sWjeUUWP3Z+BxJ2bue71m6+vrufjiizP3v/e973FEJx0odscBX/sag8ePb3ejko4ITDoBt3orTjD7/x4OPPdcPnj44Ux/+azPO+ecLj2/W/4W9EFae62cNo2106Z1qatRYNyhuEv/ReLtrvWFd/Ybm8xSZ8l+s2fTsHUrA9rZGKjDuR05mYQT6LTLigV78vUaDIc56JvfpCzHtUJW7MluW5PuJjXqxBN3/2Rj8slrd+J4WTxmheM445YuXbp03LjcF7zUpnpklmXRJ9grsZ/+kODXLySwX8/sPrU7/jJ5MuGyMj7/9NM5Z61fufJKXNdl/I9+BNh67VX87W/wxS8mb8+YARYb/KTwes3++Mc/5rrrrgNg9OjRLF++nH5a/JShO/4W9EXk1QZ5tUNubZDX7Fm2bBnjx48f77rusrzNgFu3xHETjeAmoJd+40vEYsRqa4nV1lLz/vvtbgndFWpXr2bUtGl9r9VQq813LPHidunSpdx0002Z+3fddZeC71b0uWu2m5BXG+TVDrm1QV69kbeLMM2JJzce6WoNeHcRSe1KFyotzbkftuu6bF+1qtf+rGhGLJZcgJnmC1/oubm0QyKR4KKLLiKeuhbPOussZhqX9QghhBAid/I2ADdvixNP1kI7Hha6dQfpbdj3PuOMnDuB1G/eTHznTsr23bdvtRtatAjS29kfdRRk0b4sF7rq9t577+XlVJvFsrIybr/9dqup7dH0qWu2G5FXG+TVDrm1QV69kbcBeCAQINDJFto509sz4NXVQHLxUfXKlexovXCwC9SuWoUTDNJ/zBh7r72Jbiw/ga5dsxs2bOCKK67I3L/++uuz2m6+L9KnrtluRF5tkFc75NYGefVG3hoz3xo1E4D30gx4dTXhsjIGjRtH6ciROZWh1K5eTb/KSoLhcN/ZcjaR6HD3Syu64nbOnDlsT5UZTZo0qUUXFNGSPnPNdjPyaoO82iG3NsirN/I2AI/FYpmWeSakSlB6cwBeOHBgpgduLmUo6RaE0A1eewtvvgnrUrtbfuYzcNBB5i+ZrdunnnqKP//5zwCEQiHuueceZR86oc9cs92MvNogr3bIrQ3y6o28/b92PB7PLE6zwN0DSlAKByY3t6icMYPNb71FQ2qXxK5S22wBprXXXsP8+U23Z82CHNs4ZkM2bnfu3Mm3v/3tzP3LLruMQw45xHpqezR95prtZuTVBnm1Q25tkFdv5G0Abl+CsgdkwFObxww9/HDC5eWse/55T2PVrl5N2X7JHQ77zE9N3Vz/Ddm5vfrqq/n4448B2GeffZg3b153TG2Pps9cs92MvNogr3bIrQ3y6o28DcDNicfBCeAEgj09k3aJ1NRkMuCBUIhRU6d6KkOJ7dzJrk8/7VstCN9/H1asSN4ePhyO6cJOoob861//4le/+lXm/m9+8xtKSkp6cEZCCCGE8ELeBuD2bQi7vg19d9I8Aw7JMpQNL79MbOfOLo1Tl9omvGzvvZPj9oV2Q82z31/4AnRTfXVHbtesWcMdd9zBl7/8ZRobGwH42te+ximnnNIt89rT6RPXbA8grzbIqx1ya4O8eqP3RpA5kuvW67slHuvdAXizDDjA8MmTcYJBNixezJguBG7bV6+maPDgTDBv7rU30APlJ9DS7cqVK5k/fz7z589nyZIlLZ43cOBAbr311m6b155On7hmewB5tUFe7ZBbG+TVG703gsyRoqIi2xeIx3vtAkxouQgTIFRUxMgpU1i7YEGXAvDajz5qUX5i7rWnWb8eXn01ebusDKZP75aXdV2X5cuXM3/+fB577DFWpEtgWlFaWsof/vAHKioqumVe+UDeX7M9hLzaIK92yK0N8uqNvA3A0ytyCwqMguTengFvFYADVJ50Em9cdx2JWIxAll5qV6+mbN99M/fNvfY0jz3WdPv00yEcNnupxsZGXnrppUym+5NPPmn3eQMGDGDmzJnMmjWLU089VXXfXSTvr9keQl5tkFc75NYGefVG740gcyTdk7K4uNhkfLcXZ8DjDQ3E6+vbBOCjTjiBV6+6ik1vvMHwyZOzGqt29Wr2O+KIzH1rr75RVwd33QWPPw5dqU1LdRgBTDbfiUQiVFVV8cgjj/DYY4+xefPmdp83fPhwzjzzTGbPns3UqVP1hy0H9phrdg9DXm2QVzvk1gZ59UbeBuDmLXHiMZxemgFPb0PfOgAPl5cz7OijWfPcc1kF4Il4nLqPP25RgtLrWw1t3w533AG33goe+54Dycz3aaf5MqW6ujr+8Y9/MH/+fJ588knq6urafd4+++zDmWeeyRe/+EUmTZqkzXV8otdfs3so8mqDvNohtzbIqzd6ZwS5J9CLu6BEamoAKGrWBSVN5fTpLLvvPo768Y9xdhPg7Vi3jkQs1qIEpddSUwO3354MvFPvPyfmzk3WgHtky5YtPPbYY8yfP59nn322wxXiEyZMYNasWcyePZsDDjgAx3H0x0wIIYTIc3pnBOkD6YDHLJjpxSUokepqnGCQgv792zxWOX06b/z852xdupQhEyZ0Ok7tqlUEi4ooHTGiaWxrr12luhp+9Su47bZk9rs5EyfCvHnQ1Z0iS0thyJAuT2XNmjU8+uijPPLII7zwwgskEol2nzd58mRmz57NrFmz2C+1wRFAbW0t0Ivc5gm97prNE+TVBnm1Q25tkFdv5G0Abk4vXoSZ7gHeXoa7ZPhwBo0fz9oFC3YfgK9eTdnee+82U94jbNuWzHbffjukAtcMkybB1VfDqaeabyHvui4PPPAAd999d5t2gWlCoRDTpk1j1qxZfOELX2DkyJGmcxJCCCFE76Z3RpA+YL4YoDdnwGtqWmzC05rRM2aw+vHHOWzOnE7HqV29us0OmD2+yGLLFrjllmSd944dLR+bMiUZeJ90knngDfDBBx/wrW99i0WLFrV5rLi4mFNPPZXZs2dzxhlnMLBVPX579LjbPEVebZBXG+TVDrm1QV69kbcBuHVbHLe3Z8A7CfgqZ8zgndtuazfAbk7tqlWMmDKlxbEeaze0dSv84hdw553QejfP449PBt7Tp3dL4B2NRvnFL37Btdde26K2u7y8vEW7wNLS0i6Nq1ZONsirDfJqg7zaIbc2yKs3emFtgT/EYrFMaxwTevMizFbb0LembN996b/33qytqurwOa7rUrtqVZsA3dxr24nA738PY8fCjTe2DL6nToWFC+GFF2DGjG4Jvl977TWOPPJIrrrqqkzwXVpayq233sqmTZt46KGHmD17dpeDb+gBt30EebVBXm2QVzvk1gZ59UbeBuDhcJiw4SYqyTaEvfPbXqSmhsJBgzp83HEcRs+YwZpOAvDItm1Ea2vbdEAx99qcFStg2jQ4//xkBjzNjBmwaFEy+J46tVumUldXx6WXXsrkyZNZunRp5vjpp5/OsmXLmDNnTs5eutVtH0JebZBXG+TVDrm1QV69kbcBeCAQsO2j3JtrwHeTAYdkGcrWd99l16ZN7T5eu3o1OA7999qrxXFzrwD19XDVVXDooclAO81hhyWz3c89ByecYDuHZjzxxBOMGzeOO+64A9d1AaioqOBPf/oTTzzxBHu1cuSVbnHbB5FXG+TVBnm1Q25tkFdv5K2xhoYGGhoa7F5gD64BBxh8yCEUDx3KugUL2n28dtUqSkeNIlRU1OK4udenn4bx4+HnP4f0T1r9+iU7nixZkqz37iY+/fRTvvKVrzBz5kzWrFmTOX7++eezYsUKvvrVr+L4WPZi7raPIq82yKsN8mqH3Nogr97I2wDcnN6eAd9NAO4EAoyaPr3DMpTtu1mg6Tvr18NXvgKf/SysWtV0fPbsZCnKnDnd9oXHdV3uu+8+DjroIP785z9nju+///5UVVVx//33M6iTEh8hhBBCiM7onSlcHyhqlbn1Gzce75Vb0buum1UJCiQ35Vn0X/9FtLaWcKtdH2tXraK82SYxaXz32tgIv/41/PjHLft577VXsuPJGWd4GjYSifDaa6/xzjvvZFZoZ8tjjz3G888/n7kfCoX44Q9/yLx580zbLVlfs30VebVBXm2QVzvk1gZ59UbviyB9oqNdCH2jl5agxHfuJBGP7zYDDjBs4kRCRUWsf+EF9m4V6NauXs3ok05qc46vXt98Ey66KPnvNKFQchv4efOSO1JmSUNDA6+++irPP/88ixYt4tVXX/XlJ7Gjjz6a++67jwm72bTID8yv2T6KvNogrzbIqx1ya4O8eqP3RZA+EY1GAcNvZvF4r+yCEqmuBsgqAA+Gw4w84QTWLljQIgCP19ezc/36Nh1QwCevdXXJRZZ33gnN/8OdMiWZDc9i6/hdu3bxyiuvsGjRokzAnZ6bH5SWlvLzn/+c73znOwSDQd/G7Qzza7aPIq82yKsN8mqH3Nogr97I2wDcvCF8L82AN6QD8CxKUABGn3QSr151FY2RCMHCQgDqPv4YXLfdGvCcva5YAbNmwfvvNx0bNAhuuinZbrCDldSNjY0sXLiQhQsXsmjRIl5//fVO+44WFBRw9NFHc9xxx1FeXt6lKZaWljJ79mxGjx7dpfNyRZsY2CCvNsirDfJqh9zaIK/e6H0RpE+ErIPjXroIM1JTQyAcJlRSktXzRxx3HIl4nE9ffZVRJ54IJMtPwuXl7WbRc/L66KNwzjktt5A/77zkDpdDh3Z42sqVKzn77LN5s3mpSivC4TDHHHMMJ554IlOnTmXy5MmUZOmgt2B+zfZR5NUGebVBXu2QWxvk1Rt5a62+vh4w/GbWS3fCjFRXUzRoUNbt8QpKSxk+eTJrq6oyAfj2Vaso33ffdsfw5LWxEa65Bn72s6ZjQ4bAH/8IJ5/c4Wmu6/Lb3/6WH/zgB5nXTVNYWMjkyZM58cQTOfHEE5k0aZLpAsnuwPya7aPIqw3yaoO82iG3NsirN3pfBLkH4Lpury1BybYDSnMqp0/nndtu4+jGRgLBILWrV7db/+2J6mr4+tfhH/9oOnbkkfDIIzBmTIenbd68mQsuuIDHH388c6y8vJw5c+Ywffp0Jk6cqHozIYQQQuyR9L4I0icKU/XMJiQS4Lq9swQlix7gramcNo3Xr7mGLW+/TcWRR1K7ahX7zJzZ7nO75HXpUjjzTPjoo6Zj552XXGjZSbb6H//4B+effz4bN27MHDvxxBN58MEHGdNJ0L6nY3rN9mHk1QZ5tUFe7ZBbG+TVG9qIxwvx5OK/3tgH3EsGvGjwYIYefjhrq6pwEwnq/v3v3Dfh+ctfYNKkpuA7FEp2Pfnd7zoMvuvr6/nud7/L6aefngm+CwoKuOGGG6iqqsrr4FsIIYQQfYe8DcAjkQiRSMRm8PTGLr0xA15T0+UMOEDljBmsqapi5/r1NEYiHZag7NZrYyNccQV8+cuwc2fyWEUFLFgA//Vf0EFt+ttvv81RRx3FnXfemTk2duxYXn31Va644opuawXYk5hes30YebVBXm2QVzvk1gZ59UbeBuAFBQWGCzBT7e96awbcSwA+fTo7167lk6efJlBQQOmoUe0+r1OvW7cmt5K/6aamY8ccA2+9Bccf3+4piUSCm2++mYkTJ7J8+fLM8UsuuYS33nqLI444osvvZU/F9Jrtw8irDfJqg7zaIbc2yKs3TCNIx3GCwAPAKGA18C3gIuC81FNudF33bxavbdoWp7dnwLtYggLQf8wYBhxwACsffJD+e+9NoIOMc4de33472d/73/9uOnbhhcmykw7qw9auXct5553HggULMseGDh3K/fffzxket6Dfk1ErJxvk1QZ5tUFe7ZBbG+TVG9YZ8FnAatd1pwErgdnAt4FjganAj6xeuKGhwZetyNslE4D3vovOawYckmUoDVu3dlr/3a7XP/4Rjj22KfguKIDf/hbuvbfD4Psvf/kLEyZMaBF8n3766bz33nt9MvgG42u2DyOvNsirDfJqh9zaIK/esI4g9wPeTt1+CzgTWAUUAyVATavn99+wYQOjR4/OrKqNRCIUFBQQCoVoaGjAdd0Wj4VCIQoKCohEIiQSicxjO3bsIBAIUFRURDQapbGxscV5gUCAwsJCYrEY8Xi8xWOO41BUVEQ8HicWi7V4DKC4Idnzsq6+nkTNdnasXp3ZijUcDuMmEsTicUKhEMFgMHNe+ieaWCzGkPHjKR4woEvvKRKJEAwGCYfD7b4nx3WJbN8OxcXU1tZ27T0VFzM0tRV8yejRRKPRzH9QRUVFJBIJotFoZg61tbUAlNx8M6Frr818gO6IEex68EEaJ06kMDV2+j2FQiGefvppbrvtthaBd1FRETfddBPnn38+iUQiM6ecP6fi4sxj4XCYQCDQ7ntKX1/pXqa5Xnu7+5w6ek9cFzo8AAAgAElEQVQ7UhsU5dN76g2fU319PQUFBezYsSNv3lNv+Jyi0SjRaBTXdfPmPfWGz2nHjh2ZeefLe+otn1MsFiMSieTVe+oNn1N7/+/a09+T1ee0o9lGhNYB+HLgNOBvwEnAQOBJYAUQBC5o9fxzHn74YSZOnJjzCxcUFNgt3EtnwIMh3rzmR2xcvLjLQ+x/zjkcPneur9OK7dgBiQRhjxnw8rFjGTBuHEMOO6zD54TDYYLBIIlEgvAdd7QIvhsnTSLxv/9LYuDAZJvG9LxiMf7yl79w++238+6777YYb8KECdx///1MmDAh8yWmr1JQUJD1Bkoie8LhsOoTDQiHw8k9EYSvFBQUEAjk7fKsHkXXrA36++oNx/JidJLRxM3A4cAyIA5MByYBYWABcJSbmoTjOJNefvnlVyZPnpzza6e/vVj0p0x89D6Nf7wPd87/xyPHHccJd93