Klamer_U_Net.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Klamer_U_Net.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "C78Gd98_sR4p",
        "dq_8D3kY7Ueo",
        "bmviYCbq7QbJ",
        "w6POq8z_-aSd"
      ],
      "toc_visible": true,
      "machine_shape": "hm",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/KlamJ/DeepLab_V3_KL/blob/master/Klamer_U_Net.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eZ-vxFBI_msg",
        "colab_type": "text"
      },
      "source": [
        "![Bild](https://drive.google.com/uc?id=1cWkcdnYZkzyS7pm4Tz_u8x_sMwv2PA_v)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7uwQDXN_yiEb",
        "colab_type": "text"
      },
      "source": [
        "# Cityscapes Dataset\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3MQ5IWWiqbtn",
        "colab_type": "text"
      },
      "source": [
        "The Cityscapes Dataset focuses on semantic understanding of urban street scenes. In the following, we give an overview on the design choices that were made to target the dataset’s focus.\n",
        "\n",
        "The following Code is partly from https://github.com/fregu856/deeplabv3.\n",
        "\n",
        "\n",
        "\n",
        "Dataset Link: https://www.cityscapes-dataset.com/\n",
        "\n",
        "Needed Data:\n",
        " - gtFine_trainvaltest.zip (241MB) [md5]\n",
        " - leftImg8bit_trainvaltest.zip (11GB) [md5] \n",
        "\n",
        "The Data should be in the same direktory. \n",
        "\n",
        "To use this notebook you need to Mount your own GoogleDrive with the same directory path. \n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aHRbT_1h6pLy",
        "colab_type": "text"
      },
      "source": [
        "## What is the Citiscapes Dataset\n",
        "\n",
        " - 5 000 annotated images with fine annotations\n",
        " - 30 classes\n",
        " - 1024x2048\n",
        " - Dataset size 11GB + 241MB\n",
        " - Semantic, Instance-wise, Dense pixel annotations\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-PnWRxO4sHWu",
        "colab_type": "text"
      },
      "source": [
        "###Mount GoogleDrive \n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Icq80SfN7rYR",
        "colab_type": "code",
        "outputId": "c15913bb-cd86-4e8e-fe0b-7d37dd181eb4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 124
        }
      },
      "source": [
        "## Mount Drive\n",
        "from google.colab import drive\n",
        "drive.mount('/content/drive')\n",
        "\n",
        "#https://github.com/KlamJ/deeplabv3\n"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "C78Gd98_sR4p",
        "colab_type": "text"
      },
      "source": [
        "###Check Gpu status"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "feyLL9_FEAdP",
        "colab_type": "code",
        "outputId": "4eb27c95-c8cc-4ab7-c1d6-79ddf20c91fb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 80
        }
      },
      "source": [
        "#' ' means CPU whereas '/device:G:0' means GPU\n",
        "import tensorflow as tf\n",
        "tf.test.gpu_device_name()\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<p style=\"color: red;\">\n",
              "The default version of TensorFlow in Colab will soon switch to TensorFlow 2.x.<br>\n",
              "We recommend you <a href=\"https://www.tensorflow.org/guide/migrate\" target=\"_blank\">upgrade</a> now \n",
              "or ensure your notebook will continue to use TensorFlow 1.x via the <code>%tensorflow_version 1.x</code> magic:\n",
              "<a href=\"https://colab.research.google.com/notebooks/tensorflow_version.ipynb\" target=\"_blank\">more info</a>.</p>\n"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'/device:GPU:0'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 1
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4rsm8T6EEEvK",
        "colab_type": "code",
        "outputId": "97d9ebc3-e678-4b35-af42-4170683906f0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 102
        }
      },
      "source": [
        "# memory footprint support libraries/code\n",
        "!ln -sf /opt/bin/nvidia-smi /usr/bin/nvidia-smi\n",
        "!pip install gputil\n",
        "!pip install psutil\n",
        "!pip install humanize\n",
        "import psutil\n",
        "import humanize\n",
        "import os\n",
        "import GPUtil as GPU\n",
        "GPUs = GPU.getGPUs()\n",
        "# XXX: only one GPU on Colab and isn’t guaranteed\n",
        "gpu = GPUs[0]\n",
        "def printm():\n",
        " process = psutil.Process(os.getpid())\n",
        " print(\"Gen RAM Free: \" + humanize.naturalsize( psutil.virtual_memory().available ), \" | Proc size: \" + humanize.naturalsize( process.memory_info().rss))\n",
        " print(\"GPU RAM Free: {0:.0f}MB | Used: {1:.0f}MB | Util {2:3.0f}% | Total {3:.0f}MB\".format(gpu.memoryFree, gpu.memoryUsed, gpu.memoryUtil*100, gpu.memoryTotal))\n",
        "printm() "
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: gputil in /usr/local/lib/python3.6/dist-packages (1.4.0)\n",
            "Requirement already satisfied: psutil in /usr/local/lib/python3.6/dist-packages (5.4.8)\n",
            "Requirement already satisfied: humanize in /usr/local/lib/python3.6/dist-packages (0.5.1)\n",
            "Gen RAM Free: 24.6 GB  | Proc size: 3.3 GB\n",
            "GPU RAM Free: 13MB | Used: 16267MB | Util 100% | Total 16280MB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HrQDJ7K8htKI",
        "colab_type": "code",
        "outputId": "ea2379b1-3b2b-4b1e-df5c-b1f0b8ce344b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 306
        }
      },
      "source": [
        "!nvidia-smi"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Sat Dec 21 13:02:22 2019       \n",
            "+-----------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 440.44       Driver Version: 418.67       CUDA Version: 10.1     |\n",
            "|-------------------------------+----------------------+----------------------+\n",
            "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
            "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
            "|===============================+======================+======================|\n",
            "|   0  Tesla P100-PCIE...  Off  | 00000000:00:04.0 Off |                    0 |\n",
            "| N/A   46C    P0    27W / 250W |      0MiB / 16280MiB |      0%      Default |\n",
            "+-------------------------------+----------------------+----------------------+\n",
            "                                                                               \n",
            "+-----------------------------------------------------------------------------+\n",
            "| Processes:                                                       GPU Memory |\n",
            "|  GPU       PID   Type   Process name                             Usage      |\n",
            "|=============================================================================|\n",
            "|  No running processes found                                                 |\n",
            "+-----------------------------------------------------------------------------+\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cqW7y02Yh3hz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aI97pt3KEFh_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#if not the only user and the gpu Util is >0 do following \n",
        "!kill -9 -1"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "h9mzp7ni0Cdp",
        "colab_type": "text"
      },
      "source": [
        "#DeepLab"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0d_Gwc6F2t3x",
        "colab_type": "text"
      },
      "source": [
        "##The DeepLab V3 \n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9H992mJv6K9I",
        "colab_type": "text"
      },
      "source": [
        "Import the required libraries and set the most used paths"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dcPu5-b0iIrx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "%matplotlib inline\n",
        "import sys\n",
        "import cv2\n",
        "datasetPath = \"/content/drive/My Drive/U_Net_Datasets\"\n",
        "cityscapes_data_path=\"/content/drive/My Drive/U_Net_Datasets\"\n",
        "cityscapes_meta_path=\"/content/drive/My Drive/U_Net_Datasets/meta\"\n",
        "\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VBecHJEQK2Fh",
        "colab_type": "code",
        "outputId": "4a1e348f-293a-49c0-b9d2-06b974c5d788",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 159
        }
      },
      "source": [
        "pip install pytorch-ignite"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting pytorch-ignite\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/8f/31/efcc2b587419b1f54c5c6ef51996f91bb5d8f760537d17de674c89e06048/pytorch_ignite-0.2.1-py2.py3-none-any.whl (84kB)\n",
            "\r\u001b[K     |███▉                            | 10kB 30.6MB/s eta 0:00:01\r\u001b[K     |███████▊                        | 20kB 1.7MB/s eta 0:00:01\r\u001b[K     |███████████▋                    | 30kB 2.5MB/s eta 0:00:01\r\u001b[K     |███████████████▌                | 40kB 1.7MB/s eta 0:00:01\r\u001b[K     |███████████████████▍            | 51kB 2.1MB/s eta 0:00:01\r\u001b[K     |███████████████████████▎        | 61kB 2.5MB/s eta 0:00:01\r\u001b[K     |███████████████████████████▏    | 71kB 2.9MB/s eta 0:00:01\r\u001b[K     |███████████████████████████████ | 81kB 3.3MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 92kB 3.0MB/s \n",
            "\u001b[?25hRequirement already satisfied: torch in /usr/local/lib/python3.6/dist-packages (from pytorch-ignite) (1.3.1)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from torch->pytorch-ignite) (1.17.4)\n",
            "Installing collected packages: pytorch-ignite\n",
            "Successfully installed pytorch-ignite-0.2.1\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wtNUqHto5R65",
        "colab_type": "text"
      },
      "source": [
        "####Utils"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dq_8D3kY7Ueo",
        "colab_type": "text"
      },
      "source": [
        "#####1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S5WnI-Qz6OQK",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "\n",
        "import os\n",
        "import pickle\n",
        "from collections import namedtuple\n",
        "import torch\n",
        "import torch.utils.data\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "import torchvision.models as models\n",
        "\n",
        "from torch.autograd import Variable\n",
        "import torch.optim as optim\n",
        "import matplotlib\n",
        "matplotlib.use(\"Agg\")\n",
        "import matplotlib.pyplot as plt\n",
        "import time\n",
        "import ignite as ig\n",
        "\n",
        "train_dirs = [\"jena/\", \"zurich/\", \"weimar/\", \"ulm/\", \"tubingen/\", \"stuttgart/\",\n",
        "              \"strasbourg/\", \"monchengladbach/\", \"krefeld/\", \"hanover/\",\n",
        "             \"hamburg/\", \"erfurt/\", \"dusseldorf/\", \"darmstadt/\", \"cologne/\",\n",
        "             \"bremen/\", \"bochum/\", \"aachen/\"]\n",
        "val_dirs = [\"frankfurt/\", \"munster/\", \"lindau/\"]\n",
        "test_dirs = [\"berlin\", \"bielefeld\", \"bonn\", \"leverkusen\", \"mainz\", \"munich\"]\n",
        "\n",
        "class DatasetTrain(torch.utils.data.Dataset):\n",
        "    def __init__(self, cityscapes_data_path, cityscapes_meta_path):\n",
        "        self.img_dir = cityscapes_data_path + \"/leftImg8bit/train/\"\n",
        "        self.label_dir = cityscapes_meta_path + \"/label_imgs/\"\n",
        "\n",
        "        self.img_h = 1024\n",
        "        self.img_w = 2048\n",
        "\n",
        "        self.new_img_h = 512\n",
        "        self.new_img_w = 1024\n",
        "\n",
        "        self.examples = []\n",
        "        for train_dir in train_dirs:\n",
        "            train_img_dir_path = self.img_dir + train_dir\n",
        "\n",
        "            file_names = os.listdir(train_img_dir_path)\n",
        "            for file_name in file_names:\n",
        "                img_id = file_name.split(\"_leftImg8bit.png\")[0]\n",
        "\n",
        "                img_path = train_img_dir_path + file_name\n",
        "\n",
        "                label_img_path = self.label_dir + img_id + \".png\"\n",
        "\n",
        "                example = {}\n",
        "                example[\"img_path\"] = img_path\n",
        "                example[\"label_img_path\"] = label_img_path\n",
        "                example[\"img_id\"] = img_id\n",
        "                self.examples.append(example)\n",
        "\n",
        "        self.num_examples = len(self.examples)\n",
        "\n",
        "    def __getitem__(self, index):\n",
        "        example = self.examples[index]\n",
        "\n",
        "        img_path = example[\"img_path\"]\n",
        "        img = cv2.imread(img_path, -1) # (shape: (1024, 2048, 3))\n",
        "        # resize img without interpolation (want the image to still match\n",
        "        # label_img, which we resize below):\n",
        "        img = cv2.resize(img, (self.new_img_w, self.new_img_h),\n",
        "                         interpolation=cv2.INTER_NEAREST) # (shape: (512, 1024, 3))\n",
        "\n",
        "        label_img_path = example[\"label_img_path\"]\n",
        "        label_img = cv2.imread(label_img_path, -1) # (shape: (1024, 2048))\n",
        "        # resize label_img without interpolation (want the resulting image to\n",
        "        # still only contain pixel values corresponding to an object class):\n",
        "        label_img = cv2.resize(label_img, (self.new_img_w, self.new_img_h),\n",
        "                               interpolation=cv2.INTER_NEAREST) # (shape: (512, 1024))\n",
        "\n",
        "        # flip the img and the label with 0.5 probability:\n",
        "        flip = np.random.randint(low=0, high=2)\n",
        "        if flip == 1:\n",
        "            img = cv2.flip(img, 1)\n",
        "            label_img = cv2.flip(label_img, 1)\n",
        "\n",
        "        ########################################################################\n",
        "        # randomly scale the img and the label:\n",
        "        ########################################################################\n",
        "        scale = np.random.uniform(low=0.7, high=2.0)\n",
        "        new_img_h = int(scale*self.new_img_h)\n",
        "        new_img_w = int(scale*self.new_img_w)\n",
        "\n",
        "        # resize img without interpolation (want the image to still match\n",
        "        # label_img, which we resize below):\n",
        "        img = cv2.resize(img, (new_img_w, new_img_h),\n",
        "                         interpolation=cv2.INTER_NEAREST) # (shape: (new_img_h, new_img_w, 3))\n",
        "\n",
        "        # resize label_img without interpolation (want the resulting image to\n",
        "        # still only contain pixel values corresponding to an object class):\n",
        "        label_img = cv2.resize(label_img, (new_img_w, new_img_h),\n",
        "                               interpolation=cv2.INTER_NEAREST) # (shape: (new_img_h, new_img_w))\n",
        "        ########################################################################\n",
        "\n",
        "         # # # # # # # debug visualization START\n",
        "        #print (scale)\n",
        "        #print (new_img_h)\n",
        "        #print (new_img_w)\n",
        "        \n",
        "        #cv2.imshow(\"test\", img)\n",
        "        #cv2.waitKey(0)\n",
        "        \n",
        "        #cv2.imshow(\"test\", label_img)\n",
        "        #cv2.waitKey(0)\n",
        "         # # # # # # # debug visualization END\n",
        "\n",
        "        ########################################################################\n",
        "        # select a 256x256 random crop from the img and label:\n",
        "        ########################################################################\n",
        "        start_x = np.random.randint(low=0, high=(new_img_w - 256))\n",
        "        end_x = start_x + 256\n",
        "        start_y = np.random.randint(low=0, high=(new_img_h - 256))\n",
        "        end_y = start_y + 256\n",
        "\n",
        "        img = img[start_y:end_y, start_x:end_x] # (shape: (256, 256, 3))\n",
        "        label_img = label_img[start_y:end_y, start_x:end_x] # (shape: (256, 256))\n",
        "        ########################################################################\n",
        "\n",
        "        # # # # # # # # debug visualization START\n",
        "        # print (img.shape)\n",
        "        # print (label_img.shape)\n",
        "        #\n",
        "        # cv2.imshow(\"test\", img)\n",
        "        # cv2.waitKey(0)\n",
        "        #\n",
        "        # cv2.imshow(\"test\", label_img)\n",
        "        # cv2.waitKey(0)\n",
        "        # # # # # # # # debug visualization END\n",
        "\n",
        "        # normalize the img (with the mean and std for the pretrained ResNet):\n",
        "        img = img/255.0\n",
        "        img = img - np.array([0.485, 0.456, 0.406])\n",
        "        img = img/np.array([0.229, 0.224, 0.225]) # (shape: (256, 256, 3))\n",
        "        img = np.transpose(img, (2, 0, 1)) # (shape: (3, 256, 256))\n",
        "        img = img.astype(np.float32)\n",
        "\n",
        "        # convert numpy -> torch:\n",
        "        img = torch.from_numpy(img) # (shape: (3, 256, 256))\n",
        "        label_img = torch.from_numpy(label_img) # (shape: (256, 256))\n",
        "\n",
        "        return (img, label_img)\n",
        "\n",
        "    def __len__(self):\n",
        "        return self.num_examples\n",
        "\n",
        "class DatasetVal(torch.utils.data.Dataset):\n",
        "    def __init__(self, cityscapes_data_path, cityscapes_meta_path):\n",
        "        self.img_dir = cityscapes_data_path + \"/leftImg8bit/val/\"\n",
        "        self.label_dir = cityscapes_meta_path + \"/label_imgs/\"\n",
        "\n",
        "        self.img_h = 1024\n",
        "        self.img_w = 2048\n",
        "\n",
        "        self.new_img_h = 512\n",
        "        self.new_img_w = 1024\n",
        "\n",
        "        self.examples = []\n",
        "        for val_dir in val_dirs:\n",
        "            val_img_dir_path = self.img_dir + val_dir\n",
        "\n",
        "            file_names = os.listdir(val_img_dir_path)\n",
        "            for file_name in file_names:\n",
        "                img_id = file_name.split(\"_leftImg8bit.png\")[0]\n",
        "\n",
        "                img_path = val_img_dir_path + file_name\n",
        "\n",
        "                label_img_path = self.label_dir + img_id + \".png\"\n",
        "                label_img = cv2.imread(label_img_path, -1) # (shape: (1024, 2048))\n",
        "\n",
        "                example = {}\n",
        "                example[\"img_path\"] = img_path\n",
        "                example[\"label_img_path\"] = label_img_path\n",
        "                example[\"img_id\"] = img_id\n",
        "                self.examples.append(example)\n",
        "\n",
        "        self.num_examples = len(self.examples)\n",
        "\n",
        "    def __getitem__(self, index):\n",
        "        example = self.examples[index]\n",
        "\n",
        "        img_id = example[\"img_id\"]\n",
        "\n",
        "        img_path = example[\"img_path\"]\n",
        "        img = cv2.imread(img_path, -1) # (shape: (1024, 2048, 3))\n",
        "        # resize img without interpolation (want the image to still match\n",
        "        # label_img, which we resize below):\n",
        "        img = cv2.resize(img, (self.new_img_w, self.new_img_h),\n",
        "                         interpolation=cv2.INTER_NEAREST) # (shape: (512, 1024, 3))\n",
        "\n",
        "        label_img_path = example[\"label_img_path\"]\n",
        "        label_img = cv2.imread(label_img_path, -1) # (shape: (1024, 2048))\n",
        "        # resize label_img without interpolation (want the resulting image to\n",
        "        # still only contain pixel values corresponding to an object class):\n",
        "        label_img = cv2.resize(label_img, (self.new_img_w, self.new_img_h),\n",
        "                               interpolation=cv2.INTER_NEAREST) # (shape: (512, 1024))\n",
        "\n",
        "        # # # # # # # # debug visualization START\n",
        "        # cv2.imshow(\"test\", img)\n",
        "        # cv2.waitKey(0)\n",
        "        #\n",
        "        # cv2.imshow(\"test\", label_img)\n",
        "        # cv2.waitKey(0)\n",
        "        # # # # # # # # debug visualization END\n",
        "\n",
        "        # normalize the img (with the mean and std for the pretrained ResNet):\n",
        "        img = img/255.0\n",
        "        img = img - np.array([0.485, 0.456, 0.406])\n",
        "        img = img/np.array([0.229, 0.224, 0.225]) # (shape: (512, 1024, 3))\n",
        "        img = np.transpose(img, (2, 0, 1)) # (shape: (3, 512, 1024))\n",
        "        img = img.astype(np.float32)\n",
        "\n",
        "        # convert numpy -> torch:\n",
        "        img = torch.from_numpy(img) # (shape: (3, 512, 1024))\n",
        "        label_img = torch.from_numpy(label_img) # (shape: (512, 1024))\n",
        "\n",
        "        return (img, label_img, img_id)\n",
        "\n",
        "    def __len__(self):\n",
        "        return self.num_examples\n",
        "\n",
        "class DatasetSeq(torch.utils.data.Dataset):\n",
        "    def __init__(self, cityscapes_data_path, cityscapes_meta_path, sequence):\n",
        "        self.img_dir = cityscapes_data_path + \"/leftImg8bit/demoVideo/stuttgart_\" + sequence + \"/\"\n",
        "\n",
        "        self.img_h = 1024\n",
        "        self.img_w = 2048\n",
        "\n",
        "        self.new_img_h = 512\n",
        "        self.new_img_w = 1024\n",
        "\n",
        "        self.examples = []\n",
        "\n",
        "        file_names = os.listdir(self.img_dir)\n",
        "        for file_name in file_names:\n",
        "            img_id = file_name.split(\"_leftImg8bit.png\")[0]\n",
        "\n",
        "            img_path = self.img_dir + file_name\n",
        "\n",
        "            example = {}\n",
        "            example[\"img_path\"] = img_path\n",
        "            example[\"img_id\"] = img_id\n",
        "            self.examples.append(example)\n",
        "\n",
        "        self.num_examples = len(self.examples)\n",
        "\n",
        "    def __getitem__(self, index):\n",
        "        example = self.examples[index]\n",
        "\n",
        "        img_id = example[\"img_id\"]\n",
        "\n",
        "        img_path = example[\"img_path\"]\n",
        "        img = cv2.imread(img_path, -1) # (shape: (1024, 2048, 3))\n",
        "        # resize img without interpolation:\n",
        "        img = cv2.resize(img, (self.new_img_w, self.new_img_h),\n",
        "                         interpolation=cv2.INTER_NEAREST) # (shape: (512, 1024, 3))\n",
        "\n",
        "        # normalize the img (with the mean and std for the pretrained ResNet):\n",
