notebook.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Computer-Vision-Final.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Krntud63UOUm"
      },
      "source": [
        "# Introduction"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "29RWqqVZUSPY"
      },
      "source": [
        "Now, more than ever, algorithms are responsible for life-changing choices\r\n",
        "such as recruitment or loan allocation. Therefore, the question of algorithmic\r\n",
        "fairness is central if we wish to make our society fairer and more inclusive.\r\n",
        "\r\n",
        "\r\n",
        "This project aims at measuring the fairness of a neural network used to\r\n",
        "classify pictures of humans. More specifically, we’re interested in knowing if\r\n",
        "there are biases with respect to protected categories such as gender or age and\r\n",
        "finding ways to alleviate those biases.\r\n",
        "\r\n",
        "\r\n",
        "We’re building an attractiveness classifier, i.e a classifier that should separate human images into two categories : attractive or unattractive. This classifier is going to be evaluated in terms of biases with respect to protected categories in order to measure its fairness.\r\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YnR0gq_ROUlT"
      },
      "source": [
        "**Importante note :** This notebook contains only the source code of the project. For explanations please look at the report."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9J0DLDLhUXY3"
      },
      "source": [
        "# Imports"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FuzsoKdzUghM"
      },
      "source": [
        "## Mounting the drive"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IUCZVAmcUIrP",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d519e315-ca31-4d61-c654-70ffbd68233d"
      },
      "source": [
        "#mounting the drive\r\n",
        "from google.colab import drive\r\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HspxpZf0UlnH"
      },
      "source": [
        "#drive path\r\n",
        "drive_path = \"drive/My Drive/DTU/Deep Learning/\""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Po568zVhUu8k"
      },
      "source": [
        "## Importing the images"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pzS2ghOfUx85"
      },
      "source": [
        "#unzipping the images from the drive into the local environment\r\n",
        "import zipfile\r\n",
        "with zipfile.ZipFile(drive_path+\"ProjetDL/images.zip\",\"r\") as zip_ref:\r\n",
        "  zip_ref.extractall(\"lesimages/\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bZo8jtMUU4n2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d44f07d9-d68f-4e46-8abe-d481031ba1f3"
      },
      "source": [
        "#checking the loading\r\n",
        "import os\r\n",
        "root = 'lesimages/img_align_celeba'\r\n",
        "img_list = os.listdir(root)\r\n",
        "print(len(img_list))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "202599\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YR3eRU4_VAdR"
      },
      "source": [
        "## Other imports needed"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "at1BO8xcVEE7"
      },
      "source": [
        "import matplotlib.pyplot as plt\r\n",
        "import pandas as pd\r\n",
        "import numpy as np\r\n",
        "import torch\r\n",
        "import torchvision\r\n",
        "import torchvision.transforms as transforms\r\n",
        "from torch.utils import data\r\n",
        "from torchvision import transforms as T\r\n",
        "from torchvision.datasets import ImageFolder\r\n",
        "from PIL import Image\r\n",
        "import torch\r\n",
        "import os\r\n",
        "import random\r\n",
        "from torch.autograd import Variable\r\n",
        "import torch.nn as nn\r\n",
        "import torch.nn.functional as F\r\n",
        "import torch.optim as optim\r\n",
        "import torch.nn.init as init\r\n",
        "from torch.nn import Linear, Conv2d, BatchNorm2d, MaxPool2d, Dropout2d\r\n",
        "from torch.nn.functional import relu, elu, relu6, sigmoid, tanh, softmax\r\n",
        "import torch.optim as optim\r\n",
        "from google.colab.patches import cv2_imshow\r\n",
        "import cv2"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QIiPiDoxVQsx"
      },
      "source": [
        "# Creating the sets"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LmtJpAwvVYQt"
      },
      "source": [
        "## The DataLoader"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gJsc5Y0iVbzu"
      },
      "source": [
        "**Note :** The code for this dataloader is heavily inspired by https://github.com/yunjey/stargan/blob/master/data_loader.py"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2J195GWeVfFu"
      },
      "source": [
        "class CelebA(data.Dataset):\r\n",
        "    \"\"\"Dataset class for the CelebA dataset.\"\"\"\r\n",
        "\r\n",
        "    def __init__(self, image_dir, attr_path,  split_path, selected_attrs, transform, mode):\r\n",
        "        \"\"\"Initialize and preprocess the CelebA dataset.\"\"\"\r\n",
        "        self.image_dir = image_dir\r\n",
        "        self.attr_path = attr_path\r\n",
        "        self.split_path = split_path\r\n",
        "        self.selected_attrs = selected_attrs\r\n",
        "        self.transform = transform\r\n",
        "        self.mode = mode\r\n",
        "        self.train_dataset = []\r\n",
        "        self.train_dataset_with_added_bald = []\r\n",
        "        self.test_dataset = []\r\n",
        "        self.valid_dataset = []\r\n",
        "        self.all_dataset = []\r\n",
        "        self.all_dataset_with_added_bald = []\r\n",
        "        self.attr2idx = {}\r\n",
        "        self.idx2attr = {}\r\n",
        "        self.preprocess()\r\n",
        "\r\n",
        "        if mode == 'train':\r\n",
        "            self.num_images = len(self.train_dataset)\r\n",
        "        elif mode == 'valid':\r\n",
        "            self.num_images = len(self.valid_dataset)\r\n",
        "        else:\r\n",
        "            self.num_images = len(self.test_dataset)\r\n",
        "\r\n",
        "    def preprocess(self):\r\n",
        "        \"\"\"Preprocess the CelebA attribute file.\"\"\"\r\n",
        "        dict_split = dict()\r\n",
        "        lines = [line.rstrip() for line in open(self.split_path, 'r')]\r\n",
        "        for line in lines:\r\n",
        "          split = line.split()\r\n",
        "          dict_split[split[0]]=int(split[1])\r\n",
        "\r\n",
        "        lines = [line.rstrip() for line in open(self.attr_path, 'r')]\r\n",
        "        all_attr_names = lines[1].split()\r\n",
        "        for i, attr_name in enumerate(all_attr_names):\r\n",
        "            self.attr2idx[attr_name] = i\r\n",
        "            self.idx2attr[i] = attr_name\r\n",
        "\r\n",
        "        lines = lines[2:]\r\n",
        "        random.seed(1234)\r\n",
        "        random.shuffle(lines)\r\n",
        "        for i, line in enumerate(lines):\r\n",
        "            split = line.split()\r\n",
        "            filename = split[0]\r\n",
        "            values = split[1:]\r\n",
        "\r\n",
        "            label = []\r\n",
        "            all_label = []\r\n",
        "\r\n",
        "            for attr_name in self.attr2idx.keys():\r\n",
        "                idx = self.attr2idx[attr_name]\r\n",
        "                all_label.append(values[idx] == '1')\r\n",
        "            for attr_name in self.selected_attrs:\r\n",
        "                idx = self.attr2idx[attr_name]\r\n",
        "                label.append(values[idx] == '1')\r\n",
        "\r\n",
        "            if (self.mode == 'train'):\r\n",
        "              if (dict_split[filename] == 0):\r\n",
        "\r\n",
        "##### --- Uncomment the following lines in order to introduce data into de sets (repetition of images with a chosen attribute) --- #####\r\n",
        "                # if all_label[4] == True:\r\n",
        "                #   for i in range(1,10):\r\n",
        "                #     im = Image.open(\"lesimages/img_align_celeba/\"+filename) \r\n",
        "                #     im.save(\"lesimages/img_align_celeba/new_image_\"+filename.split(\".\")[0]+str(i)+\".png\")\r\n",
        "                #     self.train_dataset.append([\"new_image_\"+filename.split(\".\")[0]+str(i)+\".png\",label])\r\n",
        "                #     self.all_dataset.append([\"new_image_\"+filename.split(\".\")[0]+str(i)+\".png\", all_label])\r\n",
        "\r\n",
        "                self.train_dataset.append([filename, label])\r\n",
        "                self.all_dataset.append([filename, all_label])\r\n",
        "            elif (self.mode == 'valid'):\r\n",
        "              if (dict_split[filename] == 1):\r\n",
        "                self.valid_dataset.append((filename, label))\r\n",
        "                self.all_dataset.append([filename, all_label])\r\n",
        "            else:\r\n",
        "              if (dict_split[filename] == 2):\r\n",
        "                self.test_dataset.append((filename, label))\r\n",
        "                self.all_dataset.append([filename, all_label])\r\n",
        "\r\n",
        "        print('Finished preprocessing the CelebA dataset...')\r\n",
        "\r\n",
        "    def __getitem__(self, index):\r\n",
        "        \"\"\"Return one image and its corresponding attribute label.\"\"\"\r\n",
        "        dataset = self.train_dataset if self.mode == 'train' else self.valid_dataset if self.mode == 'valid' else self.test_dataset\r\n",
        "        filename, label = dataset[index]\r\n",
        "        image = Image.open(os.path.join(self.image_dir, filename))\r\n",
        "        return self.transform(image), torch.FloatTensor(label)\r\n",
        "\r\n",
        "    def __len__(self):\r\n",
        "        \"\"\"Return the number of images.\"\"\"\r\n",
        "        return self.num_images\r\n",
        "\r\n",
        "\r\n",
        "def get_loader(image_dir, attr_path, selected_attrs,  split_path, crop_size=178, image_size=128, \r\n",
        "               batch_size=16, dataset='CelebA', mode='train', num_workers=1):\r\n",
        "    \"\"\"Build and return a data loader.\"\"\"\r\n",
        "    transform = []\r\n",
        "    if mode == 'train':\r\n",
        "        transform.append(T.RandomHorizontalFlip())\r\n",
        "    transform.append(T.CenterCrop(crop_size))\r\n",
        "    transform.append(T.Resize(image_size))\r\n",
        "    transform.append(T.ToTensor())\r\n",
        "    transform.append(T.Normalize(mean=(0.5, 0.5, 0.5), std=(0.5, 0.5, 0.5)))\r\n",
        "    transform = T.Compose(transform)\r\n",
        "\r\n",
        "    if dataset == 'CelebA':\r\n",
        "        dataset = CelebA(image_dir, attr_path, split_path, selected_attrs, transform, mode)\r\n",
        "    elif dataset == 'RaFD':\r\n",
        "        dataset = ImageFolder(image_dir, transform)\r\n",
        "\r\n",
        "    data_loader = data.DataLoader(dataset=dataset,\r\n",
        "                                  batch_size=batch_size,\r\n",
        "                                  shuffle=(mode=='train'),\r\n",
        "                                  num_workers=num_workers)\r\n",
        "    return data_loader, dataset"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "w3RSHPoWV9wb"
      },
      "source": [
        "## Creating the sets"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LuEv_xv2WaoG"
      },
      "source": [
        "### Common variables"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5v1sqe2pWdQk"
      },
      "source": [
        "#image directiory :\r\n",
        "celeba_image_dir = \"lesimages/img_align_celeba/\"\r\n",
        "#split\r\n",
        "split_path = drive_path+\"ProjetDL/list_eval_partition.txt\"\r\n",
        "#attributes\r\n",
        "attr_path = drive_path+\"ProjetDL/list_attr_celeba.txt\"\r\n",
        "#selected attribute :\r\n",
        "#selected_attrs = ['Attractive']\r\n",
        "selected_attrs = [\"Attractive\"]\r\n",
        "#crop size\r\n",
        "celeba_crop_size = 178\r\n",
        "#image size\r\n",
        "image_size = 128\r\n",
        "#batch_size\r\n",
        "batch_size = 24\r\n",
        "#num workers\r\n",
        "num_workers = 1"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "V_gGUOGhWf2l"
      },
      "source": [
        "### Train set"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YfUM1yS2Wi5l",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "79ea83c2-257e-4b52-d910-e034691a5e94"
      },
      "source": [
        "mode = 'train'\r\n",
        "trainloader, trainset = get_loader(celeba_image_dir, attr_path, selected_attrs, split_path,\r\n",
        "                                   celeba_crop_size, image_size, batch_size,\r\n",
        "                                   'CelebA', mode, num_workers)\r\n",
        "\r\n",
        "trainset_allattr = trainset.all_dataset"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Finished preprocessing the CelebA dataset...\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gz_Z1v2ZWm2n"
      },
      "source": [
        "### Validation set"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "j_PhIWCpWqIe",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8674aa69-14b0-4bf0-c39b-b93b2093620c"
      },
      "source": [
        "mode = 'valid'\r\n",
        "\r\n",
        "validloader, validset = get_loader(celeba_image_dir, attr_path, selected_attrs, split_path,\r\n",
        "                                   celeba_crop_size, image_size, batch_size,\r\n",
        "                                   'CelebA', mode, num_workers)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Finished preprocessing the CelebA dataset...\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DwLYLCNJWxQo"
      },
      "source": [
        "### Test set"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3yrOgOfuW2Wu",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "16fad0e7-23df-4431-f739-b2403d98bacb"
      },
      "source": [
        "mode = 'test'\r\n",
        "\r\n",
        "testloader, testset = get_loader(celeba_image_dir, attr_path, selected_attrs, split_path,\r\n",
        "                                   celeba_crop_size, image_size, batch_size,\r\n",
        "                                   'CelebA', mode, num_workers)\r\n",
        "\r\n",
        "testset_allattr = testset.all_dataset"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Finished preprocessing the CelebA dataset...\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UXCMrgv0Xl6b"
      },
      "source": [
        "## Printing some images\r\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "d0NyP5yFXou4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 216
        },
        "outputId": "e4ace7d2-ac30-42b3-bae7-2b7e04761847"
      },
      "source": [
        "train_data_iter = iter(trainloader)\r\n",
        "test_data_iter = iter(testloader)\r\n",
        "\r\n",
        "classes = ['Unattractive', 'Attractive']\r\n",
        "def imshow(img):\r\n",
        "    \"\"\" show an image \"\"\"\r\n",
        "    img = img / 2 + 0.5 # unnormalize\r\n",
        "    npimg = img.numpy()\r\n",
        "    plt.imshow(np.transpose(npimg, (1, 2, 0)))\r\n",
        "\r\n",
        "images, labels = train_data_iter.next()\r\n",
        "imshow(torchvision.utils.make_grid(images))\r\n",
        "\r\n",
        "print('  '.join('%5s' % classes[int(labels[j].item())] for j in range(batch_size)))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Unattractive  Attractive  Unattractive  Attractive  Attractive  Attractive  Attractive  Unattractive  Unattractive  Unattractive  Attractive  Unattractive  Unattractive  Unattractive  Unattractive  Unattractive  Unattractive  Unattractive  Unattractive  Unattractive  Attractive  Unattractive  Unattractive  Attractive\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9Mj4NrV1XFpB"
      },
      "source": [
        "# Creating the model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AVoYYN8nXe7u"
      },
      "source": [
        "## Model variables"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ytLbvk4pX2U2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4d524762-ab2a-4565-dbe2-3614df4cbe0d"
      },
      "source": [
        "#getting the dimensions\r\n",
        "x, y = next(iter(trainloader))\r\n",
        "print(\"Batch dimension [B x C x H x W]:\", x.shape)\r\n",
        "channels_from_iter = x.shape[1]\r\n",
        "height_from_iter = x.shape[2]\r\n",
        "width_from_iter = x.shape[3]\r\n",
        "\r\n",
        "#hyperparameters for the model\r\n",
        "num_classes = len(selected_attrs)\r\n",
        "channels = channels_from_iter\r\n",
        "height = height_from_iter\r\n",
        "width = width_from_iter\r\n",
        "\r\n",
        "#first convolution\r\n",
        "num_filters_conv1 = 96\r\n",
        "kernel_size_conv1 = 3\r\n",
        "stride_conv1 = 1 \r\n",
        "padding_conv1 = 1\r\n",
        "\r\n",
        "#second convolution\r\n",
        "num_filters_conv2 = 128\r\n",
        "kernel_size_conv2 = 4 \r\n",
        "stride_conv2 = 2\r\n",
        "padding_conv2 = 2\r\n",
        "\r\n",
        "#hidden layer\r\n",
        "num_l1 = 96"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Batch dimension [B x C x H x W]: torch.Size([24, 3, 128, 128])\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Mhq2NjRJYKt8"
      },
      "source": [
        "## Defining the model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "T_Yi9eoZYMiG",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "894cf9b0-4164-4a73-f72b-6d5071b14864"
      },
      "source": [
        "def compute_conv_dim(dim_size):\r\n",
        "    return int((dim_size - kernel_size_conv2 + 2 * padding_conv2) / stride_conv2 + 1)\r\n",
        "\r\n",
        "class Net(nn.Module):\r\n",
        "    def __init__(self, num_classes):\r\n",
        "        super(Net, self).__init__()\r\n",
        "        self.num_classes = num_classes\r\n",
        "        \r\n",
        "        #first convolution\r\n",
        "        self.conv_1 = Conv2d(in_channels=channels,\r\n",
        "                             out_channels=num_filters_conv1,\r\n",
        "                             kernel_size=kernel_size_conv1,\r\n",
        "                             stride=stride_conv1,\r\n",
        "                             padding=padding_conv1)\r\n",
        "        \r\n",
        "        self.batch_norm1 = BatchNorm2d(num_filters_conv1)\r\n",
        "        \r\n",
        "\r\n",
        "        self.conv_2 = Conv2d(in_channels=num_filters_conv1,\r\n",
        "                             out_channels=num_filters_conv2,\r\n",
        "                             kernel_size=kernel_size_conv2,\r\n",
        "                             stride=stride_conv2,\r\n",
        "                             padding=padding_conv2)\r\n",
        "        \r\n",
        "        self.batch_norm2 = BatchNorm2d(num_filters_conv2)\r\n",
        "        \r\n",
        "        self.conv_out_height = compute_conv_dim(height)\r\n",
        "        self.conv_out_width = compute_conv_dim(width)\r\n",
        "\r\n",
        "\r\n",
        "        self.l1_in_features = num_filters_conv2 * self.conv_out_height * self.conv_out_width\r\n",
        " \r\n",
        "\r\n",
        "        #first layer\r\n",
        "        self.l_1 = Linear(in_features=self.l1_in_features,\r\n",
        "                          out_features=num_l1,\r\n",
        "                          bias=True)\r\n",
        "        \r\n",
        "        self.dropout = Dropout2d(p=0.3)\r\n",
        "\r\n",
        "        #output layer\r\n",
        "        self.l_out = Linear(in_features=num_l1, \r\n",
        "                            out_features=num_classes,\r\n",
        "                            bias=False)\r\n",
        "\r\n",
        "    def forward(self, x):\r\n",
        "      x = relu(self.conv_1(x))\r\n",