FyOOOy/r813/yE6Lbt3PcLbf4Oq/a1at54owzOHPhQkoqKnwdO03G64MPwre+1fTAOefAffdBONw0n9pa7r33Xn71q1+xdu3aFuOUlJQwZ84crr76asLNzunLWF6zfRl5tUFebZBXO+TWBnnNnmXLljF+/PjxrusuM82ApwLruQCO41xDMuCeDDQAMZJBuAOkvwXUlZWV+fLaphdEPA6hAta/8ALBoiKGdTFjXzx0KDUffOD7tBq2bQPwtAgzWyKRCKG//hUuuqjp4Be/CPffn6mJX7t2Lbfddhv33HNPplQlTUVFBZdeeikXX3wxgwcPNpvnnoj+iNkgrzbIqw3yaofc2iCv3rDugjIc+BOQAKpc133BcZxHgFdILgC9y3XdhMVrm3dBCYVYW1XFyBNOINjFDG7JsGHUb9rk+7QiNTWESku7PJ+uEH72WQovuqipxOS00+B//gdCId555x1++ctf8qc//Yl4ukwnxdixY7nssss4++yztYV8B2gluQ3yaoO82iCvdsitDfLqDesM+KfAtFbHbgJuav8M/zCtSYrHaMRh/YsvckyzGuhsKa6ooH7TJlzX9bXmN5cOKFmxcCGFX/86Tjq4Pv543L/+lecWLeLmm2/mn//8Z5tTTjjhBC677DI+97nPqa5xN6iOzgZ5tUFebZBXO+TWBnn1Rt5+bWndecRX4jE2bdxKIhZjZAcbzHRGybBhJOJxItXVFA0a5Nu0vGxDnzWvvQaf/zxOererI4/EffxxvnP55dx9990tnhoIBDjrrLOYO3cuxxxzjM188hDTa7YPI682yKsN8mqH3Nogr97I2wA8kTCpbAHAjcdZ9/E6hk+eTEFpaZfPL04tkKzfuNHfANzjNvS75d13kztcptrnNB54IMGnn+bnd9zRIvguKSnhm9/8JnPmzGG//fbzfx55juU125eRVxvk1QZ5tUNubZBXb+RtAG65GCARjbD+32s49Ctf93R+uLycYGEhuzZtYuBBB/k2L5MSlA8/hFNOgepqANx99iH+1FP8z1NPMW/evMzTLr74Yn72s59pYWUOaAGLDfJqg7zaIK92yK0N8uqNvA3ALdn6wf8R2VXPqKlTPZ3vOE6yDnzjRl/nFampof9ee/k34Jo1cNJJkJ7nyJFEn3qKqhUruOCCphbuX/va17jrrrtU4y2EEEIIkQV5GzFFIpFMXZLfrHv7HYZUjqB4yBDPYxRXVLBr82YfZ4W/NeUbNyaD708+Sd4fMgSefZY3t23jq1/9aqbLydSpU7n//vsVfPuA5TXbl5FXG+TVBnm1Q25tkFdv5G0G3GoXTNd1WfvOUvY/fHxO45RYZMD9WoRZXQ2nngrpXuVlZfDMM6wtK+NLJ59MXV0dAOPGjWP+/Pn6+cknzHZu7ePIqw3yaoO82iG3NsirN/I2ALfaXXH7hx+yc+s2Rh18YE7jFA8bxvaPPvJpVkl8qQHfsQM+9zl4553k/eJiePJJtu+3H5897jjWr18PwMiRI3nqqacYYLjpT19DO4LaIK82yKsN8mqH3Nogr97I2wA8Go0C/rfFWfPcc5QPr6BfxVBP5ycSCV566SWqIxEKfMyAN0ajxHbsyC0Ab2iAM8+EV15J3i8ogPnziU6cyKzTTmPp0qUA9O/fn6eeeooxY8b4MHORxuqa7evIqw3yaoO82iG3NsirN/I2AG9sbDQZd+2CBYwa+xkIde1Cc12XZ599liuvvJK33nqLgONw7fjxnO7TvKLbtwM5bEO/bRucdRY8/3zyfiAADz+Me8opXHDuuSxcuBBI7nj14IMPcuihh/owa9Ecq2u2ryOvNsirDfJqh9zaIK/eyNuVc4WFhb7XJu9cv57qFSsYdcB+OF3YevW1115jxowZnHrqqbz11lsAJFyX/121ikafFi5EUm0CPWXAP/wQJk9uCr4B7r8fZs/mqquu4g9/+EPm8N13381nP/vZHGcr2sPimhXyaoW82iCvdsitDfLqjbwNwC1YU1VF6ciRDBgyELIIwJctW8asWbOYNGlSJoPcnHd37uSVBQt8mVukuhoch3BZWddOXLQIJk1qWnBZUAAPPgjnncc999zDddddl3nqT3/6U84991xf5iuEEEII0VfJ2wDcoi3O2qoqKqdPx2ls7LQE5eOPP+b8889nwoQJPProo5nj/fr145prruHCCy/MHPvFLbf4MrdITQ3hsjICXcjM8/vfw8knJ8tPAAYNgueeg3PO4YknnuCSSy7JPPXCCy/kqquuUrshQ+TWBnm1QV5tkFc75NYGefVG3gbggUDA197UDdXVbH7zTSpnzIB4rN0M+KZNm5gzZw4HHHAAv//97zPbs4bDYebMmcOqVau4+uqrmTdvHqHU+Y9XVfHhhx/mPL8udUBJJOBHP4Lzz4dYLHnsgAPgtdfghBNYsmQJX/nKVzLzP+2007j77rtxHMd3r6IJubVBXm2QVxvk1Q65tUFevZG3xvyuSVr//POEy8oYesQRuPF4iwx4bW0t11xzDfvttx+33XZbZkVwIBDgG9/4Bh988AG33norQ4cmO6eMGTOGs88+G0guzrzppptynl9DdTVF2QTgu3bBV74C11/fdGzaNHj1Vdh/f1atWsUZZ5zBrl27ADj88MP585//nFndrFovO+TWBnm1QV5tkFc75NYGefVG3gbgsViMWDq76wNrqqoYNXVqssQjHodQCNd1ufPOO9lvv/34yU9+wo4dOzLPnzVrFu+99x6/+93v2Kud7eEvv/xynNTtBx54gHXr1uU0v0hNze4z4Bs2wNSp8Ne/Nh274AJ4+mncAQN49913+exnP8umTZsA2GuvvXjyySfp379/5ul+exVNyK0N8mqDvNogr3bIrQ3y6o28DcDj8Xhmu/Scx9q1i09ffjlZfgIQj+GEQtxyyy1897vfZcuWLZnnTp06lVdeeYVHHnmEgw8+uMMxDzroIE4YOxZIXry33nprTnOMbNvWeQD+7rtwzDGwZEnyvuOQuOEGXvnmN7n8qqv4zGc+w6GHHsoHqcWYAwcO5B//+AcjRoxoMYyfXkVL5NYGebVBXm2QVzvk1gZ59UbeBuB+/iSyYfFiCAQYfuyxyQPxOKs/3cS8efMyzzniiCN45plnWLBgAZMmTcpq3P+cOTNz+ze/+Q3b0oshPRCpqem4B/iTT8KUKbBmDTHguXCYb59yCpW33caxU6bwi1/8go+a7cpZUFDAo48+ykEHHdRmKP3UZIfc2iCvNsirDfJqh9zaIK/eyNsA3E/WLFjAiClTCBUVAeDGovzXdTdSX18PwIwZM1iyZAmnnHIKjuN0NlQLjpk4kQmDBgGwc+dO7rzzTs9zbHcRpuvC7beza+ZMHt2xg/OAYY7DydEov37mGTZs2NDi6ZWVlVx66aW8/fbbnHDCCZ7nIoQQQgghOiZvd8JMt8TJ9VtZIhZj3fPPc9SVV2aO/flf7/HMS4sBKCoq4je/+Y2nFcDFw4ZxxsCBvJvKfN9+++3MnTuX0tLSLo8Vqa5ukQF3YzH+esYZPPzPf/I0sCvzgNvivAMPPJBZs2Yxe/ZsjjzyyN1+gfDLq2iL3NogrzbIqw3yaofc2iCv3sjbALwrmejO2PTmm8R37mRkKiO8bds25j5ZlXl83rx57L///p7GLhk6lIPDYY447DDeevtttm7dyn333cf3vve9Lo/VehHmtw89lN+sWNHuc4866ihmzZrFrFmz2i0z6Qy/vIq2yK0N8mqDvNogr3bIrQ3y6o28DcCLUuUiubK2qoqKo4/OZJevuOJyNu1M5pPHjRvHZZdd5nns4mHDcByHS88/n2+kgu6bb76ZSy65hHA4nPU48fp6GhsaMgH4feed1yL4DjgOxx9/PLPPOoszzzyTMWPGeJ6zX15FW+TWBnm1QV5tkFc75NYGefVG3taA+7Eq13Vd1i5YQOX06QC89NJL3Hff/8s8/tvf/rZLgXJrCgcOJFBQwIxDD2VsqiPK2rVr+eMf/9ilcSLV1cnxBgxgyQMP8F8PPph57OJDD+XTjRt5ftEiLr300pyCb9BqZ0vk1gZ5tUFebZBXO+TWBnn1Rt4G4H70pdy2bBm7Pv2UyunTiUajXHTRRZnHvnXuOUyZMiWn8R3HobiigsiWLVx++eWZ4zfeeGNmF8psiNTUAFC7eTNnXXAB0dTxaQMGcMerr2Y2APID9fu0Q25tkFcb5NUGebVDbm2QV2/kbQDuR1uctVVVDBo3jtIRI7jppptYvnw5AMP7lXLDT67xYZZQUlHBro0bOfvss6msrARg5cqV/P3vf896jEh1NYlgkPNOOYU1jY0AVAaDPPzSS5nOLX6hdkN2yK0N8mqDvNogr3bIrQ3y6o28DcD9YG1VFaNPOokPPviAn/3sZ5njvzxtKgMGD/HlNYorKqjftIlwOMzcuXMzx6+//nrcVh1LOqKhupq/bdxIVaoUJQz87d57qRg3zpc5CiGEEEII/8jbADwSiWRa43ih9t//ZvtHHzFq2jQuvvjizFinz5jOF8cdACF/1q8WV1SwK7X1+4UXXsigVF/wJUuWsGDBgqzGePzee3k0NQbAHV//OhPPP9+X+bUmV6+iY+TWBnm1QV5tkFc75NYGefVG3gbgubK2qor+e+3Fo4sXs3DhQgBKSkq449qfJFvuhAp8eZ2SYcOoTwXP/fr149JLL808dsMNN+z2/JVPPcWVjz+euf/Nz3yG/2y2CFMIIYQQQvQu8jYALy4upri42PP5a6qqKJ00qUWbwZ/+9KfsPWJE8k4omOsUgaYSlDTf+c53MhvxPPfcc7zxxhsdnlu3fj2zZs1iV6pU5ciSEu56/XUcD5sCZT3fHL2KjpFbG+TVBnm1QV7tkFsb5NUbeRuA59IWp37zZra+8w6/XbKErVu3AnD44YcnN8iJxyAYxHH8UVcybBgNW7fSGE32Lhk8eDDf+ta3Mo9ff/317Z7nJhKcP3EiK1PnlQcC/O2f/6So2W6YFqjdkB1ya4O82iCvNsirHXJrg7x6I28D8Fza4qxduJAPg0H+/MQTAAQCAe655x5CoRDE476Vn0AyAw7QsGVL5tgPfvADCgqSrzF//nxWrlzZ5rybZ87kb+vWJecH3HjmmeyVY1vEbFC7ITvk1gZ5tUFebZBXO+TWBnn1Rt4G4OFw2PMmOauefpr/lwpuAb773e9y1FFHJe/EY74twISmAHzXxo2ZY5WVlZx77rlAcjOgm266qcU5C375S/77qacy988ZM4YTzzzTtzl1Ri5eRefIrQ3yaoO82iCvdsitDfLqjbwNwAOBAAEPtdDRujp+++STrNm2DUgGw9dee23TE3zOgIeKigiXlbWoAwf44Q9/mFzsCTz00EOsWbMGgDWvvcZXfvhD0tv0zBoxgpl77UVRaht6a7x6FbtHbm2QVxvk1QZ5tUNubZBXb+StsYaGBhoaGrp83qI//IHHNm/O3L/rrrvo379/5r4bj/uaAQcoHjYs04owzdixYznrrLOAZH3VLbfcQqS2li/OmMGW1KLLseEwv3vtNaLbt1PYTQG4V69i98itDfJqg7zaIK92yK0N8uqNvA3AvZBIJJj7s5/RmApwZ82axec///mWT/K5BAWSu2HWNytBSfPf//3fmdv33HMP5x9xBK/v3AlAKfDIX/5CcXk5bjzebQG4EEIIIYTIjbwNwIuKiijq4jbs9/z617y3fj0A/fv354477mj7pHgcx8cSFGg/Aw5w5JFHcvLJJwOwa9cu/vTRR5nHfvf973Pw5z9PJLX7ZaFx95M0XryK7JBbG+TVBnm1QV7tkFsb5NUbeRuAJxIJEonE7p+Y4tNPP+WKZhnn6667jlGjRrV9okEGvHjo0DY14GmuvPLKNscuO+oovnTLLQBEamoAui0D3lWvInvk1gZ5tUFebZBXO+TWBnn1Rt4G4NFolGiqR3Y2zJkzh9odOwA45phjuOSSS9p/okENeMmwYS26oDRn6tSpHFNWlrk/bcAArn/xxcz9SHU1waIiQt3UBL+rXkX2yK0N8mqDvNogr3bIrQ3y6o28DcALCgoyvbSz4T/+4z8YOmRIpud3MNjBTpc+d0GB1G6YmzfjpmrPm+O88QZ319YyHDgUePjFFwk1+6knUl3dbeUn0HWvInvk1gZ5tUFebZBXO+TWBnn1hr+p3F5EqItZ6i984QtMnz6d559/ngkTJnT8RItFmMOG0VhfT6yujnCzbDcAN9zAEcAGgO9/H8aPb/FwpKamWxdgdtWryB65tUFebZBXG+TVDrm1QV69kbfW6uvrAbr0rax///7MnDmz0+e4jTYZcIBdmza1DMBXroT585O3CwrgBz9oc26kurpbA3AvXkV2yK0N8mqDvNogr3bIrQ3y6o28LUExIx7DCfr7vaVo0CCcUKhtK8KbboJ0Wco550BlZZtzuzsDLoQQQgghciNvM+CFhYU2A8fjUOLvtzwnEKB4yJCWrQjXrIGHHko9wYHLL2/33Eh1NSUjR/o6n84w8yrk1gh5tUFebZBXO+TWBnn1hjLgXcWgCwoke4G3aEV4yy3J1wI46ywYO7bd8xq2bevWRZhCCCGEECI38jYAj0QiRCIR/wc2WIQJrXbD3LIF7rmn6cFm/clbE6mpoagbS1DMvAq5NUJebZBXG+TVDrm1QV69kbclKGaLAWIx3xdhQjIDvnPduuSdO+6AXbuSt08+GY48ssPzunsRphZZ2CG3NsirDfJqg7zaIbc2yKs38jYAt2qL48bjOEYZ8C3/+hfs2JEMwNO0sxNmmkQ8TrS2tltLUNRuyA65tUFebZBXG+TVDrm1QV69kbfWGhoaAINvZnGjDHhFRXI3zHvugerq5MGJE2Hq1A7PidbWgut2awbczKuQWyPk1QZ5tUFe7ZBbG+TVG3kbgLe3q6QvxOM4BTYZ8IatW0n88pdNhflXXpnsgNIBkVSg3p0BuJlXIbdGyKsN8mqDvNohtzbIqzfyNgA3bUNoVAOO69KwaRMlAAcdBJ//fKfnRGpqAAiXl/s+n45QuyE75NYGebVBXm2QVzvk1gZ59UbedkExw6oLyuDBAOxK/4RzxRUQ6PzjiVRXU9C/P8Fw2Pf5CCGEEEIIG/I2ALdrQ2iTAQ898wwFjY3Uh0IwZgx87Wu7PSdSXd3tPcDVbsgOubVBXm2QVxvk1Q65tUFevZG3JSgWq3Jd14VGg414XBeuv56SWIxdoRBcdhlksZihu1sQglY7WyK3NsirDfJqg7zaIbc2yKs38taayWrcxtTOlH5nwJ99Ft56i+IxY6gvK4MLLsjqtIYeCMC1ytkOubVBXm2QVxvk1Q65tUFevZG3AXj65xBfL4zU1vC+9wG/4QYAiuNxdh12GJSUZHVapKamR0pQQP/BWSC3NsirDfJqg7zaIbc2yKs38rYGPJFIkEgk/B00Hkv+288A/LXXYOFCAEoch/rUYsxs6IkSFBOvApBbK+TVBnm1QV7tkFsb5NUbeZsBN2mLEzcoQUllvwGKjz2W+m3bsj61JxZhqt2QHXJrg7zaIK82yKsdcmuDvHojbzPgJvidAV++HB59NHk7HKbky19O7oaZJZGamm7PgAshhBBCiNzI2wDcpC2O3xnwG29suv2Nb1A8dizxXbuI7diR1ek9UYKidkN2yK0N8mqDvNogr3bIrQ3y6o28LUEJBoO+j+lmAnAftH38Mfzxj8nbgQD88IcUl5UBsGvjRsr79ev09MZolPjOnd1egmLhVSSRWxvk1QZ5tUFe7ZBbG+TVG3kbgIctdodMl6AEfdD2y182ZdS/9CXYf3+KGhtxgkHqN2+mfL/9Oj09vQ190aBBuc+lC5h4FYDcWiGvNsirDfJqh9zaIK/eyNsAPBqNAgZtCIMhHMfJbZzNm+G++5ruX3EFAIFgkKIhQ7KqA4+kFmt2dwmKiVcByK0V8mqDvNogr3bIrQ3y6o28DcAbGxv9HzQe86f85Pbbob4+efu00+DwwzMPlVRUUL9p026HiNTU4AQCFPTvn/t8uoCJVwHIrRXyaoO82iCvdsitDfLqjbwNwM3aEOa6ALO2Fu68s+n+lVe2eLi4oiK7DHh1NeHycgLdXHuldkN2yK0N8mqDvNogr3bIrQ3y6o287YJigh8Z8AcegFT9NpMnw/HHt3i4ONsMeA90QBFCCCGEELmTtwG4RVscNx7PPQD/5z+bbn//+9Cqnrxk2LCsS1C6uwMKqN2QJXJrg7zaIK82yKsdcmuDvHojb0tQAgGD7xa5lqAkErB4cdP96dPbPKW4ooJdvTgDbuJVAHJrhbzaIK82yKsdcmuDvHojbwNwmxrwGE4uGfAVK6C6Onn74INh8OA2TykZNoyGLVtINDZ2Wt/dE9vQg2q9LJFbG+TVBnm1QV7tkFsb5NUbefu1JRaLEYvF/B40twz4iy823W5V+52muKICt7GRhq1bOx2qp7ahN/EqALm1Ql5tkFcb5NUOubVBXr2RtxnweHqTG38Hza0G/KWXmm4fd1y7TympqACgfuPGzO326KkSFBOvApBbK+TVBnm1QV7tkFsb5NUbeRuA27Uh9CkA7yADXtCvH6GSEuo3b+50qIYeCsD1U5MdcmuDvNogrzbIqx1ya4O8eiNvS1BMiMfA605Pa9bAxx8nb1dWwpgxHT61ZNiwTnuBu66rNoRCCCGEEHsoeRuAm7QhbMyhC0rr8pNOtrPfXS/w+K5dJKJRtSHMM+TWBnm1QV5tkFc75NYGefVG3pagOJ0EuJ6Jx3CKSrydm0X5SZri3WTAI6mNfIp6IANu4lUAcmuFvNogrzbIqx1ya4O8eiNvA/CioiL/B82lD3jzDigdLMBMUzJ0KNuWL+/w8UiqlWFPlKCYeBWA3FohrzbIqw3yaofc2iCv3sjbEpR4PO7/ylyvW9FXV8PSpcnb5eUwblynT99tBry6mkAoRKi0tOtzyRETrwKQWyvk1QZ5tUFe7ZBbG+TVG6YBuOM4Qcdx/uA4zkLHce53HCfkOE6l4ziPpY79xOq1TfpSeu2C8sor4LrJ21OmQCcb7ECyFWFnXVDSPcB74mcf9fu0Q25tkFcb5NUGebVDbm2QV29Yl6DMAla7rnu24ziXA7NTxy5xXXed5QvbtSH0UILShfITSGbAY3V1xHftIlTStua8JzugqN2QHXJrg7zaIK82yKsdcmuDvHrDOgDfD3g7dfstksH33sAvHcepAK5yXfflZs/vv2HDBkaPHp35QCORCAUFBYRCIRoaGnBdt8VjoVCIgoICIpEIiUQi81hdXR3BYJBAIEA0GqWxsbHFeYFAgMLCQmKxGPF4vMVjjuNQVFREPB4nFos1XVyRCPF4nHDqnFgsRjgcJhAI0NDQACRroRKJBNFoNDNvZ9GijOjoxIk01NZ2+p6c/v0B2PLvf1NSWdnisWAwSP3WrYTKyqitrc35PaVXLhcXF2f1nlzXJRgMZs7L9XNKv6dwOOzb59TV95Sed319fY++p7q6OhzHIRAI5M176g2fU2NjIwUFBcRisbx5T73hc4rFYpnuB/nynnrD51RXV0cgECAQCOTNe+otn1NjY2Pmms2X99QbPqe6ujqAFv/v2tPfk9XntGPHDtJY14AvB6anbp8EHAgcBlwOfA24rdXzz3n44Yd9eeFYLEY0GvVlrAyNHkpQGhoIvvkmAG44jHvUUbs9pXDwYHAcGjpoRRipribcAy0IAaLRqP9eBWB0zQpds0ZEo1H97GyA/g7Yob8FNqgExRuOm65Nthg8WaR8M3A4sAyIAqe4rntI6vGXgRNc142n7k96+eWXX5k8eXLOr11bWwtAWVlZzmOlif3qZwSnnkrgsKOzP+mll5raDk6Z0rIdYSc8cuKJHH7ZZewzc2abx1783vcoGjKEo+fNy34ePmHhVSSRWxvk1QZ5tUFe7ZBbG+Q1e5YtW8b48ePHu667zLQExU1G93MBHMe5BlgAHOk4zgAgBhSmg+8UdX59gMXFxb6M0wIvXVBab8CTJSXDhnW4GU+kpobyz3yma/PwCROvApBbK+TVBnm1QV7tkFsb5NUbpgG44zjDgT8BCaDKdd0XHMf5EfA4EAautnrtdEucAq9bx7c/aNcXYXZhA57mFFdUsKuTEpSeWoRp4lUAcmuFvNogrzbIqx1ya4O8esM6A/4pMK3VsZeB7CNRj6TrkXz9ZtbVDHgiAYsXN90/9tisTy2pqKC+g17gkZqaHtmGHoy8CkBurZBXG+TVBnm1Q25tkFdv5O1OmOFw2NfxXDcBjY1dy4AvWwapbeMZPx66kLUuHjaMbStXtp1HIpHpA94T+O1VNCG3NsirDfJqg7zaIbc2yKs38jYADwR8bvASb0z+uysZcI/lJ5AsQWkvAx6rq8NtbOyxANx3ryKD3NogrzbIqw3yaofc2iCv3sjbADzdp9G3b2bx5E8sTlcy4F3cgKc5JRUV1G/ZgptI4DS7uBtSGfWeKkHx3avIILc2yKsN8mqDvNohtzbIqzf0tSVbUl/5UuAAACAASURBVIsMPGfAuxiAFw8bhhuP07BtW4vjkepqgB7LgAshhBBCiNzI2wx4UVGRvwOmMuBZB+CffAJr1iRvjxmT/KcLlFRUAFC/aRPFQ4Zkjke2bSNUXEzI7/eXJb57FRnk1gZ5tUFebZBXO+TWBnn1Rt5mwBOJBIlEwr8BMxnwLEtQcig/ASjo359gcTG7WtWB9+QCTDDwKjLIrQ3yaoO82iCvdsitDfLqjbwNwH3fcrarGfAcyk8AHMdJ1oG36gXekz3AQVv5WiK3NsirDfJqg7zaIbc2yKs38rYExe+G8G4mAM9y3Bw6oKQprqhomwGvru6xBZigRvuWyK0N8mqDvNogr3bIrQ3y6o28DcBDXd0yfnfE44ADweDun7ttGyxdmrw9YAAcfLCnlyyuqKB+8+YWx3q6BMV3ryKD3NogrzbIqw3yaofc2iCv3mjXmuM4w1O7WO6x1NfXAz5+M4vHIRTCcZzdP/fll5tuT5kCHntkllRUUPPBBy2ORaqr6VdZ6Wk8P/Ddq8ggtzbIqw3yaoO82iG3NsirNzqKDP/tOM7jjuPMchxHX22ga9vQ+1B+AslWhLta14D3cAZcCCGEEELkRkcB+H7AYuA6YL3jOLc6jnNI900rdwoLCyksLPRvwHi82zqgpClpZzfMnl6E6btXkUFubZBXG+TVBnm1Q25tkFdvtBuAu667znXdG1zXPQj4AlAMLHIc5w3Hcb7tOE7/bp1lbyDbDHh9PSxZkrxdWAhHHeX5JYsrKojW1hJP7TIFPR+ACyGEEEKI3MimOHk9sAHYDgwAzgLWOI5zruXEciUSiRCJRHwbz802A/7GGxBLdUyZODEZhHukZNgwgEwrwkQsRrS2tkcDcL+9iibk1gZ5tUFebZBXO+TWBnn1RrsBuOM4JY7jnOs4zkJgKcmSlPNd193fdd0ZwGeBX3XjPLtMQUGBvwsCUoswd4tP5ScARakdMNMBeGT7dqBnt6H33avIILc2yKsN8mqDvNohtzbIqzc6iig/BVYC9wNnuq67vfmDruu+4jjOa9aTywX/2xDGcLIZM8cNeJoTDIcpGjw40ws8UlMD0KN9wNVuyA65tUFebZBXG+TVDrm1QV690ZG141zXfbezE13X/azBfHyjIVU37W8bwt2M1djY1ILQceDYY3N+2eJmu2FGqqsBKCwvz3lcr/juVWSQWxvk1QZ5tUFe7ZBbG+TVGx3VgA93HGds8wOO4xzoOM7J3TAnX3BdF9d1/Rswm0WYS5dCqkyEQw5JbsKTI813w4xUV1NQVkagBy9y372KDHJrg7zaIK82yKsdcmuDvHqjo4jyDmBGq2M7U8cPNJ2RT/jeEiebDLiP5SdpSppnwGtqerT8BAy8igxya4O82iCvNsirHXJrg7x6o6MM+AjXddc2P+C67hpglP2UeiduNoswfdqApzmtS1DUglAIIYQQYs+mowB8neM4hzU/4DjOoSRbEu4R+N4WZ3clKK7raweUNCXNdsOMVFf3eAZc7YbskFsb5NUGebVBXu2QWxvk1RsdRZS/Bv7qOM7VwP8B+wNXp47vEfjfBSWOU1La8eMffwzr1iVv77UXVFb68rLpDLjrukRqaigaNMiXcb2i1c52yK0N8mqDvNogr3bIrQ3y6o3OasABfgzsBXwM3AXc3h2T8gPfV+PuLgNuUH4CyQx4IhYjUl1NZNs2BhxwgG9je0GrnO2QWxvk1QZ5tUFe7ZBbG+TVG+1GlG5yOevt7EEBd2vSP4d0WxtCg/ITSGbAAeo3bkwuwuzhGnDfvYoMcmuDvNogrzbIqx1ya4O8emO3vxs4jtMfcNL3XdetNZ2RTyQSCX8H3N0iTIMOKADh8nIC4TC7Nm/uFYswffcqMsitDfJqg7zaIK92yK0N8uqNdiNKx3H2Au4FjgNa95cJWk/KD/xvQ9hJCcrWrbB8efL2oEFw0EG+vazjOJQMG5bMgPeCRZhqN2SH3NogrzbIqw3yaofc2iCv3uioC8pdQDVwLLADOAJ4DLiom+bV63A7K0FZvLjp9pQpEOhIqzeKKyrYsWYN8fr6Hs+ACyGEEEKI3OiopmIysJfrujscx3Fd133HcZwLgReB+7pvet5J1yT59s0sHsPpKANuVH6SpqSigur33wfo8QDcd68ig9zaIK82yKsN8mqH3Nogr97oKABvBNJNHWsdxxkEbAdGd8usfCAY9LlSJh7rOANu1AElTXFFBZveeAOgx0tQfPcqMsitDfJqg7zaIK92yK0N8uqNjgLwt4HpwDM0Zb13Acu7aV45Ew6H/R2wo0WY9fWQCo4pKoIjjvD3dUn1At+8GScYJFxW5vv4XcF3ryKD3NogrzbIqw3yaofc2iCv3uioWPkCYGXq9veArUABcF53TMoPotEo0WjUvwE7CsBffx1iseTtiRPB4CeYkmHDgGT22/G5vryr+O5VZJBbG+TVBnm1QV7tkFsb5NUbbSJKx3FCwKXAPADXdbcA/9nN88qZxsZGfwfsqATFuPwEmnqB93T5CRh4FRnk1gZ5tUFebZBXO+TWBnn1RpsA3HXduOM43wSu6IH5+IafiwHcRCMkEu0H4EYb8DQnkwHvBR1QtMjCDrm1QV5tkFcb5NUOubVBXr3RUT3D/wJf786J9Gri8eS/W5egNDbCyy8nbzsOTJ5s8vK9KQMuhBBCCCFyo6NFmKOBCx3H+T7wCZDZ5sh13dndMbFc8bUtTioAd1pnwN97D+rqkrcPPRTKy3N/rXYIhsMUDhjQKzLgajdkh9zaIK82yKsN8mqH3Nogr97oKAB/I/XPHkvAz8WKHWXA09lvMCs/SVNcUUG4F2TAffUqWiC3NsirDfJqg7zaIbc2yKs32g3AXdf9SXdPxG98/SYWT3U5aR2Ar1vXdHvcOP9erx0OmzuXfpWVpq+RDfqGa4fc2iCvNsirDfJqh9zaIK/eaDcAdxznhI5OcF33Bbvp+Ecs1RqwoKCDzXO6QiYD3mqs6uqm28blISONM+zZ4qtX0QK5tUFebZBXG+TVDrm1QV690VEJyt9b3S8jWQdeBwwynZFPxNNBsy+DdZABr6lput0L6rO7A1+9ihbIrQ3yaoO82