        "        img = img/255.0\n",
        "        img = img - np.array([0.485, 0.456, 0.406])\n",
        "        img = img/np.array([0.229, 0.224, 0.225]) # (shape: (512, 1024, 3))\n",
        "        img = np.transpose(img, (2, 0, 1)) # (shape: (3, 512, 1024))\n",
        "        img = img.astype(np.float32)\n",
        "\n",
        "        # convert numpy -> torch:\n",
        "        img = torch.from_numpy(img) # (shape: (3, 512, 1024))\n",
        "\n",
        "        return (img, img_id)\n",
        "\n",
        "    def __len__(self):\n",
        "        return self.num_examples\n",
        "\n",
        "class DatasetThnSeq(torch.utils.data.Dataset):\n",
        "    def __init__(self, thn_data_path):\n",
        "        self.img_dir = thn_data_path + \"/\"\n",
        "\n",
        "        self.examples = []\n",
        "\n",
        "        file_names = os.listdir(self.img_dir)\n",
        "        for file_name in file_names:\n",
        "            img_id = file_name.split(\".png\")[0]\n",
        "\n",
        "            img_path = self.img_dir + file_name\n",
        "\n",
        "            example = {}\n",
        "            example[\"img_path\"] = img_path\n",
        "            example[\"img_id\"] = img_id\n",
        "            self.examples.append(example)\n",
        "\n",
        "        self.num_examples = len(self.examples)\n",
        "\n",
        "    def __getitem__(self, index):\n",
        "        example = self.examples[index]\n",
        "\n",
        "        img_id = example[\"img_id\"]\n",
        "\n",
        "        img_path = example[\"img_path\"]\n",
        "        img = cv2.imread(img_path, -1) # (shape: (512, 1024, 3))\n",
        "\n",
        "        # normalize the img (with mean and std for the pretrained ResNet):\n",
        "        img = img/255.0\n",
        "        img = img - np.array([0.485, 0.456, 0.406])\n",
        "        img = img/np.array([0.229, 0.224, 0.225]) # (shape: (512, 1024, 3))\n",
        "        img = np.transpose(img, (2, 0, 1)) # (shape: (3, 512, 1024))\n",
        "        img = img.astype(np.float32)\n",
        "\n",
        "        # convert numpy -> torch:\n",
        "        img = torch.from_numpy(img) # (shape: (3, 512, 1024))\n",
        "\n",
        "        return (img, img_id)\n",
        "\n",
        "    def __len__(self):\n",
        "        return self.num_examples\n",
        "\n",
        "        \n",
        "\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bmviYCbq7QbJ",
        "colab_type": "text"
      },
      "source": [
        "#####2\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5l14z8QR5Ovn",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "\n",
        "def add_weight_decay(net, l2_value, skip_list=()):\n",
        "    # https://raberrytv.wordpress.com/2017/10/29/pytorch-weight-decay-made-easy/\n",
        "\n",
        "    decay, no_decay = [], []\n",
        "    for name, param in net.named_parameters():\n",
        "        if not param.requires_grad:\n",
        "            continue # frozen weights\n",
        "        if len(param.shape) == 1 or name.endswith(\".bias\") or name in skip_list:\n",
        "            no_decay.append(param)\n",
        "        else:\n",
        "            decay.append(param)\n",
        "\n",
        "    return [{'params': no_decay, 'weight_decay': 0.0}, {'params': decay, 'weight_decay': l2_value}]\n",
        "\n",
        "# function for colorizing a label image:\n",
        "def label_img_to_color(img):\n",
        "    label_to_color = {\n",
        "        0: [128, 64,128],\n",
        "        1: [244, 35,232],\n",
        "        2: [ 70, 70, 70],\n",
        "        3: [102,102,156],\n",
        "        4: [190,153,153],\n",
        "        5: [153,153,153],\n",
        "        6: [250,170, 30],\n",
        "        7: [220,220,  0],\n",
        "        8: [107,142, 35],\n",
        "        9: [152,251,152],\n",
        "        10: [ 70,130,180],\n",
        "        11: [220, 20, 60],\n",
        "        12: [255,  0,  0],\n",
        "        13: [  0,  0,142],\n",
        "        14: [  0,  0, 70],\n",
        "        15: [  0, 60,100],\n",
        "        16: [  0, 80,100],\n",
        "        17: [  0,  0,230],\n",
        "        18: [119, 11, 32],\n",
        "        19: [81,  0, 81]\n",
        "        }\n",
        "\n",
        "    img_height, img_width = img.shape\n",
        "\n",
        "    img_color = np.zeros((img_height, img_width, 3))\n",
        "    for row in range(img_height):\n",
        "        for col in range(img_width):\n",
        "            label = img[row, col]\n",
        "\n",
        "            img_color[row, col] = np.array(label_to_color[label])\n",
        "\n",
        "    return img_color\n",
        "\n",
        "#not used _____________________________________________________________________  \n",
        "def mIOU_funk(gT,pred):\n",
        "  #mean intersection over Union\n",
        "  mioU_array = []\n",
        "  label_imgs_numpy = gT\n",
        "  for dim in pred:  \n",
        "    for img in label_imgs_numpy:    \n",
        "      for obj in dim:       \n",
        "        intersection = np.logical_and(img, obj)\n",
        "        #print(intersection.shape)\n",
        "        union = np.logical_or(img, obj)\n",
        "        #print(union.shape)\n",
        "        iou_score = np.sum(intersection) / np.sum(union)\n",
        "        mioU_array.append(iou_score)\n",
        "      \n",
        "  mean = np.mean(mioU_array)\n",
        "  #print(\"mIOU is : %g\" %mean)\n",
        "  return mean\n",
        "\n",
        "def to_one_hot(tensor,nClasses):\n",
        "    \n",
        "    n,h,w = tensor.size()\n",
        "    one_hot = torch.zeros(n,nClasses,h,w).scatter_(1,tensor.view(n,1,h,w),1)\n",
        "    return one_hot\n",
        "\n",
        "class mIoULoss(nn.Module):\n",
        "    def __init__(self, weight=None, size_average=True, n_classes=2):\n",
        "        super(mIoULoss, self).__init__()\n",
        "        self.classes = n_classes\n",
        "\n",
        "    def forward(self, inputs, target_oneHot):\n",
        "    \t# inputs => N x Classes x H x W\n",
        "    \t# target_oneHot => N x Classes x H x W\n",
        "\n",
        "    \tN = inputs.size()[0]\n",
        "\n",
        "    \t# predicted probabilities for each pixel along channel\n",
        "    \tinputs = F.softmax(inputs,dim=1)\n",
        "    \t\n",
        "    \t# Numerator Product\n",
        "    \tinter = inputs * target_oneHot\n",
        "    \t## Sum over all pixels N x C x H x W => N x C\n",
        "    \tinter = inter.view(N,self.classes,-1).sum(2)\n",
        "\n",
        "    \t#Denominator \n",
        "    \tunion= inputs + target_oneHot - (inputs*target_oneHot)\n",
        "    \t## Sum over all pixels N x C x H x W => N x C\n",
        "    \tunion = union.view(N,self.classes,-1).sum(2)\n",
        "\n",
        "    \tloss = inter/union\n",
        "\n",
        "    \t## Return average loss over classes and batch\n",
        "    \treturn loss.mean()\n",
        "\n",
        "\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "raYB_3GT5LP1",
        "colab_type": "text"
      },
      "source": [
        "###Preprocessing"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fmNmdrnezYiX",
        "colab_type": "text"
      },
      "source": [
        "####Directory Tree\n",
        "\n",
        "To use the notebook, the data must be stored as shown in the following structure.\n",
        "\n",
        "\n",
        " "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "p_3Umiqs1yxE",
        "colab_type": "text"
      },
      "source": [
        "U_Net_Datasets/\n",
        "\n",
        "    gtFine/\n",
        "        val/\n",
        "            frankfurt/\n",
        "            munster/\n",
        "            lindau/\n",
        "        test/\n",
        "            berlin/\n",
        "            mainz/\n",
        "            bielefeld/\n",
        "            leverkusen/\n",
        "            bonn/\n",
        "            munich/\n",
        "        train/\n",
        "            bochum/\n",
        "            darmstadt/\n",
        "            bremen/\n",
        "            hamburg/\n",
        "            cologne/\n",
        "            erfurt/\n",
        "            dusseldorf/\n",
        "            hanover/\n",
        "            jena/\n",
        "            aachen/\n",
        "            monchengladbach/\n",
        "            strasbourg/\n",
        "            weimar/\n",
        "            krefeld/\n",
        "            stuttgart/\n",
        "            ulm/\n",
        "            zurich/\n",
        "            tubingen/\n",
        "    leftImg8bit/\n",
        "        train/\n",
        "            aachen/\n",
        "            bochum/\n",
        "            bremen/\n",
        "            cologne/\n",
        "            darmstadt/\n",
        "            dusseldorf/\n",
        "            strasbourg/\n",
        "            monchengladbach/\n",
        "            hanover/\n",
        "            krefeld/\n",
        "            stuttgart/\n",
        "            jena/\n",
        "            erfurt/\n",
        "            hamburg/\n",
        "            ulm/\n",
        "            tubingen/\n",
        "            zurich/\n",
        "            weimar/\n",
        "        test/\n",
        "            mainz/\n",
        "            bielefeld/\n",
        "            leverkusen/\n",
        "            berlin/\n",
        "            bonn/\n",
        "            munich/\n",
        "        val/\n",
        "            lindau/\n",
        "            frankfurt/\n",
        "            munster/\n",
        "        demoVideo/\n",
        "            .ipynb_checkpoints/\n",
        "            stuttgart_00/\n",
        "            stuttgart_01/\n",
        "            stuttgart_02/\n",
        "        .ipynb_checkpoints/\n",
        "    meta/\n",
        "        label_imgs/\n",
        "    pretrained_models/\n",
        "        resnet/"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5xbNRRwL1_C8",
        "colab_type": "text"
      },
      "source": [
        "To Check your Directory tree you can use the following code."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tq790jlzzdv-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import os\n",
        "\n",
        "def list_files(startpath):\n",
        "    for root, dirs, files in os.walk(startpath):\n",
        "        level = root.replace(startpath, '').count(os.sep)\n",
        "        indent = ' ' * 4 * (level)\n",
        "        print('{}{}/'.format(indent, os.path.basename(root)))\n",
        "        subindent = ' ' * 4 * (level + 1)\n",
        "        #for f in files:\n",
        "            #print('{}{}'.format(subindent, f))\n",
        "\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0QpeZojD0MLK",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "fbaef559-3fca-4fc8-9028-bba6d2ee1513"
      },
      "source": [
        "list_files(\"/content/drive/My Drive/U_Net_Datasets\")#set your root directory here"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "U_Net_Datasets/\n",
            "    gtFine/\n",
            "        val/\n",
            "            frankfurt/\n",
            "            munster/\n",
            "            lindau/\n",
            "        test/\n",
            "            berlin/\n",
            "            mainz/\n",
            "            bielefeld/\n",
            "            leverkusen/\n",
            "            bonn/\n",
            "            munich/\n",
            "        train/\n",
            "            bochum/\n",
            "            darmstadt/\n",
            "            bremen/\n",
            "            hamburg/\n",
            "            cologne/\n",
            "            erfurt/\n",
            "            dusseldorf/\n",
            "            hanover/\n",
            "            jena/\n",
            "            aachen/\n",
            "            monchengladbach/\n",
            "            strasbourg/\n",
            "            weimar/\n",
            "            krefeld/\n",
            "            stuttgart/\n",
            "            ulm/\n",
            "            zurich/\n",
            "            tubingen/\n",
            "    leftImg8bit/\n",
            "        train/\n",
            "            aachen/\n",
            "            bochum/\n",
            "            bremen/\n",
            "            cologne/\n",
            "            darmstadt/\n",
            "            dusseldorf/\n",
            "            strasbourg/\n",
            "            monchengladbach/\n",
            "            hanover/\n",
            "            krefeld/\n",
            "            stuttgart/\n",
            "            jena/\n",
            "            erfurt/\n",
            "            hamburg/\n",
            "            ulm/\n",
            "            tubingen/\n",
            "            zurich/\n",
            "            weimar/\n",
            "        test/\n",
            "            mainz/\n",
            "            bielefeld/\n",
            "            leverkusen/\n",
            "            berlin/\n",
            "            bonn/\n",
            "            munich/\n",
            "        val/\n",
            "            lindau/\n",
            "            frankfurt/\n",
            "            munster/\n",
            "        demoVideo/\n",
            "            .ipynb_checkpoints/\n",
            "            stuttgart_00/\n",
            "            stuttgart_01/\n",
            "            stuttgart_02/\n",
            "        .ipynb_checkpoints/\n",
            "    meta/\n",
            "        label_imgs/\n",
            "    pretrained_models/\n",
            "        resnet/\n",
            "    training_logs/\n",
            "        .ipynb_checkpoints/\n",
            "        model_eval_val/\n",
            "            checkpoints/\n",
            "        model_eval_val_for_metrics/\n",
            "            checkpoints/\n",
            "        model_eval_seq/\n",
            "            checkpoints/\n",
            "        model_eval_seq_thn/\n",
            "            checkpoints/\n",
            "        model_2/\n",
            "            checkpoints/\n",
            "        model_5/\n",
            "            checkpoints/\n",
            "        model_50_b3_w2/\n",
            "            checkpoints/\n",
            "        model_101_b3_w2/\n",
            "            checkpoints/\n",
            "        model_101_b3_w2_from_148_LR_001/\n",
            "            checkpoints/\n",
            "        model_101_b9_w1_from_148_LR_00001/\n",
            "            checkpoints/\n",
            "        model_101_b4_w1_from_148+22_LR_00001/\n",
            "            checkpoints/\n",
            "        model_18_B4_LR_00001_W2/\n",
            "            checkpoints/\n",
            "    vids/\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "w6POq8z_-aSd",
        "colab_type": "text"
      },
      "source": [
        "#### Unique execution to create labels and class weights\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P-AVpbPI5W_m",
        "colab_type": "code",
        "outputId": "8bd10a22-baff-4882-ddf9-3fd94c637f98",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 938
        }
      },
      "source": [
        "\n",
        "\n",
        "# (NOTE! this is taken from the official Cityscapes scripts:)\n",
        "Label = namedtuple( 'Label' , [\n",
        "\n",
        "    'name'        , # The identifier of this label, e.g. 'car', 'person', ... .\n",
        "                    # We use them to uniquely name a class\n",
        "\n",
        "    'id'          , # An integer ID that is associated with this label.\n",
        "                    # The IDs are used to represent the label in ground truth images\n",
        "                    # An ID of -1 means that this label does not have an ID and thus\n",
        "                    # is ignored when creating ground truth images (e.g. license plate).\n",
        "                    # Do not modify these IDs, since exactly these IDs are expected by the\n",
        "                    # evaluation server.\n",
        "\n",
        "    'trainId'     , # Feel free to modify these IDs as suitable for your method. Then create\n",
        "                    # ground truth images with train IDs, using the tools provided in the\n",
        "                    # 'preparation' folder. However, make sure to validate or submit results\n",
        "                    # to our evaluation server using the regular IDs above!\n",
        "                    # For trainIds, multiple labels might have the same ID. Then, these labels\n",
        "                    # are mapped to the same class in the ground truth images. For the inverse\n",
        "                    # mapping, we use the label that is defined first in the list below.\n",
        "                    # For example, mapping all void-type classes to the same ID in training,\n",
        "                    # might make sense for some approaches.\n",
        "                    # Max value is 255!\n",
        "\n",
        "    'category'    , # The name of the category that this label belongs to\n",
        "\n",
        "    'categoryId'  , # The ID of this category. Used to create ground truth images\n",
        "                    # on category level.\n",
        "\n",
        "    'hasInstances', # Whether this label distinguishes between single instances or not\n",
        "\n",
        "    'ignoreInEval', # Whether pixels having this class as ground truth label are ignored\n",
        "                    # during evaluations or not\n",
        "\n",
        "    'color'       , # The color of this label\n",
        "    ] )\n",
        "\n",
        "# (NOTE! this is taken from the official Cityscapes scripts:)\n",
        "labels = [\n",
        "    #       name                     id    trainId   category            catId     hasInstances   ignoreInEval   color\n",
        "    Label(  'unlabeled'            ,  0 ,      19 , 'void'            , 0       , False        , True         , (  0,  0,  0) ),\n",
        "    Label(  'ego vehicle'          ,  1 ,      19 , 'void'            , 0       , False        , True         , (  0,  0,  0) ),\n",
        "    Label(  'rectification border' ,  2 ,      19 , 'void'            , 0       , False        , True         , (  0,  0,  0) ),\n",
        "    Label(  'out of roi'           ,  3 ,      19 , 'void'            , 0       , False        , True         , (  0,  0,  0) ),\n",
        "    Label(  'static'               ,  4 ,      19 , 'void'            , 0       , False        , True         , (  0,  0,  0) ),\n",
        "    Label(  'dynamic'              ,  5 ,      19 , 'void'            , 0       , False        , True         , (111, 74,  0) ),\n",
        "    Label(  'ground'               ,  6 ,      19 , 'void'            , 0       , False        , True         , ( 81,  0, 81) ),\n",
        "    Label(  'road'                 ,  7 ,        0 , 'flat'            , 1       , False        , False        , (128, 64,128) ),\n",
        "    Label(  'sidewalk'             ,  8 ,        1 , 'flat'            , 1       , False        , False        , (244, 35,232) ),\n",
        "    Label(  'parking'              ,  9 ,      19 , 'flat'            , 1       , False        , True         , (250,170,160) ),\n",
        "    Label(  'rail track'           , 10 ,      19 , 'flat'            , 1       , False        , True         , (230,150,140) ),\n",
        "    Label(  'building'             , 11 ,        2 , 'construction'    , 2       , False        , False        , ( 70, 70, 70) ),\n",
        "    Label(  'wall'                 , 12 ,        3 , 'construction'    , 2       , False        , False        , (102,102,156) ),\n",
        "    Label(  'fence'                , 13 ,        4 , 'construction'    , 2       , False        , False        , (190,153,153) ),\n",
        "    Label(  'guard rail'           , 14 ,      19 , 'construction'    , 2       , False        , True         , (180,165,180) ),\n",
        "    Label(  'bridge'               , 15 ,      19 , 'construction'    , 2       , False        , True         , (150,100,100) ),\n",
        "    Label(  'tunnel'               , 16 ,      19 , 'construction'    , 2       , False        , True         , (150,120, 90) ),\n",
        "    Label(  'pole'                 , 17 ,        5 , 'object'          , 3       , False        , False        , (153,153,153) ),\n",
        "    Label(  'polegroup'            , 18 ,      19 , 'object'          , 3       , False        , True         , (153,153,153) ),\n",
        "    Label(  'traffic light'        , 19 ,        6 , 'object'          , 3       , False        , False        , (250,170, 30) ),\n",
        "    Label(  'traffic sign'         , 20 ,        7 , 'object'          , 3       , False        , False        , (220,220,  0) ),\n",
        "    Label(  'vegetation'           , 21 ,        8 , 'nature'          , 4       , False        , False        , (107,142, 35) ),\n",
        "    Label(  'terrain'              , 22 ,        9 , 'nature'          , 4       , False        , False        , (152,251,152) ),\n",
        "    Label(  'sky'                  , 23 ,       10 , 'sky'             , 5       , False        , False        , ( 70,130,180) ),\n",
        "    Label(  'person'               , 24 ,       11 , 'human'           , 6       , True         , False        , (220, 20, 60) ),\n",
        "    Label(  'rider'                , 25 ,       12 , 'human'           , 6       , True         , False        , (255,  0,  0) ),\n",
        "    Label(  'car'                  , 26 ,       13 , 'vehicle'         , 7       , True         , False        , (  0,  0,142) ),\n",
        "    Label(  'truck'                , 27 ,       14 , 'vehicle'         , 7       , True         , False        , (  0,  0, 70) ),\n",
        "    Label(  'bus'                  , 28 ,       15 , 'vehicle'         , 7       , True         , False        , (  0, 60,100) ),\n",
        "    Label(  'caravan'              , 29 ,      19 , 'vehicle'         , 7       , True         , True         , (  0,  0, 90) ),\n",
        "    Label(  'trailer'              , 30 ,      19 , 'vehicle'         , 7       , True         , True         , (  0,  0,110) ),\n",
        "    Label(  'train'                , 31 ,       16 , 'vehicle'         , 7       , True         , False        , (  0, 80,100) ),\n",
        "    Label(  'motorcycle'           , 32 ,       17 , 'vehicle'         , 7       , True         , False        , (  0,  0,230) ),\n",
        "    Label(  'bicycle'              , 33 ,       18 , 'vehicle'         , 7       , True         , False        , (119, 11, 32) ),\n",
        "    Label(  'license plate'        , -1 ,       19 , 'vehicle'         , 7       , False        , True         , (  0,  0,142) ),\n",
        "]\n",
        "\n",
        "# create a function which maps id to trainId:\n",
        "id_to_trainId = {label.id: label.trainId for label in labels}\n",
        "id_to_trainId_map_func = np.vectorize(id_to_trainId.get)\n",
        "\n",