        "      x = self.batch_norm1(x)\r\n",
        "      x = relu(self.conv_2(x))\r\n",
        "      x = self.batch_norm2(x)\r\n",
        "      x = x.view(-1, self.l1_in_features)\r\n",
        "      x = relu(self.l_1(x))\r\n",
        "      x = self.dropout(x)\r\n",
        "      return torch.sigmoid(self.l_out(x))\r\n",
        "    \r\n",
        "\r\n",
        "net = Net(num_classes)\r\n",
        "net.cuda()\r\n",
        "print(net)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Net(\n",
            "  (conv_1): Conv2d(3, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
            "  (batch_norm1): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
            "  (conv_2): Conv2d(96, 128, kernel_size=(4, 4), stride=(2, 2), padding=(2, 2))\n",
            "  (batch_norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
            "  (l_1): Linear(in_features=540800, out_features=96, bias=True)\n",
            "  (dropout): Dropout2d(p=0.3, inplace=False)\n",
            "  (l_out): Linear(in_features=96, out_features=1, bias=False)\n",
            ")\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VPCEPMRnYa5m"
      },
      "source": [
        "# Training the model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7HesUcSqYdS4"
      },
      "source": [
        "## Defining a loss function"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "F1snY1imYfwW"
      },
      "source": [
        "criterion = nn.BCELoss() \r\n",
        "optimizer = optim.SGD(net.parameters(), lr=0.001, momentum=0.9) "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NsVqxwY8Yt90"
      },
      "source": [
        "## Training the model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mc4BY5SaYwFP",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "178cc0b6-03db-4064-db98-68e02850e36d"
      },
      "source": [
        "num_epoch = 4\r\n",
        "\r\n",
        "train_iter = []\r\n",
        "train_loss = []\r\n",
        "train_accs = []\r\n",
        "valid_iter = []\r\n",
        "valid_loss = []\r\n",
        "valid_accs = []\r\n",
        "\r\n",
        "log_every = 1000\r\n",
        "eval_every = 1000\r\n",
        "\r\n",
        "from IPython.display import clear_output\r\n",
        "\r\n",
        "for epoch in range(num_epoch):\r\n",
        "\r\n",
        "    running_loss = 0.0\r\n",
        "    net.train()\r\n",
        "    correct = 0\r\n",
        "    total = 0\r\n",
        "\r\n",
        "    correct_valid = 0\r\n",
        "    total_valid = 0\r\n",
        "    for i, data in enumerate(trainloader, 0):\r\n",
        "\r\n",
        "##### --- Uncomment this part in order to do the validation as well --- #####\r\n",
        "        # if (i % eval_every == 0):\r\n",
        "        #   net.eval()\r\n",
        "        #   val_losses,val_accs,val_length = 0,0,0\r\n",
        "        #   with torch.no_grad():\r\n",
        "        #     for j,data in enumerate(validloader,0):\r\n",
        "        #       inputs_valid,labels_valid = data\r\n",
        "        #       inputs_valid, labels_valid = inputs_valid.cuda(), labels_valid.cuda()\r\n",
        "\r\n",
        "        #       output_valid = net(inputs_valid)\r\n",
        "        #       predicted_valid = torch.round(output_valid.data)\r\n",
        "        #       correct_valid += (predicted_valid == labels_valid).sum()\r\n",
        "        #       total_valid += labels_valid.size(0)\r\n",
        "\r\n",
        "        #       val_losses += criterion(output_valid,labels_valid).item()\r\n",
        "        #       val_accs += 100 * correct_valid.true_divide(total_valid)\r\n",
        "        #       val_length += 1 \r\n",
        "\r\n",
        "        #     valid_iter.append(i+len(trainloader)*epoch)\r\n",
        "        #     valid_loss.append(val_losses/val_length)\r\n",
        "        #     valid_accs.append(val_accs/val_length)\r\n",
        "\r\n",
        "        net.train()\r\n",
        "        inputs, labels = data\r\n",
        "\r\n",
        "        inputs, labels = inputs.cuda(), labels.cuda()\r\n",
        "        \r\n",
        "        # zero the parameter gradients\r\n",
        "        optimizer.zero_grad()\r\n",
        "\r\n",
        "        #forward\r\n",
        "        output = net(inputs)\r\n",
        "        loss = criterion(output, labels)\r\n",
        "        #backward\r\n",
        "        loss.backward()\r\n",
        "\r\n",
        "        #optimize\r\n",
        "        optimizer.step()\r\n",
        "\r\n",
        "\r\n",
        "        # print statistics\r\n",
        "        running_loss += loss.item()\r\n",
        "\r\n",
        "        predicted = torch.round(output.data)\r\n",
        "        correct += (predicted == labels).sum()\r\n",
        "        total += labels.size(0)\r\n",
        "\r\n",
        "        if i %100 == 99:\r\n",
        "          train_iter.append(i+len(trainloader)*epoch)\r\n",
        "          train_loss.append(float(running_loss / 100))\r\n",
        "          running_loss = 0.0\r\n",
        "          accuracy = 100 * correct.true_divide(total)\r\n",
        "          train_accs.append(float(accuracy))\r\n",
        "          total = 0\r\n",
        "          correct = 0\r\n",
        "\r\n",
        "        if i % log_every == 0:\r\n",
        "            fig = plt.figure(figsize=(12,4))\r\n",
        "            plt.subplot(1, 2, 1)\r\n",
        "            plt.plot(train_iter, train_loss, label='train_loss')\r\n",
        "\r\n",
        "##### --- Uncomment the following line in order to plot the validation evolution --- #####\r\n",
        "            # plt.plot(valid_iter, valid_loss, label='valid_loss')\r\n",
        "            plt.xlabel(\"Number of batches\")\r\n",
        "            plt.ylabel(\"Mean Loss of the 100 previous batches\")\r\n",
        "            plt.title(\"Loss of the model over the batches\")\r\n",
        "            plt.legend()\r\n",
        "\r\n",
        "\r\n",
        "            plt.subplot(1, 2, 2)\r\n",
        "            plt.plot(train_iter, train_accs, label='train_accs')\r\n",
        "          \r\n",
        "##### --- Uncomment the following line in order to plot the validation evolution --- #####\r\n",
        "            # plt.plot(valid_iter, valid_accs, label='valid_accs')\r\n",
        "            plt.legend()\r\n",
        "            plt.title(\"Accuracy of the model over the batches\")\r\n",
        "            plt.ylabel(\"Accuracy\")\r\n",
        "            plt.xlabel(\"Number of batches\")\r\n",
        "            plt.show()  \r\n",
        "            clear_output(wait=True)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ipEyd1kO96DN"
      },
      "source": [
        "# Saving the trained model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JcNV5bOGCn8m"
      },
      "source": [
        "PATH = drive_path+\"ProjetDL/testdata_without.pth\"\n",
        "torch.save(net.state_dict(), PATH)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OGT2O1nQ9dEd"
      },
      "source": [
        "PATH = drive_path+\"ProjetDL/testdata_with.pth\"\n",
        "torch.save(net.state_dict(), PATH)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SOJayo9jZEEl"
      },
      "source": [
        "# Loading the trained model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DrVG_hjRZGvu"
      },
      "source": [
        "The training take more than an hour each time. Therefore, when post-processing the results it is easier to load an already trained model."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vPgDmxkgZPUh"
      },
      "source": [
        "## The model trained without added data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_-1zJ26pZR2i",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "85439548-51d7-49cd-a8f2-93b6c5074b65"
      },
      "source": [
        "PATH = drive_path+\"ProjetDL/testdata_without.pth\"\r\n",
        "net.load_state_dict(torch.load(PATH))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<All keys matched successfully>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kYeBtULjZnOM"
      },
      "source": [
        "## The model trained with added data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QASZya0lZpWK",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "94d44eff-f516-44a1-bb84-adef17c53361"
      },
      "source": [
        "PATH = drive_path+\"ProjetDL/testdata_with.pth\"\r\n",
        "net.load_state_dict(torch.load(PATH))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<All keys matched successfully>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YoAkHSx5aBHI"
      },
      "source": [
        "# Viewing the results"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Yli1JV8paDsE"
      },
      "source": [
        "## Testing the model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-_QolCvXa1HC"
      },
      "source": [
        "The model that needs to be loaded here is the one trained without the added data."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kxy8CO28aNLC"
      },
      "source": [
        "### Creating the dataframe"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jcQE7jlCaF6J"
      },
      "source": [
        "##### TEST DATAFRAME #####\r\n",
        "test_df_without_data = pd.DataFrame(testset_allattr)\r\n",
        "test_df_without_data.columns = [\"image_link\",\"attrs\"]\r\n",
        "test_df_without_data[[key for key in testset.attr2idx.keys()]] = test_df_without_data['attrs'].apply(pd.Series)\r\n",
        "test_df_without_data.pop(\"attrs\")\r\n",
        "test_df_without_data[\"status\"]=[\"Not_known_yet\" for i in range(len(test_df_without_data.index))]\r\n",
        "test_df_without_data[\"predicted_value\"]=[str(0) for i in range(len(test_df_without_data.index))]\r\n",
        "\r\n",
        "##### TRAIN DATAFRAME #####\r\n",
        "train_df_without_data = pd.DataFrame(trainset_allattr)\r\n",
        "train_df_without_data.columns = [\"image_link\",\"attrs\"]\r\n",
        "train_df_without_data[[key for key in trainset.attr2idx.keys()]] = train_df_without_data['attrs'].apply(pd.Series)\r\n",
        "del train_df_without_data[\"attrs\"]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yDdMCjz5aZar"
      },
      "source": [
        "### Accuracy of the model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GPaCs3BmacJU",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "02409ee5-b87e-488d-f00d-6845c5113c66"
      },
      "source": [
        "correct = 0\r\n",
        "total = 0\r\n",
        "idx = 0 \r\n",
        "not_predicted_as_attractive_idx_without_data = []\r\n",
        "\r\n",
        "net.eval()\r\n",
        "for data in testloader:\r\n",
        "    images, labels = data\r\n",
        "    labels = labels.cuda()\r\n",
        "    outputs = net(images.cuda())\r\n",
        "    predicted_raw = outputs.data\r\n",
        "    predicted = torch.round(outputs.data)\r\n",
        "    predicted = predicted.cuda()\r\n",
        "    total += labels.size(0)\r\n",
        "    correct += (predicted == labels).sum()\r\n",
        "    for predict,predict_raw, real in zip(predicted,predicted_raw,labels):\r\n",
        "      test_df_without_data.at[idx,\"predicted_value\"] = str(predict_raw.item())\r\n",
        "      if predict==real:\r\n",
        "        if real==1:\r\n",
        "          test_df_without_data.at[idx,\"status\"] = \"True Positive\"\r\n",
        "        else:\r\n",
        "          test_df_without_data.at[idx,\"status\"] = \"True Negative\"\r\n",
        "      else:\r\n",
        "        if real==1:\r\n",
        "          test_df_without_data.at[idx,\"status\"] = \"False Negative\"\r\n",
        "        else:\r\n",
        "          test_df_without_data.at[idx,\"status\"] = \"False Positive\"        \r\n",
        "\r\n",
        "      if predict==0:\r\n",
        "        not_predicted_as_attractive_idx_without_data.append(idx)\r\n",
        "      idx+=1\r\n",
        "\r\n",
        "print(total)\r\n",
        "print(correct)\r\n",
        "\r\n",
        "print('Accuracy of the network on the {} test images: {:4.2f} %'.format(\r\n",
        "    len(testset), 100 * correct.true_divide(total)))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "19962\n",
            "tensor(16000, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 80.15 %\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1b6Q2ZKEdxOZ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "outputId": "ec75c364-53e9-47a1-e32b-d97b5df9ed8f"
      },
      "source": [
        "test_df_without_data['predicted_value']= test_df_without_data['predicted_value'].apply(lambda x : float(x))\r\n",
        "attractive_predicted_df = test_df_without_data.drop(not_predicted_as_attractive_idx_without_data)\r\n",
        "attractive_predicted_df.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>image_link</th>\n",
              "      <th>5_o_Clock_Shadow</th>\n",
              "      <th>Arched_Eyebrows</th>\n",
              "      <th>Attractive</th>\n",
              "      <th>Bags_Under_Eyes</th>\n",
              "      <th>Bald</th>\n",
              "      <th>Bangs</th>\n",
              "      <th>Big_Lips</th>\n",
              "      <th>Big_Nose</th>\n",
              "      <th>Black_Hair</th>\n",
              "      <th>Blond_Hair</th>\n",
              "      <th>Blurry</th>\n",
              "      <th>Brown_Hair</th>\n",
              "      <th>Bushy_Eyebrows</th>\n",
              "      <th>Chubby</th>\n",
              "      <th>Double_Chin</th>\n",
              "      <th>Eyeglasses</th>\n",
              "      <th>Goatee</th>\n",
              "      <th>Gray_Hair</th>\n",
              "      <th>Heavy_Makeup</th>\n",
              "      <th>High_Cheekbones</th>\n",
              "      <th>Male</th>\n",
              "      <th>Mouth_Slightly_Open</th>\n",
              "      <th>Mustache</th>\n",
              "      <th>Narrow_Eyes</th>\n",
              "      <th>No_Beard</th>\n",
              "      <th>Oval_Face</th>\n",
              "      <th>Pale_Skin</th>\n",
              "      <th>Pointy_Nose</th>\n",
              "      <th>Receding_Hairline</th>\n",
              "      <th>Rosy_Cheeks</th>\n",
              "      <th>Sideburns</th>\n",
              "      <th>Smiling</th>\n",
              "      <th>Straight_Hair</th>\n",
              "      <th>Wavy_Hair</th>\n",
              "      <th>Wearing_Earrings</th>\n",
              "      <th>Wearing_Hat</th>\n",
              "      <th>Wearing_Lipstick</th>\n",
              "      <th>Wearing_Necklace</th>\n",
              "      <th>Wearing_Necktie</th>\n",
              "      <th>Young</th>\n",
              "      <th>status</th>\n",
              "      <th>predicted_value</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>199242.jpg</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False Positive</td>\n",
              "      <td>0.514208</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>189733.jpg</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True Positive</td>\n",
              "      <td>0.727866</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>185254.jpg</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True Positive</td>\n",
              "      <td>0.744254</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>187591.jpg</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>True Positive</td>\n",
              "      <td>0.616067</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>183249.jpg</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True Positive</td>\n",
              "      <td>0.836090</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    image_link  5_o_Clock_Shadow  ...          status  predicted_value\n",
              "0   199242.jpg             False  ...  False Positive         0.514208\n",
              "1   189733.jpg             False  ...   True Positive         0.727866\n",
              "6   185254.jpg             False  ...   True Positive         0.744254\n",
              "8   187591.jpg             False  ...   True Positive         0.616067\n",
              "12  183249.jpg             False  ...   True Positive         0.836090\n",
              "\n",
              "[5 rows x 43 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EYhAk7q0cWmv"
      },
      "source": [
        "## Best and worst classified pictures"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TB6uSRMzd8DZ"
      },
      "source": [
        "**1. Least attractive**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-RfZiXMod_jv",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "outputId": "401a2b0b-222a-4764-8194-c9f448af8c10"
      },
      "source": [
        "test_df_without_data.sort_values(by='predicted_value').head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>image_link</th>\n",
              "      <th>5_o_Clock_Shadow</th>\n",
              "      <th>Arched_Eyebrows</th>\n",
              "      <th>Attractive</th>\n",
              "      <th>Bags_Under_Eyes</th>\n",
              "      <th>Bald</th>\n",
              "      <th>Bangs</th>\n",
              "      <th>Big_Lips</th>\n",
              "      <th>Big_Nose</th>\n",
              "      <th>Black_Hair</th>\n",
              "      <th>Blond_Hair</th>\n",
              "      <th>Blurry</th>\n",
              "      <th>Brown_Hair</th>\n",
              "      <th>Bushy_Eyebrows</th>\n",
              "      <th>Chubby</th>\n",
              "      <th>Double_Chin</th>\n",
              "      <th>Eyeglasses</th>\n",
              "      <th>Goatee</th>\n",
              "      <th>Gray_Hair</th>\n",
              "      <th>Heavy_Makeup</th>\n",
              "      <th>High_Cheekbones</th>\n",
              "      <th>Male</th>\n",
              "      <th>Mouth_Slightly_Open</th>\n",
              "      <th>Mustache</th>\n",
              "      <th>Narrow_Eyes</th>\n",
              "      <th>No_Beard</th>\n",
              "      <th>Oval_Face</th>\n",
              "      <th>Pale_Skin</th>\n",
              "      <th>Pointy_Nose</th>\n",
              "      <th>Receding_Hairline</th>\n",
              "      <th>Rosy_Cheeks</th>\n",
              "      <th>Sideburns</th>\n",
              "      <th>Smiling</th>\n",
              "      <th>Straight_Hair</th>\n",
              "      <th>Wavy_Hair</th>\n",
              "      <th>Wearing_Earrings</th>\n",
              "      <th>Wearing_Hat</th>\n",
              "      <th>Wearing_Lipstick</th>\n",
              "      <th>Wearing_Necklace</th>\n",
              "      <th>Wearing_Necktie</th>\n",
              "      <th>Young</th>\n",
              "      <th>status</th>\n",
              "      <th>predicted_value</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>8616</th>\n",
              "      <td>186706.jpg</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True Negative</td>\n",
              "      <td>7.850181e-09</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4856</th>\n",
              "      <td>195079.jpg</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True Negative</td>\n",
              "      <td>1.184525e-08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1689</th>\n",
              "      <td>184915.jpg</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True Negative</td>\n",
              "      <td>1.374571e-08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4264</th>\n",
              "      <td>197504.jpg</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True Negative</td>\n",
              "      <td>1.812096e-08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5602</th>\n",
              "      <td>185465.jpg</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True Negative</td>\n",
              "      <td>1.946956e-08</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "      image_link  5_o_Clock_Shadow  ...         status  predicted_value\n",
              "8616  186706.jpg             False  ...  True Negative     7.850181e-09\n",
              "4856  195079.jpg             False  ...  True Negative     1.184525e-08\n",
              "1689  184915.jpg             False  ...  True Negative     1.374571e-08\n",