iCvdsitDfLqjY5KUFpEk47jlAI/B97tjkn5ga9tCDvKgDcPwHtBfXZ3oJ+a7JBbG+TVBnm1QV7tkFsb5NUbWVXOu667k2Rf8GtMZ9NbiceSbQZbLzTogwG4EEIIIYTIja4sXd0bKDWah+9EIpFMa5ycicchVIDjOC2P98EA3FevogVya4O82iCvNsirHXJrg7x6o6NFmPMBt9mhEmAy8NvumJQftAmWcyEeb1v/DS0DcKMe4L0NX72KFsitDfJqg7zaIK92yK0N8uqNjhZhvt3q/g7gRtd1FxrPxzeKior8Gywe6zwALy6GPlID5atX0QK5tUFebZBXG+TVDrm1QV69kbd9wNOrcn1rQ9h6AWYkAvX1ydt9pPwEfPYqWiC3NsirDfJqg7zaIbc2yKs32q0BdxznCsdxJrY6dozjOD/snmnlTiwWy/SmzBW3vQz49u1Nt/tQAO6nV9ESubVBXm2QVxvk1Q65tUFevdFRCcqlwN2tjq0A/gb8wnRGPuHvTpjtZMD74AJMULshS+TWBnm1QV5tkFc75NYGefVGR11QSoBdrY7tBPrbTqeXEo/jdLYJTx8KwIUQQgghRG50FIC/D8xsdewM4APb6fiHv20I2ylB6aMBuNoN2SG3NsirDfJqg7zaIbc2yKs3OipBmQc86jjOn4APgf2B/wDO6q6J9SraK0Gprm663Ue2oRdCCCGEELnTUReUZx3HOQ74T+AE4GPgeNd1/9Wdk8uF4uJi/wZTBjyDr15FC+TWBnm1QV5tkFc75NYGefVGRxlwUsH2t7txLr7iZ1ucdrug9NEAXO2G7JBbG+TVBnm1QV7tkFsb5NUbakOYDfE4BNUFBdRuyBK5tUFebZBXG+TVDrm1QV690dEizEtJth1szgrge7bT8Y9wOEw4HPZnMHVByeCrV9ECubVBXm2QVxvk1Q65tUFevdFRCcoe34YwEOjou4UH4jH1AU/hq1fRArm1QV5tkFcb5NUOubVBXr2Rt20IGxoaaGho8GeweBxa1zb10QDcV6+iBXJrg7zaIK82yKsdcmuDvHoj2zaEnwG+Sp9tQ6hFmEIIIYQQwh+ybUP4b+B4oLz7ppYbRUVF/g0WjysAT+GrV9ECubVBXm2QVxvk1Q65tUFevbHbNoSO44wEvgH8BRgBlHbP1HIjkUj4NpYbj+N0VgNevsd8L8kZP72KlsitDfJqg7zaIK92yK0N8uqNjtoQBh3HOdNxnMdJZr+vBf4HGNONc8uJaDRKNBr1Z7B4rOMuKCUl0IdW//rqVbRAbm2QVxvk1QZ5tUNubZBXb7SIKh3HGQtcAJxLMtM9Hzgd+ANwl+u6W7t9hh7xtSF8663oGxqS/0CfKj8BNdq3RG5tkFcb5NUGebVDbm2QV2+0LkFZAWwFrgD+13XdnQCO47jdPbFcCbXOWOdC60WYzctPBg7073X2AHz1KlogtzbIqw3yaoO82iG3NsirN1pb+z3wJeDnwEGO4zzguu7Sbp+VD9TX1wM+fTNrnQHvowswwWevogVya4O82iCvNsirHXJrg7x6o0UNuOu63yS50HIeMAV4x3Gct0huwNN3Vhq2pnUXlD4cgAshhBBCiNxoswjTdd0druve57ruscAhwEKSu2C+4zjOfd09Qa8UFhZSWFiY8zhuohHchALwFH55FW2RWxvk1QZ5tUFe7ZBbG+TVG53uH+q67nLXdecCo4BzSGbH+xbxePLfKkERQgghhBA+0GkAnsZ13bjrun9zXfdz1hPyi0gkQiQSyX2geAygZRvCPhyA++ZVtEFubZBXG+TVBnm1Q25tkFdv5O3SVd8WAygD3gItsrBDbm2QVxvk1QZ5tUNubZBXb+RtAO5bW5xMAK4MOKjdkCVya4O82iCvNsirHXJrg7x6I2+tNaQ2ysn5m1mqBEUBeBLfvIo2yK0N8mqDvNogr3bIrQ3y6o28DcBd15+9g1yVoLTAL6+iLXJrg7zaIK82yKsdcmuDvHojbwNw31riKAPeArUaskNubZBXG+TVBnm1Q25tkFdvZNUFpU8Tj4MTwAkEm45VVzfd7mNb0QshhBBCiNzI2wDct7Y4sVjL7Df06Qy42g3ZIbc2yKsN8mqDvNohtzbIqzdMS1AcxwkCD5DcyGc18C3XdeOO4+wFfAAc6bruUovX9rULSmcBeHm5P6+zh6DVznbIrQ3yaoO82iCvdsitDfLqDesM+Cxgteu604CVwOzU8cuBxZYvXFBQ4M+K3His5QJM120KwPv1axuc5zm+eRVtkFsb5NUGebVBXu2QWxvk1RvW0eN+wNup228BZzqOswRwgU/aeX7/DRs2MHr06ExRfyQSoaCggFAoRENDA67rtngsFApRUFBAJBIhkUhkHquuriYQCDBw4ECi0SiNjY0tzgsEAhQWFhKLxYjH4y0ecxyHoqIi4vE4iR11FASDTT+x1NdTFo0CkCgvJ9rQQCAQyLThKSoqIpFIEI1GM/Our68HyPk9RSIRgsEg4XA4p/cUi8VaPAZQXFyceSwcDnf4nmKxGMFgkEAgkDfvqbd8TtXV1TiOw6BBg/LmPfWGz6n54/nynnrD51RfX08kEqGkpCRv3lNv+Jya/78rX95Tb/mc0n8PiouL8+Y99YbPqTq1Lm7w4MF5856sPqcdO3aQxjoDvhyYnrp9EjAQuAK4uYPnn/Pwww/78sKu6/rTGqdVCYqzfXvTa/Sx8hOARCJBIpHo6WnkJa7ryq0BumZtSCQSaj9mgP4O2KG/BTb4Fm/1MRxLaY7jOCSD7cOBZUAlsNF13Ysdx/k9cHPzGnDHcSa9/PLLr0yePDnn105/e8m1PU7j4oW4y98h9J9zkgdWrICDD07ePu44ePHFnMbf0/DLq2iL3NogrzbIqw3yaofc2iCv2bNs2TLGjx8/3nXdZaYZcDfJXNd1pwNbgX8B4xzHeRo4GfiN4zhFzU6pKysrs5xS12m9CLMPd0ARQgghhBC5YxqAO44z3HGchY7jVAFR13V/6rru8a7rngY8C1zsum6DxWv71hYnHlMA3gy1G7JDbm2QVxvk1QZ5tUNubZBXb5guwnRd91NgWgePfcPytYPB4O6flA3xuLahb4ZvXkUb5NYGebVBXm2QVzvk1gZ59Ube9tALh8P+DKQMeAt88yraILc2yKsN8mqDvNohtzbIqzfyNgCPploF5tqb0m2dAW++DX0fDMD98iraIrc2yKsN8mqDvNohtzbIqzfyNgBvbGz0aaAYhJp9u2ueAR840J/X2IPwzatog9zaIK82yKsN8mqH3Nogr97I2wDct3Y48ThOUUnT/T5egqI2Q3bIrQ3yaoO82iCvdsitDfLqDeuNePZ8tAhTCCGEEEL4SN4G4GpDaIPaDdkhtzbIqw3yaoO82iG3NsirN/K2BCUQ8Om7hTLgLfDNq2iD3NogrzbIqw3yaofc2iCv3sjbANyvmiQ3HsdRBjyDar3skFsb5NUGebVBXu2QWxvk1Rt5+7UlFosRi8VyH6izEpSystzH38Pwzatog9zaIK82yKsN8mqH3Nogr97I2wx4PB73ayCcdAmK6zYF4P37twzM+wi+eRVtkFsb5NUGebVBXu2QWxvk1Rt5G0H614awWQa8vh7S3/L6YPkJ6KcmS+TWBnm1QV5tkFc75NYGefVG3pag+EbzRZh9vP5bCCGEEELkTt4G4P61IYw3ZcAVgKvdkCFya4O82iCvNsirHXJrg7x6I29LUBzHyXkM13WTJSfpALy6uunBPrgNPfjjVbSP3NogrzbIqw3yaofc2iCv3sjbALyoqCj3QRKNgKsSlGb44lW0i9zaIK82yKsN8mqH3Nogr97I2xKUeDye+8rc1PmOSlAy+OJVtIvc2iCvNsirDfJqh9zaIK/eyNsA3Je+lPHU+cqAZ1C/Tzvk1gZ5tUFebZBXO+TWBnn1Rt6WoPjSFif9jU4Z8AxqN2SH3NogrzbIqw3yaofc2iCv3sjbDLgvZAJwZcCFEEIIIYQ/5G0A7ktbnHQJSoEC8DRqN2SH3NogrzbIqw3yaofc2iCv3sjbANwPXJWgCCGEEEIIn8nbGvDi4uLcB0lnwIMKwNP44lW0i9zaIK82yKsN8mqH3Nogr97I2wy4b20IA0GcwP/f3r3HR13f+R5/fZNJSMK9yhGLN9BWBGqpAoLSZkIXWvVwq10pRcTdba2V4wXbbUHocdXFO667trTd43mo9YIFpMfjtqtyMUDRVnfVegS0BbUuBSkK4ZbbTOZ7/phLJiHX38yHJL++n49HHs59fvNKiN/55TvfXyqTBuBabsiQ2tpQVxvqakNd7aitDXUNJrQD8LwtQxjJ+iOBBuBabsiQ2tpQVxvqakNd7aitDXUNJrRTUIqLi3N/kHi8cQUUaDwUvXPQv3/uj98D5aWrtEhtbairDXW1oa521NaGugYT2gF4QUEedu7H4417wL1v3APerx/k4/F7oLx0lRaprQ11taGuNtTVjtraUNdgQjsAr62tBXJ8Z5Y9BeXoUWhoSJ7+C51+AnnqKi1SWxvqakNdbairHbW1oa7B6G1LG3z2FBTN/xYRERGRPAjtHvCSkpLcHyQew2kN8Cby0lVapLY21NWGutpQVztqa0NdgwntADyRSOT+INoDfoy8dJUWqa0NdbWhrjbU1Y7a2lDXYEI7BaW+vp76+vrcHiT7Q5gagAN56iotUlsb6mpDXW2oqx21taGuwYR2D3hRUVH7N2pP9ocwNQAH8tRVWqS2NtTVhrraUFc7amtDXYMJ7QA8EsnDS9MUlGPkpau0SG1tqKsNdbWhrnbU1oa6BhPaajU1NUBu78y89oAfIx9dpWVqa0NdbairDXW1o7Y21DWY0M4BzwvtARcRERGRPAvtHvBevXrl/iDxOK6sd/K0BuBAnrpKi9TWhrraUFcb6mpHbW2oazDaA96W7CkoBw40Xj5wYNdsj4iIiIj0eKEdgNfV1VFXV5fbg8RjmoLSTF66SovU1oa62lBXG+pqR21tqGswoZ2Ckp9lCLUOeHP6kIUdtbWhrjbU1Ya62lFbG+oaTGgH4PlYFsfHYzjtAW9Cyw3ZUVsb6mpDXW2oqx21taGuwYS2Wm1tLZDjO7OW9oA7B3375rh1PVdeukqL1NaGutpQVxvqakdtbahrMKEdgHvvc3+QeDy5B9z7xgF4//5QENqp8+3KS1dpkdraUFcb6mpDXe2orQ11DSa0A/D8LEOYWgXlyBFIJJKX/QVPPwEtN2RJbW2oqw11taGudtTWhroG85e7K7cj0gfi0fxvEREREcmT0A7A87cMYUQD8CxabsiO2tpQVxvqakNd7aitDXUNJrRTUHL9VK73PmsP+EeNV/yFD8D1aWc7amtDXW2oqw11taO2NtQ1mNBWy/nTuA0Nyf9qD3gT+pSzHbW1oa421NWGutpRWxvqGkxoB+DpP4cE/sGIxwBwkYgOQ58l567SKrW1oa421NWGutpRWxvqGkxoB+CJ9KolQcXjyf/qQ5hN5NxVWqW2NtTVhrraUFc7amtDXYMJ7QA852VxUnvANQWlKS03ZEdtbairDXW1oa521NaGugYT2lVQcqY94CIiIiJiILQD8JyXxckMwLUHPJuWG7KjtjbU1Ya62lBXO2prQ12DCe0UlMLCwpzu7zNTULQHPFuuXaV1amtDXW2oqw11taO2NtQ1mNAOwIuLi3N7gMwe8EINwLPk3FVapbY21NWGutpQVztqa0NdgwntALy+vh7IcRnCwkKcK9AAPEvOXaVVamtDXW2oqw11taO2NtQ1mNAOwBvSB9IJKn0UTNAAPEvOXaVVamtDXW2oqw11taO2NtQ1mNAOwPOyDGH68KrpAXhBAfTpk9vj9nBabsiO2tpQVxvqakNd7aitDXUNJrSroOQsvQc8kYCDB5OX9e+fHISLiIiIiAQU2tFkrsvi+PQe8MOHk4Nw+IuffgJabsiS2tpQVxvqakNd7aitDXUNJrRTUApy3VMdjx+7BvjAgbk9Zgjk3FVapbY21NWGutpQVztqa0NdgwntADwfc8Cd1gA/huZ62VFbG+pqQ11tqKsdtbWhrsGE9m1LLBYjFosFf4CW9oBrAJ57V2mV2tpQVxvqakNd7aitDXUNJrR7wOPpA+kEfoCYBuAtyLmrtEptbairDXW1oa521NaGugYT2gF47lNQ4joMfQv0pyY7amtDXW2oqw11taO2NtQ1mNBOQcmZBuAiIiIiYiC0A/C8LUOoAXgTWm7IjtraUFcb6mpDXe2orQ11DSa0U1Ccc7k9QDyOKynVALyZnLtKq9TWhrraUFcb6mpHbW2oazChHYCXlJTk9gDxGET6agDeTM5dpVVqa0NdbairDXW1o7Y21DWY0E5BicfjuX0yV8sQtijnrtIqtbWhrjbU1Ya62lFbG+oaTGgH4PlZB1wfwmxO633aUVsb6mpDXW2oqx21taGuwYR2Cko+joRJJAIHDjRepkPRa7khQ2prQ11tqKsNdbWjtjbUNZjQ7gHPldcecBERERExENoBeM7L4jRfhrCwEHr3zs/G9WBabsiO2tpQVxvqakNd7aitDXUNJrQD8JzF47iCAjh0KHl+wADQUjsiIiIikqPQzgEvLS3N7QHicaitB++T5zX9BMhDV2mV2tpQVxvqakNd7aitDXUNxnQA7pwrBB4FhgDvATcAz6SeNw78jff+jxbPnV4Sp6ioKOADxKCmuvG8BuBAHrpKq9TWhrraUFcb6mpHbW2oazDWe8BnAu95769wzn0vdf4K7/1u59yXgL8H/kfW7fvu2bOHU089NfOp2rq6OoqKiohEItTW1uK9b3JdJBKhqKiIuro6EolE5rqqqioKCgqIRCLU19fT0NDQ5H4FBQX06tWLWCxGPB5vcp1zjpJ4nFjVQdKf7U3068eR1HSU0tJS4vE4sViM4uJiCgoKqK2tBZIL0icSCerr6zPbXVNTA5Dza6qrq6OwsJDi4uJgr6mkJLPd2dd15jWltyG95FAYXlN3+T5VpT5vEIlEQvOausP3qaamhqKiIhoaGkLzmrrD96m6upr6+npisVhoXlN3+D5VVVXhnCMSiYTmNXWX71NNTQ11dXXHPEdPfk3d4fvU0v+7evprsvo+HTlyhDTrOeBnAm+kTr8GjPPe706drwcSzW4/96mnnsrLExcVFQV/N+Y9NMRxRxv3gHvtAQdy7CptUlsbxcXF6mpAP6821NWO2tpQ12CcT89xtnhw56YCX/bez3fO3QWc6r2f45wrBtYC3/De/yHr9uNfeumllydMmJDzc9fX1wPJ//l2lo/HiC9dSOGQsyn45reSF/7d38FDD+W8XT1dLl2lbWprQ11tqKsNdbWjtjbUteO2bt3KqFGjRnnvt1rvAf83oNY5twHoDXyYuvxfgeXZg++Uw/369cvLE9fW1mb+TNBpqflM7ujRxsu0BxzIsau0SW1tqKsNdbWhrnbU1oa6BmM6APdJ3/HeTwI+Bp5xzt0CvOu9/7nlc+ckfUjVw41zdTQAFxEREZF8sF4FZTCwguRc7/UkV0LZAPzaOTcJeNl7v8jiuUtKSoLfOZ4egB9uvEyHoQdy7CptUlsb6mpDXW2oqx21taGuwZgOwL33HwIVx/M50xKJ5p/v7ITUFJQmA3DtAQdy7CptUlsb6mpDXW2oqx21taGuwYT2SJj19fWZDwZ0WnoP+CENwJvLqau0SW1tqKsNdbWhrnbU1oa6BhPaI2HmsiSOT+8BP3Sw8UINwAEttG9JbW2oqw11taGudtTWhroGE9oBeCSSw0tLD8APHmq8TANwIMeu0ia1taGuNtTVhrraUVsb6hpMaKulj0YU6J1ZPAaFEVzq6E6ABuApOXWVrnSg4AAAIABJREFUNqmtDXW1oa421NWO2tpQ12BCOwc8J/E4FBWBBuAiIiIikmeh3QPeq1ev4HeOxyASaRyARyJQVpafDevhcuoqbVJbG+pqQ11tqKsdtbWhrsFoD3hL4nEoKIRDqTngAwaAc127TSIiIiISCqEdgNfV1VFXVxfovj4eb/wgJmj6SZZcukrb1NaGutpQVxvqakdtbahrMKGdgpLThwHiMaiPNZ7XADxDH7Kwo7Y21NWGutpQVztqa0NdgwntADzXZQhdvfaAt0TLDdlRWxvqakNdbairHbW1oa7BhLZabW0tEHwZQleb9eeUgQPztFU9X05dpU1qa0NdbairDXW1o7Y21DWY0A7AvffB7xyPQ13WYVW1Bzwjp67SJrW1oa421NWGutpRWxvqGkxoB+C5LUMYh+w94BqAZ2i5ITtqa0NdbairDXW1o7Y21DWY0K6CkpN4DFJ/UgE0ABcRERGRvAntADynZQgb4rga7QFviZYbsqO2NtTVhrraUFc7amtDXYMJ7RSU3FZBiUF1TeN5DcAz9GlnO2prQ11tqKsNdbWjtjbUNZjQVsttHfA41GgA3hJ9ytmO2tpQVxvqakNd7aitDXUNJrQD8PSfQwIvQ3i0uvG8BuAZOXWVNqmtDXW1oa421NWO2tpQ12BCOwBPJBLB7xyPQ7UG4C3Jqau0SW1tqKsNdbWhrnbU1oa6BhPaAXjOyxAeOdp4XgPwDC03ZEdtbairDXW1oa521NaGugYT2lVQcuFjMTiqAbiIiIiI5F9oB+A5LYsTj8HhI8nTxcVQWpq/DevhtNyQHbW1oa421NWGutpRWxvqGkxop6AUFhYGv3NdHS49B3zAAHAuPxsVAjl1lTaprQ11taGuNtTVjtraUNdgQjsALy4uDn5nzf9uVU5dpU1qa0NdbairDXW1o7Y21DWY0A7A6+vrgYDL4hw50nhaA/AmcuoqbVJbG+pqQ11tqKsdtbWhrsGEdgDe0NAQ/M76AGarcuoqbVJbG+pqQ11tqKsdtbWhrsGEdgAedFkcn0jgdBj6Vmm5ITtqa0NdbairDXW1o7Y21DWY0K6CElhDHGprG89rAC4iIiIieRTaAXjgZXHicajNup8G4E1ouSE7amtDXW2oqw11taO2NtQ1mNBOQSkoCPjeIh7DaQDeqsBdpV1qa0NdbairDXW1o7Y21DWY0A7AA89J0h7wNmmulx21taGuNtTVhrraUVsb6hpMaN+2xGIxYrFY5+8Yj2kA3obAXaVdamtDXW2oqw11taO2NtQ1mNDuAY/H40HvqCkobQjcVdqltjbU1Ya62lBXO2prQ12DCe0APPAyhPEY1GStgjJwYJ62KBz0pyY7amtDXW2oqw11taO2NtQ1mNBOQQlMc8BFRERExFBoB+C5LEPo6uobz2sA3oSWG7KjtjbU1Ya62lBXO2prQ12DCe0UFOdcsDvqQ5htCtxV2qW2NtTVhrraUFc7amtDXYMJ7QC8pKQk2B3jWUfC7NULgj5OSAXuKu1SWxvqakNdbairne7e1ntPLBajoaGhqzclkJqamq7ehG6lsLCQoqKiVt+ghHYKSjweD/bJ3JoaXH1qOR3t/T5G4K7SLrW1oa421NWGutrpzm299xw4cKBHTuUoKCjQwXhaUFdXx4EDB/Det3h9aPeAp9ekLC0t7dT9/MGqxjMagB8jaFdpn9raUFcb6mpDXe1057axWIyioiL69u3b1ZvSaek99oWFhV28Jd3P4cOHicViFBcXH3NdaAfggZfFqdIAvC1absiO2tpQVxvqakNd7XTntg0NDUQiPXNIpjngrYtEIq1OKdLfDJrTAFxEREREDIV2AB54WZyqg42nNQA/hpYbsqO2NtTVhrraUFc7atuopqaGaDRKNBqlb9++mdP79+9v835Lly7lgw8+aHKZ977Vec7jxo3jjjvuyNt2h0nP/HuHpUOHGk9rAC4iIiIhU1paSmVlJQBjxozJnE5LJBItfrBy8eLFHX6O999/n2HDhrF+/XpuvvnmXDa3Q9vW0/T8V9CK0tLSQB+0cAez9oDrMPTHCNpV2qe2NtTVhrraUFc7PaWtTzTgq/bn5yvR8SUN161bx7Rp05gxYwaPPfYYd955J9FolPPPP5/169cDcMUVV/D222+zbt06LrnkEmbOnMl5553H22+/fczjrVq1irlz53LmmWeyY8cOAPbu3cull15KNBrlyiuvBODZZ59l/PjxRKNRnnzySdatW8fChQsBeOONN/jGN74BwOc+9zmuu+465s2bxxtvvEF5eTkTJkzghhtuAJID82uuuYbPf/7zVFRU8Oc//5mJEydmtmfWrFm8++67Ab4jNkK7Bzy91FBRUVGn7ucPHW48oz3gxwjaVdqntjbU1Ya62lBXOz2m7aGDxP95aV4eKnLDYhjwiQ7f/siRI6xfvx7nHNXV1SxatIgPP/yQ2bNn88UvfrHJbRsaGvjFL37BM888wyOPPMK9997b5Pr169dz4403UlZWxqpVq1i0aBFLly7l6quvZvr06SQSCeLxOEuWLGHLli306dOHRCLBhg0bWty2jz/+mJtuuomhQ4dSU1NDZWUlzjkuvfRS3nvvPV5//XVKS0vZvHkzkByQn3vuubz22mucddZZ7N+/n2HDhnWyoJ3QDsCDLjfkjhxpPKMB+DG68zJOPZ3a2lBXG+pqQ13t9Ji2/fonB855eqzOGDNmTGZVk0ceeYQVK1ZQWFjI3r17j7nt6NGjATjllFM4cOBAk+vef/99tm7dytSpU0kkEhw9epRFixaxfft2brvtNiC5fviePXsYOnQoffr0yVyWvapK9tzyE088kaFDhwKwc+dOvvvd71JTU8POnTvZvXs327dvp7y8PHP7goIC5s2bx+OPP85nPvMZLrvssk61sBbaAXhLay52yGHtAW9L4K7SLrW1oa421NWGutrpKW1dQWGn9lrnU/bc6uXLl/Pmm2+yd+9eotHoMbdND5Sdc8d8CHP16tU8+OCDzJgxA4Crr76anTt3cs4557Bp0yamTZtGIpHgpJNO4v333+fo0aP07t2bRCLBwIED2bVrFwC/+93vWty2H/3oRyxcuJBoNMoll1yC9z7z2Onn9N5zwQUXsHDhQrZu3cqKFSvyEylPQjsHPPCRmY4cbTytAfgxdMQrO2prQ11tqKsNdbWjtp0zfvx4Jk6cyH333Ufv3r07dd81a9Y0GbRXVFSwcuVKFi9ezI9//GOi0ShXXXUVkUiE2267jUmTJlFRUcFTTz3F6NGjqaqqYsqUKbz++ustPv60adOYP38+X/3qVzOXzZw5k6NHj2bmgKdXdJkyZQplZWV84hNd86amNa61pWO6gnNu5FtvvfXWyJEjc36sQ6nVTPr169ep+yVGf4aC372VPPOb38AFF+S8LWEStKu0T21tqKsNdbWhrna6c9uamhqgB0yPaUF3PxLm3Xffzdlnn53ZM348Nf++bt26lVGjRo3y3m/VW8Fm3FHtARcRERHp6W655RYqKyuZOnVqV2/KMUI7B7ykpCTYHY9WN57WAPwYgbtKu9TWhrraUFcb6mpHbW1050PR33rrrV29Ca0K7QA8kUgEu2N1TePp/p379PBfgsBdpV1qa0NdbairDXW1o7bSnYR2Ckp9fT319fWdv2Nqvg4lJckvaSJwV2mX2tpQVxvqakNd7aitjbYORS+tC+0e8EAL7dfX4+qT64Rq+knLuv0BDHowtbWhrjbU1Ya62lFbG915Ckp3Fto94JFIhEikk+8vdBj6dgXqKh2itjbU1Ya62lBXO2rbqKamhmg0SjQapW/fvpnT6aX7WrN06VI++OCDDj/PuHHjuOOOO3LdXBPxeJx+/fpx3XXXAbBkyRLOPfdcotEoX/rSlzKr5rTkiiuu4O2332bdunUsXLgQgDlz5nD66ad3+PlDOwCvqanJLP/SYdlHctIe8BYF6iodorY21NWGutpQVztq26i0tJTKykoqKys5++yzM6fTa2W3Nl9+8eLFnHbaaU0ua20Kyvvvv8+wYcNYv359Xrc9n3P5R4wYwYMPPpg5f88991BZWcmECRM6feCeJ554gkGDBnX49qEdgAdSVdV4WgNwERER+Quxbt06pk2bxowZM3jssce48847iUajnH/++ZlBdPae30suuYSZM2dy3nnnsX379mMeb9WqVcydO5czzzyTHTt2ALB3714uvfRSotEoV155JQDPPvss48ePJxqN8uSTTzbZq/zGG2/wjW98A4DPfe5zXHfddcybN4833niD8vJyJkyYwA033AAkB+bXXHNN5kA8f/7zn5k4cWJme2bNmsW7777boRZVVVWZNxXr169n/PjxjB8/nieeeCJI2haF9m8xvXr16vR9/P6Pycxk0gC8RUG6SseorQ11taGuNtTVTo9pO2gQVFe3fZvUwW9o7+A3ZWWwb1+Hn/rIkSOsX78e5xzV1dUsWrSIDz/8kNmzZ/PFL36x2SY08Itf/IJnnnmGRx99lHvvvbfJ9evXr+fGG2+krKyMVatWsWjRIpYuXcrVV1/N9OnTSSQSxONxlixZwpYtW+jTpw+JRIINGza0uG0ff/wxN910E0OHDqWmpobKykqcc1x66aW89957vP7665SWlrJ582YgOSA/99xzee211zjrrLPYv38/w4YNa/P1f+973+M73/kORUVF3H777QAsWrSI5557jj59+jB+/PgmR9/MRWgH4IFkz33SAFxERESOt+rq9gfgRsaMGZP5UOUjjzzCihUrKCwsZO/evcfcdvTo0QCceuqpHMiewkty+snWrVuZOnUqiUSCo0ePsmjRIrZv385tt90GQEFBAXv27GHo0KH06dMnc1n2hzqzp7aceOKJDB06FICdO3fy3e9+l5qaGnbu3Mnu3bvZvn075eXlmdsXFBQwb948Hn/8cT7zmc9w2WWXtfv677nnHiZPnsz06dP505/+xPDhwwEyU3OGDRvGhx9+2O7jdERop6DU1dVRV1fXuTtpAN6uQF2lQ9TWhrraUFcb6mqnx7QtK8vvVycUFDQOC5cvX87GjRtZsWJFi/Ou0wPlluaAr169mgcffJDnnnuOF154gZEjR7Jz507OOeccNm3aBCT3UJ900km8//77HE0dhTyRSDBw4EB27doFwO9+97sWt+1HP/oRCxcuZOPGjZx77rl475s8dnq7LrjgAl5//XWeeuopLr/88g41KCwsZPHixZk3Ct579u/fT319Pe+++y6DBw/u0OO0J7R7wAMtN7RfH8Jsj5ZxsqO2NtTVhrraUFc7PaZtJ6aMWBo/fjwTJ05kwoQJ9O7du9XbtbQM4Zo1a/jVr36VOV9RUcHKlStZvHgxV111Fffffz+nnXYaP/vZz7jtttuYNGkSZWVlfPOb3+RrX/saVVVVTJkyhXPOOafF55w2bRrz589vcv3MmTN5/vnn+fznP08kEmH16tWccMIJTJkyhVdeeSWzF7sjJkyYwIIFC9izZw9Lly7l4osvBuDGG2/M21Qm150WT3fOjXzrrbfeGjlyZM6PFYsl1/PuzD84v+BG3AP/nDzz05/C1VfnvB1hE6SrdIza2lBXG+pqQ13tdOe26dVZSktLu3hLOi+9dzx7D3V3cvfdd3P22WczY8aMJpc3NDQwYsQIpkyZ0mQllKDmzJnDu+++y8svv5y5rPn3devWrYwaNWqU935raPeA19bWAp38h6ZVUNoVqKt0iNraUFcb6mpDXe2orY3utCO3uVtuuYVXXnmF7373u8dcV1hYyDvvvJO35+rsCimhHYAH+oHQALxd3fkfWk+ntjbU1Ya62lBXO2r7l+fWW2/t6k1oVWgH4IHm6GgA3q4es4xTD6S2NtTVhrraUFc7amtDh6IPpntO2OkqB7MOO6oBuIiIiIgYCO0e8PRSQ516x3vwYOPpgQPzvEXhEKirdIja2lBXG+pqQ13tqK0NTe0JJrR7wCORCJFIJ99fHMraA96/f343KCQCdZUOUVsb6mpDXW2oqx21bVRTU0M0GiUajdK3b9/M6f3Zx0NpxZo1a/joo49avO6Xv/wl/fv3z3zgVVoX2p/EQJ9yPnw4+d+yMiguzu8GhYQ+PW5HbW2oqw11taGudtS2UWlpKZWVlUDy6Jfp0x2xZs0aRowYwYknngg0nQO+atUq5syZw/PPP8/06dPztr2JRKLbLnMYVLheTZZOH/GqthaXvr3mf7eqxxxJrAdSWxvqakNdbairHbVtm/eea6+9loqKCiZPnszu3bv56KOPKC8vp6Kigq985Svs2LGDtWvXcuWVV7Jo0aLM/bz31NfXs3v3bm6++WaefvrpzOM+9NBDjB8/nmg0yoYNG6iurmbWrFmUl5czZcoUACZOnJjZaz5jxgx27drFQw89xOzZs5k6dSrr1q3jxhtvpLy8nHHjxvHmm28C8Jvf/IaJEycSjUa5//77uffee1m5ciUA27dvZ968ecczYaeEdg94S4dNbVP2/G8NwFvV6a7SYWprQ11tqKsNdbXTU9om4nGq9+7Ny2OVnXQSBR2cdvPMM88wePBgli9fzpYtW7jnnnu4+OKLueiii7jjjjsye6EnT57MkiVLGD58eJP7r1u3jilTpnDKKadw4MAB6urqqKqq4uGHH2bz5s0UFRWRSCR44IEHuPDCC7nhhhva/Z6UlJSwYsUKIDlILysr49VXX2XZsmU8+uijLFiwgNWrVzNkyBASiQR79+5l/vz5XH755Tz++OPMnTs3WLjjILQD8E5/yEJLEHaIPrxiR21tqKsNdbWhrnZ6StvqvXv5v6k9w7ma9sIL9BkypEO33bZtG6tXr2bDhg0kEgmGDRvGpEmTeOmll5gzZw5jxoxhwYIFx9wvPQVl9erV7Ny5k3Xr1rFr1y6ef/55TjjhBMaOHZuZ/lNQUMD27duZP39+5nz2Y0DTD3WOHTs2c/quu+7ixRdfxHufObJkPB5nSOr1FRQUcPLJJ9PQ0MC+ffuorKzk9ttv73Cr4y20A/BO0wBcREREuljZSScx7YUX8vZYHTV8+HBmz56dmVoSi8WIx+OZg9lMmjSJWbNmUVRURENDQ5P7xmIx3nvvPTZu3AjA7t27+f73v899993Hf/7nfxKPx4lEIiQSCc455xw2bdrE6NGjM3vVBw4cyK5duzjjjDPYtm1b5nHTA/S9e/eyadMmNm/ezG9/+1sWL14MJOf179mzh5NPPjnzWLNnz+b666/nwgsv7Nbzxk0H4M65QuBRYAjwHnA1MBNYANQA87z3uyyeu9PLDWkA3iFaxsmO2tpQVxvqakNd7fSUtgWRSIf3WufTzJkzuf7666moqABg3rx5nH766fzgBz+gsLCQ0047jcGDB3PJJZdw3XXXMWXKFBYuXIj3nrVr13LeeedlHuuTn/wk77zzDgMGDGDu3LlcdNFFlJWV8YMf/IBrrrmGefPm8fTTT9OrVy9eeOEF5s+fz8yZM/nsZz/L4MGDj9m2QYMG0adPH6LRKOPGjctcvmzZMi677DKKi4uZNm0aN910EzNmzOCaa67JvBnorpzl+o3Oua8Cn/Xe/8A59z3gfZKD73JgLHCl9/5bWbcf+dZbb701cuTInJ/76NGjAPTu3btjd/j5z+FrX0uevvZa+NGPct6GMOp0V+kwtbWhrjbU1Ya62unObWtqagAyUyt6kvTe8MLCwi7ekqTa2louvvhiXnzxxa7elGO+r1u3bmXUqFGjvPdbraegnAm8kTr9GrAI2O69rwe2OOfua3b7vnv27OHUU0/NvEOtq6ujqKiISCRCbW0t3vsm10UiEYqKiqirqyORSDS5rqCggFgsRn19PQ0NDcdc16tXr8yfWEr37SO9QFFdWRkFqctjsViT+0EyZPq64uJiCgoKMp/eLSkpIZFIUF9fn9nu9DcgH6+psLCQ4uLiDr2m7Oucc5SUlOT8mpxzFBYWcii1ZnoYXlN3+z4VFxeH7jV15ffJe09RURFHjhwJzWvqDt+n9HUNDQ2heU3d5fuU/j0QptfUHb5PzrnMz2x3e02xWIzCwkIaGhoy86G9901OAx2+Ln3e4rrWnj/9gcpctjvX17R9+3a+9a1vcdNNN5FIJHJ+Tblel0gkiMVimX9PR44cIc16AL4N+DLwNPBXJJc9zDraDc3fLs196qmnmvx5IahYLNapuT8uaxUUr4PwtCoWi5FIJJp8YELyIxaLqauBWCwGdJ+9M2ERi8Uy/2OR/NHvATvpn1n9Lsiv7nQkzOHDh7Nx48Ye8W/IegqKA+4DPgdsBU4BDnjv/zZ1/cve+wlZtx//0ksvvTxhwoQWH68z0nto+/Xr17E7/OlP8PvfJ5cjHDECPv3pnLchjDrdVTpMbW2oqw11taGudrpzW01BCacum4Lik6P77wA45/4BWA3c5ZwrBsYAbza7y+F8/cPo9IcshgxJfkmbuvuHV3oytbWhrjbU1Ya62lFbGz1hb3N3ZL0KymBgBZAA1nvvNzrnHgAqgVqg+x6iSERERETEgPUe8A+BimaX/Rz4ueXzQs9ZbqinUVc7amtDXW2oqw11taO2NrrTHPCepPuuUJ6jgoKCbr0Ae0+lrnbU1oa62lBXG+pqR20b1dTUEI1GiUaj9O3bN3N6//797d53zZo1fPTRRy1e98tf/pL+/ftnVqnpbpYsWcK5556bOThQv379iEajjB07lscee6zV++3YsYOvpZaqnjhxIrW1taxdu5azzz6bn/zkJ4G2JbQfX9c7XBvqakdtbairDXW1oa521LZRaWkplZWVAIwZMyZzuiPWrFnDiBEjOPHEE4Gmc8BXrVrFnDlzeP7555k+fXretjd9lMt8uOeeezj//POJx+OMGDGCyspKjh49yujRo5k7d26HH2fy5Mn8/d//PfF4PNB2hPatYHq5IckvdbWjtjbU1Ya62lBXOz2l7aBBg+jdu3ebXyUlJZSUlLR7u0GDBnX4eb33XHvttVRUVDB58mR2797NRx99RHl5ORUVFXzlK19hx44drF27liuvvDJzyHrvPd576uvr2b17NzfffDNPP/105nEfeughxo8fTzQaZcOGDVRXVzNr1izKy8uZMmUK0LhXGWDGjBns2rWLhx56iNmzZzN16lTWrVvHjTfeSHl5OePGjePNN5NrePzmN79h4sSJRKNR7r//fu69915WrlwJwPbt25k3r2MfNTx8+DD19fUAxONxvv71r1NeXs7UqVOpyj5Seh6Fdg940Hck0jZ1taO2NtTVhrraUFc7PaVtdXU11dXVx/15n3nmGQYPHszy5cvZsmUL99xzDxdffDEXXXQRd9xxR2Yv9OTJk1myZAnDhw9vcv9169YxZcoUTjnlFA4cOEBdXR1VVVU8/PDDbN68maKiIhKJBA888AAXXnghN9xwQ+bgPa0pKSlhxYoVQHKQXlZWxquvvsqyZct49NFHWbBgAatXr2bIkCEkEgn27t3L/Pnzufzyy3n88cfb3aO9bds2ysvLef311/nhD38IwOrVqxk2bBhPPvkkDz/8MMuXL+fyyy/PoWzLQjsA15+abKirHbW1oa421NWGutrpKW3Lysq65LG2bdvG6tWr2bBhA4lEgmHDhjFp0iReeukl5syZw5gxY1iwYMEx90tPQVm9ejU7d+5k3bp17Nq1i+eff54TTjiBsWPHUlSUPNZ4QUEB27dvZ/78+Znz2Y8BTT/UOXbs2Mzpu+66ixdffBHvfWZd7Xg8zpDUEtIFBQWcfPLJNDQ0sG/fPiorK7n99tvbfM0jRoxg48aNbNq0iYceeogrr7ySHTt2ZJ537NixLFu2rMMNOyO0A3ARERGRnmbfvn1d8rzDhw9n9uzZmaklsViMeDzOrbfeCsCkSZOYNWsWRUVFmYPvpMViMd577z02btwIwO7du/n+97/Pfffdl/nAYyQSIZFIcM4557Bp0yZGjx6d2as+cOBAdu3axRlnnMG2bdsyj5seoO/du5dNmzaxefNmfvvb37J48WIAioqK2LNnDyeffHLmsWbPns3111/PhRde2OF541/4whe48847eeeddzjrrLN45ZVXmD59Oq+++iqf+tSncgvbitDOAa+rq8ssOST5o6521NaGutpQVxvqakdt2zZz5kx2795NRUUFFRUVPPHEE5k51uXl5QwZMoTBgwdzySWXcN1113HXXXcByT3Wa9eu5bzzzss81ic/+UneeecdBgwYwNy5c7nooouoqKigsrKSa665hs2bN1NeXs6Xv/xlAObPn8/MmTO56qqrGDx48DHbNmjQIPr06UM0Gm0yv3zZsmVcdtllRKNRHnjgASA5h/zf//3fueKKKzr1+r/97W/z4IMPctlll7Fz507Ky8tZvXo13/72tzvdsiNMD0XfWc65kW+99dZbI0eOzPmxDh8+DEDfvn1zfixppK521NaGutpQVxvqaqc7t9Wh6POntraWiy++mBdffPGY65YtW8bKlStZvnw5559/fk7Ps3btWhYtWsSCBQuYM2dOi7dp61D0oR2Apz/pnJ53JPmhrnbU1oa62lBXG+pqpzu37ckD8PQHKbvDGutvv/023/zmN/nOd77DjBkzunpz2hyAh3YOePrTzt3xH1pPpq521NaGutpQVxvqakdtw2/48OFs3ry5qzejQ0I7AE+/0+2J7ya7M3W1o7Y21NWGutpQVzvduW1RUREHDhzI6wFnjpfutAe8O0kkElRXVzNw4MAWrw/tALynLDfU06irHbW1oa421NWGutrpzm0jkQgnnngi9fX17a6N3d2k/7JQXFzcxVvSvaS/p9lLLDa5/jhvj4iIiIg045zr1m8SWpPe890Tt70rhfbvBVpuyIa62lFbG+pqQ11tqKsdtbWhrsF0tz3gxTt27MjLAx05cgSAPn365OXxJEld7aitDXW1oa421NWO2tpQ145LjXGLofutA94fGAbU5+Hh+gLcVlNGAAAKC0lEQVRzgceAw3l4PElSVztqa0NdbairDXW1o7Y21LXjioF3vfcHu9UecO/9QeD1fDyWc24AsA9423tflY/HFHW1pLY21NWGutpQVztqa0Ndg+lWe8BFRERERMIutB/CFBERERHpjkI7AHfO3e2c2+yce8w5p8NedZJzrr9z7hXn3BHn3KjUZX/tnHvJObfeOXdK6rLhzrlNqcu/2LVb3f0558Y5515ONVvhnCtS1/xwzp2U6rXRObfBOXeyc25i6rJfO+c+k7rdYOfcC865Lc65K7p6u3sK59xs59y+1Gn9zObIOXeGc26fc64y9TVIXfPDORdNNXzROTdTvwfywzk3Ievn9ffOuX9S2xx470P3BXwWeDx1ejEwu6u3qad9AUXAIOARYBTJFXNeJvkBgouAn6Zutwb4FNAP2NLV293dv4CTgdLU6TuBr6pr3toWAgWp01cBS4CNwEDgNOBXqev+CfirrJ/pkq7e9u7+lWq7BnhNvwvy1vQMYHXWeXXNT9dS4FmgOOsy/R7If+dHgHK1Df4V1j3gFwIvpE4/R/KXmXSC9z7mvd+XddGngO3e+3rv/Rbg3NTln/Te/8F7fwjY75w78bhvbA/ivd/jva9Jna0HzkZd88J73+C9Tx9Cri+wE2jw3h/w3n8AfCJ13Thgg/c+DvwHyTeY0rbZwCoggX4X5NNFqb/U3oG65ssEoAZ41jn3C+fcyej3QF4554pJ9vsP1DawsA7ABwKHUqcP0vhDIcFlN4XkHjFo+jOk1h3knDsdmAL8GnXNG+fcaOfcb4H/AbxE07bx1P84irIG6mrbDudcIXA58PPURfpdkB97gLOALwD/DfgK6poPJ5HsOhX4X8Ct6PdAvv0VsB7oj9oGFtYBeBXJP9dB8gdkfxduS1hkNwVoSP03kXWZWneAc64fyfVSryK5dJO65on3/g3v/QXAD0hOP8tuG/He1wMx51z6d5/atu8KYGXW/1D1uyAPvPd13vujPvk3+zUkp06qa+6qSE7VqSc5SPwc+j2Qb39N8i9izX8XqG0nhHUA/hLJd2gAXwK2dOG2hMUfgHOcc8XOuQuBN1OX73HOnemc6wt8wnv/UddtYvfnnIsATwG3eu/fQV3zJrXnJe0gcASIOOcGOOdOpfF/Aq8C0dT34nxg6/Hd0h5nBHClc+45ktMkrkM/szlLdUr7PPBL1DUfXiXZ0QGjgW3o90DeuOSiFmOBX3vvq1HbwEK7Drhz7l5gPPAB8Depd2XSCc65X5H8BfZH4Kck59XdANQC87z3/+WcG5G6rhC4xXu/tqu2tydwzs0FHgD+X+qiH6f+q645cs6NA+4jueewFvhbkgPGuwAPXOu9/11qTujPgN7AT7z3P+uiTe5xnHP/4b0f45ybhX5mc+Kcuxj4R6AaeI/kz+tlqGvOnHPzgVkk/93/LTAE/R7Ii9TP7cXe++tT57+A2gYS2gG4iIiIiEh3FNYpKCIiIiIi3ZIG4CIiIiIix5EG4CIiIiIix5EG4CIix5lz7h+cc/+ni7dhvHNum3PusHPu+q7clrY4537inLu7q7dDRCSfNAAXkb9ozrlK51