        "train_dirs = [\"jena/\", \"zurich/\", \"weimar/\", \"ulm/\", \"tubingen/\", \"stuttgart/\",\n",
        "              \"strasbourg/\", \"monchengladbach/\", \"krefeld/\", \"hanover/\",\n",
        "              \"hamburg/\", \"erfurt/\", \"dusseldorf/\", \"darmstadt/\", \"cologne/\",\n",
        "              \"bremen/\", \"bochum/\", \"aachen/\"]\n",
        "val_dirs = [\"frankfurt/\", \"munster/\", \"lindau/\"]\n",
        "test_dirs = [\"berlin\", \"bielefeld\", \"bonn\", \"leverkusen\", \"mainz\", \"munich\"]\n",
        "\n",
        "#cityscapes_data_path = \"/root/deeplabv3/data/cityscapes\"\n",
        "#cityscapes_meta_path = \"/root/deeplabv3/data/cityscapes/meta\"\n",
        "\n",
        "cityscapes_data_path = \"/content/drive/My Drive/U_Net_Datasets\"\n",
        "cityscapes_meta_path = \"/content/drive/My Drive/U_Net_Datasets/meta\"\n",
        "\n",
        "if not os.path.exists(cityscapes_meta_path):\n",
        "    os.makedirs(cityscapes_meta_path)\n",
        "if not os.path.exists(cityscapes_meta_path + \"/label_imgs\"):\n",
        "    os.makedirs(cityscapes_meta_path + \"/label_imgs\")\n",
        "\n",
        "################################################################################\n",
        "# convert all labels to label imgs with trainId pixel values (and save to disk):\n",
        "################################################################################\n",
        "train_label_img_paths = []\n",
        "\n",
        "img_dir = cityscapes_data_path + \"/leftImg8bit/train/\"\n",
        "label_dir = cityscapes_data_path + \"/gtFine/train/\"\n",
        "for train_dir in train_dirs:\n",
        "    print (train_dir)\n",
        "\n",
        "    train_img_dir_path = img_dir + train_dir\n",
        "    train_label_dir_path = label_dir + train_dir\n",
        "\n",
        "    file_names = os.listdir(train_img_dir_path)\n",
        "    for file_name in file_names:\n",
        "        img_id = file_name.split(\"_leftImg8bit.png\")[0]\n",
        "\n",
        "        gtFine_img_path = train_label_dir_path + img_id + \"_gtFine_labelIds.png\"\n",
        "        gtFine_img = cv2.imread(gtFine_img_path, -1) # (shape: (1024, 2048))\n",
        "\n",
        "        # convert gtFine_img from id to trainId pixel values:\n",
        "        label_img = id_to_trainId_map_func(gtFine_img)#(shape: (1024, 2048))\n",
        "        label_img = label_img.astype(np.uint8)\n",
        "\n",
        "        cv2.imwrite(cityscapes_meta_path + \"/label_imgs/\" + img_id + \".png\", label_img)\n",
        "        train_label_img_paths.append(cityscapes_meta_path + \"/label_imgs/\" + img_id + \".png\")\n",
        "\n",
        "img_dir = cityscapes_data_path + \"/leftImg8bit/val/\"\n",
        "label_dir = cityscapes_data_path + \"/gtFine/val/\"\n",
        "for val_dir in val_dirs:\n",
        "    print (val_dir)\n",
        "\n",
        "    val_img_dir_path = img_dir + val_dir\n",
        "    val_label_dir_path = label_dir + val_dir\n",
        "\n",
        "    file_names = os.listdir(val_img_dir_path)\n",
        "    for file_name in file_names:\n",
        "        img_id = file_name.split(\"_leftImg8bit.png\")[0]\n",
        "\n",
        "        gtFine_img_path = val_label_dir_path + img_id + \"_gtFine_labelIds.png\"\n",
        "        gtFine_img = cv2.imread(gtFine_img_path, -1) # (shape: (1024, 2048))\n",
        "\n",
        "        # convert gtFine_img from id to trainId pixel values:\n",
        "        label_img = id_to_trainId_map_func(gtFine_img) # (shape: (1024, 2048))\n",
        "        label_img = label_img.astype(np.uint8)\n",
        "\n",
        "        cv2.imwrite(cityscapes_meta_path + \"/label_imgs/\" + img_id + \".png\", label_img)\n",
        "\n",
        "################################################################################\n",
        "# compute the class weigths:\n",
        "################################################################################\n",
        "print (\"computing class weights\")\n",
        "\n",
        "num_classes = 20\n",
        "\n",
        "trainId_to_count = {}\n",
        "for trainId in range(num_classes):\n",
        "    trainId_to_count[trainId] = 0\n",
        "\n",
        "# get the total number of pixels in all train label_imgs that are of each object class:\n",
        "for step, label_img_path in enumerate(train_label_img_paths):\n",
        "    if step % 100 == 0:\n",
        "        print (step)\n",
        "\n",
        "    label_img = cv2.imread(label_img_path, -1)\n",
        "\n",
        "    for trainId in range(num_classes):\n",
        "        # count how many pixels in label_img which are of object class trainId:\n",
        "        trainId_mask = np.equal(label_img, trainId)\n",
        "        trainId_count = np.sum(trainId_mask)\n",
        "\n",
        "        # add to the total count:\n",
        "        trainId_to_count[trainId] += trainId_count\n",
        "\n",
        "# compute the class weights according to the ENet paper:\n",
        "class_weights = []\n",
        "total_count = sum(trainId_to_count.values())\n",
        "for trainId, count in trainId_to_count.items():\n",
        "    trainId_prob = float(count)/float(total_count)\n",
        "    trainId_weight = 1/np.log(1.02 + trainId_prob)\n",
        "    class_weights.append(trainId_weight)\n",
        "\n",
        "print (class_weights)\n",
        "\n",
        "with open(cityscapes_meta_path + \"/class_weights.pkl\", \"wb\") as file:\n",
        "    pickle.dump(class_weights, file, protocol=2) # (protocol=2 is needed to be able to open this file with python2)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "jena/\n",
            "zurich/\n",
            "weimar/\n",
            "ulm/\n",
            "tubingen/\n",
            "stuttgart/\n",
            "strasbourg/\n",
            "monchengladbach/\n",
            "krefeld/\n",
            "hanover/\n",
            "hamburg/\n",
            "erfurt/\n",
            "dusseldorf/\n",
            "darmstadt/\n",
            "cologne/\n",
            "bremen/\n",
            "bochum/\n",
            "aachen/\n",
            "frankfurt/\n",
            "munster/\n",
            "lindau/\n",
            "computing class weights\n",
            "0\n",
            "100\n",
            "200\n",
            "300\n",
            "400\n",
            "500\n",
            "600\n",
            "700\n",
            "800\n",
            "900\n",
            "1000\n",
            "1100\n",
            "1200\n",
            "1300\n",
            "1400\n",
            "1500\n",
            "1600\n",
            "1700\n",
            "1800\n",
            "1900\n",
            "2000\n",
            "2100\n",
            "2200\n",
            "2300\n",
            "2400\n",
            "2500\n",
            "2600\n",
            "2700\n",
            "2800\n",
            "2900\n",
            "[3.362088928135997, 14.031521298730318, 4.986657918172686, 39.254403222891234, 36.5125971773311, 32.89620795239199, 46.286660134462245, 40.69042748040039, 6.698241903441155, 33.55545414377673, 18.487832644189325, 32.97431249303082, 47.676506488107115, 12.70028597336979, 45.20542136324199, 45.78372411642551, 45.825290445040096, 48.40614734589367, 42.75592219573717, 7.912219457368151]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2INXlplU73ZG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#einmalige ausführung\n",
        "#import zipfile\n",
        "#path_to_zip_file = #'/content/drive/My Drive/U_Net_Datasets/gtFine_trainvaltest.zip'\n",
        "#directory_to_extract_to = '/content/drive/My Drive/U_Net_Datasets' \n",
        "#with zipfile.ZipFile(path_to_zip_file, 'r') as zip_ref:\n",
        "#    zip_ref.extractall(directory_to_extract_to)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DHSYi2oj5fxc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# this file contains code snippets which I have found (more or less) useful at\n",
        "# some point during the project. Probably nothing interesting to see here.\n",
        "\n",
        "import pickle\n",
        "import matplotlib\n",
        "matplotlib.use(\"Agg\")\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "import numpy as np\n",
        "\n",
        "model_id = \"13_2_2_2\"\n",
        "\n",
        "with open(\"/home/fregu856/exjobb/training_logs/multitask/model_\" + model_id + \"/epoch_losses_train.pkl\", \"rb\") as file:\n",
        "    train_loss = pickle.load(file)\n",
        "\n",
        "with open(\"/home/fregu856/exjobb/training_logs/multitask/model_\" + model_id + \"/epoch_losses_val.pkl\", \"rb\") as file:\n",
        "    val_loss = pickle.load(file)\n",
        "\n",
        "print (\"train loss min:\", np.argmin(np.array(train_loss)), np.min(np.array(train_loss)))\n",
        "\n",
        "print (\"val loss min:\", np.argmin(np.array(val_loss)), np.min(np.array(val_loss)))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cfcowAUH32SX",
        "colab_type": "text"
      },
      "source": [
        "###Model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HwBiOz6m39og",
        "colab_type": "text"
      },
      "source": [
        "####ASPP\n",
        "Use of Atrous Spatial Pyramid Pooling (ASPP). The idea is to apply multiple atrous convolution with different sampling rates to the input feature map, and fuse together. As objects of the same class can have different scales in the image, ASPP helps to account for different object scales which can improve the accuracy.\n",
        "\n",
        "\n",
        "![image.png](data:image/png;base64,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)\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "e8lPxBUz2251",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "\n",
        "\n",
        "class ASPP(nn.Module):\n",
        "    def __init__(self, num_classes):\n",
        "        super(ASPP, self).__init__()\n",
        "\n",
        "        self.conv_1x1_1 = nn.Conv2d(512, 256, kernel_size=1)\n",
        "        self.bn_conv_1x1_1 = nn.BatchNorm2d(256)\n",
        "\n",
        "        self.conv_3x3_1 = nn.Conv2d(512, 256, kernel_size=3, stride=1, padding=6, dilation=6)\n",
        "        self.bn_conv_3x3_1 = nn.BatchNorm2d(256)\n",
        "\n",
        "        self.conv_3x3_2 = nn.Conv2d(512, 256, kernel_size=3, stride=1, padding=12, dilation=12)\n",
        "        self.bn_conv_3x3_2 = nn.BatchNorm2d(256)\n",
        "\n",
        "        self.conv_3x3_3 = nn.Conv2d(512, 256, kernel_size=3, stride=1, padding=18, dilation=18)\n",
        "        self.bn_conv_3x3_3 = nn.BatchNorm2d(256)\n",
        "\n",
        "        self.avg_pool = nn.AdaptiveAvgPool2d(1)\n",
        "\n",
        "        self.conv_1x1_2 = nn.Conv2d(512, 256, kernel_size=1)\n",
        "        self.bn_conv_1x1_2 = nn.BatchNorm2d(256)\n",
        "\n",
        "        self.conv_1x1_3 = nn.Conv2d(1280, 256, kernel_size=1) # (1280 = 5*256)\n",
        "        self.bn_conv_1x1_3 = nn.BatchNorm2d(256)\n",
        "\n",
        "        self.conv_1x1_4 = nn.Conv2d(256, num_classes, kernel_size=1)\n",
        "\n",
        "    def forward(self, feature_map):\n",
        "        # (feature_map has shape (batch_size, 512, h/16, w/16)) (assuming self.resnet is ResNet18_OS16 or ResNet34_OS16. If self.resnet instead is ResNet18_OS8 or ResNet34_OS8, it will be (batch_size, 512, h/8, w/8))\n",
        "\n",
        "        feature_map_h = feature_map.size()[2] # (== h/16)\n",
        "        feature_map_w = feature_map.size()[3] # (== w/16)\n",
        "\n",
        "        out_1x1 = F.relu(self.bn_conv_1x1_1(self.conv_1x1_1(feature_map))) # (shape: (batch_size, 256, h/16, w/16))\n",
        "        out_3x3_1 = F.relu(self.bn_conv_3x3_1(self.conv_3x3_1(feature_map))) # (shape: (batch_size, 256, h/16, w/16))\n",
        "        out_3x3_2 = F.relu(self.bn_conv_3x3_2(self.conv_3x3_2(feature_map))) # (shape: (batch_size, 256, h/16, w/16))\n",
        "        out_3x3_3 = F.relu(self.bn_conv_3x3_3(self.conv_3x3_3(feature_map))) # (shape: (batch_size, 256, h/16, w/16))\n",
        "\n",
        "        out_img = self.avg_pool(feature_map) # (shape: (batch_size, 512, 1, 1))\n",
        "        out_img = F.relu(self.bn_conv_1x1_2(self.conv_1x1_2(out_img))) # (shape: (batch_size, 256, 1, 1))\n",
        "        out_img = F.upsample(out_img, size=(feature_map_h, feature_map_w), mode=\"bilinear\") # (shape: (batch_size, 256, h/16, w/16))\n",
        "\n",
        "        out = torch.cat([out_1x1, out_3x3_1, out_3x3_2, out_3x3_3, out_img], 1) # (shape: (batch_size, 1280, h/16, w/16))\n",
        "        out = F.relu(self.bn_conv_1x1_3(self.conv_1x1_3(out))) # (shape: (batch_size, 256, h/16, w/16))\n",
        "        out = self.conv_1x1_4(out) # (shape: (batch_size, num_classes, h/16, w/16))\n",
        "\n",
        "        return out\n",
        "\n",
        "class ASPP_Bottleneck(nn.Module):\n",
        "    def __init__(self, num_classes):\n",
        "        super(ASPP_Bottleneck, self).__init__()\n",
        "\n",
        "        self.conv_1x1_1 = nn.Conv2d(4*512, 256, kernel_size=1)\n",
        "        self.bn_conv_1x1_1 = nn.BatchNorm2d(256)\n",
        "\n",
        "        self.conv_3x3_1 = nn.Conv2d(4*512, 256, kernel_size=3, stride=1, padding=6, dilation=6)\n",
        "        self.bn_conv_3x3_1 = nn.BatchNorm2d(256)\n",
        "\n",
        "        self.conv_3x3_2 = nn.Conv2d(4*512, 256, kernel_size=3, stride=1, padding=12, dilation=12)\n",
        "        self.bn_conv_3x3_2 = nn.BatchNorm2d(256)\n",
        "\n",
        "        self.conv_3x3_3 = nn.Conv2d(4*512, 256, kernel_size=3, stride=1, padding=18, dilation=18)\n",
        "        self.bn_conv_3x3_3 = nn.BatchNorm2d(256)\n",
        "\n",
        "        self.avg_pool = nn.AdaptiveAvgPool2d(1)\n",
        "\n",
        "        self.conv_1x1_2 = nn.Conv2d(4*512, 256, kernel_size=1)\n",
        "        self.bn_conv_1x1_2 = nn.BatchNorm2d(256)\n",
        "\n",
        "        self.conv_1x1_3 = nn.Conv2d(1280, 256, kernel_size=1) # (1280 = 5*256)\n",
        "        self.bn_conv_1x1_3 = nn.BatchNorm2d(256)\n",
        "\n",
        "        self.conv_1x1_4 = nn.Conv2d(256, num_classes, kernel_size=1)\n",
        "\n",
        "    def forward(self, feature_map):\n",
        "        # (feature_map has shape (batch_size, 4*512, h/16, w/16))\n",
        "\n",
        "        feature_map_h = feature_map.size()[2] # (== h/16)\n",
        "        feature_map_w = feature_map.size()[3] # (== w/16)\n",
        "\n",
        "        out_1x1 = F.relu(self.bn_conv_1x1_1(self.conv_1x1_1(feature_map))) # (shape: (batch_size, 256, h/16, w/16))\n",
        "        out_3x3_1 = F.relu(self.bn_conv_3x3_1(self.conv_3x3_1(feature_map))) # (shape: (batch_size, 256, h/16, w/16))\n",
        "        out_3x3_2 = F.relu(self.bn_conv_3x3_2(self.conv_3x3_2(feature_map))) # (shape: (batch_size, 256, h/16, w/16))\n",
        "        out_3x3_3 = F.relu(self.bn_conv_3x3_3(self.conv_3x3_3(feature_map))) # (shape: (batch_size, 256, h/16, w/16))\n",
        "\n",
        "        out_img = self.avg_pool(feature_map) # (shape: (batch_size, 512, 1, 1))\n",
        "        out_img = F.relu(self.bn_conv_1x1_2(self.conv_1x1_2(out_img))) # (shape: (batch_size, 256, 1, 1))\n",
        "        out_img = F.upsample(out_img, size=(feature_map_h, feature_map_w), mode=\"bilinear\") # (shape: (batch_size, 256, h/16, w/16))\n",
        "\n",
        "        out = torch.cat([out_1x1, out_3x3_1, out_3x3_2, out_3x3_3, out_img], 1) # (shape: (batch_size, 1280, h/16, w/16))\n",
        "        out = F.relu(self.bn_conv_1x1_3(self.conv_1x1_3(out))) # (shape: (batch_size, 256, h/16, w/16))\n",
        "        out = self.conv_1x1_4(out) # (shape: (batch_size, num_classes, h/16, w/16))\n",
        "\n",
        "        return out"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XO6CrH4C4MqC",
        "colab_type": "text"
      },
      "source": [
        "####ResNet\n",
        "\n",
        "Various pre-trained ResNet models for training\n",
        "Models are pre-trained on ImageNet database"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "z40_2NTCF4VO",
        "colab_type": "text"
      },
      "source": [
        "![image.png](data:image/png;base64,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)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kSFBJVdj4BJz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# NOTE! OS: output stride, the ratio of input image resolution to final output resolution (OS16: output size is (img_h/16, img_w/16)) (OS8: output size is (img_h/8, img_w/8))\n",
        "\n",
        "preTrainedModelsPath = datasetPath + \"/pretrained_models\"\n",
        "\n",
        "def make_layer(block, in_channels, channels, num_blocks, stride=1, dilation=1):\n",
        "    strides = [stride] + [1]*(num_blocks - 1) # (stride == 2, num_blocks == 4 --> strides == [2, 1, 1, 1])\n",
        "\n",
        "    blocks = []\n",
        "    for stride in strides:\n",
        "        blocks.append(block(in_channels=in_channels, channels=channels, stride=stride, dilation=dilation))\n",
        "        in_channels = block.expansion*channels\n",
        "\n",
        "    layer = nn.Sequential(*blocks) # (*blocks: call with unpacked list entires as arguments)\n",
        "\n",
        "    return layer\n",
        "\n",
        "class BasicBlock(nn.Module):\n",
        "    expansion = 1\n",
        "\n",
        "    def __init__(self, in_channels, channels, stride=1, dilation=1):\n",
        "        super(BasicBlock, self).__init__()\n",
        "\n",
        "        out_channels = self.expansion*channels\n",
        "\n",
        "        self.conv1 = nn.Conv2d(in_channels, channels, kernel_size=3, stride=stride, padding=dilation, dilation=dilation, bias=False)\n",
        "        self.bn1 = nn.BatchNorm2d(channels)\n",
        "\n",
        "        self.conv2 = nn.Conv2d(channels, channels, kernel_size=3, stride=1, padding=dilation, dilation=dilation, bias=False)\n",
        "        self.bn2 = nn.BatchNorm2d(channels)\n",
        "\n",
        "        if (stride != 1) or (in_channels != out_channels):\n",
        "            conv = nn.Conv2d(in_channels, out_channels, kernel_size=1, stride=stride, bias=False)\n",
        "            bn = nn.BatchNorm2d(out_channels)\n",
        "            self.downsample = nn.Sequential(conv, bn)\n",
        "        else:\n",
        "            self.downsample = nn.Sequential()\n",
        "\n",
        "    def forward(self, x):\n",
        "        # (x has shape: (batch_size, in_channels, h, w))\n",
        "\n",
        "        out = F.relu(self.bn1(self.conv1(x))) # (shape: (batch_size, channels, h, w) if stride == 1, (batch_size, channels, h/2, w/2) if stride == 2)\n",
        "        out = self.bn2(self.conv2(out)) # (shape: (batch_size, channels, h, w) if stride == 1, (batch_size, channels, h/2, w/2) if stride == 2)\n",
        "\n",
        "        out = out + self.downsample(x) # (shape: (batch_size, channels, h, w) if stride == 1, (batch_size, channels, h/2, w/2) if stride == 2)\n",
        "\n",
        "        out = F.relu(out) # (shape: (batch_size, channels, h, w) if stride == 1, (batch_size, channels, h/2, w/2) if stride == 2)\n",
        "\n",
        "        return out\n",
        "\n",
        "class Bottleneck(nn.Module):\n",
        "    expansion = 4\n",
        "\n",
        "    def __init__(self, in_channels, channels, stride=1, dilation=1):\n",
        "        super(Bottleneck, self).__init__()\n",
        "\n",
        "        out_channels = self.expansion*channels\n",
        "\n",
        "        self.conv1 = nn.Conv2d(in_channels, channels, kernel_size=1, bias=False)\n",
        "        self.bn1 = nn.BatchNorm2d(channels)\n",
        "\n",
        "        self.conv2 = nn.Conv2d(channels, channels, kernel_size=3, stride=stride, padding=dilation, dilation=dilation, bias=False)\n",
        "        self.bn2 = nn.BatchNorm2d(channels)\n",
        "\n",
        "        self.conv3 = nn.Conv2d(channels, out_channels, kernel_size=1, bias=False)\n",
        "        self.bn3 = nn.BatchNorm2d(out_channels)\n",
        "\n",
        "        if (stride != 1) or (in_channels != out_channels):\n",
        "            conv = nn.Conv2d(in_channels, out_channels, kernel_size=1, stride=stride, bias=False)\n",
        "            bn = nn.BatchNorm2d(out_channels)\n",
        "            self.downsample = nn.Sequential(conv, bn)\n",
        "        else:\n",
        "            self.downsample = nn.Sequential()\n",
        "\n",
        "    def forward(self, x):\n",
        "        # (x has shape: (batch_size, in_channels, h, w))\n",
        "\n",
        "        out = F.relu(self.bn1(self.conv1(x))) # (shape: (batch_size, channels, h, w))\n",
        "        out = F.relu(self.bn2(self.conv2(out))) # (shape: (batch_size, channels, h, w) if stride == 1, (batch_size, channels, h/2, w/2) if stride == 2)\n",
        "        out = self.bn3(self.conv3(out)) # (shape: (batch_size, out_channels, h, w) if stride == 1, (batch_size, out_channels, h/2, w/2) if stride == 2)\n",
        "\n",
        "        out = out + self.downsample(x) # (shape: (batch_size, out_channels, h, w) if stride == 1, (batch_size, out_channels, h/2, w/2) if stride == 2)\n",
        "\n",
        "        out = F.relu(out) # (shape: (batch_size, out_channels, h, w) if stride == 1, (batch_size, out_channels, h/2, w/2) if stride == 2)\n",
        "\n",
        "        return out\n",
        "\n",
        "class ResNet_Bottleneck_OS16(nn.Module):\n",
        "    def __init__(self, num_layers):\n",
        "        super(ResNet_Bottleneck_OS16, self).__init__()\n",
        "\n",
        "        if num_layers == 50:\n",
        "            resnet = models.resnet50()\n",
        "            # load pretrained model:\n",
        "            resnet.load_state_dict(torch.load(preTrainedModelsPath+\"/resnet/resnet50-19c8e357.pth\"))\n",
        "            # remove fully connected layer, avg pool and layer5:\n",
        "            self.resnet = nn.Sequential(*list(resnet.children())[:-3])\n",
        "\n",
        "            print (\"pretrained resnet, 50\")\n",
        "        elif num_layers == 101:\n",
        "            resnet = models.resnet101()\n",
        "            # load pretrained model:\n",
        "            resnet.load_state_dict(torch.load(preTrainedModelsPath+\"/resnet/resnet101-5d3b4d8f.pth\"))\n",
        "            \n",
        "            # remove fully connected layer, avg pool and layer5:\n",
        "            self.resnet = nn.Sequential(*list(resnet.children())[:-3])\n",
        "\n",
        "            print (\"pretrained resnet, 101\")\n",
        "            \n",
        "        elif num_layers == 152:\n",
        "            resnet = models.resnet152()\n",
        "            # load pretrained model:\n",
        "            resnet.load_state_dict(torch.load(preTrainedModelsPath+\"/resnet/resnet152-b121ed2d.pth\"))\n",
        "            # remove fully connected layer, avg pool and layer5:\n",
        "            self.resnet = nn.Sequential(*list(resnet.children())[:-3])\n",
        "\n",
        "            print (\"pretrained resnet, 152\")\n",
        "        else:\n",
        "            raise Exception(\"num_layers must be in {50, 101, 152}!\")\n",
        "\n",
        "        self.layer5 = make_layer(Bottleneck, in_channels=4*256, channels=512, num_blocks=3, stride=1, dilation=2)\n",
        "\n",
        "    def forward(self, x):\n",