              "4264  197504.jpg             False  ...  True Negative     1.812096e-08\n",
              "5602  185465.jpg             False  ...  True Negative     1.946956e-08\n",
              "\n",
              "[5 rows x 43 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "p5Fi8EOucu5v",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 235
        },
        "outputId": "4b877de1-b3eb-4998-8799-ebaeefbb5848"
      },
      "source": [
        "#Least attractive pictures\r\n",
        "path0 = r'lesimages/img_align_celeba/'+'186706.jpg'\r\n",
        "img0 = cv2.imread(path0) \r\n",
        "cv2_imshow(img0)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAALIAAADaCAIAAACNT48IAACjL0lEQVR4nOz96ZIkSZImiH3MLKKqZuZHeERedXV1V3fPUs8sTgLREPAGAJ4JD4LXwdKCCFhazOzQYmimj+o6syozMg53NzNVFeEDP0TN3MLDPTIiM6s7a5FMQR7m5mZqaiqsLMwff8xM/4//+/8NAICIOP5sQgc5PhMRifj4V8CPfw3Yg8/T3fHeTx55w+mJvcfLwcwPH+fwbe5939Ovee8ND16He2+8uw4hDz5/8oLTz3XAH/wOEXF6hNN3xUHcvf2896e71wBE5IcT55PPac9R3L8OWiYA6eFr8YN8j4WI7t29j90z75DlLW+pRdPCh2+pH+R7K6cm5FEL935yalqOTzIzM/9gLf5c5dRIfAODsWyLHvfe3v77QS3+/OTU2zh9/EHSfC+iu/e2B2bAD2rxZyff0kicHgcAnRzk9Gg/qMWfq7RVPP78xse5514wJfygFn+O8i8QidypxVvx9wOvfusMTk6O31DYuzfiW3nLR2HmY7x+75wfFHd/DG9YYjBmvPU1H5TTdx2P1s7n3nvvfdw7jnx4GT32ycdL3WCJ46eb2VsHAQARiRNZznDBPw6vf/vE4q1DEeOHAPXPUY7K9y0D1HfID2rx/ZXj2j+IVXyHAMbb8oNafE/lwZV+/ye/zecS0Q9q8f2VU0/lwcd/sk3kkXzSD/L9kXtu7J96EyGSH6zF91fe4Uk8+MrvVv5EasF/yoN/z8UB/1N/xnuqwte+KGhhGhz5Bk7u74CzHoubj6HuEvce0bF7SkCEINB3BVu8kQLACd7wjg94GG8ALd+gHbCdLO5zGu5wgrs1ppOfiDjwUUDuHmgfRI/pRMTJ9Tk+JmJOD78lFqzpCDsBWD4BaEyL43fEYa+5x7dopytvXrHjR999VrskdLgUFH8ylPPuKvC/wN3zfREKWm6Ptip3iBm/AfcdsaO72+w7O4UPBz3j5Gc7tf//tPN/QvkXwJre/ekPPv5Q+SEn8h3LW3vug97inbWI8AcNxocu6r0M+10K7RulWH9Qi+9YWi4GJ0zY9qv/S22k9/Tjm8kPavEdy2N3+RvrdPe4uSDfjbU45tm/2RFO5Qe1+I5FRE5/PUn5/sldDfqOGDr4QS2+c2lq0bTB/S7NTXSiLifW4k9xDt9SJ/AOtXjQFkWEx6M8htPXL38lcl/gne/KP79/HP/AupLHT+DeN2o/H+ORnNahMPPxvbXWk6Px8WUf6luc8ire/PQ4vf73OBbv/lLvI81l/iFA/VZyXIzHVuVb3rX/WvKDWnxDeRtPvHfLfq3GfD+l2aEffIuvkXcsKj3EmW7ewtHCn4SL/wro1jvk8c3lTS7nD/Kh8rbGnKZs7mnGn5f8oBZfI1+7Bdy3GXFqKk6P8P1Sjndbix98i28o7/At/ky9iib/IqS9+D6q3SnD4Gt/BhDksSS63/q3pL0p/IS88EafgqMsjwMAeXvt91Zr0tsMhvbrMeC+97w3TYoD++CATxyP2LgLBxqEJ1nqKfyU3/AO8CAe/tOjcTk//HoCPfi5xAFq/IUA4IiW204iHu7ujoBHEBgEIqcI8lOKxkHXl4Md2Q7tJXzs7bGclweBFvymfWKjGixfRx65Eqfw+OnXJyzZr/v8iUYZOcmSx8nzuMN77l/PewSwdjn/7H2Ld8A49GYh3rLlNxoLYnkcyzMWHhFgogCE6Ph6mANMTkRNIRgCwKxhdHIElxgCeEAPSw9gUbGj8XAwFrvhQY2J8i6f408NhzyGMf4vVi1a1depP4jlpo6IxSIdHwBwsyPb6ug8BsHVghAgHK+ge7MTaFmuxpgiRDg4iHhhph3PjY6qgIWY837e52M+ygf5Lu+Og+7dMziYkz8btXi354xHwsUHrMVb1/TUGp8iyo0YJ3RH8lu2T0RE5NYapNmP9py7u3OixRId8wOHgOShk+fH+BaPnec7XvD213/w+bdfc/rKPzO1eEyOpaSnmFK8tfveU5G3r+lx08HdEntE5JzvXubtrw4A0TaR5rgIEYHBxO6Nw0pHxkWAichM2/m+9Q3elSz5UJ04/TrvtiWnOZ1T5fgz8y0e0/17T59er7ecsuUvD774bY0hIga5WtO75T2+sDW1zER0TJIRMzMHIeQ+6bj9xoFjazNqDk0woTF132Ut8E0NxodvH3c+6Z+NWrxDTmGDozac9p/DiTkRuouwjoRvAp047kREIkuMRotruYhg2Rpc7c4kHCAsMBGzH88q2CMAjwiKdnseEqrvDXF9Sw/j8duJ7j3+X5RaHJf/Hqx0LOw/ffHb1+hBj4SIREREEvGQMgAhJuHE0swD3zmvCDU/CmE0cyyuiVtE+OF8+NQoLGF1vC908ScKSd62GU2+nm9x7233jPPxNnrk/ML9DSN5fO9jfTMf0+62zHjLKpw2tTw6EA0naKt7WL8ws4hwcDP+hxNYlpCIcs7r9Xq1WvV9n1ICwIFMzKBm9mkp3HCKcPd2kPb51iR8DRQ3VVdVd1c3rW5m8zRSEpEMoLktzEmEillrY3pvF/MDQeOxGPIxeWtT+Cag+321+FpX5buSt/fyd8s9Y3D89bFvnXN297ZYAEQkpSQiFGRmqqqqzJxz7rou53x2diYiOWcRaRrj7uQR6nFQvRZwULgfVi7Mj6fHARCziIQYmycxh5lZMrMQ5nZtTcPNKYJCg0SIT+PVR/yh9025PXb3v7+0N77RDedoAP5lNOP0PO7kkU8+3SxwcvneXbbWXn8HWhCVuZAHAyKp7/vNZrPZbJp5OEandpAwX6XheJLkBlCAgSCEuyPszkVlbviFRzDAzEIIJhUyjSxrdy9ai2tYeEPPLCgRTiKgdyjEPc24F1u9fT2/sWbgCH7fU4hvc8RvJnfG4xF22z0P62sVt5TSNosj57atNAdy1w3DsF6vh2FoCtG2g6MRam8EAOKUEt2pxcn6uTVn89R6EeCmHEFBREEsESEEC4ewhjORgBJBVc1hcHc3vAlvv7kieKdmHK/Jg89/g3V8w1r8y5uKe1fh/V//nnLvMh1/vbq87Lqh+RANk2jO4rA+QwQOKNZyNQI6VgaBHABTUx2LaLAnLQWerk3n3D2ztM2EKEgkIszg4FlrAiRJInTCtXKtVsKLmZ22IDvI8Vs/qBD35Ajpvv3dv5mktxXiVFv/dHJafXWU91fKr/3mOedlI4gQkb7v+77POZ8N5ymlvu+bSwGReyhku8TH2l8iOlIojrgDAF6ikgZzLQvJIEQQwG4QJreIgLq7kWmLP5JwJmKCBNwcXmlJpL1xBd5WiLcN+WNB5vtcn8ekfUrCv4apwKH66hg4vP+nv/WFH35Lu3cBNJ1Yr9fr9brve9Il+BQRpAQRMINIb26Ww0XcqYVHTpkCLd6ICESwHyIgOiQzmTinjhkeOo1oihNOIA+Hm5s2X2TZaQIMF6YMHqgri8G6s1LHM3mHqXgs0LinGd9sQe9vIn8aOY1Fl8Tyt9kUTuXto7QbeJ6mFlkMw7BarVer1dAPKaWUU4s+KA8gQlGrRVU55/Yp3D7IvZV5MGUOgBwgCkMEwgBXrRzmYe3VzARhgDIl8gARSBDBZtUD4GJKCDNtBRWIoHBGDCkRXFUt0D5yaQCxVLsjOE4R0ncTWB6yFvfecUjsPyKLtXhbp752we6tUNPxe9VUx3Pye/UOBARAOHahJ6J2lovFeuRD/SFXdCGzEBjEIKagIIqggBPO+tX5+fnZ2XnqBmpPl0AOq6qQZJUkIQScJA2pXx29RydQAkfDvwVu4UphgCMcXtyVhIOcwUQtRaIAEE4rgVnUGl4BBFOIR3gK0jJ7VeZEoHCnoJ6EoIJQ5qpRHdUQgFOw5CA4ecCDW16lXStuY0HuWYt7q3b4lekBJhS/9frTa/tIgHqyWm983jcNUt40FUsaku/NVnnPuPxtsXACBRABc4NHQoDo6vxi6NfrYd2ngTkDCFAgMicIEySIKSRIKBjMgARxIJzCoikXOOAQIIAczIwIVFACeQSICeFBRuFAwzaNJCOMMlMEDLVorWWeZ4FEREPRKAggYpHMOtUEg3OgJdVglIKgQQFvbkx8SybXXZeVIxGEgQfKk47ywCZyGhDHm5nJd330Y+nd93jjwef7JppBQcS8oI9mDGJJKaUnT57k1HfDkCQzpwiyIAJS1wUxMxMY4CXLwS1spGhmmxDkaPiVBwFgYmpwRUI4SAAKp4h2o1CLYYAgVYAABjcnwhoQUkzDScA4+D0i7TQAgAEGMTOICOyAhS2KTHfNCFoE9C8g9zeRo0481qfhMXlMLe7Ztnt/eMDZ/sC7om1eDHg4g/qcN6v1uh/6YUick3TMiUjAkqndoF20nR0AhIhJEkjAgoggJvKIZrCJwimckQIGYiDg2sDTcAecEEQMOC1WnkONAoCBBQBzEpGc+3G/bVi5mWm1WmvOwUmIiJwpIKAgUiDcCcQWLoTm2jK5L8yf74oc++Zq3l/Z+z2/j6v7WJ+GB2sjv7G8Z1z+mFAseQpzQ0QSGYZhs9lshhXAvnyLCAILJ8mUMkQoAkEBBqXgRJzACURAsDcqJgUx0O7fQBi5R3i4hZtVM3cygMBE8AALqOXbnRjegiADAaYRYCx1yUQkEaGuRc1DOZWF9eqBphmO8DBTIVAQnBrbk+gA9MlRMejkJx6Fh7+RpKN5uIcvPer5f2Aq7x3W4kQhTo7wgd/O1SIi4Ill1Xerfhhyl1JXqyK47R1MHMTBwpywXH8KMB11ggXAklOHRQSBggkRYIeFG8w9XM3M3cIMAQHaPRx+1GkCJQJMzUzhYRa1aq0qkhrx8whLuLtWJw51jUV7mBFwhzvaq4U8Dtcw7ojBHyCP2JevtxZvOxZ4V5+GRz790b8+agO+pak4/YAkad0PZ2dn69UqSXL3ruuYErhjTgBHUDi5w7m59UIQUAoSogQ6NAV0IoAITk7t1N3NrJq6VrgtAWRLjEeEu1G4I8Jbeo0BVS1FzSIiTF1rrcU8iEDVIgLhhGB3V1UPNTMYkURKTEQJFAiP8AgKJLATWUSAH50n8F3LGxnUU/14s0/DfaTlu5JvqRBoTiNxTmm9Xp9tNn3fw7xWHfoekggZxAYC2ALhrZ6DmJhZgphYmv94wKaCIAZDMHEQyCzMTNXNgiIIQcFBTOHeqNuGBpnDPGCu5lUbukpE7nCDO8wiojII4OOgSma2auY1nFIkcBJhZ5gRmtcry0bJDcP44Ov7DV2R9NjecfQh7r3gQ+XNDOfp4w/bjFJKqmpmdODINIUNtfXZ+eXlZd/3AtRpFknDMFBaFjuImDJxCmYDexAxBzGIwUIQRIAYYETzCjQ4mFmYQUxJMnfMPE1eptlqYYKkKHMRoVXf9ascVsdxP+sMD6HUDqmqNhdVb9kwOU4LDJKO+pTMjNy32y0zt4ydmeqsInK+Wt3sd7XWOkO63K83kpNrrWWOZfYgN6/vOCfmeG3fWEfCHRAGnLIQFkbqI2uR/hQ24BvI15qNU1pNC5QaFL0eVjlnxlItRI1fCQEY1OxBIs7BzJwcHB4nvWsZBEjjcBOYhUiCAmYRViqAeZoapARwTn1KicmFoAuWQNY8UW8gXZrmOczNXNVD4X5wooIawo22WO5tE6HG72AmIoqAu0ZQmRvYGuaqSqVQeBAL5/e8XN9S/vXV4j1xiyMb6nQ7Y+bNZtPnLMQUEBZmluYwQUDSnMoQEclBWZis6BsIMR2AItfjzRTuZmq1mGsZCyg4QBQpSSKGG8JT6tyqVjMz11qquQVRlLn6gf+zJGVARBRgdbNqWFjXpNWrVmYmEWZpit20xUOJU4uhVVXLzHCWTMKNa/anzl79ydXizePff/w2bvGYXhy1oT0gopxzl/Oq71MDsFJaiFVBEUTeQGkCN0xTSBIRMTsRMS1cqWPzolprS4cGzMxUS9Xiam7h7kIQESc4IqrWOveJalW3KkC4lqKuBvKItrtJ+ALpEnMzb808AGiMwBaJgEKAILg7DqQvKyYJQdH8GzVzlUQQYSBOM0rf2GY8FonEMYP6ryjvH4zQgZLZznvJgQ0DEQlxYsmSGlPSTqkxCADcXAwgsGxDbRNZ8iyN9+Xqx5vV1VwbTnDIxSwnGeGmWmuFUtUKNwhRUARV8zDNnMEIDiY0Sp8DBHJzN6gFETVuuCMc4RZMSA0hjaXPmLtbncCJkghLMLd8TLhnIYAP6SNaPIw3L9XJT3wzPONPrhbvsBYP4haPkfCOagEgpdR1Xd/3fcpRDYwjZTciiJip0b5bzWB7G4VHyzI1bVlsjxMa54pavKlm1cPITQAQ9OCvMDMtYLSk1KkWt8bPYSYPJzeoubS6ZkQzFe7aPtFVG+kXh8U8AoZBqG4cQQdo292rKWUn7oOFjrunOw7AwbcNBd5tLT4UtvoO5YNwiyPsJiJd13Vdt1xZd/LgAMwDcIKkJLlza66DoO0gh09UVT5AkrTYDQLBPaj9i4A11CA8GtMqkRAR3FxNcZIWMDPAGV5Na62qdkwBRiyFAi1Bb6WaHUgVC9cERHJgecHMKEAcEdaQ1Gi5YRCJLLky98YHu4czHccsflfyPdlE2pb5Bg/gnpfRMCQGNVORUyJAVfPBvFcP98qUUibJ2eHgJRsdJ4dtasHU6N1GJI21WWuNiIAtFR6mbuYRkgdmMLOFqc5aK0wJYBYDzELdKHxuSlG1mAuBueVdYeZh7qFWLcIiyN0iGLRUQosICUe4RnAEe2jAHRYUCocxOSciEZAE4MgLVH4oUWAixmkhKx2z1rTMm7jXZuPrK+VT+GPW4viQ8MZOZad39ul+/6DQY/M17krvCMDCpiYyNxHhJIalEBRMzGzhiMgsh9yMM3Fi7vMAZg1QEEhCZDaU/SxdTyTEDA6Ch1cmEPNmtT4GikC4L8hCw7o8zMKdAEmJkxOaywnAY4ksIMygUidilr5HOHl0xIjENJMHhbmh6tx2jUbj4pRVEW5E4gGtRgFJlLvO3SOMmM2sFi2lqHpRhyQiDiWdLbSC3IjMh2JKFP2Q+o6BcK/BrdIE0uC9BYptGVeP+2qBR7uZNoT9X91a3MkBjyOSIBgiIhxxRzQgAkWr7BARd2dEEOWUmJdcVBCbO4WFRCqJO0qU0BBqco2QEHc+grbNAp8S5lpU2X5tcF5OHQAPU1VTNTOmcCJmDiJyc4dGVQeYKGWU0sqLCMIEWgxBAOEeRzc2YuGIFtOUOIm06pW5FJg7WESCxYNMw0KN3IONWFGrukcpWmMzrNZJOMMLIg61rUsERtH4Om9PQ3pXMXS76b43anEUJpxGEocULpZKLB6GQZjn/YhAK+5Y4Ie7uC0iotaahJI0RzIiolnW4waPwyVov6rqUUWONSkACBoR5lprDa0RkZoL2LyiFmeqmhkimNkIjrDwoKUSzsIaOGtmrosHGhEQYmZdKt+pnUOtFQoIE4m7W6CqVyIFebCClJIhVMs4zeG16y9ykoPtewvqpndpwDvkXxi3+HppiOCycofthSLUjAJZUt915DF7BCDES/aBm3owEzFLCJfZZFEsamagtZwwo3sK137V0wU7eTCOIwAPa4HNkWDQdMhUSym1FFUVCqGFNd70gFkWw2MGQFXjjVHYDOEEDoojSR1LKaKDKIhraA1UiAIOMmKfptR3EbGfZncbhpTSGZEA9p2gnou1eM/OLHfypwdFj3jDUgnuEQir2qfcSn1g1sq2UkrLdnDydjpwaI5Ha0+2O9tsMQ+nC3zqHp1ioMcnGzgtrSokLCLMzcxqKfM81zK7u1MYiA/2hohqsyVu6gZf1p6I5NABgYhSStVKrWYWTClJTDTVWnOXm4/qAXPTCPUoENPdGTMx11pLHYdVWq2HzbqzOh6QjO9AHrUW3xXk/qHHPyxRHFtHtOQkzPtVN/Q9eWitRJRY7hCFEyVw96AFSTQzLOQ8Ohj8JXt5PIc4lCYfnzmaECLq+z5iqULltsk0yh1R2z7a7hMRagqPTDh6J1iAqdrqFg9cPTnqBJGQcGhUszA7Yjnq0RERiKn5Wa6qpcYUdfQISTlzVZ/L+PomX1ycrdeDgwnkFPIm1ftDb/v4PqCcb4u6tzUgotSquDzcPLOcDatV17vaPM/wkMRhHtS06I1CUxAxi0XUWimkEfYQ9wtZcdCJtlqnmnH8q4iYmbVCINy9MaVUa/VDKcohRW5GCIQjzM2quntYbdZF3Vres1UzBhMRqar7QZvN1dQd7XkjdiKAG+iu4eo+a9BulzupVctcb262r1/frIZ+6PJCRV1O35cWgt9Ivndq0a4+HaX51+5Z0tD1XcpTqVaVsUBAEVVEWEhEiIlFRBKSuFGEuxkQmdMx6BDJR5eifeLbv56eT7MHVUtEJMJCDmZu21kzV8IUEWFVGwJ2sBMLEyCMjinfxQPiw5dDrdUiwOSAmtZS3IOZx7mAJSSZEBETCVEjjkdRUzc3K+rYj69vtqtV/+knHzuCKBzR4FgAD4Uh73P9keKRvpb+kPFfeoe99Sd6PKH3js3i3q/N2SQiSsItnRjeSPKJeD2srNayH9nRpRxqME+ZG/xEHG2fBrO711mZuqAAyMxsmogInIio1tJsQDMD7Y5vEe+xhUFbueYqEtMwDCseIsLKrKpmGhG3t1NENJ6E1tKOw8xdkuaBmhkJCxOc3L2a1VoZAaYWQJmZmhFHqdVV4UjSoePZpzLN5uwRSzKHBMxgckVVi1C1Elpvb66fXp1/+fzFZ599liQnYS17Ne+zeHhYXa1WUxlPL/jb2n9qOPHNemc5gR+xS4+qxTsPeOrfoSk2L7avqSARCYhZcrMGAXjQic4ee1GYGVTBHMEWXm12Rjg7IWAAghTAMJwdff6j12lmjY12usssx2zlA0ItXDlwvhdUFG5HB1ZEGMRMXdc1zyYi5nneb2/neT56M3xIqJpXjwgOWxwLxAE6Y8kNudFixqHkNeAtiGGywFzUtE6l7qd5muLV65vPPvmIGUES5MerekDtvp6Be+8F/2qbyD2FwJFizgxauj+0ZqcJJCllkhREHi3LgEP0cXQPzSxIWyGQI0QyCyHECQ2ysAPw11y/dssef7VjX87DdtO0R6u1TYqIYOruhCMkFVarqtYyl1KszmE+5BSHvH9Oqeu60mWupcWoTScs3MxqLS38iAgORASZmpk5iKjr8mwGQ6OcLz3bmAWp1MYE0bnqOBUr81dfvah/+fPEslS+BBO5hx0Tde0Kn679208u5gTAOyKRx553BD30l28T0ZxqRkN4iJYuxuQhzFlSIibAzVQ11LzhW+5JSHGILKKVaxG3DSUxIUH4qBYRYXbnch75qkdw8+g0HM9nvdocrQVMI6KpRVjfItRaa5lT13VWs5n53NIjKiJ91x2jUC31GFgt/kqthgg7tA42Ny2h5g4nCCSCDjzf8HB3OJC6DubWUieg6jGN86vr26lq1ychBpG3DO5J3T0eNwz3/nRQi0dW7cFFflAhvrG8bTACEUytDG4x8O5EyJJaHiQ8Qg2NRu1vfG0ctEokMVOZC5yZgqKReOFL1nTxE1NKDUfHgULX1CLnTERHX6HlRBYqRrvpw4hIy9Q2kWYGUkqp6VLO8zxP02RmVTUiyjzXWkmYEBZeTMVoSbLzoV2wh6u2siKYOwGhNWAgBVeiGijqUyDOkrlbNJKghKMU3e+nMldfDSlRg/GXBH3YkSmyXOF3Ggwc9p0PtBaBxwjIj1mFdxuLe5pxBBKYSFgirH1UZiGPoKBDHyt+02k5AhgL8YJJtSzz59zQCjRIcIgsWgTY7qf2oHmgR1BynufmcGhd6lBUVedJVdGyJa7u7lrNLNwamhURXRLyaN5rRDTLUWtdOpM043QANoiYhBtaqqbefBc1dRC5ETvxkgeOiCD3qHMtpm4AsxNbYK5lHMdqGg1gaZzyQ0bzfTaON172r4tbPOZb+MEZZObgsCPW2eoDD8EhgMQid3Am0YmfZYGcs3PLbkaEL2gBkao1M9Awx1qrmbVfW1+ldoR2u+eczzbnEeFhtdaKEJFwBSCUzUwL1VYG4g63iLjZ71JKfcrHWAaAiDQXVbjl2MLdYqFOsOnS4o2WyBaqmnN3vCbMLIQUSB57U0dQEg4wsxu0Wql2GEhIbnD4230M7/sQj2nGe6iFU9zvhUCBt0Jhxrt8iHdtPG8pB3OERlBAiHEolzkemplbXA6AhCGNHNfKL9p2gIAZSHJH7nYwpt6MDFGfJCIxs4i0rhIAMajdws2raH9tbqO5RkSoeS2qqlYaFa9adffWG2PxbCAgG4Z1SpwoNZWrtZojwKqNuc7EQgAtOwEFwcKLtTQIh5CzO6KGe4S6O7cUMLsHIBRgSCcJQSJCsfB60JK+LtFoRBSteyjetPpv4wIPPp+OMczdUh1yeiBnoGWYl4guPN7uvHGSjTw84JNfH65g4bf1GeBwgAbpKWLaz2TeUR4OXberabVKSVpF7+jVxjKAwYmZJy2qqh5EREslWXi0GmvinLo85NznxflgN4dHYkHji4usVqtSyrQfW1QC9/3tNqVU6+zWWp1oKcWrEkdDJrQsDmZESCuCZl6t+tQnInixWa1WBUCc+r4/giURDOpqmJpNbrObluLu3Cqd+1xqC7ekwTfqVtRnj+CeghNxmF2sNnXeXV1cvH7xVZ1HxHliMQoWJBZTV9XW1u1oIY6P72nGPTzjUWvBS8Fj8/CdjnDHG6aitaNrzwjeZTPeU1prsCCixmijAz7dcmMQckRRbQWkKfX7UkXQwgfJXeYDxkFCRB5kZrVq2zCsmKzO2v1fban9apR8VVW3tsaNAIaWUgkjIGchSg1v4MQppVKKu2uSUkrbIKhl74giaHE5ayVqyRM30KxG5ouaipjZXFzDqrpa1ECEI0AeABQIcLQnQbr4142I2oacQEAaS0RmpmGNo+VwODdsxuGLG/a2QrxDvt63oEMlxTugzLtXvvkAH64oB0tOzBRcIxaoiht0wNxqMBAkuc8598OKJB/yT42An5i5ViMSBhExICJuEQDPtcShY6u7B1FKSRpJohV3EDin1Hdmxm5dWmxJRETsVDVR7rpusSgp4aTDMBNl5tLSqrXagdzRXNlGBGvurQBVdarFotWamB9YnwdvVFqBlLuDwry1bVqy8C1AOzqVxxg7Fm6Rm3m4U9tkPmgBADxI8V2mSh1UDCcKQUSPFQ8en38wuHhbHnzeCS34WDxJovAwCyZberBbKyNdzk1VuTEezY5qEVCAU+oi3A/pMRFJIiK5FA2n5C6tWZZwA8KJqLaAwCylJDmJGoA+SVOXFlDUWi1wjF/80BX6+HVSI+se8qt2WPJTj7gpenMzfSGXtGslERrRvjIsQmHROjKA3KFuHmEeEd4xQ7gxGqWli+OQtkWQe7iL3PVDurdZvFu+xlqcKsT7HPFBg/FBsti6RSkpwhXOSgZz9zAiZkkCJjOrNrEQOKWUU0qSWoMZiQg0cIIFgPodqn1+fu5BC2u/KWIEgFJK7rvU5VJK23EIIOGplnk/zvPcTsyDPMxqGbquJV5xmDfW1hvmIimlzFzN5lLqYkiYUkoicJ9LKcdtPvDGpk4kQBC5hZvDIjzcwUFsDiNuiX2Ex4HGllJar9dHAKZtZ34oRjr9iPeXd1mLo0IcNxGcWIXltW++6/Qb4iTp8LY8+HzzVo9dApqLTRxOoeFYcqtIxMRk4Wq2GhKC792L7m5TGYZhvd70fR8kutybrWsiHKEHi21m6jbPc9d1zFxqPXYARvMzVVuw2vc9ETUazqrr1crinSDcvdkSjaUrr4a7u4aTB4QbvZMDhjAzDecAqKV5l5YW8MNgvSBHGKAW6q3/YzhReDNpRvDSupebpS73q9ytBmaubhHObSDjwodd5P0dC7zbWnyoqTi+6/TBh5oNImntKNwR7owlX4BovgUCUDWKIJGccwvTLapEc7OitnJeiHs4pFq0rn7tAm232+oLcxNYqvyaLVGzVuGaUqIkRARzjg5rXp2d933fHaCnCCNqMxXLsp2bNdw6gFLrfhqnedZw4iXM1nBXDfdSqxPopCBkKT5wPdwJzZ+QCDeERWBpIIggTKWUaksipVYtZejT2dlqGAZmXhJGwg0B+hAv8w15303kQ+UbbyILqPVGNLVkLjiJiGjrExLepdT1fcqrONYU5YGZLWAa19c3ThjHcbfbuYOZJSfhbKC7kDIvlasi0jVL0D4op5SSu1uptZRWrZRzhul+v29bAFGYmVdrOf1mdSyW5Opuv5+miYiaN+ru5t7yag1jPYKqzfdoFB4ArZbEPDhTgMMjiFpXt2AOQq21mjNFBLlWNV1RzkOfcybhpnkuj7Al3lvu97c4eXD3PJ1giEe0ISKA+0pz2DXouH08plWP6c2CXnt4hFkgXAkRTHmpsnNHIJIkALXWrmd1r8UQSux9yh1nZKS+B9Ai/qV1SXVv3coOPL90mBpBRJJSu9tqrere4k8rNacU5vM8l1KYGZJy17UoxqGtkIRax5yDqAdJIkmqqr4UyFsgAtW8VCWiDCJiyR2ZR6i6ASTEJFLVIpyZRVhA8zQbkPpORMxpc37m2z3ChqH/6strYX/58uUnH18RUa010aHHXE4iolaIHr7z7637cTniFOV8Qyc+xDx8Y3PymESERfDB/LmDyJkbu4ms9d9mJhHJKUiCwMwtF3AAmC2Y1utN8xAlUhy6WRARpdwWb4kOeEG7cfQz2v7SnNeUcs6Qu35Ax+RcRBs3s/TBKaVM01RKwYHql3NuetbkePXpkM0HsATJjVWyNHMjIg+mUo2TkKQgnavNtidJRky5Z2ZTtXDOKXQ2t77vQcRMIPeAhXN8KwQp0X0j8YBr+YabedABogWa/m41w4LIQU5gcaqIoCA/+JKC1vYkiWSRDE4AgZgEHmGl1FrBlVmYsiMWheB0jFN0rhSRWeIwQMTdjWiJKg/E/MwszJGoYdKSxMx8MTgMIY5gZjcGQkOrW7WYq4mQA5xSJqK2Z5SiByi5ceg8lr47tVaOY3N/wcINE0JM1TJbsHASLy2TA7AQKiXR2VS167qic621Xw3cZhIQNdDL/Wsmrbxb3ohE6ABW3MMn3n78Np7xXRUKLFtPq70RsWrHu9Mi2AnLrA7gQGs+IWh5OFFygLb7HYAFhud0THPM+5mZhbgh6C0uOFoOPxB8jrtmG1J0zHvZoSlWzjki7OC6tow8ALN6/C7HqPjUxhyNx/FoEbS06iKyFjswRYQuVU+S+yFUgyhAqipdb2atkQ6YimnbEw/nvYxHitM1fW+520TeYSfesXjfIEh5H6G7ynQOzgZ1WNCRLdGK0O8uMbdNmJO7N0Yhs6SUWicJapxvNKfBYJ4lMTMDqqaqcy0NqNhsNi36aAnbRrJQN8QdWauqtQiUiNB1xbxFs4hgTl2HnPNci6pGreoRxI7Wu4SrVgDmYYEAMTEYRA2cCzQ3qNEWgoIk9zKbWitazTmxWFU19SA5UIdIFn0tpWh4OviBwN208G/WU+0w1vEUacC7NpHjSPvTTQT4IIfkXXKnFizOdyAdWuzKjUYTB14FH1nUraAZx0kffKcWjXHfiOAMoghtwWEp8zyrqoXXWlMEEakbADlUIHb94IfHC6VKa1uJeZ7LuDczYRZpw8joWJ+ycOxkwc7necbJOIvlUguj8TkbuOnLWgLouq5MoaU4i7NV92pa1aXr0eB2CjNrUdLt7a2ZRatpYEajZx0+5RuswjfxLf50pgKLDYjgxZZ6q+5wX9rqErUpcSJdTp2z1NIaI2qbECIi5m0YUBBRqLfGuc1gw30/btsoVAtXVbhnkcw5zBXqJ6wcANWtqrWcRCxNUQ6otWqt1WsB4CKBZWRVS2c3l/MYi9GBSo5D4H3KC2z7iJlrKACSxV+MCEOrkoGZA+i6jlIioq7rAlanaXM2DMNwc3Ojqp6IKBEzlnbz37yN5zsiET72QliWZIkMmLFc62/4me+UdlMHLx1A3B0GZzi32gBhFpKUc885C6ftrqgDQEop5xBJ3gzs0k+Z/W5OwtK1sP0ahzQWs5CglQ+qLqFpS8la4HZ7fZffEgFg4RY+zpOZwVv1GytRKydMib1BbYego729RSU4qIUeW+EwRct3hJtFkAuhZXSbC0KtGimi64auXxWPiJCuD62z73POq66fxz3atMSmZ0TUWoYEfzO/Mx2nCN/tJrR4cO2yHl/K1H5duikfTfvyrsNc+ra1nejZg3NG8GZfyOOzcK+JkWUQRLHaAgROyYNBidKKkoSgBpVSq9X9rpZq7ioiQ99zK84shSCNx5Zz369WOWez8OoDp7DQRrmDAR5MqAtPP5wMobWO49Tu8u24a8am7TUAmIST7Pf7JSQSYfXWkgXk5906wkstVjUR992KgsPAkDj07wUTpwRAPVLOKNVctX0EIoM7YVWvtbp56gdGihrkIKJJzYOsxO769un52qqy43JzloM2Q9+K3jeblVkdx3EYBrg+eP3fvbm8G+Xkk58NPPZFDb8u8DgEXR8qvtmsdvvbm5vXiRIMKSWORJIFQktVACO8FqtWpqqvX+3MEbZMmKXG/Y3Y76am2bnvVqtN7nuRTB7OHRwWtjSw4ECj0RMBbEG11jppKaXVZezn0Vxr0Vpr0Wp2aP2fGhLIrRwodd2Qh5SxHfcMd7V7Caqu61S1qkK4MTOCwCFlKo0oEEzByzzmluZyD1UPMki7NVLXDR3sZre3ai3wGbe7abejq7NaJtRVbi1mFWHBwYlZ/V2zhh6TO7WI76jb41EhvplmtIYCOefMeQEE3KupdAtHK9zVylzKbtyO4/zyxY0TkS90KTQSdkqllGASzrkULdYNfd/3WbrKC6EffDIKMBCgCHcNr9rsjVaPiKloVV0czJZcjeaBeoPHmhuRcu5zLwmbVWLBMm2bhJm11pasP3UpGloXEdO0J6LW/isRg5YRbqrWUDUSRkrsMNBctRU0pMymVEphVWZGhBXTWVOfhTjc4ZRTEuKHbcXjQm+zs04B0Uf04119M3Gycx8ffK1duSelFOGccxdB+/1uvx9F8jAMYGrxggNTGbfb7eubV/v9VOfWsX+hSgQhS+J07NRJAFSVZmIQcohwC0EJERytpsIAdzjCNBoBsxartVaPaa5VWzmIqR4yYu77/b7lVDmJBLv6VPYBD8uSwCB4JGIRQcQBzQwslKplnHPTtpRSWghEAZaDWiiIUur6YR2SZ7W5aNnvlDoRYfJXNze6fX25HtaZ6lzWuQ9zV2MQ4B0LQK6xdMh5b3lALY5yDKI+VL6lqQC471dTLVbmMKgbCeecOXUkUs28lIgYp3G73d7e3k77OXEOPc4T4TbNuM5m1aMDQ4jI3OFKYaE9J5bWCFzQyrojwiOKaUSoharPVUvRUrSqV41irrO1NIhZ+2er1aYNzuz7PksCtz4c6loQZB61VvJIKSU+lqIEc0qpCzcLr9Wmqbg7B8DuoQ4kImauHk4ULCBxIrBAyMiqeurp1Vev6jQ+f/58//LL+fzsL3/8ydPLp08unsCKTRqwEORhYE7jPCLT1829uydv9c46tRCPWYvHDva2QhyO8AGpsqDmkBsHrzdn52cy7+dpP++n8fLy0qq2MGGcx2ne11o9tM5KwULEkpnZENVVVdt8J1rYl7KclVOiFBBKBGeiCI+2KVipGlCLam4WalHVq5pBzN2CnBAsckjTpJQAr2UyLa0SKaUk0ubeUasGaoYFKbd8afvWzMxYsudHCLWVnUSE5LQUmZFEhJtrVYnUqJ0ePk3Tr371SyvzwLxZry8uLn7+k7/43/+v/jeusbvZ2ryncEn85MnFer22UiUN8SGtku5bi2NU/f6HuL+o38V8kDJrgFM3NEbMdtzPc02UzDGVOk17MytlmqapYVMRuur69Xqd+iHAUy2l1E7dLAJQC3NNFtEaZiYSKsK5Ge0j+TwizMLaXeyhBguo26w1yKqrh5NAWBjMOalSYvYIV1OtWkphTpmZ0Q05UQYIQWYOmJAA7THUTdUWZEXd3Ycuc053OHjrF8im4aBEnEhygKra9W5/s73d7Xa3NzcX69Xl+dmnlz/5d3/7N//mpz9ZD5tf/cM/37x67lYSoesF9tNV12dJH2r237WJfGP5NgrRJKUEFjX7/R//+Pnv/zjtpk8++ewv/+Lns9Zxmvbjvu3z8zyrVgZdbM4uzy+eXD3rVutSbT/Pc1FzvHr1qpjXWrXOglbx0XU53JBYs0nOuXkezVo0lmZpfdiNxrnsp6mUYojm5olInzpmIQmWvF4PHPDQBeIqVbW4a8gKTETiCAuCWW0oq8NhtdhUiwXUq3qYB6fMSSzgCA9483aIqzp31PW99GsFbq5vXr+++eL5H8f9drMafvzZZx3hk48+/m//3b/bMP39f/mHrz7/7e3rV4mj63i96i/Ozn/06Wd9N4z4UKcTeLBO5LjADz7t/oYX+baNuffMg/UgOAyCOH0LMxPzfi7PPvr4P//n//wf/sf/KCK/+Mu/vnx6Nc7TQoAjckSLFBhYDf1nn3wyDKuUcyllX+o4zXPx6qGQzdkZM+/32+3tbZ2nxuDteyR2kr7re07s7gmNysBeS5Ri4Qo2RvHYzROxg2M1DDmnnOXi/Bzw65evgrRbDV23ziwATOt2u91ut6kbbnd7IjnfrIT95vrVONvl5bm7b8dR1UmgFmOpRHJ59Sxlnvf7eS7SD+t+YJFa61jq5vxMg6vb9vbmZr9//ur18xevd7vd08vzq8uLJxcXH11c/PxHP97fbl++vv7yy+ef//bzdd9dffSEyef9fntzU0rZrM/C6tKR/mRnf/TuPTRB+J51wwnucv/3f//3/91/99+fn13+n/8v/9cI+uIPf3zys5/tdjtOQiptIw8MXUqb1WpYr4SklPJ6t3vx+nY7zaBMKXug7EYRYk5nT67Mahmn19vdJ8OZolUiKVgi2rzRWG02qF3B7nq8mYo5MXd9TzaPNylRv84ppTKX69tXEbEbdzyP+3lcD0PXdV3KRNQPA+X81YtXF1dPh66bxyn3+Mlf/Hwep5cvXzJzgCUxJ+EOyKKqxYMclPuupfRaAUOgo7Sd5tvt7vnr11+9up1K6dabTz756G/+5hfb65efPH16dbZ5sjnfrFa7293zzz9/+cVXl5dXibzrhpwoYOpRSundDh2S318Y8O+ZWgC3u+3/63/4H7th/e//T//Hfr158dXL1HXX222XMmdPmsJqy2L3XTcMXa315e2r23HiNFw+++hC0nY/39zuX7x83ZJeXcqrPgNe51JVX95eC2KzXner4eLsYrPZ9KnnnF69viHrbue6n8qrmxsj1sA0bWH7rhfOLCL7/cSBruscoaoONCTeoh6y89JvzvbTvN3umblPOUrNufv4089qrWOZb29vd7dbQ0hOLQoNlgh2t1LmaZoal0cttuP+Zr+/3e2r8+b8/Oqjj5599MnZZiVeri7Pz/vVZlh1Kd9e39ze7ITz+flZnfYpdetN7zBHFK1fe7XfloYpfGh/iz9JHuRUcu6fffzJ3/3d3/31L/72H/7hH1X1008/dTUG2IVTtnlvrXrErNb6uxef78fJkS6frtOwqk43480Xr17v57nM1cwkcWJKjPWwOtuccZJa5ttxn26yZC5aGRJAUDIiC9cIZ4EkV5vVos7TrPM8AzyOo4ishnUElVJy6s/WupoVwY0fSkT9erXf771ql3Mbxrnuh4uLi7Pzs6xDALOpayVmN5tUX97uVHWe9vv9frvd7ve7ltQllurWr9Y//fFnP/6Ln59fPilqu9vt5fnFZlgJONx1LmUsWfLVZ09snkq1apb6LluvoBpBuUP5EN8iuNUIfu+sxTzPT68+mqf6X/7h7yPw9OOPHLzarPbbWzAdepYEADMbx/Hl61eSh9Tlr15f/6d//Off/P6PL292Nejs/HKeKzMuzy+GLoWb0O2qzzfXL6f9lsI/evrss88+u7i4SJSq2/nF1Vj1i1ev/vD8xajRrTfOPJtLSJ3HcSpmXqvlnG9upuvr22mcc+5T33Xd0OVhGIau65lp9/s//OhHn642m+vr6zrNZ+sNML/4zW+ePn06rFbS95uLC9ttb7fbV69evXh9fbvdT1OZxl2tlby14gx3Pzs769ery6unH3/88dOry2G13k8ztH+yXp0NvQQkqAGvfd+fnV28nOdqVtQMBEnmXsKWWpEFtzil6AYezqEtjuY3sBbfklT8LnHCb3/z+//hP/zHv/tv/93f/u3f5tw9efKkzHp9eyu0TAYBICJEYarjWDjlsczbm+3Lm+2vf/flP//u85v9TLlfby4t/OmTq6uPPl6fn93evH718qXW8fJspe5Dl66ePf2bf/O3f/HTn20250H8T7/85+nFy5vb3edffvH6dh9d1w2bJHi2zu7UUKwur9dn57e3u+cvbmoxTkVVS7Xz88uf/PhnV1fnBBurfvniq81mwwEnzLWszzZ/8dmnYb7arEVkdt1/Nf7hyy9evny5n+YXL27HMqtWAvok61WfcxbE1UfPrq6uPvrk08vLy9z17rrqu03fo05Wdb1adyzT9VZE1hcX0uWFohah4dWNOdRsrh+8jxAR8Agt+F9RIuKLL774P/z7f/+Xf/lXv//8j9NcG7GFWr/VCF9IlDGO4+3trfT9q+vb691okPOrq0+c4sWr3VSNaL05/9HPfvazv/yrLklrjbbu06rj7c1rgq/PNpvNZrVZ574zi7/5b/5N/v3nr6fyfLu9nkpRXXX5k4+f9l5Y+2kq+/2+X5198smPh377hz++eDXfJl5plP24G1apX10MmyelbC+eXL346vPr6+tPnn3EzLf73dVHz/7ir/6SPKTLqjqbuvs4jmB6cvXsy5e73K37PlLi9dCfrVeb9dB13ZMnF2dnZ5dXl6vVijhpICIENE1W5t0qdyK4ubnJxOuz81AlCEkOFnMoIgtH4sZA/KD7+JupBbfOWRQMNC/3u5ytzYHV0P3iL34+73b/8//0nzT81YuX0zT9xc9/mkiUtA0La/P+xt3+1atXLtmInzx50q/Pr3fTftab270ZaSnrp1dPn1zmnG9ev3r9+vWT87O//utfsE2vX67rPJ2fXw7DOksHD1W9vrm9ubnxqn1Oq2EQj88++ehvfvHXsb/1eX718uU0zhxY9wMu+Hxz9tvf/P5SJBH6nM43q4+unp5fnN28nrdTOVtvODCXiYJyTvM8/epX/8zM6/W67/tS52ker29eT/N8fh6deN+vNqthte5X/XB5tnny5MnmbN2lPM7T7vXNPE5nZxf9anCjUqd1312PO7O6n6fnL768XG38yeW03zuWlnARAVDOfU4rSdm1AQrHiS2GNowHb4zCPAoRgSg9OljiISWjBtGgTVZjBN/B7X6nH6fNDujw/B2fAwDwRgan9QAydbV/+zd/c7m+6FbDxeXV8xdfffqjH8+1TNN09dHHO9XMktebQlPdbzP42dXTzdNnljoCm/NqsFW/+tHHn273Y9cPFxdPzi8vxXTdyV/97KdnZ5t1P8zbiortq+22331y9bEV21xs1msRTq9vtjC3ubB6BtXdeP38xWeXV/36MkF2u52773fXOae//svPdL559erVbpo+ffb0Fz/76MkZdVzkLK/kss55u93WMg/9eujSqxfPf/vrf661PnnyRIQgPPTyyUdPPv/8c593f/fzT55cXJ6fn4lIm+DadR1IEN6Yhdnj6WZzdXW13e4//+K19bla2c/bVe4vn1wkcCkzyLqc+5RhHoaeBlLy2RP3s81OgTYO+sCDaex6HEgIC3bFzJzabf9h1mKp4AQ7YcGOD77rdyWdpGeXF9IP/dB1Kd/cXv/+8z9M0zT0nQCbzcbmadrtAXr27Nnl1ZPo16PZPJVSdJXofL350UdkjmDK3QBgLPMqnQ3rfrM5z8Jf/ea3L754/vz58x99/MnQ9VdXVy0Lv2fpmHrhTMRmxazs99N2W1ebTMiSnpxfhPnZul/1w9Dx2fDvrm9vpn3t+v6jp5+cnfVWzVzFDUxnq8Esp5RyEkH40KezzZOri5xzqx8R8k3OWeR83Z2vV6thiEZ9S8xAwMaxSMSqH1bDKhG31Oi6H17vXq/W/X7ej7e789Ww6np3ZWbpUsptNIZz4q7rRcSqMydQa6wWh58gaq3IQIQ2RhoAE4Fbx5jvXyQyTvvUJxYCx7DuxzK/fPmCmUspmYWFg4mYc9ddXFx8+ulnk5qMxfZlnEuZ1UCExCTBorqf53k37kWEfY2iZvVss9qfb3a31x89fcrMqz7P81znUSISS59yn7supaIWpqZld3OtXQe3LuVgt6qRfOj6Jz/96SelWAWYEnUWrtMUpqucu1UGsHQxiMiJuw2tVqvLi4vVatX3PTOvu/Rsc8GM0KnrUuJcTVsZiro2CnHXdathtV6vQTRN0zzXYNpu95eX57evb/c3+/VPfiac6zRthp4Tp65jNOJgau2Fa505Rcu8NbVYmkW1tDERDhTZVrDJzDkP+P6phe+nsjk/r2ZRa9d1Ld99cXFxfn6uc3HTnPPV1ZWuho55nufN6owhydFL2u+n/VTmSavNasvYq3VKAMp2azLmnP/qL352db653Gz+5hd/SWFlmudxNFWrld27LOer4Xy9IqJ1361zrqXoPDEQHOS+3+5snlLfeakQzpLBidyrmlCschpS12ViTvM87vdTKROYpBv6ruuF110/9B2AWK1yENx2+2hE0rCl/1WE2OGezlm6oSeicZ7GcZ7KBGC3G00jpTRNkwStJfd9r3MZhiHgBkMEJeIsDkuLv6EBAE7e9hIA4KUuoVl+Fs6UJElHJH96tThuMc2ruIPiA2/A8gGEgySnPHQ+F2bOWWIfrTPy2dlZkQlhm66TwLi9nbbbeZ4YBENPRP3Qk/Sc9lGmqlPM6tH3fb/qx3G82e679fDp0yfPLi/Yzebp6snFkFNOTF3m8NfjHu5DkrOhv1gPQKy7bt3lXrpp2od5liQ5F51dTVGud2Puuz4PkjoCEOhzSn2mOmcmIDQCWqHa9/2q71NKHTAwd62XRsQ6pXmu6y6rqtZjXwMKQEBdWuZm55wl5cmqhkZETt3tzc2Ti8tnT66m17fjOF08WTX+cDf0xD7Po9eZCJKZOAKKWKZsAmhEXQGIBcJMy/gmTiKSSLJID0rfN2uBgBWtQZH75OBu7FJmIprnWZhXw3A+DF7qPO1JWEDTfgwzra7qczWdi5tBddN1s1YOs3kq+53Xad1ffHR1sV7lF1/NwqDw9dB1TCnl61evrVSY9iKbvtv0naoOKfUim3XP5F5VOukkZZOwCubiEwVCTWMWtKQ7C0FD6+wRoWWCW2J0idd9NwxDSmndd5JIiydmzqiz95zMq9Vq7m3uoTmcWHL2ABGlTrrcTbW05rI3NzfTNK0//vQnn/3oudPL51/VWrf7nTAP3RDQ/byN8NynbsjgAJzgsTj7zIdW1yll5gRmXsayJBahlFmGAH/f1MKDY5r2EbFar1mwPlt98sknT548cVcCuafGcEwpSc4ssRlWoT7XWorK7CKJpeZsnGTAUErZj9su07OPPv3Rj3509fQyJxbGs8uLIUnPvL+9SZyn7W3UIo5e8qYbzleDmfVDGvo85JRo5WptvCQLGJmEox8cAQ9HCISkUYGQkmhVNQ2O1KcI6rqUO9mcrYhoNXREMSKICSxM0FrrNM37MQCS5A4LgJFZauBYheBuIBqGgSDn63Mttrvdn202up9KKQS/PL9IvTRWPmdebdbdaihlFCIQeCm6bMPGmUg4da2+iTkFwJxSSpQyOAX/C2wiHyjn5+e7/b6aqRVEylmuLi+eXj29uLiYdnst1bphNfR0fu7udRwBosSd9Lkf+sG7uco4laKvbq67rlMru3G7Xq8/+eTjp0+v3E1n7yQ9fXZ1sTkT5t3N7eXlVZfz7nYvgY5p1XWbbtBB+2E4Xw05p9XQhaHorKUGpSxJckos6mZVg5BYOAkFR5gzZgpwRLBkd3dOHESSc8Akc6PaBLA0oTD3qqHW5miKgJEiMTMlQBIxSFWnaTJV5Hy+WV1uzvY3t8+fP/+LH//o6urq9VdfDcPQ6CNqsPCcun61yl0317GRP1rfQabUhuMhOKVMwkwdETkxMYMTsVBORpxKKceeYqcU38fpnCdzfk7z94+9Oo54Bp38bOOils+6+8koOluoCJtZL92s+uzqap72N69eI6LPXSsHbYy37X7UlOZ5BqjvVlXt1c319c2tu0/TdH2rOeef/vxnP/nJT548uQAwz7OZXj25XPUdIbTUi83Z7etr8hi6bhxnnYvVucvpYr22iPH2Zv3smSTJXdboyjhtx/2022v4Zx9/knPOK6Ekidhg5GSwsYSLoZKZQQsFrdeby8vLftV1XUci4XZ2eT5N0/52izYiL4KZHCCixFTUp/0cNJ9fXZHHV8+/cGpdoOL58y+H3JWprNcDTLXUp1dXUA3XonVzcf7i1Yub293HP/n0/OmTCkdKnDtuS8bC1EbjsIOlG1JKoGRmQtJ1XZLOGCFgRGoEfDq0ivr21L13y9fSt3zpExLuShybvm8R1pPzCwAtgEyM2ko7hG/2u91uV6um1Llhv5+qKYtA+PLy6aeffPzpjz67urqMsNvb27mM634d5EIc5l5bM0616lZbOjrgROauBgLcd7ttzl0Mw3q9Pl8/u9TLcRzned7tdiJiOedO0GZQ4FAY2YremRoBfdish826H7qUkoCqHgaSHfjfS7Ov1nKPpVq4Bfc9RwhxEg5JnKQT40AtZbNeDSnrXIiWTm3RRllEWHjqu83FpaTOHMN6I4STqY+tpSwJCYhdOiIBOSCQLiQ7WeoSEVKtFYd2HKcEnncs3rGipCnPe+rQ+zI9PZKQO6wqPNabVZbsGjDvUu5zt+p7ZiSWTvjs7OyXv/xlNduPc8TU+m6t1uvVZv2jTz9bn22eXF70wxCwcdyR8LDuV10fhp47bm2/m4OvoXNptV8CgkeYc5IsUudSSlGtKclmszm/2Kyn9X6/v7m5YeaUj/0bAYAI3ZCpaz5dGnK32WyeXl0+ubj84osvBOFmYRHqXs1Ka3hUWycuU23K6OAAr3KXU8oiJAIWIiEh67rb168uLy+HodtN01zGec5tPPMwdBEx1wqSzdmFpL5YyUPvi3lnAjGxkzAlgLvcceoAtFomyhmUKFRBRr6oRWvx9D40zAdfdlp8CODNgqrH6kceOTyCWSLcrbrVLMSROqb9zXWsNplYiXOWLuecztfrNX6Bm9vb/Xas5kKpNW8/Ozt78uQJM4OilGkax1pKEurXm1W3ouBESSSHNZolRVCbSNRJ7nOf09y7kUjf9xqtfGja7/fDMDQlaKhU63WwUHW95Rr88vyJemvAalb19vqm7Kfrr15uVusIWLE6zfM4lWnWonUuAEiYLCKi1BoRLL1kToQw1zI7SB0WXmady1jqxNgImVqZppj7oZRSSlmvh2CprRBiWJ1dPtlPO0q569YgYkjr4EltTmybuZYEwRwhItxlhmjUbt05kGqtRGRmD472e0wzTlf6Pdo+xkOm4uF3tbas0WaMlVqmmZxCzTXCVEsdzUuiIXfM0PAwJM59DhFjSM5dlk4ojft963IQYUKcU4JpG7ScOCdKcHiN1vnQLRAkkruMYbAzUzAFo2tzWzIzc+qTE9Q9pZSke/36da11mvbTPJYytb2PiErRruvW6/V6vT5brTdnq1U/dJLmaTItZZx0X3xSFKca5AgnDWt94lqxfOaOiGqtMU1azcJL0bmWeapTnS6vLiQxMx1b67TPTalj5iSdJEvd8OTq2Uov5hbrI6HBVkt3FApiSglySGiJoMsRCI3q2SJSK1U4Drw4thv72q3huImc6sry+G7JW1d54L1MxfKWCDAQFFbLuN0xBObCkkBwb9N5XNqcXL+6uLy6uIygcIoAM3epTynN82hmtcyqxcyIIjGCyTWiB5PYoWOEqZtFSp2TBjDQWi1quFp196Il9Wm9Wp+dna3XZ13XMWUi+vGPf1zrPM/zNO9KKe4KcoKs8tDloes6FiRiBlnRsdwC8KK2L3U/WalkLg5xmJm6llLmOqk6M4NcQExEbiBZuv+bh7kQMblrVSvh5opaJnftuk5EprEQUdcNt7fbL796Cab9OF189GMgNYWgNpqEiYmq2XEZhMg93DBV/3//P/97D1qsxdJ5LqV0aGL9Dm2gQ2eSZY3pXZEI3t+rOL7YnIJTSnAfd3vmJCShZlXD4O45S5cyRTL16iXnPOSehKLtCGpV7WzYjNNuHOfr61fTNKXM55v1arM2s3zwpQ4tlQlATn2QBLhD5Jx5Zp21lLqt26Q9SHLfdW6ZQMJENNcS4WBKXQaTGZtXtiCnHNSBBdJJEiHnVJ32u51N87wb6372VlZUNNTQOm340qipeakiklJqLQE7ZnIyMXSUQVbnUvd9FrOqc6Gg0HjyZK3uX371/Ga3p0z/4T/+p//89/+Vk1Tnig7U0WE+Cx8049Wrl0TUUPacc85iFtNcvvjytQUls1B1ERdZzKAwI+5y8UEnladLosU5mMiBr02fLuXSH1A/4uQEaUF2RK01iff9upRSlaDzPM8ps2th5nmqT548kdxlSSLiLbGgZmZffvllrXXcb3e3+3HaMYMtIuj8yZqjORYmApfQ5CRBAg5n5tZ5CB6llLFOc6tSntyqw4WDsSKRvF6viYIYEaFaSpmqzlG8R7/KXUOjtVSjsFKncWyua53nNtcoworWWquIkC+dmSRxW6HcSWh1bp2U2U3DKrQGt3I0MIiZq01lGpPk9XqYtT5//nw7jdT3v/6v//DbP34eTKlbSXfpS1+QRaGbWojQUS3aKEYzG6eS+zOA036cV+DcLZWKBJA7H4xFtN40rQcGEUUbKLFwYY6rfm/B/YSFIUSnfz9uIK2N7ZvPNKVMYFkuk7BwEFDr7O67mxt4dDnnYBsnSmnT5dUwdFkcVuZai2o4DKq63e/2+/329np3ezuXccjpcnO27tccqYwataTUSaIp6qR1dh2GoZZpO43jvN/tdtvtdne7HedpKqVfrXpC3dcx7TrJIpIGYV5oIu4OD6EU5EjIktQ05jZkRs1qmEXEtB9LmUabK5mRzTaXKJapzGO4dolFVtxiGyZrUHdKRF60MHmXkUCjVrNcZr/xfSfd1bOz882Zqb746ou56vOXzz/6yc+u5/LHm33tzkxy4RTemjRQqEU1AK0qM+e03ADg/RREwZxIcsuepmLOar26WgiFkTqRa0ib5desbDhYWgruYDIcOP58tzzaX+HowdwZkmAiiTem1BB7gIjhYDDzkFPfZRHJLJxyYiJqs4PU3T28VC2lvL698arb7fb1q1dD11198uzJkycAxnHs+k0iWGgptWgNsKS0H8dpnkudpmna7W/329v97XaqRVJGjXk/wylzS4V2DWA+RvXNW2++joBA7ETirmWepqlMU611HHeqWmxpt1JcZ9dq1SLafE9GEMkCPRHUKjFAHqZL29Boky4QzgTuurzZbFbrfndbyzzvxuni4oxz+vKPX7x4fevDyoFaap8WU9EuJzM7mMFzqS0aIXInJgJzMFPfSpPnqRDRnLIIcSQQty92GMLABzPhH1j4/GGybPYH0tjp8IqjzhBRTnkYhqHvRSQRk2SKqLW6oWg1hZmN+2m/31+/fBURr169evnixY8/++zjZx89vXzy6vo2hBGGMK1WplnLjAhhjLW2XgOmWqZ5mqYG9A3DQCTmvt/vg0C5Y+ZcCvHSdDGllJmYGhFEYN7oLDCvhzZ40zTt9/vWAT4igryNtKlaKcKCHEytWdxBGl2jbVJNHNwK3ymCmYdhODs7W/XDuNvPpaj7J5/9aGd4/vz5btoN/eDutWpoCyPuRn8vc8jdlviQKLixtphIanUAqbWmnnLNWYTBlpzdHcECOAhCi/tA7wwiHnEd5MABeeD1bzscRBS0dPlsm+DRw3V3OrS/bK38OQDmUkq1qNWqegSq6na7v7m52W63dS7Xr16H+tl68+TiIuc+7HVKYA8tU5l1nmvV1hgj3NWtWtVa53meXC0nHtJweXFJkqZabrfb29vbYAJ5v1oRo/W3WHU9JW73Y5tYuuh4kvYCq5WZVYuqulaLNvz3mDRgIqc35cF7htyZuZYKoEu57/uUuoO1ko5ktVp9/vs/vnz5UgEzqx6+jKGPCFpc2hbquFvrH83EzB7L3F53bzdkmkol4b70tctJJIm5SARZ+OKsn1SKfoM5qI0l+IBaLPNvDx37AMK7nFIzO7YHOyhKuGupbha1WlF1QFX3293udjvu9tvttpTy7Orqk08+Wa1WDOpz5+bVJ1Wt1dAml0fUeVTVeR7HcTfu9vM0hXufu9X6bLM57/rVWMdpnnc3104QIVUdui46F7RBzZ1Ic7eitfTEgSApIi2lrlrmeTaE6xL6tWYnVvVuBiBLq38EYB4coNb7sfWncm9uLBH1fZ9YzEwRktLm/CKZ7cf5V7/+7fV+TOB5nucIyr1Qa5Rtrc7X3eEuIuqtrbjRoWtPaxTZpwwg7acC5mHQqZaUOCWu7qws8sauT3dYxKOOwiPPv0tpTn2Iw+M2L/Q+uN5+Wy6NLtMGI6gazFGraa3VfK5lv9/v9/tpmso8D3n4+KOPnl4+gZOZdinvxglkrhrmnLqh79qIw3EcG1zYGoenlFarfr1eCyil1LUeNGjdmLztMhGRmLOkLKIRaPsDRUS06c7HuYftK9xFd8ICZhH3PPl+aY2HRsNZxJZ5SmRmTS1UtY0YYFBKiVnMjBE59+dnF2vif/r8D19+9TyAYb26KepEmYFwQJqOHpoVs7eGicdrexg7DaCFG6kVYkyl9DXXZDmpJa5MOe6gi4Nv6Acn9DsQeggxo8Mc1APh8E4tsDwTdhhe6lUjyILDqarN8zyVOo7z7e3u9nZX55pSOtucrVYrM9tut1bLPM/jOEruwlHUyAJgtaiu4zxPZZ5rUTc6EBuJaKqFSlF3IlmtNqvNMKxWAHa3W6uaghKJBESkcS+ooSfmjQg416KlqrYeYI6lMJ9igX4hOZEzDt0t3N087lo1kli04piwgBtAcexWS0Sccmb0ESGd2u/HuXacJKWYa+66WHpRLn7hohaRIiLJXclytHGzERHRkiGp1UUVtaJe3dSSWoh4+GOtE/+kwg/AZQfYgw4jgRtO7KrNWiC4qs1T2U9zq+Tc7XYMagg0gne7fevNPI67aZqGYQDLOM3qyP1gIeM03Y67cRz30ziVuaqZKYgCe+4GqqXdu+uzzfnlZdenUsp2u3X3zJJSYkJKCUuL7mbr4BG2tIDzo0JwEnYK8oiopu5+nPPuJ0MIzCznTEkIzIc+r8sFAnGbAt28l9S5u1hXnIppa+GoRatph77OM1FKxM1LO1xVby4t8+LW2OJ6EjM3dm+aSuU0D9Mwd3UW7oRzYkkopsvUuMO2EQRhpoerTh7PpDzW3yIOrtmRvdFo6Q15PDksYemLK8RLt6HWlbfWWq1UUo95KuM47ucyzbOZpdSt1+snF+eX55uIePnypTAlltZkWVUtUKoqCNM8Vd/P0zSX/X5q6EJDGthd3Nli9+p6v9+r25Onl5vNpus64l0bhN4GW5bVahiGoetF2HzhcbVk2XHwXSkl57zKaa616ExEmXNKCeFtDGIpxTzaBp9zLuqNpu0gPxluv+r6PudaK4NWXU8kGk4po7paXFxe3b5+MdXS5jTnvndttuZYoEwBM3fhrm15x4vfLFDT4FTUknk1r6bqqXpYuC8m7Q0D/ujC/0vJcidhUQtVrebqZBbeApXFLc9dR33f537ohnUWMi11mne7XSklJRYRDxprrUZOPLtPpc5Fp1omsyAKEZjP1YwqNNRjrsUjSpvikWS12rQaBa0VtWrOB3KQgQkcy6CwCKcFIm7OPw5jIoIgxO5eDzPu3F11mbxNSRZbvUyDeGPDPQ6ab41ES7Wx6lhtnmsxQ7AkaWFkhD9WNhxH3+0hSVOprX6mlr70rtYmqrm785Hp2ZqkfKsJFd+NuLsF/HAn1mpqXC2qW+uJDyYWycz9MKxWq9Vm3XfJa7mNuL5+/erVq65PnJKDitrsZMwO0cBcdTvP+3FyrdTi4QA3xDgnEUl9l7o+SJwA4c3FJd3etPnquVpXXUiDlsYVHHd4A4AgtJZ+cXA/l4CQAGLilgdks9LwiczEIkdevoODyOEA+tz1OWWWRExJUkoAl1Jvdrvdfj/X4gROwgG0UbwBOtqaaJa7jdltLicDaNPrA4hYQotUa51rnqsWrUU7VdVwR/JlTC9HRBvWtpiaR5TjUVvy3jbm4Ea0X4A7lxPA4tM1/hMOoz3cvXpUc3Uzb33QPJiYpVsNueu6oV+v1gQnou12S7c3N9s9mIyoeBSnYAnOIamYbed6u9vVWsOcPIgoCYPp4uLibHO+PjvbnJ2RcG2DDruOcydd9qruXmslDj6kXpqh9cOAi3b+zC0ts1zKYIK16ntqR2jtYNs+cmgs3C7sUcfiAGG4qpqGJVTzuejNdr8dp7kUMyOnQ4d257uW5zgc6Bjhe9OYY0VRkDdFSUU9qRatxY7jXBevp42CkkNXwWiN6z+4FJkfjGjjQM85eMMHOudjPAyRgMNPbsPlXqRWetOGAFo4kUju+tU69R2JcJIsedisN2dnm935WGs1naqOVafqo5lTgiTzGOdqzLJaCTF5wIMFWVK/2azOz1ebTRq6IKhHgOZaiHkYBhdtWJiocLfAR0fHPg7ATBsY4HZ3+Q4LcydtxtbiZi1eXWtlf+eNumprRAkgdXsHTdM0juPNzc1uv5+1KpzCFRFkIGm13IdrbER0HAkYEY0m4gd8NQ5wS6qm1bw1HlfNxZpmsBqL8PFbHU/0cXzioQ3sdPjd10mcGoe3rAUzH8YtOU4ud4AN6gQSBlOwiHS574f1Ovd9EFczogBJ6rt+vXoqz653+3K7s2qz2b6Uqc7WnJLUnV09vTg/z5K8qqlyeM7SdV3K2RDFNJi6rsu5K1oyS+47F7FSG4+SnWxJNBIaynJIRouIA0slz+JfB4CjW3o3g3MZnyOn2nMqzT00i1orl1KK7ufy1YtX2+2uDQzH0nA3IBF2B0+fIpNHC4TDLdqO3J5JpocBKkVLtaWpcUKX5G3y9/HL/GvJohOHWC5aY4Pm1hFzTinQdf36bJP7jlOiZRRlAEhdHtar1dm5p1c14H0nG09zGS0UqCWG1ebJ5eX5+bmAyjSbqiCkzaXh5U5aAijmlBKDKBhMlAIRCKhq8ILgnvqJDVI+3otLKEhw92pWa52KFl3mSwSxxQIMn9rFphNtfnMSMdNaq2O8ub29vr7+8ssvb3bbQCTJECHzlhi7x51r+xsRRZu12EZpotUu81FRUiOAVDP15ttrVa22dMlpqt6qDJrCPQaABz0SiD6ywI+l3Qx8/GuQU8AJHG6IgLW2qggPhJMjmMLhQRRJkjty8q5LfZ+ZQYLgAJMTInHKOa8GyV1Xp6GqBM5SfhIoHgG+vt2nrh+GQYQaVwoRnLKF8YG0Jm0uXTi5rfrO1awqyCkJhbmGmSZuGcm777f4m0crF20YIwO8WJAI8+ptwBYzHdhxFgFQ4zG0fwYat9vdPHdap2rks4/1q+vrL16+/urmdjdVheTcUU4EgwdBAnq3iTzCeuFAEMiNQCBHcCpRufJgtbpWt6K1LgPWwgMGFHWE55yFWE+aad5XiwNVh+7vNf6gEhwv3NFyoiVsSaJVx8EpiOAc7gi1AjjY3M0X5ysgzBqZKIiSEByulpINA3mUADgNYGiEao2UVucXr7c3JkSrROpu1dQZBEnrngPmZRwJ4cSZI0JDO+JeOumyCLUsnVPUWsOqVXXXPkufO6ZkUU2NrDWOPnADiFgERKYId+IkiQGomql5WM65lNlr1TIZWo9wiohqgFkQexBxCgsPCubR9Nd/+P3NXPv1xnx+fTv+9g9f/O7LL7+82ZokSb0xm4NyJwgKkPARaVp8MSAihAWBCD0kr1sRBlLqQZGam9ashR12MQv3oAg6ONUACMFovtTD9zm9+eCgJW/xLRZzQwc9PXEgFkpPLG9xcoY7nOEpJZBp69ENM29WwwR9l4WYwVQMzJRz6vvOI1qS3Dn5kiY1C5fc972rgzC7V2psBveemERIMgmHt5FVHuYdSZIkxAFyNWUlgpqSkIUiQsOTazCzUNclAhhvQLTLkpxAmct90QKGg932Fhsu3UkgzEEcQYQwRHUrZnMtJez1fj/qc6OvSo2b/fTly+svX79WkmB2FqcGg1G7VUUyaFnXx6zFsihLiBkApZZyVVWtVtSLepv47NmPcCwW+/OwOryHLMEI33mToDv250lICki4kwPBB8rP0iaODvMeDl+DwgPiBKFMwhaOIJGUUpaUrahZAEaExp0kB0CJUpd69JTRCSZxUXdOubXtTpIhHIRq2rg9w7BqsaWGH1daQC3ioJbCjmifmlI7fWuX7sHFONUVOuTGbPnt+CImktaTorlS2pI+8+xB17v9/PJ6O5a5+KS+necJlmWNA44gQNsP4JFSajN4cDKA8x1hwOJbAHCD2eJ4LgUOGhpu8YZm3H2V91aHey9/440P2RwGHCZ37K+7fkuq2vAZikNjhsQRosYeMDU1iwjpsoi0mSApdQfSEQON8AxiMFJHXeHKIaSk4Uwp5ValKe5wb7MPYQ5qk9o84B4eTsQH3hozUwCu6gogsUhKJAJbSDSnl+44bB2njiTc3e2YsgLacDUiWHgAhxlmdhiaZ0h52o+vt9ub3V6NjCRADIEsLBsnirZ9YKE10YGAc4Bi/ZiLeVQtGgvLzKod34PTIrh23tbi4KUt1zcUP2wcd8/cP6JLOIUfTcnhAS99nw5DAtGwuxDivmrUUs0jpS6vhiTdNJWIYAIlSZIJIA93CNDlLpEVFI4ambyLWiuRXKzPlyScmbomlj5TJDgtjJ92lQzROkR1XWJuLco8/O5adTl7Y+OEB5MfIwFhBIdT4/f7YkpaZnVRH1pSYNQ+y4PcQ8O1ulqohYZL1+k0lQgDOTNEhBOYPShoAT1AQuTkQSB3Z0hKy0wTIq21xtL0uQkd/t2tQyIiXcarqJk1naiH72mtc8CJ1j/ucj7scrz9+qP/88CLAwIHDG1+x6IQACCpC3Nx8og2HaYh/n3fk5BamKEb+vXmPAjTNCHY2c2sgcy24HRA6Kzm1SDU5ZVI9oMfraolSpgHSCRRIjCZE3c5mKTWqkvuztBik3Zy0nJjMFfSoe8jlgKL4xofvQo6mXzZImc3tC1nwRvARACxWWiYmatZadu6u4HMfXKfPVQElAKsOOzxfICXCBwcIIpYpssd4Pb22nc09m0uQ/KgcGsNWapaVVdrg7Jo2SDvbSIfKKcEjfbITxiadPA02wcIQOEMN+Kj/vgSsnIQQMKUgoMBZjiYKXVddkfMc5Iupc4jEKVWM4s6axy491bVAlPxaS5h6Ffden2W0jKPZuFalrnWJVYUEQhXjdymv7QeZEwgpDY/6pDabVNuCXScJUOHhGREtCcbuyeWgdBLaSe1aXVBy94BYm4+KxM5gtyg6tp8PotqUWrdFZvclakVZahHIBITmILvrjYDEXys3TpCJo1/Wh/p5LpsIhHkBiM3a320F/HWZji++wG4XyMnPLBAMloGmZoFO0VAGuGNQwCiNJqTcOo6NnW6oya0ZXC1Usq8n0opTS32s+/Gqdba9/16vc6S2hdl5jBXq20VmbmTxMw3+2m1WXfDEC1ESCJZwJQSu0dL97elXO76E1z4DtVus8fa9ztRmjbKxAkWbUI6gjjaDdBogASNpaTV3c1jqjGpjm3+qiQQ2BFLX5MjRxpYMKfFrzqF0VoW5mG1OMxNTeFEJERsAVWvVc2i9Tb3fvFKzN3cSagBc8evh5M94t5mcQqh3D3ZXtmOefhgYo6I8KVo3+fZAxYESUh9ALVGrbXPw2zzRb8edzfZou/E5jl3+XgaOWcD7ff7uZRxHG9utrvdrs6FiMo4X796Pe52NfDHF9epG1ojuqb1gDPzbre7unwiiW9vb9394uJi1Q+11s36otYqsg2CkVOSftX3fW+1gDzn3Pc5SWtI1YYZQ3LiJGACk0Rq2HbuF65U09d2BUop6ggwSSLziKjFXBDMBBmnuRStZvNct/sxIvr15nqcXt3uVe3y8oxIxmk6joawhvcSCSUQzMPcTtqiUgQ84OYAVsPm6IE2WjLaJFciAKnFgH4APpd4aWm9EOoLE7yB8QDkEVrNh0rf96pqVY+Oi7kqgoNSyizdbDGNVUn61eb8Yv36xct5P3WS1LDqemb0K055qHMKliCQktZSq+3Hcbfb3d4uJUDzOE676fr6end7O871i5ubYX12dfn07Hw99OvcSUpdoxbsxn0pcynl7OyMRW73uxdfffXRlTarCyYlAxFnSV3u+xywrusuLs7OzzY5ZyZW1a7rJB1cmQhVbfzQtqEc/Zi2lUjOVssC6DcONR0DFA8mMJvWolXNADL1cTYPAZFaEJmINP6DPtR3l+hRY992sXZvHMXd2/qmVgYcEbYMcIvmb97Fqs4S7GER30QhTkFxOsEttLo7HAxCPiSX3SqRTGZqLt2qW63J6fV2d/vlH7743e/mcftv//qvevjFkGvZp8S1VnCGBBOIw6zup/F2e7vb7cZxZsZq3XeJz9abq8uL/XZ3vdt98tOf5tX6bLXpVn3OXUrSVv3i3/5dI/qq6nq9Hobh9vb27OzsD7/6vA22gbBCrZXOMF9cnpnVnHOYCtN6veaU0Zh5fGdHF8/d7BgQniIZRC0wIT+UVkebrMw0z4VEgrhUq2rMAnC1mKYCkEhq9PK2KZi9lddeIEE6teJvwic4bihxoDU1TxItErEId3I/xsfeKKl6gGW+lct5+t6Th9M8t/Ugopa4d3cP8mClRLnnYaOUv3j58h/+6Ze/+c1vXr94AZ0/fvrsZ8+epNyNu9uEiHCTHqQRzfmyUqZSJnffnK0uzp9cbs4SSycp51xn3Y37F9e36qi17qdxv9/P075MYEplnLquA6Bzebnbo1VYWPz0pz/NKaWUnGnpl8MAk4aWMgEopex2OwC8WrexNweq5eJDtGrvliA9XuQTh4MjtHkYHoFAwCOoqDJiLtr6T0qXw6kUnaeKYOY2Fs+aEaq1ppQXnCSoeRVHt+bB4PG4Ls35WPxf9zabJzW/NyJMQ3WBTdrMz6MElvHZHwhlvXUqJ9EHJ0ldZlCzsUQkIAM5ceo3RukPL27/6Tef//I3v//i+VfTNG1f37JNu8mGs8sgJknMSKlTazdYBNrGPUdYSvzZZ588ffrR1dkFBeCeJQmlavq/Xp9bRCnlZru9vr693W3neQ6n1oJNcldrncZGCE1dyhIsWMZFOXkIQQhMs87zPJZS2rirVmaYs8xaIWDm1n0iiKJx8AEQBZFF1APSpW4BBFMsJa2NIgX3IJJabb8fx3EKkg6iZvM8l9k0QogoeMH8Y9kRBEtWNBB0COWa5/bgchx3ND4IgCgzmrUA4Dgqsp/CnfqmtfgGavHAOQWAAyvJvAWEfd/3KRulyej56+1vPv/iH371u199/uX1dgyWnHOkYZ7nV6+3JGk37nvOgHVdNxVW4MilVVUR6fv07Nmz9TCAvBbTUmbQ0rV5v+v7frNZX56dffr06X6c5nlWCzObqs5FzaxWHedZVeGNwgUwGSE4SLjFj+fnm6lO8zyP465Mo5m1MpPGVqEloW/tlBak/E0ko93Zx0sUy4SZUG89vGia52maiqkwqapFiOQsMqlptZxza4RFRPlA7+M3ZyQTLWbj7bU47nHtPBrAz8zNZCa/i2tRzSxcG9J30Izqlu0O5H9H64vH5JgoO/VNFjvkLiJd1+Wcx2m63r7+7Zc3v/3jy3/81W++eHk9O0c3gLgUDcPNze43n39e9H9X9vvzy3XdXZdSzGX2NhV9KnUGRT90Z2dnSWguoxZiCOC1qmlxyahhfY/1OqUkjj7AklxAspbdKKgASjZ47Kubx9l6Qy2/KBzkRrCwNmOxk8Qr7vtc56GU0qK+eZ556W6gJ4piTV+Py/CmHF0QN7PWEskd81zMIkknkjnJWddd9Ju9D+X5C41KAIFcjQgppQgsLaE9Fh7dyfLj4GM+aDnigLaZmeQEIHG4k0dr0xbUFksD5jCNBsO1Pz6IS36tUPiJHh3LnVnhQKIkLhEku3n+ze+f//LXn/+H//nvK7KCrdtEkEI8goj3082r2+2XL14HOIjW683r2+v9VAqkGqZSSimuRkSr1eri4qLdvhl8vt6I9Puq43a3N79Yb2ys+3Fbi5WiRauDSaSoW3jKveS+Vh3H0cJyn4sbKJi4AYBqOtVSraywAjxn7lZrWq/b9DkPrfMoKSL6VuNkhztqmqaUEpGYxdKT1RFBzc0z9xqo3lIQ4e6tLwNB+j4TJ87p4vLy4tmnL2d+udvrrjGpvGjNFMsknkCQnzpwtNCmTpORcZfcPrFeR30lYQBJdU8UzPAwVahZUZ/mauvNXL1UXxvDJYpDPOUcbiTcgFY/JGOY+eAMt9x4Y+yAA5ncVD2IJJOQQizYOVPuZ8kGeb3d//4PX/zyn3/969/+9vmr13OlQNSwGkap4wx4aC2qtj7bfPn8+a9/99u/+vjq5evtajjb3l7LcJYY5DSOs5mv+lXfDW7BOXWZyXwcZ8LoZoCr112d9/vp9evX2/0kIv2wRiNmSrbwcOKcum4Ak7ptd6MzB4OIWJBZKFHupZdBhBuEFK6t2QqZms47m2sdh2EtIiwgDqttOw5XYyZOOQVbLWrVwqexjFrGuVRzDQ4mc4tguGiptdSnq/NK5EFPP/3Jxcef9l9eT64Kc0Kz57njgB0AKyGGwyMIlNq4kEOCtunCkRUZBGISHNijCwZtAJCCfMnxswTBiZ2SBdcgDVbnau4aLggDJO76cB3sz9FA4fDhx0dEblWzJOLkIA1GY1pTSqvzr17e/NNvPv+nX//u91989fr2dq4WfFYwB5M5KTlTCqRAVfNg2o5TZ9Pzr17+7NnVze1uihpmTLM5aplCrTWNy4dWT3xgOcRSRxIAvnr1cprrdj+O00Qko1k4itZutWZm4gStU11cbmMgiYMSQUiCXFwiTJzCTESyEIG8jTuGMxEnYWbAzVwPPA9dQKWFyWGBCHJQAzcjwkHFTQNu4epQm+cqoG61SqmbS11fnF8++yhv1nk1pGE1j3tFBBO3/h6tzgho9F1aqBtLDxLiA9FpmZPbnMUT4OAtlzFppNYOCszOYhAFG6QaVYNaqLc+d26t+j0Y3tKudzrh7nxivE6iY+a8CubqPteAEEs2l+04/+N/+Q9/fHn9m9998fmXL653ozmDE0nDfCJimcgGd1dztS5nM9tP9vkf//C//W9+MZbZXBNT3e0CqPNsWgLN6ydhorBAu1hxiAzd3ad5tgAnSplbtSiIm7VrHavMbJ7mUgqEU5fDnMBgAQUIAQ8KBSwi58zI1Pp212oOELXiKysEQKvXYnXZRqiqNjfCLOZaj/GeubdkWPUw9dCwqqXUvuv61aoFCFdXVz/+yWc79c3mbH222d28Ngtp5SS0YO90VxBypNS3Jip3DKk4UqPj7pWnmhEeAJJRxy3vxHDplLMhFSQD12YznGvAHBpgRwriO5rdST/okyTNHZuEaFQXSoZkCUFSZv/tF89/8/s//H/+p//vWGNf3YK7fuPgolpKEeFggocDdKAVmZmREXO4/vOvf/v6+nYFh4e7z/MMeK0WWtG0waoVgTjFYiKxaJep6jCsg5Bz5pTMgpOIZItQtdz1KXXVNKsbQkRSl4vX1qgqibSsWoQDMNUGC7h7nacGRZCglBIazBoRbrBws3BEqbWppoabLvCRupUDJjBr1VaKoQYNFrAImCxUOnn28cfPPvpofPFyfba5OH/yVXpetVDHwnJoWUQ46SIEcloG0d2pxbFCBA/JEiI1a2HcBQMizOKSnVfKuSIp98ZZISWoBmpQ3+yeWZAs+9HCJnACEZpn+cZHOvFWbej6nFfjOD7/6tXv/vDlP/zTr371uz9Y8FgxFXNKJESc3DlIWnwS3oZrWaMJRMR+nJhTRP31b37797/85S9+8tma4fPIXolgZgFjcJhqmRsrKaVOWNp90wyGhqfUaj38GLyllBIzs5IgwpjRD1k6IRIwuTozhCE4WPxDoZ6qtWi+lBIRXZdSSuQSCD3wH9Wtue2qVt1Mm89P6qEeql61zrUUm1XVDjUdQRiGQSQVrSJ0+fTZR588G9areIFuGM4uznPflZsxkkBaTTofN4U3zQPeIH8/ohOnvmdbwWTcQVgkR2JP2aU37oyScVJKtdkMp9k8m7FJAIm9pZ5pUTF6WwcX3gqkO7sYnf/41fU///o3//Uf/vE3v/vjdj87ccqr4l6d1B1qYAIoghrBzxwWxkJ+wFuKGacs7Ne73X/5x39+evXEVylq7UyFA4AQSUpdYpCbVWa4syMoWO+aCCRi+NGcMTHz4hZxg1kDAEnqUnJQhPXpjhrTFs3j0ArCPcLNrOlW1+XUdYxco2qxotXMVX2eajEFSM3C4Agi8VZ06j7NcwP0qlks0DmDKeeeOe22+8snFz/++c+uPvn/lffmv3Ld2Z3YOee73HtreSvfxsd9EyVRbKn3tls93S173I6dsRMEGBhGBjDmh0kwyV8VBAhgBHGczEzHgacbVtvutqRWt0RRGylSJEXyPb61lrt8l3Pyw/dWsfgepfHvuSAK9YpVt27d7/me9XM+50QEASKbF/3+fJZ33PAwghAIc9QTPXHEiKQLn7LXzaiNo8IxNSXpLZoxI0WsrFIk2rC2gbJIOkWJXsgzOAYT2EcmH7KW5Slh4AQAZluEpsAZQQChgGro4faDLz795Pa9e/cfb++Mykrbztz83N7+kNCAMloZBIqQ2sRBGAViCByFFZBSCkUp1IpMkApJC6lbn98/f/6sn+93SIL3mQKtU/kqy/NUGgUiQE6oouB9iDGi0pkhQGSQCJHFRgattRAJIya5wcRzm1BuCEBpNGRrklkiSIK1BAkAIIDK2Ewr0zIn0tiVwoqBA4OP0vhYuaZxnkgnFkBJgZuk/3XOuSZ4l5jWtEKtCBUJRQBrDGk1N794/sKl3sJiHaLKrNK20+sXRTEkjahaKVVaYoBnnUcCBHnqW3yF7Zh1L9KjFiJBDUqJ0kya0USyAVXE5HtS+nlOi2NRHJWk9i4RaRE2T9N2bVoCGSnl7zyYjz+7987NT2998qkgdedPmF6syuZw5JTJARWiEYaQciKoiEhiTK1jwizIwIITWFHtPCJrxIdP9h7t7geOK3Pd3AcCRQZRa53ZVDFP215EYvDOhRAYAK02WmsWQZBMG8lEGFERMzgJmgiVQlCCwNwuntKYYPPJMWTkVGhmhJS/Srm4LLNZlimlWKErhy6xvMcQo/gYfIhN8FPIHypSKiUCvPPRR0451hQQIilShkDFJoLRaGx3fuHE+potuvvjobY5amPzQmc5pC46xNR2Mp03OuN4yuyfRx3SZ4XkqLYQRkiBjjKktKCOjI4xohYiUZaJInCI7ANrQuWjAmxxP239V5RSPnCn06l8GI3LTm9ufnHp0ePttz/47bu3vihFL66edCEGzyKSdXPLGKNEwUQ+qphjEBZgDlapEKJKVUIywCjAOsFHyDShMnlhM/vWb967cvniwktXQ+ON0uhFaVHaoNaMYIucfYiRgdlaa23CqLQYSSJK40ZDIm0nIKO9i0kPxhaFy4kU3GhFRNpaVNQ0rqwqpVSvKBIwAAiTIIYYy7p2Elzg2gfvQ1k3ofFN8J5FgJiBlGGGxnlEFsIQufY+Co7rhkWKfg+tdk1gkH6nYAhb2zsvvPDCd3/ne93+3H5VNZEDi7E2y7L5hcWHqKILKtPRhVlNgIgaCJFkgtKdUurOZjwTQixlNlPyego0BACNEgkMChAoIk1akbZKK0YVQTEQY8u8EQRijKLIew8SQwgmDYciQ0oVOh+UlYCaW1pxjO/e+OTGRx/ffbQ7htwRIGXKCJJQ64uBEmKWFtXhGTWjMIJBabGJwBIhTj0YIgJCBnQsDBJc2B6MPt9+stnJI5AHUoyNF9RBI2ite71ejDH6GALHICKtWFhrSStFBhETc1wUBMIYJaRIgWMMElIfisTC6Jho3klRlqe7SUoZZjQtfNI5VyfSLYWN8z742rnGO+994513MXDUZAQIMNmrCAw+igCxQLfTE1KskEVlhbUm19ru7g/XNjZPn71INnMsyhiSCMygFGlrsgK0BR+BFBCl3zAhTE6USFGAcOJu4qSX6bgRmSbF2ywnEgBojJE4aLBGodEqMzbLsjwzigwoAlSgAEgEI4v4KN57oZbLQ6EBgCjMARkpMlKWB8webj959+atj2/fHzvmXIHVRqFSWiukmKD9SVohJo6wlDZgFCGJiVcjMjNBFCJM6H+ttbYhuIgqivgYtg8O8/uPuqc3jdYFalFAdeOCzzOTZRkZTUSatNYCjAgqzWrDCbU5UfpVLIKCEAVdiK7xniMAxTYs9gaxaaoQmBCzzBijo0jrmSY7F7lqfFlWIQTQUFVVHX2d5pY2wQUfArsQekXGCAIoQBMoi4hgCEyZAaVZWGmb5wVHGAwrUNmVl16+dPVFYwsfAa0lYfQeSdus6Ha7Jst9XYEgkQkhaEro7TbD3QYBE9+itRH0FGw121888yKrRKmG7EmUQbGG8swUiS0kMzrTxqAyREqQBAlFYmRuGk+Z1Vk2Ga2WkNgSUYreQs1489M7Nz68ffeLrSqqQNZ7IIhgSIBUG7AkyiYCkLQ8MeUlYjt/LbFXtAFeCAQqIZlIG9QGEEkTM4183No/uGcz4KWlhXlGitKUGBf73cVFqutaASpUxiiF2pjM6EwpzT5opZRSAhCjuBhiSEQUiMIIopHSEIa2UOicKI0cAYkAM2MCS+AoIsG7GIIPIWkF51xoZFhXLrrEexSjBOHADC2Yvc0txhib4JGREVBlHIEFtM2s7QhS5epxWb34wktXXnhpZXVjWJWiNJAJ7IQUoFJGd/tzWafwg/0ozKnZCc1MRjENMkr+7Wz1+2nMKG031DNiAVOIrxGvwVolGVFhTJHZTm7zPMssZZq0FqVAUQQCAQaByIHQECCH6GIAIGMtagOoS8d3v9h6+90PPv78QROQVV553+v3ERQwcpgkaEUIkENARIWotWJCVhQCMjOTAFhgBMbEQBJARKT9eUIRERUpYxhx5OT+4x2jNJlckEIIJKEoclImsmhFLdk+mTQexijyAUgEWaJw9ByDTwOBYpQYvPg0zi0iAgGygEKyxiiimEgLEZVCIPAhOOeqqonMgYUjuCaUvhq72klMaAFhRFBKKVTCAi1hHggjpZYGVCrvZCFG0EYXWRO5GtXW5mfOrX/r299bP7mJpIGMMjmj+DYjq4l0p+jN9RdG24/jhEmMSQmBmiBoAQBZGIUm9ClH3ItZPwNnQpUUyWsFUWO0RLnBzOgis3lm8ywrMqUxKiWY2lAQERAEEhwoEZMRASptiJD0aFTd+vyL9z+9+8WTQ6EMMhMiodbGZJCgNy0/4QTmzwwt22K6uIiIRCCCWmuw7dtijCk3kHoABNFFRq1Ia1QqSNwv652DkbaZj6GXqVxJ3TjnQtHNtdFGGwKUKM7V3nuNmhK2GIihRbqn2yIhplZnAQBOsAcQjppIkYlIPgZAQqWicPQxsf1WdR1BAFXg2HhX1k3laz9BtSEoVISKKElDiC2SPpUxCIkUgjLGqswGwdHoIDKcPXvya9dfu3L1JRGsqgqVJmUAIqCKIqhVFMk7xcLCwqO8I24EwkIpzQYsCerdmhIBAAbGFv+NMwJBrUGBqZ6YSgkAaA1gUDSyJcoUWa2MSf8MCQCEtjEgfS2ARpLEvxaCtbmIlGU5rgc3P71z68Hj+08OWHdsd4EjAuiuMRI8AdKELkiYW24X5sSVF0IAkQS1SmRniMoYQkQScM4lASKljDHkKMQYBLTSoDWwEpS9YeljGI87q4v9fq6Lkd7bP7RGIYBGAqQYJDovgk4g1xkBCgODxBijAIISQmFWiKRVMs0xzV1hJqNSopqFtVGk0DuuqiqRgjMzg0Tmpkl2wzvnUp8mACAICSGqKKC1TnlMohbIl/z/EBFR