yDc+7crMsGOOe8c+6Mrtsyc/8IrPDe9/Xe/0vzK5u/SUj1GG25QS29MfHeX+O9/77l84qIHG8agIuIwAHgzq7eiCBcUmH7tzzGUBqXwzSXWj9YRETQAFxEBGA5cFFqTdtjOOcecc49kHW+yR7y1PX/2zm32jl3xDm31Tk3yjn3LefcLufcPufctc0eNpK6zyHn3B+cczOzHt855653zr3tnKtK7aU/J+v6951zi5xzvyG5jvSIFra5r3PuX51ze1JfP3HO9U5d9yHJAfiK1PZ+uq04zrlXUidfSt3+5tTlZzrnnk29vj8655akj4DnnLvKOfeGc+7W1PM95Zzr45x7xjn3Z+fcQefcJufcZ1O3nwHcDPz31HMcaaX9GOfcllSXbc652VnX/UNqe36Yuv6D1Jrl6esnO+feTE272eucS6/DLyJyXGkALiKSPILb3SQPKBHUXwP/BAwgeSS4Z4AzgWHA14B/cs6dlHX7LwOvAJ8AbiI5GD4zdd23gb8DpgInkjxU+bOu6dE+rwLmAX2Ad1rYnn8GzgJGAZ8Bhqe2D+/9YJIHKZvtve/jvf99Wy/Mez8udfLC1O3vcM6VkTzU93qSBzr5fOp1/k3WXUcBceA0YC7J/+c8SXLwfxLwOrDSOee89/8HuAP4t9Rz9Gm+Hc65AcBzJI8mOyjV6X855y7KutmXgE3ACcAS4CHXeNTJR4F7vfd9SX5fHmvrdYuIWNEAXEQk6QHg9NSe2CB+6b3f4r2PAyuBM0geubDee78eOEhyIJz2e+/9T733ce/9s8CLQHpv7nzgf3rv/5C6/l+AUuCCrPv/2Hv/jve+ofmRflN7oecAi7z3H6cOX34zyUPK5+v3/qXAAe/9A6nX+AHJQf/Xs25zEFiaur7ae3/Ie/9z7/1R730tcAvwaeCTnXjOfd77B733Me/9RpID+nlZt3nNe7/Se99AcoBdnHoOgBhwlnNuUGobXgr64kVEcqEBuIgI4L2vAW4luRc2yJzqvVmnq4HDqcfMvix7r+4fm93/jyT3JENy8P54ahpFlXOuChgInJJ1+w/a2JZBJAee72dd9i7Qi+Qe9Xw4AxjVbBuXAYOzbvMn730ifcY5V+qcW56aQnMoa/s6uk2n0PQ1QfJ1ZXf5MH3CJw/1XAOk94DPJLlX/h3n3OvOucs7+LwiInmlAbiISKP/TfL34rxmlx8ByrLOn5yH5zq92fnTgD+lTv8X8Nfe+wFZX2Xe+xVZt0/Qun1APclBctoZQB3wUcDt9c3O/xfwn822sZ/3fmQb2/gd4Hxgove+X9b2uVZu39wumr4mUud3tb/54L1/zXt/GckB/+3Ak82mBYmIHBcagIuIpKSmLSwmOV0j22vAl5xzJ6fmE9+Sh6f7tHPum865iHPuUmAS8PPUdT8CbnPOnQ3gnOvnnJueNZe5vdeRIDk1Y6lz7hPOuRNI7tl/LHuPdCftJTmnPe3fgJOcc9c650qcc4XOubOdc9E2HqMfUAsccM71SW1T8+c43TkXaeX+vwL+W+o5I865z5OcavOz9jbeOVfsnJvrnBuYalCVuire3n1FRPJNA3ARkSze+6eBHc0ufhzYCLwNvAH8Mg9P9RwwnuQHQP8ZuMJ7/4fUdT8EHgHWpKZqbKfp3OqOuIHkdI1twFaSr+mmHLb3B8C/OOcOOOcWeu+PAH8FfDH1PB+THPQPbv0huB9oIDnQfgt4udn1q4BDwL7UlJYmvPcHgIuBK1LP96/At733v+7ga/g6sMM5dxh4EPi69/7jDt5XRCRvXHKKnIiIiIiIHA/aAy4iIiIichxpAC4iIiIichxpAC4iIiIichxpAC4iIiIichz9f5tP1IP/Cs2lAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "'''PLOTTTING THE ACCURACY GRAPH'''\n",
    "\n",
    "plt.figure(figsize=[30,20], dpi=30)\n",
    "\n",
    "plt.plot(iter_list, train_acc_list, '-', lw=3, c='salmon', label='Train Accuracy')\n",
    "plt.plot(iter_list, train_roll_acc_list, '-|r', lw=7, label='Train Accuracy [Roll]')\n",
    "\n",
    "plt.plot(iter_list, test_acc_list, '-', lw=3, c='brown', label='Test Accuracy')\n",
    "plt.plot(iter_list, test_roll_acc_list, '-|k', lw=7, label='Test Accuracy [Roll]')\n",
    "\n",
    "plt.title('ACCURACY vs ITERATIONS', size='30')\n",
    "plt.xlabel('Number of Iterations', size='30')\n",
    "plt.ylabel('Accuracy', size='30')\n",
    "\n",
    "\n",
    "# plt.yticks(range(91,101,10), rotation='horizontal')\n",
    "# plt.xticks(range(0,len(iter_list)+1,inspect_size), rotation='horizontal')\n",
    "\n",
    "plt.grid(True, linestyle='-.')\n",
    "plt.tick_params(labelcolor='k', labelsize='20', width=3)\n",
    "\n",
    "plt.legend(fontsize='20')\n",
    "fig1 = plt.gcf()\n",
    "plt.show()\n",
    "\n",
    "fig1.savefig('acc_vs_iterations[margin=1].png', dpi=75)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "vps3zKEnsKX9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy Analysis : \n",
      "====================\n",
      "[Average]\n",
      "--- Maximum accuracy on test-set  : 99.90\n",
      "--- Iteration at maximum accuracy : 450\n",
      "[Absolute]\n",
      "--- Maximum accuracy on test-set  : 100.00\n",
      "--- Iteration at maximum accuracy : 105\n",
      "\n",
      "Loss Analysis : \n",
      "====================\n",
      "[Average]\n",
      "--- Minimum loss on test-set  : 0.00\n",
      "--- Iteration at minimum loss : 450\n",
      "[Absolute]\n",
      "--- Minimum  loss on test-set  : 0.00\n",
      "--- Iteration at minimum loss : 105\n"
     ]
    }
   ],
   "source": [
    "'''ANALYZING ACCURACY AND LOSS'''\n",
    "\n",
    "### ACCURACY ANALYSIS\n",
    "print('Accuracy Analysis : ')\n",
    "print('====================')\n",
    "\n",
    "# Getting maximum accuracy\n",
    "test_max_acc = np.max(test_acc_list)\n",
    "test_roll_max_acc = np.max(test_roll_acc_list)\n",
    "\n",
    "print(\"[Average]\")\n",
    "print('--- Maximum accuracy on test-set  : {:.2f}'.format(test_roll_max_acc))\n",
    "print('--- Iteration at maximum accuracy : {}'.format(iter_list[test_roll_acc_list.index(test_roll_max_acc)]))\n",
    "print(\"[Absolute]\")\n",
    "print('--- Maximum accuracy on test-set  : {:.2f}'.format(test_max_acc))\n",
    "print('--- Iteration at maximum accuracy : {}'.format(iter_list[test_acc_list.index(test_max_acc)]))\n",
    "\n",
    "### LOSS ANALYSIS\n",
    "print('\\nLoss Analysis : ')\n",
    "print('====================')\n",
    "\n",
    "# Getting minimum loss\n",
    "test_min_loss = np.min(test_loss_list)\n",
    "test_roll_min_loss = np.min(test_roll_loss_list)\n",
    "\n",
    "print(\"[Average]\")\n",
    "print('--- Minimum loss on test-set  : {:.2f}'.format(test_roll_min_loss))\n",
    "print('--- Iteration at minimum loss : {}'.format(iter_list[test_roll_loss_list.index(test_roll_min_loss)]))\n",
    "print(\"[Absolute]\")\n",
    "print('--- Minimum  loss on test-set  : {:.2f}'.format(test_min_loss))\n",
    "print('--- Iteration at minimum loss : {}'.format(iter_list[test_loss_list.index(test_min_loss)]))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "yEKvVWfaCIYM"
   },
   "source": [
    "# SAMPLE ANALYSIS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "FG8WAInKxt42"
   },
   "outputs": [],
   "source": [
    "'''MODEL TESTS AND ANALYSIS'''\n",
    "\n",
    "def process(matrix_path):\n",
    "\n",
    "    \"\"\"\n",
    "    Processes an array of images.\n",
    "    Takes their path as input and returns thier encoding.\n",
    "    Helper function for `assess`.\n",
    "    \"\"\"\n",
    "    \n",
    "    matrix = get_image_from_path(matrix_path)\n",
    "    matrix = cv2.resize(matrix, (global_image_width, global_image_height))\n",
    "    matrix = np.rollaxis(matrix, 2, 0)\n",
    "    matrix = matrix.reshape(1, 3, global_image_height, global_image_width)        \n",
    "    matrix = torch.from_numpy(matrix)\n",
    "    matrix_enc = get_encodings(matrix)\n",
    "\n",
    "    return matrix_enc\n",
    "\n",
    "\n",
    "def assess(anchor_path, image1_path, image2_path, threshold=triplet_loss_margin):\n",
    "    \n",
    "    \"\"\"\n",
    "    Takes in three image paths  : anchor - Image corresponding to data in database.\n",
    "                                : image1 - First input image\n",
    "                                : image2 - Second input image\n",
    "    \n",
    "    and prints which among image1 and image2 is close to anchor image.\n",
    "    \n",
    "    If difference is less than `threshold` then notifies accordingly.\n",
    "    \n",
    "    \"\"\"\n",
    "    \n",
    "    anchor_enc = process(anchor_path)               \n",
    "    image1_enc = process(image1_path)               \n",
    "    image2_enc = process(image2_path)               \n",
    "\n",
    "    image1_dist_list, image2_dist_list = return_diff(anchor_enc, image1_enc, image2_enc)\n",
    "    image1_dist = image1_dist_list[0]\n",
    "    image2_dist = image2_dist_list[0]\n",
    "    \n",
    "    diff = image1_dist - image2_dist\n",
    "        \n",
    "    # Return output\n",
    "    \n",
    "    print(\"=======================================================\")   \n",
    "    \n",
    "    # Found forgery\n",
    "    if(abs(diff) >= threshold):\n",
    "        if(image1_dist < image2_dist):\n",
    "            print(\"VERDICT : Second signature appears to forged.\")\n",
    "        else:\n",
    "            print(\"VERDICT : First signature appears to forged.\")\n",
    "    \n",
    "    # Cannot finnd forgery\n",
    "    else:\n",
    "        print(\"VERDICT : Unable to determine authenticity.\")\n",
    "        print(\"\\tEither provide different samples or lower threshold.\")\n",
    "        print(\"\\tAlso, please cross-check the Achor provided.\")\n",
    "\n",
    "    print(\"---> Current Threshold \\t\\t\\t : {:.4f}\".format(threshold))\n",
    "    print(\"---> Difference between Encodings \\t : {:.4f}\".format(abs(diff)))        \n",
    "    print(\"---> Distance between Anchor and Image_1 : {:.4f}\".format(image1_dist))\n",
    "    print(\"---> Distance between Anchor and Image_2 : {:.4f}\".format(image2_dist))\n",
    "\n",
    "    print(\"=======================================================\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 680
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1909,
     "status": "ok",
     "timestamp": 1562562178239,
     "user": {
      "displayName": "Aditya Ahuja",
      "photoUrl": "",
      "userId": "08809216472288903233"
     },
     "user_tz": -330
    },
    "id": "I_7-Hq21Jnka",
    "outputId": "1d228a53-d03c-4c23-bb23-c8631c0b1c85"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=======================================================\n",
      "VERDICT : Second signature appears to forged.\n",
      "---> Current Threshold \t\t\t : 1.0000\n",
      "---> Difference between Encodings \t : 4.5911\n",
      "---> Distance between Anchor and Image_1 : 7.9553\n",
      "---> Distance between Anchor and Image_2 : 12.5464\n",
      "=======================================================\n",
      "=======================================================\n",
      "VERDICT : Second signature appears to forged.\n",
      "---> Current Threshold \t\t\t : 1.0000\n",
      "---> Difference between Encodings \t : 4.7556\n",
      "---> Distance between Anchor and Image_1 : 7.2346\n",
      "---> Distance between Anchor and Image_2 : 11.9902\n",
      "=======================================================\n",
      "=======================================================\n",
      "VERDICT : First signature appears to forged.\n",
      "---> Current Threshold \t\t\t : 1.0000\n",
      "---> Difference between Encodings \t : 1.1851\n",
      "---> Distance between Anchor and Image_1 : 9.4213\n",
      "---> Distance between Anchor and Image_2 : 8.2362\n",
      "=======================================================\n",
      "=======================================================\n",
      "VERDICT : First signature appears to forged.\n",
      "---> Current Threshold \t\t\t : 1.0000\n",
      "---> Difference between Encodings \t : 9.9845\n",
      "---> Distance between Anchor and Image_1 : 15.7653\n",
      "---> Distance between Anchor and Image_2 : 5.7808\n",
      "=======================================================\n",
      "=======================================================\n",
      "VERDICT : Unable to determine authenticity.\n",
      "\tEither provide different samples or lower threshold.\n",
      "\tAlso, please cross-check the Achor provided.\n",
      "---> Current Threshold \t\t\t : 1.0000\n",
      "---> Difference between Encodings \t : 0.2178\n",
      "---> Distance between Anchor and Image_1 : 18.2131\n",
      "---> Distance between Anchor and Image_2 : 17.9953\n",
      "=======================================================\n"
     ]
    }
   ],
   "source": [
    "# TESTS :\n",
    "# Note : Format from dataset is [Anchor, Real, Fake]\n",
    "\n",
    "threshold = triplet_loss_margin\n",
    "\n",
    "# Test1 -- Signature2 fake.\n",
    "index = test_master_index[0]\n",
    "assess(index[0], index[1], index[2])\n",
    "\n",
    "# Test1 -- Signature2 fake.\n",
    "index = test_master_index[432]\n",
    "assess(index[0], index[1], index[2])\n",
    "\n",
    "# Test1 -- Signature1 fake.\n",
    "index = test_master_index[100]\n",
    "assess(index[0], index[2], index[1])\n",
    "\n",
    "# Test1 -- Signature1 fake.\n",
    "index = test_master_index[2313]\n",
    "assess(index[1], index[2], index[0])\n",
    "\n",
    "# Test1 -- Incorrect Anchor provided. Should complain.\n",
    "index = test_master_index[15353]\n",
    "assess(index[2], index[1], index[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "RrHFA982xt2O"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "RYL95nDIzj5m"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "JZ2wgGYTsLDp"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "dB2b9VgwsLKk"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "0BtZaOcSsLPi"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "5D4K-N5rsLU-"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "IKx9c5zlsLS3"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "yeQxp1s0sLNV"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 289
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 5628,
     "status": "ok",
     "timestamp": 1562543575442,
     "user": {
      "displayName": "Aditya Ahuja",
      "photoUrl": "",
      "userId": "08809216472288903233"
     },
     "user_tz": -330
    },
    "id": "2rrlQ2Jv1YdD",
    "outputId": "21e978d1-309e-4665-f768-4e375f9fc991"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1570,
     "status": "ok",
     "timestamp": 1562561604362,
     "user": {
      "displayName": "Aditya Ahuja",
      "photoUrl": "",
      "userId": "08809216472288903233"
     },
     "user_tz": -330
    },
    "id": "oXsgoWsZsKO8",
    "outputId": "571afa11-c0a8-447f-a765-353ba910ac08"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "-ZaMk8fQhgyq"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "RQDUhIs8bkcy"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "Notebook.ipynb",
   "provenance": [],
   "version": "0.3.2"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}