        "        # (x has shape (batch_size, 3, h, w))\n",
        "\n",
        "        # pass x through (parts of) the pretrained ResNet:\n",
        "        c4 = self.resnet(x) # (shape: (batch_size, 4*256, h/16, w/16)) (it's called c4 since 16 == 2^4)\n",
        "\n",
        "        output = self.layer5(c4) # (shape: (batch_size, 4*512, h/16, w/16))\n",
        "\n",
        "        return output\n",
        "\n",
        "class ResNet_BasicBlock_OS16(nn.Module):\n",
        "    def __init__(self, num_layers):\n",
        "        super(ResNet_BasicBlock_OS16, self).__init__()\n",
        "\n",
        "        if num_layers == 18:\n",
        "            resnet = models.resnet18()\n",
        "            # load pretrained model:\n",
        "            \n",
        "            \n",
        "            resnet.load_state_dict(torch.load(preTrainedModelsPath+\"/resnet/resnet18-5c106cde.pth\"))\n",
        "            # remove fully connected layer, avg pool and layer5:\n",
        "            \n",
        "            self.resnet = nn.Sequential(*list(resnet.children())[:-3])\n",
        "\n",
        "            num_blocks = 2\n",
        "            print (\"pretrained resnet, 18\")\n",
        "        elif num_layers == 34:\n",
        "            resnet = models.resnet34()\n",
        "            # load pretrained model:\n",
        "            resnet.load_state_dict(torch.load(preTrainedModelsPath+\"/resnet/resnet34-333f7ec4.pth\"))\n",
        "            # remove fully connected layer, avg pool and layer5:\n",
        "            self.resnet = nn.Sequential(*list(resnet.children())[:-3])\n",
        "\n",
        "            num_blocks = 3\n",
        "            print (\"pretrained resnet, 34\")\n",
        "        else:\n",
        "            raise Exception(\"num_layers must be in {18, 34}!\")\n",
        "\n",
        "        self.layer5 = make_layer(BasicBlock, in_channels=256, channels=512, num_blocks=num_blocks, stride=1, dilation=2)\n",
        "\n",
        "    def forward(self, x):\n",
        "        # (x has shape (batch_size, 3, h, w))\n",
        "\n",
        "        # pass x through (parts of) the pretrained ResNet:\n",
        "        c4 = self.resnet(x) # (shape: (batch_size, 256, h/16, w/16)) (it's called c4 since 16 == 2^4)\n",
        "\n",
        "        output = self.layer5(c4) # (shape: (batch_size, 512, h/16, w/16))\n",
        "\n",
        "        return output\n",
        "\n",
        "class ResNet_BasicBlock_OS8(nn.Module):\n",
        "    def __init__(self, num_layers):\n",
        "        super(ResNet_BasicBlock_OS8, self).__init__()\n",
        "\n",
        "        if num_layers == 18:\n",
        "            resnet = models.resnet18()\n",
        "            # load pretrained model:            \n",
        "            resnet.load_state_dict(torch.load(preTrainedModelsPath+\"/resnet/resnet18-5c106cde.pth\"))       \n",
        "\n",
        "            # remove fully connected layer, avg pool, layer4 and layer5:\n",
        "\n",
        "            \n",
        "            self.resnet = nn.Sequential(*list(resnet.children())[:-4])\n",
        "\n",
        "            num_blocks_layer_4 = 2\n",
        "            num_blocks_layer_5 = 2\n",
        "            print (\"pretrained resnet, 18\")\n",
        "        elif num_layers == 34:\n",
        "            resnet = models.resnet34()\n",
        "            # load pretrained model:\n",
        "            resnet.load_state_dict(torch.load(preTrainedModelsPath+\"/resnet/resnet34-333f7ec4.pth\"))\n",
        "            # remove fully connected layer, avg pool, layer4 and layer5:\n",
        "            self.resnet = nn.Sequential(*list(resnet.children())[:-4])\n",
        "\n",
        "            num_blocks_layer_4 = 6\n",
        "            num_blocks_layer_5 = 3\n",
        "            print (\"pretrained resnet, 34\")\n",
        "        else:\n",
        "            raise Exception(\"num_layers must be in {18, 34}!\")\n",
        "\n",
        "        self.layer4 = make_layer(BasicBlock, in_channels=128, channels=256, num_blocks=num_blocks_layer_4, stride=1, dilation=2)\n",
        "\n",
        "        self.layer5 = make_layer(BasicBlock, in_channels=256, channels=512, num_blocks=num_blocks_layer_5, stride=1, dilation=4)\n",
        "\n",
        "    def forward(self, x):\n",
        "        # (x has shape (batch_size, 3, h, w))\n",
        "\n",
        "        # pass x through (parts of) the pretrained ResNet:\n",
        "        c3 = self.resnet(x) # (shape: (batch_size, 128, h/8, w/8)) (it's called c3 since 8 == 2^3)\n",
        "\n",
        "        output = self.layer4(c3) # (shape: (batch_size, 256, h/8, w/8))\n",
        "        output = self.layer5(output) # (shape: (batch_size, 512, h/8, w/8))\n",
        "\n",
        "        return output\n",
        "\n",
        "def ResNet18_OS16():\n",
        "    return ResNet_BasicBlock_OS16(num_layers=18)\n",
        "\n",
        "def ResNet34_OS16():\n",
        "    return ResNet_BasicBlock_OS16(num_layers=34)\n",
        "\n",
        "def ResNet50_OS16():\n",
        "    return ResNet_Bottleneck_OS16(num_layers=50)\n",
        "\n",
        "def ResNet101_OS16():\n",
        "    return ResNet_Bottleneck_OS16(num_layers=101)\n",
        "\n",
        "def ResNet152_OS16():\n",
        "    return ResNet_Bottleneck_OS16(num_layers=152)\n",
        "\n",
        "def ResNet18_OS8():\n",
        "    return ResNet_BasicBlock_OS8(num_layers=18)\n",
        "\n",
        "def ResNet34_OS8():\n",
        "    return ResNet_BasicBlock_OS8(num_layers=34)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8l35iRa74B86",
        "colab_type": "text"
      },
      "source": [
        "####DeepLab V3\n",
        "\n",
        "Creating the DeepLab V3 \n",
        "  - here you can set which ResNet is used\n",
        "  "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-Yc2wbsyF_sg",
        "colab_type": "text"
      },
      "source": [
        "![image.png](data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAvwAAAGICAYAAAA9AzTFAAAgAElEQVR4Aey9Z3BWyZYl2tEdEx0TMf/e/OnofzPRPR0Tb153T7++/eb6W+7eqrp1iyooiqqicIX33ggQ3khIQsILL0Agb5H3FoQEEkYCgRASwhsZQF5ivbXy0weqKpBASCBBnuDwHZ2TZufKPHlW7ty5869gD4uARcAiYBGwCFgELAIWAYuAReCtReCv3tqS2YK9NAKNza24efcB7tbU41FDM27ff4jquno8qG/CjTsPUPugEfdq63GDYaofNOBO9SNc53Xto0bz/G7NI3P/5r2HDNeAOt3n87pHTSaewj/gdXWdI259Y4tJQ3/rvHmPeTN9XVfdqsV9pqG0lLfuSR5dKw2F1T3loWvJVsv7txj+YX2zyV/hHzY0mXh3mbfkuM74KkPdw0Zcu12HWv7qb5OGrlmGW5RZ5X+ah9Jg2Zif8lYaSuuW5OmQ4zbTd6Zx7U6dkV0yKYziOWV71NBiZHNicZ9561phnbIpvHAThpLtOtPTc4XT/RpeCxfJVPeQdcN79xnnIetJ5ZfstawfyaRr4XiLcj7i8yqWWfWkNEx+5pp436419aw8dJo64zNdS2ZnHdQwXZVTcuu8xni1lOEOZVFY4akyG9mIkWSrZv5qL5JN5XBeSzaVVWUUroqna6WhdniP7VB/qy0onMojLCSD6taUqaOeVB6l4czX2WYlu8rrrH/JKfkfMp7SUrrCx+DN/PR35/pVmZSufp1YqBzKz4RV+UwZOur/gaNtmrISF/2qHCq36kzxhNete49MO1VckwfrRmWVfMrPyMO/jWyUU+1SOKqeDMYsg8qh9ISL8hGeqie1f6fM+tWp5/VNLUYGYSd5dE95K72ret+EBd93Z7qSR3WiduUsv0mro56qmYZpe5TL8a7yPVRbMPLomrKxPMqvcztV/d3ifclg6prlqGc9KQ29/840DC6U3dEX1Js6VN2o3ak+K2/WmDIrfdWxKTPTVfoN7Fv0buqeeRcYRtemTB33JZvyqKHM6seUhjBQu1AeCqvy6V3oXH7lrfowbYxpqQ4d5ee7zrwVz7yzTFdYmjxUTx35Od9TySNZneVUn6R2r3TVtpVHje4pDeHF9ISFcJMM13hdw7wlt3BReoonedS3mnpmfN3XtTM/tUNzj7/OMMpD7UHyq190tGnWP+9LRpVPz9W2FEeySR61PZ2qR+e76Ww3TiyEocqkspmwHVjovZAckl3vo7PedU99n/BSm3bGk0wqv75PkufqTbZZyqa/nd8FYak61Huva8WRzCqf0pVMSlcYqo5VPiMbf531L3lNXIatusO+x+Sh91vvr6N+Fc+BMdsCZal96Ohjzbuu9mT6Kce7qfSUn9J0YmGwZjzJpGuVXd8qI48pP9/Du/p+Ce+nGDtlU5t29otqC8JQZVUZVNb6xmbz7kp25aG8JbvBU2XmveaWNp7taGt//NJcwUYYmAhYwj8w663XpFaHcrLkGs5euomyqvs4HHMKMZnncfbiTQQnnEby8Ys4df469oSdQE5hhXl2MPoUUk6UITTlLPZFnMCxM5XwiyxAbNZ5pOVfxuHYQsRnX8Cx0xV8no/jZ68ijn9HpJ7DiXPXkHT8EkKSzqC04g7CmEYq00rmPb8oRxpJx8uwNSAHSccu4UhsEfZH5iMlrwxBlEdy5J+rwqGjlIFxjp+uNPlJtuNnriIgrhCnS28iIbcUwYmneX3D/EaknkUe5dhPeeMoZx5l9g3ORU5RBRKyS025cylvTEYJ8zyF4rJbvHeS+V7CibNVxOKMkbuA8u8NP2HSCkgogl90gZEtOPEMDvI6l+ntCjmOxGMXkXmqHAHxpxmvjPcrcSSuCGeZrgOLYuJ+HfE5pQhLPotsyi+s4ilLMsPvN9cXDN5KL/PkFYQyD+GcxXSPEOMQ/q0yHyBuqScusazXmR/zYF1mn7qCcKZbfOkWkilLIHE5S1x2BObiKMuYUVBu8lN5s06W8/4xFLKeVUfhrBNhpXDh/Luk/I6RJ5V1kN5Rv8I3gbLvCD6GHNaBZAlNOmvklAzC4zhxU50m5l40dazyZ56qQHrBZdaT5KRsrMOjmSXIL64iPoUmT4XxiygwZY/JuoAtR3JMuDCmvzs0DxmUITixCP5sq2qnagvCRO1CWEp2lWV/eD7OX76N6PRiRKWx7fF+GOUKIHZnGG8v21LaicvE55LBUGVSm1IeqvNA1m8ky690Q5murs+pXTB+JjFL4ruh9q840XxnDkafRFbhFcpz0tEmiYt/TCGiM86bd0fxMliPagtBbBdFbJtq1+Fsm3rH1C6Fo2mnbPNqG8Jub3ieebdUj9tZf8dMmz+FQ2xvKndY8hkExhei+PItto98nGG6J4uvIYL1eJK4nuL7rTJfqrxn2qDeo2xiLHlCGVdtb8uRbGJ8ke221JRDbTeW2AfEnTZYKpwTC9WT3u90YrA/8qSRUXKqHlQ21YHewyzmoXYQxPcw9/RVtgWFvUQMyg222czjmN4LylF65Q4SWd5Y4pjDe8Jb7fAEsd/sn23aWlRaMfaE5DHcBRxlnXodzOD7xH6KbUHtuIBhI4ll5snLpqx6H9W3ZRZcYd9RhAJiIrn9+V4XFF/nPUeZlJ/6GLV3vbd7w5SHQxa132N8x/QeBLPPUlvQO6Kwpy5cxwG9/6x/lV/tW/1fHOXbxTakdA8xrxDzXlSwzGcQwnKpbQorpSUs9J5KNvVlh4iR2oLatsqk/uQQ21UY8xZW29gvqq9Tm/YNyjXvQnRGsWmHwkrx0tifllbcNf33McYvZD+jvlDp6v09xDB6FwLZBtXHq684xD5dbTKVz335TqttO7A6ZeIIF7XlwvM3TN1FUzaVLyTxLNtZMbKZz27ipnfb0W6KzDsfwza9n31WAWU4yLKpbWQxP7ULPVP97GN/qv5W5VcbK2CbVR+pPreEbfpgVL5pC+l8V7ez/PrOJBHvLYezKFuZaTN+7C/zWSa1mWTKrr5F9aS6Kbpww7yTeq76UbqSQf2Xvnd6P9Smo9LP8btx0fRZan/CQv2bvnVRLK/qV/XlxOUCv18qh+LovVca+nYKF/XZ6mP13NFnV5h61rXqX++H3me12SD2l8f5PSpS/x3HdsG+R32B3gv1U+oThJvwOcByRqUWI4fybz2cbb5Pei+EofpStYUjTKOk/JapX9W3sA9nnAzWzdmLt0zbU5kLiUt0uqMOWtvae41T2IT6LwKW8PffuukzyfRySztwjp1iCT+0IlQb9qWbDmPltkTTkaaw01yxPYEfruOmA5+2Jtx0aN4HsjDf8yj2sINZsysZM9aHm05ttnsUfPyzsI8fzsWbYplGtvnIzlwfYT5ym/i32540dsjF7ESPY+WOJH64rmDN7lTsJWnayQ/57A2ONHaGnsCYpQHm3pJN8ZjlHsmPyQks35qIqWvDEckObr7XUUPMNHBY7BNrCI0I59LNcYa0iBit3J5oPtwr+Ou+J9V0uLMo7xbKqQ/ohOVBhghvY/ldfGKMvN4kEUpPH00X7xjzEYtgZ7mCuOyk3Opcp60LNx3wkq3xmOMRZWRTXvN5LTI1aWUwP5rHzYd16ZYEluOE+Zgu3hxvPoKbDRapJLrnsYUfsLXEUR+TGcRKdSGyMNON1wzncyiT6YWYTnrV9iTMJhYiLEuI8SrfJEOy5xG3fSSFIilLt8Sbj4A+ZOt2pSKdhHIXBzbLiEsyP0zjXAPhfSCDROUkVDe69idBHbcsCLH8yG7Yl4b1xCqMH8uNfLaO1yKpM90izWBLH28X5r2N+G4NyMXYFcEI4AdU9bnaNxkHSWSWUIaVki35nKnT7SQme0m+l7D8h0ga9pNE6VpkV2X1Yj6RJOSLN8WZ64MMM8stCt6HstimsjHK5QjLkIe1O5MxZVWY+YCuZNtcuDHGtNmFG4+SvJ4ydbOOWIpAJbAss1m+LH50Pfens1yphgCvoVxLmY8+jtPYljTIUtubQwx3BBJ7Xk9ZHcp2eg7LtibAfW+aaW9rmLc7sTHtgu1DpH4nP+pq/xogePK9mOsZDX9+pNU2XdQm2Rb063kg0wzEXJiv8AnkB15pJ5AYiiQJY2G/ie1S7UikUG1eg94dQceMnJuIg/L7YWmQed/URhcwPw1K1+1KYnpxhpjNdo/gQLnUEGK33cnmN5YDKrWXPBIGvU/L2ZZF9l184rB6Z5J5r0e6HDYkxZf56f0WKVSeSzcn8P0vNeFUfr17qie1TT+2oVnu0dgedJwn47EeAkksV7MOVL8iiCt2JGI5yxTEd3P2BoYNzGO8U1i2Ld7UmQinZFNfsI3tyYd1LtxW70phuVJMPYxYdARbj+SaPmrqqlBs5+DEyy8dw+b6mYHuKqbvczAT0ZTNfW+qGQjlkaTrfdTAQOR6+bYEQ5Il9yK+11EkOupPNrBP0js7e0OkaXt6b6evDTMDYMmiPimIpHY9w6kskey/1I9tZJ2qzsz7z/pX+VVmtadNlG/ymjAcIYFeyLxW7UimDIVwZX6rWKZYDuoXeMUggm1eg0S9/1GUU8RuAbEXYVa8jeyL1FZ0T+1WWI1ZEmD6LC+2afVfB0iGNx5I5/upfvGcKZvInwYHakuKE5d5wVwrXQ3QFvC90QBG8qi/Vz1JnuVsk+q/J7D/2sG2tpnlWEA5dE91vNA71si+jHE8KZuUGav43ruxDz9COacQNw3qfIjxUmIhxYg32/QsfhuiOOid5xFtyqpBifpkvd8auMxgf6pBfBjbzga+b7qnPlKyibDOYz2qLUsJMpbfhc1M05f9zygXfyoGTmAz+8jZ7C9F8NdRFpHtVCpq3Phe6d2UEmUB30k930dFgvpyvad6v9QWNHhTm1bfp35Ndb6a77vqcjz7xd0sh4cfvwvs19TmNThexPcvu+iK6efUv/mz/1H/btovcZm0KoR1w3JuSzJ9i/JbuzOV/XeUaTdLtsTxfc5lmz1H3OPN90T9t94Fyaky6lr91GrfRKyjPFL4zGE5PfitDqAibNSiw2Zwqr+FofpST74XSzbHmoHJMrZ5P9a3sFefLsWQ0lM96tuj9019lQYq6ovVVq7erEZzazvaHz+G1fvjrTss4X/rqvT5BdKUvqZjL1+7bz7c6pT1kdFH0JWdmTQnIn/SDErLsi8iz2gMpFH05gdOnYU+juoo1ElKu+XNzlZaE3XQ0uTpo7Iz5JjRHEozsomdurQ+IdQQ6sMrTbVIgzrTM9REi3hKKxKZXmLIrrRo0qasYwevj6A6b2n4pImQRlZaPWlnJIO0TEp7J7Uw0tKmk9wZ7ew5aXrYuZNkSluqD6C/8uZzyROadNp8SETmUhhP2hNptaTBlrZF16eoFZF2R3lkUWbJK0IhbZiIsLQ3u0jWRNCjKbueb+OHQITQY38aP+bFRusnefRcsxg7OXiSRkuDFGn+NLugWRJpIqXV1wdQmlSF9/HPNOHUCYuwSvusPIRzIq+F8X5+vNLzy0mGqPWinNLaKT9pC6UBE5mQJkfa7T3MW9oktz0p7NiLjPZYWIhQKO31e1PMTI1mN5z1JA2gCKpmSTSYkyZe9evLvIVvGE83Eix9LKTp9iP5UFr68EhjKowkr8iF6lj3peEVURIWBSVVpqyaEZCmWoRWsybSmJn2xIGENGQi2xpo6YPl5ZdptH+q0+2clVBZNbhLorZMpFFYqp5PUGO2+XAmTlMLq9mNw7EnzX2loTaleGpL0ihHimCw7tTuo4i9Jz/u0uRpNknkRO3tAAc6IslqF2ofKqfap9q/CIpw3MoBg2al1DY1qBUuCnuEZUomAfFlmRVPs0ZKO+9spXm/DvKd0yxZCNul6ljteDO1l2qXGmxuPJhlcNA7sp6ERu+b2rzISRLTOxidz/rNQxE1hyIK0hznkowcOlrANnaF5E+zTseMFlWDG71Hel8kowZfhlSQ+Ks8ym875ZfGWvKoPUnDLU23BlXCwpf4hbINx5GgOOq32GgQJU8KNa6a9VJdCwuRxb0sUxrf081sp0pf8USiEqlUSOPMlGTTrJTakzTRwk39gtqv2pAGkho8Cn8R/XC+IwGc0dAgR+1bmKlNZ7P+NQDSDEAJZ3a2BWSbNuGYJdEM5RXzTmkQJS25cNDAWO+m6l9pioRv5OBBcurdE7mTplfvhNp09skK05do9kd1pgF7NN8LaVB3sszSWquv82Qfkaz6Z14aJEsG1bmwEZ4aIDi1zCqnZllEboWh0t1BEqn3Txpq3VMa6ls0CNI9zdi470nje3yB4Qo5OMs0fZPixZDgSZkjgqw4Iv+a2VQ/HsfBm94X9cN7QvNNH5nIetK9PezP9A0Q8VVbUzm2Uc4Yvq+mXfBaM4p72K8a2Vg+yaXBr8rqRdxUt4qndqN3XvWpNiK81U/p/VT/pcGAZl1UJ96MV8j3NINtR/XhrKf9DCNtuOL5M57eVZU/hGlGsK2s5iBI8oZwhkftXv20sFQ7Vj+rOlNfqm/A9qAcM5Op/kvtUf2FtPgi65pFUd+j9q02KHnVlymsBpBHiadIsepXs24qn95rfb/UJjRDp/T0bmn2QO1XxDshRzOA+aYN6bnIt9JW/ar/i+SMQk6h+hnOWqrPomZf74JmXPVd1rX6Kc1eHCTOah8qpzT/znc2hm1Pf0tRpXJLcaSyFF28wXrKY19Zar4zUhJpZqKAuEh2M6tHWaRsURwNAiauCDGzDhp8lV2993wiYZ8MWAQs4R+wVffygl+5Xs0O8ZQh6urcpC3RtKg6KE1vanr88rV7tOetMXaRFyvvOmwmaf8ncwHZD1YwjQucFZD9aFnVPTPlKtvGkvLbUPqyEdd0csWNGmODWMwPr2wU9bcGGrIJVvqX2KHIBvgyzYiqbtUZG+LzTFfhZE98mh2Wwimtkst3TJhLlEdyKA2F1SyF0lYYySb7x4u8lh2j7HtlwiD7TpWjTHlLTspTcaPa2IWKZJgyMU+lIVvRK9dYPk7Vyi5TvyYP5nfpKrFgmhowGRmUF9M9T3MXyavynCcGSkM4VlJ22VVKHqWh+8pD9q0VTEdYyJ7VXBNHySGsVH6Fd8hZQ7xrjBmJ1gUoP92XfajSUvm03kAyqD5Ubt1XmZX3Zaar/ISj5JB9qaaIVUbZ6Sq/cl0zPX28FL+ccRRPWKk+1R5kB6p8lcd12pArD+GrUxiqbMKnrOqukU15lfFv1YfahcKpTIonW1vlrWvJpvuS53ZHPSpPlc+Un2VXe9OAQ2VQexP2WjegtnDhiqNMao/X79Y5BrOsB8mushRzWlt2rVeItWl7vK96MnkTC6Wl/JV2yZWncspsR2moHKaeWOf6AJazrcp2WPElo+rdkUaNwUpyqB71K3M1Z52Xq0y874yn+6pLyVjFvJWH7HBVVuFo2ilxU1sQdqqnKx3vhepP8dU29Q5IDuGu9GTjL9xU/3eqNbC/xzbmsNNX3rLhLWU4YSc5dU946NqJsfIzeDJdyaMwdyins/xq/7qndqu81eYVR+1d8SSb6knYKV2V52KnMinctTtsC5RBbVR2yg7ZGk06ylNpqF/QKSxO0/SgkuXX36qbSr6/qtNTF66ZNq301YbULwjL62xfspm+wDpV+YW9yizZhekFlYnXuqfwys/0X7yupIZTdSo5JYvKd4syKpwwkD31Zealdqr6c/ZDlbQnVznUNpXHOaZh2gLrSfFMX6D8WFfCUzJIXr17KpfSFRZKT8/VflQmyaZ7qie9T6c76l9tUW1Bazr0Dktmpefss1TX6h8UX/bbKqvk1bunenK2b7Ub4aN7utZzvdNq206sdM+0C8rs7E+dsqlswuYGZZcMwl7xDBYmjWqDrdYgqJ9S36J2I3mEoepH8WR3r77MgYWjnhRG91U3alP6Tqj8qhfJozar91d/K4zKdJny6H1W3+L4ltV24K3n6nvUV6ttOuQUXsJH8qocSlfvkPovgwX7cvV75XymMJLdiYvWz6j8iuN8n9RfC3PVvykn01EbusF4SlNy6l1XWpVMV21S74pkE7bOduHA8I65V0YcVC7hI9lU9853VuWXbMJQ5b5MLJR2dV2jqU99j1U+xVF59F3Tc7UVkX+Z9WggoZkJKUJkaqnZTin0moyNfxuo7LfHW4KAJfxvSUV2VQwtztECI9lwagpWU6QxtPWT9knaFpGiBmr/FU4LeMx0Ht/ytjZdw7zwreaa93jDcV/XXPDTEUbPTdyO57pWR6H7+jXxzD1H+q0dz00cBlCeSsv5t8yO2jrumbyVLs8ncnTka9I210/TMPcU3pl3R7qPTXqORUrKT3mYfDvCOq4dZVIazrycYdoZjv86ytQhj+J2nM7wTtkV3omVI7+fY2HKyHDO5xQJhPUJFk/L3HHPKbPK1FFux+/TepLsTpl17cTRWaanGEt2x3PJbMJ2lEVYGfn5N5N4gpWzTE/TULynGJh8O+EtTJxhnfI6wyi/J3lIzp+EVTydrZxiduLkxLiN8iq80nDe65yuZDJlouzOdPRc7e5n8TruOeV8gkWncjhlM/kxPP9Rpo68Fb8jDYcMkq0Dl5+USfGUhsIbjPmH41r3nt534qZn5lrheEo2Zzmc5TbvodIz9eSQzVknCst/Jg09N3IyrENOBxYmPeGpPPhMZTXPO64Vv3P5zTVvPklD4RSX8Zz5/igNPlcaznI446md85FDNv4+bQsdcjjjycRAefA05Xde876RzfztSEthlL4zP/5p8vipbE/j/aQcjKA8OmPROeyT8jHc0zw65O2E25N66lT/pu39KN6PyyxZhZHJryOe8tM9R3/ZUX7ec2LhlNPRFhz15Cy/sw6MzB3pKo8nWCgPns76cGLnlF1/P3nOa4dsjntPwioMT/4zz01auqd09dshv8qk+M54Jm0+d97Xr1OOJxgzPG+btJzxTN08Sfdpu9D75oyvPJ+k0amN6blTPuevA4sfy2bk7MDlp1g40zVYPOt9MmVyYOXIo53lfirbjzBk4VRuJ8ZKW8+ZhMHpR31Ihzym32MApW2w4K+wMTLz1/lNVrrmHi+MzB3xCUdHHo9putPG6zYzOL7KQcsZrvHazlnKOTS9k8mP1qtppqSJzjzs8XYgYAn/21GPXZZC2iFNyWtaUtOTslnXglppVTUQsIdFwCJgEbAIWAQsAu8uApoVkzMLmRbJTEjEX0pBzUI0N7e9u8C8RSW3hP8tqsyfFkUje7kn03SoFjNpUaFsiTUdLndu9rAIWAQsAhYBi4BFwCKgWQO55tbMQCTXpGgdghaCa22bTlkC2GNgI2AJ/8Cuvy6llx2mFh1qcZIWP2oBpmy17WERsAhYBLpCoI1T/devX0duTi6uX7tuzAK6Cv+sZ5cv0/NHSAh8d+3Cnr17sXf/fuzavQd7eV1UVERTBqs1fBZu9p5F4E0joD0OCs5fM+v95LVIi8G1DsMeAxsBS/gHdv09V3pp97UISx5H5EJMrr60mEj2eE1NwL37j3HjZjuu89Sv87x+ox23brWjvt5hX/ncDF7DAxYBt++040xxOy5caseDh29eptdQbJuFReCNIyAyXl5ejri4OPOr/uRFD4UtvXABPlu2YMZ8F8xfsQ4uq90xceZ8/PJ3f8A//+u/Ytu2bWhosBrDF8XUhrMIvG4EtLBXpsByaStvS/L2JA6hNQP2GJgIWMI/MOutS6m1cEeLcOS+TS+qXI9pca52eNQqqCsV7QiLasH23U3w3taEjVufnp5bmrB9TxPOFrdx0VWX2byGh4+RlduKVe5N2La7GZcua5HWa8jWZmEReMcREGlvbuaOvQ8emN8XhaOdL2hhYSHcNnhg/vK19E+egYKKasTnlWCZ+2b8adBX+N//3y/hs2kz06570WRtOIuAReA1IyBzYHm0k/toufpcS5eocvGrhcP2GJgIWMI/MOutS6lbuPpevnwXenPDFvowl19uvbxi+y38yS9oxaIVjfh6dD1GTWrAlLmOc/KcBkyc1YB5SxuQc4y2fFycL4KtUwq+ds7ANzWDBIB/P+OdV7jGRlBzR08Az1nY38L7DQ2P0dBIjwLPmdFXmEaOTRoZJjKmBT9MbcDilY04c45ehDriSB7J0cD8FL7z0VlmQmFkdsbrHM5eWwQsAr2HgMh+Scl5rFy1BjMXueJwXCbO3qrH8bLb8NkXiG/HTMaYafMwZPho7Ny1F3V1tb2XuU3JImAR6BMEtHO9NmvT5l3aO6WJH1x9f+0x8BCwhH/g1Vm3EjfSxeZJbpyl3XO1w2ARNzVxHiL8x/Ja+UFuIOFvwJ4DzSg604bC0204VdSGgsI2nD7bZkx+Hj16jMvl7UbbX3aZv+fakZzWisTUVlwobTfEXumKYN+799jES+KzuKRW5OXLBpguyTrIuMh55dXHyD7Whpj4FiQkU0bmd+u2YxCidDqHUR5KY9+hZoye7CD8ZzsI/yOaG5VebEdaRitiE1uRy/JcrWo38VtbHuPqNZkBtaH4fBtOsjwZ2a0mb0v6hbI9LAK9j4Ax4ykthc/mLVQYrCLZz0BhVS2OcwMon30BGDV5JqYtXIGV3ruoZJhBwr8HdbWW8Pd+TdgULQK9i4B892uDR206FsfNu7Rng3z022PgIWAJ/8Crsy4lbuF02wlq9OVOS7viZnCHy86r6w3hP+Eg/CPG1yOZxPoxbfLkC7i11XGKwOuoJIne5deMeYsbsd6rEes8mzDbpdFo3NdsaDSDgxYSbNn9HwpsxsLljVikk7MHs10asGVnEy5f0cDgsRlUeG5uwsyFjVi8qhEuDDOHYXbsbTL2+RoYlJa1Y9OOJkyd5yD4HpuasMC1Ad+MqYfrmiaUkMA/pB2/BgMr1jvyWrq6EQuXNcJnezPOlbRzoOIwV1rg6shj2dpGrPFoxDGWuanJqiUcNWv/twj0LgJnz5yBh9dGLF7thpBU7vR5/QFyuXneui27+f5OxJQFy+CxJwAeu49g5MTpJPy7LeHv3SqwqVkE+gQB+evPo6tO7Xg9ffeCmuAAACAASURBVF242b1Y+/bYY+AhYAn/wKuzLiXWdFtkWjHGugZx6/kEnORK+1Yng2dMJ+EXIZdJjxdt9mMTWozWPTquhRrzFmr72/FQ2n2S9bUk+Z9/U4/pJOF7DzYbjbvMfkTCDwc14yYX+CalthizIN0/EtICpbNwGc2E5tQz7WaUkcjv3N+MkTQfWunWhLTMVkTEtGIa0xwzpR7B4S1cNPzYyDByYj0mzW5AVCxnElJaIMI+ZEQ9lq9rwjlq7S8yLdc1jRg3rQH7/ZuRnN4Kd+8mTJjZYGS7dLkNezkr8OXwehKLemza3oQ4lqnyqjZD6hI6+9AiYBHoAQJnThdh1eo1mLVoGQISs40Zz8nKGrjv2G/I/oQ5SwzR3x4Uiw27DmPE+KmW8PcAZxvFIvAmENBGZnIAcpge/wJii3Ceu/02c9M7eww8BCzhH3h11qXEDdxI69jpSuwIPo4DXGGvbbQ7r6p/QvgXN2DQd/UYPaWBH+oGat4bMH1+A+YuEZEWAW9HBUmyu0+zGRiIVJdzAHD3bjsX+zZjGAcLW3Y2G7t6Xy7y/W5sPdw3NqGK5jTN1KTnn2wjaafZTiFNbo63Mt1GjJ/RgBQSdB0PHjw25kSK58Z4JwrasGvf03QfPqTtPe34/Q43m7RXrG/CKaYlkyKZ+IybWo+AkGZk5LTCm6T+u3ENnF1oYDqtOHCE6XBAotmGIpoqaV3Bs9YcdAmkfWgRsAh0i8Clixfh4elpbPYPxaTj9M2HDjOevQH4nsR+4rwl8NoXRLd+sdgWEAO3nf6W8HeLqg1gEeg/CGhevLqunmY9l+EXWUDSX4gSkn757bfHwELAEv6BVV/dSltF7zyB8UXwJeFPppee+7X1P4rTmfB/NbKeGvdGY45zIKDZkGt/muZkkkTX1j3Glcp2rCcZHzHeoT2XaY4W/opof09zIJ9tzcijqYyI/vBx9dhNLf7duz/uBDjhgFTa2o+fzsEEZxVkh69DpkCaWRhNrf9imvckpLRi4xamy3QCQ1ueLMSN0qJdDkpE+I8zr6AwDgBI7ocz/+XU/rv7NGEpNf7T59N0h2ZGWp8gu//hDLOBz25zjYA9LAIWgd5FoJ3TZcXF57DRZxNcVq2nZj+Hmv0GZBZfhdv2fXz/pmLS/KXGjGdHUBzJ/lFzWsLfu/VgU7MIvA4E6h41Gk9/K3ckGm89eWeu/kiR+DpksHm8OgKW8L86hv0qhfPlt7mrbgqmrg3H7rA83Ljz44VxTsI/i+R7xIR6muO0GlMXEfNmmuXpeTtt+nVcpvtOEf6RExpwgJp2EX4tzgsIpXkOzWU6E/5vqanfsafZLN5VXC3AraMWv7rmMVKp1Z9Ik5tZnEXQgmAdWi8gDzxKewlt+pOouVd63/1AU6HglifmN+HRHBRQoy/Cn5ffipAIkfl6jGd6WnAcE99qZhJkRpSR3YLSSzTpoemRBg5yMXqbi4ntYRGwCPQeAvLGc4Y2+2vXrcOcJSsRlHQM5++3Iqf0Oty37cPQEWMxce5SbKDNvjT7W49EW8Lfe/DblCwCrx2BRw3NOFlyDaEpZ7l49yIquTFXB0147bLYDHuOgCX8PceuX8a8zc0yknIvUst/GlmFV1DzgPYsnQ4Renm1mUHy/S3JdVhkC7Xg7cYW37n51g3a5cvkRvby67yaSJ4b4Ed7eSfhPyINPwm1NPKn6eFn136Z1Gi2gAtw6b2nju615RnnSFALsulHX1p3LdIdS7v7o3GtxruP8trs65gZ0F4A8hQkoq51BVrcK+8997k5mC8HEcNGNdCW32HSk5HVYmYLZi5qpMefVtTUPkYFZyJk3y87fcmugcA3LJsH07n9kxmHTlDYS4uAReAlEdCA/0LpBWzw9MIsF1ccik3nAl2a8ZQ5vPEMHzcZUxcup2Y/ENsDRfaPWsL/khjb4BaB/obAw3rO/J+8gm10BrI79ATOXLxpNfz9rZJeQB5L+F8ApIEUpJIbZeyPyKeWPxXBSaehAUDnw6nh16JdLYadR9t6LdwVyRZBlhmM/k7LaDEkfMMmavOfpeHn7MBGhpNdv8i9vPfIbEda/sOBWrTbaNJOTmuhy8zH8Oc92fDLn74W6crsRot6NfCIT24ludfi31az+FYLcGW7H8AFwIuYzleUU4t3i0tI6knutZ5Ai3a9tzZDMwBaS7DWowkxNBGqqGwzgxMNQIyG3xL+ztVvry0Cr4RAKcm+z5atNONxw5H4TJyqqkE2fXO7UbM/fNwUkv1l1OwfwTaSfZnxSLtvNfyvBLmNbBF44wjIpCeObjlX+SbBjf74885ak543Xik9EMAS/h6A1p+jVF6vxt7wE1i9K5luOYt+Rvjl/rLkQptZeLt4lezfm7Cc5jLO05WadGnq5dnmPLX1wREtdMnZZDzmaBEtFXxIpZcd3ZM9/T1q4e9XP0Z8UquZDTBuOblYVvb18thzjS47ZS4kon4ooMWY7yiMNP7yACQ7/lt3HDvoVtFvvwYGxqUmBwabd3DXXy4QVjh53rlS+dhsyKW9Arb4NpvBg9KRDb/ch2pAcL+6nQOIFrripHwcWFRzBsAeFgGLwKsjUFRUiLXr3bBg+RqEpDhcb0qzv9ZnF74eOc643pQZz/bAmB+RfUv4Xx17m4JF4E0iIDec2oArjCY9UenFKLt6/4np75uUy+b9cghYwv9yePX70HerHyGNq+n1YspbT93DH5v0iLDL5aZ87Jde4s6YJPXFF56e5847ruVuUz7vb9K0Ru4579IW3undUwRf92Q+IzKvQ5t0yX4+jSY3ctMpW32Fe3LwUmloI6xELtDVDILylyzOQ24zZcojjz2pfH76LDfMopxaPKzBQD138NVhNujifW3ilUDXnccZ/jrderZRFqVxh/mUccOwm7eeburliGn/twhYBHqCwIXzF7COZH/6/MU4HJthFujmV9yH165DNA2chPFzFsNrr9Mbz1PNvtXw9wRtG8ci0L8QePCISj9q+NdyfaDH/nTkn6OGv8265uxftdS9NJbwd4/RgApRxcU0h+k2a+PBTERyJH635tFrlp8Le7tZzdOudbsaeXRxdJeGM2pnl6POe/a39xGQ7TYdNNnjHUTg0qVLZgfd+a6r4U+yf5o2+1klV7HB9wC+nzANkxa4wqMLsm81/O9go7FFfqsQeMhFu5ncxHPrkVzsDDlObf8N2vBbwj/QKtkS/oFWY93IW3WrFke4OYb3oSxEcQMuafztYRF4FQQauZHB+ZLzuFbFTdxkE/aCRxunWy7ST3tERAQSExO5OPz2C8a0wfoDAi1c8FNSUgJPb2/uju1Gbzy5Dteb1O6t3bybe3FMpOtNkX0u0O3wxuPU6P/017rl7A81amWwCPQMgYbmVmPSE5FajKOZJSirum8X7fYMyjcayxL+Nwp/72curzwFxdeQeqKMK+lvQCNze1gEXgWBOrpdio6ORn5+PjdD40KOFzg0SMjJycHcuXPx+9//HnPmzKHf9uJuY2omQQMFuX5saZFnqEfv1Kky94dD+JeWlsJ12TIM+W4E1m/di/gTxYjMKoT7dj98PWoCF+iugKfIPm32f0rwf/r3yxD+qqoqFBQUoLy8nHXfYcfXH0CxMlgE3lEEarloNyr9PJZsiod88ecWVaDVmvQMuNZgCf+Aq7KuBb587R4XzOVgsU8sDkYX4Oa9B11HsE8tAt0g0MxFExUVFbhz544h4t0E507LzUhPS8fYsWPxxz/+EbNmzUJ8fDw9Md3vLirXZzSbWYHTp8/g+PETSE5Oe6fO48fzcP36dTPo6RasPgxQX1+PyMgofPXNcLrKnYCFazyxevMuzHBZiSHDf8D42S5w33X4mQt0f0r29ffLEH4NLkePHo358+YjKioKtbU/3kukD4ttk7YIWASegcCjxmbkck3gvvB8HDx6EsVcrO/cr+cZwe2tfoqAJfz9tGJ6KpZMeLQFdgQX7R4/U/GzRbs9TdfGe7cRkMbX2PF3A4PCZWdlY/So0fj8889x4MABSGMrjf+LxH/woA7+/ofh4rIKy5Zvh7tHIDw8Q3mGvPbTk/nqfB15e3qFws09gBr1LfD23oaiotNmhqMbuPvscU1NDUJCw+GyxgNb6V4zMPUkQjIKsY7a/WFjJsJl/SZsORzVrSmPk/y/DOE3m3qtXYsvBn2Bjz/+GFu3bsWNGzf6rKw2YYuARaBrBAzhP1OJvXT5fSD6FIov37aEv2vI+uVTS/j7ZbX0XKg79x8h+dgluuQ8jWxuvFX7k423ep6yjWkR6B4BmWHMnjMbv/rVr7B502aIOL7MUXHlChYtWoZ5C7Zhn18xN2+rfHIG8DowuMqcuu78rLvrgOCrCAypgn5fJGxQyDUEh143p64V72Xz7C6fzs8dZbuKbTuOY5HLNnht3GY0/S+DXW+Gram+j/DIo/DZG4CjJy7ieOVD5F9vwL7oNIyZNgfLPLdjMwn/thcw53lZDb/MxjTLIe3+qFGj8Lvf/Q47duzA3bt3e7OINi2LgEXgBRGopbe/SK4JXOQdC9ctCYZbWJOeFwSvHwWzhL8fVUZviFLBjbeMH37fDj/893688VZv5GHTsAg8D4HAwED8+te/xqRJk3Dp4qXnBXvm/dbWFuRkZ9PufzW2++Yh5ziQns0zC8jMAbL5d3J6KxKTm5CS0W7u61lXZwbjZx/jzs8Ml5TajGS6jE17ThyFzcp1nLEJDxAUqsFFBXeHrjHx9ayrvF71mcqYmt6OLdtyMIcYlHCh9Js6aqqrEREVg00k/DEFZci/1oCTt1rgF5OBH6bPhavHVmz2j+wTwu8sc/tjbuqXkYEhQ4Zg0KBBSEhIeCGTMmd8+2sRsAj0DgL1XAuozbb8ok7SU1chSsqthr93kH29qVjC/3rx7vPcbnFn3cTcUhyJK+RW2JdR88Aueutz0G0GBoGrV6+aRboyw0hKSqJJSstLIdPQUI/YuDgsW7EdB/wvIJMEOzmtlWcLjsbWYMfOPMydvwMrVoUYMp5C8p+Szj0bMtu5b0M7rxW21fydlvmY99qQlNJELf1VrFwVjAULd2Pn7gLuA9H0JF2lbU7GTWM6kUfvYr17NMZPXIfvhi/Ed9/Nx7jxq7F6bTiCOUOgsMpL6RuCz0FCWiaJOu/pmWTQs87PFU7PJY/uS2ZnvrpWvnom2UX6/Y+UYeGijThztvtFzi8F8EsEdhJ+H+6aKw1/3tVHKLjRhP1H018b4Ze4WtMRHh5uCP+CBQvMQt6XKIYNahGwCPQCAiL82YUV8A3JM2Y9Z8tuWpOeXsD1dSdhCf/rRryP89Mi3djM81ywm4+0E5dI+Ov7OEebvEXAgUBoaKjRxq5evRoPHrz8YnER/ri4BJJzX0P4M6TxJlGOia+jH/h0fEvy/T//5y/w2V/GGeIukpyQ3EgNfC1/65FCwizCHZ/0iHFquVv0QxwOKKN50E78n19+gl/84kO4LDngCNuJdDtJfFTMPSxfGUCvQl/id78fjKHDZpH0L8CHH36D9z8YShK+hztD10EzAZoBCAm/hqCQSkRE30FiKmcdmHdCEgctfCaZwqNucWByFZF6nsKt6SmP5IpnGMmqfDX4cMqqGQgn4V/k4k3Cf67XmpYGX6qTF/Wy1F8IvwCQKc+a1WvMAnA/P78XWgvSa8DZhCwCFgEYk570EmPSs2xbInLkpYfrtewxsBCwhH9g1Ve30l6qvIuNfhmYtT4Cu7lBxvU7dd3GsQEsAq+KgMikq6srBg8ejNTU1B6ZXvyU8Iv8SiMeFnnDaOjHjF2B//iPj/CXv4yltv+Y0eSHRd6C96Y0bN2ew1mAapLqBzSJyYKHVwJt7q9gy9YsjJ+wFh9+9A3+8IfBmE8tvwYEqZ0Iv4i6CPimLRn482c/4JNPR8N9Q2zHQKLB3P9+5CJMmLSWstw0swCe3olmADBvvi8X2h7Gzj0nDdE/EngZGzzjsWZdhJF57jxfLF56AHv3n8YuhnH3iOXahDNmICOCHxBcDi/vJOw7cJa7Rjciq0PD39uE/9q1a2Y/hCh63iksLDSLYJsan+9itT8Rfu39kJyUjPfeew8TJkywC3hf9WW18S0CL4lAQxN3vi+9ifCUc4jKOG/88LdrM0Z7DCgELOEfUNXVvbAy4SkorkLa8Us4a/3wdw+YDdErCJw+fRrDhw/HuHHjzAZdPUn0p4RfGn6RcZF4meXs9zuLL7+cTO8/42njf8zY90vDPnvONnz77Txs3pJu7o8Y4UKSvwZHAssRFnETBw+XGlOgz/7yA+Yv2IX4xEc/MqvRoEKzCHr2u99/Adflhw3Z14DDaPPjH8DvUDH27C+iVv86Vq4JwbBv5mDMDyswbfpGDPlqOoaPWGRIvadXohk0fPrn0Zg1ewvNgVbhj38ajukzvM0AYchX08x9h8b/kTEVGv79AjPAkPmR1gnIpKc3Cb+8I+Xl5Zn6+Y9f/AeGDB6CJUuWIDAoEBcuXMCjR49+NkDrT4RfbUleeqZMmWK0/JGRkVbL35MXzMaxCPQQgYf1nMEkp/D0S8cm/yycLLlG18FWw99DON9YNEv43xj0fZPxpavU8B/IwGy3COwJPY4bd6wP675B2qbqRECEMiQkBJ999hnW0p2iNs7qyfEzwk/TGaetuyHlNJMZ9u0c/OXzcfRmk8vFt48NeffenIa/DBqHr7+eyXMGhvJ0cz+KuISHRpOeQtMfkXTFm0/znp8SftnYS3M/fuJavPf+UKOtd9jpP7XLN+sBqJE/4H8e33w3j+R5AWU4Rg39Fcyeuw2/+e3nxnTIZYkffvmrjzF4yFQ+o80r1x0MHjKFMs3EEtdD+JL3vxw8hYuBryCUg4cxPyzHN9/MNgMVlbE7kx65PdVGaJWVlbh8+bKxaZdnpOed2j9B4bSYWp6T/uZv/gZ/+7d/i//6f/1X/MM//gPef/99TJs2zTzXrsgPHzoW+Vd3LNp90zb8znakcvsf8scfP/ojvTgtsh57nMDYX4vAa0DgYX0z0uju2/tQFt305qLwAvcKsSY9rwH53s3CEv7exfONp1ZxvRp7wvKwirvhBXLh7i3rpeeN18nbLsCDugdYs2aN8Zl+8ODBHhf3+YS/1SyQ1cLdr0mOnYTfQcpBbXw1Fi3ZRxv9D/Av//JrQ8C1+Nb5XL8rVwfT9n/scwm/yPdYLs794MNh1NTT5IbechwLaR0LcUXENTDw3Z2PT/88BpOnbEB0TLXxJOS1MRG//e0gjBy1hHlvpenQl5g4aR2O58PY748Zs8yQ/I2bUjF95mYT38MzzpghSaYZM70RGnHdpO8k/C604T/7DBt+bYilBdGzZ8/mguLvMHLkyC5PbWClMJ988gn+7u/+Dn/1V3/1s/M//af/hH/6p38yszNaIFt9vxoi/JHRsegvhF+N6kr5Ffww5gfO8nwJ+erXQNMeFgGLQN8j0NjUijOXbiIqrQQxWRdQfu0+rElP3+Pe2zlYwt/biL7h9O7QS0/SsVIExhch62Q5F+02vmGJbPZvOwJVV6swcfxEDPp8EFJSUnpc3OcT/jaHaU1cHTX8c/H5oPHGY48IuExuNAuwzi0a//zPv8L/+Md/gctiP7rubCRhd3jSkeZ81ZpQM1BYuGj3k2edZw+0YHfKNC/6fP8CPpvSzWBB5DuVcdM6Fg4fob291gt8Shv/6TM2mjUDMsHRguL33vvKLPCdOXszF/kOw9Tpnsg9wZkDEvnRY7S2YSp27DqBjd4pNAeabTz/TCfR//iTEcYrUBIX/Wpg0h3hb2hoMK4qly5dauzZp06diq7O6dOnG1MYra34+7//+2eS/f/23/4bFyr/3nhYiomJQU11DUn//X5H+KXl14Jw2fLLRefLeoHqccO0ES0C7zgC0vCn5JXBi+sDN/lnO0x6rIZ/wLUKS/gHXJV1LXB51T1sO5INF++jOBiVj5t37aLdrhGzT18VAdmBfzXkK4wYMQLnz/fcd/zzCH8SfefrDI24YUxlPvlkpCHeIsmyhd9DjbxcaP7pT9/jww+GYfj3C7lR03HjASclvcUs0l3iehB/+ngE7ec3UzN/3zxLTqNffg4WpMmXlx253nyfJj2jRrvC72AJ82xCEp9r8e+SpYeonffmrrsJ1NZPxlguINa6ArnVXLk6hOYyf6ZWfz3mLvDFH94bjEmcAcjJcxB+af4HfTHJmO0Eh1Zi6jQP/OY3n3G9wCAMoyeg/QfOdQxcnrrldFn8bA2/SK8WSGtjKpnryLSnq1OLda9wM7Pg4GD88pe/xF//9V/jP//n/4z//t//u9H6u7i44NChQ8jNzYXcqiptac5F+OWHvz9p+NVO/fb7mZmk3bt3PzE/etX2a+NbBCwCXSNQ39iMXHrm2ROah/2RBThHt5zWpKdrzPrjU0v4+2OtvIJM5VX34RuUy93w4nA45iRNel7ePeIrZG+jvoMI5OfnG/IoW/A7d+70GIGfEX5p2EnGo0jQt9MWftmKAONp51e//jMmTXbHZnrg2UuPN3Pm7SAJHIHlywOoyQ/B51+MxwT60T8SdNm495TXnO+4MPaXv/qUC2ynYu36SBw4VGLcYyp9LQzWLEAAw08iaf/Vrz8zC3IVbgMJ/jQuuFX6WgwsTzvTqN0fOnSmmTXYuj3XhP3oj98ZTf2K1UH44IOvTRwH4b+BHzg4+GroDEP4kzjIWLEyCP/6r7/B//pf/8csFI6mdyHnTMUTDf9zCH9PwT116pTZDO2jjz7iYGUs9u3bZ2z7n5def1u065RT3no0sFy1ahXu3bvnvG1/LQIWgT5EoLm1DTIXzjtzFfnnqnDr7gPrh78P8e6rpC3h7ytk31C6t2nSk3zsIoITipBzqhy11qTnDdXEu5GtNM7x8fE0Y/kQCxcuRG1tzxeJi/DHxsbTneVO41lHG2+JiAdRK76APvC1WPZjmtPIbaY05tNn+BgXmHK1KdeX8tgTRt/3Cxfvo0nNRmP24+4Rh7HjVtKcZzzNZ0YbDzrS4Mttp/zfa9MraflF+qXxl7Z94iQ3hp/APCaaGYXBQ6ZhytQN2MVNu+TNRxuATeSAYzRt8yfQ5ecoavAXuexDcFiVeabBhtxyytxHpkLL6PVH7js1yJCJ0M7dJ0laF+MbehbauafA7AAscx7JoTiHA8pplvRsDX9PWpU09rdu3UJmZiaOHTtmrrtbWN1XhN9912GMGD+Vi5l3oa4HbaWoqAgzZ87E/PnzrXvOnjQGG8ci0AMEHjzibGruRazfnUpPPRnGE2Bbm11D0wMo32gUS/jfKPy9n3nJ5VtYvjUeIxf5w5veeq7eqOn9TGyKFoEOBOQjPSQ4xHh7WbZsWY823HKCKcIfH59ATT132j103pBfEWFtdnXw8AX47so3i2Z9Sbx37OLU8sFzhmQflvtNetmR9ly71cp15qEjlxwbX4VcxZ59RcZ+XvG27zyBXXtPISTsWofJT7sh+w4tf7sh3xo4bKPmXj7z3eiPX954gkIqjO2/THji6Nbz4OGL8KHffq+NSdhJucIibpm40tYfDrhk3IFqMKGNtZRXID32qBwqjwYachUqjz8xdPmpNB2nw6RHsi9y2chFu723064IvmzeX3Sha58Q/sAYuO30x/fjpvSY8F+6RGzopUcuOq/QVMkeFgGLQN8j0NjcYsx4YjJLEJdzHuXU9lOPYI8BhoAl/AOswroT935tPY5xC+x4rqQ/RV+5D+g/1x4Wgb5CQIT/sP9hfPD+B1i1etUr2VW3tbVC5kELFqw2m1GJLMu23vzqmqc2q9L5s/vUjjufOWz+GYb2/U/iyh5f8TrSed6vYyEvd8zlot9YknFp9HXtuM+dcTvs+pWWY0fdGrO7rp478u8k43NkdoRj+I44TlmURtTRe3D3jKWGfz1KSy/2VbV1m25vE/7tJPubDoZjwdqNGDF2AgICAtHY+PIOBbReQQuWf/jhh1daL9ItADaARcAi8ASBR7Thzyav2MHNPPdE5OMsPfZodtceAwsBS/gHVn11K61s62Izz+MgF9ak5V1CdV1Dt3FsAItATxEQ4ff39zca/lcl/JLh/v172LvXjxtDbcHqNWFwo0mOO+3oN3gmYgM3tdrglfTj345nCuPuEW9O2d0rvCMerzvF83DG7xTeGc/5q/gK57kx2Zwmzk/CK32PTs+f5N8p787pKc2fhnHK6AznzrUGK1YFYMFCD3j7bDGEvzvTm57WW3fxepPwG7J/iGR/jRfNeaZwrYM7SkpKXni2obOsWlgswj9q1CjOgJzt/MheWwQsAn2EQB1NemKzSrFyRxLW70lD3tmrdtFuH2Hdl8lawt+X6L6BtLWwZm94Htb4JtE1ZyFu37eLdt9ANbwzWf6I8HMhpXPjplcBoKKCm1KFRmPz5v3YtNmfO+gefifOTZsP0buQPz3qhCEnJwdVVVchfN/E0VuEf0dQLLYcicZ8kv1vR4+niZSH8aHf03JZwv8mWoPN811HoL6xBQUl1xEQfxqhyWdxsfKu9cM/ABuFJfwDsNK6ElkmPXKfFZd93vjKtSY9XaFln70qAn1B+CWTBg5yKVlefgXltNV+J06W9caNm2YdxKNHjyCf+29q2rw3CL9Dsx+Bhet8MHrSdM7YrO3SMxCrvdujqqoKrq6uVsPfLVI2gEWg9xB42NCMLJn0BNOkJ1wmPbesl57eg/e1pWQJ/2uD+vVkdP1OHSJSzmIXbe0Sc0qhAYA9LAJ9hUBfEf6+ktem+2IIvCrh3xZwFD4HwzB/lSdGjJuMdevdOHgrf7HMuwhlCX8X4NhHFoE+QqDmYQPCyCvmboiGi08cN/W8gtY2a8PfR3D3WbKW8PcZtG8m4YobNOkJc5j0BHG33dv3Hr4ZQWyu7wQClvC/ndX8KoRfmv3N/pGYv9oLI0n2165bb+zte2O2whL+t7O92VL1bwQaaNJTeP46ghLPIDz1HC5dvWdNevp3lT1TOkv4UjVUhwAAIABJREFUnwnLwL35kF55zl++jZPcHKOs8h4amt6MDfDARdBK/jIIWML/MmgNnLA9JfzS7G/2j8JSz+3GZn+dmxvK6EqztxYfW8I/cNqQlfTtQaDmQYOx3Z/tHoVFG2OQebLcavgHYPVawj8AK60rkcuvaafdY3DdHA//o9pp12r4u8LLPns1BCzhfzX8+mvsnhB+afZ9DoQZbzwjubmWvPFcuHChV4toCX+vwmkTswi8EALaeCuJG3q6703HxoOZ3HjrGr30PH6huDZQ/0HAEv7+Uxe9IsnVW7U4HHMKG7kbXiSn3u5WW8LfK8DaRJ6JgCX8z4RlwN98WcK/PTDW+Nmfv5o2+2MnYc3adcb1Zm8DYQl/byNq07MIdI+A/PAfP3MV++nu2z+mECXlty3h7x62fhfCEv5+VyWvJtDt6kdmC+yAuCIurCmHpuLsYRHoKwQs4e8rZN9suiL8kdG0xd8XiJj8S8i7+ggFN5qw/2g6fpg+F64eW42dvrT6TtebC9ZwU61xk7Bhg2efbYplCf+bbRc293cTgZqHjbTdL8aCjUeN9UB2oV20OxBbgiX8A7HWupBZi2k2HsjCLNra7eHi3et3rB/+LuCyj14RAUv4XxHAfhpdhD8sPBJuW3YjPOccjlc+fCbh3xEcC2+a8SzkDrpjJk83mv2yS2WmVE3Nzbh2/QYu8e+ysssov3wZFXSxqt1ya2tre7TxliX8/bTBWLHeagS08VY8vf6t3ZUCj33pyOcawTa70+6Aq3NL+AdclXUt8LXbtQhOPM3NbnIQk1mCezXWLWfXiNmnr4KAJfyvgl7/jVtTU4PAwCAsWL4O/onHcaziAU7e7Kzh34ath6OwyT8Cc1ZswPdjJ2J9xwJdlUrtIjs7G75bvLHZbQV8vdZim8dquC6chzlzZiM+IQHNHBC87GEJ/8siZsNbBF4dgUZ66Sm6cB1h3HQrKrUEZVfvWz/8rw7ra0/BEv7XDnnfZugg/GdI+HNJ+M9bwt+3cL/zqVvC/3Y2gZqaagSHhGK5mw+C0k45NPydCP8yrx30sx+ORW6bMHL8ZLi7b0BpaakBo/3xYxzPy8O6FYux0WUc4rcvRKLvYuxcNhEf/fp/43/+0z/Cx8cH2lzsZQ9L+F8WMRveIvDqCFTTNDggvhBTVodillskUvPK0NZq/fC/OrKvNwVL+F8v3n2eWym3vHbfm4apa8JpW5sLDQDsYRHoKwQs4e8rZN9sujLpCY88Cq+dhxCdd8Fhw28IfwbGzpiHxW6bsWj9JowYPwVuJPtlFy8agVuo2T927BhWLl2IzUsm4FzwOjzM2YmioNVYMWkw/vLb/wd/+MX/jR1bN6HuwcubG1rC/2bbhc393URA7r1PX7yByLRixGSU4HKV1fAPxJZgCf9ArLUuZL59X4t2S3EkrpC+ci/bRbtdYGUfvToClvC/Oob9MQXHot1YbNkfjPhT5cZ+v/BOGw4l5OL7iTMwfMJ0s0DXfYMHXW+eN0VoampGRnoa1i6dT7I/HsWh69B+8gAuRXliy6KRmPv9H7F24meYPPRD7Nu1HbV1dS9ddEv4XxoyG8Ei8MoIPGpohhbq+gYfx97wfJy9dMva8L8yqq8/AUv4Xz/mfZpj1c0aHImlW076yo1Ko1vOGuuWs08Bf8cTt4T/7WwAIvzBoeFY5bUdfjHZOFpQhoTTFfDcF4z3PvkLPvn8C2r23XGxw4ynta0dGZmZWLV4Lna4jsfFKHfg9CGSfnesnjYUU4a+hy1zvkLQqtGY8e1H8N3qg5qal599tIT/7WxvtlT9GwFt6JlyogxedPe96XAOTnHXXbtot3/X2bOks4T/WagM4Htl9NKzxT/LuM/yi8zHzbsvP20+gItvRX/NCFjC/5oBf03ZPXz4ELFx8Vi8Yg1cacfvtnUvPH0PYMbCZfjLoMFYtWo1LnaY8TQ0NiE9IxOrXRdiy9LxuBLrhceF1OxHe2L99G/ww6DfwmvmEMR4TkTw6tEYO+g38PF0w7371S9dGkv4XxoyG8Ei8MoINDW34fyVO0jm5lupJP5Xb9ZCa3XsMbAQsIR/YNVXt9LerXmE9PwyhKecxbEzlah72NRtHBvAItBTBCzh7yly/Tue6rWMbjRDw8Kxb58f9u33g9+Bg9i5azcOHz5CN5sdrjebmpCalooVLnPgs2gcLkZ7oPHEXpRGeWDj3OGY+vX7HWR/MhJ8piCQGv5xHABs2uiOe5xFeNnDEv6XRcyGtwi8OgKPGlpw/OxVHIgqMBt7nufGW+12p91XB/Y1p2AJ/2sGvK+zK79+H74hx7B8a4Ix7bl1z5r09DXm73L6vU34H1Nr1EQS2dbWZmDV342NjXhEjfPDhw/e0vMhGhrq+QF1lLm/tCdhr/ptoftMudDUb0tL85O60QLdtLQ0mvHMhu8yafY3oiFvL86Frsf6GcMwY9iH2DpvGGI3Osh+vPdkQ/inffNH7N6xxdrw95eKtnJYBLpB4AH98Mszz8YDGdh8JNth0kMzPnsMLAQs4R9Y9dWttJU3q80o3G1PGkKTzuAOF/HawyLQVwj0NuFvaWnByYICVF29SrLZRrORS4gIj8LOnXuxfftunnveutPXdw993gejqKiIrioHxgBdA4DsnBwsd5mL7a4TUBHjhbaT+3EhYgPW0GZ//Je/c2j2vSYhcdMUiOw7Cf/0b/+EvTu3WcLfVy+lTdci0MsINLe0ofJGLQqKr+EU/fHfrn7Uo43zelksm9xLImAJ/0sC1t+D36l+aGzsRPZziipQ+6Cxv4ts5RvACPQ24Zdm//z58zh79ixiYxOxbv1WLFu+E2vXBXJjp1C4beg43fn7Mqcznn5fNF5P4rxo2s5wLNO69UFYsWov7eI3IyAwxOxE+/hx/9WePaqvR3JyMs145tJmfwJKIzfQZv+g+fWY/R2mDXsfG2cMRtSGCYbsx1HD/1PCv8d3qyX8A/i9t6K/WwjINDg+uxRrdibDfV8aTpy7ahftDsAmYAn/AKy0rkSuoEnP7rA8rNyRxI0yinDrnl202xVe9tmrIdDbhF/S3Lt3F/6HA7F4sTcJfjT2HyhBaMQtRMXU4GhcHWLiHyA6tgYRUXcRFnGbz26bX107zyf3Iu8g8mg1w9cxruLVIjL6PsIifx7PxOH9CD5XOIV/Eueo4tx5bn7OfHvyGxJ+E4cOl8F7Uypcl+/ATl8/XLlSThOf/kf6G+l6M4Vkf+WiWdi5fCIukuy30fXmmRCHGc+sb2TGM9Qs0E3ymQqRfUv4X+0ds7EtAm8agXrutHuy5BqCyCnCqEy8WHHX2vC/6UrpQf6W8PcAtP4cxbnT7taAXMRm2Z12+3NdvQ2y9Tbhlx17LjduWuSyAavWRCA+qQGZOUBKeiuCQq9ix848bNuRi4CgciSlNCEjG0jPAtIydT7uOB33zDPejzp6D7v3FmLztizsP8iNY+LrTBxHPGecxyYtxYlLfIiDhy9g2/ZcxsnmgOMcYuJqkcF8nsZxXOvvn+bvvNf5N61Dxp+Gldy6pzKmZrRjK/OcM3cDF8YGorb25d1W9mWbkjeelNQ0LFtIsr9sAq7GeQFFB3E2xA3LJ33Fxbi/w9a5QxHrNZma/adk3xL+vqwVm7ZFoO8RaGxuRemVu0jLu4TMgnJU3aqzJj19D3uv52AJf69D+mYTvEPbOrnNCks+i1z6za59aE163myNvN259zbhr75/H+7unli6dB9CqfkWEY6Jr4XPlgxMnOyOb76dh6+HzcYPY1dgg0ccYhMePCH6GhToFPEXOY9PqifRP4WZszfju+/mY+jXM/Hd8AVY6nqIswI3kJzmCJ+c1mLiJac143BgGRYs2ovvR7jgm2/m8JyL779fhHnzfLFnXxESUxqRlvV0kJCW2Y70joFG5/wdxP5puNQMXbcb2UTslafO1Iy2jnttJP6PcZQzC+4bjmLZsg3c0Kq033xUHz56hKTkJKykGc/mxWNRyQW6bSf9cD58A9xmfoMJg/8AL5rxxHhNpPnOlCeafavhf7vff1u6dwOBOi7aTcy9CPc9qfA+kImTtOW3fvgHXt1bwj/w6qxLiS9X3cPWw9lY5B0Dv8gC3Lj78rtZdpmBfWgR6IRAbxJ+eYW5QPv9ObMX07wlBVkk+yLO/gFlmDBpHb74cjJmztlCDfh2vP/BUHz2l7FG4x8cVkUTnGoSaJqbkMRHRN9BCO8dOlyKGTO98cc/Dcf48WvgsuQABn0xCR9+9A08NyaZAYGTfOs3MKQC02b44Le/+8IMKha67OXgwJ+EfzHjfIsp0zxpVnSPhB80E7pDzf9Z5n/c/CpPEXgNMkLCryEwuIJyX8KuPSexz+8MwqPuUK5bnKWoNLKmpLeZ8BqwhIRXMb27SEptNuU96H+eMxwbkZWVy4XLzZ3QfjOXDfSaFB8Xh+ULpmMXzXhE9huO76E3HjdD9md9+yG20RtPDDX7stV3kvzOv9aG/83Unc3VItAbCDQ0taCQm22FJp5GBF1+X6q0Jj29gevrTsMS/teNeB/nJ688KcfLEJx4hlthV6DGavj7GPF3O/neJPxasFtADz1z5i7HNt9c5Bx3mLkEBF/B8pUB8PCKpx39TWPaM/z7hfjDHwZDpNxl8T6sWh1sSH4EibXr8sOYPXcbtmzNxrIVR7jo9wj8j1xCfOIjkn4/fPjhMCxZdoj2+TWGYIvsyzxo7fpIfPDhNxg1xhX7SOYTSN7jE+tx4BAJOPNYsHCXyV+zA0uX+ZsZgEGDJmDYN7Mpgx9nJK7jMAcns+duxQ/jVmLyFHcMHjIVX301jXIEYOWaYEznAMTLO9Fo9hOZ5y7OQMydtwM+m9PNYIEc3ww8lizdgsTEFOOu8022MLlIzcjIwGq63ty7ciJuJm1G04l9KAxaSzOewdxB9w/YNv9rknx54vm5Zt9J+i3hf5O1aPO2CLwaAg1NrThdesOQ/aPpxbjMDT7txluvhumbiG0J/5tAvQ/zvMlFukczSrCP2n0R//u1DX2Ym036XUegtwl/fv4Jms+swHbfY8g6hg6teQM16/eNdj0opJLa/zQMJon+cvBkbPCMMwT7L5+Px8pVQVizLsLcHzdhjdGyS/MvYi+SLpOcCZPW49M/jzGDBw0AnHb/0Ux/BrX7v/v9l3D3TDCzBTIn0gyDNPfS7IdH3aKW/i5WrA6C8hs1eqkh/jITev/9YRx0hMKHC28/+ug7/OI/PsKs2Vswb74vPh803hD/seNX4eNPR2Dc+NVmnYBMlRa67MPgwVOx3v0okpiP8gxkGZe4bkNCQhIa6t+cW91HNONJSUnF0vmzjOvNK8b15gEUh7kb15uTh/wBPrOG4KgnXW92WqDrJPmdfy3hf9d7Clv+gYxALb30RKYVY7FPDFZsi0dO4RW0Wj/8A65KLeEfcFXWtcCVN6qxNyIfa3alIIjTb7fvDwy/3l2Xyj7trwj0NuEvKMj/GeE3NvlcTCtN/5SpHiTlX+BXv/4zZs7yQXjkLaMl/5Y2+p98OgoffzICw0csxOatWTSRaYI05vHJ9STQ/g4i/osP8e13c6nxv0jznzba7st2HhwQXDdE/IMPhmHP/tNmAa3s60X2ZZuvxbwOMn7VEH2RdO9N6bTpf0BPQjGcbRiC4bT1l+ZfgwbNFPhxgbBMeGZxDcHHH480A4Cvvp5hBgsHaLajhcfffjfPrBHQouJU2vA78qh444S/ubkF2dnZGDNqFD577z+wfMIXOLxuMna5joUrr2fTjGfLnKE46jGxW7Iv4m8Jf399g61cFoHuEWikhr+47Dbi6AgkMbcUFderrZee7mHrdyEs4e93VfJqAsmkJ+n4JQQlnEYWR+HWD/+r4Wljd43A6yD8It46w2nOs9Y9CmNpLvMRbepFlv0OlZiFtB5eCdSqf4h//udfknQfMuY4Tu19Agm/BgBTaYP/50/H0I5/Ak1o0p7Y8Ivwh0XewPgJa/HeB18bjz7OPJ2EX950ZPojoj7kq+kY88NKEvYKnDgJI4NMe7SmYPbcHST7w4y5j/JNSGnAcpoV/ZmzCpp9WLhoDzX+E4ypkQYKn3wy0iwS1gyGc2CjtQRvWsNfV/cQIUEBmDlqMH3qf4HQdWMRuWEi3GcMwejPfgWPaV8asp/0E288nbX6na8t4e/6PbJPLQL9GYGH9c3IKLiCrUdysDPouDHvaWt/3J9FtrI9A4F+Q/jlc1o2vP3R9/QzcOu3t67frkVw0mnILWdM5tvvllPtRbuzytZYpwhod0djYyOqq6vx4MED0+Y6h1d6z2qDWlDqvK/r57VVZ3yFeReOvib80q7LraYWssrsRmY+8pQj8v6rX31qbPJlbiP7+1/+8mP82//+nSHVR40pTy3t70ugRb3S0OfkAW4esWYWQDMFWg8gsi+ifZRuN2X3/5vffo7Va8PM7IDWEGiGQLMAR6iN37mnADt3F+CroTMwZswKEv4rhvDvP1RstPYi/HNoj68Fvt9wMBJHk6HYhDozAPnzZ2M4ExBrXIqOHLUYQ4fNwoiRLhwIjDYmSqkZrWZQIzn7A+G/X12D6LBgBHrMxuXgJahPdUNTugfSt8/A/O8/wi6X4cb9ZkIXdvuW8L8LPYAt47uAgEx6YrIuYPn2BKzdmYLjZ7Tx1rvxjXub6rdfEH6RJO2umZWVhYqKircJ39delvJr97EtMBeLN8XhYHQBbt59uzfeunz5MsLCwnDgwAEcPHgQwcHBSE1Nhe6L2Hc+7tPlY05Ojgm7detW7Ny5ExERESasBg06CwsLucNrLC5evPhkMCDyXl5ebhaU3rp1C3V1dcjLy0McPZfovvNobm5GWVkZcnNzcffuXeftt/q3Lwl/Nsm9yPgBEuqp07ywZMlB4x8/mP745bVHhH8mF8F6eCWShE83dvKDB08x1+vdoo13nJmzNmH6DG8S9XxEx92mb/8Q/Onj7zFxkpux6xfhd3rM8fJJpn3/aOMNaAMXCGugINt/+eLXAtzJUzfAl+nIFl8mPZvoKjQlrZ42/7F47/2vSOCXmAXD770/FF99PcvY6YvwL3H1Y54jqOGPMqZDWtT77//+Hn7x7+9j5MjFxhWoiL7TdKg/EP7qmlrERIQg2GsOKsOXoT2LPveP+eDYnjlYOPJP8F34HTfXmgRL+N/q19sWziJgEGjgxlunS68jPPUsorVot+o+F+1acAYaAv2C8Ddz90aRr8mTJyMqKspoaTUIEJmQ1rahocGQMREvkTL9LXKFTg1O4UXw6rntu3n2k5rQPT1Tej/Vzipdpfm8uD9Jql//ebemHlmnruBoZgnyi6/iwaMfk95+LXwPhAsOCqZZxCf47LPPMHz4cHMO/WooZs2aRS8niU9Ivwj47t276Y/9O3z99deYOHEiRtE++auvvoKrqytKSkoMSXdzc8Pvf/97rFq1ClevXjUSqb1oILF8+XJkZmbi0qVL3AWWrho//BAbN258Qu5rampw5MgRLv6cYQYHPSjOgIvS14RfhFzuNUeOXmIWxn43fCFGj1lG15mD8OfPfqAHHn+S9/UYOnQmvd+kYAvJucPf/jx4b04hUd/A+hxMgj4F4yasxp/+9D1kp7/OLeqJD3/545e5jlxrLlnqj9/+dhAJ/FCMHLWE5H4N8xmLP/7xOzNzEEJPPOs4mPiMi3bH/LACK7lQV5p6uf5c7x5tBgEi98Npz68NvET4l688gkGDJpqFuYk08dFsxP9Lwv8v//IbPgsycqicMh/qLxp+J+EP8pyDijBXtGZ64nGuN3J3z7aEf8C9pVZgi8CrISDTYJH9+Z7RWLopFtkny+2i3VeD9I3E7heEv7GhEevXr8eQIUMQGBhoNKilpaVGiyqXcEePHjWu4c6dO2fu6e+0tDTcuHEDjznMfPjwIc6ePWs0rhowSMOrmQKRER0Kl56e/iSdEydOcFObC8akQwMA5SVtrbS96WnpuFZ1rd9sePOyraKMfvg3085u/sYYeurJf7v98HPAJxL/5Zdfmt+ioiKjwZf2XkR+/PjxOJF3AvI2Eh8Xb+6NHTsWSUlJkKZes0rbt2+necYYhISEGO38ihUr8A//8A/4/PPPERoaauJqMLlnzx5MmULXg/HxKC4uNoPTf/u3f6Ot9jfmngaNmkHw9fXFsGHDzILHl627gRi+Lwm/zHdkSy9f9Vt35GDcuNUk1t/TB/8wfDt8vtGs795XaMx6NnjGG//28rzjRo83s+dso4/8E2YB7kx6y/mEtvsi8YPoy99l6QF63LlJgv104y2RbSfpX7kqmH7455iFt+9/+B3dbs4zHoCk8Vc4LRReQY9AQxnmk09/oPnOXP4daDz5yK5/DjfpWsE0EpIbzFoCeRxatHg/FxcXmtmEA/4X6O/fG5MmbzCDGWn2dVrC3/0bUFVVZQboGqyrz7eHRcAi0PcIaOOtuOwLWO2bhA170nBCJj3WS0/fA9/LOfQ7wh8eHm5MLDw9PQ2pWrt2LUTCJk2ahDlz5sDDwwPLli3DtGnTEBgQiNs3byMzI5M7cy6Fi4sL1qxZw2l+avu8vFBZWYl79+6Z2YOpU6di9arVWLlypSGC7u7uRqt78uRJE8d1qav5nTt3riGBijsQj6pbtTgSW4iNBzMRkXYOd6vfXi89Itm7d+2ml5PvkZycbAZpuidN+4H9B8z9TT6bcL7kPNatXWe0/+Fh4U8Ggqrf69ev0/1hAk6fPm3andqbNPdjRo/BggULUHK+xMwI7du3zxB+DQxFNKZPn44RI0YYwq82JyKidQGaqRLhl3nau3D0NeEXCRYZFnmWzfzWHbnYtDWTpj2lZiAgO3kRcA0KFE4DhJj4OoRF3DR2+Yn01CNb/R278o2ve3ngkc1/Z5KtPHQ6Fuq2M606k/4WLvTdtCXTLNSVC02Z/piTAwW5+9xNN58+9NSzz+80/75v4ickUxMWddusO9BGWjolj2YPJKMGGfFJDSZMGOVOZHgn2e8rwq93QkqRmzdvPpnx6q5tWg1/dwjZ5xaBdweBepr0FJRcQ0B8ETffOouLFfTD397JxOLdgWJAl7TfEH6ZUkgrKw39mTNnDLmS6YVMKUSyRo4cachVRHiECSOTjPnz5+NM0RnkZOdg65atiI6KNmYcImqDBw82mteU1BQTV0RettXR0dH44osvDOmPjIrEpk2bMGHiBBM3Pz/fDCikPZI9+LNMg/p7bd+4U4fwlHPwDclDfE4p/fDX93eReyyfiMye3XsMsZf5jnNGRwleLruMRQsXmcGftPuTJ02mDfgSlF8u/1l+Tvt9kXYRfpmWKV3NEOzdt9fMBvj7+0ODRmn41T41qNTgUe1W4Q8dOoQ7t++YmQZp/eXS8F04Xgfhd5DxdrPAVqYvOmX6Ivt+aeWd107irJ1wnfdE4p1xnPf1tzOsk+w7f3Xfmb4zntLqHMcZxjyn954Mpqe0HTMGbSZvPXOm6fS+I1kdZWEY+tuXC07J5wynX+XVFzb8MllT2160aJHp2zRbqkHA8w5L+J+HjL1vEXj3EKh90ISI1GIs3HgUrlviuannFavhH4DNoP8R/sgonC46bWywZ8+ebeylZX8tm2lp+CsrKnHt2jVjTz1hwgRjsqGFk9LchoaE4oDfAUyaOAm//vWv4efnZzSugwYNwv79+031PKh7gPnz5mP6tOnYsWOHIXHffvutWfgpraxmFj799FMzbSyN7UA7yrgD3ib/bMzzPIp99Md/487bu2i3M+GXlr7zAE3aTNnha2ZIpjmjR4/GunXrTNt5Xp1K279y1Uqj2S8qLDL2+WpjGiTu3bPXaPVF+DUbIDt9tRXZ9LsuczWkPykxycwOaS2BJfzPQ/n597VW4ll++DsTYpFtB1l3mMB0ftbVtSOeNPQvGs+RT1d5OdJ0yuOYJehKhhd59iKEXyZmGqSqvXd3akCmtU0pKSl477338F/+y3/BP/7jPxqlh9akyIxRpnC3b982aTprxxJ+JxL21yJgEXhYT2VEQTm2HM6Bb7DTLWe7BWaAIdA/CT8JlbT3MtGR1xN9jPRx0uLKqmtVxlRH2lWROdnri+wtc12GtWvWYv269fh++PeG8O/ftx8y6ZCNt2YKdOgDKfMOkX5vb2/Iplsa/w0b6IGD9tdOk6GgoCDU1tYOsOoEKm/QnCX6JP5/9s46Oqqs2/b9173j3ffu931tuLu7u7tbgODBXRIgBAkSnJAAgQQJENzdIbi7S0MjjTTujbTMN+c+dSI0BEmAQFeNscc5dbxOnVR+e+255hoevAWLNx6HfPm/1pcN/JLWSJcvYLRfctnRqI46iQJxReF7evY0z5O9jT01CdvU+SvvQ8+Vnr0b12/g0KFD1I23MJ1PD3cPA/m2/McGfiX2CqY0GiCNf5/efUxisNyA/gmvTxXhfxdY/tBt3r8zEBnuFbnfRvtO2XjKWSjiaMD7XtPbgF+/X4rOz5492wQsJk2ahKhacHCwCXpo1DN9+vT45ptvwprgP0+ePOZ51W/r1GlTTY6U7GqdwP9P+Ot1fkbnHXi3O/D02e84cPIa5q45ioUbjuPspdtOW853u3WxaqtYA/yKvkqGs2zpMhPhF6zpn5SAXwmW0u1LkiHZhcBMun7JeiS1UMdAwKUEX0Vf1TmQ04p010GBQQbo7Qj//Qf30bNnTyPJEOAL3JS0KWhTMubu3btNgu9P536KJBGJVd9aFBdjJD0cepu4aB8r4p376iU90vBrhEayLyVgK/qpjppsOqXtHz9+vEnKVlK4a31XE+0X4OulbRXVV97INj47JwlSGhXQyNLdO3eNa1NIyHQ6wNRB3rx5jWbfBn7lkCgPRCMJTx4/MdBVpkwZusCUo7d6JSMfi+Jr+mpWxUbgjxh5tyL0FqBrXhKaTWphkf7wSP4m6uvDAV3LrYJf4dtGBn1tq22UX7Bg0S+YGnLcaP9VRCv8OH/fJ6p1bwN+JaBrxEkjVmVrzz2dAAAgAElEQVTLljXPmp63NzW5V2ldwYIF8cMPP4TBfkTw/+///m8kS5bMmCaog6C/Cfnwr162iLac7k6Xnq/mr9X5QZx34MPuwIPHzxySnlWU9KzDTkp6fncm7X7YzfyMe8Ua4A/T8FPSc+zoMRNlld7UBn5BvBJzbeAXmClqK5BXRFY2jPJQF7ALxnLnzm0iYIrsC/y0v+wUJcGQG5A6CNL8y1ZR83LxUbR21apVGB8w3ri9KIL8pb1OXmDnKGAdGnnOxujpW3Hlxv0v7SO8+/Xy65HWXo46Y8eONd+9HJjkqCNIV8RdCbaSNIRuCoXsOps2aQrlgcih58CBA2ZUR/kheo7U4ZPOWaMCto/+jRvXzchS8uTJDWApV0AafnUUNSokONJLz6k6CmnTpmUBqALmWt79g3y5W8Ze4I8I9RaYRwR3KwHXgn2578yYfcYk2m50SIZMx8DRAdB+2v5VUNfyNUzEHR+4G25u/VGmbH1W23Wj7/4s47f/qj7/1f1f9/5twC8pj541OZXJAlauZlE1/f7NnTvXBEgyZcwYBvz/9V//hf/85z/Qc63nVSOdixctNh1YdYSdEf4v92/SeeXOOxDTd+DZ85c48dOvtPs+jTXMDbx47Z4zaTemb/InOF6sAP6XL16aiKwi74rSC8yVTDtu3DgD+LbTjjT3ivbLZlOgp06Ctpc1oxIqBWD+/v4mWVOjBVq/nnIf2/FHEh4BnaKwGh0QHO7Zs8eMHkjKofP19uoN31G+OHnq5Ce4/TF/imtM2pWUJ2jBbibtnvmqI/y6e8vZQZS3vhKtBeFyz3FzczPJiZtp3Sog1Uv5GErEbtmC/ul04OnWtZvZXs+BZFxnz501kK9nSc9JxPyNHdt3mFEgyXv0vAi49EzpeLdu3TLHF4gp+i/9viRG0kX/E16xC/h/N1V45ZizaOk1Rt2vYN6Ci8YlZ+2GpyykddUU41L13aV00tlIB53VdN/p3nMiXGnzOdJ3XZjbj1x4Zs4+h2mM2s/ncbS/OgM2pFudhxd0/9mLRqycW6VqK7i16E+rzs608eyK0WM2maTiiJ0Me9+opm8Dfj1Tkq5pNEsd2Xdpqi8iFyuNQCnKLw1/+fLlzSinZI96ptVxff7sedgj6wT+sFvhnHHegX/8HXjy2wvsPnoZwUv2GcmwAovOSrtf3mMRK4BfESVBlCKukkhIQypXCUVhNYStf27yzVezC2SdO3vORFrVGVBVVbn7KJIlLbe81xWp17wi/rNmzzKdAVVkFSAKCBXF1Tn0D1NabSV2an9FfxXBtWUfX9pXeok97ymL9xm/3HlrD+Pmna83aVffjUZl9D0vW7bMJCBqqsRafbcRAUbbKmovpybpnyUFUlK3qupevHiRVQOtREjVZFCkX8+c/ZKbiZ5N6fIF+Ho+paPWM6vnx36pk6DnTaNI8uT/J7xiD/BbUfhlK29h+MhV9LgfyqJZA9gBHI1gWnGqIFer1kNQq3YHun0xt8N9vKngO542n6XL1EOK5OkNqE+kfec82n96ek1FvfoeVsGuloNYxGuzseu0ov2Whac6C737zkDV6q3hTR/+hUuuYvqMkxg5eh2msTrwx4jwf+gzpedanVSNkkq2oxFNPb/qDLzu5QT+190V5zLnHfhn3oFHxof/LAZN3GTyA/cdd/rwf4lPQqwAft04yWfsFvG9fVNfty6i5Eb/uARZmmq5ktsEavpH5zfaz8h/JPuQhEcRYY0g2FCm7QVx6jyog/Elv+TS4z9zOzxG0qWHvXFp+r/ml747dRgFnoqy2xH9qD6zOnP37t4zBd5e3V7HU3vdK+Lyd93udcf5mpbFFuAXiG/e9ifmL7xsKvHmzFmM1W57cJSPfwchtPntMJK5Hh4E9Jno3GUcpTfNWZxrHEb7b2J0viWyZy8MdxbMmjj5INePZeXm5qzSOxJ9+s00x3Ft0ANBk5gXs+m5OY+sNhctvsYCWqNZDMwVbdsP535j0IVFvnTM5SvvIPQ1MqCoovta9y4R/vd9fvSs6rdQckiNjr4J8iMe92MB//xBTdGxfjlMDhyH+w/e/7dJn0EJxs7CWxG/Lee88w583Dvw7MXvOHH+JlZtO4N1zA28dO2+U9LzcW/5Rzl6rAH+j/LpeNA/mViyOXSzqeQ7YsRISnqGUO89Dmc4QvA1vgT4S0NPYBJ9+PWH+TX78H+N39+X9pliF/D/hbmU8NSu3Qk5chbHkGHLjQRn5Zr78GJ1XQG99PZ9vWejWPEaJtqvQl5uLQaiEgF/cvARSn6Om3XFitXEoMGLMZ2RerfmA5A/fwV07zHFSH4E5Zvpv69CYM3cvOl0U4pOYG2YN9KReR6uqFmrA0aN3kDJEBN2I8iA3gb7Hwv4P+SZimngX+fXDuv82mKWdyO0qVMGwRMD2Ql5/ehCVNfrBP6o7o5znfMOfJw78ISFt3azuu40Wn3PWHkIJy/cNIG2j3M251E/1h34Rv+wb1IXf/6nszjLqqJnKaOxmubf3M5w3RnKJiK2s3qv5a9ZF7Y8imNGPF/E7f82T329Oa+OdeY0fjp3Bj+fP4fLFy/g6i+XcY3WnVevXcXlK5chj34lAa9dsxqzZ87A/LlzsG1LKPc/gQvc73yEduGnMzCNyy7wfqj9/Go7z2VsF9/ULnDdhXPh6817ba9l7BlznZrZxkyt4136+Sf8euMqfmeUOjqvS9fvGf/9wZM2YcG6o7TlfHNxneicx7mv8w7oDsRG4HehZKdESRcC/CEm2/5uKu4OGLSIEh8ftO8wGo2b9kbBQhVRo2ZbVtLdiqaE9oqV3FjFdzer8W5GrtwlUKhwFbRoxQRuj0C41OuG8hWa0NZ1mpH12MA/Z97PjP570gGnIvoPmI9Zc87R5WkhChSshPocEVCU30r2Ddf+vw36P0aE/0Oe1JgE/rWE/Q1j2mPR4Gbo06wC2jeta36PX7608mve5/qcwP8+d8u5rfMOxMwdePT0BTbu/Qkjp2+BP+v8HGTVXaeGP2bu7ac8yjf3793FzGnBaNeiKRq51ELjunXYapv5RnVq4dXWsE5NNOTyBmq11fi+dg22mmikea43y2px3rQacK3Fxm0amH2tbazjhM83cuF8pFaL79nq8hrCWm1eWy00qVebjY4rri5wa1gPrZo2RPsWjdGtfUv08uiIXj270tJTSZkd6D7hhkaU8DSqXxctGtVFu2auaN/cFR3Y2rvVRwfTrPcdW7iiU4sG6NyyoWldWjVEt9aN4N6mMdzbNoEHW/d2bB2aoIej9eRUzbNDU9N6dmyKnh2bwZNT0zo1RS81LuvV0Q1enazWq3NzeHZ247ZNebym6O3eCtMm+uLenZvR+v7P0R935LSt6Dh0KaP8TMZ7i6RHchjp1dUiSlYiXsTLly+M1EkSKOVYXLt23dliyT1QEvs9/g0/ffqEkqYX/NpeL0eK+H3G5HxsBH5p70uXqY9p1NNv3PzCSHFq1GiH2i5dmJi7Ad6EcysS3y4S8I8P2oMxY7eYiH0ZRuq795yCwRwlkOtOv/5zMYlyH1lwSs6jNm/BZTRvMQhVqrQweQJbuCxk5mkWuKqDChWbcgTgZ8sG9D2i/F8b8Av217MtHtIcQ9tVR7O6VWhhG4jrdL/6kJcT+D/krjn3cd6B6N2B5y/+wE+UC289cAHbD13CtZsPmfcWvWM69/70d+CbX6np7N+rJ7KnTYmkP36L5PF+ZPsByeJ8H9aS/vg916l9hySmfY8kP2j+W9OScaqm/ZN+z2UR2w/fIvEP/zFN2yflfkl/4LHYdCzznlNzvrg8r1qc75CcLUW875EywQ9IleBHpE4YB2kSxUG6JPGQMVlCZE6ROKxlTJEEGZInRPpkCUxLx2nKxPGQJH5cJIj7IxLG4THjx0GmFAmRM21i5ErLaZoEpmk+d9pEyJ2OLX0i5MmQGPkzJkGBTElQKHNSFMmaDEWzJUex7ClQPEdKlFDLlRIlc9stFUrlToXSeVKjTB5rWjpPGpTOmwZl8qY1rSynEVu5fHyfXy0NSnG/4rlTo0LRLPDq5obrVy9H6yl48Og3HDx5FVv2n8fxn27gMbPro3pJzytXpIs/q1R2eOEq7fPbb09Z5OwSi/HsZ1L0OgQGzcDAQX7o3WcYayIMd7ZYcA8GDPBlYaUZ9GZfT7eV/aZGxbtotKN6Jt5n3ccAfnd3b0wI2o3tLGL1toi4vd7W8M9d8DNc6nZDqdJ1DfCv2/gU/bznoHTp+kzUDcSK1fcx2m8jSpWqQ+1+c4zy28CIfx/q8BtgbMA2BE0+QHvNepTltKf8ZxeW0VN/6vRjmMD5hbTv3EBnH0XtlS+wfNUdyoQC6XPvhsFDl5n1StjV6EDt2h2xeNn1DwN+diR69wmg69NG/MaO3Od4xVSEX1KeRUPcMKhNVbRpXBszOcp6m7lSbwouvO2zOoH/bXfIud55B2L+Djx4/BxLN59Ej9Gr0XvcWuw4fMnpwx/zt/mjH/GbG9evobeHOzIkTYR4//s/SPDd/yLBt2z/+V/Ed7R4nJr2b07V/vX/uO3/ZdP0/yE+5+P/i+///T9c938Rl+/j8lhx/vf/IM6//g/imvY/iKv1/+a2/+Y+//4Xz/EvJGRL9C3bd/9C4u//bTXOJ1H74d/sRPybHYD/mJY8Ln2j432HlHG/Q6r4PyB1/O+ROsH3SMOWOv53nP+W82yc134/8rq/57X+oGvmOVKyA5EpaRxkSfYj2w/IzJY1udqPyJaCLWUcZEsVBzlTx0WeNPGQL2185E+fAAUyJEDBTAlRKFMiFGYrkpktSyIUzZqYHYIk7BAkQbFsSVGC0xKaZk/GDkIyTpNbLQenEVrJnMlRMhdbTm2XFIW5T+kC6dGzc2MC/6VofemXrt3FlEV7MTBoA+atPYJf3yLpUXKznGfu3I78T1jR/PWsUOs/ZhKGDJ1MoJmL4aNWMEK6BqNoX+hsseEerMXIUWswgt/LkKHz4DNkMiPU1JmvXktHIssuNFoP0zvsHJPALwg8fvwYa2p4Gfjewcq1NtC/bRoO/BcjAP8JY9Ope1S2bAOzfMCghXTrGYz8BcpS1lMZPTynwJXJvfkLVECXbuOMxr9N22GU8DRE67bDTWS/dZuhlAKNYgfiBIH/OSH+TyMV0rz/mM3MBehokoCl51cCcLFi1eHlNc2cW9f1tmuPuH4bCzTPnnue8iF/bGGiwIsX4S5Q7/B1xNgmMQH8NuwPbF0FHZq5YEbINNxhsnx0Xk7gj87dc+7rvAMfdgcUONx28GcEzN2FiVQOHD133Snp+bBb+Vn3MsDfp7sHMqZIiviE4sSMwic2UXlrmojR+YQE8YTf/weJGLlXS/gd27dc/q2mjnkCekJ2FBJp2++0j7ZnZJ8tCY+RNGKkn/NJ1MLO9W/Oc5s4HCngyELyuN9bQM/ofmq2NAl/QNpEP1otMaP8bOnZMiZhI8BntluyuAT5uMiaLB4ycF1SHis+zx+P15yA50vB0YtMyeIjeyq1eKbl4DRH6vim5UybALnSJWCkPwHypU+IghkSEfQTM9JPsM9CsM+SmC0p4T4ZigvoCfHFCe8lcqZASc6XYivNebVSHAUolYsRf05LczTAanzPaH4ZMxrAKUcGSnJ9sZwpUa5wJvTq0izaEf5b955gy77zWLLpOPbQN1cV8qJ6CbIU2Y8Y3f/llysYP2ECevUeSYvBFZg05SBmzrnAiOUtRkgf0Nf8kbPFknug70Pfi76fSdSsjxrNSoj9/FmPYgwdqk5H9dXHyLqYBH5d0HUGIDw8emPAoHkEZks+ExGIo5qXDaaq3A4fsYaynXm8LzcI5i/pv/+Lsdls2LgXi7GNgGevqcZNR6DvSTBXEm+jxl4E/GFG/iMbz169p6FlKx9KAgeiXftRGDpshfH2l0TIhnhF+peuuMXO73q0ajOEHYouRtPfhxIg+f+rYxDV9b66TjIhjSCMHb8d7u4+LCR4hJHwP2Pke3rfg0QH+NeObov1/u2xcmRreDUtS6lnRQRPmcQaKtGTK+ozOIH/fb9J5/bOOxD9O/Dkt5fYc4w+/EraXWUn7To1PdG/s5/2CN/cYMGVPt3dCfxJHMD/PYH/Owv6Be2EdYG7gX1CcyI2Ab/eJ47YGI1PQtAXyKvDoO0idhwE/aYZWY9D/mMAn5AfT1IiSngYsU8lyHcAfjpCfrrEVhPAm0a4z5g0LsE9LiP0VhPkR2rJ9T6e6TQI9OML+HWtlA5JGpQ5OWE/pQX8An8Bf840agR+QT+BP0+6hIT+RJT2JDLAX9jAvoCf4E+Zj6BfEfyIwC/oLyn4N8DPKYG/VC52AAj1ZQz8C/hTEfRt6CfwE/YN8Bch8HeNPvDfuf8U29kTX7HlJPafuIKHT6IG/oiPmzTgp06eQsC4yfDoMZoJjFuNBGLjZkkYgC3b5U4i/TKdj5wtFtwDS0tufy+btkhmch8BE3bRUcafBcSCWFPg1AfLJyI+G2+aj2ng1zO4ePFSRt59MSZgh5HP6PMJrgXab2uC6LXrn2D1+scGnu39BOaTpx42ibyLllzHMr5XxH4mE23VMZgx67RZt0QFuXgMFd5S0a2giXtNxH0Nj/f3c/9hoF66/nkLL5nOwvQZp0yhL2sU4N2uWddovkN+zqkhp9ghGY8JE6bg5s1f33TbP/ryDwV+afYV2V88tAV8KONp26SOKYx4g7kmMfFyAn9M3EXnMZx34P3uwENKelZvP4P+gRswZHIo4Z8+/H86gf/97uLn3zoM+DMxwp+QEf4k1OobnT5B2UThpdWX3t5MrWUCeWtdONRb+2hbNSt6b2v8jXafEfxkBHyjzWfk3ejzjSSHEXyCvd3SEezV0if5kVF8NjMf18C+IvqS5GQi8Nuwr2kk2Nd7An3WFPF4nB+RmPIfG/gF/Uk5epAuSVwD/DntKL8N/KkJ+w7oF/DnJfDnzygpT2IUzpzERPlNpD9rUkJ/UhQ3kX5LtlMyh6L8Kaxov4Bf4E/QLxkF8Av+1WEoym1jKsJvXHoW74VP0EYsoKTn5lskPREfwUsXL8HfbyJhcSxh6BSB1oJ8O6L5akTS+Z62i4TD2NL0PQkcQ7f+RceY8/DqPZnR5/EsQvbxqkbHNPDrebx//x4TZGWj6Ws6nYuW/krY/hNbaIepz/e2JlmMWsTtt+8Cdu5l2wNo3rzn/A7mCdjzWq/9lDi7Tdtz/S4u0zbmeG8491buo220rfbRvm+7xojr1TlYvuoBJkw8SPvQyaz8PN4Uf/vUCdgRfws+BPht2F/o44YBLSvTjccFs2fNwoOHMVf8zwn8Eb8l57zzDnyaOyBJj3T7E2n3HbxkP/MDf2XSrhP4P83dj7mzGElP3x4eyJwyGRJ/R808YTwpE2aTMRpuJekK9i3gtyP2JspPeFakXxCd6Hsl8FrN7OuQ5UiaY8BeybdqiuBTe68ovh3JT5MwHPaNbMch2bFlO2HAr6h+hBYR+O15RfWzMLqf1QH8GZPFYbKuZD3WdZpr5WdLzXNkZYQ/Z2pG9yNIeqxIv6A/oYnw56WsJ38GSnsySrtvQX9havaNbp/TYlkt6C/BpN6STOpVM1p9wb46AAJ/B/AbaY/kPYr0M8Iv2JfspxiBvwhHBsoWykjXnuhH+G/ee4yNe85h3roj2HHo57dKeuxHSZA1Z84i9PIKQDCTFCUteF9JwruAr2QXGiV43bHtCOq7HOfjbvO7ub7NEQBP89Y1v58m++Ne5+s7G7qP0pfPnH3WRPrHj59IOcUN+6uO0enHAH5d4LlzP2HKlJnwZKS/d9/pGEHZjN+YrWxbotG0f8RjvDof8b3O8+r2bzv3+26/BaN5jpH07PceMJeuQP6sETIF+/btt/JqmNyqYnKf4/W+wG+sNynjWejTDH2bVUTrBtURQvc3JejG5MsJ/DF5N53Hct6Bd7sDTwj8u49eMbAfsuIQThkffifwv9vdiz1bfSPNbL+eHsiamsBP6Y7lkkOHHkb6E7MlYsTebgk5L4lMXEp34lDDrxZP8h5ul4xwn5zRdAF+Crr8pLSbAD9ic0h2JNtJxZbaAfyCfUF+WgfwS6OfnpF4yXgycioZj5HyRIjuW5H9eEaiI5mOgN/AvoCfTR2BlDxHQsqIdN0Cfs2n4LLMKeIiBxNzczigPzuj/Go5HO49uanll46/AKP80vIXsqGf0p4iRt4jTX849JekvEetRA5KfZSU65D3KNIfDvvS9bNR0y9nn5JsRdkxKMz9yhTKwKTdptHW8N95QEnPkUtYuf00Dpz8BY/eQdIjaAtlcbKuXQcyWXI7vcafvBbIYwJeBaJrNzwx9oavHm/T5tgA05ZsRB2eVWsfYNnKW9Rp38ZKauXXb3rGSPMfjOjHhut8Pezb99RIWTj6MCFwD92U/CmTWYxnz36L8V+ejwX8yilR5esN60Ph7T2UibbuaNmyG1q1+rqa/Zl6enpjwcIluEbXNFX7Pn/+PA4dOoQHDx7E+Hf2Lge0gX