+Sieuej0lk6sXHv5+ksvXSuK4vBwKIDG5h4ByeR5Z9wMtbIAmOed+cWFoihKVwIAKTVZn2QjnooFswA9B7o3gRw/NTFtGCLQrjIiagJFbCha8gVyjiFHRIjIQTiKCAMzUgSOqViEioxReTdAtn9YPX6y/+ubt+uIojsBdFNVQYzOjTGmrirAFliYAqJJaT8m8tS6roVDose21pLKlErzb4mtIQkYMAJYawN702Q+RhRUoLS2KkaxMGyaw3JUeUeGIhW2rB8N9k3HuiwTkYw0BuAYkQWERElkltQ8F4OIpEmQITKpRG2WeHhTGgUEyAeoG88gSltEFUJVlmWSVyEUxoZd5UMdgheIUSJLaJOqqVIaRURry8whsDFGqRZbi0SKbADyPoA2G5sbL738ypWrL80tnRiNRqWP2hqV5TE4VJTnORwOk0unNXW7fWttCQBAWhmJEZhjW4OYZG4Tno8ZAFMz4kQ+iAiZhSiRx0kiomg/A6BBG9QKFBIGA85KlTPmYvNoNREpiTGCsKAOCiVybuzYV1GUzrKyoZ2t/U9u37tz7zHrPKIW1IJkiAhEQl0NKpIUJOFUXzHzlFfT+dqHBgAkgKD4GLQKRGRMZozKCmtzFVx0oVGa/NDXdZ1YbEPtlVCWZbbIHKpy7IdP9rZGBxsrixc2V2ol/ZUVo3VP28CCwSuOGShSajA4RDUpgwjFlhO4ViaLzCGIAEQQSCPIyVRVrBxXjgEZfdTS1FVTO7e3v49Wi9IeuHShdGEcxXv2AoAt+JsxVacAUTkXsqxQKsQYgSXNwrFFxirznm2vd/bixetfe+3MuQukzfZhCaTU/DyLOAmRoGnKqilzq2p0mVLModcp+v25gydbQMa5qLVBYCBhAZU4YwEBmDi2HmhUgKyUQm3SlMPUoKQQASWhRhgEjRWANE9AAJiASRrNxoDLAVWIWZ4bY0KQ5EABGSJzODoocquy7qjhrSeP7z3e2d4ZVmyQszhJoQAkuxMBIHCbdZj2QCqFIip1WSVMushkWBKIcy71VotYY4SISKMlO+26SQKtUWvUhNoF71g8qkagqZq4s9eE6tETe+LECYVUkA0cO4LdTt8IVlXFzCIhFTJwUl4SxKZpIhKQD4CtO2i0ItuEGIWEVAhxMCqRpK5L7z0Zzah84Kpx47IcVmVZeceOKJFWtHkCAUhzx2JkCKGlTjakQAMhGh2F8rn++XMXX/vG1zdOn4silfdZp1c2NUMUECaUCdGaQkhhFAJba/M8B9LgGa2BpPaFYRKSJtPgvQeQVCFXihKvFCbakETFkdLkSKgkCiTkqtYImkBJIPYoGQIrxFRKtlpl1nhsG9gh+IggQE1A56onu4O7Dx4/fnJYe0CTtw1Jz4BxnnJ6PBsC8TS5JjPA1/Sndy0Psvd+2jEAAMnJnUoGTuarkTImS9xtMXooy+phPdpFXlv40AIsFt0EkstMBNJNZFA69aUDgFKJ/Q6iiHOBCVOnLyeES4yK2HtBIFTkm9A0TYw+snfes6ALflQ3g7I6HI3KqmkcB4wm04zclpNTBQsVIjJABPAAWmvMLMYIIaCxHNWZ06e/9tqrFy5f0qY4GAxEhDlIixdppwcQoibFiMaIUgokFkW3358nY9g1WhvgdiQvQJr2kSRjMgVGMHWVCjBC4hJjFEAUhcAomAhWgEA0IGpFTIljhBkkEjAKK4TM6Nb+IepJH2OI0Wad3f3B1vbu9t7hsPSgbZZZSDCqNLwLAFrC6+R1P41/JtLQElnGlqsQ4JmMSot3res6RZXTLsppd+9UmNonymQZkOp5p2Ij0Zd14Pc+uLnY728srS0XHd80zu11tCUBIQzc9uAxEABJiL4d8UGYNjvH4EFCRGQkjQghRudj412MPsbomMumrpwflNWwLMu6ScTMWmlWODM9PKkkRETSOiHEUBs0mgEjizZmbW39xevXzl68EJlHg4ExRhkzGo2QBRGQUBAVkkZipKhQa1FKccQsy+bm5vI8L8sREYlwYrRI2VtqxUJltmAQaCe1YRSUyBAnxI/CkYAEQBg5ApAyIIKagJVEElHCxJE4ojBINJoQOAYHINYaC1A73zTxycHo8ZO9Lx5tj0unTGGKnoDygZEUtoTdCWva6gaeNCqJTPueW1SSTJBC8Gxb3FSXtNiWiVS1tmxGhkQkrSgIMyQ63IyjB5Lt3ebjW58v95evnj3T16Z2fsi1URqBkQUAWmUMwEFCFCKFQKkVk2P0IbrgI7MtCmZumrqqK+8dg4QQa+/GZV36ZlRVZVU1IQIpIkKNHhim2wMVIAJpQBRSIASEoK1HdCJoTNbtvfrNb5y9eCHvd4fDYWDOTaGxNaytdwCIAApJEWlKtEfaeySt+vPzvV6v3N99eicnZLAtgTQAaZMq2i3ZwgTeiUSAnCrdsWWF4Cn2T2NSEsIITMAiKeBiiYFBSBsChZp8ZO/9aNx89NnnVePrSFFlDDo2CWkG1urEztBKArSBKDMcWcgkE0fjoqnpwacGYiouIpKgLlO5SZ8KAgLEkAquIKKENOoMBNDU9x8+fhvfDy5ePn26Z230LOWYXaUJU9+pwibN3wZBkUhKUFEUrKOvG+e9DxyV9y6GpqmbphGIANA4V9ZVEC6db5wPnPYpIEKKS9PekJSNxpaDh0WINJEC1FEQSPcX5k+dO3fl5ZdFqVFdGZtrI2loSGodaLNSkOoMqEARRETUyiQy8F6vN7e4sL31yEfRBCDtWIfJjUIAiAEkqQ7SgqBmMJxCadiTAEchVIlyRgRAtASPkRBYtVwTkrCT3keljCIDAHXjDgajnZ2dJwfDnf0xa4MqQ4UxgE9wVaU5YYmZJ+RgktKrNBmnM01mTyevxBgjTyHLmIQn2XuaNECn0kR6c7IpMjPDN3X4RgZgBZT47hkFABSBGjl358FjEXAunDm5kSvDjQMXMoNGCHzABEhXRmnrvRdyCCowuxBciDHGAOKbioHbjg9hEam9q+qaERrnfQwBJPWcpUoAKsI0iE+AUAlimkHGgkoZaqF1qtez585e+NrXv1l0OpWPgYU0EGBgl36vmtS+JclcS/9OqWkq3Z9ut7+wsJhlRVOPUvGjRSgAAaRhlpQwQ9TyZSRlzixCKpXtEDAx9QgSoYiECBA1SBRJ6fpJ67AgM8QogMoY64IflW77ye79h4+3d/btwgkR4yKLCBBppdNCxjD5DkBstROKiDFmahR4ptEqFR3SZOKJ2gARKfKcJuMLeHKksGXqT0yb4xgxMnOUkIQTCZWVCIDsxec6K0O482CLRTGqjdWVDJXJDJCKRIkimQQzAxlSE6LIZBIeJ5JoFMKyqlAhCzuJiVG7Cb7xPnB0gX1kTmQlKg1NSkzM0Lr3AEoIBFCIlDbKQpszUL3ewplT51588eWt4TDv9Ji5Go+BpZPnrON4NEJUqSCOgCyCknpCA2IaPUTMnGVZt9s11jYVAhADK8GIApC4cBAAgBKVMwIk8HTy9KMYQkrtj9KO3wNBBCGPAnraXdiO9GFhSYU+bFwcVgdb29t3732xd7CvtF1cOTkMyErR1IVkjCEkKyqJrqrVna2RS4h451wzQafF6GVmlNfUsU2bP42TTNfDzIiY5zkRdTqddJIEmEi6RKVpUILIKAIcKUH/UVRezNWjQTfraqtvfnp7bzD83e98O9e40it2Bnu9Tsdqg4hk7Ng1h1VF1Lb0A2JMPnOInmPtascxYahS1trF2ET2MTKSKE1Tp5KICKPzRilBYuYoHrVkRTfPO+OyzsgCUtX43lz36uWrl85fgghK2yZ6iUxaE0uiZ0lj5iZWFQO1TPKKtLVmUB50et39/d1Okb3wwgt3bt0a7WyHCLbIKLngiHVdA2O/3/cB5ucWHty7m3U6J04s7TzZUgrzzJKCEBzHoIyyWgNA0zhXl/NFMdfvakLNDLUPWRN8zoBKaatNNqzc8MnB7v7BYDRuXLDdBSLtBJk0AyIhiCAL6ta1MQkyz5EnbGHMgRkaF5K2kEmjo8jTed3JWs+KRRp/krgDpookpTemHmgSpqROkNIMTxJRiFFQtW1NiDrvuBj2xxUrszccf7G795Pf+/HW3dsj54ZVlSAmnSxP3InRO2ZOISUqiiIhhCb6YVWm5v0QYwjRpcQ9Uu2DMcbmmbKGBDynGD4WWQ4AiKqdNq20EhVdXFpYFsDDw6EPvHpp/ZWXr68ur+/uHEi/M1l+QUycYUyTaQ8yzayk1iBEItTaOh+stYhISi2tnLj3uYU23IjIQEpprSVCiKKUfnDvzsLyiX63s3+wm1vV6xQsoSqHmdXaaufrajDUWi8vLi6dWf+LP/+zXrfQDBA59YMQg/YspYte6ie7eweD4d7+sIls8yIrMibtEx8PkpAgt00Ak5GpnNLkMUaWGEKIrf5v47WpQ5Dc5DYSmem0n7gKLa9UWvipTyoyqfdMBokBQIye0ExGsaR9IoBKhBsfFSpR0DRNp+gdjoef3nvw56c2V04s7G892nr0eHBwWA7LweEeshCg0Zo5AJBSqIwWhASuGTdOFJqssN0CXSjL0rmASkcka6zNO1mRU9u4EEgmExSEgJBICVAUjAJK6ap2SpmV1ZUXr147e+YcKu32DxS3fNmt593eoJbJNd259hEJSJAoy7K6LouiG70zJjtz5uytTz4qR0MRSdUcANHaIkmMUWsLgEWR7R/s1qPB2TMnR4P9vb0n/SJDVMCUKzh5cvXVV1994403vvHa9cVep8iMrnywSBkoUSYADMvmsHQxyP7hgIEw62ZAEakK6cK0JMdkGhFAmqjDddNMPFwQiDwJmVIH9GxeK8VAqTlR4GnqogVX89OxBNPu+ukZZjNgzOyDN0pPyToAABhFJAIGF4TD8uIcACijQ13e29r69O693//+d06fPv34i4f7e7uD/YMn29tNVUbvrVYSfVJpymoRqZ3GRseKahdSKxgYY2wuxCazc8Yoa6y1SinkqAAUWUOqqRwCCJC0URJpICI1HIxRm7PnLrzytVcvXr5SuhhC05tbGGO78ml3MWAaA6MQIwIIRGRI6WxK/bySelmVVSE4rfXm5ubi8spoMGQJhtQkRyxaKe997evFE0vleFiNB/P9Yny476rhN195aXNj9eTG2pUrly5dvrh58tTC4nyn08m00hKMVjoyRSYXuQlxXMfau8ZH55wLQeeFtgWRiSyR8emMRInIIqndnmPk1PYY2gQDIcpkN4OoSa6Tn3KFPkNOPvscAOJEUKat8emzU3qFqaZpLVX0SboS2aCkpp0YtbbjQUnaZrk4X5o8G4+Hf/t3Pz+9sXbh9Mnu4mLW66xvnFzdWNcEvqljcCF4iR4AhCSEUNdl3cRB6XZ29/cHw6qpQRuV2U7XGJtlWUbtSNkY2znewsxAyIJRIMQQoggpazrKGEt6bfPU9Vdeu3z1RZsVg+GYmbO5PsR2O6VYg2CKUaTUQpX8RUltvwo5jWOc7Bel1Nzc3MmTJx/cvSMuQqaMMa5uYoxWWcRQl2PT7zTVeG11CWOzs/3wJz/64f/w7/7t69/7DkePEgFYKZOWFqJn5xywRmWZlI84boJgo4xBRNA2s50mSlU1qEHbAjKDrAI3yI6YOTIk4RAhARDIjAUAQWBM7a0ShBEwhkldfiZpMdUcqYV/qn0AQFk7Y1baRFaMMdWtpx+cqg0Rmdy+mewIY1ZkY6VCiIFj1fisyHQMv73xwftf+3B9Y1UZC8IMQpnJM6sUAmScBjwGF4ILwMqiQZqzfZ0X88sOiEhbxzgal6Ny7GNAbt1MpS1xAA6BIzPFGHzkhOnNik5WdG3WWVnfuHzl6rnzl9BmZQhUZARYNjVZ1XIRCadmbiASmkxynLQSC7SkjOlnWmtDaIiINBDImTNnPv3w5t7jx4FjoQuPbsqoQQokRiL2rnLl8Pvf/fb/9O//3bUXLkNsukaRLoCDq+tyWIZ0e32UGDWSVsbYPFfGApGgAtIJqQccQmABYY0kJIJRCENQECCRVEJKqiNNMmwCScYTtKKFE04XeJKQoGk2YioWaTEBQFGYLm7aRikqmdZEJviRqf/BlPZTopIBTMi4EEK3261dwxyVtkKoMzsqy49u3f7BD15f6HWDBO/r2jsiGI4PrdEkLBAYYoQoik1GOs+KYrnoLhSdHllL2tSOH21vf/FoazAYpAowJJYT0hIligThJvgo0il6SyurS8srRa9Pyp4+c2795Km8162c98JZXoigD15P+lCI22wPIpImkQljLaXsAUwdEBHJssz7WmutFTXleHNz8/Tp03u7u+Jq6AKRDiEwQkwZ6kwT6v393ZcvXfiLf/Pff/PV648e3EXf40x38wKBY12HauzrJkZxdQzOayalTZF15jqdjlIqtpMaOTKjNl1rAmCI3IQaEQmFY0SJUzZpRYpAEYprQmKuTGX4tKwiYrSetQUxxikPioiIEGLgCIIt3XCLu4FJfQ4gpZJkZoTT1IcFBJHImNAEIgo5QbxRxuPx0tLCuBwSYaffa5oKAI3Nf/3u+4939nrdeVv0tM2I9FynUEZH77RJA5ciSwBEa622Wa+zPL94Iu90AwMpo3R2tiy3nux+9tlnBwcHT55sDQ8PG1dLBPYYArLSNQgg2LmltdMXN8+c7fbmGLA/vxDJVlEoK4yBOngRsXmHQVBYIPWbsWq7BZUwIzLjtM2cEKGdrywxZYNQGaXIhbi8vHRi/aS2H4WqYpBpZo9SdtKJAjixMP8HP/7hqy9f3Xn0YLnb7WfGVeXO7k50PjG+N41val+WtW+8XljZSP2XY8cpg9AW7RJLdAgAQixmsn2tJoTWi2QAiTEAAzBOSp0kksReT9IAre2QNkPNSEKotEkkMSwkhNOwFqAFlzFH79rwVWc2xujqmplzo3Vm22HfIoipWpxCOkFCUCggRnA0HiQlVY9LIsp0hgb3B+P/+z/+7f/8P/77U6fOlOMhgp2bn0PdoWTViZRSZKzSmrQCpUDIk/IBBBUEYu8FzPLKyYWltWE5jlF++ctfElFd19uPHh8OxyGE/nLBEX74B/+q6PScDzc+/ezGBx++98GNP/uzP3/55ZcvX768s7OjEYuiM6pqbfIYBIgya8W5pq4IsZNZ7z0Co4QEl2JmjCyBNXAnN5Vrsizz3rvAnbn5OsjpsxcXTry3MxqNDocrq8u7O09ChNwoCWQUjA4OXn3t+u//4HcWrH7y8F5ncW5np7ZEwBCa4FxoHJdlNS5r52pjjDZ5MU0a8FNYsKg2bk5FZkBIZk4mLchTx78Nn562QE5R6TNO5dNI4dmGOGjdcJmyoLccHc9+fPrBlMBIAWqb1eBjND3IE50iE/wiSEoUAsaAN25+8g+/evuP/+gP6iZm3TnUGdk8ebiEGhWBykQrVrqdRzG9ZiFASHeJrO3rzBj7g3/x46rxiPj48Xae51vbO9bahcUTT/Z233r3o5/94u8QVFnXT3YO/pf/9X+7fOnSn/7pn3zt+rWiKFxVE5FzLhU4fOoyi8EQ+qckmwQpjSGJED3xSING8kSkFTAmTou1k5svvPjSzqNHqMAFn3cy4Jjq5gZhbXHxX/74R2c3N+vx4PTaWqxGoXEBkAM7F8rKj8qmHLvaNcoQIutpemB2JaaPR9by2Io+s35TD3FWGo4cRyRj1ihMnuPsBcw+TsWCpiHRBPL1ZZc6vSScSLwy+t69e2+//fYPXv8egnSyLIJ0e3N1XUJymUFQODKQAEriypieqsWRAEAC1Ii49fX1J7v7WZaJ4OrqqgD9/Oc/v3P3/t379/q9RSLiKN6717///e3t7cPDg7/8y7/87Pan169fP3tqc3FxcfvJbsKyMbPWZHUBMTRNlWlz7EYJIBMRsSYSpTiKIBIDIMHcXO/K1Zf+4RdviqtGVdmx2gfvPeeAoYk//MH3f+/HbywvLB5sj9sqdgghcoxclX7/cDwYlj6yoOp1cgeiZ0oSOPuEZ/hmZ6/vyBockYzjgjVbCv+yDx45cEZpTT+YHqeZDJkENWqGymP2kSZE/CKSzpT+yxjTNM3d+/fev3HzxatXvBEJvtfvcNPOR0tpGIiSskpWZdPPzopdURTee2ZJ87RFpN/vhxB++tOf3rx5s6pqQ8qHZuPkqa0nu4uLi6PRKMa4sLB0586dqhzdv3//6uVLly9fvnD5CiKGpi59jQLWKEbwvnn6kzExnQO34asgIgEqIIWaIWErRGu7trG+uLy09/kdJu1jIK2CqwHN4lz/v/vTPzm9sR59tbi4eO+Tj1YW50QAAX2UsqmH49GwrrUpik5RzC1UrnnqDx6/uc/df1+2urOvH1/O5+qML7Evz4EpH3mbTKAbIs8M652V7GnK6+gXKcrz/IuHD//DT//T5qmNosg9cNV4pScOE04nEyLQU5kTmaU2bH+XMebw8DAvutbaojCffPLJ44ePDg4OLly4+MXDxysrK7u7uzHG4XA4HpchhHNnzlw8f2FvZ/tv/+b//eC3v/kXP/whI5w4cWK+18+yrCmDc04RZVkGrR3hlo5MmBBUgpKLYNtTTyAoaXw3qX5v/sUXX/77rcc2I1ePe92MHVpNP3r9+9/6+teJJTifF2Z6T4CQQSrnS+cCCBEGwvdv3R4Mx3raojqTPwAAaMnbvtxq/BfFAo7J0BHJeK70wIy2mC6zzMw9gCl548xJjl9n0hZJMlJtNv2ZxtTu7+3+4y//6Y033jhx4kSyIK3zkJIpMnGWZqYaI7ZOU8LSlWWplNLaMHOKCwaDwS9+8Yt33333woUL9+7du/bKK//H//5Xr33nO/sHu69c+9r84uJnn92+d+9elpuNjY1OJ3/08OH/+Vd/9dZbb33jm699/3d+99y5M8XifF1WwTfQjkd/Ku7TgwATXohIkwiTELZut7X5tevXP7z5/ujgCbs6GLIKz2ye/OOf/AFxrKpSsd/ZPjixtNxUVWQgQh+4CZ4JydhBNT7Y2fnZL365PxzpWet75CKOiMU/0wocWfLjFmGaiTq+uiITSMDMxaQnPBkLO/stiCjPorzgqR185spnfggrQ2TNaDT6zz//2drayksvvYTaRJd8BUg+M0/Y9tNVTb73qVgkhK0ILC8vR0ERefTo0fvvv3/h4rkn20+WV1Zu3Ljx3/7rf33v/v3r169/+OHN8ai6du3a4uIiS3jw+b3RaKCUyjLz61/83aOHD/Z3915//XdffOFKURSI6FzNLUUzAACCEGLCcWgkpjQwT1iBABIRKBEEY+2VK1cuXjz/1t/f05lh75eWF3/yez968eKl2NQQQox+PBgu9nulO0x3ybN4AdvpAMvDLx68897Njz6/Oxg3NNXGaYel0mWqzv3zVQU8b+s/u0hH33bkgGNi126OZ48jPtDUgTh+HDdD6bGuawDo9XpZp/jPP//5+x/cHI5LFiBryWhlDRlNRicakyME+tPvEpEEJhoMBsaY4XBYVdXjx4+3trYA4Nz5MyiytLh447e/2dvdZR8uXbr07W9/Wyn17m/eOTg4sFYbY5yrB4PBxoULOzs7P/3pf/zrv/7r375/oyxLpY0xWUp4HvttQgoIUE9uzKTgbrTWvaKzubFx/uw5kLi0sJBZ2lw/8cc/+UlhaLHfs8ZA4E6nU5alMZlSBlAJEJLKOz2dFzsHB7/94MbO4WDg66c04DSZyjqFvRxZyFmw3fFDa30kRphCsGbN0/RsKWWZvlpmTNisCpn+8nTyaVw63b4ucZoc85qnSmVWpGCiNia1b4xR/sN/+n9u3Pwo7/SioDZWAEPkpMmi8xqf+q0pF5cybEqpPM/ruu52u/v7+4uLi2VZvvnmm/v7++PBcH9/P8a4srJSFMWpU6c+/vjjg929W7c/qerxpfMXfN18dvt2DEEjjUYjAOjN9S+/cHV1fePT25+9+ff/cO+LhyzQ7fWBlI+MpLUxMmFUDcGlotKMUkcRUUhNVWvCr19/5fL586Pd3V5m//SP/vjMybXM6roqJXgCRgGjtLU2xmiLDiPUrnEc9w8OHm9tNZHRWpt16MgNfe4mO/Jkuoozeevn7Cf459mdf+Z7puwDUyD4JGf6nFPNWsDZX4SISqk0jzqdau9g/5NPPtnbO+A0LlOZ2V34FfpyMBgURRFCWF5eRsQvvvji8PDw2rVr165dm+/1T548+eGHN61R9+7dXTmxlGXZ+fPn67p+8ODBgwf3Xn/9de+b/f39laXl2jXW2vX19asvvHTp4pUs7z16vH3zk9t7w2HW7c8tLAaOo7JSStksExECRBJSSAqIKN0RhZRp0+vk7NyZkxsvnD8H3n3zlWs//N3vcnQcmsR8HmNoyWNj1MpyBO+CMpm22eFotLWzWzvnfHQh6iNWfPrncwXiyCrOvjKrJ5670sdff+4rX/bnEXsx3buzI+qfKwfPbizQBL6pNBmPopQ52Dv81VtvX7rywtUrL+Q5KQUhoIhQGhr5vNkt6VhbW2PmR48eiwiQ/tnPfvbOO+8sLy/vFHld10K4srKyvr6xenJzb+9gd29Ha725vnFY7Gu9/vOf/3xpfi63mbXWSZxfXN48ffb0ufPGmK3t7dFoNKxqMnZtTc11O2RzjMwwYfRSoBmY0tTPBM8hJHFNtTS34Ef7S73ON1659vDTD/6rN3505fzZen9PxLM4YQ+TXQyCeV7UwdfOpyk1jx4/frT1WGvNSiFNmOW/2sGclYypC3nk1h9RGM/9+Kw0fLVAHPkvRGz7wGYikeO26blKYtb9nF6q1jrUwRhTluUHH3zw1ltvnTq52e12tdIoxD7RpXFEDy275dHjwYMH8/PzFy9ejDHe+uzuO++8s7Gx8fLLL+8/2dZal1Wzv7fz0Ucfnzt/vr+w/K1vfcu58PHHH/d6neD9icWFBw8erK4sI+LC/NK5cxdWVtdDFEDudOeyvBtBdnf3ysatLC8tLy30rR0PB3XdKEUQGVGIkAAUqMnsEnFllWtVgfiqvH71avHf/MkPf/d79eG+IhBsWeJJQapsRmZjTOW9MLJgNRw/ePBw/2DUX1x8Mqo5jcedXe/nSsbsmn2ZXj2eITj+CF+ibJ5rcY78meY6Td2XdMw+n/3IVIAmv+ipieQQjNKKKMZobZZnRbl38O7b7373W9/t9Xr5wjwRcUhw62fmfh85NjY2RqPR7u6uiPzyl79smkZr+/bbbyvhfr9/+crV/sL8y9e+NhyNnuwd/uLNN/tzc+fPnx8Nh7/+9a+X5uc2NjYQODKsbpy8fOXqytr6qK5FsOh0ik4fAA4Oh2Xtd/cOlTLzcx2T5TGK9w2RCAKyEE7UBiEhmE7BwfeLPAx3z2ysXl7/4ekTy3u7292iiBwie+ZAqSVZUIhjZACw1lYx7uzsPHr0KE3zNJow4VqP3NnpDf0KyXiuwoCv1ATHPY/nisIRqZu+YVZbPCMT8vRL2zc8J/k9casRhcUYxT4ggALSWoMx9+/ff++99/rd7kK3b0hFUol5QSGFmcLK7PHo0aPV1dUsy8fj8c9+9rOiKE6cWC2KwhIeHBx8+OGHniOIOn3ubCezC+fPk1Lb29tPtrdfffXVTz/6cDg8jCF873e+v3HuwoWLl5ZXVg8OhzEKg65d8DHML64gSairre0nzvVXV5Y7Xb23W5JqUZ8IifUJkEiDdHqZK8vMQq6VJVruzrnxsJ/Zpqmdc9459kEJpH4mEXHOxSha23K0e/fuvb29AwCoqipGYRY9u0JfrSpkJpl9ZL2/TIy+WkkcWf7jdmRWRnFmWCPMJleO2ZHn+hZTEVFEIFJVtbGGmaP3/bkF7/0/vPkPK8vLmxsb/W5GRDQRN/gS7bi+vg4Au7u7b7755p07d/r9/vb29mg06mZ2bW1tc3NzYXnp/fdu7u3tPXz4eH3z5Jkz57a3tjY2Nm7duoWIvV7v2ssvf7G19S//1Z+undxQyljrgTQDNE2DijqdgggbpFEIVdWUZZ1pZa2V6CadpYwokKafEioQFHBlvdLrKolK2Fe11cqHxgfnXC0+AhlCgplYfXg4uHPnzo0bN8rRuKsMIrGveFZbHNcBXyEcz13dr5aM4/LxvBf5yNacLqqaHMdWWuCYTCfVkECRiCiSYgpEFFQmRG6aZq7ouhCqslldXWywufn+B19/7bXR4bCbWUzzwyIICuiEpQQAQGBAJGEAOjg4AFLz84vv/vb9ubkF75sLZ88xc2b1Bx98UBTd3/72t69+/Zveh4319caFd/7pV9df+7pzrtvt3r9/f2lpYVQ3P/jxG5dfuEomPzw8ZIQ8z0IIIqY/P+cb17gKFCwszAHH/YODLMt63U49roVl0icIiKgAUYBjXOp1q71Bv9Ptol/o5VKHw8EeGsscOFFUIQspAhBGVKpumt39vc8+++yzW7eZw9ryigNeOrHEgForKym1LNN7SogYOU52o4g81dQA0yVppwS06yqE7ZMUuLbbOHWVTQAXMGlnmx6MmOhBAADT8ER5FsqbjpSkT7GlzMbGE2zONA2qlNJapYpXDBInUf40gRFBbKdX1R5R9XpzVdUAS6c798u//+VLL1xdWvxOlllFetSMOr2uYxFgEiFIzWMEwJG5NzffBPnHX/36i0c7zsfxqHxw//Mss7m1r15/hYhu3Pzw4f07VdnMLy4V3d7v//iNjz69de/Bw6Lb684tjZqwevr8mYsvDOoafUCrO6rwvgGUvNAgjaLY62QKM++9d3UEqqMPo2ap3/WubhwTIJLSicRQxECA4DfXV7qWOyDlaEfHqrAwjo0mYEMB2LOLwpoUah1joEI3oVxbP/Fv/+LfLCwsRAAyujffF4RJz8yxesSX7ez/4nFMPcCMUD39riOmR0QmHMpPkd+zJzyeKGul81loOMxgg9uhj4lYSkAAGRhIM6bhcqARAQgTewRDVTbvvvObc2fOXrx03gfO804MIkQg0E7XYwTFqTjXNL6/vHIwHn92/36oqrXV1V5uMqMPDg4Gg4PRqHzpxZdXVtZE5IMPPx0NDt9+++2i09vY2KhDPBiNin4/K3rDqt47OLTWWpsSEICCWqNWFEFIQCQqhWAtEgXvY0Igt9uYNEKmlVXaIOgonYx6OfVy6RHoGCkKc5BAjiVRNAmnbWOINCpzYn1t6cTqd8tyZWVtfn6eBbUhF50AHHU5Z2/xEd9z+qI8G6bOPnmex3As5TU52bGTT9lGn0mspf9L3brTo3UhiVKOb5qBxWNltukrnCbNPFsRxEmrJxlVNvU//uqXl65c3Dy1gQh5bsdVicbOehdJXwJC0zSFjzs7O8654FxkOByMVleWsqLY3Nzc3z94tL31wUcfnzx5stfrLS4vDUbVsKzu3bs3rOqsKC6eOXPq1Cmr9Hg4qBMtGBEiGqWsNan+IJEBxWidWc1sXIWuid5HYdSkDCpFaCVqYSPSzVQvV71cdwrKlaIIEEiYFZkgmHLNIEREqA2hFsF8bm7xxDo7RzYHxOgaETFkYCoWz9UQR16cPh6PUZ8rELP/dURVHDltK/4TaaRJm1B626yTO03MP62SwNOSW3pnKmTArFs6/dKZK0GcYltQGLIsa5pma+vg3Xffffmlq2fPnkmEaWp6qZMfLQAikhX5/fv3nXPf+NZ3Hty9U48Gw/2dTpF579977721tfW1tbUrV65+/NGnFfjAkUExh16vZ7sFoFKAWZatrq42rpLUweZ9jFEhJbHodDqawBhDShtSpLRByrRqqpHROtNKRDh4cQ0SKoVLC0udDHKrlAJQJKAQM5CoVKZk+vMJSYMyQFSPSnAehEIIEEpsmwPTzE14CsOZvfXPXfXpvZ7djrMCMbvMcFQTHJWJyameeb+ITPEWMqOTptt9Viu07gJOTyVTtTH7nvS2yY8SZp7O15gIDSYeMWbW1n706Sfvvf/+xuYGxqBMNpWJ6Q0QAUDo9Xq/fv/DTm/uD//wDz/+4Mb9u3ce3LszroZZli2vrASOv3rrn06ub26ePD23sHj79u3+/BIZbUrnPedFsba+ura2MtfvMhfMnKDtTVNzCBxCE4Kry9xmRVEIhxiMUVprnXc6oa6sMpowBkex0cgLWbHQtSfmtFWQKVYokRmRSBlEy5HbOZiISApS7Mac9/vAAkGIKEbRWpMxIBwhwqwROb7Mx5XBrKo4/rbjL86ebSaCeIqigGe9mdnLmD0bzrQMHfmKWbc0fV3qXZ6qnOmTtLqBp4P3sGUHEQZIHFPQ7XafbO/+5v0bL157ee3EysJyp/HuqBFJTFG139/f73a7p8+cy409c+bUgzsnf/Wrf/R1MxqNtNbf+fb3Hjx4+Hh7a+/gcP3kJinjY7j/8NH27t658xcvX7589uzZwWBgSCljqegws3dNWZZlWbq68cE1LMIh1DrNIeh0OpSZLCsUAkK0wMbq+Y49udxfmu9oYE1MiCKJrloJZUhAEKEdd0htsC0IgpDlUJaAqJSq6yoNPKiqChVwcjmfaxSOLP/spp8NEY9oiOMCMbvqk3POJqePeiqz0jnrHxzJdn+FZKRyyRGxaK8HQWIEkJmKdRpkA569tjoKR+Fbt++8+5v3Xv/+9wvvGYFSo9uzwnFwcMAMWaeoqqrT70X2Fy9dcd7f+vTj/f39wWCws7/HCKdOn1JkvA+Pt56cWF1ZX18na6zVVTVumloTJvi0IjTaWKU0Ka0wWON9FmMM3pV1QwQhy4QDu8wYEyUqkF6RL3ftyly+1LXWiDR1ulmAFIEEdaBEBY0IE7Fo40QRYb+7W9d1t+gZY9JEN6UsETnfwFRbHHEhj6/x8cX4iuP4+5+VvNksO87Yi2c8xHQZU4zFEcXzFVclEydj+k6adBwlWwDQQv1acjoRACaiPM+Hw6HWen9//8MPP/7GN741HlVFP5/5UU9/RdXUeZ7bLBtXtSbcebK32O98+9vfvXr16l//X3/lY3DO9Xq9D29+vLFxcm5h/pVXXvni0cO7d++Wrul0et57ibHT6WAbgYlEBoAiM5mZFxHfuHE5HAy8C0EjRCSPxD4URZckFLme73ROrvSWewZCBeUA2INGAAPKIJAABiZGjCFS68sHEZHQ2lkESjibEMJoNAohGJNVVdX4GgDwxes/+rLVnd2mU+cfZzqGeQYZNa10HxOvVNyKUxPAEgBgwmQyq4bSZ59WSnE22ZA4f33bPZwwMkopQyqdrWmaxKybUhdFUSSWvsTb0f4WSe2vs6EKJQlyLoiIMZlSCoRWV1d/9KMfvfF7P8otdgqTWw0SFAMoQjIB7e17j7f2h/NLK8x8uH9gjdIQFXK3W3z22Wd/8zd/c+uTTzc2NpeXlweD4Wg06hQ90urJ3r4XttYurZz4oz/6r3/84x+XgzJ6z2nIF0r0gZmRWiKapmnG46Grm+lW0dqun1g8tb642s/62mdQm1BxGFsNQAQ6AzJRZQ60B43CJjWHMidvPQ1PY2ZkiTGGyQQ97z2H4EKomhqmpbL/Hx+t4OZ5HmMUAedcXTkievx468GDB5cunPnqz/OzadnhcLy2tvbaa6855/Z3D8qy7PX6586dcz4GYTMc1VU5GAya4O/cuXP79u2NlY3pZyeVOQZobR9RG3MddbZEABiBCQUhkDAIgghMEKBpKBhDmnY9pRyRyWg9dlUdQpiKRQgheu9CGI/HgvD/AbDWXMUOTHe2AAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<PIL.Image.Image image mode=RGB size=178x218 at 0x7F6F800CF9E8>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "E0OfXGP6eVM7"
      },
      "source": [
        "**2. Most attractive**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fiQIFulqeX5w",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "outputId": "cf6743fa-f033-4bc0-ab5f-6bc43465d152"
      },
      "source": [
        "test_df_without_data.sort_values(by='predicted_value', ascending=False).head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>image_link</th>\n",
              "      <th>5_o_Clock_Shadow</th>\n",
              "      <th>Arched_Eyebrows</th>\n",
              "      <th>Attractive</th>\n",
              "      <th>Bags_Under_Eyes</th>\n",
              "      <th>Bald</th>\n",
              "      <th>Bangs</th>\n",
              "      <th>Big_Lips</th>\n",
              "      <th>Big_Nose</th>\n",
              "      <th>Black_Hair</th>\n",
              "      <th>Blond_Hair</th>\n",
              "      <th>Blurry</th>\n",
              "      <th>Brown_Hair</th>\n",
              "      <th>Bushy_Eyebrows</th>\n",
              "      <th>Chubby</th>\n",
              "      <th>Double_Chin</th>\n",
              "      <th>Eyeglasses</th>\n",
              "      <th>Goatee</th>\n",
              "      <th>Gray_Hair</th>\n",
              "      <th>Heavy_Makeup</th>\n",
              "      <th>High_Cheekbones</th>\n",
              "      <th>Male</th>\n",
              "      <th>Mouth_Slightly_Open</th>\n",
              "      <th>Mustache</th>\n",
              "      <th>Narrow_Eyes</th>\n",
              "      <th>No_Beard</th>\n",
              "      <th>Oval_Face</th>\n",
              "      <th>Pale_Skin</th>\n",
              "      <th>Pointy_Nose</th>\n",
              "      <th>Receding_Hairline</th>\n",
              "      <th>Rosy_Cheeks</th>\n",
              "      <th>Sideburns</th>\n",
              "      <th>Smiling</th>\n",
              "      <th>Straight_Hair</th>\n",
              "      <th>Wavy_Hair</th>\n",
              "      <th>Wearing_Earrings</th>\n",
              "      <th>Wearing_Hat</th>\n",
              "      <th>Wearing_Lipstick</th>\n",
              "      <th>Wearing_Necklace</th>\n",
              "      <th>Wearing_Necktie</th>\n",
              "      <th>Young</th>\n",
              "      <th>status</th>\n",
              "      <th>predicted_value</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>17858</th>\n",
              "      <td>190946.jpg</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True Positive</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15910</th>\n",
              "      <td>192599.jpg</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True Positive</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2612</th>\n",
              "      <td>187635.jpg</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True Positive</td>\n",
              "      <td>0.999999</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18267</th>\n",
              "      <td>195818.jpg</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True Positive</td>\n",
              "      <td>0.999999</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2934</th>\n",
              "      <td>198191.jpg</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>True</td>\n",
              "      <td>True Positive</td>\n",
              "      <td>0.999999</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "       image_link  5_o_Clock_Shadow  ...         status  predicted_value\n",
              "17858  190946.jpg             False  ...  True Positive         1.000000\n",
              "15910  192599.jpg             False  ...  True Positive         1.000000\n",
              "2612   187635.jpg             False  ...  True Positive         0.999999\n",
              "18267  195818.jpg             False  ...  True Positive         0.999999\n",
              "2934   198191.jpg             False  ...  True Positive         0.999999\n",
              "\n",
              "[5 rows x 43 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YR1x63JXc5cj",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 235
        },
        "outputId": "c2d8e128-4eba-4250-e96e-0e46ce796f6c"
      },
      "source": [
        "#Most attractive picture\r\n",
        "path0 = r'lesimages/img_align_celeba/'+'190946.jpg'\r\n",
        "img0 = cv2.imread(path0) \r\n",
        "cv2_imshow(img0)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<PIL.Image.Image image mode=RGB size=178x218 at 0x7F6F800CF0F0>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dDNmBf4netif"
      },
      "source": [
        "## Attributes amongst populations"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "12ATtfZsMyqE"
      },
      "source": [
        "**1. Proportions of False Positives in the classified images**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VOW2CJvI67OG",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 431
        },
        "outputId": "59c503ec-da78-4b94-fef7-f836876c6a24"
      },
      "source": [
        "### Proportion of False Positive regarding attributes: \r\n",
        "\r\n",
        "fig = plt.figure(figsize=(10,5)) \r\n",
        "\r\n",
        "interesting_features = [\"Bald\",\"Young\",\"Heavy_Makeup\",\"Male\",\"Pale_Skin\",\"Smiling\"]\r\n",
        "\r\n",
        "# Plot accidents depending on type\r\n",
        "i=1\r\n",
        "\r\n",
        "for attr in trainset.attr2idx.keys():\r\n",
        "  if attr!=\"Attractive\" and len(test_df_without_data[(test_df_without_data[\"status\"].isin([\"True Positive\",\"False Positive\"])) & (test_df_without_data[attr]==True)]):\r\n",
        "    plt.bar(i,len(test_df_without_data[(test_df_without_data[\"status\"]==\"False Positive\") & (test_df_without_data[attr]==True)])/len(test_df_without_data[(test_df_without_data[\"status\"].isin([\"True Positive\",\"False Positive\"])) & (test_df_without_data[attr]==True)])*100)\r\n",
        "    i+=1\r\n",
        "\r\n",
        "plt.xticks([r for r in range(1,len(trainset.attr2idx.keys()))], [attr for attr in trainset.attr2idx.keys() if attr!=\"Attractive\" and len(test_df_without_data[(test_df_without_data[\"status\"].isin([\"True Positive\",\"False Positive\"])) & (test_df_without_data[attr]==True)])],rotation=90)\r\n",
        "\r\n",
        "plt.title(\"Percentage of False Positive regarding attributes\")\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "X3KvU1nzM9FF"
      },
      "source": [
        "**2. Proportions of attributes in the train and predicted attractive populations**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 337
        },
        "id": "NOYAmURnYANw",
        "outputId": "48efb507-d679-4b1f-973f-8d307ee690f6"
      },
      "source": [
        "train_df = pd.DataFrame(trainset_allattr)\r\n",
        "train_df.columns = [\"image_link\",\"attrs\"]\r\n",
        "train_df[[key for key in trainset.attr2idx.keys()]] = train_df['attrs'].apply(pd.Series)\r\n",
        "train_df.pop(\"attrs\")\r\n",
        "\r\n",
        "fig = plt.figure(figsize=(10,5)) \r\n",
        "\r\n",
        "interesting_features = [\"Bald\",\"Young\",\"Heavy_Makeup\",\"Male\",\"Pale_Skin\",\"Smiling\"]\r\n",
        "i=1\r\n",
        "barWidth = 0.25\r\n",
        "\r\n",
        "for attr in interesting_features:\r\n",
        "  if attr!=\"Attractive\":\r\n",
        "    plt.bar(i,len(train_df[(train_df[\"Attractive\"]==True) & (train_df[attr]==True)])/len(train_df[train_df[\"Attractive\"]==True])*100,color='#B22222',width=barWidth,label=\"Train Attractive Population\" if i==1 else \"\")\r\n",
        "    plt.bar(i+barWidth,len(attractive_predicted_df[attractive_predicted_df[attr]==True])/len(attractive_predicted_df)*100,color='#F08080',width=barWidth,label=\"Test Predicted Attractive Population\" if i==1 else \"\")\r\n",
        "    plt.bar(i+barWidth*2,len(train_df[train_df[attr]==True])/len(train_df)*100,color='#8B0000',width=barWidth,label=\"Train Population\" if i==1 else \"\")\r\n",
        "    i+=1\r\n",
        "\r\n",
        "plt.xticks([r + barWidth for r in range(1,len(interesting_features)+1)], [attr for attr in interesting_features])\r\n",
        "\r\n",
        "plt.title(\"Proportion of selected attributes inside different population\")\r\n",
        "plt.legend(prop={'size': 12})\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlAAAAFACAYAAABz3EJDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXgV5f3//+ebgCEQICFJWUPC4gL0W0BwB6UKVhTcZV8VK21BrVKrxJ/ARxDX1lZsXbCCgIjSVoliaV1Q3FBKY1FwAUsIBDCAYGJAWe7fHzM5nBxykjOQkACvx3XlypntnvfMmTPnPfd9zxxzziEiIiIisatV3QGIiIiIHG2UQImIiIgEpARKREREJCAlUCIiIiIBKYESERERCUgJlIiIiEhASqDkqGFmQ8zsn9Ww3nPM7EszKzKzyyupzEwzc2ZWuzLKO4w4JpnZnCO8zh5m9nnY8Doz63UkY6hIZIyHUc6nZtYzyrSeZrbhcNfhlzXSzN4JGy4yszb+6wQzyzaznWb2gj9uipltNbPNlbH+Y8Hhvh9m9piZ/X+VGZPUbEqgjnP+l9cu/4S7xcxmmlliDYjroATDOTfXOXdhNYTzf8B051yic+7Falh/VNWZfJjZEjMbHTHOmVm78pZzzi11zp1cSTHMNLMplVFWuMqK0TnX0Tm3pBJCCrreROfcV/7g1UATIMU5d42ZtQJuBTo455oe6dhiOUZqusiEFcA5N8Y5d3d1xSRHnhIoAejnnEsETgW6AXdGznAka0qqu1amDBnAp9UdxLGgBr63x4MM4Avn3F5/uBWwzTn3ddCCzKPvDRGUQEkY59xG4FXgxxC6UvyVmX0JfOmPu97M1pjZdjNbaGbNS5b357/RzL7ymwceKDnZmlktM7vTzHLN7Gsze8bMGvnTSmqbrjOz9cAbwNt+sTv82rGzymimONvMPvKbJj4ys7PDpi0xs7vN7F0zKzSzf5pZarRtj7ZdZrYWaANk+3HEl7Hsb81so7+ez83sgrBtvt3M1prZNjN73swaR1l/IzN7ysw2+WVNMbO4iPhW++tYZWanmtlsvC/Dkthu8+c908zeM7MdZvZxeBOSmbU2s7f8cv4FlLdPks3sZTMrMLNv/Nct/WlTgR7AdH/d082s5D372B83wPxmEX8fbQaetrKbSk7zt+sbM3vazOr66znoSr+kBsPMfg4MAW7z15ftT29uZn/14/6fmd0YtuzpZrbczL41r8b1d1G2vVSM5tX0jTez//rH2/ywGFP9fbPDP36Whh33oRpC85rSZvrbuAo4LWKdUeMuI74U/zj91sw+BNpG2UeTgbuAAf4+ugH4F9DcH57pz1/eMbPEzKaa2btAMdDGzE4xs3/52/u5mfUPm3+mmT1qZq/4x9kyM2vrTzvoGClj20aa97md7u/rz8z/TIXtp4X+uteY2fVh0yaZ2QL//Sk0sxVm1ilyv0TEWmYNph347JZ85q7wx7cHHgPO8rdhR1llWcXnyjHmdQ3Y4e8vKysOqcGcc/o7jv+AdUAv/3U6Xk3L3f6wwzvZNgYSgPOBrXg1VfHAI8DbYWU54E1//lbAF8Bof9q1wBq8ZCQR+Bsw25+W6S/7DFDfX1fJuNph5Y8E3vFfNwa+AYYBtYFB/nCKP30JsBY4yS9vCXBvlH1Q0XaF9lEZy54M5AHNw7alrf/6JuADoKVf7uPAvIhtru0P/92fXh/4EfAhcIM/7RpgI94XrgHtgIyyYgNaANuAi/EukHr7w2n+9PeB3/nxnAsUAnOibFsKcBVQD2gAvAC8GDZ9Scn7G3EMtAsb7gnsBe7z15ngj9sQsX8/wTv+GgPvAlMi3/Oy1gHMLJnXH64F/BsvaTgB73j7CvhZ2PYP818nAmdG2fayYvwQaO7HuBoY40+bhveFWsf/6wFYGZ+ve4Gl/vLp/jZviCXuMuJ7DnjeP15+7B8f70TZR5PC3+Mytq2iY2YJsB7oiPdZa4R3zI/yh7vgfX46hL0n24DT/elzgeeiHSNlbNtIvGPm1/7+HADsBBr7098G/gTUBToDBcD5Ydu6B6/Zsg4wHvgfUCfK8TmTA8da5H65xn+/a/kxfAc0K+e4DC8rlnPly0AS3rmyALjoSJzz9Vd5f9UegP6q+QDwTvBFwA4g1z8xJfjTXMmJyR9+Crg/bDjRP1llhs1/Udj0XwKv+69fB34ZNu1kf9naHEgm2oRNLxkXLYEaBnwYsS3vAyP910uAOyNi+UeUfVDRdq0jegLVDvga6FVykg6bthq4IGy4WRnbXBuvf8r3Jfvdn3cQ8Kb/ejFwUznvX3gC9Vv8xDRs3GJghH+i3gvUD5v2LFESqDLW1Rn4Jmx4CbElUD8AdSPGRSYnY8KGLwbWRr7nZa2DgxOoM4D1EfPfATztv34bmAykVrCtZcU4NGz4fuAx//X/AS9RRlJA6QTqK0p/Pn7OgQSq3Lgjxsf5x9EpYePu4dATqKjHTNj7/H9h0wYASyPmfxyYGPaezIh4Pz+LdoyUsX0jgXz8JNQf9yHeZz4d2Ac0CJs2DZgZtq0fhE2rBWwCekQ5PkPHT+R+KSOuHOCyco7L8LJiOVd2D5v+PHB7LJ9D/dWcPzXhCcDlzrkk51yGc+6XzrldYdPywl43x0uyAHDOFeFdabaIMn+uv8xBy/qvS5KHspatSGR5JWWGxxJ+h1Ex3kmswrKibFeZnHNrgJvxTtxfm9lzYVX1GcDf/Sr6HXgJ1T5Kb3PJfHWATWHzPo5XEwXel8baimIJK+uaknL8srrjJW/N8RKg78Lmj9yHIWZWz8weN6/Z9Vu85CPJwpoWY1TgnNtdwTzRjpugMvCap8K3fwIH9vl1eLWSn5nX7Ns3QNnRjqcH8GpX/2le8/XtUZZvzsHbGWvc4dLwPjvRygqqvGOmRF7E/GdEzD8ECO+QHutnL5qNzs8sfCXHRHNgu3OuMGJamecg59x+YAOHcDyZ2XAzywnbxh9TTpN3hFjOKYe7j6SaKYGSioSfxPLxTp4AmFl9vGaejWHzpIe9buUvc9CyHKgN2RJlXeGvyxJZXkmZG8uYtyKxbFdUzrlnnXPd/TIcXnMVeCfyPn5yWvJX13l9zcLl4dVApYbN19A51zFselvKFrmf8vBqE8LXWd85dy/elXiyv30lWpWzabfi1RSe4ZxriNfkB14zYlnrjiaW+aIdN9/hNSF6KzaLvGusrO3/X8T2N3DOXQzgnPvSOTcILzm9D1gQsT8Cc84VOududc61AS4FbgnvsxNmEwdvZ0xxRyjA++xEKyuo8o6ZEi5i/rci5k90zv3iMGKI1CKiT1DJMZEPNDazBhHTyjwHmdcXrSUHjqdiwo4nSid9IWaWATwJjMXrFpCE1+Qa67F/WOcUOToogZIg5gGjzKyzeZ2p7wGWOefWhc3zG/M6H6fj9QGaH7bsr83rxJzoLzvfHbgzKFIBsB+vL0hZFgEnmdlgM6vtd0btgNevoCq2q0xmdrKZne8vtxvY5ccNXr+Yqf7JGDNLM7PLIstwzm0C/gk8ZGYNzet83tbMzvNnmQGMN7Ou5mlXUiZeAhq+j+YA/czsZ2YWZ2Z1zesQ3dI5lwssByab2Qlm1h3oV87mNfC3Z4d5nd8nRkyPXHe0cbH4lZm19NeTxYHj5mOgo//e1MWr6StvfR8CheZ1Wk/w98GPzew0ADMbamZpfs3EDn+Z/RwGM+vrvyeG11dnX5Qynwfu8D8fLYFxscYdzjm3D68P4SS/lrADXhPtoYp6zESZ/2W8z94wM6vj/51mXufqWMRyjPwIuNEv+xqgPbDIOZcHvAdM8+P8CV6tYvizzLqa2ZXm3fF5M97FyQf+tBxgsL+dFwHnUbb6eElSAYCZjcK/uSZsG1qa2QlRlj/kc4ocPZRAScycc68B/x/wV7yr6bbAwIjZXsLrDJsDvILXFwDgL8BsvGag/+ElG+OIwjlXDEwF3vWr0M+MmL4N6ItXS7INuA3o65zbWkXbFU08XufgrXhV8j/C67sC8AdgIV7TTiHeSfyMKOUMx+s8vAqvM/wC/CYU59wLePviWbxO3y/idUQGr//Hnf4+Gu9/wVyG1/xTgFdb8BsOfNYH+zFsx0uIniln2x7G6/S91Y/9HxHT/wBcbd5dZX/0x00CZvnx9Cd2z+IlkV/hNVdO8bf9C7w+Rq/h3Qn6TsRyTwEd/PW96CcXffH6a/3Pj30GXsdngIuAT82syI9/YEST9aE40Y+vCK8f3p+cc2+WMd9kvGad//nbOrtkQgxxRxqL1+SzGa/vzdOHGnwMx0zk/IXAhXifkXw/hpKbBGIxiYqPkWV4+3Ur3rF/tf+ZB69/YKa/7r/j9b16LWzZl/D6aZXcZHKlc26PP+0mvIuGkmbHMp/r5pxbBTyE935uAf4f3s0NJd7Au+Fms5kddM45zHOKHCVK7hQROWxm5oAT/X5BIiKBmdlIvJsTuh/CspPwOokPrey4RCKpBkpEREQkICVQIiIiIgGpCU9EREQkINVAiYiIiASkBEpEREQkoCP6y+ipqakuMzPzSK5SRERE5JD8+9//3uqcSytr2hFNoDIzM1m+fPmRXKWIiIjIITGzqD+TpCY8ERERkYCUQImIiIgEpARKREREJCAlUCIiIiIBHdFO5CIiNdGePXvYsGEDu3fvru5QROQIiouLIykpidTUVGrVClanpARKRI57GzZsoEGDBmRmZmJm1R2OiBwBzjn27NnDli1b2LBhA61atQq0vJrwROS4t3v3blJSUpQ8iRxHzIwTTjiBFi1a8N133wVeXgmUiAgoeRI5TgVtugstV8lxiIiIiBzzlECJiBwH+vTpw6xZs6o7jCozZswY7r777uoOo1KMHDmSO++885CXT0xM5KuvvqrEiKQs6kQuIhIhu23bKi2/39q1Mc2XmJgYel1cXEx8fDxxcXEAPP744wwZMiTmdb766qvBgixDz549+fjjj9m8eTPx8fGh8ZmZmcyYMYNevXoBsG7dOlq3bs2ePXuoXbvyv2ZmzpzJjBkzeOedd0LjHnvssUpfD3jbtmXLFuLi4qhfvz59+vRh+vTppd6b6tSzZ0+GDh3K6NGjQ+OKioqqMaLjhxIoCaSqvlhi/UIROZ6EfxFGJinh9u7dWyWJSrh169axdOlSGjVqxMKFC7nmmmsOq7wjEXNlyc7OplevXmzcuJGf/exnTJkyhXvvvbe6w5JqpiY8EZGjzJIlS2jZsiX33XcfTZs2ZdSoUXzzzTf07duXtLQ0kpOT6du3Lxs2bAgt07NnT2bMmAF4NTjdu3dn/PjxJCcn07p16wprqJ555hnOPPNMRo4cWaopcNiwYaxfv55+/fqRmJjI/fffz7nnngtAUlISiYmJvP/++8ycOZNzzjmHX//616SkpDBp0iTWrl3L+eefT0pKCqmpqQwZMoQdO3aEys7Ly+PKK68kLS2NlJQUxo4dy+rVqxkzZgzvv/8+iYmJJCUlAaWbvdq3b8/LL78cKmfv3r2kpaWxYsUKAD744APOPvtskpKS6NSpE0uWLIlpv7do0YI+ffrwySefALBw4UI6duxIUlISPXv2ZPXq1aF5MzMzmTZtGh06dCA5OZlRo0aFnjNWsv/DmRlr1qw5aJ3lva9ZWVksXbqUsWPHkpiYyNixYw8qa+fOnQwfPpy0tDQyMjKYMmUK+/fvLxVHkONADlACJSJyFNq8eTPbt28nNzeXJ554gv379zNq1Chyc3NZv349CQkJoS/UsixbtoyTTz6ZrVu3ctttt3HdddfhnIs6/zPPPMOQIUMYMmQIixcvZsuWLQDMnj2bVq1akZ2dTVFREbfddhtvv/02ADt27KCoqIizzjortM42bdqwZcsWsrKycM5xxx13kJ+fz+rVq8nLy2PSpEkA7Nu3j759+5KRkcG6devYuHEjAwcOpH379jz22GOcddZZFBUVlUq4SgwaNIh58+aFhhcvXkxqaiqnnnoqGzdu5JJLLuHOO+9k+/btPPjgg1x11VUUFBRUuM/z8vJYtGgRXbp04YsvvmDQoEE8/PDDFBQUcPHFF9OvXz9++OGH0Pxz585l8eLFrF27li+++IIpU6ZUuI5I5b2vU6dOpUePHkyfPp2ioiKmT59+0PLjxo1j586dfPXVV7z11ls888wzPP3006HpQY8DOUAJlIjIUahWrVpMnjyZ+Ph4EhISSElJ4aqrrqJevXo0aNCArKws3nrrrajLZ2RkcP311xMXF8eIESPYtGlTKCmK9M4775Cbm0v//v3p2rUrbdu25dlnnw0cc/PmzRk3bhy1a9cmISGBdu3a0bt3b+Lj40lLS+OWW24Jxfzhhx+Sn5/PAw88QP369albt+5BtTbRDB48mIULF1JcXAzAs88+y6BBgwCYM2cOF198MRdffDG1atWid+/edOvWjUWLFkUt7/LLLycpKYnu3btz3nnnMWHCBObPn88ll1xC7969qVOnDuPHj2fXrl289957oeXGjh1Leno6jRs3Jisrq1RSF6ug72u4ffv28dxzzzFt2rTQg2JvvfVWZs+eHZonyHEgpSmBEhE5CqWlpVG3bt3QcHFxMTfccAMZGRk0bNiQc889lx07drBv374yl2/atGnodb169YDonY9nzZrFhRdeSGpqKuAlKIdyR196enqp4S1btjBw4EBatGhBw4YNGTp0KFu3bgW82p6MjIxD6ifVrl072rdvT3Z2NsXFxSxcuJDBgwcDkJubywsvvEBSUlLo75133mHTpk1Ry3vxxRfZsWMHubm5/OlPfyIhIYH8/HwyMjJC89SqVYv09HQ2btxY5vZmZGSQn58feFuCvq/htm7dyp49e0rFmZGRUSrGIMeBlHZ09OCTY97OyZOrrOxGEydWWdki1SXywZ8PPfQQn3/+OcuWLaNp06bk5OTQpUuXw26O2bVrF88//zz79u0Lfdl+//337Nixg48//phOnTodFEu0h5JGjp8wYQJmxsqVK2ncuDEvvvhiqHkqPT2d9evXl9nZPJaHnpY04+3fv58OHTrQrl27ULnDhg3jySefjG0HRNG8eXNWrlwZGnbOkZeXR4sWLULj8vLyQq/Xr19P8+bNAahfv36odgy85thoKnpfy9sXqamp1KlTh9zcXDp06BCKIzxGOXSqgRIROQYUFhaSkJBAUlIS27dvZ3IlXZS8+OKLxMXFsWrVKnJycsjJyWH16tX06NGDZ555BoAmTZqUeu5QWloatWrVqvBZRIWFhSQmJtKoUSM2btzIAw88EJp2+umn06xZM26//Xa+++47du/ezbvvvhta34YNG0r1N4o0cOBA/vnPf/LnP/85VPsEMHToULKzs1m8eDH79u1j9+7dLFmypFSH+1j079+fV155hddff509e/bw0EMPER8fz9lnnx2a59FHH2XDhg1s376dqVOnMmDAAAA6derEp59+Sk5ODrt37w71+4q2j8p7XyP3fbi4uDj69+9PVlYWhYWF5Obm8rvf/Y6hQ4cG2lYpmxIoEZFjwM0338yuXbtITU3lzDPP5KKLLqqUcmfNmsWoUaNo1aoVTZs2Df2NHTuWuXPnsnfvXu644w6mTJlCUlISDz74IPXq1SMrK4tzzjmHpKQkPvjggzLLnjhxIitWrKBRo0ZccsklXHnllaFpcXFxZGdns2bNGlq1akXLli2ZP38+AOeffz4dO3akadOmoWbFSM2aNeOss87ivffeCyUu4NVAvfTSS9xzzz2kpaWRnp7OAw88ELozLVYnn3wyc+bMYdy4caSmppKdnU12djYnnHBCaJ7Bgwdz4YUX0qZNG9q2bRu6S/Ckk07irrvuolevXpx44onl9u2q6H296aabWLBgAcnJydx4440HLf/II49Qv3592rRpQ/fu3Rk8eDDXXnttoG2VstmR7G3frVs3t3z58iO2Pql8VfUcqHOHD6+SckFNeFKx1atX0759++oOQ44h5T23S2qeaOcAM/u3c65bWcuoBkpEREQkICVQIiIiIgHpLjwREZFKtm7duuoOQaqYaqBEREREAlICJSIiIhKQEigRERGRgJRAiYiIiASkBEpEREQkICVQIiJyRGRmZvLaa68BcM899zB69OgqX+eSJUto2bJlla+nKvXp0+eQfry5JurZsyczZsw4pGXXr19PYmJiTD+kfCToMQYiIhGq8setIfan4ycmJoZeFxcXEx8fT1xcHACPP/44Q4YMCbTenj17MnTo0KiJy7p162jdujX169cHvB+jHTNmDLfffnug9cRiwoQJMc03cuRIWrZsyZQpUyo9hhLOOdq2bUvdunVZtWpVqWlmxpdffhn6MeIlS5YwdOjQwL+dF6tJkyaxZs0a5syZExr36quvVsm6zIx69ephZjRq1IgBAwbwwAMPhI6x6hb5NPdWrVpRVFRUzVEdoARKRKSGCv+yOJI/DbJjxw5q167N+++/zwUXXEDnzp0P+g22vXv3Urv2sfEV8vbbb/P111+zd+9ePvroI0477bTDKu9o2jcff/wx7dq147PPPqNnz56cdNJJjBkzprrDOiqoCU9E5Cizf/9+7r33Xtq2bUtKSgr9+/dn+/btAOzevZuhQ4eSkpJCUlISp512Glu2bCErK4ulS5cyduxYEhMTGTt2bIXrOeuss+jYsSOffPJJqCnsvvvuo2nTpowaNarcOABmz55NRkYGKSkpTJ06tVTZkyZNYujQoaHhd955h7PPPpukpCTS09OZOXMmTzzxBHPnzuX+++8nMTGRfv36AZCfn89VV11FWloarVu35o9//GOonF27djFy5EiSk5Pp0KEDH330UYXbOWvWLC677DIuvvjiUk1l5557LgCdOnUiMTGRWbNm0adPH/Lz80lMTCQxMZH8/HwmTZrE1VdfzdChQ2nYsCEzZ87kww8/5KyzziIpKYlmzZoxduxYfvjhh1DZn376Kb1796Zx48Y0adKEe+65h3/84x/cc889zJ8/n8TERDp16gQcaPb6/vvvSUpK4pNPPgmVU1BQQEJCAl9//TUAL7/8Mp07dyYpKYmzzz6b//73vxVuP8App5xCjx49QmU/+eSTtGvXjsaNG3PppZeSn58fmtfM+OMf/0ibNm1ITU3lN7/5TejHmCPf13Xr1mFm7N2796B1rl27lvPPP5+UlBRSU1MZMmQIO3bsAGDYsGGsX7+efv36kZiYyP33339QWfn5+Vx66aU0btyYdu3a8eSTT4bKnjRpEv3792f48OE0aNCAjh07Utm/xasESkTkKPPII4/w4osv8tZbb5Gfn09ycjK/+tWvAC8Z2LlzJ3l5eWzbto3HHnuMhIQEpk6dSo8ePZg+fTpFRUVMnz693HU453j33Xf59NNP6dKlCwCbN29m+/bt5Obm8sQTT5Qbx6pVq/jFL37B7Nmzyc/PZ9u2bVGbvXJzc+nTpw/jxo2joKCAnJwcOnfuzM9//nOGDBnCbbfdRlFREdnZ2ezfv59+/frRqVMnNm7cyOuvv87DDz/M4sWLAZg8eTJr165l7dq1LF68uMK+Q8XFxSxYsIAhQ4YwZMgQnnvuuVCi8/bbbwNeLU1RUREjRozg1VdfpXnz5hQVFVFUVETz5s0BeOmll7j66qvZsWMHQ4YMIS4ujt///vds3bqV999/n9dff50//elPABQWFtKrVy8uuugi8vPzWbNmDRdccAEXXXQREyZMYMCAARQVFfHxxx+XijU+Pp4rr7ySefPmhcY9//zznHfeefzoRz/iP//5D9deey2PP/4427Zt44YbbuDSSy/l+++/L3cflLxfS5cupUuXLrzxxhvccccdPP/882zatImMjAwGDhxYav6///3vLF++nBUrVvDSSy/xl7/8pcJ1RHLOcccdd5Cfn8/q1avJy8tj0qRJgJd8t2rViuzsbIqKirjtttsOWn7gwIG0bNmS/Px8FixYwIQJE3jjjTdC0xcuXMjAgQPZsWMHl156aUwXDUEogRIROco89thjTJ06lZYtWxIfH8+kSZNYsGABe/fupU6dOmzbto01a9YQFxdH165dadiwYaDyU1NTady4MaNHj+bee+/lggsuAKBWrVpMnjyZ+Ph4EhISyo1jwYIF9O3bl3PPPZf4+HjuvvtuatUq+yvn2WefpVevXgwaNIg6deqQkpJC586dy5z3o48+oqCggLvuuosTTjiBNm3acP311/Pcc88BXkKRlZVF48aNSU9P58Ybbyx3W//2t78RHx/PhRdeyCWXXMKePXt45ZVXAu0v8GrrLr/8cmrVqkVCQgJdu3blzDPPpHbt2mRmZnLDDTfw1ltvAV4tUdOmTbn11lupW7cuDRo04IwzzohpPYMHDw5tK3j7bvDgwQA88cQT3HDDDZxxxhnExcUxYsQI4uPj+eCDD6KWd+qpp5KcnEy/fv0YPXo0o0aNYu7cuVx77bWceuqpxMfHM23aNN5///1SP0/z29/+lsaNG9OqVStuvvnmUkldrNq1a0fv3r2Jj48nLS2NW265JbSPKpKXl8e7777LfffdR926dencuTOjR4/mmWeeCc3TvXt3Lr74YuLi4hg2bNhBCenhOjoaaUVEJCQ3N5crrriiVEISFxfHli1bGDZsGHl5eaEr76FDhzJ16lTq1KkTc/lbt24tsw9PWloadevWjSmO/Px80tPTQ+Pr169PSkpKmevLy8ujbdu2McWWm5tLfn4+SUlJoXH79u2jR48eAAetNyMjo9zyZs2aRf/+/alduza1a9fmqquuYtasWVxxxRUxxVMifJ0AX3zxBbfccgvLly+nuLiYvXv30rVrVyDY9kb66U9/SnFxMcuWLaNJkybk5OSEYs3NzWXWrFk88sgjofl/+OGHUs1vkVasWBHqIF8iPz+fU089NTScmJhISkoKGzduJDMz86DtzcjIKHcd0WzZsoWbbrqJpUuXUlhYyP79+0lOTo5p2fz8fBo3bkyDBg1KxRHeTNe0adPQ63r16rF79+5K7Z+mGigRkaNMeno6r776Kjt27Aj97d69mxYtWlCnTh0mTpzIqlWreO+993j55ZdDV+VmdljrjVy+vDiaNWtGXl5eaN7i4mK2bdsWdXvWrl0b8zpbt25dap2FhYUsWrQI4KD1rl+/PvESNz4AAB7HSURBVOr2bNiwgTfeeIM5c+bQtGlTmjZtyoIFC1i0aBFbt26NKZ5o43/xi19wyimn8OWXX/Ltt99yzz334JwLbcNXX30VqPwScXFx9O/fn3nz5jFv3jz69u0bSiLS09PJysoqtW+Ki4sZNGhQuWVGat68Obm5uaHh7777jm3bttGiRYvQuMh9XNKUWb9+fYqLi0PTNm/eHHU9EyZMwMxYuXIl3377LXPmzAntIyh/XzRv3pzt27dTWFhYKo7wGKuaEigRkaPMmDFjyMrKCn3JFRQU8NJLLwHw5ptvsnLlSvbt20fDhg2pU6dOqIaoSZMmUb+4KzuOq6++mpdffpl33nmHH374gbvuuivU0TjSkCFDeO2113j++efZu3cv27ZtIycnp8yYTz/9dBo0aMB9993Hrl272LdvH5988kmos3j//v2ZNm0a33zzDRs2bChVGxNp9uzZnHTSSXz++efk5OSQk5PDF198QcuWLUNNUpHrb9KkCdu2bWPnzp3l7pvCwkIaNmxIYmIin332GX/+859D0/r27cumTZt4+OGH+f777yksLGTZsmWh8tetWxd1X4HXjDd//nzmzp0bar4DuP7663nsscdYtmwZzjm+++47XnnllVJJRiwGDRrE008/TU5ODt9//z0TJkzgjDPOCNU+ATzwwAN888035OXl8Yc//IEBAwYA0LlzZ95++23Wr1/Pzp07mTZtWrn7KDExkUaNGrFx40YeeOCBUtPLO17T09M5++yzueOOO9i9ezf//e9/eeqpp0p1YK9qSqBERI4yN910E5deeikXXnghDRo04Mwzzwx9AW/evJmrr76ahg0b0r59e8477zyGDRsWWm7BggUkJydX2DfocOPo2LEjjz76KIMHD6ZZs2YkJydHfaBlq1atWLRoEQ899BCNGzemc+fOof4q1113HatWrSIpKYnLL7+cuLg4Xn75ZXJycmjdujWpqamMHj06lNBMnDiRjIwMWrduzYUXXhja9rLMmjWLX/7yl6Hap5K/MWPGhDqfT5o0iREjRpCUlMTzzz/PKaecwqBBg2jTpg1JSUlRm64efPBBnn32WRo0aMD1118fSjAAGjRowL/+9S+ys7Np2rQpJ554Im+++SYA11xzDQApKSmlmtHCnXHGGdSvX5/8/Hz69OkTGt+tWzeefPJJxo4dS3JyMu3atWPmzJlRtz+aXr16cffdd3PVVVfRrFkz1q5dW6rfFcBll11G165d6dy5M5dccgnXXXcdAL1792bAgAH85Cc/oWvXrvTt2zfqeiZOnMiKFSto1KgRl1xyCVdeeWWp6XfccQdTpkwhKSmJBx988KDl582bx7p162jevDlXXHEFkydPPiKP+Shh4dVlVa1bt26usm8jlCMr+xDb7Sty7vDhVVIuxP7QQjl+rV69mvbt21d3GCJHhciHix4Lop0DzOzfzrluZS2jGigRERGRgJRAiYiIiASkxxiIiIhIzI5k15+aTDVQIiIiIgEpgRIREREJSAmUiIiISEAxJVBm9msz+9TMPjGzeWZW18xam9kyM1tjZvPN7ISqDlZERESkJqgwgTKzFsCNQDfn3I+BOGAgcB/we+dcO+Ab4LqqDFRERESkpoi1Ca82kGBmtYF6wCbgfGCBP30WcHnlhyciIpWhT58+oadrHw1GjhzJnXfeecjLJyYmVurP1ohEqvAxBs65jWb2ILAe2AX8E/g3sMM5t9efbQNw5H7BT0SkCj14mD+6W5HxMd4GnpiYGHpdXFxMfHw8cXFxADz++OMMGTIk5nW++uqrwYIMk5mZyZYtW4iLi6N+/fr06dOH6dOnl4qvOvXs2ZOhQ4cyevTo0LiioqJqjEiOB7E04SUDlwGtgeZAfeCiWFdgZj83s+VmtrygoOCQAxUROd4UFRWF/lq1akV2dnZoODx52rt3bzmlVI6Sda9YsYLly5czZcqUKl+nSE0WSxNeL+B/zrkC59we4G/AOUCS36QH0BLYWNbCzrknnHPdnHPd0tLSKiVoEZHj2ZIlS2jZsiX33XcfTZs2ZdSoUXzzzTf07duXtLQ0kpOT6du3Lxs2bAgt07NnT2bMmAHAzJkz6d69O+PHjyc5OZnWrVvHXEPVokUL+vTpwyeffALAwoUL6dixI0lJSfTs2ZPVq1eH5s3MzGTatGl06NCB5ORkRo0axe7du0vFEM7MWLNmzUHrLG/bsrKyWLp0KWPHjiUxMZGxY8ceVNbOnTsZPnw4aWlpZGRkMGXKFPbv33/Y+0KOb7EkUOuBM82snpkZcAGwCngTuNqfZwTwUtWEKCIikTZv3sz27dvJzc3liSeeYP/+/YwaNYrc3FzWr19PQkJCKJkoy7Jlyzj55JPZunUrt912G9ddd11MT5jOy8tj0aJFdOnShS+++IJBgwbx8MMPU1BQwMUXX0y/fv344YcfQvPPnTuXxYsXs3btWr744otDqrkqb9umTp1Kjx49mD59OkVFRUyfPv2g5ceNG8fOnTv56quveOutt3jmmWd4+umnD3tfyPGtwgTKObcMr7P4CmClv8wTwG+BW8xsDZACPFWFcYqISJhatWoxefJk4uPjSUhIICUlhauuuop69erRoEEDsrKyeOutt6Iun5GRwfXXX09cXBwjRoxg06ZNbNmyJer8l19+OUlJSXTv3p3zzjuPCRMmMH/+fC655BJ69+5NnTp1GD9+PLt27eK9994LLTd27FjS09Np3LgxWVlZzJs3L/C2Bt22cPv27eO5555j2rRpNGjQgMzMTG699VZmz559yPtCBGL8LTzn3ERgYsTor4DTKz0iERGpUFpaGnXr1g0NFxcX8+tf/5p//OMffPPNNwAUFhayb9++UMfzcE2bNg29rlevHlB+x+sXX3yRXr16lRqXn59PRkZGaLhWrVqkp6ezceOBHh3p6emh1xkZGeTn58e6iSFBty3c1q1b2bNnT6k4MzIySsUYdF+IgH5MWI4DVXVHVax3UolUBYs4rh966CE+//xzli1bRtOmTcnJyaFLly5V2hTVvHlzVq5cGRp2zpGXl0eLFgduys7Lywu9Xr9+Pc2bNwegfv36FBcXh6Zt3rw56noq2rbIfREuNTWVOnXqkJubS4cOHUJxhMcocij0Uy4iIseAwsJCEhISSEpKYvv27UyePLnK19m/f39eeeUVXn/9dfbs2cNDDz1EfHw8Z599dmieRx99lA0bNrB9+3amTp3KgAEDAOjUqROffvopOTk57N69m0mTJkVdT0Xb1qRJk6jPfIqLi6N///5kZWVRWFhIbm4uv/vd7xg6dOjh7wA5rimBEhE5Btx8883s2rWL1NRUzjzzTC66KOanzRyyk08+mTlz5jBu3DhSU1PJzs4mOzubE0448MtegwcP5sILL6RNmza0bds29HDMk046ibvuuotevXpx4oknHnRHXpBtu+mmm1iwYAHJycnceOONBy3/yCOPUL9+fdq0aUP37t0ZPHgw1157bSXtBTle2ZG806Bbt25u+fLlR2x9Uvmy27atknLPHT68SsoFeLKcK9vDoSa8Y8fq1atp3759dYdxzMnMzGTGjBkH9Z0SqWminQPM7N/OuW5lLaMaKBEREZGAlECJiIiIBKS78EREpEqsW7euukMQqTKqgRIREREJSAmUiAjopztEjlOH+tlXAiUix724uDj27NlT3WGISDXYtWsXderUCbycEigROe4lJSWxZcsW9u/fX92hiMgR4pyjuLiYjRs38qMf/Sjw8upELiLHvdTUVDZs2MDnn39e3aGIyBFUp04dmjRpQsOGDQMvqwRKRI57tWrVolWrVtUdhogcRdSEJyIiIhKQEigRERGRgJRAiYiIiASkBEpEREQkICVQIiIiIgEpgRIREREJSAmUiIiISEBKoEREREQCUgIlIiIiEpASKBEREZGAlECJiIiIBKQESkRERCQgJVAiIiIiASmBEhEREQlICZSIiIhIQEqgRERERAJSAiUiIiISkBIoERERkYCUQImIiIgEpARKREREJCAlUCIiIiIBKYESERERCUgJlIiIiEhAtas7ABE5MnZOnlwl5TaaOLFKyhURqclUAyUiIiISkBIoERERkYCUQImIiIgEpARKREREJCAlUCIiIiIBKYESERERCUgJlIiIiEhASqBEREREAlICJSIiIhJQTAmUmSWZ2QIz+8zMVpvZWWbW2Mz+ZWZf+v+TqzpYERERkZog1p9y+QPwD+fc1WZ2AlAPmAC87py718xuB24HfltFcYocN7Lbtq2Scs8dPrxKyhUROR5VWANlZo2Ac4GnAJxzPzjndgCXAbP82WYBl1dVkCIiIiI1SSxNeK2BAuBpM/uPmc0ws/pAE+fcJn+ezUCTqgpSREREpCaJJYGqDZwK/Nk51wX4Dq+5LsQ55wBX1sJm9nMzW25mywsKCg43XhEREZFqF0sCtQHY4Jxb5g8vwEuotphZMwD//9dlLeyce8I518051y0tLa0yYhYRERGpVhUmUM65zUCemZ3sj7oAWAUsBEb440YAL1VJhCIiIiI1TKx34Y0D5vp34H0FjMJLvp43s+uAXKB/1YQoIiIiUrPElEA553KAbmVMuqBywxERERGp+fQkchEREZGAlECJiIiIBKQESkRERCQgJVAiIiIiAcV6F56IiIgco3ZOnlwl5TaaOLFKyq0JVAMlIiIiEpASKBEREZGAlECJiIiIBKQ+UCIiIkeJ7LZtq6Tcc4cPr5Jyj2WqgRIREREJSAmUiIiISEBKoEREREQCUgIlIiIiEpASKBEREZGAlECJiIiIBKQESkRERCQgJVAiIiIiASmBEhEREQlICZSIiIhIQEqgRERERAJSAiUiIiISkBIoERERkYCUQImIiIgEpARKREREJCAlUCIiIiIBKYESERERCUgJlIiIiEhASqBEREREAqpd3QGIiIjIselBsyopd7xzVVJuEKqBEhEREQlICZSIiIhIQEqgRERERAJSAiUiIiISkBIoERERkYCUQImIiIgEpARKREREJCAlUCIiIiIBKYESERERCUgJlIiIiEhASqBEREREAlICJSIiIhKQEigRERGRgJRAiYiIiASkBEpEREQkICVQIiIiIgHVru4AREQkugfNqqTc8c5VSbkix4uYa6DMLM7M/mNmL/vDrc1smZmtMbP5ZnZC1YUpIiIiUnMEacK7CVgdNnwf8HvnXDvgG+C6ygxMREREpKaKKYEys5bAJcAMf9iA84EF/iyzgMurIkARERGRmibWGqiHgduA/f5wCrDDObfXH94AtKjk2ERERERqpAoTKDPrC3ztnPv3oazAzH5uZsvNbHlBQcGhFCEiIiJSo8RSA3UOcKmZrQOew2u6+wOQZGYld/G1BDaWtbBz7gnnXDfnXLe0tLRKCFlERESkelWYQDnn7nDOtXTOZQIDgTecc0OAN4Gr/dlGAC9VWZQiIiIiNcjhPEjzt8AtZrYGr0/UU5UTkoiIiEjNFuhBms65JcAS//VXwOmVH5KIiIhIzaYnkYvIYdGTskXkeKTfwhMREREJSAmUiIiISEBKoEREREQCUgIlIiIiEpASKBEREZGAlECJiIiIBKQESkRERCQgJVAiIiIiASmBEhEREQlICZSIiIhIQEqgRERERAJSAiUiIiISkBIoERERkYCUQImIiIgEpARKREREJCAlUCIiIiIBKYESERERCUgJlIiIiEhASqBEREREAlICJSIiIhKQEigRERGRgJRAiYiIiASkBEpEREQkICVQIiIiIgEpgRIREREJSAmUiIiISEBKoEREREQCUgIlIiIiEpASKBEREZGAlECJiIiIBKQESkRERCQgJVAiIiIiASmBEhEREQlICZSIiIhIQEqgRERERAJSAiUiIiISkBIoERERkYCUQImIiIgEpARKREREJCAlUCIiIiIBKYESERERCUgJlIiIiEhAtas7ABGRIym7bdsqKbff2rVVUq6I1EyqgRIREREJSAmUiIiISEBKoEREREQCqjCBMrN0M3vTzFaZ2admdpM/vrGZ/cvMvvT/J1d9uCIiIiLVL5YaqL3Arc65DsCZwK/MrANwO/C6c+5E4HV/WEREROSYV2EC5Zzb5Jxb4b8uBFYDLYDLgFn+bLOAy6sqSBEREZGaJFAfKDPLBLoAy4AmzrlN/qTNQJNKjUxERESkhoo5gTKzROCvwM3OuW/DpznnHOCiLPdzM1tuZssLCgoOK1gRERGRmiCmBMrM6uAlT3Odc3/zR28xs2b+9GbA12Ut65x7wjnXzTnXLS0trTJiFhEREalWsdyFZ8BTwGrn3O/CJi0ERvivRwAvVX54IiIiIjVPLD/lcg4wDFhpZjn+uAnAvcDzZnYdkAv0r5oQRURERGqWChMo59w7gEWZfEHlhiMiIiJS8+lJ5CIiIiIBKYESERERCUgJlIiIiEhAsXQiFxGRCuycPLm6QxCRI0g1UCIiIiIBKYESERERCUgJlIiIiEhASqBEREREAlICJSIiIhKQEigRERGRgJRAiYiIiASkBEpEREQkICVQIiIiIgEpgRIREREJSAmUiIiISEBKoEREREQCUgIlIiIiEpASKBEREZGAlECJiIiIBKQESkRERCQgJVAiIiIiASmBEhEREQlICZSIiIhIQEqgRERERAJSAiUiIiISkBIoERERkYCUQImIiIgEpARKREREJCAlUCIiIiIBKYESERERCUgJlIiIiEhASqBEREREAqpd3QGIiIhUZOfkyVVSbqOJE6ukXDn2KYESEZFKk922bZWUe+7w4VVSrsihUhOeiIiISEBKoEREREQCUgIlIiIiEpASKBEREZGA1IlcRESOWw+aVUm5452rknKl5lANlIiIiEhASqBEREREAlICJSIiIhKQEigRERGRgJRAiYiIiASku/Cqme4AEREROfqoBkpEREQkoMNKoMzsIjP73MzWmNntlRWUiIiISE12yAmUmcUBjwJ9gA7AIDPrUFmBiYiIiNRUh1MDdTqwxjn3lXPuB+A54LLKCUtERESk5jqcTuQtgLyw4Q3AGYcXzuHLbtu2Ssrtt3ZtlZQrIiIiRx9zh3i3lpldDVzknBvtDw8DznDOjY2Y7+fAz/3Bk4HPDz3co0YqsLW6gziGaf9WPe3jqqX9W7W0f6vW8bR/M5xzaWVNOJwaqI1AethwS39cKc65J4AnDmM9Rx0zW+6c61bdcRyrtH+rnvZx1dL+rVrav1VL+9dzOH2gPgJONLPWZnYCMBBYWDlhiYiIiNRch1wD5Zzba2ZjgcVAHPAX59ynlRaZiIiISA11WE8id84tAhZVUizHkuOqybIaaP9WPe3jqqX9W7W0f6uW9i+H0YlcRERE5Hiln3IRERERCUgJVABmts/McszsYzNbYWZnx7BMUZTxM/1HQRy3zPOOmfUJG3eNmf2jOuMSORRm5sxsTthwbTMrMLOXK1iuZ0XziFQWM8sys0/N7L/+91lMz280s+ZmtsB/HTpmzezS4/Wn3JRABbPLOdfZOdcJuAOYVt0BHc2c1348BvidmdU1s0TgHuBX1RvZkRWZZJvZSDObXo3xTPKTgXZh4272x5V767KZrTOz1KqPskb6DvixmSX4w70p49EuckDYReknZvaCmdUrZ95D+lyYWRMze9m/8F1lZov88WUmrsdyQmBmZwF9gVOdcz8BelH6gdhROefynXMHXfQ75xY65+6t3EiPDkqgDl1D4BsAM0s0s9f9WqmVZnbQT9r4tS3T/R9ffg340ZEOuCZyzn0CZAO/Be4C5gAP+VdHH5jZTyD0pT6+ZDn/hJvp/602syf9q6p/lnyBmdlpYVdZD5jZJ9WwiUerlXiPJilxDaC7bCu2CLjEfz0ImFcywcxON7P3zew/ZvaemZ0cubCZ1Tezv5jZh/58x/rPY5VclP4Y+AHvgqqy/R/wL+dcJ+dcB6Dc5OgYTwiaAVudc98DOOe2Oufy/Qufaf65crmZnWpmi81srZmNAfDPtQedQ8MTW79l5Y/+8f1VSSuLmdUysz+Z2Wdm9i8zW3QstMAogQomwT/APgNmAHf743cDVzjnTgV+ipcAWMSyV+A9ib0DMByosPnvODIZGIz3w9RNgf/4V0cTgGdiWP5E4FHnXEdgB3CVP/5p4AbnXGdgX6VHfQSYWZqZ/dXMPvL/zvHHl/ll7CedHcOWX+LP+6WZpfnjapnZmpLhKF7E/21LM2sL7CTsycNm9mf/RPupmU0uI+4EM3vVzK6PlhRE1ij4tQQ9/ddFZvZ7v/zXK4i1JnkOGGhmdYGfAMvCpn0G9HDOdcG7WLinjOWzgDecc6fjnUseMLP6VRxzTbEUaGdm/cxsmX+svGZmTSJnjPa5iKIZ3k+NAeCc+28Z5Z3mr69tLAnBUeyfQLqZfeEnNOeFTVvvnyuXAjOBq4Ez8c7PQTQDuuPVdJUkolcCmXjff8OAsw51A2oSJVDBlFwtnQJcBDzjJ0oG3GNm/wVew/udwMgP/bnAPOfcPudcPvDGkQy8JnPOfQfMB2bjfWBn++PfAFLMrGEFRfzPOZfjv/43kGlmSUAD59z7/vhnKz/ySlOSmOeYWQ7eFXOJPwC/d86dhpcYzvDHR/syng/0BzCzZkAz59yHeDV7Q/x5egEfO+cKyonpWyDPzH6MVxM1P2J6lv8k4p8A55lfU+hLxKtVnOece5JDSwrqA8v9pPgtYGIF89cI/pdzJl7tU+QjXhoBL/hX8b8HOnKwC4Hb/eNgCVAXaFVV8dYUZlYb7wJqJfAOcKZ/bD8H3FbGItE+F2V5FHjKzN40r/9P84h1nw08BlzmnCvrR0/LSgiOSs65IqAr3s+rFQDzzWykP7nkQdgrgWXOuUL/HPG9fz6N1YvOuf3OuVUc+B7sDrzgj98MvHm421ITHNZzoI5nzrn3zevrkQZc7P/v6pzbY2br8E58Erv9/l80eymd8Ifv3+/DXu8DEji67PKv/ACvZgYo6WvUC+gQVqHZ0Ly+Yo2AWWZ2IuCAOv705/GuMifiJVIL/PF/AV4CHgauxaudq8hzeMnTz4ALgFFh0/qb9zuXtfG+YDoAJVf2LwH3O+fm+sMXApfagSbYWJKC/RxI2uYAf4sh3ppiIfAg0BNICRt/N/Cmc+4KM8vES5AiGXCVc+54+M1Q8C8e/NdLgafwaurn+xcAJwD/K2O5Mj8XfoJQinNusZm1wbvo7QP8x78wAGiP90yjC/0L27K86JzbD6wqqzbsaOOc24d37C0xs5XACH9SyXl0P6XPqfsJliuELxvZEnNMUQ3UITKzU/CewL4N78vsaz95+imQUcYibwMDzCzOPzH89MhFe1RZil9T4jfnbHXOfQusA071x58KtC6vEOfcDqDQDtxhMrC8+WuwWnhX4539vxb+l0TJl/GPgX74CaVzbiOwza8RGoCfhDjn8oAtZnY+cDrwagzrfhmvun29/x4AYGatgfHABX5T6yuUTmjfBS4Ka8YuSQpKtqGVc2415SfFkY6mB9b9BZjsnFsZMb4RBzqVj4yy7GJgXMm+M7MuVRJhzbEr7LgY55z7AXgEmO6c+3/ADZR9XET7XJTJObfdOfesc24Y3s+QnetP2oTXBaO8/XzMJARmdrJ/0VWiM5B7BFb9LnCV332gCd7FxVFPCVQwoaYWvC+mEX42Pxfo5mfzw/GaVyL9HfgSWIXXr+f9MuYRmAR09ZtD7+XA1dFfgcZm9ikwFvgihrKuA57036/6eP14jjb/BMaVDJhZSU1VeV/G8/GaPRpF9PeYgVeb84J/3JbLOVeM17l/asSkhnh3nO30T4Z9IqbfhXeDxaP+cLSkYB3Q2T+ppuMldiVq4fXBAK9/3DsVxVtTOOc2OOf+WMak+4FpZvYfol/R341Xm/hf/1i/O8p8x7LwY3tElHmifS4OYmbnm393n5k1ANoC6/3JO/A6/U8r6X93jEvEq7le5Z9jO+Cdc6vaX/H6oa3COwet4Og8H5eiJ5HLMSu8St+825KbOeduquawDmJmRc65xLDhkUA359xYv5n4UbymhtrA2865MebdjjwLL5F5BRjqnMv0l2+C9wV0t3Nucli5dfBqTE93zpWV5JfMNwkocs49GDF+CTDeObfczGbi3QiRh3ciXOicm+k3X3fz1/MXvH4WE/GaDs/GS4z+55zr6ydUc/D6ZKwGkoFJzrkl5j3a4Qm85r+vgQEV9NmSo1Dkse+Puwyvj9g3eH1FT3PO9YzlcxFlHb/Ba34uqfF82jn3kJ8wjfePxVZ4tbLX+mWWrGcm8LJzruT5RwfFK7EpOR+bWQrwIXCO3x/qqKUESo5ZZjYA73ldtfGqqUcez1/C5j3D6ffOuR7VHUtF9EUlcmzxL8CS8Pq13e+cm1mtAVUCJVAixwG/Bu4XwBDnXI1vDlMCJSI1nRIokeOUmWXhPSAz3AvOucg+TyI1npmNAiKb6N91zh1Xv2wgR44SKBEREZGAdBeeiIiISEBKoEREREQCUgIlIiIiEpASKBEREZGAlECJiIiIBPT/A7tWggLr3JBOAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xaXtIkHNNK9b"
      },
      "source": [
        "**3. Best and Worst attributes to have to be classified attractive**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "N-JDV0RIYm-t",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "46cf9cdc-bd5f-47e1-9047-7683780b5528"
      },
      "source": [
        "test_df = test_df_without_data\r\n",
        "\r\n",
        "test_best_attr = {attr:len(test_df[(test_df[\"Attractive\"]==True) & (test_df[attr]==True)])/len(test_df[test_df[\"Attractive\"]==True])*100 if attr!=\"Attractive\" else 0 for attr in testset.attr2idx.keys()}\r\n",
        "predict_best_attr = {attr:len(test_df[(test_df[\"status\"].isin(['False Positive',\"True Positive\"])) & (test_df[attr]==True)])/len(test_df[test_df[\"status\"].isin(['False Positive',\"True Positive\"])])*100 if attr!=\"Attractive\" else 0 for attr in testset.attr2idx.keys()}\r\n",
        "top_5_best_attr_test = sorted(test_best_attr.items(),key=lambda x: x[1],reverse=True)[:5]\r\n",
        "top_5_best_attr_predict = sorted(predict_best_attr.items(),key=lambda x: x[1],reverse=True)[:5]\r\n",
        "top_5_worst_attr_test = sorted(test_best_attr.items(),key=lambda x: x[1])[1:6]\r\n",
        "top_5_worst_attr_predict = sorted(predict_best_attr.items(),key=lambda x: x[1])[1:6]\r\n",
        "\r\n",
        "print(\"---------- Statistics in Test Dataset vs Network output ----------\")\r\n",
        "print(\"\\n\")\r\n",
        "print(\"Top Attributes to have\")\r\n",
        "print(\"--------------\")\r\n",
        "print(\"\\n\".join(map(lambda x,y: x[0]+' '+y[0]+\" = \"+str(round(x[1],2))+\"% vs \"+str(round(y[1],2))+\"%\",top_5_best_attr_test,top_5_best_attr_predict)))\r\n",
        "print(\"\\n\")\r\n",
        "print(\"Worst Attributes to have\")\r\n",
        "print(\"--------------\")\r\n",
        "print(\"\\n\".join(map(lambda x,y: x[0]+' '+y[0]+\" = \"+str(round(x[1],2))+\"% vs \"+str(round(y[1],2))+\"%\",top_5_worst_attr_test,top_5_worst_attr_predict)))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "---------- Statistics in Test Dataset vs Network output ----------\n",
            "\n",
            "\n",
            "Top Attributes to have\n",
            "--------------\n",
            "Young No_Beard = 92.1% vs 93.14%\n",
            "No_Beard Young = 92.06% vs 91.77%\n",
            "Wearing_Lipstick Wearing_Lipstick = 76.1% vs 78.06%\n",
            "Heavy_Makeup Heavy_Makeup = 63.6% vs 64.88%\n",
            "Smiling Smiling = 56.49% vs 56.75%\n",
            "\n",
            "\n",
            "Worst Attributes to have\n",
            "--------------\n",
            "Bald Bald = 0.15% vs 0.03%\n",
            "Gray_Hair Gray_Hair = 0.18% vs 0.14%\n",
            "Double_Chin Double_Chin = 0.26% vs 0.42%\n",
            "Chubby Chubby = 0.31% vs 0.53%\n",
            "Eyeglasses Eyeglasses = 1.04% vs 0.6%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LLhbcmABb6JF"
      },
      "source": [
        "# Fairness evaluation : fairness metrics calculation"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gbTe4d-tcCHU"
      },
      "source": [
        "### Attribute \"Eyeglasses\""
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "08raA9F_b9bK",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "fcf7c95d-423a-4d11-a048-b41e6c75d7bc"
      },
      "source": [
        "print(\"----------STATISCTICS WITHOUT DATA - EYEGLASSES ATTRIBUTE ----------\")\r\n",
        "print(\"\\n\")\r\n",
        "print(\"Statistical Parity\")\r\n",
        "#P(M(X)=1 | A=a) = P(M(X)=1 | B=b)\r\n",
        "print(len(test_df[(test_df[\"status\"].isin([\"False Positive\",\"True Positive\"])) & (test_df[\"Eyeglasses\"]==True)])/len(test_df[test_df[\"Eyeglasses\"]==True]))\r\n",
        "print(len(test_df[(test_df[\"status\"].isin([\"False Positive\",\"True Positive\"])) & (test_df[\"Eyeglasses\"]==False)])/len(test_df[test_df[\"Eyeglasses\"]==False]))\r\n",
        "print(\"\\n\")\r\n",
        "print(\"True Positive Rate\")\r\n",
        "#P(M(X)=1 | A=a, y=1) = P(M(X)=1 | A=b, y=1)\r\n",
        "print(\"--------------\")\r\n",
        "print(len(test_df[(test_df[\"status\"]==\"True Positive\") & (test_df[\"Eyeglasses\"]==True)])/len(test_df[(test_df[\"status\"].isin([\"False Negative\",\"True Positive\"])) & (test_df[\"Eyeglasses\"]==True)]))\r\n",
        "print(len(test_df[(test_df[\"status\"]==\"True Positive\") & (test_df[\"Eyeglasses\"]==False)])/len(test_df[(test_df[\"status\"].isin([\"False Negative\",\"True Positive\"])) & (test_df[\"Eyeglasses\"]==False)]))\r\n",
        "print(\"\\n\")\r\n",
        "print(\"False Positive Rate\")\r\n",
        "#P(M(X)=1 | A=a, y=0) = P(M(X)=1 | A=b, y=0)\r\n",
        "print(\"----------------\")\r\n",
        "print(len(test_df[(test_df[\"status\"]==\"False Positive\") & (test_df[\"Eyeglasses\"]==True)])/len(test_df[(test_df[\"status\"].isin([\"True Negative\",\"False Positive\"])) & (test_df[\"Eyeglasses\"]==True)]))\r\n",
        "print(len(test_df[(test_df[\"status\"]==\"False Positive\") & (test_df[\"Eyeglasses\"]==False)])/len(test_df[(test_df[\"status\"].isin([\"True Negative\",\"False Positive\"])) & (test_df[\"Eyeglasses\"]==False)]))\r\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "----------STATISCTICS WITHOUT DATA - EYEGLASSES ATTRIBUTE ----------\n",
            "\n",
            "\n",
            "Statistical Parity\n",
            "0.04809930178432894\n",
            "0.5462432389010872\n",
            "\n",
            "\n",
            "True Positive Rate\n",
            "--------------\n",
            "0.2524271844660194\n",
            "0.8241960183767228\n",
            "\n",
            "\n",
            "False Positive Rate\n",
            "----------------\n",
            "0.03035413153456998\n",
            "0.2395809867087182\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "M3bXJEpYcF0N"
      },
      "source": [
        "### Attribute \"Bald\""
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TaKLbzczcJpD",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "33a81818-0beb-4a69-9a57-b3d399f6483b"
      },
      "source": [
        "test_df = test_df_without_data\r\n",
        "\r\n",
        "print(\"----------STATISCTICS WITHOUT DATA - BALD ATTRIBUTE ----------\")\r\n",
        "print(\"\\n\")\r\n",
        "print(\"Statistical Parity\")\r\n",
        "#P(M(X)=1 | A=a) = P(M(X)=1 | B=b)\r\n",
        "print(len(test_df[(test_df[\"status\"].isin([\"False Positive\",\"True Positive\"])) & (test_df[\"Bald\"]==True)])/len(test_df[test_df[\"Bald\"]==True]))\r\n",
        "print(len(test_df[(test_df[\"status\"].isin([\"False Positive\",\"True Positive\"])) & (test_df[\"Bald\"]==False)])/len(test_df[test_df[\"Bald\"]==False]))\r\n",
        "print(\"\\n\")\r\n",
        "print(\"True Positive Rate\")\r\n",
        "#P(M(X)=1 | A=a, y=1) = P(M(X)=1 | A=b, y=1)\r\n",
        "print(\"--------------\")\r\n",
        "print(len(test_df[(test_df[\"status\"]==\"True Positive\") & (test_df[\"Bald\"]==True)])/len(test_df[(test_df[\"status\"].isin([\"False Negative\",\"True Positive\"])) & (test_df[\"Bald\"]==True)]))\r\n",
        "print(len(test_df[(test_df[\"status\"]==\"True Positive\") & (test_df[\"Bald\"]==False)])/len(test_df[(test_df[\"status\"].isin([\"False Negative\",\"True Positive\"])) & (test_df[\"Bald\"]==False)]))\r\n",
        "print(\"\\n\")\r\n",
        "print(\"False Positive Rate\")\r\n",
        "#P(M(X)=1 | A=a, y=0) = P(M(X)=1 | A=b, y=0)\r\n",
        "print(\"----------------\")\r\n",
        "print(len(test_df[(test_df[\"status\"]==\"False Positive\") & (test_df[\"Bald\"]==True)])/len(test_df[(test_df[\"status\"].isin([\"True Negative\",\"False Positive\"])) & (test_df[\"Bald\"]==True)]))\r\n",
        "print(len(test_df[(test_df[\"status\"]==\"False Positive\") & (test_df[\"Bald\"]==False)])/len(test_df[(test_df[\"status\"].isin([\"True Negative\",\"False Positive\"])) & (test_df[\"Bald\"]==False)]))\r\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "----------STATISCTICS WITHOUT DATA - BALD ATTRIBUTE ----------\n",
            "\n",
            "\n",
            "Statistical Parity\n",
            "0.0070921985815602835\n",
            "0.5250524591841957\n",
            "\n",
            "\n",
            "True Positive Rate\n",
            "--------------\n",
            "0.13333333333333333\n",
            "0.819285642011535\n",
            "\n",
            "\n",
            "False Positive Rate\n",
            "----------------\n",
            "0.0024509803921568627\n",
            "0.22390223695111847\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aSFCG-qfe79X"
      },
      "source": [
        "# First fix : Adjusting the thresholds"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AG7LMMgmfcHQ"
      },
      "source": [
        "## Calculating the FPR and TPR for different thresholds"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xzn0njEFfjBS"
      },
      "source": [
        " def get_rates(treshold, attr):\r\n",
        "  #create the dataframe\r\n",
        "  test_df = pd.DataFrame(testset_allattr)\r\n",
        "  test_df.columns = [\"image_link\",\"attrs\"]\r\n",
        "  test_df[[key for key in testset.attr2idx.keys()]] = test_df['attrs'].apply(pd.Series)\r\n",
        "  test_df.pop(\"attrs\")\r\n",
        "  test_df[\"status\"]=[\"Not_known_yet\" for i in range(len(test_df.index))]\r\n",
        "  test_df.head()\r\n",
        "\r\n",
        "  #do the test\r\n",
        "  correct = 0\r\n",
        "  total = 0\r\n",
        "  idx = 0 \r\n",
        "  not_predicted_as_attractive_idx = []\r\n",
        "\r\n",
        "  net.eval()\r\n",
        "  for data in testloader:\r\n",
        "      images, labels = data\r\n",
        "      labels = labels.cuda()\r\n",
        "      outputs = net(images.cuda())\r\n",
        "      predicted = torch.round(outputs.data - treshold + 0.5)\r\n",
        "      predicted = predicted.cuda()\r\n",
        "      total += labels.size(0)\r\n",
        "      correct += (predicted == labels).sum()\r\n",
        "\r\n",
        "      for predict,real in zip(predicted,labels):\r\n",
        "\r\n",
        "        if predict==real:\r\n",
        "          if real==1:\r\n",
        "            test_df.at[idx,\"status\"] = \"True Positive\"\r\n",
        "          else:\r\n",
        "            test_df.at[idx,\"status\"] = \"True Negative\"\r\n",
        "        else:\r\n",
        "          if real==1:\r\n",
        "            test_df.at[idx,\"status\"] = \"False Negative\"\r\n",
        "          else:\r\n",
        "            test_df.at[idx,\"status\"] = \"False Positive\"        \r\n",
        "\r\n",
        "        if predict==0:\r\n",
        "          not_predicted_as_attractive_idx.append(idx)\r\n",
        "        idx+=1\r\n",
        "\r\n",
        "  print(\"treshold : \" + str(treshold)+'\\n')\r\n",
        "  print(total)\r\n",
        "  print(correct)\r\n",
        "  #ici nous on a pas testset.data\r\n",
        "  print('Accuracy of the network on the {} test images: {:4.2f} %'.format(\r\n",
        "      len(testset), 100 * correct.true_divide(total)))\r\n",
        "\r\n",
        "  #false positive rate with attribute\r\n",
        "  fp_w = len(test_df[(test_df[\"status\"]==\"False Positive\") & (test_df[attr]==True)])/len(test_df[(test_df[\"status\"].isin([\"True Negative\",\"False Positive\"])) & (test_df[attr]==True)])\r\n",
        "  #true positive rate with attribute\r\n",
        "  tp_w = len(test_df[(test_df[\"status\"]==\"True Positive\") & (test_df[attr]==True)])/len(test_df[(test_df[\"status\"].isin([\"False Negative\",\"True Positive\"])) & (test_df[attr]==True)])\r\n",
        "  #false positive rate without attribute\r\n",
        "  fp_wo = len(test_df[(test_df[\"status\"]==\"False Positive\") & (test_df[attr]==False)])/len(test_df[(test_df[\"status\"].isin([\"True Negative\",\"False Positive\"])) & (test_df[attr]==False)])\r\n",
        "  #true positive rate without attribute\r\n",
        "  tp_wo = len(test_df[(test_df[\"status\"]==\"True Positive\") & (test_df[attr]==False)])/len(test_df[(test_df[\"status\"].isin([\"False Negative\",\"True Positive\"])) & (test_df[attr]==False)])\r\n",
        "\r\n",
        "  return tp_w, fp_w, tp_wo, fp_wo"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3mpo-N9yfu4Q"
      },
      "source": [
        "## Plotting the ROC curves"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RfAwoKJNfwvD",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "353ba565-8b9b-4868-b465-8a0befa0f09f"
      },
      "source": [
        "category = 'Eyeglasses'\r\n",
        "i = 0\r\n",
        "x_w = []\r\n",
        "y_w = []\r\n",
        "x_wo = []\r\n",
        "y_wo = []\r\n",
        "interval = np.arange(0,1,0.05)\r\n",
        "for t in interval:\r\n",
        "  result = get_rates(t, category)\r\n",
        "  y_w.append(result[0])\r\n",
        "  x_w.append(result[1])\r\n",
        "  y_wo.append(result[2])\r\n",
        "  x_wo.append(result[3])\r\n",
        "\r\n",
        "plt.plot(x_w,y_w, label = category, color = 'pink' )\r\n",
        "plt.plot(x_wo,y_wo, label = 'without_'+category, color = 'blue')\r\n",
        "plt.legend(title='ROC curves for the Eyeglasses attribute')\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "treshold : 0.0\n",
            "\n",
            "19962\n",
            "tensor(9905, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 49.62 %\n",
            "treshold : 0.05\n",
            "\n",
            "19962\n",
            "tensor(13298, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 66.62 %\n",
            "treshold : 0.1\n",
            "\n",
            "19962\n",
            "tensor(13924, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 69.75 %\n",
            "treshold : 0.15000000000000002\n",
            "\n",
            "19962\n",
            "tensor(14386, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 72.07 %\n",
            "treshold : 0.2\n",
            "\n",
            "19962\n",
            "tensor(14740, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 73.84 %\n",
            "treshold : 0.25\n",
            "\n",
            "19962\n",
            "tensor(15104, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 75.66 %\n",
            "treshold : 0.30000000000000004\n",
            "\n",
            "19962\n",
            "tensor(15350, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 76.90 %\n",
            "treshold : 0.35000000000000003\n",
            "\n",
            "19962\n",
            "tensor(15621, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 78.25 %\n",
            "treshold : 0.4\n",
            "\n",
            "19962\n",
            "tensor(15811, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 79.21 %\n",
            "treshold : 0.45\n",
            "\n",
            "19962\n",
            "tensor(15932, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 79.81 %\n",
            "treshold : 0.5\n",
            "\n",
            "19962\n",
            "tensor(16000, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 80.15 %\n",
            "treshold : 0.55\n",
            "\n",
            "19962\n",
            "tensor(15955, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 79.93 %\n",
            "treshold : 0.6000000000000001\n",
            "\n",
            "19962\n",
            "tensor(15837, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 79.34 %\n",
            "treshold : 0.65\n",
            "\n",
            "19962\n",
            "tensor(15621, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 78.25 %\n",
            "treshold : 0.7000000000000001\n",
            "\n",
            "19962\n",
            "tensor(15360, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 76.95 %\n",
            "treshold : 0.75\n",
            "\n",
            "19962\n",
            "tensor(15048, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 75.38 %\n",
            "treshold : 0.8\n",
            "\n",
            "19962\n",
            "tensor(14715, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 73.72 %\n",
            "treshold : 0.8500000000000001\n",
            "\n",
            "19962\n",
            "tensor(14236, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 71.32 %\n",
            "treshold : 0.9\n",
            "\n",
            "19962\n",
            "tensor(13663, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 68.45 %\n",
            "treshold : 0.9500000000000001\n",
            "\n",
            "19962\n",
            "tensor(12835, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 64.30 %\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pggUhVPwgYOC"
      },
      "source": [
        "## Finding the tresholds"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VZQKZoQEgZ_F",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c2ccab48-07c6-4331-cb86-4c8bb6a1b8d3"
      },
      "source": [
        "#Printing the rates\r\n",
        "print('TPR for eyeglasses')\r\n",
        "print(x_w)\r\n",
        "print('FPR for eyeglasses')\r\n",
        "print(y_w)\r\n",
        "print('TPR without eyeglasses')\r\n",
        "print(x_wo)\r\n",
        "print('FPR without eyeglasses')\r\n",
        "print(y_wo)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "TPR for eyeglasses\n",
            "[0.9966273187183811, 0.33389544688026984, 0.24957841483979765, 0.19392917369308602, 0.16020236087689713, 0.11973018549747048, 0.09190556492411467, 0.0699831365935919, 0.05480607082630692, 0.04384485666104553, 0.03035413153456998, 0.024451939291736932, 0.016863406408094434, 0.01433389544688027, 0.012647554806070826, 0.00927487352445194, 0.00505902192242833, 0.003372681281618887, 0.002529510961214165, 0.0]\n",
            "FPR for eyeglasses\n",
            "[1.0, 0.912621359223301, 0.883495145631068, 0.8155339805825242, 0.7087378640776699, 0.6504854368932039, 0.5631067961165048, 0.5242718446601942, 0.4174757281553398, 0.32038834951456313, 0.2524271844660194, 0.1941747572815534, 0.1553398058252427, 0.14563106796116504, 0.13592233009708737, 0.1262135922330097, 0.11650485436893204, 0.06796116504854369, 0.019417475728155338, 0.0]\n",
            "TPR without eyeglasses\n",
            "[0.9996620860554178, 0.6979049335435908, 0.6296463167380041, 0.575580085604866, 0.5278215814372607, 0.47814823158369, 0.4375985582338365, 0.3898400540662311, 0.3439963955845911, 0.2982653750844785, 0.2395809867087182, 0.19103401667042127, 0.15307501689569722, 0.12198693399414283, 0.0975444920027033, 0.07591799954944807, 0.056206352782158145, 0.040774949312908315, 0.026695201621986932, 0.011826988060373958]\n",
            "FPR without eyeglasses\n",
            "[1.0, 0.9935681470137825, 0.9857069933639612, 0.9778458397141399, 0.9677386421643696, 0.9555895865237366, 0.9415007656967841, 0.9236345074017356, 0.9007656967840735, 0.8713629402756509, 0.8241960183767228, 0.7754977029096478, 0.728534966819806, 0.6781010719754977, 0.6291985706993364, 0.5774374680959673, 0.5251659009698826, 0.4625829504849413, 0.3917304747320061, 0.2936191934660541]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dvKHJLpzgqtX"
      },
      "source": [
        "We choose tresholds so that the rates are equal"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "H-WwAcAWgjLr",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "048798e6-4fd2-43d2-9d56-6e1f4c143c2a"
      },
      "source": [
        "#Choosing a treshold of 0.4 for people with eyeglasses\r\n",
        "print(get_rates(0.4, \"Eyeglasses\"))\r\n",
        "#with eye glasses :\r\n",
        "#true positive rate : 0.4563106796116505\r\n",
        "#false positive rate : 0.046374367622259695"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "treshold : 0.4\n",
            "\n",
            "19962\n",
            "tensor(15945, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 79.88 %\n",
            "(0.4563106796116505, 0.046374367622259695, 0.8967840735068913, 0.32608695652173914)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QNPcbT4zgn0S",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c320df40-4090-4a9b-8d9a-2e975a50adf2"
      },
      "source": [
        "#Choosing a treshold of 0.85 for people without eyeglasses\r\n",
        "print(get_rates(0.85, \"Eyeglasses\"))\r\n",
        "#without eye glasses :\r\n",
        "#true positive rate : 0.46248085758039814\r\n",
        "#false positive rate : 0.04190132912818202"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "treshold : 0.85\n",
            "\n",
            "19962\n",
            "tensor(14225, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 71.26 %\n",
            "(0.02912621359223301, 0.0, 0.46248085758039814, 0.04190132912818202)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0Tt_HAEUbxkb"
      },
      "source": [
        "# Second fix : Adding data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4gA8hORtag5j"
      },
      "source": [
        "## With added data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "n5c85NSNap4E"
      },
      "source": [
        "Here the model that needs to be loaded is the one trained with the added data."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rEEmkEK4ajyZ"
      },
      "source": [
        "##### TEST DATAFRAME #####\r\n",
        "test_df_with_data = pd.DataFrame(testset_allattr)\r\n",
        "test_df_with_data.columns = [\"image_link\",\"attrs\"]\r\n",
        "test_df_with_data[[key for key in testset.attr2idx.keys()]] = test_df_with_data['attrs'].apply(pd.Series)\r\n",
        "test_df_with_data.pop(\"attrs\")\r\n",
        "test_df_with_data[\"status\"]=[\"Not_known_yet\" for i in range(len(test_df_with_data.index))]\r\n",
        "test_df_with_data[\"predicted_value\"]=[str(0) for i in range(len(test_df_with_data.index))]\r\n",
        "\r\n",
        "##### TRAIN DATAFRAME #####\r\n",
        "train_df_with_data = pd.DataFrame(trainset_allattr)\r\n",
        "train_df_with_data.columns = [\"image_link\",\"attrs\"]\r\n",
        "train_df_with_data[[key for key in trainset.attr2idx.keys()]] = train_df_with_data['attrs'].apply(pd.Series)\r\n",
        "del train_df_with_data[\"attrs\"]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZBn4wa3ca8ab"
      },
      "source": [
        "### Testing the model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "suIK1uyta6Tz",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e0031537-ad64-4782-e21f-4930f94840a4"
      },
      "source": [
        "correct = 0\r\n",
        "total = 0\r\n",
        "idx = 0 \r\n",
        "not_predicted_as_attractive_idx_with_data = []\r\n",
        "\r\n",
        "net.eval()\r\n",
        "for data in testloader:\r\n",
        "    images, labels = data\r\n",
        "    labels = labels.cuda()\r\n",
        "    outputs = net(images.cuda())\r\n",
        "    predicted_raw = outputs.data\r\n",
        "    predicted = torch.round(outputs.data)\r\n",
        "    predicted = predicted.cuda()\r\n",
        "    total += labels.size(0)\r\n",
        "    correct += (predicted == labels).sum()\r\n",
        "    for predict,predict_raw, real in zip(predicted,predicted_raw,labels):\r\n",
        "      test_df_with_data.at[idx,\"predicted_value\"] = str(predict_raw.item())\r\n",
        "      if predict==real:\r\n",
        "        if real==1:\r\n",
        "          test_df_with_data.at[idx,\"status\"] = \"True Positive\"\r\n",
        "        else:\r\n",
        "          test_df_with_data.at[idx,\"status\"] = \"True Negative\"\r\n",
        "      else:\r\n",
        "        if real==1:\r\n",
        "          test_df_with_data.at[idx,\"status\"] = \"False Negative\"\r\n",
        "        else:\r\n",
        "          test_df_with_data.at[idx,\"status\"] = \"False Positive\"        \r\n",
        "\r\n",
        "      if predict==0:\r\n",
        "        not_predicted_as_attractive_idx_with_data.append(idx)\r\n",
        "      idx+=1\r\n",
        "\r\n",
        "print(total)\r\n",
        "print(correct)\r\n",
        "\r\n",
        "print('Accuracy of the network on the {} test images: {:4.2f} %'.format(\r\n",
        "    len(testset), 100 * correct.true_divide(total)))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "19962\n",
            "tensor(16038, device='cuda:0')\n",
            "Accuracy of the network on the 19962 test images: 80.34 %\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RoZRzlWCOPPq"
      },
      "source": [
        "**1. Proportions of False Positives in the classified images**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D_4lPnSG7pDC",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 431
        },
        "outputId": "f238e05d-c71a-49bf-ce1c-c2ecf168135f"
      },
      "source": [
        "### Proportion of False Positive regarding attributes: \r\n",
        "\r\n",
        "fig = plt.figure(figsize=(10,5)) \r\n",
        "\r\n",
        "interesting_features = [\"Bald\",\"Young\",\"Heavy_Makeup\",\"Male\",\"Pale_Skin\",\"Smiling\"]\r\n",
        "\r\n",
        "# Plot accidents depending on type\r\n",
        "#for attr in trainset.attr2idx.keys():\r\n",
        "i=1\r\n",
        "\r\n",
        "for attr in trainset.attr2idx.keys():\r\n",
        "  if attr!=\"Attractive\":\r\n",
        "    #print(attr+\" TP+FP:\"+str(len(test_df_with_data[(test_df_with_data[\"status\"].isin([\"False Positive\"])) & (test_df_with_data[attr]==True)])))\r\n",
        "    plt.bar(i,len(test_df_with_data[(test_df_with_data[\"status\"]==\"False Positive\") & (test_df_with_data[attr]==True)])/len(test_df_with_data[(test_df_with_data[\"status\"].isin([\"True Positive\",\"False Positive\"])) & (test_df_with_data[attr]==True)])*100)\r\n",
        "    i+=1\r\n",
        "\r\n",
        "plt.xticks([r for r in range(1,len(trainset.attr2idx.keys()))], [attr for attr in trainset.attr2idx.keys() if attr!=\"Attractive\"],rotation=90)\r\n",
        "\r\n",
        "plt.title(\"Percentage of False Positive regarding attributes\")\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ha9B17XXOD0K"
      },
      "source": [
        "**2. Proportions of attributes in the train and predicted attractive populatioins**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NTa_35Gf81jA",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 337
        },
        "outputId": "348383c6-ad06-4759-d9cc-3c8a26477c11"
      },
      "source": [
        "# Set up a grid of plots\r\n",
        "fig = plt.figure(figsize=(10,5)) \r\n",
        "\r\n",
        "interesting_features = [\"Bald\",\"Young\",\"Heavy_Makeup\",\"Male\",\"Pale_Skin\",\"Smiling\"]\r\n",
        "\r\n",
        "# Plot accidents depending on type\r\n",
        "#for attr in trainset.attr2idx.keys():\r\n",
        "i=1\r\n",
        "barWidth = 0.25\r\n",
        "\r\n",
        "for attr in interesting_features:\r\n",
        "  if attr!=\"Attractive\":\r\n",
        "   #plt.bar([attr+\"A\",attr],[len(train_df_with_data[(train_df_with_data[\"Attractive\"]==True) & (train_df_with_data[attr]==True)])/len(train_df_with_data[train_df_with_data[\"Attractive\"]==True])*100,len(train_df_with_data[train_df_with_data[attr]==True])/len(train_df_with_data)*100])\r\n",
        "    plt.bar(i,len(test_df_with_data[(test_df_with_data[\"Attractive\"]==True) & (test_df_with_data[attr]==True)])/len(test_df_with_data[test_df_with_data[\"Attractive\"]==True])*100,color='#B22222',width=barWidth,label=\"Test Attractive Population\" if i==1 else \"\")\r\n",
        "    plt.bar(i+barWidth,len(test_df_with_data[(test_df_with_data[attr]==True) & (test_df_with_data[\"status\"].isin([\"True Positive\",\"False Negative\"]))])/len(test_df_with_data[test_df_with_data[\"status\"].isin([\"True Positive\",\"False Negative\"])])*100,color='#F08080',width=barWidth,label=\"Test Predicted Attractive Population\" if i==1 else \"\")\r\n",
        "    plt.bar(i+barWidth*2,len(test_df_with_data[test_df_with_data[attr]==True])/len(test_df_with_data)*100,color='#8B0000',width=barWidth,label=\"Test' Population\" if i==1 else \"\")\r\n",
        "    i+=1\r\n",
        "\r\n",
        "plt.xticks([r + barWidth for r in range(1,len(interesting_features)+1)], [attr for attr in interesting_features])\r\n",
        "\r\n",
        "plt.title(\"Proportion of selected attributes inside different population\")\r\n",
        "plt.legend(prop={'size': 12})\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Wt4M7D31bCKj"
      },
      "source": [
        "## Comparison between with and without added data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LJ8ENvQVbHtw"
      },
      "source": [
        "### Proportion of attributes\r\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "25bwPM_JbFnE",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 431
        },
        "outputId": "0194b6bd-0423-4152-b111-ccc9e9309688"
      },
      "source": [
        "# Set up a grid of plots\r\n",
        "fig = plt.figure(figsize=(10,5)) \r\n",
        "\r\n",
        "# Plot accidents depending on type\r\n",
        "x_attr = list(trainset.attr2idx.keys())\r\n",
        "x_attr.remove(\"Attractive\")\r\n",
        "\r\n",
        "i=1\r\n",
        "barWidth = 0.25\r\n",
        "\r\n",
        "for attr in x_attr:\r\n",
        "    y_percentage_without_data = len(train_df_without_data[train_df_without_data[attr]==True])/len(train_df_without_data)\r\n",
        "    y_percentage_with_data = len(train_df_with_data[train_df_with_data[attr]==True])/len(train_df_with_data)\r\n",
        "\r\n",
        "    plt.bar(i,y_percentage_without_data,color='#B22222',width=barWidth,label=\"w/o adding data\" if i==1 else \"\")\r\n",
        "    plt.bar(i+barWidth,y_percentage_with_data,color='#F08080',width=barWidth,label=\"w adding data\" if i==1 else \"\")\r\n",
        "    i+=1\r\n",
        "\r\n",
        "barWidth = 0.25\r\n",
        "\r\n",
        "plt.xticks([r + barWidth/2 for r in range(1,len(x_attr)+1)], [attr for attr in x_attr],rotation=90)\r\n",
        "plt.legend(prop={'size': 12})\r\n",
        "plt.title(\"Proportion of selected attributes inside different population\")\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rmvPBwkPyMu5",
        "outputId": "3ea80916-e58d-4dc8-c67c-5eb98aa60676"
      },
      "source": [
        "# Train with data\n",
        "test_best_attr_with_data = {attr:len(test_df_with_data[(test_df_with_data[\"Attractive\"]==True) & (test_df_with_data[attr]==True)])/len(test_df_with_data[test_df_with_data[\"Attractive\"]==True])*100 if attr!=\"Attractive\" else 0 for attr in trainset.attr2idx.keys()}\n",
        "top_5_best_attr_test_w_d = sorted(test_best_attr_with_data.items(),key=lambda x: x[1],reverse=True)[:5]\n",
        "top_5_worst_attr_test_w_d = sorted(test_best_attr_with_data.items(),key=lambda x: x[1])[1:6]\n",
        "\n",
        "# predict with data\n",
        "predict_best_attr_w_d = {attr:len(test_df_with_data[(test_df_with_data[\"status\"].isin([\"False Positive\",\"True Positive\"])) & (test_df_with_data[attr]==True)])/len(test_df_with_data[test_df_with_data[\"status\"].isin([\"False Positive\",\"True Positive\"])])*100 if attr!=\"Attractive\" else 0 for attr in trainset.attr2idx.keys()}\n",
        "top_5_best_attr_predict_w_d = [predict_best_attr_w_d[tu[0]] for tu in top_5_best_attr_test_w_d]\n",
        "top_5_worst_attr_predict_w_d = [predict_best_attr_w_d[tu[0]] for tu in top_5_worst_attr_test_w_d]\n",
        "\n",
        "#Train without data added\n",
        "test_best_attr_without_data = {attr:len(test_df_without_data[(test_df_without_data[\"Attractive\"]==True) & (test_df_without_data[attr]==True)])/len(test_df_without_data[test_df_without_data[\"Attractive\"]==True])*100 if attr!=\"Attractive\" else 0 for attr in trainset.attr2idx.keys()}\n",
        "top_5_best_attr_test_without_d = sorted(test_best_attr_without_data.items(),key=lambda x: x[1],reverse=True)[:5]\n",
        "top_5_worst_attr_test_without_d = sorted(test_best_attr_without_data.items(),key=lambda x: x[1])[1:6]\n",
        "\n",
        "#Predict without data added\n",
        "predict_best_attr_without_d = {attr:len(test_df_without_data[(test_df_without_data[\"status\"].isin([\"False Positive\",\"True Positive\"])) & (test_df_without_data[attr]==True)])/len(test_df_without_data[test_df_without_data[\"status\"].isin([\"False Positive\",\"True Positive\"])])*100 if attr!=\"Attractive\" else 0 for attr in trainset.attr2idx.keys()}\n",
        "top_5_best_attr_predict_without_d = [predict_best_attr_without_d[tu[0]] for tu in top_5_best_attr_test_without_d]\n",
        "top_5_worst_attr_predict_without_d = [predict_best_attr_without_d[tu[0]] for tu in top_5_worst_attr_test_without_d]\n",
        "\n",
        "print(\"---------- Statistics in Test Dataset vs Network output WITH DATA ----------\")\n",
        "print(\"\\n\")\n",
        "print(\"Top Attributes to have\")\n",
        "print(\"--------------\")\n",
        "print(\"\\n\".join(map(lambda x,y: x[0]+\" = \"+str(round(x[1],2))+\"% vs \"+str(round(y,2))+\"%\",top_5_best_attr_test_w_d,top_5_best_attr_predict_w_d)))\n",
        "print(\"\\n\")\n",
        "print(\"Worst Attributes to have\")\n",
        "print(\"--------------\")\n",
        "print(\"\\n\".join(map(lambda x,y: x[0]+\" = \"+str(round(x[1],2))+\"% vs \"+str(round(y,2))+\"%\",top_5_worst_attr_test_w_d,top_5_worst_attr_predict_w_d)))\n",
        "print(\"\\n\")\n",
        "print(\"---------- Statistics in Test Dataset vs Network output WITHOUT DATA ADDED----------\")\n",
        "print(\"\\n\")\n",
        "print(\"Top Attributes to have\")\n",
        "print(\"--------------\")\n",
        "print(\"\\n\".join(map(lambda x,y: x[0]+\" = \"+str(round(x[1],2))+\"% vs \"+str(round(y,2))+\"%\",top_5_best_attr_test_without_d,top_5_best_attr_predict_without_d)))\n",
        "print(\"\\n\")\n",
        "print(\"Worst Attributes to have\")\n",
        "print(\"--------------\")\n",
        "print(\"\\n\".join(map(lambda x,y: x[0]+\" = \"+str(round(x[1],2))+\"% vs \"+str(round(y,2))+\"%\",top_5_worst_attr_test_without_d,top_5_worst_attr_predict_without_d)))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "---------- Statistics in Test Dataset vs Network output WITH DATA ----------\n",
            "\n",
            "\n",
            "Top Attributes to have\n",
            "--------------\n",
            "Young = 92.1% vs 91.73%\n",
            "No_Beard = 92.06% vs 93.75%\n",
            "Wearing_Lipstick = 76.1% vs 78.88%\n",
            "Heavy_Makeup = 63.6% vs 65.65%\n",
            "Smiling = 56.49% vs 57.19%\n",
            "\n",
            "\n",
            "Worst Attributes to have\n",
            "--------------\n",
            "Bald = 0.15% vs 0.03%\n",
            "Gray_Hair = 0.18% vs 0.11%\n",
            "Double_Chin = 0.26% vs 0.51%\n",
            "Chubby = 0.31% vs 0.56%\n",
            "Eyeglasses = 1.04% vs 0.58%\n",
            "\n",
            "\n",
            "---------- Statistics in Test Dataset vs Network output WITHOUT DATA ADDED----------\n",
            "\n",
            "\n",
            "Top Attributes to have\n",
            "--------------\n",
            "Young = 92.1% vs 91.77%\n",
            "No_Beard = 92.06% vs 93.14%\n",
            "Wearing_Lipstick = 76.1% vs 78.06%\n",
            "Heavy_Makeup = 63.6% vs 64.88%\n",
            "Smiling = 56.49% vs 56.75%\n",
            "\n",
            "\n",
            "Worst Attributes to have\n",
            "--------------\n",
            "Bald = 0.15% vs 0.03%\n",
            "Gray_Hair = 0.18% vs 0.14%\n",
            "Double_Chin = 0.26% vs 0.42%\n",
            "Chubby = 0.31% vs 0.53%\n",
            "Eyeglasses = 1.04% vs 0.6%\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}