+BrzsuL+mL37eNwl+7/LBrclf0bFwOQT1dsWpUGyPfWcuCWrLeXDJU1ps14FavGiZNDDIWvu9yrvfZ5kOAX3lK6ji9aj2sZ1fLNLVf2lbL1GRZbL+0jaoVP3r0kNNH5juSG9ajR4/x8NEjPOK6x/zennCZvj8tf/SYU7YnT7TsKffTe2u5ttX+mj7i8R490f6PreP+9pu5Xl2L8+W8A7HhDtyj+9+C9cfQccgSeIxcYew5/2BRU+fry7oDVtJuD3dkSpmEEXtKehidtxoj9oyGaz5xWPuOwPwdgf9b/PDtv00T/CfiMgP8BvbtaD6hn6CvFjGqL8hXUyKuWmqHnCetLDfVwoA/rpHzZHDAvonuE+Cl3bclPTbwm8i+ovsmwu+Afkp6BP3qSOhzxCfwm0g/p0nZMVHnISeTdyXrUZQ/O8Hf6Pm5LFzW44B+Je8S+I20x0T6LegX+JsoPyP9JaTpZ6Tf6PoF/QR+NTnylKIjjyL9pajXt4Bfkf9U7BhYwF+EwG9F+KMP/OfplTt21nb08F3BAlz7oMq7b3udOXOabi8B6MNI6qq1D00EPqah1oo6vzCVS2VhKJ25dNGK+Jt1hGkVKpLWWp0CO9pvrbe2EcxquZbZy+3rtLe339vbannYvGM/+5xa/moTKAv0lQCrqGvXbgHo3GUM5TEhCJy4h0Wc7juA39aTW9fy6jGt69F1v7re3i9ypyH8+v9+Ta9e47u+1zWtoZ59lN9qjBoViDNnzkRKzn7bc/Eu6z8W8NvnfvHiOez6D6oB8bW227dvGxvc8M/9gu9/i/Hvyz7+26Y28M8b2RUXF0cB/P7toOi+ZDyDWldF5xYNsGDuXH5n4W48linAn/wsfxqIjg7Ifgjwy4lMz/6xY8foehRuU3ydnasjR47g6tWrYbdDHYODBw8yYfqwgW+teMbvQctmUZ4UMoPFGxcsYI2F+Zg+fQaCgoIwNXgK5s6ZjQXz5xoJ07Sp0+hINN0UeZw/by7mzZ2NkOnT2aZhzuxZmD9vjmmzZs3kvlMxY3oI5s+ZyzaPxSG5bUgIdu3ebToYYRfmnHHegc94B549f2mceZaEnsSKradx4eo9p6TnM34fH3rqCMBPDT8ddpLKB5/gnlSNUXsL+JnES2iWHj4R4V+uNz8wuq8WxwH8ku+ESXgI+UrAVbOAn1IeB+jbU0X4DfQnCnfgSUPNfppXovwG+An9tpzn1YRdS8MfDvtZHKCfjdIeNeUCJGVCsGBfjj2K8kt+pJEDwb4N/EbaY+CfnQBp+Zm8qyh/XibwWtacEaU94cBvknjtSD+TeS1dv2XNKXtOK9Jvgb9J6HXo+qXtL8FWlLCvCH/ZGIrw32WEfzcj/Gt3nsGh01cZOXoe5bOhf3Arlq+g77e/8S0XJKq9K1S+y3aCWUunfMcAdO48JVGvvjsWLPoFW0yewF8E/Zu0l1yDYcNXYtWaB6bTYeucJQEyEWvCs5U0/Bc10lbCqnWthGhes72drsk+p85rz9v7amovj3j9WqZOyPBRq00V1kqVm6FxEy/IbrFyZTdUr96G1pdLwpIyrcRl6zqsY1vn0vkinkvXqs9in1NTNfvc9vbWssgdAXubD51KijQ15ASGDQvBhg2hMR7l/9jAH+XD61z50e7AfY4sLJw/B6M9m+NIiAdebB0J7P57hF8VdBcOdkOfphXQumEtAutM3Lt/31zXvXv3zMjhjJAZBpYFzLMJvEuXLKZk65zpALzvB/gQ4FfHScC+a9cuPv/hJgYXL17Etm3bcOHChbDLUGR/y+YtZrki9cqjkFtS/4E+6NzdC97DRmNEwCS0d/dE+fLl0IQOREO6t8CEQV0xwqsdGtSqDJcqZTGsR0sE+XRF0OBu6NKyPmpWKoOOzGkI6N/JbNu/WwtUq1AGrhVrwqdldwS6D4Zfx35oXbMRCucrxJoR3Ux9lLALc84478BnvAOPnz7HBlbaHcGinv4zVWn3GpN2nRH+z/iVfNCpLeCnpCdzKkp66LSjolvJCfvJVHyLzYZ+E/VXB4AdgviU9vz4OuC3I/yOyH4Kh1bfeOaHSXlURMtqNvDLjSetgX0L+MP0+4T/SBH+VzT8kSL7JrofP0y/r+h+Nsp2Mib90dLx85oTOIBfbj1KCA4DfiXtRmpW8u7fgJ96fiPtMTadkaU9Rt5D8C8eFu0n7BvgF/Rb4G8i/IzqWwm9Sup1AH+2FChTkJKeTk2iLem5cuM+Zqw4iJFTN9Oa8xhk0xnV6/79u4w8zcEgn1nGclPwGTFK/qGgGXE/wbqAd+Lk/ahKn/IkiVOxOFFlU+VUMp4Nm57T538pvc8boXqNNvBlYSR5kU9mlF3LZYk4Z94F47wykYWReveZTm36RFNASSMCskqcv+gyI/B7IQcWfQaNVITMPIUZrHy6et0jzF1wiV71y02hpD59Z2IirUYF9/ZnlX2jkpR1jbqGwnAQHvoAACAASURBVEWq8J/uBOO8MkU2jV4hrJ7aBPVd3TFr7k8E9j+wZPmvrEuwDu7dg2iFOQtTgg8buZIkS1PoCDOO9orqxHh6BZvPMXveeeMIM4Wfa9acs2EjGSrgFDRpHws8HTejCzHZ4Qqlm9KCxdd5T5czR2MJHj6MWYmIE/ij+uv6ctcpwj9vziz4dGmM/VO6/Q34J3q6Yg3tN1eMaEXNfgUj45kyZTLrP1ia/YcPHxHuZ9JquA0GdWOBs97tMLh7SzRxqYKypUsieOpUjmCEw/e73qkPAX5bHqZ6I5q3X5LoqFL1o0fhScV//vGnGUXSiIskPYfYUejVpy+69x+CqUvWY9WeE5i8eB0atWyHuhWLYcHwDji+YDCOzSewd2+IeuULYGDbWjg8dyBOLRqKJaM7o3HVomhRswSWju6CM0uHY8sUL3g2q4ZqBYtgVEMPHKAF7YkRSzC/0yjUyVMGmROnhmudemaUwb5W59R5Bz7nHVDgcO3Os/CZtIlssRX7TvzCwltOydnn/E4+5NxW0u47A/8PbwR+S9LDyD6hP0zK8zfgt0Hfniq6LymPQN+C/nBJj+W1n56Qn4EtTMNPSY/mI0X2HVKeLCkiAD+TcrOljG+2S8HCXOqkJGBysUncjQD8dtJuZOCnrIdR/tyM8lsRfmr5lbxLTX8Y8Id580vLT9ce2XUK+Dlvy3uMrt/AvoDfIemhrKe0GiU9xoNfEX5Kgcqo8FbH6AP/xWv3MGXxPgwMXI95aw6/1aXn4sULCA6ezwJC6wmr0qjHbIRZ4C/gl0xHYFymjCuqVmtlChWpoNFCRvnXbXiKjp38kTFjHhQtVoO2hNNMddTaLp1RsXJTdOrsz4JSWwnNy1CPwN2gQXc0bOSJWrU6sEBRMDsH58y6+vU9TCdCwDx3/kV085hgQTshvlOXMWjY0BNNm/WFa4MeaMOiSpO4fD07GybizmuUvWL3HpORP3956NoWLb1u3Fnk0GIqtPqHmvMsWHwVs9kB6dBxNIsudeC19OKIhQcaN/YiWG8gzJ8z56lUqRk154N53p6oUaMturoHUMq0E215bhV5Un6Azj923DbU54hH7z4hXHbbdDwidpiiMx+67Q/KkB4yAXU9E0KnmmTQD/mheNM+TuB/0535spcL+JcunIeJ/dvi9BxPvGSE/y9HhN+zSXlM6kXpziA3Wm9Wp4zH1chQbhKo9ZIF51x2FgZ0b4PpPm1weM5A7JzaG77d6qNioazInDY5fEeOMNr1971LHwL873sOe/v9+/djoM9QeHgPxpItB3DiJn8fj/6MwaPHwatLC6wL7IkX+6bi6e7JWBPggX6ta2B8r6b4eaUvXuyfioNzB2F4N1f0a1MTmwn5fxycjtubx2OmT3t4uTZAcNsBuDR2PZ5O34+9g+bAt4EH2hWrjXp5y9LOtCUOMofD+XLegdhwB54+e8mcwKuYQ+e/BeuP4+xldoidOSax4at5r2sIA/5MjPDLc9+K8FvRfSXwSuJjafql5f8hTNLzIwtuqcWxk3Y5GpCcha2UsJvCEeGPnKwryI8TlqRr23AK+NX0XlH+yMBPSU4E2DfQ79DxW848Ssy1mpHyOOQ8suSUC48Bfkb6U1I2lMAB/LaOX+exYN/hw88If7isRw4+lpY/jyN5N78jgbdQRrr1SMfPVsRh1WlF91mQi7CvVtxIe6jhd2j6jZbfAL/t3GNp+SXxKcZtilD/X7YAI/wxAPxhhbc2Hqe05zLknxvV6+RJeY3Pw4SgnYTP8KJC0YHMiPuqA6E2d/7PqFuvK+rVc6f94B76/E808O9LhxJF6L0HzkeRItW4jTs7H1sI6GORO28pA/hBBPORo9YZaK5UxY2a9A2YMu0w6rp05bB6Q3j3n0Od/VgULFgJPoOXmOi7/NAbELQVkffuP9dUPxWUjx23lZ2HLUwEnYvAoD0G8m2JzVJ6sLs1H4BixWtiNOF+Q6ii/n/x+qxOi0YpZOGoEYN+3nPYMShnJD/TQo5i0JClqFKlORoR/lWIqUTJOsibryyvZymtJbewk9KDVVhbYNiIlea6qtVobdyQlBDcpes4FC9ewxxTmvuIcp+I9/JD5tX5WceO3GiOmgwaNI6Fra5H9Ti89zon8L/3LfsidhDwr162EItGe+CXpf3w53ZfSnr8sYfR/l4E/hEdWE2Y0w7N6lKPPtckpOqDXb9xE1MmTYRXp2ZY6tfVwLBAd8tkL/Ryq4KW1QvDpVwBBE8cb5Jd3/dmfArgV2R/957d6Nt/AHoP8cWGg+dw8jbrO+w/C6+BQ+HeyhUbgnridwH8lglY5NsJ3RpXxJiejXFpjb9ZvndWf3i3rQ3PFtWwYxrv36EQXN0wBpP6tkDXWi4IaT8Y18ZvwuNpexHaNxie/F3rXKoe/Gp1gXvZhujUvA0OHDr4vrfHub3zDnyUO/D46Qts2f8zxszaiQnzduPIGWel3Y9yoz/yQcMlPamT05WHwE9Yjwj9Av6kBP0kbFEBv2Df7Mf9DfAzup/SjvBr6pDxGO0+ATyNaZYlp5x5Xm3pEkt/L5ce6ffjRWoW5CvKbznxZFahLWPHGW7JmY3R/uxs8uRPxc5EAkb1BfvKP1ACr84XVnDLIecJB35L4pOLxbgstx7p+OXYwwg/E3gL29D/GmlPMUeUP8yykxH84krgjeDcY7T8hH11BIoS9ouYCH9GAn/0k3Z/vnYXQfN2od/YNZi98iB+vRM+XP26Z+n48aN0cplN4N/xUYDfACcj58NHrELJki5wde1u5C7de05GzpzF0b7DKJMIO3b8NgiC23fwNfIdrS9fsREjbAsNlA8avBjlCPfu3ScauY78zr0J3RUrNqWDyxBG231RpFh1I9vROUMo5Wnc1MuMBAzjuatQSlSREfcePadgfOAu41OvqL1dOVXSl8WUAzVp2helSruYBF0Btg3f6rTYwL9wyVU0bzHAQLoAfu8BmCqsHTv5mtELdTDKl2/E6H9HSo2u0GP9DqVEE9ghaEELxvW0X5xoZENDhi03xZ8aNLI6JpIwqYMRk6MsGl1Zv/E5gT8U/QeMw7VrTuB/3d+Bc1nkOyDgX7mEyakjuuDiwt6M8I/Anzt9sWtiZ7StWQTNqxVGM8pzZoRMw11q9fW6fecuk1ZppdqhKaUsXXBv6wQ8ZwR886Te8CTsezYph8AeLujsWo7AH0B//rcbCkS+qo9faVeSnwMHDhjY9xw0HGv2nWRk/zk2HvkZvX1GEMTrYuWYbvhtzxTc3xaEJX5d0KVRBUzwamZg/89DM3Bong8GdqgDrxbVsX1qX24bjBsbxyHAqylaV62Cqe0G4ZeADXgyfS92DZgJr2ot0a54bUx09cS8ZoPQvVxjdHRr7QT+V7985/vPdgceUcO/cc9PGE45j98MavgZ7Xe69Hy2r+ODT2xF+D09kClNcrrtCPgJ7o5mNPyM2CdVM5F+Qb987b+jJWfkCH+KuFZ0Xx0GAb8d3begn9F9Qr9g39bvW1NF9sOtOK3oflxG+a1m6ffDo/jGoYcRfllwCvYtDT8hn8uy2u+5ThH+bI4mLb+8/mUpGo/XLeCXRaeAXy49uRTJdwC/fPmNtEfuPSbCT+BPG5/2nAkM7BtJjwP6bWmPKvDaNp2CfbsVp09/xMJccu+xXHusKH8JAr86AkU4GlAoa3KUZoS/RwxE+AX8kxbsRv/xazFn9aG3Av+xY0dY1dMC/g0xHOEXuAq+l7CSaes2Q5A9e1Fq46tRUtPT6OHTp8+FSpWbI5gaef+xm1G9ZhsD/NLv9+w1hRH9NkYDv27jM/TpNxtlyzGaP2Ce6QDs3gcTLa9cuQWaNOmDdu1HGjmQ9P5Kpp0x6wwaUWLTiEm305i0KrjW6EGp0vWZK9AU7djRCJ52jJ2cZ+YaBfZyCWpJCU7RotXYQVnJ3IJnpgiTGQEgOCv6LpeeGbPPmCh96dL1zGjBfo68L6cUR6MWlZjc27vPDPP5JB2Szn/Fqrsm50DAP9p/E/zGhqJOnU5oQ9nQQHZkqvIffo+ek3ifbppOxYdE8t+0jwF+yoZ8ed7+A8bSkcQJ/B/8a/kP2vH+g4dYsXQRgga0x5YJnXF6bm+cX9gP8+mzX6NUbtSrWZk++9MoEbtt7oo073Kp8XZvhZXjPHB363g83RuMTZN6oUfTyuhSvzRm9m+MeYOaon3dUggc58cCa++fT/IxI/wyMJBrTx/v/vDo54Ol2w/h6PXHWHfgjJHxuLdphPWU8RjY3x7Ez+mOnm5VMaZHY1xY4Ytn+4JxYM4g+HRwgUfTStga3AePd07GdcL+1P5t0LlmLUxq1Y+wvxH3p+zEwSHz0L9WO3Qo4YLxdT2wsu0oLHQbDI+yjZzA/w/6W/sSPuqzF7/jzEVKUHefQ+i+C1CuYHTctr6Ez/w1XqMV4ffs/jfgT8aIvZod4TdRfkK/AX7Cs6Q8dkv4g5J81VGwbDgF/BGhP1WCH8NceizffUvCY2Q8r0h5wivtKmHX8uDPzAh/5uTU4xPiDexzPouaI4KflfNWs2A/DPgp69G8ziM70XgEfQP8hH51NAT8uR3An0uQb5rl3JMzFRN3jY4/frhTjyPKX9Ak7yZ0FONKFAn4LS0/pT0E+WKqxBvJrtOCfhv2i0m/H8PALx/+nYd/xuqtp6i5u/JWH/6PCfxWdPw5E1L3EoSbMRrfDJ27BjCJdTq6egQZAC9WrJaBeAF5derc27UfZRJ2e/YKNu+Hj1xj9P8DfRahXLlG8GAewKp1D7GDEf6+3rN5zCbU4w9FJ+YAFClSHYOHLMNmwrsSbV3qUkLk6sFk2GOYOecnkxQ7Ztx2tKaGvhwj8BpNWEIZj6Lq6pgoedir93TkL1CBkf4+mDvvZ9N50HpBu0YKenlNM8m4bs37M8JfyyQUmwj/kmtG06/8BO/+88zIQ33mGixmZ0cdCV23nH4E/HMXXkSLlgNNkrJkR5L6SPZj5xO8Cd4/ZLkT+L/Gn+2P/5mUULtjx3b4D/WGX78uCPTpjolDe2CgRyt0btcac2m9eYcRfb2uXruG4MmT0K9bS6wK6M7IfiDhdyrh2BPdGP3u5loa0/s1Mvad6jA0r1oYfiOHhu3/Pp/mYwG/YH/v3r3wHuSDAb4BWLH7GI5Ts79m32n0HTwcnhx93TCxFx7vmoQH2ydi/siOJrI/vncz/LRsFF7snYp9MwfAu00t9KVmf3dIPzzdMxm/bhqHoD4t0bV2Xcp4huDK2I24O3kHNvedil7VmqMzgwYT6nXH8tYjsLa9H+a7+VDS4wT+93kmnNt+/Dvw4MlzrGJtH+8J6zCYibt7jl12Rvg//m2P8TOEAX/mNCkI898iOeHcNEb5kxHcFd2XM4/dJOtRAqwi/KYR/BU912iAHd2XO08kOY8D+MMi/ATwMNjnvCL7KrSVkYCfiS0LK+RmY9Q+O2E9Z0omz6ZJxIJYCZGNQC6tvuDfJOgaoCfsK1nXNEu7bzT83C6bDfzMDTDAT9CPxzyF10X4TZQ/EvBbkX+TuGt0/JasR1H+MOCXvCczW1iUX9F+Ju9mZRJvGPDbmn7CP117FOVXZN80An+RLIzwZ6FPf/706N6hMa7/cjlaX/JPTKbxnbYZnWgfOZmR/uu3oo6ifTzgt6wpV1DO0tNzCuU8ddCDCbHSrC8lAAu0lYirxNYWLX3YCZiKmjXbGg293Gq695iEatVb0aZzlZHVjA/ahWqEaQH1JDrszF94ibr4nuwENIDkPr3pvFOoUGUm6gYaGU0/6vqLMQG4NiPpWt+FHY1x43cwQv/AJN6WKetK/X1vuvpcNVBvS3amTD1iOhrZcxRjtN+Hhci2YiIddDwoJVInoWFjTzqMHKHefhZzBipSTsSEvuXXmRQbaiL10vArL0FSo7r1uhH4rxvgl+OPXH6U1Ct3oN59Z1DjXwa5chU3un85/8S0nEcdBCfwR+vP6R+7s3TsstU8feokDtOpxjRqyg/s34fjx4+H6e8l/ZEPfc92TbDYtwtlLoF4eWC6kbJ0b1LJwP6Mfg2xjn7969nmDWyKtrVLYAIj/Nr3fV8fA/gl49m3b5+J7PceQokhYf/sPdrmHr4AL8p6OjapbZJyH++chCe7JmPhqM7o2qgixno2oYzHz2j2jy4YSs1+LRPx3zq1j5EyXd8wFhPYIWhXrQYmtxqAqwGheDJtH3b0nwFPdvLblaiDcS7uWNFmJFa19cWadqOdwP++D4Rz+09yB6yk3V8we/VhzF93FGcvycXK6dLzSW5+DJ7EAv5e3emcYAF/igRxHJKeOAbiLWtOAr8N/g7gj2uAn+BPiBZMy8pTGv6I+n1Bf0rKaaTfl/9+OPBHLLj1owX7BP/M1PWrZUn8A7LRPz8HoT8XQT53qoTIlyEJpS+p6FmfBnkyJmdl4ATIkpJe+3TiCQd+zjsg30radUT4w4BfUX4CP683HXMDFNWXTj9c1mNH962pWU6nnjwR3HoM8Av6Mzr0/JkcwC89v+3cI+BXAi+j+3ZhLtl1mmi/DfuaOoC/IKG/VL508GjfCNeiCfzXbj7E4g1HETh/J9buOI27959G+bh8LOC35Dy0hKQLT8dOfkZ2I0vNHbsR5nwjuUvXbuONY05/Ju0KsCsz8bUHnXdkc6nI/Rgm2WqkQD79fQnZVau3Njr75i0GGfiXg88CRtcF6koIrlGzHSP4w+iS04+++a157tEGst3c+tMtpxc6cnszT5efocOXYzXtOxXdFxxLCmQKVRHYa9fpjNLGUag1atbqSFgXwLszIXcd93lAl57z7Jz0N1r8lswhaNi4F9d3pRRoNWbOOmO8+5WfsIwdnBWr72HAwAXmmsYH7ea5/qD950EetwNKlKqD/oMWGBcgnf9DovhR7fOlAb9A8+bNm2FFj37//SVtEq/h7NkzBj5Pnz6Fr66dOY0rVy5HKrwV5R9tLFl57959zJ09G/27tSLsd8YDwv5Tats3UsbTvUlldK1fCtP6NoS8+gX8qsgr4O9YvxymMGk3Nmj49bzt2bMHXn290WPAUCzddhAnbsmN5yL6D/eDR5uGWEvpzuMdE408R52aro0qwa9HI1xa7Uc3nmlGs9+vdS30YJ5C6OReFuxvHIvJfVuhffXqlPF4M7K/Ac9DDmIP3Xj6MlepQ8m6GEcZjyL7gn1BvxP4Y8mD7byMv92B3+jSc/jMNSzYcIx23yfw0+U7BP6/beZcEMvvwFuA30rEFfTbEX4l7yqib4DfaOKt6rtRA3+ccOhntD81bTJTsyhXWnYI0rNlYmcgG0E/R6LvkT3x98hB4M9N//y8jPTnJfTnYUJuwbSJ6HqTGuULZUelEnlRqmAW5GEnIDsBPzvBPTslONnYbOBXdP/vEX5Lxy/HHgP8Sso1wC/5jnT7kYHf6Pi5XFH+PKYAl2XNaUO/0fEb4KesJyLwE/YN8BPkFe0vls1qtrzHRPeZqFuU2v1C3KYgtxfwu7driKvRBP4btx9hxeYTmLp4LzbuPot7D36L8hH8mMAvj/vllMLIH19Rec0L3uV5rw7B2g1PKbU5R8nPfnrbn8d4uuYoqdaH3vvTZpyki81RKEFWx9H2GhUYRbcZRdu7MmIuj365/6i4lEB9AmG6B0cTJJ8ZMnSFkd5Mn3HCVO/VObyorZebj6w8xwZsM8m0lg9/uBWprk/XJZcf7wELKNPxQ3vab/ZmB0Sa/w3MJ1AkXpCtPIH+BPnOXcfRSnSqsdwU3K+hi4/8/KdMO4K1vC4dT3kJchtatPSGibqvZMLwBH7ekYz4L1j8C2VLMe+QpGv80oBfrj/Hjh3HL1eu4Mrly1i1ag1HWGbSXWkGbUVnYlzA19fG8LNNmDCDGvglOHHiOKvBRu2sFeUf9Cda+csvV1khdhoGeLTG8jHueLJ7ktG2r5vQAx50rOlSrxSCe7sS9NsbKY9gPxLwB33+pN0/+KztYUXbAT5DjBvP8p1HcezGU6zecxJ9GelX8vFGuvG83D+VOQmBmDeiA9wbV6JmvwmtN0fjOWU8O0O8MYia/T4ta2DrFCY3c3Tj8lo/BHq1gAe99Ke2HYQr4zbgUfAeuvFMQT/mJXUsWQ/jKeNZRthf7YD9zwn8GuGIWJ9Aj9Aff/yO58+ffVXtS/i7+kR/vu99Grn0hO47D9+QbRg7ZycOnXIm7b73TYwFOxD4r6MvI/xZGOFPwmq6KWidaaL8BPNkRtbzirTHAP/3iCsNvwH+75nsK0lPhAg/pUApub/dUsUn8EvWw+Vp2NJy2/SsfpsxwXfImvA7A/p5k/yA/El/QIFkP6BQih9RNGUcFEsVF0VTW60YYbxE+kSokDsd6lem3KJhVdStVJTSmNSE9oTISdlPdkb9JfsR6Odg5D+HmY9v9PqWht8Cfs3LAShXGsmFrBYG/IJ+A/9W4q40/nLqEfDnc2j45dZjEngzJkIhAb9D1hMu7UlCqY6gX/IeR1O0n5BfnE3afWPHKf0+tyuYOSlK5rWB/1K0Houzl25h2JRNaDtgISbM3YmrN6MeNv9YwG9HnQXqAnI53Ag+7eUWjHId4VnrLL3/SybkPqOW/Tfz3loevo+20fbr6fojz/zQrdb+9rEUqVcSrtbrXNrfksn8wemfJhdAchptI52/HdmPeE32sVSES+dTjQA18zm4TFNre+vcei+gN/p7rec1aJl9bruwl44V8fPY17d1R/hnf/U6YuL9lwb8irgq2r1r5y4EBk5j58yXnSp6lLNjpNEeybLGTdhhOmxK9JaU6m3N2m+7tR+LofmP2xLlPlqvDqHOY871Dvu87RqiWi+bWZ8hi9nZHYsRIwKwe/ceglbshf7rN24Y603P9k2wzL8bde2TDehunOhlNPseDcpghjcTXE1Uv50p0BXbgP93avYF+336eaPPUF+EHr3AyP4LrD/4E3p6s6ou3XjWs/Min31Zby4c1RHt65fFuF5N8Mu6MfidYL9nZn+j1/dqVQP7Zg3Anwdpvbl+DAI8m6F1pWoIaTcEV2W9OXUvNfvB8KjUBK2ZszSxvieWGxnPKBPZF+x/LuBX4uU15mCYQmME//uUcanTGRq6GatXr8XqNV9J42dZt2499jFP4+pVFo3iyKHz9e53QJKe/Scp6Vl1GAso6TlDznBW2n33+xdbtvzmVwJ/P8/uyCrgjyvgj2tBP8E/3K2HybtK4HVE+hP+aAF/XAE/o+WJ+D6iht8qvEXgZ4fBzHO/1GwG9BnZz8SWlVH+HIzq52Y0P59AnxH9gsl+ROHkP6JIijgoyiq5RQn8JdLEQ6m08VA2fXxUzJgAlQnRLsWyc6jVFaMHeqBz02ookysdZTdJqPO3oF8Rf5N0q6h/igRmdEFRfdlx6noT0mkoAxOB1VGQXEdOPFak34J8G/iNLaejAJcN/LY9pwX8iQn8VivMCL+aIv0CfaPlp7bfSuIV9DPaL9hXZN80K/pfiF7+BTMR+POkdUT4owf8suFct/MMHXoOY9uBC7j/6PNE+CPCqgA4HJRf/g367fXW1AJme5kNzDZkh28Tvp19rtet0zKtj7zO2tfe73XTv2//933+vo3dGbDP9+b3Ea8p8meMfH9ed23vs+xLA379MKr66YQJBKTuflBF5LUbfjPfoUZxNEq0jDkg6rhpVERNna3XNa3TvVInUiMvKmqmCsza1to38n729pqq1oK2lyRrFSVc61k4TsdSATn7XJq3m73MOra1jb0u4j4Rtwuf1+d4bka5+vabhsGD/XDo0GEokTS2vW6xku6CebPRt0tzLCXsy3rzGd14NjO67UnHmi716Mbj3ZiRfUX0LdhXRd7YBPwaRdpLGU8fb8l4hmDlrqMmQXfDoZ8wcNRYdGtVHyvGdDEjFrf5+RZRruTexNLsX6Zm/w966u+fPRCDOroYzf4Wfvanu6cQ9scisHdzo9mXjEduPI+m7sHugbPQtyZlPHTjmVCvB1Ywsi9HHhv07ennkPSog33mzBnTTp8+zZoKSzBkyDiaE4xiB3SU6XCr0/01tB6eI9Gvny8mTuT3R+vVp0+jrkIf2/72Puf1PH3+EvsI/LNWHcJ8Ft9SYNFZeOtzfiMfdu5vbty4Dm9G+LOlS2Gi9KkSsVCVDf0OPX+YY48D+AX4cruJS3g2AM33tn7fJO5yO0l8UrADkYotDaP56QX6hPzMCb838p2cDtjPo8g+Qb9Q8jgE/bgoQl2+QL9oqnhQVL8UgV+wXyFTQlTJkhDVsiZGjdyp0L5OKcyaMAjTxg1Ec5dyKJA5BXKlS4wcaRMS/C2HHbnsZGekX448gnwlG6vJaUgWnwL9PGqS9nBbW8tvovpcLimPRgHy8JiquKsmi04D/XTqkY6/ICP8gn4b+G3ol1VnYYK/mqBfBbkE/SqyZYCfch4l7BYk8BfIlAQl8qRBt7aS9EQP+BXRX7j+CMZz2G31tlO4cz/qH7WPHeF/Hzj9NNuGQ/inOV/MAvz7XvOnAP6ZM2cyKbskC3sNYgGmxx/2S+TYS5HtxYuXUiLlj7HjdxKyJXX6k3KuXxmZ30pZVggtW4NNPQPlh6jTpVEcjeBoFEWjNtZIkDVyoiTtiVMOoB9rI6j+wSAWZguZeYrQLmmWvb320SjQX2bkSFItuUbpPD1YD8KHCfBTQ46bToDupzkP97XOa40iRT63ridys9Zb+4Rva1+rta2uYfqM07R1nYhx4ybFeJG0aH0x3Pn69RuYNTME/Wm9uWZ8T9xh5FsWlesCe6Bbwwq03qRmv09DE9GXZl+gb7fYAvyC/YMH+TwMGAgvynaW7ziCI9ceY+2+M/Ah7Ht2bIbQSV54ylGLO1toADCiIzo1qGCKav20fJT5vIrmD2gnn/0a2MmiWo928Ltigq40+51r1sE047NvufHsGjgTfWq0QodSdRFQlzKeVtTst/k77H+uCL+eiadPnpik5dF+Qfz7CqQ0cRtmzj5r6qFIivhVNFZGlzHCbIfO3AAAIABJREFUlKmH+RswhwURx2HDhk3MFYre71V0/6a+lP0fsIDnsi0n4eW/xjj17DpyyenS86V8eRGu0wH8PZE9XSoD7akTx2ehKhv6Ge030K9ovxXxV5Q/EWU98rSP+z3BXxFzG/i5LllcRvspDUrJlpawn57gn9EB+lkcsJ+DSbS52PIwMTc/o/lFCNYlKI8pRSguRc27aZwvTZAuTSec8lxXOUsiVMuWGNWzJ0H1HMlQrzB965kYNm5IT3QlKBfPm4ma/qQG+nNS75/LEb0vRqAuRhecLCkTGulRAnZUkvIaVZBLVXQN8AvuCf15Cff52ApyeTFCeFFKbfKks0YB8nKZDf0W8CeirIewHybrEeBbTdH98Gbp+aXpl+e+7d5j4N8Af1ICf1KUyJ2WwN+AwH8xwtfz/rMXrt4xsO/lvwozV+zHr7ejLm7zzwP+zwvg7wvs0d3+YwO/otDTp01HqZKlGJkeHJZs+/5PrrXHL1d+Qa9e3vyHvMhE5rftBJ2OfqWL0zRUYkK3CqjJxlTF0WThauVEWPUelBui0RJ9ZgH3ilX3DKxrezkkVWeyZMlSLqbq8XTmiKgzYcu6BOGSbqkKc/Xqbc121am3rl27A+ToJMvYIazxoFEGS5oVueOoc6rzYX1fugb7ObNGGey8FWtUyN4u8khQ6Na/GOl/iQAWhuvpORyHDx+m13XsyIxT/YZJQYHwoj3lSlpvPtgWRG37NPrsexnNvqw3Q/o2csh3IsN+bInwS8YhN56+/Rmd9wvEahbVOkkZz+o9p+BJGY9HmwbU7HsaJ56HhPjZQ9sT9sszau+GCyt9maA71UT2Zb0pR569tOF8xqJa1wj7fh6N0b5qTUxvN9jhs7/LaPa7V6EVcan6JrK/vPXI10b2P2eEX391ypXx85tAZ7QAk1O1et1jbAy1/o5kMPBVNP596rdBo2r6zRjos4CFG32xY/t2p7znHX6sf6MP/4nzv2LlttNYu+ssLl67h7/+dLr0vMOti1WbOIDfEznSpTYSHgF/agJ/qkRxwyL9yQX9YdF+q+JufMG+Af7vDfALopMxcp6coJ+azYA+o/oZqdnPxMTczEzMzcKofjaCfk5ab+YlcBcilBcnHJdhEaqyeVKhbN7UKJ03ldX4vkzulCiXMyUq0M6ycrYkqEbYr5UrGWrlToHquVKgVv60aFevIpq4VESJvFmQN2My5E6fGDkZkc9NUK/AbTpUzAzvennQsnw2uvkkoCXndyZXQUW78hL0Bfz5GbUvwkh9KQJ7eUpyKlKTX5PnrsLjF2RVXZO0y+PlpXZfzZL1OICfHRLp+I20h/sWYivMToaAvzAhXxH+iPBvQ78F/NLvJ0F+di6K50qLrm1cow38d40P/0Ws2X4aB0/9gsf0z43q5QR+G8y+zunHBv4XL14gcEIgSpcqDV9fX/z2LGoJWVTPouD26JGj6NTZy2j1d+21cj9mzDoNtxb90cA4K63ESNZmKM+qyyVoaziKjkrTGH3XNpLrCJoXLL5qInkB1Pu3aDWIHYTm6MWRAVVYVjG2IkWrG6vW1aznoPwSQbiAXwXgXF17oHDhqizM5mt0/Eqs7uY+njUTGhvnJY00bN7+J6Of501dhj60V1V9BkUPJQVSYTYdJ2jSHiYab2Wxs3mmQNyc+Rd4XVcwmRFGRUztjsEqjkBYdSLOMtn7MbbvAubMv8gRDj9sCt1ikiajumefYp2KcClBtxc1+4tH041ne5BxolEVWQ8msUrGI9hfzwTddRFkPHZ0/2MAf+PGjY096Lt+fsG+8dkfOAh9Bo/C2v1ncI6xkM3HL2HQiNHo3qoeNnKk4hlHLJSToErBHrQVDaSt5q3NE/DX4RAcXchCee1qmwq6u6Z7Q1V1byiy790K7arWQHDrQbgVtM248eweNAte1VuibfE6JrK/ihKeiAm6NuRHnH4OSY/+fpctXcr6I340TdjHjnJ4vpXdOf2apvo7V97UrLkX0av3RIyfMJkF4O6/62P0j93uEZN2N+45j5HTt8J/Nivtnrrm1PB/gU+DBfxejPCnt4A/TeIElMAQ+lntNhWlPZL3JLcbo/yS9yRhVd34LLalCL+mStoV6Eu+k5bgn4GR/kxM0s1IO84M1PFnIPRnJPBnVVSfEpuClN4UY/S8dA6CPgG/QoG0qFjIamULpkHp/GnoWuNo7ASUyUPoJ7xXyZmcIJ4ctfKmQDVOK2VPihqFMqFsvkyE/eTIRdCXBCe/RgTypkSPmjkxvXNxrPMuj5CuJVG3WHqTSJzkx2+Nx39RwraphsvE2jIE8xrsUDTJkwytCqZEx8o50LRibgPvuTgCoUh/ZOBX4i6hX8DP8xVWpF+NoxICfiPxeS3wK9JvRfsLc1qQ2xdgp6J4rtTo0qZ+tIFfFfBmrThovPiXbjqG2/diTtKzgVEfZ4u99+B1owEfG/glwRk2bBgBvDyCg4P/5vbxPr+JcgbZv38/PDwGGAcjWbjq+hctvW4Kk8klaROjdKrtoIJlJUrUNuAuqHZrPoBJvdsM7EuG07zFQPj5h5pk37FMuhXc6x+9V+8QE73v5z3H1GQQ8AsC1m54wgTOmazdUNNYtyoKuHMPTFMV5VF+6w24yx5WQN+u3UhTP6EVLVlVUblFy0Gsz3DUSH+aNfdmTYl2xo5Wy2vX7khXqfG0gV3Bug1DjBXtitV3DfTLTrYVq1DLtlXOTQL++Ysus05DANau2/DZdca3bt3C3DmzMMyrPb3ou5tiUk8Z1V7HirM9m7KCLmE/2KuBsd6MqNmPCPsxCfzXmXPmTe19gwYNmOfAEtfv8JIDjSK5HTp2RuOWbeEXPAfLdx7BvPW70KX3QFYNroQRXephx7S+2EdtfmCfZqhXoQBa1iqBRQT/IwuGsKCYh0narV02H0a5N8TuGf2xbWofjOjqipolCqNrlcZY1n0cDg2bj6UeY9CbMp7OZeojwMXhxtPOst6MCPivzn8O4L979w7lY4F0NZvHvwd1mGPeGvh1v0ufc5kkgor2+43ZyMKJAbh4UTJaZ7Q6qj+lMOCfRuCf5QT+qO5VbF73za1btDocNhCF8mRBKkJ+miTUvCdJQOBXpN+CfuPcY6CfwM9Iv4BfTjfys0/CloKwn0agz4i+4D69GoFfLZ0D/FVIKy918pK2lMrFyH2+1KhQUC0NKhdOh2pF06M6gbxK0XSoQPgvVzAdyhZIh5IE/+LsFJRk00hAeVpcVs+ZFLXzpURtdhTqlcyB+hUKokntcmjTuAZauVZG8xpF0KVGbgS0KYoV/Sthx4iqCB1aBWO7VYZbnTKoWCIPz5kZlQtmQEmOIBQlcJelJr8uJUNt8yXF4PqFMNq9Ftq6FDURfuUCvAvwG+iPBPyWhj9yhP9NwJ8KXVrXizbwa6gtmJacPkHrmVxzGDfvRq1RfNcI/6YtLxj9eYYtzhYr74G+G31Hr/4j/djA/4T63549elIGUx3Lly+P1m+dgP8Ak+ks4N+N7QR+fR5FGPUPeiltWVXnQFH7MuXqm4i7tMb+Y1n0jFIfRe87dPQ1tRFasKaDNP6S7Kzf+MxE9zt1GWPqPNSq2R6qg6DhfQGO5D9LGblvw/oN6kQoqi+3JS23opuS60gS8BJLKC/q3HmsqRwtC9mptF7tQlvWIkWqcvkYo/1XpyF//vKEieXGEUjF2CpX5igDZUmq5VCJFZcV1VcysKpFqwq1OiDrNj41wD+PReUE/OsI/L99psRCubfIejNgXABaNOQ9YaT79JJhLDTlT/2+NPsVTVEtafYV1X9TZN8G/5jS8N+5cwdDBg+Bi4uL8c9X0unbXg8fPuBI0GgUKlkGbh3d0Wf4WPQdOQ6NWlGuVbwwmlUpBH+P+hjfqxEGtKmBsvkVRMoAL7eKZtmwznXQtFoRGiukQ6uaxeDf3ZWR/8bo0bQSSubKhmq5isGb8q+AJqztUbcT3IrXgGv+ChhfnxV0X+PG8yro2+8/B/DLocffPxAjR61gp1e/IeFys1d/S76m9+r8T5y8j3+jQTh16hSlc3+87TH6R6//7dnvOHTmOhasP4alm07i/BX68PM3wvn6su7AN3du38bECaNRvXIJ5MiUCumTM8IfBv0OaY+R99Cuk849yRjllyd/MkbzU7KlVlRfNpsO0E/LaRpCflrBPiP8mVhBVxaZBTMymk47ytKM1pfPnxoVC6ZFlSJpCfrpULNkRtQqlRm1Smc281WLp0cldgCqFMvAaQZWoU2DokzULcYofwVKfRqVzAZ3akZ9OjeA/4BOCAkaihWLpmLN8llYEOKHMX1bwLtRMYxsVgQh3ctjuXclbBpRG+uDOmJmoDe8ujZFff7Il+APePbUiZCbVXkrpYuPBgT+TiUzIKiXKwIGtUS98rmRl5F9JfNK729H+GXPKVmP8eM3ybuK7jOyr2aAX1F+NiPrsaDfkvZYen5ZdlpNDj2JOSLBCH9OAj+Hla9euRitJ+g+ffcPnLiCzXt/wrFz1/H4txdRHu+dgJ8guW3HU+zZdx+79z5wtlh4D/TdbN9J7S116RH/MX9s4FeRrBbNW7CoWUMDYFE+bG9Z+SbgF3hLNz+LNRuaNu2LHASsfPnLMVIewKj4PVM8bSCLl5VkEbOsWQsx4t6TdR8OG5hXDsCatY/gztoNgvBsXK8KzbOZxGfp6i37VlVc1qhAqdL0TmelZ3UEbH2/IEiAoGPNnnvBQHtDQnzIzNPYsx+mxoRyBGrWak/tfTATmOuiRo12HIm4izUbrKrKFSs2Y17CUiYOT0ZFVpdW8rCuwY2jAfVdu1MCRDnF9r/MOSzgH/9ZgV8jN6tXr4ZLjcqE3OII8mqMhSPaYwKry3Z2LYuOLiUcMh4V1Pq7Zt8GfXsaU8CvpHB/P3+OmtTG1q1b32lE6fq1q/ALmIAmHbvDZ/x0DJ84GwPHTYVL4+Zo41IWc/qzuBbrBCwf0QojO9ZA1SKZ0b95eWwc14E5CW0wuVd9NOf/C68mZTHfpxnWj+F3O6oVBrWqisq5CmBo9fZY3HKYKaAVULcHWhapQd1+PSxpNcwU07KB/m3Tzwf8QRjpu9y4Yf3jgH9oIE6ePIk//3QCf1Q/zw8pDV678yx8Jm7CiKlbsP/EL5T0OIE/qnsWG9d9c/vmrwjwHw7XOhVQvmQ+5MueHhlSJka6ZAkN+JtIP6P9xrmHsJ+Ckp6UlOzIT1+Qn45NkG9An/KdNBwBSMeOQUb63GelTl/6d0lYSlF3X4Yymwr5U6EKI/jVijCqz4h+7VKZULdcVrhQY1+7bBa4aL5CdtQqmxW1OF+b0xplsqIm4btR7eJMrKqFsQM6YvG0Udi5cS4O7FyO3duXY92q+ZgbEoDJozwxuEMtdKmWF+7V8mBos2II7loOi7xrYP7wZvAd0BautSsge8ZUSCrrUV5vzqRxUJVA3zBnMgxuXg5BPi3QoWl5UxRLPv151Pg5XgX+AoJ9SnoKUsojcJfeX+AvHX8hRwLv66Bf2n5L4085jw38uRThp6QnmsB/8RcWupq3C33GrsHMlQcgm86oXm8Dfkl4Qrc+x+49D7HvwB0D+7s4/+W2R/wMf29f7ud5yM/zEAcP3ca+/fcIsZ8W+BUdq1mjJjp16oQLFy5E9ai9dd2bgF/aeNU6WLrilini1odF0CSZqV69DQtxbTcdHHUGSpeujxQpMpjKzctW3DHRSllyKglxasgJDB6ylBWfe5vRgEGDF5sIu92ZkFSnTdvhKM4I//jA3dzX6ggI+u1Ok6xAp/E4tSjRcXMbwKJvFw3wT595EtVY1blylRbU+zOfgZ2GuvXdCftPmOT7yFSIrsgo/rDhK6G8AkmAWlPaM2zESpOA3LlLgJEtxSbgf0Dd/sK5c+DVpi4meNTBwiFuWDK8JXw71kSLaoUwqlMNrBjZmlKet8N+TEp65LIzffp01KlTh25Oi/Hs2bO3PldXf7mCkLmLMHb2Cmw9dws7Lz7AxlM30JeR/jG9muPqkj54sXUkHm4cilBCfsc6RbBsmBv+3OOP51tG4sTMnhjSphrmDmiK6ysG4PdtvngcOgwrh7dFm9JVsLadP24PXYeno7bgQPfpGFi1DTzKNsQSdgJWv4OUx+4I/LOB37Zu/jQjDGERfifwv/XvRxs8fEqr4V3nWOOHMsmQrThAi84//nj76No7Hdy50Se7A9/cv3MbswJ90b5BJTSjvaVLjbKU92RGxlSE/qTxkZbRfpPESzBOzZaWLT1BWU1R/NSE/FSOlprv0zOir+TYPOksCC5KG8rSeRiZJ+hXpHxHsF+d0ftaxTMwqp8RdcpkZiRdwJ+VBbVyoW2j0vDqXBd9uzVAz/a14c4oimcnF4we1g2L5wdge+gC7N2+1ExXLp6EMSN6oVMrV9SrXgZ1yhdA03K50bx0drr4ZEKDEtnQrXZBDHUrgZEtS6Avo/71K+RD1vQpkICypLh0F0pAx6E0lCIVZrEu1yKZMMGrIQZ0rkUJURorWZd5AVbSrgX8dnTfKsKViJH+RCgg6LebgJ8QbyfvWom7St5VxF9N81YrxBEA5Rvk475FKXPqHAMaflXaXb7lFKYs2Y8Nu8/h3sOokyjfBfg3E/h37RbkP2DBKg77MoossIzpFspjhm59Q3vD+d5ne+v4LynVkJVj5Ba69SW0/tXP9Mbj8zpf3Vbv33f71x3jfZepQ7aXsC/g12f4VBF+ySlCQ0NRsWJFk7AbXUvOV4FfGn4B94LFlynJWM/Ky8eMFl+JrpLRFC1W3VRdXrTsOgY4IvzZshdGPcK2ZDnzF13hfuuYjLjTeOrru9FxqlZrRS39UEL2NYfjjqXh780E3GLFa5ok3cXLboStW77yLvzHbGbl5fnmuK4Nuhv5kCL8SiyeRNvPiiyqVKNWO3RnpWeNEhjgZwdBuQN9+s2gS1BTDKbLz0Kes4t7AN83hkvdLihXvhFGjFprEn71WQUisSHCf+/+A6xYvAAzh3XGqdk9cW+tDx5sGIqt4zuid5PyGO/hgpWj2tBv/9MCv/4zmpEHSnr69e2HO6wL8LaXgH8mqxgHLliLXRcfYv+1Z9h2/h4GjKYFZZ+W+HW5N/7aORrPCPfbAjuhc91iWDmyBXBgHP7cMRpn53phRPsaWODjhlurfYBd/nixjYm/vu3Rrmw1bOgwDg9GbMRL/x045jkLg6u3M8C/uOXQLxr4LTmbJaeL+Jvytnlrv4iuVW82RAjf1upYq3P9KUYZnMD/tr+ayOt/ow//oTPXMH/dMSzZdAI/XbntlPREvkVfxLtvnj28i1NrZmDXtP5YEjwEg316okmDqiiYKyMyJE+ENIzuq6Wjc08GavwzcJrOwD9Bn4BvFdci+HNeUf0cLHYl15mi1NqXypWcWn0rql+Jev3KBP6qhdOihiOyX6d0JriUzggXQX9FFtNqyyjKtKE4sn8tfjqzFyeP7cSB3Wuxf9cK7N2xFNtCmRC1MBBjRnqhe6cmqFutNIrkzUyHoeTmvAVZoKscIboqtf5lsqREpbzp0bx8TrSrlBP1i6RHWUqK8tHFR7kKAv74cawmX351UhpXzIN+rSvClfsI8vMSxPMS6E3jfD7OWy0hp2oEfi4voAi/Hd13TCXtKcT7UIjJyYVN43yYxCcpl/O9ovuOcxRlLkGXGHDpufOAOuCDF7GC0L/v+BVoKC6q17sD/yPs3P3IwP76TR8pcZU67XWbXjIy+iysreb8uo0qcKSiRxHPq38i2v4Ft30etr3mw7e397Gm63mMVYz2Ll1+C4uW/BrWlrC40pr1Ku6k40feR8davf4Z9/sNK9k01fXpWNb1vDpl5d2Nuqbwz6BrWs/PFfn4ET9L9Ob1D3gPZUafGvjlbuHv72/0+ytWrKAONnpDvK8D/i2UuUxi8a06Ll0YVe9P55ztmBJ8yMh2ChWuQp/8SRg+crWx62zatA+6dBtnNPzNmvWjTGEtXBt2R83a7Qnby+jMcZYdhCCUojVnp07+HDG4aaBegCEJjzy669brhqJ08enmMcH490tr36ffHEbw25hEYSXZdmYuQNVqLTHIZzHrA1yHnHqKsfPRtv1IDOIogqRFul657gj4e/edjgoVGlHGsxTrQp9hyPDlKFCgAjJmzIPaLp1phXjWgL6+x9gE/KuWLmTSqgeuLvNm9vJoYO8YHAnpgYEtKyOAwK8I/+cAfuV5NGrUCNWqVnunUSUB/4x5ixEwdxW2n7+LPVceY/PZW/AeNR5jvVrg2tK++H37KDwJHY4t7NB0Yu7W8uHNzed9SbA/NbsXhrWrjnkDm+HGyoGmE/DblhFYPbId2papinXtx+DusPV4NnobDvecAZ9qbb944NffhEbWrMrnVtG5t4G+1lvA/pK/d6qG/ixS8OHV/Q3scx/Z4a5cex8r1twzBfWs2htWzsyr+8TUeyfwR0UFf1+nAp7z1x9FR9Yk6c58j637z+N3JsM7X1/WHfjmxaN7uL1vGR4cmIlfz27Bzs0rETLBBz27tkCJgrmRIVkiZGCUP2PSBGaaLjF1/QT+FAR8tZRsKmyViTaXOSl9KWC0+ozqE/bL50uBSgUY2WcT8FcyCbppKOVJhxqM8Av8a1CvX49yno5NSiFweAdsXRWI80fX497N0/jj98fMnX+ByxePYv6MMejj3hyNXcqjVOEcyJkxBdImjcdz0+aTGvxiBPSStNcsS8iuIL/79KxemzUlajIxtzqj9QXTJmKHhHIjyndSchQiKT9DIsqTErAlSxgXBbOmoC61ANrWzE070BQG9gsYy8zEJgKfj/IdS9KjDkAE4Lcj/Ab6HeBPkDfAT2mPbDcN3JupIvsCfwv4JQPKlyEx8vB4RelY1DUGJD0/X72LiQt2o1/AWpbBPhgjkh4rwm8BvyLIAt2Y+uG1j6Mf/9WE7pAF1+A35TRGTz5p2vDA4wicyQ7M2scm8iOXFu0j/bVgfNq8qxgedAyjJh232sTjmDrnCo/1NCxSZHugL2eF1sFDl6Fxo76UoXSlLKSbaU2byuFlu3FqsXXbobye9fyHFzznMo9/FMMCj3B6xEz9p57BguV3zXod23wGTjfwuhavuo/xM87D13H9vpNPwD/4DGYsvGH+AYZtz23tzx7d6SYe63MAvyp0NmvWDO3btzc62Oj+9L0K/Era3UL9/tx5PxvPfWnfpZOvXbsTvfEbMGm3v4m8e3oFGx2835hQLKQWv6v7BPN+IL38FU2vzCRZ+fcL5hWJdyGMj6Nzz7qNTwzo6/4L+CXZ8fXbyA5MW0beG6N2nU4cLfBAVcJ+nbqdMXDwQnYSbhuf/CbsXCgXoB2ThOu5eqB+Aw+OJOw2lYGVmNum7QispWe/JD1D6M7TmFIiX/9NVseCLj9y7ilYqLJxDZI7j3IUdB2xDfjnj+qGS4sJxATfv3b54eBUD/RvUemzAv+VK1fg4e6OPLnzYPPmzW997JzA/+ZbZCXthmv47SR22cvK3WoEO82z5zPfhc+mAXT+Lmpq/2ZFXGb/DU2cfMAkr8vNShWkX/ebp221bsbsM6YoXpt2w9Gi9WB2tDnqMmGH6Sxbcrr/z957gFV1tlnDme/7579m5pqZf9r7JrFjARuKiGJXbGDvvVew966IYkUFsaGCCkqvIsUC9t5rLLFHjRpL1KhRE7P+tZ7jUTRINJqoGeB6rr3P7mefXda9nnWv29LDoEBC67yYZnl+vti/5rN3wEjwrOtwmOFYdczW5bMB/+uviczm3Kct57YD57EodheWUj1w9PS1bFvOzE7URz7ts59+/AG3z+7C3QsbcffqMZzdlYZ9sX7YEj8PU8YPJbh2JtOfA0Xy5YAdLTsL5ZKe/wuCfTr45PocRQi6SxZk4SpKeFQxtir1+jVonynQXMdo9pWkS+vN8pZkXTeNE/w3pIa/bR1H9GtbDWN71ce88UwKmz8Yq5eNQdoKL+xMDcDZo5sI9o8hZNlcWqe5wrlkIRRjAa1Ceb4wYL9Ivi+ZEJzDFOiqT7lMLYLy2gw4XAWoeTyVC+dBLRbdqmGfjw5BORm0MECgNWh+JhPnJ8jPz96KfGx2+XOy8BVtMds4Y3DbsqYXQr0U5QXIydgrqVZN0htr8S2rtEfJu+Wl5bcCfq4j5r6iAfvWoRX0C/BzvkD/s2DCWYCficGVHfK/Fw3/le/uInHjMSyhU0/6rq+hCnlZ/b0tw/9HAX4B7JhVt+FDYO05ewOmLtzM5LqtGDVzHcb4bkNw9BXLw/oZ4N/AF0Bi6j1Mmb8PfSfEY5xfKsbOSsWQKYnwnrsLsQTkqqRqAXOWbmI5ujSh7KJgkQJwYrJ45bolUZoOUXZFihj9thxYrMBrE9ddTYbK03cHBniv5LaT4emXghF0Xhnhk45FDELE4j8PEPiiWcPqrUEMEMaS6ZswbwN8Ardi+uIt8GIxp8nzD2LV6nvPl7e+MN/H8EMA/ocPH2LFihVwcXHBwoUL37nCrq7RzAC/XtDr0h8a73olug4cPBf9Bswm6x5Ge0x2L/M3Wxy01yS9rqLVpX7zsMgzDOA2I4QVdSNjLlDGk2a09fLWHzEqyIB9ARptO+P5F6BIXXsPCxftNhV9+w/05778Wfl2BQIW7zIe++qlkbvOwsV7yOyHYdiIIBZyiuD83SZgSKSFp3ohAukCJECjir4R0ecQwLyAqNiL3OfPWMVk3vkLt1PDn2qCmTS6BWm6jiUb8Gf1tLLMk25/2dKlKFOmDGbNmoU7d+5kudL/JsB/+/ZtXL58+Y3vx4yAXwSKrkHViRg6bBFc3ToYB6mZfmss98ozcsPC4luAv2Wcz1eupxoV43kvNGrsgeo1WmDUmGWGvbc+I633mu4z5eTMnb8FHTqOMhI7JcF3Ya+c1m1MxyPd3/G0xtWyyrMR66+eAKuzltkm71/t34yrNyLtIe85SzE9K7DXe8K6X+ux6rOe88alJ1vDn+W9Y515n+YfBvDHCPDvIeBweI6fAAAgAElEQVS/ykTnd+vRtW47e/jnnYHPnv7yEx7dv4HHP17FtYvHsSt6EXbO7Y8zST7YuXopRg7yQLnSxWFLwG+bR2DbYtkpfX9xymAcbQmCJWMhiK1GZt2FRbIE+GspSZfFq1zpmS8Nf62yDAIYANQnwGrlVgq92lSD96DmWDqlO6JYgn0Vk6VSFvRlFcc+iJvtgeXTeyJgcj9MGjsADepWR9FC+VCQ+ywoByHtm8C/uA2r49Iys24JVuNV8SvbL1HFAP1chvUvx3nVGABULc5qtpTyONAtqAhlR7ZsciNSnkIRJigXL8TiWGTYPZo5YVJvF/Ro5Ehgns9IkwT6TWItgblA/q8AP8G+AL/R8fM8mORdK+DXkO0Fw28F/goEpN/Pze2psBcDhpI2GNDz3ZN2r9+6j7VMrolIOYBNe0/jNiU+Wf19TIA/LO46pizYA79lmxG+ai9iUg9gdsgmjJ61AfOC5arygl3Sgzw26XtM8N+KYdMiMW/ZasxZkopxs2IxfvZGhMcRvL8C+MU4Va/ZAsXL50T38ZUxamFddBlVAcXZGyVf96iYi8+lFVo3lS+QoVPWY+S0OMzn9gOC18AnIBljqLmeF/I1UtgjYX2ZbSTgVw+Fpk+YuxGBUdsRs/oAYlfvx5wVCly2Iy75zvPlrS+h9zH8EIB///79Rlbh5uaGnTtJxb+Hv8wAv86PfnfreRYLL1CuF7v8tK3XhObrha7ldW2oqTfI+pIXwEjhegLglmkvg33r76DtbCTIEMDIuC8BD+v2NdRystFUsS9JF6z7sR6rddnnx//seJ5/5vpmmxxa963hHwH4nzLXQm5K27ZtM7an8rJ/8CDr54I0/JL0fIwMvy41efC3bNHS+PGrdkNWf58q4D94+FBWXyvTeZI7TZ48GZMmTUJycjLUC3fr1i1Wk/0p0+UzAn7JGdUr5eefbnqgypVzRcWKdeE1IdIE0nKdCqScTrK2FEpwdI8puBajHxt/BSErjrEehTetbZvDiU5aSoLX9iz37ov7TW5Xy5YfYfJ7P1SoWA+yy13O4Fw2unKyUjK+5gVxX7pfVdTOhwnunuPDzfylZv+67x6aJPo5DBz82Hs2gbUspk5bZbYVGXOO0r+DLJB31txfabyfVbF7AXN7VoSdNPdr4JK9mJYN+DO9Ll6dePvej5T0HMEASXpmriK2kKTn6auLZX/+yM/AZ79QNPPU/AOXzp/B+pDZSJvcBUeXDsSJxJlICJyCjq0aEHDnoRwmJ4EymXICfQfb3HAqkocMOK0lyagrOVe2mwL81R0tzYWuN5bxvKZ4VsuaJdGvXQ1MHdoKwT69kLRoCNYEUusa0A9J83pRL9oDwZM7YcaQFujb2gV1KtrDvlBe2JCFtzT2KhDwFyEjr2MoShlROVpm1ibgrkawX4aJt07MIVAV3SK5KPWRHSiBulo5BgKl2BtQnM5BRdnsacdZqogNShTOhxJ2tM50LIDODRwxrV9NDGtf3lT+rUTQr0BGzL5YeFlzlqNsyKLff6bhl6TnmazHquU3en4BfQVCaho3jYCfwYeCAKs7j/IDVNirAnsh+vdogyuXVATk9/+du3wbc8K2w50PxwFei9kdm8iKlEcg/2oVoHn17+jRwwgICGcy4jakUQ8vwJIRgEi+82dIesTwh8Vfx7RF+xAQtgMJaw8jecNRLIrcTsZ/M+ZnAvhjCPi9527D2JmxCApLw+IV6zDJfyW8/bcgnNvKDPDXrN0KZarnx9A5bpgR3woDfWrBkYnk8krPCPjFFAvwD5+2EeNmrkRg6Dos4T5mB63GeCZ+Llh+OlPAr+lTArZgxco9PP5jSNl4FIujdrGXYgfi/yKA/+7duyZJt1KlSqYIkq6t9/H3OsBvuR5fYRQzBH9W0P/iurUGhrqWX11PgcHL1/iL9X4tE7AAeev2XsiwLPu0BBmv7v/Vz9ZjyLifF8u8fCx/BODXb3Pw4EH07NnT2Fl6enoiNDQUO3bswMWLF6FaCq/+feyA/w4DkgXzF9ACtYbpYcoqgPlUAf/O3btYafkR1Jv2W029HmpxcXFwcHDA//k//8cMu3fvjtmzZ0M5NocOHYIKqT158uT5z50R8Is9X0MppKpW+81ej959ppvk8vETwg0jLrmcbGe7dZ/A98VOUyBu6PCFpnCccl0E+pXAruT3erSp9WBdi9WvAH5d9wqovdgT4OTkgs5k9SOjL5jti3VXb94i9pZ5T4o1NrzqpevTZwYaNuxugpBmzfqZ+hty0opNuIK+/WZRptcFkti1bNmfx9sRvXpNY4XrSPRkLY5Ro2nXTZJAx+HDCt2S6XmxUJ9634KW7qMPf7Yt5/OLIYsRMfw7Dl9gjZ89WL5qH746czVb0pPF+fpYZ32mCnPWZLsbVy9j16pQbJo3GIcW98d+Mu77l0/ArHF9UK2iE+U0uVCiUG6UYqXcMkXzkgHPR7AvUJzPOM9UlXbeNBWXIrPOSrg1yfbXp+d+h7plMK5nPQRN7obYeQMQP7cvAb47wn26IWhCe/gOa86S5W7o2rgiC58UQymCcFX7zUcHIHn/S2dvQwlRAU6zlTSHLL0Af3my+C4EzOULfgFHavkF6ovm/RyFKNsRm2/PXgAx+0omdmCQoF4ByZAKk+EXs1/CLi9K2hGAU1LTrIY9xvaohjFdK6N59SJGb/8C7Fs1/C+AvknaJWAvT5a+PIMKtQpqBPnWJvBvkfuI7ZfvvqVJ/iSGvyzXVe5DeQYC/bq1ZuGtdwP831PCs/crPngj09CyU186mdSkR3p7zPDxoa/3WhbTuWReDNbf/GME/AtCXwD+hRG/DfjHEPAHhqZh0fJ1BPtZA/4atVqhtAvzJWa7YXpcK/SfXguldH12GJUp4B/2DPArmAjiPvwC3xDws9vTCvgXRe7C6L8I4Nd1s27dOvrMN0Xbtm2xfft2du2+H6Yna8BvBdsWEJ8RPL/5eEZwnXHcuu1Xh6/fl4CL6VUwDP3L27L2AGi+WeY3AoyMx/9bgF/n//Hjx1DQJfmGmpKns2pieOVsU61aNfzTP/0T/uM//gNFi7LGSf36GD58OJYtW2aYfwVuApjax8cO+PVCPf7VcWPPqUBGYPZ1f58a4B/m1gm9WSMgaEkQEhMTjXROAVpWLTw8HBERESxaNwz58uXDZ599Zto///M/I1cuElZly6J169aG+U9JScGFCxcM8Fdvj7//Cw2/JUfKUs16+IjFBvCrONzGLTCM+px5mwxo7t13hkleb9VmEGtcLDRGCLonxMjP9F1rcmUsgP+Hlxh+9WoJgPcbOBuOpSnnpVwulXUyBPbNPcVeAwUe6jVTztUwHoNsbpUTI2ZfFakbNuqBAZTbBS5hL0+rgahUuQEBfjQr525Apy5jUKt2G+Pi1YY5O7LAjYo5T4OGy/BgReuatVqbPK7N22QGQIY/G/C/7rZ5aboA/5b9lCay5zqIOn7V+Pn5PT33X9pR9oc/9AwQ8BPy8wEvPdajhz/i2rmjOJnOF8DSEdgyrQsOBQ3B9jh/jBzigTIlbFG2WD6UI9CvYG9DoG9pFelCIwArpl8JqbKdrOZgQ92+LTrVL4ORXWtj2sAm8B/RCnNGtcHUgU0xvHMt+jlXYLEtR9SuUAyVHGwZSORjrkBO2BDgyzknL8G+in3ZsKnCrw1Bf34mCBcge68qwAL0TgTzku5oWIqAv6QNAwEm52p+IS5rywTdogT4DgwEShfMiRL5ZTXK4EE6fs4vSmmQAwF/SbuccGFRrwEdKmO8e020o/+/QLvsRcsS1KuJjX/u1KNeA30m+28F+wbwG4AvcG9tAvgWsC/HHtN0np6dM+eieSz5DzyfvTq1wJFDB/Adi6GJeXvb9u2VS9i5/zj8V2xEj9FL4FitNf71//sb/oUP/S8+5/miVl0veW9vb6xfv54P/fNk+bbRy5wa5Y+A4Q+P/47s+F74LiVDTsAcmbSP8p6NlPRsIqN+zrwQ9FIQQFJ3cpyR9GzDsKnRmLtkNWYHSscfDy/2CERkxvCTharp2gbFnPOg86iKGOZfBx2GVEAxOiR16DQG0bHf/ErSI8A/Yno8/INSzT6mLkjCmJlrMT8kE4afPQILlp+hpGcTFkVsQ2TyfkSm7IN/yDbKgMjwp3y6kh71Dv346EfDCnfr1s1YcSYkJBjm8X09od4M8L8Kyn/fZ8t19DJQzwi8sxp/se5jghMlsb+wQtU8SSMkNxCrKABk2dab7eu3AL/Avph5ATsFXHKr6dSp02tb586dzby6deviiy++eA4EBQj/8R//0YD/3Ln5nCMg7N27t5GBmGCCDHrqyjjE+H58SbvW602sd0BAAF2XahoZi9jrzP4+JcAf030KhtfpjB5tO5HJ7mae1+XKlaOrU4XfbBUrVqTzUzH827/920u/s37rf/iHf8C///u/o0ABVrmnDE+s/7ffXmG7irlzAzMU3rLUn5BsbviIQNShu5QAv4JYgXLJegTSXWhfW9zemQB7nJHc6LqX1n7D5p+M9a1sZzMD/FpG+TPuvafBqUwNSnQSjPZeAYX1PtEykv2ox1XSHje3DgTzG7Ft5y9GjqPe2LZMlPf1W2+S6ps26w3lZyWyTsdQunBVrdoYI8ns92FQIjctBSALmYPTmOPtO4ww0qPtu7I1/JndK6+bdvf+j0jechyeNAOZtDgdu8j2Z/vwv+5sfbzTDeB/+gtFPdRjcYCfHv2A62d2Y1/MDKyf3h17/T1wLtkfEQGT0LSeC5nwQihHcFqeEhTJUMTyO1PWI7cZMf21ytKFp2oJtK/rjH5t6H3fqyGmskiLV5+GGNzRlUW2yqI2GfwKdNApTv28Emdz0ynnS3rif/Hf/4Uv/ue/zXhOfs7L6QL8BvQL8MsdiEBeycJi/+3oGFSC8hxV8pV8p4yShynxcSDoLyyWn2BfrTABvj0ZfUcC/tJM3rUn6Nf6CiSUh1C2GFl+WzL4PP4uzSpgtEdddGzgTAY+jwH8xp6TwF4JvFYG32j7JWeiFt8S5KiXIx97NtTLYUOJkw2qObLRItSFgFKVdGs4FWIBMkurwfPkws9VHQsy2MnPKsKUlbAg2IzpU2k1OJIWhN1MF7zYqzdtvTx6ol0nd7iwyqdDzR7IWaQa/u//++uH/3/+53/C0dGR2+3BF+Uk2gvOI+DfjnSji34ZmPx5kh4B+DvGBWfk9GSM943HxNkJGDQpDqNnMgCgvv+FA4MF8CeteYDJ8/bCY0woRk4NN62fV6RJ2lUwYJX0GFkG2daw6HOGESpKaVqNIp+jWdm8cClKeVeBgnRb8UMCX2Z6qenFo3Vl+Tlxzl708YrF4ElkzyaHYZB3OIZPX8vk3G+YpPuIgYc1AJFF6CMsjbxEvf4GjJqRgIn+CQw+Evg5FVMXHELS2gd8IVoSibMClG8774/W8N+7dw/p6el8wfqiefPmtJisi8DAQMMqv89H258B+AVMxB4msUKvmEZp9S0A/s0CB11L+n3iV35rknuDWNE35VkCsLYjkK/iXEowFjCaNiPFACLt06Jlzno/bwL4VWG2V69eBgw2adLE9LaoxyWz1qxZM2ObWq1qNfz973//FRAUGBQQtLOzMwGEErGvXbuGm7e+fwb4h350Lj0Zr7kTJ06wiJmHSR5fxOTxe+z5ePXvUwL80d0mQwy/Bxn+2XP8sXTpUiOf8/X1zXIoAD9nzhy4u7tDAZyV4RfQ/6//+i8ULlwYrq6uGDhwIIKCgkwOxIMH93Hlyrdcb/FLgN8KysWuvwr4JU9Tsm2FCvWQJ08h9Os/2wS3GwzYt+S+CGC/DvDr+ad8mv4D5xiGf/SYEN4/d1llmgy/5Hd8Tq+jdW0SE/CXhRwyPQVy2AqkHn/3PgUBF4xUp0WrAdT1rzFJv+0I4iU3UiAxYvRS9mo3pZ4/2hTaa9LEnXkFk4wblqphS26k5bZlA/5Xb5MsP//4+CccP3sNa7edRPrOU7j47W3ixeyk3SxP2kc48wXgZ2npX9SYxPvg7nWc3p2CjfOHYvPkTji2YCBS/Eegd9dmqOBohzLFbeBESU9pavgdyY47EzDXci6M1nXKwqNNLQzp2gCDutTHsB6NMJEFtHw9e8BnbA+M7t8eXdrUR53q5VCquC3y5c5Ba8zPkZN++Dn+9j+mCFbOz/8HOT//b+QSw0/AL0mPDa0/8wvsG8AvW80vjLVmQQYLRejY40AwL8BflmC/AnX2zhzak+0vQslPYbL7Av0FyeZLzlOGcqRyBOlODFDsuZxyEGqR2S/H4MWZPRbN6zljiHtDdG5ahT0WedhzQTae+QmVCeCrlMpv7DOrcViNSb4C8tVZMKs61xeAr+lsx8RkO7iWs4Nb+cLMQSiMupWKoF7lYkxWLo4Gle3RgMW9GrJ0e/1q9qjHoVul4qhRrjBqlS+KTqx2PHP6NIwaNRrdu/cwLzK9zN609e7lgU7dPFC/dS9U4Esjd2FnAv5/ef7wt74E/vVf/xWlS5c2250yhQleAvwLt30wwL+eLwG9ZFal3MCcRZsxZlIIxk8NwbjJSzBs/GJM809FzMpLxmJtA5mgDZvFJD2lb/NN+C/ciGFegRjjvQjjpizByIlLMWt+GuJXXTEAa70APIsZpbElM/lrQbcxmGBbDFO/yIFZufJh8pdfwtu+NIKHzMFq+rKn7bAsrxfQmvQfsGDpLoyatAJDxy/CKO8gjPZegokz4li587jRgSpA0LGIoRLTGxb9NSb5JmLUxEBz/KO9eWyTlyNg6W4mm7JwGZfT931bUJ/V8n804D99+jRaNG9hgOG//Mu/YPDgwSYR8H0/z34P4Lcyi2IEt/K3U7P+JhpXYSw1jW9gEKfrTOBgxixWoo44ZX4HLa+mQl9i/gQGNvF6yRgIaFxNSYiBTFLsRztOFdPy8JhODfI5c62tJpCZM4/OIx1GGd/+NrTnlOtI955TyCoeN+tn3GZmv+lvAX7JpwTIlSitIEyFz7JqCg60nP9sfziUdHj+LJC0p2SJkibpdfz48UbWs2nTJuNrL03/rdvfP0va/bgBv3o80tPSabXaygDaGTNmmPMhpxqr1OxTA/xDXTtS3unBCtq78PjJY5Njod8kq6YcBvV4REVFGXCvZ716dBUEjh071vy+W7ZsgSxNX6fhX8MeSj1bdQ/omTJ6bIhx6VFVat1Dmp7MoncjWFxOrHvZsrVMAbqQ5V/x2n5qWHlZ6SogaNCwGx2u/J7fV5qv613X/6atT6nRj0FpavhleysXK217J/chwD8/YDuGjwzE9JlJ1O53M/fZ/IAd2LMfdN06T93/OGOxO8s33QB+WeJGx1007lfDRwWiCmtieE+OY8BwxDD6LtWboU2bwSQr+pPY2sVj+Blb+DzIdul58yf4Pdpypu86jVnBmzAndAv2H7+cLel589P30Sz5mWj9p08fsz1i5bQnTN/9mT/kE9y6fhH7EhYi3ccDm707YMPsAQTuHqhb3RlO9vlRmoC/LJn+GgT6rdyc4dHODYPcWfmyT0sM7NkMXdvUQc+ODTFmcFf4TOyPWZMG0uazL/39u7Gab204ORRBIZvcKEDXn/ws7JU3x+cE/wL9/2UAf26C/jwE/GqS9hhJjwA/WfmCAvwE8GL8C1LmYy97Tmr0HZm0K4ZfwL8kmX+j1+e8QmT4BfqVrOtEzXxlJhZXKWOD0nT2qUIw34yVf2uUKYgqZNwbujrDZ9JgzPUZiXYNK7M3gmC8nC0bgTzHXSsUtoB5DutWLGIAfd3KRVG3KkF9VYJ5tkYE841cSqBR9RJoXL0kmtRwMK0xh41rlOQ0Djm9kUtJrlMCtbmd2hWKoHeX5tizczvOnDnDRNujOHbs2Fu148e/Iou8BYO9l8K1nScKONbB//NP/2YSuP6HPSfOzs4MJLqbYknS82sf69enk+EJMYD/QyTtKnkqde1tRMccIWucRBY5CDNnLqTMKILdzsHw8ZmHWb6LyKYmMIlrL3Wdl8kIXSfTc5TAK5nLLGUy1gLMnLWQDFggbfoWm20ELIrni+QAklPZM8Dqpjv91+JIHy+ccq2PkxUr4USlyjhVvgKOV66CU+wuP9moJQ4NmobtAZuwNvESq7CeppPEBsyZuxwzZgbwOAKYyBbMc7WCnxdg3rxwLF+xjd3IZPrX0Qo1+RJWMPdg/oJoc8y+vkFm2VmzWBhqxgKe82C6RqyjbOgrssDf8+X3ck9KZuDvTaf90YBfOl+5fkg6IX2wZCSHDx/mM4Ndgu/x720Bv8CDmHPZ98nlQ/acSiaUZlfe9nIUmUYb1SkshrU0+JCps6BCWNIBi+2TvafWU6CWsOo67Tq3GYnBbDp+xLBKr8UJyMLoa19qYvbHjFuOVgQQjo7VUKdOJ8Poi72XC0h/brNhwx4sALbABBXy3y9Xri6ZxVBqle+y50gWg69n+X8L8P/e060AQb0B9vb2Zujl5YXYmFjzDFAAIe1+xr9PQcNvPd7Hjx5DwUqb1m0Mu21NJtd1q79PCfDHiOEn4O/fzR0HDh20fsU3Hu7Zs8ck0o9kL7F0/fqs86CE3sz+MibtypZTxeJCQo9h8tR4Vo7uS+vT6kY+M2e+xeZWVrJNm/Ux17lAeYuW/dCvny8B9yXT5JbTvcdElClbw9xjU3n/LQs5TFcf9hLy/lFTcB5Kpxwx8+VZgK5LNy/jDLQwcI8JMho2cjeF6/yZL6AidzVqEFMMmsdAnTVRZqawZ6urqcuxOHC/qYXRuu1gPlcvEPBbNP+VqjQw93gSnwHDRixi8nIlvvtqmXXimFis4F73WTbgz+yKyHyaNPxbD57HwuidxvL7yNfZtpyZn6mPe+pnv/z8BI9+vMWo/zZ+/uWx4L7x7Hn0+CEundiD3bFzsNGvHw4Gj0H68ikY2qstmrqVR3O3cujUpBr6tKuNIWTy+3Zvho7tGqBNyzpo24KNbHXnto3QtUMTdGzN6rXNXdGqSS00cK0CZ8diBPu5kP8Z2C+Q50uC/i9MIawv/vZfBP3/jdxk+nMZtv9vZnoe6vrV8n1pkfSIsbdhIKBgQJp8JecWz6ehReZjT3a/KO07i3J6SUl57JhwTMmPPccdlXRcIi+KU8ZThr0UDcnEu7E4Vx2y8G2auCA+dgnOfL0foUGz0KdzQwPOXzD1YustrWGVYgT4bNUygHwC/cZqBPsC/JZWEg0J8Btwupj9BtVKkOnnkIx/PTL+rgbwF8YAd7r0XL7wTlfMxW/vIDB2L/p5R6JBu2Go4uKK5s0oqfIab5i+H3744aXtHzt2hC4X9BH/IBp+JXmxgl/sUfjOCoCn50hMmzoT8wn2584NIZhfgrkMRnym+8Fr/BgOZyM4ZD3CwnbBz28Jv5MnbdVmmuVm+S5koMDKmQToPj6z6QftiZmTfRA/aRH295+M89Vr4TqD1Jts52u64mzdRrhcsTzO1GuGS1Uq4wZ7rb5jPsVZt/rYOHQKAqb4YyLt7bwJdHUcCiz8ZgdhwYIoBhULmAcxkW0SggJTkLDyDAFlmvk8YcJ4TOfxLgiIMsei4587dwW8J07EhAleJiiJTThF8CjG6/2A/j8a8FsvmHPnzhngX7VqVfTv398Eo8r/eV9/bwv41bOi7nm5b7TvwIq6rLHQk8V7Zvmu42+VTkAxkgl+3U3hrU6dxpB53Gw0wJ27eMLW1oGMcHuyiCn87Q7TzWMp1++H+mTtZQs4aNBcA+QVkBpJmMAKgbqq84rBnMl9SDvs6tqOgSEDZ85PZFEu2QLK91+BxcYtPzOAiCfYoJSC21vF+ZsJNv5swK/fSCRCcHAwYqJjDMv7W7/ZpwT4ZTl5/Phx9O/XHzlysGYMmW1JIs+ePWu+5hUaFXwKlXbX9PWDAP+gmm3Rt2tP7DtASvst/5R/kZnr0us28wLwJzHA/cnUtZDsRYmxFSvVR1kC5UpVGqI9dfOjqIvv28+PgN6bCbQC8XcN29+l63gTaEtn37b9cFShhr6sc22UL18HjQjeR/Leikv49rkkUyy+7oHF7ClTMTxVzFZxPBXUc3FpYRj9aSxWp3oXYRGnjQOP3HVUUE/BhPT7CkDk0jOA91U/yjHjGIjrWaDeiNZMJNY9qOdDwOKd7G3zgGudjpjFZ4KsdNXLlw34X3dFZD79yU8/4/L1Ozh86gqOMmH3xu372ZKezE/VRz31syePHuDGlRO4efU4fnx4Ez/98ojAn7Cf0p4fH97FN1/twr5oXxyNnYoj6wIRuXgy5k0eTEA0AHMm9MLIXi3QqqELqlYqgzJODnCpUo4Mfh10YwJq44a14FSqGIG1DZNi8xvPexW6UoXbnGy5v2ByLocC8kbHT4D/BRl+K8sveU+Ov0vio2UtTYy/JD4FqOOXxEeAX0y/knTtmKxbhNp948JDKY+GpeiAU4XgvipbKer3C1MCZMsehYIcahsqKqZE5KqU6TStUQqDezZFYsxifHf9HJmhkwgL9kPXli5k8iXPKUp5TnEj0ZFMpx6Z/XqcVp9DSxCgeRnna5kXrS7lO3UrFTOtDofWpp4D9SQM6NkGl785/04XzM3v2S279yyWxG7BLNptRseuxIkTxxnUZc7wfFCXHnat6gEcFrGbQHihSbyzMuHTp/saML948SrDpgsw+/j407kihaB/owHhEznN128xeyfEqvtzeT8GL7GY7b/UgO/ZAwdgdYO6uFCqKG6yZ+cmrVUvVyiLEy06E+g3xlXH4gT4DXCqaXtcKVsKt2ntKtC/t4ozFnfvgsljR3O7c7A4cJVh+qdO9SFrv5y9DfHc1yxMnTrL9EpEsichMCiFxz8NU6b40FZuJQG/NKQzuLyY/bUMRuZx/hT2RCxgAaYjBiBa9eBZAcA3mfdnAX5dmEqMnDZtmrFDlN/3t9QAv6+/twX8FhnYTQM6SpWqTLa9I6b5JBs2fxxBi4CICgF5soiPm1tHY+cXuGS/Ad+lS7uYz4vYK6Dqns7ZNsQAACAASURBVHL2EJiYTs29HEBq8/OgIfMNANJ+Xv4dfqIF4WW0aj2YFX/bYimlAwreLBp9Ff55agCFPMB7uk+hL3lLU3nUCjZe3tbLbP8fwfAL8Cvp+m16ZP4IwL/Wvw+iJ3Wh5XJtBAbMxZ1M9PZvey3pe+3bu88kHNesUdPIzcRsy4bU2mvxKTD8yb1nIbW3L5a2G4MeVZtgzNAROPX1qbc9HW+9fEbArwR0gWz1lHkzd2q8VwTtM6OYtBvJ+yrFSG1UVE7XuyWg/YU+9+dM5Wklza4IO2VyVjzHW9eLMNIaMfcC4wqKrSSHnn1KbA+L+Jq9CSvRl6C9d++ZGD1mOd1z9pt5uo+0/DLmxKjInfIFRo0JNvMt9S8eYAmD9aClh3g897gOi9xFnTE9dVFk/HU/qhCe5EAKRuITr3J7ytn5ORvwv+WVcu8Bi7KxkOes4I3wD92K/V9R0sO8z+y/T+sMfPbgh7s4upfVdTdF4sr5vXj0+A7B/hPy/D8Z4P/DnRs4tycVZ7aH4cqJ9ThxIA2HtidjQ8JizPDsjaZ1KqOUPauVFsyPgvltYF/UDuXLOKBKxbIoTUBVKH8eJtrmoOb+S2OrKeAuEG9tVsCvAEDs/uf/w8Tdv1kSd7/kckrmtYJ+K/DPS4AviY8SeK0OPgUo9RHTL8CvJN3CTMqVM08lugfVoISnJps88B3I8Ns9A/uSBhVU8i8/qwegfsWSGNW7FULmjcXuTXG4dvVrHDqwGRNH9USzmo5wo4ynDtl4NTc1yXvU2EOgaXUrUtrDZuZToqPplmW0bFEy+UU5TfM5ZKCgoVrt8tT+sw3s0RqXLp57pyvo1IUb8Fm6AR58UPuGpOPcpcydK6w7OXLk0Afy4dfD/zGlF9f4omA34dI1BPOppgXSEWfKlOkQoF8clESAv8kA+YCFCQZMBy1dbQC2mPNFlPosphRIYFzrCGwvJECfPzMIIf1Y56G2C044Fia7nwc3GNidak45So+RuFC9Jr4jq3+heg0c9hiH8671cIs9PxdLFMDWymUR2aIZlgwchiULYrF0xSbD7k+cOMEMly6jD39AjGH1Q5ZvxhIet0C9jt9Ij8J3m8DAy2sc5TxzzbGqB2DhopU8vlWIiNzPrvNbz16ALwO+rMDg6+b9mYBf142kYB7uHrTLq4X4uHgGky/LQazX1tsO3x7ww6LbHRnEnJSqBOqzTQKtwEjQ0v3Gfm/KtJUEEcFkDpuTke/L30KFieIpT2pFP+4IhBMgqCegcuVGBBWhlB8cNQBEVUbbU4u/nJ7kmylBsJ57gQVJAuISrhLwD3oJ8Kt2g5U9lMxo6HAmPVLy05dARRV/rQDGuq3Mhn8E4H/b30HLv2/Av3Z2H6z264VlY9rBo5UrgpcE4of7WRf/epPjliZdGnWrjEfOZgL6GXuePnbAn9xrFlb38UNkF2+Ma9AdvVt3wcr4BJ6fl3tj3+R8vO0yLwD/KkreHprr3OowpQR0Jdia4bNxBa1WVyoBeI3LiWodwbaGa9LuW5Z/NtS62p51nYzXvNZXk4xIYDyWvQBKppd5hMC6Zf7PXPcR3YFuG0Z/JXvJtB/N07Zk4alm3a6W1XwNrdMsy1jdsizP22yG/+2ulIePnpDd/xYr1x9B8uavcPbSTUrA31/v7tsdTfbSv/cMfHbvzi1sTA1H7JIJOLwzBg9/ULITpT2G5X/KampPcP/7y7h38zR+fHANDx/coL7/LNJTI+HepQVKFC2A/HnptpMnD5NwcyFvnlwc5uRnDZmUSwmOEnPzcpg3xxfIm/NzMvpM1CVrL0Cf8xmD/yVlPFbA/zn15mpfMJFX4N8EAJyvZcT65yarn1u9AhwK/OfhUMm9tgTvxcnYF6MLjwOtOisymbiyknEJ9irb54WrkmyZfOtEplfLqNpuUWr8xfoXo/SnfLH8cG/LgkyjOyLEdxB2ro/EpQvHsHNzEryGdmMPAEG/AfzPQDtBvTT9rmZIEE9Q7yagbz5rnmXcMp2A37qcgL4AP5vm1TQ5ArYY0KPVOwP+qzdZiZSZ9KGstLvlwHnce/CiyEpmF8mHAfx6UdCJIeVbgt991Fwf4YP+mtG2J6xk7wQBtSQ60ynNWbg4AXHxrGpLLb6WWxyYaAB+ZPQhej9fIPA6ZZafMXMutfXzCbrXIpr6++S409g0diE21auPdTWr4IhzSZypWQv7Ry3A4b4T8S2rR9+g5eplZ0cc7j8JR3qMwolqVbGjshM2Vq+EA251sclzEbfzNRN0Wf3XfwmPx4+5BZFkkQ6YXIJk5hKEhe8xQcfy0O38DlfN94jnMUn6o14KyzGtMXkHqWtumpyCiOiDLF+/30xT0GN9Mf3e4Z8N+MWqxsfHkwWvjS5duhi5SGbX1ttO+z2APzH5ppEMVKxYh84cUdhCba509kOHLYIbLQUbN3anVKcvSpWqygRadywKVHXNJEpx2hqdcmj4SXb590Y+m2Jwqd7ajNet1xVVqjZlsu1kIykQOLD+NgIaAvWSKDwH/EbSY6mcm05Lzqjo8ww+5tC6tBuGDF1owP76TUpYfLmqrnWbGYd/RcC/hmA/1a83lo1rjyEd62HMsAHYs3fPW/U4ZHYtSZceEhJiAk9p1uUtn9nfRw/4yexHdvXGREpPBnXyQMqq5PfS+5HZuXh12quAX9eoAlOZFiiwtTZd85quxHdLT5YFcFuXl0xHTctZ19FQJgVZBbpaX+tkbJblXxAh2t/L83UfWfZvjsn0BFiWN8fDY9RQ95XuVy2j9a1BgqZnA/5Xr4SsPz+iS8/XF25i054z2Lqf6gfKe54+zQb8WZ+1j2+uAfzrk8MQ7D8cW1MX4vvvjuOnJz88A/yS9iiJ9xE1/veY+HMdt2+cwe5tSRg/qh8qlXck2Cewz5PbtLy5ciInQX1OSnVyPWs5COzVJN+xAv68RsLzdyPrEbOv5b+UnOcZuBfQ/5zjal8wINBngf0cXEZyoDwE+bm4Da2bi9tWsq96CpTIW4yVgIvSatOe7H5Z6vMr0Da0EkF/XedCaFe9OFpXI6NOZ52yZHNLcLmSsvRkwm9JLl+Knvxt65eHZ+/6tBKtj8BpvbFncyy+/eYrpMQHo0+nRkbWI5Bep2IxA9YF4pVwa20WUG/9LJDPRgce1/Kc9qxZlzFDTqvBY6vJHIL+3d8D4L/xA1K2nMTypP3YyJvz+7sPs7zq/mzArwexGJiVTHINZxJuZJRsDb8zD+gNm+l0o3mrLtJXeRcZ9PVG368kVwUEUQT5YscjozjkeisTz2N12l0CPGrol601Up/4pAtYx7Lvu+ZtwInOg3GkYQukN2uNtLqu2Nh7FLZNj8XJTgNxnWz/DQZ+10ra4UT73gbcr+3sjvWu1bG3fmNq/JvgRLdh2L5oqwlMQiJ2ch9plHKcMknGa9PvE/R9TcC/2xyrvos+r+HxrOHxxsR9xePZQHeWrTzOC+Zls5HfTy4YCnTCGejou6xKucwXE7uZn73AMoK/Nx3/swG/Lii5oIwbN47MeGVTxVM66nf9exfAX4l6Y+/JsQaQzKCmX/p8JRdOmZZIGdVauFLSI8ccAf6pUxMZrLSl3tdSzbNFy4H0BK+J3n1nGgnCRFb5HD8h0kgDlAiYEYBoXEmHiUk30abdULjV6WA8wOX6IWASSolCXyYxNqM1rlxOElZdMw4klqDBAlKy+l3/aoBfYH81wf5SMvsC+6OHDcTGDRvxgI4y7/ongN+JNQjk0LN1K9Hla/4+VsCf3GsmUgj2V3Qaj1F1O2NIB3fTY3bvlTyr13yt9zL514D/t6/RrK7f3z/PwvZnvf77O7ZswP92l8+9+5L00KVn2Ubj0nPAuPRkA/63O4sffunPHty/g/07VmNV6EzsTg/BrauHKev5nkCfLP/TB4zi9GD+Cfd/+BbffL0Ju9bTOWV8b9SqXBq2lOvYkN3PQzY/F2U7uQj2c5O9F4NvQDmBvJh8AXNNs6F+Xwm6+XJSx6/ptOPUfAUEAvaf8/OXbAbgG/BP6Q/nqSdAAUS+XF9y/ZwowGTc/GTlJRMSw2/dhpj+QpTpSLtfsiCLZRWmpSZZ/tqONgT6ReDRuDS61y+FpqysWo3Ty1DTX5Ya/6pF6MtOq84a9MR3p15/fO96GNutqinAFUym/6v9aTh9cg+de2i3R4cdeem7EfBbk22lvzeNbL9lXOA+w3g5fpb1Joe1BPo5nrHVpKWnbD37vQfAf/7KbSbt7oZXQBoiUg/i2s17WV5lfybgtySqPmLi4wWj24+OOUawf9MwMVbGxrLMY8Ogx8SfZHfvbeowv6G0YrsB0XL0kStOfMIZuuJsZzcv2f/VN6iLP0brzq8N2N/jl2p0+V9Tm7/fcwm2Tg7F6kHeiJy2FGtH++JsjRq4bZeDEp6c1PbnweHarojzXoi4CQuwZeAk7PEOxcGR8/B1w9YMBvpgR8BmgvnT7Dn42gQka9bdo6vQMROwJCSeo/vKLdPbEB6xj9OP8vjumGOMY3CgXgjLS0wvq2eNXdZyD5L2P5y9B0kpVwj4xUZJs/qC2XrT8Q8B+MXyp6WlEfC6EUBPxfXr17O8zt5kpgD/XuqxhwyZwGS7XQYoZ3UOxB6qi38YCwQ5M0nQa2KkCaxGMUmwXDlX45STSrmAdPqVqzRm4l6n54C/GgsHCdhLQuDea7qp1jmRVoFi4sNp3yoZkC+T/KQ9tgB+C9iQxlk++0rOrV6jOQOF6vTbX85epgPGHUhVRytUqEOnnu7M4Uhg0LCFCY07GKCeNfIIK/P4uu+lYCKaDkFjPRewMnYae1Tvv8mpe+/LvA9Jj2Q86XP6YdnoNhjYtjbGDh9IYL6N9pGZ5xO9zZcQu6/Ksy4uLkyin4XvaSP6ur+PEfBLs2+R8UzESNeOzGvogtjIaNy5c+d1X+MPma7A3VJpN4nX50Nz/7zu2vwrTRfgN5V2py005gNPn/78h5zfv8pGf3z8Mxn+G9i4+ww2M0fw0tXvKen5q3y7/z3f47NHD3/A6aPbsHt9GL4+lIK7t07wgXwN977/BpfO78Olc7tx9dIRnDqyHmnxs+Hv7Y4WdSvCvnB+w+7nJQDPmfNL5PiSzD5BvQH8OSyA30hvBPop6dH0vJwuC04D+DVdAF/s/ed/NyD/i2eAX0y+CRQ43Uh3uGxeBgkFWTDJLl9uFGI1XjvKcQrxs6ZrGTH+uQn+bbgPWwYEpexyswJuXlSlhr+2Qx40r1gQHo0c0beFM7rWK4XG1NfXpq6/tkNeNON47xZVMZqWolNGtIdX77oY1rY8hrcrh0l96yF22WScPLoFO7YkYfKYvmhB604x9obdN8DeAuRrCdTTplTD2hxmBPWaVlOA3/kF4DfLc7rAfvX3BPivEuCnbDthGP4NZPhvfyQMv4XZf4iEpPMEynuMPGc1gXv6xhfds3qhaDnpL1fS7lLSmciYQwYUS8KzmkBaoNlsi7pQse1i1iWRWcHE39gVO7BjagRON++E0w2bY++UMGwgoNuwmrrQsKOIDdmJOE8/7GxQDxdLF8M1+wI4Vb4U1rVqg5gZYUhdsR+bqOleT/eJjQmXsG9cEM40aI7jbdyxZuJiRFO+k8TeiJi4E2T2d5HlvWiOdQOBuo55VZKl1yKawYeAvlj+hJXnzLyXX5YEj+xqTl1709iRiu1fRdCvbfwepv9DAH49Ii/R/WTY0GFQNVdVf33Xv19oEawE80GDxlBCtd5452fshn/5HD4x3fQC7AL6det2YmJhkpk2Z95mI+Np2WoQ/byD6NwzGVVdmlL60dYk/02espKBiiWJd1HgHibqJhuHHzH9I0ctNYm2cgWZTP2/rDQF+K1APSL6rLEKlItJadpy2ttXMMnCHh5TIRegLt3Gm+BDOQLNmTPQhL0KbdoONT0Nkh9JEvHq97B+1neVDGI5kx9HjvLFFuqTnjx5/K6n9Xet/66A38rsS7M/pEMdjBjYB9u3b+f3yVpi+KYHK8viDh06mPoQWzZveUmz/+o2PjbAL7AvZj+q6ySMp4Xr4I4eiKV70p8N9nWeZNk5e3YALWQTTdJrVveb9Tr9KwxVZ2NR4G7K+wJw8uRJXj9PX71ssj9nOAM/PlLhreuWwls7vsYFEos/ZyP+DGfo0xgl4L9vAP/eTStw4cQ63Ll5HOdO78K2tDBEBE1A6IKRiAr0QvgiuqR49kbLhlVRkrp9G0p58hDs5yLYz/mlmgXwS7OfW4Cfw3wcitXPw3kC9TkI6OXII+Cfh70BWlYMvgIFK5NvPkuqw+XMNqj51/IKEsToF2QzQ1p62pDxz8P923CaLD5t2NQLUDBvDhRmQCA7zvJ0Z6nNarcd3EphdDdXTB3UBF59G2Bgexd4NCuHvq2qYFy/FgiYMQTL5o7CrLEdMaJDJQxoSseelk5mfOrQFkgM88WFswdxcO8mTBrdh1actPWjVEdVg90qyFqTjD/HaxPUC8DXKGtHmc4z8K9AQICfYL9WWYF+S2CgQl012FzYY6DWt1tLfPOOSbvXbv+ANGbTR605RK3deXx/L2s27Y9n+C2stiQw8dTnh4bvpEznCBlw+jJbGe8MrLaV4U9IPItlwenU5KcQONMrfb0SyjJqoC3bTaQv/zIWAlkSEI+kMb443rg1TjftgF0zE1iQ5z7SKaFJl8aTwzQ6OcSH7UcMC3RtbtsZ+2q6YE3PAYidFYU1ESewgQlq6QRcWj5Nmk+CyX0TV5Dpb4UdTVogmr0A4QT9YQwwEinFSdvwInnMvCgF+pOvsPdiDxORt5mgwAL4M9fp67sqiFEwEx65l0HCOSa/PXgOLt/0xfqhAL9Y1uioaFpZ1seyZcuME8y7PvZu3bpJZ6SZZM2XsvfmGrXEL2tvM54TSWiUJLhs+RFT2VYJtpJMidVXwCD7TXl9j6KdoPzAVQhoGBNpVdhHVoGdu4yjX3c093OJeSF70av3DKPL79ptAoOAVGPBabEQtASZkl7Jn9+Tzj8evXxY7Zat90yOTydAX0YJ2hECpzWmiqg756koV0/3qcZdRJ7iCk4s+ufMe3EE9pNSWW2aAYin50ycOvXHu7S87vd6F8BvBftLRrfFoA71MHJof2zesvW9XB863p8pHxO7r7oiU6ZM+c0icB8T4BfYT6KUZ0XH8RawT81+4srE9yJxet1vmdV0BRmBi4PY2xXC5/N1Bswv8lUy3mt/pXEF8KvXPaTULxUzWCPl8uVLWZ2i7Hk8A8IRcelHMMQnEaP9krFl31nmd2YHSZ/axfHZY9pynjm+HdvWBOLIzmgc3ZeCyOAZmDS8M/p3rIX+7apheDc3jCL73bFVfTiUKEK9PsF+rpzITRmPNPti93OoYi5bDsPoWzT7Ns9AusC6pDpqkvrkI4gX45+TQYCYfMPQcz0D9vnZytYL6AvMKzjI8yx4sFTZleOPBeQXJNtvywJeBfPlotRHcp+cHOc09QKI6bfNxcJgTpg5tisiA8djxfyR8B3fFVMI4meMbocF0/tiecA4hMwfA+9BzTGgJYOARg7o17gUBjV3wtA2zhjdhd3GoztgY0owzp8+iNWJYRjQvQUaupSitMeeun5571MqRFvPJmymmBaBvwC/QH1Ngf+yYvKftTIcliHYL0Nmn61q6YKoxtaHSdDvCvjPs+R1YPxuTFi4DhGrD+I6k3iz+vtjAb9AuezXLHp3MfuxlOlI5575C0TA6hcufw8RXFaFtSKjD5v1rQxrxvXWE5SvI+iOI4gPHT0Na5q2xMG27tjtm0yWnr0BAvrWoILsabryBzY+4ovtNMLHz0ZEdw9E+0YhNcWSQ2ACCi2ndQgm0wnINyZ9hz1TIpHWtS8iuXwMQX9S3EluhyDebD8jeFMyMh0lmLwruZEJVlhoS9NeDlas62hfj8kik+lnr0AY8wRWUiKUtt7i/Z7xu2Y1/qEAv5xQdu/aTQ/tBqbYjyw73/VPsp4dO7ZjzNgZ1NMnUFJzzzDsekln3uTmYekl0tCyDK+L9T+aYlry51697gfj5CGGXV75chqRNEfzBMK1noIvFeqJjb9kZEJyFVGi36v7VGBndSbR0DJudQrR7/kkwzSLc8k6/p6a/uL4fv1dtK+09U+MheDw4TMZWP758o6Mv93vBfyy3pQbT/DYdgT7dTFu5GBs27bttYWfMu7zTcdVKEz5I05lnJCampolu69tfiyA38rsr+jkhdF1u/Ld2gMJsXFM0P1zZTwZz7OkeRs3buD5nMVgdS2vXUvPmeW61/30V2q67yz3dNCSQ6xoPxfLQ8Lxam2ajOcne9xyBn58/BNOnv8O6WT3Jev5hpKejE5Y2efp0zgDn/305CEunt6DPWT4E0JnYKZXXwzu3hjurVzQtm5ptHdzQPfGrNDagl70NcvBtlB+An0Lsy/NvmHnxdBbAb+YfLH2ZOnF7hdScS0OBeLF6OfjOmoC8PqsoQICsfkKGLQ9syzXlz7fUoXXkhsgFl/svZHxcD19tmUBL0l9bHJzP5pP4G9XIA+KFMxLaQ8LbhXJiwHdGmJl6HSsjPKF35R+GNGrMYZ1q4NJg1pixtjOmDK8Hcb3a4KBbSrBvX4J0/oQ9Pdv4ohBLQj6WxH0d62GeZN6Ys+WBJz9+iCiVsxHz7Z10KAqC2pVdTDVhts3KA/3NjUwqCsLjdVzpjd/cQP6XZyYM/CsuRDoa7wqW5XStqjsWAgVHPKjokMBuHck4D9/7p2unLv3f8RhFsbYzqp4J85eI3OUtSTgjwT8Akfr1j94prffQbedU4bBtoBwK+h9MRQgWiepTsJJ43yzLHiD0bpr+q8Ar4AXQdIWVtHd7B2MpDYdEd+qPVZNDcGapG8JxsnsG/D+YvsC/OmUVCSsOo/gCf5Y5tEXYQtWGZZ9PRNqf3VcXD6N21iTcB4hfmGIpnRlHxOAd0+NYvB3y+zfbDNDD4WAfRpBfCIlPbIaDVm+ieBSeQrPAomXln1xbJI3WaRCuykDOm+s6X51PK9Z90MBfl2oZ1jQqXXr1mSzPUzxo3e6eJ+t/IC69djYBL6QZxmmXqBc1n6va1ZAraFlGYstnxWQW0G4AecMvrScdZ7GFRzos3U561DTX92npmm+JUiwBBuW61PXumW7uu5fnm8JEuQT/ur2rJ/VUyE/8dFj5rPuxCJcvHjhnV1s3uW3+D2AX8z+Gibohnl1wiBq9kcP7mfA/pOf3j2hO+N3OcQqz127dqWUqo6RY2Scl9n4xwL4U2m9KRnPhEYe6N++B+Jj4mh/+vrcg8y+yx8x7d69u6wnsoTJ5t5YTDtbWWnKk99yb1jul7/CuO5xVROOibvM+2wxJkz0MRbDb1Of4o84/5/CNpW0u2bHKUxevB4zlm3C3mOXsn34P4Uf7pVj/OwJAf/xw1sQvNCLVWUbMCnVEY2rlUTL2qXRrEZJNK9uj9a1HNCiliOqlyuGwoXyGhlPDgJzgfyMzQB2w9o/s+KU3IZM/HPNPuflYsvN9fJIqsOmoCAfl7MGAHLsEfuvdTTPhkm4ZtkcGue2uD1roJCfIL8Am6bbyOufMiMD+JlMXIisv/ZtR9eeJnXKY8LwLhg/rDO6NHdB+7pl0KORM/X8ldCPkh73JuXQtYETujcojR4NSsG9YSn0buyIfk1KY0AzJ0p7ymBo23IY617buBnJo39Daii82QvSsVEl2muWIIgno1/BHm0aVsLgHg0wyqMBujauAFcy/RVKFkA5NWrGnen97lQ0HwORPCbPQD0Q9vlZHZg2oR2aN8CFs2de+Yne7qO0dYFxZPgD1iEy9cAHStoV4KG/MkFsLOU4qowrB5u1aT9wupV1fwF2LQBKfsx3DOi1yGF2GjcegeVfAX5uW0B8Y/w3OESbzaONW2F1196ImLGCJdu3IyrhuEnk1XYzAm31NCQwYTiCspw1w6dgZ6PGSPTyRwwdf1LT7jzTz1tYWrMeXxBJq68hnHkEC0PSED1xDo62ZH5Ak/bY5xWMDfSOTlMXOI/Fsi/L/nS8qfTZV0GxoKBU4+STnHLVvEAzB/Fanz0hPD+xzA8INefrtJGrZDz+F/t45dzxe+7afYdWh5SMbHpZPqRj0cvalxKXCRPnsvv627e7oH5j6du3b5uqu3JK2bBhw28s/eazb9y4gZSU1ZRszGHxn8WsbJtI8L/mrdps/7Uswrb2+ToZP1vHX56/xiyfcdrb7vP3LD+LXeSSVHhNmENmP4YVYs9QpnITYrKtxaPe/My9nyXfFvArQTfVtxeWULM/mG48Iwn2pa1/9DhrwuH3HO2atWuMM8/IESNNIbjf2sbHAPhTKOUJ6zQB45n7MbRrH8p4VuL7PzlBN6vz9O23V7BiRRR71mZT/haLOZSgzQ/YhvkLtv5lmhLoff1Ws6DYIlY/D8SuXTt5/VzHvXv3PmhwndXv8rHMU+GttJ1fmxo/fiu2YF+2S8/H8tO81XF89uDBD1iTHIG+3ZuYxNEKtLCsQra5hpMt9eZ2cKUkpT4TU+vRgUbSk+K2eZCbYPxLgvYvJcMxEh4LM28A/98F6inZ+UKsvJh8ee9bGHxp+JWYKymPtPxqYvtVmEtAPi/HrUm/1kAhL5n83JyuPIC8DAzUFAwI7Avoax2tK4ZfDj4C/AL7qgGQl9Ml8bGnW09VZ1a2rVySFXULwq1sQTSvUgRNKxdBfX6vuvTBdy1nh0ZViqNNbQd0YVKve6NS6NfUCQOZ5Du0dTkMIss/oG0lA+SnjurGAKI7erSriwY1nODsYMsehTwozF6FMg52qFfdidV5q5HlL2fOYTlaQKqar2xCnexywZE1AkrRDrQUgb4Dm33ev6FE3r+jfdO6OP+OgP8iu9pkyemzjJUFqbn77tafLekR2JdG8nvKVL4ySbWyzRTYfh3YVzer3Hjksy/9u7z1lfQazWBBbjYvAX4C2HQ6qWyKYlG0wTNwtkFL7O/YGwk+IUzwPYy45AtGEy9//6RUlnPng/N05gAAIABJREFU8gb0cv+SE4XSFScm+gj2jvTH+dp1scVzHiJD6aISe9TsSwBbxyk2KIn5ARFR+y1JwVF7ERW6E9uYG3CiXS8mBbdgUu8ibGAPw8ug3wL+BfiVtCvnnojIA8aCM5FJvUauQ3D/a/BuYYVXM7chlpIh2X0K/K8z503LZwwqPh7ALx3/hAkTjKwnKirqvXbz3rvHIGbPLrL98UxuXsWWypbyF2v8TtHJbPFGWvHdd9fMA/x7Mr+yTNT5/RB/VsAf4zsUFxM88dOWWfhlx2zsXzYME3rUx/xhrZA00wOS8Ajsr+Nw6ei2GEwZj+eooUaz//jx+0nQzfj9JUFZEboCLVq0YOAegQcPfrt414cE/HLiMdab1OyPcOuEgR3dkcAaFncJMj+2PzltpaSo+GEoe5lWwH8Om//yv0ybw+8TsDCU9s5xzI85yWvnPq5+e9UE1rqusv9efwYe/PiYrD6NNFL2I3rtIZy88F22D//rT9dHO+ezOyy8FR48F52a17AUhqImXcC4TqWSTEItTm05pScMAKqWtEF5VqotRr/6PDn+Ziw0DXgn4FdhLUviLf30nwF6sfAKAIwfP4G9fPjls/85XXkk+ZEtp6YJ8BckUC9AOY6AvKWg1jOZj8A+t2MF+wLvAvPS7BcisM8nfT/XsQB+JetSu092vyCdfPI+Y/yNdSiXs82Xw0holDBbpWQ+VKR7j76PEyU/Ytwd2So4FGQSbhEC/6JoQRvPtrWKo2t9R/RqWhZd6pdGuzpl0Ny1DGpVLIHyjkVQumRhlCheEIXtbJDfJg/ysOBYPn4XOwL/0vYFyeoXQpliNnBmQa/KJdi4v/JFcsHZ7ks4F7K0srZfwqnA31GmwOfo2qI+Lp57N4b/+i1WNqTOLmL1IWxhgYw/14ffApZT1wm8H7Y42dBTXzIIC2h/FbRaQK586+WrL917It15FBzIpUdWmy8AvyQ8ZNMp49kccdIUyzpbqw6+YtXc9EWbEUmAHB3/lckPSKaTTngUgTbBeuoaVmbk9qTbD4vYZaw716y+hQOeQThXrwn2+MQgmX7+SsQNo+NPcip1ipTkmKJgDD6iGRykcBsKQqLij0Pr7pi/EV916IszdRrhwHB/Vp0mkNexSff/DJin0IFHycYJ3HYKbTtNMENnH2PTSVnJ+l/p/60gntIQc7xnLKCfQcra9Nf3jChw+JCSHjHQfr5+9L13ZTXkhe/NhSXjE1Mv5lu3buDmze/Yrv/Fmr7TDcMyZtTEys1G5/ZDyQ0E+JMSYthrNgjnYse9FvCnzemLNP++xnpzUDtXjGaC7q7du54n6P789CmDlsemPWK1zp9++vmdgkIB/Hlz56Fli5amR+lN6j98KMCf4D4d6/rNMTKeUXW7oB9lPHExsfyt72a8vD+q8UePfqT71kUjdTl4kCYVf3A7dOgQjhw5YtphSrX+yP1p+6e//ppuSBYZ1c9MOr17567R8H+o++yj+vGzOBgl7a7ccAyj2Bs5fv4abKNkODtpN4sT9pHO+uzu97eQELUYI/u1Rq8OrujXtQGGejTH0N6tWPm1Bbq1q4+2TWqgeZ2KqFetNCqXKYZiBLhK2P2SbL4AvgC8mH458VgBv9h/+elLoiNgnp9AWI46uQjic7BQltaRtEcMfX568xegNl/LqVdA0wzDT4beCugLKBE3f26jz7d7Buq1Tcl4pN3XNqTnNwCf0/M90/Nb2X47rlunij06NSqHBlWKoSo99yuXKoCKlNqUL1kQpQnMHYvkQ/kSNqjiyIq87M2oV6ko2tRxYi5DGdSvUgIuZYuhLBn8orb5UIDHkJ+tQMF8KFhIn/MiH4OQvMa9iMEIhzYKUPLnglOxgvTuZ9Vd6vddGGxUou9/efrAl2cNADXnQn9HuUKfo0fLdwf88uFfHLObN2UawuXDfyNrJun9afgF5mk1mQHcJtB1Jo1g//UJq08sCatMzA0No3Y98aIBr0pyXUnLyyh62lsAPxl39gIooWxL+Fc41H8yTtdryuq4I7At5BBS6HQTEXOQPQJHTbCgHgOBfiX8KlE4kvae4QTbcglas4Fe0zymA56BLK7VGHtnxGN92j2sIkBXISyx8QLqpgIw11+9hkmd/F6SJCmIWSO3IPpVb6O3+vH2BP31m+HAMH/2OJzmMRLwMyDReZAUKVq2nErCZW+BGP8oWYxS5iM7z3XcjmRPv2b6Lfpwzde6YXQ1io0/YXpMDLjPZJ0PCfgFTIOCglCzZk06Xsygpe+7F1T6SJ+V/6sO63sC/tjoCMwa0Q0Hg4fh8aYZgBh+jlsZ/uSZvQy7LxnPEDL7owb1pZXoFtr1PSWoB47TYnVNyiqkJkQhLSkO61bFISpsOUKCgw2wexOw/upJv3XrFqZPnY5WLemcRSvYNwFqHwLwDyebL8Af1dUbE1jheUiXPgT7cbh39+MF+6+e6z/r81N64Mt5Kfvv4z0Dd5gbmLL1BCYuWIupgeux68gF/MT7PPvv0zoDn93/4Q42p8diIbtu/SZ50J5yIIL8R2Dp/HEIXuRNu0NvBPiPw6wpw+A1ujcG9m6Pti0boGb1SijtYA/bgjZ07CH4F8BnAKDiWcZPn+C/AJn2Mo7FUbs6vaprVUFNlwqoWK4UihUuYMB/Tkp/1DsgmY4FtEumk5NyHTH+FscdSXNsNE4wXVDMPpvAtmxBC5JVL1wgr2H08z9j/7WMWHbj2MN1zJDzituyAFfFoujRvCL6tq2Kns0roVODsmTtHdGaoL4RpTmVHO1Quiir8zKJ1rVCMbL5ZdG+cSU01LwyReFQvBDs2YoWLUiQbwObAvmQn4C/AMcLFynI6QVgZ2tjwL8CgPwMMgrweIrb5UfVMsXRmMm9jXkMLvZ5UYFFnyoQ9Fdkq2D7OYH/F0yUfnfAf/Hb7xGSuBfTgjYgNu0Irv8pkh4LcBUTHhklkE1fecphLO40vwa1FrZfLPpVCxgns56Ucs1IaTRPdpcZAb+xyiRI3xq0iyB/NMF+MxwaMI1M/yms3wqCafUQHDBVecWGC/BrOytXfUOnnxQEBEQbO01p5DfQfzmdCaAHxj0D/D5xBPy04+T0JFpqBods5PIxWMYqucks6KXtGMDPhOMoBgDahsXq82dsYbGlo+5jyPQ3weF+k3g8JywAnvtPIcAX4FeQYUkIlUZfcp2vjGzJInMic296BDI/RwqWZNUp0K9gQz0FloTQl5f/kIBfoG3FihWowWJmkyZNwv37H6ZQ1Kf12P34j/YOmc/o8HCaGbTHjsUD8UiAf6f/c8AfMLw1kma4Qz77st4cM3wwffYttRjUU3Ho8BHMnDoR4/q0wfxRLHg2rivG9WxCs4IisOOz0dd31u+6ViTBmDhhItq1a4f9+/e/UW/Bnw34pzbpi2GuHRDUbjQ8WYBtcJde1OyvynaDyeSyl2RNLL/yPZSzkv33cZ6Bh5TnHT19Fcmbv8LqrSdx7vKtbEnPx/lTZXlUnz18cA97tyfTc388ls0ZgrCFYxC5dCKiQ6bR7tAPceH+iIuYg9iwOYgM8UNIoA8W+E/EFO/hGDSwO1q3aoiqlZ1RvJgtChUqgLy07MzP6ruOJYqieaPaGD6oO6ZOHIJp3mSGxg7EkP7d0ayxG+yL2RkZkAC/AflcT8N8DB7yCfQT0CtgENi30TY5bkdwXYQBhi2HktAU5IvDjoBfoNoyn848lNMoECik+QoK2DNQmPr6EoVtUJGOOE1rljIe/OPc2f3coxaGdamOEd1qYWi3emjqWg5l7POjGu0zm7uVRZPaZeFapRTKlSqCkmTpixHoFytOvX6xQihoVxAF7Ajwi9rCydkRtVyroK5bZVSp4AAHewYGXL5okfwoxOOx4fGV4LpulR2ZI+CEOmUKokqx3GT6c6Ey2f6Kdl+wfQmP1vXxzTtKeq6S0U/ecgLBSfuw/k8pvCUZi6wlBfbJkLPJljKdLL1V3vIyiy2wSgZ89XUm8+5kcLAXqwj85WxjwPVLgJ+SHoLcdDL72wj2v+oyCGdrueHAUD9spobfMN4C/FzmVcC/xiTAHjcuOcuZZKRk3lVJV4zsZj3dUjICfvnvr6PERuA8ZPlmuuusNsBf8iJ9t/UbfzYOQ88BPz8bS05+x60hB3HUYyxBf0Mc7uOFLaHHzHGlpDGH4Tngt56Lnwj6aVHK4CGcEiTJjOS7bzlPL4N4yzmT88tDBk+XTJCgHg/lRlh6Bl4s/yaA389/AyZOnIcr7zlpV4BfnuhWwJ9tcZfl8/aTmSlJz6q4GKyYNgBnYsYaSY8A/wEy/N7uDTB/aEsD9ge3d8P4kUMo49nNypuW0ptHj7ES8SQvzBjRA3tCvfD9lgAcjZ2KaQNaoT4NDiqXLoIFc/1/l45dlWE9PT3RqVMnIwN5kxNqBfwLolKx7dz32H35ATadvgkv3wDMG9MDVxM98XSbLx5u9MHmgP4Y0Koqg5nuwJ65+HmrL05GjIZPnyaIntQV11O8ge1+eLR5BlbP6oPetRshre9cfO+Thsezt+LwqFBMa9wX7lWaYDiLwQ3p7IFV9Nm/90PWuVRv8j3+issob8DHRzUrPN5L8b6/4jn6GL7T7XsPTW2fvpPjMHh6IjbQmlPyvOy/T+sMGMB/cPdqJIRNRfSSsYgNnoi4FVP4eQYSI+dgZfR8JMYsoKVlAFZGBiA+YgHB/3yEL5+LpUEzMcePzgNjBqKPR0d0bN8cTRu7ogODgJEE+rOnj0Hg3EkInDeZvQQTMWfGWExnoDB0QFfUc6tqgHueHErEldWmkmxlr2kB+wa0E9CL2S+QL49ZtohtfhShLagdQX9BAn31LhQmu65AwI6ymmJk0tUE+m3JvheiPWdRMvsORWxQkoC/VFEbVGXhq7b1ymJsrzqY69kK88a1wvQhjTG4cy0CfGeCe3rnM4ehqZszalYsCWeCfXuC/GJi8IsUIti3NSDftqgd7B1Lwq2+KwYM6ccEJx/MnzMVwwd2Q9N6VVC9ogOcHArDjseTTxIkHm+5UsUYcJRF06olUNPBhmBfgD8XKhT6giz/l+j5HiQ9SqaZumQ9ennHYkHUdly+nrXH87tIegQ8JbNJJniXBMaA9+RvDaOduVxFy5PZZ0Bgkc/sh9xr0jnN6nTzEsNPwCyWffvSvTjWZTBOU4JzcOgsbIqhTEbyGSbXyi9agD+SUpxoJv2uIYuuHIAYkzC8x1S/TV0tPb0lgXglK9qm05f9AKvoWiQ9cUinbCeeMhtVz5WrUBK/g4qDhTPBVxVwFbwYSY+V4adu3oB06fBpX7c57BjB/gScda2HIwT9W1ccsUiVqL+3MPwWH3YLsH/mxhN73AQh2q6sSy3zrDr+jEOLy456TEyFYuYaSB6k825d560A/5X369KTDfg/rQf+mx6tNWk3lj2/3ySMx1MCX+ycjYMhw+HVvR7Gda2Lfq1rUcbTB1tZVOunZ0mPX311HFO9veA9sCN2Lh+PJ/uC8d2mAIRP62uW9+xaB+7Nq2Pp4gX0n397eYuqOwvwq7rz0aNH3+jrCPCHRMRi9goClRPXsPXsbaw7dhljp/vDb0RXnI8ejQcbpuP22ilIm9sbfdn7Gze1M37e7ov766fj0PLhmOzREKFenXguvPAjA4O7aVORON0dPWvUR7KHL65NZpX6Geuxd3gIxrh1QR0H9iS3747EhBduPJIfvSpBet003VevLvtGX/YTW0gB3NChQ1G/fn2kpaV9Ykf/v+dwH/z4BHuOkXhKOUDgfxgnz9/A058tAf7/nrPw6X/Tz358eA9H96chJW42EiOmIylyJlZF+hLsqxHwR85lddI5ZPrnss0j2J+HmNC5iDHDeYhaPgehy3wp//HBYoJ7gfo5M0ZjoZ+naXOmj8LsqcMxc/JQ+HgPxpTxAzFmaA90blsfzk72tPjMwcJd0vZTry/5Dll5w+6TobclmBd4F8AvTDa9GAG3PYG3JEFFmEdQ1DRO5+fiZNBLknkvQVlNYYLsIpTWODCh1rmUHSpQquPMBNrS/Fy2RCHaZxZDD1pyThvSBEundMR8z9Zwb0nJEZOUJd2pTVZfzH5ljpdhcm5JJucWt7djK4xi7LkobF8UpZyd0K5zByxasgh79++ho8YlXL32DQ7u22KCm47NasKlnD0cS9qhMI8vPwMQWwYsFZyKo4mLI2o7FUKlohaWX2C/PFv39yDpufLdPSSsP4rFsbuwZvsp3LqTtab69wJ+AXMBdYFjgXex30mp18w0M8+A4ozAVWD/EfXrl0xVWSNRIZi36NK5rWfLS44jVn0lWe3osL3YMCseJzr2x9cNWmD/2EXYFHeRYF+aftl7PrEAfgJg6fS1TZPsSyZcwcdzhp7Hqoq2cg0KYw9EQtwp7BuzAOcYQOycEoa4aLoDcX25+Eh2o+OXTEhafsmTpKWPYQJxFJczkp6M343HsY7e/psjTxnXoNOs9PtVt6HYuGCdWSeObkMWaZP1XDBRmT0c6oGII8MfwQq7phgZ3XleZe6tLL8B9jzXAv0KrCKZ4JxE9yLLfEmOftuW8znDnw34P/2n9p/wDW6R4U+Mi8KyyX1xZPkw3Ej1NoB40/y+6EmZY3PXShgzbCC2s6jWo0cW600lX/pNnwzfkT2xe4UXHu1eiktr52CxZ3cMpC//dNY6We7ZAYPauSFo4VwmT2ZNRmT2NX8P4L9OqUjQsuXoN8oL0xcux6wlEZgeEIKuvQegb6dmCPLshugZ/RDOIozTB7elxLMSPD2aIt5voJk+f3QXvh9cMZJOdksnuCN2Zn9E+vSFV6+WaFS+KsY1dUeIx2REDZiBeV3HoGXFOqhVtTrCQsNovWlJEJXr0jH2fJw7d+4lIH/h/AWcPn36uduQ5FCStmzdupXJs5cyOwV/qWlWwN+gXoNswP8R/7I/sJ7P1gPnmR+4C0tX7qW851q2D/9H/Hu97tAI+H/AsUMbsS5pIUH/XKTGkdGPoowndCZlPTMQtcwH4UHTEBo4DSsCpyN0iQ/bDI77YPlCy7Tw4FmIXj4b4UtmYtmCSQjwHQs/AvzJY3rDc1hX+t93wbghXeA5vBu7f3uSFeqCHh0bw6VyGVO1NwfBfk4C/9wE/AL9eSXtoYxHOnhbMvdywZFuviSr/DqRJS9N1r2EvS2n5Udxgmn7IgVQghIaBwL+ktTYq5Viq+hUFFXLFUf1CiUI4h1QmUm3TiVtGQTQhrNicfRoWpEP+MZk+Vujf4eaqEONfS26E7lVpUSnsgNcKpREhbL2cCpNgO9YDCWYs2Bfyh4O5UqjRYc2WBEViouXL7Ir++nz8/sLWJHu+F7MmjTUJDpX5T7L8HgL8xht2BNRvGghOiBxP7Q9FeCvSIZfYL88XXu6t2rwzi49l64xgTXlIPxWbMbKjcfw3e2sNdVvDPh33cP2nfdYfdRadIgAlKx4OD3tBYRTyKILmGYG9gVkN7AlJp1HqPToDA4kARKr/2J5yl0otUlaTSkBCy0lx5L5nrAARxo0xpkGzel7vwwbV117ZoH5coAgmUtUtBJzdxi9e3jEfiQyYDByHIJ37V/7UYXfFREHsWTxOmwZ6IUzrnWQTt/piNBtiOOxSW6j5SzHpCBBdQROcLu7EBycRqC9j2z8ffM9rQGKhioAlrYV2Miehz2DfHCyRi1W/O2OlFmRhuHXcchKNOM6sv5UleCVq84beU8sA4p1rAWgHov1ryyr9aznSb0jCm4knVrNc2g9t7/lw58N+J/fotkjb3AGJEFJXrUSnv07w6d/cywc1R5B4zphHKWPjWtWxOCBA7Bj587n4PXgoSOYPnkCZo7ojsNRk4EDy3F9w3wILHdvUhWTWJdk1Ux3xEzphh6NK2P2zGl0J7r1Bkfy8iK/B/ArkXznTgKVkOW05Y1mcB+LMA4XBy3B4oULELE8iARWMAmGYIQGByKAcqPgwADEhoewBSM8JAiLA+aZXonI5UufT1cvhc/kqZg/i/UTArlsSAQiliyHr89MLAgIwJVvX/Smqa6EtOpnWKguI3Ovz8ePH///2XvPqCyzNVt0d/85d4zuO8bt2yf2OWfv031Pn929Y1ftXVW7otlSq8w550QQFck555xBkSgikiRnRMQA5pxzQlFRRFR03vmsj0+BslSCAvb7jrHqTSu9a31Ycz1rPvN5ye8XwH+d5SorK3Hx4sWOH/8R3gngNzMzw7ix4+hbVfoRfuHH8UmNDLyVX3VSxffxWC9Ou5c1wD8Ap1YB/mOHthPwRxPoB5LKE4CUjV6Ij3DC+hA7xATZ0lJvizBfG4TSch/B66hAO4T5WCHc15qUHUdsjHTDhlAX3tvAn1Z8N5uVsF+3EBZGM7F2+SQGopoI0+WTYWYwjWB/LixWL4DB4qmYMm44hg36C4YM+hLDhnyDoUO+xqDvvsAXn3+CTz6htf2bzzFhwhjMmzcdhkZLYUmuqIO9FRztLfmPhDEWLpqJ8WNHKkdgAfzKys+FwGek5XxJsD+YQH/Ql79XwH3M0M8wegidb7/4HT6hlV+080V5x2T2YHibTiLFZyxm839E40m5GTPkT/h+8KcYPoiqRF+Rw//5H/CnP/8Bf/iE6c//hjGTxiIgIgRnL5DH9lynLtBK0C9JNrmaWxqxqyof1iYLMXH4FxjyJcsT9P+WdKBPuFMwnJb/4RJl97e/agP8OrWeZb0A+M9cuo2A+O1Y45WF9em7cK3+zdvmXQH823c2IjO/Adn5D7CFUpQbE3ZTCYcUE1JidPSSdrxyguyishZk5DKYVk4D4jefxYb4GqWRL5r77cGvXAvQ3cbotbEJRxEeWUOgEIW8sVTiGfUD9jnHo7zgns6qz3r1ZUtYLq/kMVIyriEqthrBjJqbyAVFLoNl6UDyq7wKGBOYr4/bD1fHWGygHnYupSRTHIKpi83FhXD1CcL1dUt++SbpU0zcHnj7b0bkejryFjUqC70+n/5cxroL6WS7JbIaW2euRMlXX2LL7KVIiubfFsdBT1nS55ezhK3PK3qAuOTD3CmqQHQcdx/yGDyrbfeifV65FtCfz+/dnH6BfdqJiPXVSM3SUaj27G18Y+AtBfidyOHXLPwD8J/pD99lUV86fuwYAfAGhAd6IzLIF1Ehvgj0diN1MRC1tXVKjUd6Jg66isazZh72pbigdV8cbpPGE+u8HCumDoXTsh+R5bVMafWnui7EYgYrDPiAgF/6KBKn9xgk7i5Vfhq40LjLwGbquuEu7jTc06W799DQLnV8fpfvXuVtuHsft1nP9Rs3VKpnfQ33mIepnuBe1ITay6w+efJESa92VrESmVHxexEKj/6QvBLQrnNe/fuP6awB/oExm0LpqTt2DZsp9y1iIKcuktLT5rMzML5A66WMADn8D3FwbwkyNvkhIcoRG4JtCdwtEOyxllbqVfB1ZnIygbejCbwcjXk2ho+TsXoe6mWO6CCh7tgiyMOMz03gZr2S1vwFMDeajtVLJ8Bo4Y8wXjgWq5dMwBqqNJgZzMA6Kj+sXjkLxstmYK3hXEZlXApbCyMGbDGGvY0xrC2NYGNtwoAfHsjMTGEQlxIcOrwP5y+exdVrl+l4eAlnzxxDzc5ypKTEwc5mHSaOH4Uvvvgj/kwKzpeUDhVgrwf83/J6KC33AvqH0rr++af/gq9FJvO7P2Lu+K/ouDsSZotHUwd/CKaO+QpjR+gWB8MH/xmDv/szvv7qT/gzFyEC9r8ZPgQW9taoqd2Fhxw7gfiSFOCnvBhhv9yh4fY17pCEwmDeePzIhYbsFnxNpR6x+H//l99gxJ/+GYMpATqI2vzf/Mv/UFKdK2aOpdPuuR79Mq/eolWaf5Dhm2voTX8Cd95C6Tly5DAio1IQEVVNCU0Bva9AsoDMknJGtK1qwS5a+Ct33ENS+jXEJF+Go18lXAPKSL2hug6BsQ7wv7JiC5gVi31q5m1EJ1+AmXsBy5QiPfsm6xNg3bEdeZZDy35s/GHY2SXge3Jgl//jPyHJyA1lOaSv0ILeuYwA/uzCh4hMPAtrrwKYOm1GvDjz0kIu7bcHzDrA/xxxicdgYR6BOcPGYfonf4blai9kZF9nnzrWrx+HnMImrE8+BzuvXFi6ZSGKbUm9nesv52JBFjhJqWfh67IF674ciln//GtYGHqSZqRT++kM+qUNAfAJW67AM6QKqxzSEbiBln7y84Wm077/ci3584qbsTnrNoI2HIKh7Wa4hVDbP1+i7D74ecBfSq38wDI4u/R/wC8gSZQ72gMgAZ8PHz4gYLqvNMxFx/xjSg8fNhKQ9k2ArTf9YyNz0NT0iNSbRlJTJN1XSR/sSubqyJGj8Pdygce6Bdib7ITHpPFcKgxCvMtKrCKNx91gnArQJYG5JEBXqstCGM0YiRgaTLoTabY7Fv43faP2rm9HQAP8fTv+79p6EwG/WPXjs+vI4z+A4+dvccGvcfjfdfz6Sz7KcjZiR2kGogOtEOBiSNBuAA/bZXCneoGLNZPlYqYlcLVexrRUJU+7lQT4pojyt+biwBL+rmvgwWdOzGdnOh+2a+fBymQ2zI1nEtzPgtmq2bTqz4OFyRwCrPm8XkinXsq0mbMdh1XwcTVDELn/0SHOSCRdKDdrA/btKcKVS4dx/95VND++h6fPHtGaTiDw/AmdRZ6q9OL5U4KDBzh96ijiNkZgyeIZGP39V9w1IMD+9hMMptOtgP4hysr/ewwj6B/O53IeNfgTjCewnzL6L5g78RusmDUMy2YOw7xJgzDtx68wbuSXGDXsCwwb+gW+5a7D519/hs+//QrzVy7BVupJX6+/iqdsX1a5escr3Vk811/wf+CUftxbCS+ntZg5fojaNRhFWtEYOgSP+ZJ+Ap/9Hwz/lPr8lAAd8vtfYvDv/ycMZo/H5Qs9A/yX2yLtesVWqki7t95C6ZGIgzHrtyA0QmeJ1gNdPdjUA/4aAv7qmgcEqA8REHMc80wTYGCXgcStNxQQ7VxOgWIa2iq5AAAgAElEQVQCVLFuxyRfxCqXIqy0L0BU0gVUEtx3zi/3YtlOz7oB41X++Pv/5+/wn/7vv8WyJU7qWaUC/B1BsJQRy/yGTRex2ikfq2wzELLhOEHxo58Ach1gFppOE6xtE6i49Cn+89//Hb4bNIFb+/sI+HVSnvrvVvnbQHZ67j0C8j1Ybp4GS8/tyC548Jr6SRviIqOQqjshkTu5qByE/+uvf4HB3/zIsa1R3/Y6y72Uyci7B1u/XVhmkw2HwN1Izb6jKFDt+6K/rmD+nKImeEceZv5tMHWvQMLW69hT24jaOu4OVJI+1G6xIAuTAo6Hf1AJvH0iceNG70rfCSjsTZUeqe/kyZMQCoRcX79+DdXVVFjalsfF/zZkZeV8dCk7O4dBpCoUl7s/Av/X/c9K/t07fPgofNxdEMD/NxxJd8ezunhcLghCuPUiLJs0BB4rx2Obt86yXxhoAEkC+FfNGoUNUWEfzGn3df3XnvWPEdAAf/+Yh7f14j4pPTmVx2EbXAjnyBLUHKIOPwOXacfAGoFfPGi8i6JtSbTYG8DJbK5K9qaz4cDkyHsHJkfz+XC2WABXy0XwsluBYLc1iCDQD/MxJ1d/NbzsDdTCwM50LmzXzCKdZz4cyeV0sTWAG5109bKcns7r1LXcexHk+7tbIoQ0oTA/O0QFOSKBfgKpCQEM1BKNvTsycfbYDlw+V4c7t04xYMk1Av8GPHnahGdPm8kfa+GWqWyDMmhH6zNcunyWUVHjYMsdAmPDeQSK0zF3Fq24k7/H5LGDMWHUXzDh+y8wbew3mDt5EBZNH4Jls4Zj+dyRDDb2A6yMKKNGnqkEH5tNzulEOm79MPIrjBj+Fb6T+AGMOzB22kQER4fh9IWTaOJC45kEDGlTXmgP+sXyJX26RSfenIwELnQWYPrY75Se/1j6Cfzwze8wmqB/1Be/Vs67wz/5Rwz/5H/BeN4kXOkh4BdKj2/cdqzyzKLjLik9b1HpuXbtKhdLqfD1y1FguDMQV4B/ewsE8NfsbkJB6UM4BlRiziofLDWPhG80tem5M1BOukl7oCnXwtvPoQXeO/IQXINrCE5L4Bq6Xz2TdtpbseV+ezXoXHeIig3z8cdf/4oRin+JH8dMRUDIduygxHfnvgn43cb6BeS7eBbAz7cQnsF7sJm7Cp0t8FJWAHd2bgOWLnOlI/dvKRH4W3zx6WewtUsiLUl2MMTK/2pRoWvvGTamXoFfyE44uWTB2rWQOxYXSd9p+UkbUj6X1KNVa4IYo+KP+OP//k/4kjSwZSvc1cKnc/0yPsV0xo3bcg22vtVwpqyfU2ANIhLOtTk/d9wFkb7JN6dk1nMOdsMzdBfs/XYgLOE0qmruoK7u3msA/3PuMDQgMKiQAbKSFdWgN/+J6m3AL3SGPZR4FF7zvn37ERISTfqeK8zM/WBhGQJLq9C2FKLuLSyDeX5b6mo5yd/VMm/rw8+/NzMPgOk6d3h4BDKoWxnEwbM/H/JvnUhverm5wH3tQtQmO+MFOfs3ydkXy/6ySYOVdGeGx1JF4ykM0IF9DfD351ntm75pgL9vxr2rrT5+0oqT5+tRXHNGyX1foq+gRunp6ij2fX5G2r2D3LQNBPTzYEnQa00HLSsm29XTCPpnEvAT+JvPI6BfCG97gn1SfcK8zRHsuY6W/VXwclhBzv4SLg7mEezPhB0Bv6P5QrjarIAnaUB+bmZUbrAm99NGWfGDfMjz97JUgbx8yff3cVwFbwcj+LmYcMfAFrGhTohnwK9M6v+X5sSiqjAJB2pycObYdty8dpRbzDfwuOUu0308efKQwPopR/EFnnIBcIs0miNH9xEwVGHH9hLk56ZjS+pGREX6woExAMyM58LFYiFCGWAs3H0pojwZOMbPGDFe3LFwWAiH1VNIPxqD+dOGY9LYQRg1QgD/N/h+zDBMnj0Vdm72qNpVicamu3jCdluFwvMawK971sr+Pcb1a+dQXLAFHo6rsWjWaEzmomP84H/Dj9/+AaO/+g2+J+gf/qf/j+mfsIqKEVcunu/Rr0ICbRXvYlTYooOo3n8e9xkS+01HU9NDpGxKhZPTBmRuu/UTa70e8O/a8xA76LSbuPUK1pI6Y2TmCCNzT5h75EEs4ELraQ/I5VoPZi29tyN0w34ExdTBjsA2IpFjQqDbPr9QegrpNGthnUgQ/iUs5o6A+4oxGPLZH7BkhYdSutGB+FcgWO6TM6jjHLILUcGMLBxdAXe/ckQnnUdRp/qF/15BBaDg0B2MlzAdKyd/g0T7mZhCH4uZs1YjIemUWnDoaUOyGJH+yW6BT8RB+PoXIymiFIH+JXAL3ov0PHGwffXNsuCR/NEb9mHIkCmYMuwzeBv8iCmDf4evv/6BvgWn1AJByugXFfLN2woewj/6ONxJzdlIeU+vsN3wCD+ArDz6CrC+9mMkuwEFpABFcvycAncijlGH/aNq4RNJP4qSywTI0qfOFv7nVEU6zwVdGjIycl46B77pN9GVd70N+GWxLNrcubn5cHMLJQ0pmXSzXYhPoozpptNISjnTlk5zTE8jIfnUG1NisuTTlUluK/umcomsT97r22lfRup6W3tdf38S8YnHERO7H17eWVxUhlHdKRP19be7Mg0fLK9s4h89dhR+nq7wMl+CmngHPNkbi/rycCS4G8B45gh4GE7AVo8lKAoyUlb9Ag3wf7D5GWgNaYB/YMyYOO0WVJ+CW3QpvDdWKIlOwTnaMbBGgID/NrJSImBnMhWmS0fBkpxLsXYrwL9uJhcCs+FqRaUGh+UI8VhNsL+OoN+UPGVjeHIB4E6w72y5gGo8c0jlkQXCXN4vgrMV6T8C+u2N4OHAvAT1no5MPLvbLVd5HNfNg8PaOdwRmAMXq8UE/asRQr+AyAAbxEW4IGm9B1UPKBW6JRQ7y5Jx7kQF7t89TQmz67T219Pa/4Cgm7QaOs4+xzPCfh2jvpX3T5+0cHHwkPzTO6QFXMHhowcoIVeMmsoc7K7YirJt0chN8cO2JB9siXFGpLc4Gy+hU/EcrFoxFfNm/0DlgCEYPWYops6k07H1WmRs24prNy6RXvTktZb9zuBfrPyPm+/jxvVT2LOrAOujPLBmxXRMH/MlxinQ/3uM+fq3GPn5P5Pi879hsnAqaUw9A/znrtxBaEo1LANzkbCtFhKI602H9HnnTlqWnfzJ8y5RYFToLXpQKoC/Ujj8BPxFFfWw99+BxaZBWGNpB6N19lhmEY0AWtiFi97eqi60HaG+OATuxQq7HALZGgTH1GINqT0W3tVI28aAWm2gWkCttCmgeByD1kjkZau5dPZb/D1+86v/F599PhKh4Tvb8ut2EhQQZrng2JNYa59N1Y887hIVwtIxEw5+1TorP9+/aoO7ATl3sGixK2Mz/BoLf/gTglaPw7hvfoPf/eu/keaTyAUHqUBt3y7Wdykbl3oNJg55sHVIQxylNr29c7HWMY+g+7wC3/pFiPgAiIPvOvMY/PKXv8bEQb+Fr9FYjPniH/E///svsWo1JW5z6tuoQzogX84xSk6/iXXulXDw345wgj5H/0qscipC4PoT3Dl53AHwywJhc+Yd2NO6b+lVgpD1dbTy12C1SzFiKdUpgL+iHeCXMZJdjYjoGnh5xXAhvJc0GZ2E4pt+E1159z4A/6FD3BFyDYKdfRKDlzEOgizWOL55hfexiT4SSSmn6e/RgIodL9QiTd69LsmOStVOcHeFztSUc01IPkHVpOtqXmU3Sd63L6fPL8+ycuvVAkOVySZtjb8FKVPRqUz78t27fqH7TXAxuo3ULt+AItjYBkBoPuK30N+Ow+Ts+3m5ItBmOWo3OeNJ7UZcJmdfaDwG04YpsJ8lNJ5gAfuGELCvAf7+Nov9pz8a4O8/c/GmnuhlOSNTd2F9xl4VdVez8L9pxPrnOwL+egbaCobVynFYs3AYAf+PtPJPhN3aaXAyn0kQPJ9W+BUIcm8D+55rFdj3INh3E54/Lf9C/bHnToD9ullw4k6BM63ojuYLeL1IJQda/2UhYGc6S+0A2DDioqXRZFgaToK18RTY87mb7VL4uq5GsLelcgIWZ+BwfxvlI5Ac7YryvPU4vj8HV89Xo/7GEXL7L/B/iDfIlW/E06ct5JPR4v5C+PM/fzxtfcJFwAMqINzEtStncObUARw7vEdx7StKspC3bRMDpSQgNSUKkeHejExqCVOqATm42CBlaxJOnT5On4FHXGC8ybKvC64iVB9ZGDx61IDGxuts8wodjUlnyYynr8MKLGDwmckj/oSx3/2eVv7/o0C/CZWLegr4L1xtQFTaLjiEFyE5fz9u3nk7aLh3j+o4sfFYZeJJYHRZgUQ9eNcD/prdD0mfuYrVrkWYvyYKBqRXLDMLxgLTDbDyqVZqPHrwK0B5O8FYBq3glj47yd3Ph4VXBexIPzEimDX1qFI0FgnCJaBUAX4CU7GOj/lhMWMy/BLfUKXp20//gP/+X/4z/vVfP4OVdZxyAtb3S8pIef+YY1hB7ru1WyHsPAphYp+Dta5liNkkUXI7Av7NWy5g6vS1lHz9J3zCmA6D/vQH/JYxHv7HP/wvLFrCgHOZ11BJUCf9F0u81C8+B8YOBTCxy4KzTwksXOgr4JgPv+hjHI+mtkUOFyz83q0sb2Doi3/6p9/h94y5MPSLf8O/MJ7Ef/+HX2HSFGNKgp5WYyu7DbodkVYkpl3HGrdy0p3KyOPfQfBfhmW2eXAKqmWU4Ycqn/RHkiyiktNvwcp3J+ehGDa+VZDdk+X2edw12Uun3XsvAb+MjwDYrZk3aDWOR1hYNPn7N/jHITba3jt6G/A/YDCmoCDSdizDKZ14sQ2wt2Bj/BH+LUZgxkxTTJ1G/XT6eUTF7FXUJx1wFx+Mtt8SF20CzkXeNHXrRTg4bMKceZaYNMkAixY5wNuvUC0eZDEkZeR3onaAWG5b3h14eudh8VIXTJlqgsmct4Us4+qegbSMqyqv/AZlgbqdiwmhmslZ5wOiq0f6oxL7IAsOSfJekjx/5cshC7LWlwsPyZfNRamL6xYuwH0o4Xiu9yaqhzVx4wWHGOjKkzQe1zXzcZDSm8/3x+PO9khE2S7BwnHfwXEp1XheA/Y1wN/Dwf+Ii2uAf2BMbnPLUxw6fQPZFfSRozzneeIM2Y3VjoE1Ar+4f7ceW+ICYL7sB6xZMBRWtPDbrp4Mh3UzaHWfS7C/nGDfhNb9tXTUXaOUezxp7Xe1FmdeAftzSP2Zpag/OsA/Vz2TRYD4BIgvgN2a6WrHwHqVgHzhyo+HGcOzW9Cpy47UIVerRVxEUJWHYD+ctB6x8If6kPZDZ2BvB0PEhdigNDscu0o3Ym95Io7tL8TFs3tx/TKdehsuo4XUHnGgFcu++hE+5w/xBa39cs+krvGcM6NPuh+q/PeF0IGePWWY93u4y8VPw91bqL91lRrIpyk3V0vgWk7uehXvz6CJqjzPaLV/zqiSL0F/qwD8dvcE+mI1F8AvDsZCO2qmlb+lhX2k43FDw3Xs3lmMjQxU5mixHAbzx2P22G8xlT4G1iaLuRDpmfayAPx8qvMk5tIRtfYc7ja+OfCW/ucqaj1ubkHw8MoktUdviW5VKj1i4RcOf2HZLYRsPEFlmDq4h9bCPawWriG1VJY5iSxSU3QAm0CZAEooLqIoE5NyiTsAJ1UKjD3N8ylEkHKTtu0u6xYg+wpwpW6l+o/zFgIsZ0ybYYXpM60wa7YVVhr6IDhsBwG/yFuyT21lpPwmUnoC1kv9p3Rp/SmExJ2lks2dDvVLO2Jhd/PIppXfBTNmWWMa65/J84KFjnDn8yx+twT+al+/9DMs/hz8aXEPjD0Ff36LfEf8lutqV0OAouSXc07+PQQGV2Clka+ufn7DjFlW/B4n2JMGpQeMpRXcpWj79qz8RkVxCpT+s16pW9pav+kS8kkn0ufTt5HNcY7hOxlz+WbpU2DsCcqlnsfuvXcV4BcQK7sHuQWNcCNQld2bmpqdHZRv9PPe03NvAn752z139izWrrWCj1+BAu0CkFPT6PRtEsTF4EJMmbKKgH8VBg+ehLnzrBUdRtSdxIovcyAAuoCO5bmM5yBg39JqA0aNmotx45dyrtdh2PDpmDBxJSKjd6syevAtZ1FTcnROxajR81WaOXMdZs+xwA8/LmLbC2BhtV75aAgwL6JzdlrGZSRy1yAt4wrvm1Xb0o/cwnvIZZ8yqUi1afMZla+QzuL5lGCVvhZxJ0m/0BCndvH7kHeiLiW7OLFxh2Bu4UPt+D3KF6inc9Qb5UWNx5eWfVdy9oXG87Q2jpz9cCS6GWAVI+46EexnUnqziGo87S37moW/N0b/461DA/wDY24fPGpBye4z8ImrRBDj1uw7dpX/NmmAf2DM3qte/uJeQz2Da/nDYsUPMF08ArbGkwjSp9MyP5vKO4upq2+MQDcTBPLs52xIac4VBPsLFeffyZyAnlZ9e9MZcJQzKUAOZrp7h7XTYb9mGmxN6A9AoC80IQsDAv0VP2LdsjEE/D/Quj9R1ePjZKTAfpgfJUFp1Q9tU/5xt+UuAhcWGwLMkBHvirRYZ2yNc1fc/prKDBzcU4BzJ3fhzu3zlOmrx8PG20z1eHDvJqk/13D3zhU01F/SpdtXeH+VcpmXuTtwg5b3+9wZeIwX3Bl4IYD9BYH8C17z/II7BZIEsLcwTzODk7XQSffJs2b6CnBhIaCfCwkB/q3PxGlYaEWS2ln+Wb6V9T1jEtpRq6pf2mkl+G+itfUC6vZuR+aWWEQGU+mCfg6b4sPYx/pXs9ONK7HwR4uFP6wImxgG+10s/NLM06dPCAp3w9EphHz1AoL+221Wy+cEkQL4G5loQa4iMKlg8KlKAvy2VMZzRdUjWiqbO6Ttbffyrn2qlLyvyS95ysrvo7j0DkGRpNsqFZc2ECA1dqhb39b2HdJux/or2p7p87w8S/0Vjay/oV39d1R7ZRX32/r0+LXttO+/XEubL+vVfzefl1c+IJAUUKfvv67+0nIZu6afllFlO/Vf6q+SsXxNG5Jf3jGPvk+VO5qwp66B2uh3+VysxqRT0VItyjw2tn7Iy8tX89uNn9Nbi/Qm4Je/n3379sHU1EHFYqjerbOGJxM0i3XfhnKtKVvOM53F9OlrMGjQRDg4pdAin4sgOnZn595Rlnt/UtPEIh8RtRvrzKIYw8MPMfQhEWC9dl0EBg+ZDHta/fO4KBAajiwUZAclIKhMgfuJk1bCL6CY1vZ65JCmFRK+A8uWu8LI2I8g/gbB/D1IGyZrgrGc/iUmq4PoI1Go2k/LvApXjwxY22xUlCRDIx/uRvhxIViGIC4GnSjZGr2eEaS5MJGAbEJPcvPMJod/n1qAyI7BZi5wbGzDGHm0jJTAR2+dg/eZ4cXzF0o1KcTfi2o8y3Fwizse792IS/lB2OC4jBKbw+HBf9e3+SxXspt6gN/5rKn0vM9ZGrh1a4B/YMzdw0dPUEkDYnAyffBI6zl48jpxjxhQtWMgjYAC/GnxAbA2Gg/zlWPgsGYqLfOzad1fQHC/nLx6Iyr4GFJ/fznVeJaSs09+vsUcWvEJ7gnwBezbr51BLr6k6bTmT1U7BDarJhHkT1DJmmcL7hyYE+ybLh2NdctHw8p4vFoceNDpN9DTDKFC3/GjYg8t+wHuplxsrCClaD7cLOchzN0Q4R5GCPMyQWyYPTbF+iAtKQjZW8JRnBuHvdXbaPUvw5G6Ui4CilC3U3j66agu3YKqok0q7Szbil2VmeTwZ+FgbQkj2h7iouA6HosDLmlBz56J9jepQUIP4rmVYL/1xRNa9AnyqQz0mHla6DPwhFb6p5KvVXj8OqDfSuqO3KtFQJuFX8CLAvysRxYOEqCrVe0M6HYi5L7lSTMt/tdwjn05cXIvLl85wbZaevT7uc5AW9nlRxl0aw+Kd55Cw/13BwwS6n7btnwFEn388imHeZMgRKghreTwN6FuX4NSgqmtu08JyFdJqCSd0552z9rn1V3/NL++vLyv29fYIdXyfi+f6/N0Pv+0/jfn7Vy/3EsdUm/7fuvb6Wr90t/Obejr19fZ/lxLdZ3ObbR//7rr9vlFnefAwbvsPxdhpCEJjcc/qJg0KPqoUM5SfFne19GbgF8W0nV1dVTlcVaOujt2Cb3qmfKtyGIshmzy6uW8nuB4xoy1VHNaBBf3dAWohXojVJ3AkApMm27CXRs7ZV2XXR3ZucnOva12ClavCcHo0fPg7rkNBUWyUNVRbfJolV+7LgyDuHPgRoWrQu4SyO6CWNzFep+eeZ3jelX5Dnj75DMg4AruNpBaZOyvdgxGjpzNhXIRwuhrIrsBn38xQqlBiUKT7DBMZ39FHeoHvpNFQDHrLCh+AHvHFEUbcifoLy5j+HoCftmZsLUPR1FRCZofvTlS9vuaV6lX4h8cO0aHchcnuJkuwp5EJzTv1unsR9ktYQTdwXDlLm12m/RmZ5Df/l4D/O9zpgZu3RrgHxhz95iUniOK0nOMlJ4TOEfD4nON0jMwJq9dL39xj5Se9ORgWuOnkrs/XgF5Z4JsNwJ7L4dlBPrieLuM2vyL4MZFgIvFXJXH3nSaDtyLBZ8Ov3ZUuLFbPRU2bdZ8S/oECD3IynA8dw/Gwkys+qQNiXXfks9kweDOev3c19Cib41Qgv0Qbwn4tQ4+zkbKD8DWhDsExpNhQ66/45oZaochMtAWMaHOWB8myQlxUW4McEXLW1IgQ5v7ITXWG8kxHkig069EC06IdEYynX9TE/yxld+ZuTkcBdmySMjB+VN1uHXtFG7fPEur/zU8fHALTY03KQF6kyDpNncB7uBRUz15+/cI+Jke3yWt5zZTg46m00z/gRbKc1Im9BlpQc/E4s9Vr9B5ZDHw9FkTWpqZ91E9dwi4o0CALxSfJ08oLcoFglCM5PoO+f0368+TVlTPcq8iLrabp3e+vE3d/Yq9Z6nBfwQ1By++VaWnc8Wif56dvQ2OjgFUSNmkKBMKGG1/TB3+Juzc1YSaTkme9VZSde9+xN2Ejqm36lf971S3tNWr9e/q2Hf1Lb04Rq/ra/XORwyc1sj5OsF52wwHx0Ckp2dR7eVW5ynu1fv3D/jFn+K5SlvJobei5XzCxBWK0mNI6pSA46CQckXZmUarvwBrofx4eG1TQF24/IWkRskuwIwZpgT0EzGD+RISj3ExQSWmNg59Bi33y5a7YeiwaUoxR4C+JEWPYh4d//4Fgf9VLGe+CRNoACHwzy1sgLPbVtY7AYvpB+Lsksadh0mqf9EbaumHcAaG3BkQOpIh/TvGs9ykyQxAlXZBLSDEP2A6+xXO3QihkwldqL8A/tNnzvDfAQdGIP8ewfTN2pPoiJoER6ynZV8cdF1WjMNWzyVtDro659z2IL/9tQb4e/XP7qOpTAP8A2Mq7z+kDv926vCHtOnwH7xEDKNZ+AfG7L3q5S/uU6UnMyWMlJwZlOPkFrk5VXlI2fGwWwoPsejbLuY9gT4XAc5852xOyg4dem1NJqtkTQqQJBuTSepeT90Ra76lwVil+qOAPi37crYwmABHyn16sn4/avgH0aKvwD7PgVT/EXlOCfwl8p6m5IUunz4IK2YOZv3T4O+2ipr9VsqpN5w8/zB/K4Z9t4EsAqL4PIJxAcK8zSgZaqrqCuQ5IsCaij+uSKDaT/JGAoTEEGRsjmDsgXjuBOTh+KEKHDtUjtPHq3Hp3D5Kfx5XFKG7DRdJBTqLOzfPkGZziek87vD+9s1TzHOMevkHcPZEDS6drVXvWx4LR/8x0xPSfsRZtx536Fx88/xu3Liwh9cn8JBBxB40XlFOvLJjIFQgUU1pJr+/mQuCJ08fKaWhV9PT9Stxponcsgt2/MNMyt3/VpWe17Vw/z4t3VR0iYxMIDUhEHYOdPqM2EmQcklZWXPybpPqMDDTNul3AaVoCxpIzbjD75A0cL9lGy3XaVSgEcuyrX085yuYkZMTFT3r7t2Gl9MrfiVisZVzbx4fEvCLU7WD82YC5DUqYNr8hbakv1wgNecenMi9/+bbsfj88xEwt4whHUdiMeicZCXwmF9ACRYssKNk6mQuGJYjKnqvotCIr4NY8tOzrikL/LDhM7GBHHqh3OhoNzp5VlHnkXxCwZk2fTXmL7DlDsJZ7N0PRoc+grHjFhPMr2S8gBgMHzadDr+rFC9fOP32jpsUVcjRMZVB5QIxlv4EPlwshIRVYdy4pVi9JlQFlxN6kTgA9wfA/7iFcSbKSmGydDYs549CrO1cZIh8MXd0DacOhd2iMdjitohgXzj7bwb7Avw1wN+bf3UfT10a4B8Yc9n85BkO08KfVX4MudtP4PwVWvifaxz+gTF7r3qpZDmzUyNI45nJ6LiTqSBD6z5BvjuVF9woteYsFn3h5pPXL4sCBzrgKrqOMek6L9NE3TWpO5YrxRl3rKLvWBD0my9/xdm3pD6z0H/cbBYrpZ8gSnCG+tkghNb9IAJ1f6r0CJXHYd1ccvzHY+m0QZg5+s9YOWsoFxrzSPWhbCfzhXEnIITgPtTXguW5K+CrA/lBdCwOoGOxrwu1/Z3pc8AAYWGse2O4K5JIA9pE6tLmhCBs5lkCfOWkR6I0Pw7lhfGoLktVPgFXLh7EvYYLuHnjJM7SP+D4gXIcqS3G/l052F+TjbrqdOwoikdhRjjjFwShKDMMh2vz6RtwgdZ6WvtbW5QT8c0rB3By71YcqUrCib0ZOH+sDFfO7sLl83tw/cpRBfoF4Au3X9SFxIdAIgf3VEXlzr0mbK87T2/6Y9hz+BIePHyzDv+rn0LHKwGGly9fQllZBcFRPINT+dGRkIlBjl4FP9IHQRogZ8tQOl6GwsDIAwsW2VJRx00Fb7KyChuQ32Qh30O1JHMLf9iSqx8SukHxvi9dusjFZMedIgnmJMGs7t69q3Ns7zjd3b57n4C/WlF6nqqAcELNEdlTscTHJRxR4HzEyFnKul5Ah1dv33x8+ukQ/Oa3n1NilWQfwtgAACAASURBVH+fdJIVKk4mKUBSLpPlU8n/t7VLJDBfRm5/pKL6CIiXJFx9cQwWSo/w98X6r1fX0Tlk31XSnhsYGE5UghYvccFm1lezF9xVOYaJdAQeS/C+1jQSI0bOxMzZZiquRH5RIwF/kgL8Xt75DIBWrhyBFyywx0oDL4wZswA+vgXKwVx2MmRHoj8A/sYHTSjMz0WEowG2hxvjVIo1Tm+2QQb19U0ZkTyUO7x5fitVYK32lvyfu9YAf7f/xD7qghrgHxjTq2Q5D1xAZNoebMisU7KcwmTQjoE1AsrCn0WaiyOdba1J63G2YlTdNou+cPlFRUdoOjZ0sLXhDoA1nW9Fq9/CkKCeFnwLw3HqXiz6OtoOOfrLRiuuvukSniXx3pyLABvSfpwt51OLfyV8SeUJlABetOwH8RzAey8HA/L65zPvRKycMQQzvv8UM0Z9irW0Jnlw8SH+BAHiQMykzgT3QZQJDfRkXYwRIGBfdgG8nQxUG35cQIRxQbEh3A1JG3yRGOuHuBhPRQWKCXVAbLgTEmNckRrvg5y0cMX5P3u8BhfO7KVUZxG2l6RyJyAWWZsCsXmDO1I3uGJTlD1ifKla5M7gXZ4G1O83Qnq8GxcLudwROE/rfjMpPHdw7nAhducHo3pbIPYUr8f+HSn0McjFmeOVuMhdgetXjzKi5lVSg+4xP6lCpPM0N91RfgA9+QmpwFs1p7C5YD+q9p3DvQfvptLzpjZv3LhG4M/AVpuSEB0djYiICKbIAZiiVJ8XLVyMzz77DHPnzmMk1zDERK8fgN8i4x/FvscgOTkJ5eWluMIF2s8d4p9x4sQJFcW1N+XU3jfgFxAsVnWh8ggQFy1+0cRfszYUf/nLKFrMKc8ZvQfz5ttg5PdzMHzEdMyea0Gn31107j0HG/sE5dgrSjo794D0nwoFshctdlLqOgL2pY0S8uc9yKMfNnway1vST2C/2gEooYqOBOOysoljNNyIl20JfUh2Avbsg3IwlnZncOfB2jaBfZihqEUFxQ9p5W/kzksC43ksYP052KIUhwLx1Vdj1I7EdMqMJjComAB98VfoT4C/rKgA6cGWuLbNGdgVCOwNweFESxVFN9x8OnJ8VyhVnp8D+e2fa4D/5/4y/30/1wD/wJh/feAtl6hSeMZWYM+Ry4q6PDB6r/VSPwIK8GenRtFJdz7saMl3Jk9fQLkT6TsO3L4VkG5NB1wB+tbk3gsn35Ig34pOvnJWzrgrhZtPoC8OuTyvXfo9gf4o3uvUeCy4MLCW3QPuFrhTv9+HQDzAax3pPBYK7Pu767T9RepTNPpXLxyNeWO/xORhf8SiSV+x7Umk+SyGj9MKOhAbKC6/KAb5u1JByJ0LAKYAlVgvFwO+fO7DFEy1nxiC/fj1voiP8cX6CHdEUN8/2NeKOwTcHeBCI5Q7BrEE/1kpIdhelKIce4tz4rkLEIiEaC/ERjjTkViiChsgwIUOxi50ZGZUXi/rOUyz4M2t7jD35QT97jhxuJI8//u4f+ciDlUloDLdFZWZ3qjOj8Ceyk04eqCINKAdOHawHPu4oDh8cAedhw8qWtDR/aU4c3IPuf49c9K7cK2BDru74RxRhJSCA++s0qP/Qbzt3NzcTEWkxgGaHqh+R0VGYuzYsQgODsKdO/TTePRowH6PzMe7HAOR0iMWfgHkG2nRnzXHXPHmRRnHzCKGAHo+LemzuMOxXlnmx40jDZCBuoTaM2HSCuroO9OaXkZLvCO+HzUHS8m7t7FPpDSnqZLmtGVeFbyLTrs6nj6DmtFiL0o8YuUXvX9px8omnrtBjsoRd+VKLwbjOqXaGTeeCjXGAXSOZtwEKvWI3KcNdxZ86Dgs18LLfwX4ExWHX6Rf5ZkjaUl/+MPX+C13I6SNbZTqlO+UXYX+Bvi3BFjgQro9nlX54kVNIPZtNFMSnGFmGuB/l787Lc+bR0AD/G8en/7y9tHjp9h3/BoNiQeRVnIYpy/Wa067/WVyutCPXzQ9pAIKdeHTkkOwKc4PKQm0ZscHYBP57snkvSeuZ6JVPDHKHYmRbkiIckU8HWUTo9uuyY/fGO6MDWGO2BjKREfa2BAHdR0XTsdZKRPtQWUdX2xJJlc1NRpZabHYlhGnglBtS99IacoYpG+KQArjASREeSLS3576/2vJ81+NCF9b9sOb/QtmHvLv6W+QycjAkrK4UMneEq1LaTynkbubtl6X0mNRmLcZFaXZqCzLRklhOvKyk5GTkYBtW+NYZgPLxzBFoygnCft2F5PLvwN1u0pQXpSBvKxkZFEyMz0lEmkMTCZjsplOwRmbqA7ExUF6oj/jF/hga4IvsvmsOHsDTh/bTcDeqAD/sb1ZOFCVhKN7MnDyUAkunqmlg/BxXL1wCEcJ9PfuLkPt3kqcOlmLU0d3YveOXBzev0NJgHZh/n6S9R519/cdv4JKOu4ePXMDD5t7pvrzkwY+ggeJCYkMwjQJMTExytn6I/ikPvuED2HhF+16T588pcwzbPgMDB46lU66y5RMZyh9S9asCydtZrPi4UvgLBvSdlYQnIeGVyMotJJxF5y4OJhN/v4UpZgjoH7T5nOk0UjgN50kpwL9jF+QknpO7R78OHYxhtCBd9iwmWxruVLwEUu80IRSUs9TJjSSzrdG5OAvVyo7puzDlq2XGC36JFas8FQLEdHqLyxpoiRnOSMtByKCuw7STkzsAcydb8vFhznWbzyggL7eX0AD/G/+KV+5cgX29vb0x1iAIwwEph0DewQ0wD8w5u8hdfjL9p5DQGIVQlN24sCJa5qFf2BMXYde/kK290VhpqXlMVPzq/SY1x3SI96/Pkn02Q6J2tFy/yq/vt7HlMBsS09ayHlnartX7av2WJblRY5OklyrfrT17QnP+jJvPLPup5KoLy9JtdWuPX1ZaVfeyRi0Munz6vrDMek8Bm3tq7GSd+pe+tSsyst4iq6/kvekxKZIdup0+oWjT64+udXiqKvvl1L34f2TJ3T4ZT97Sre4eqsR6SVHlONuYfXJLslydvhlfMQ3CfEJCvALPUkcWbWj+yPwPgG/TpZT5zRbVNJMQH0ZYZE7qYtfpdRvxFouADyvsFFJdyrVHQLqAvL3RXO/kAGuxGlWAmpFb6hTdB7h/8sCQsC+5G+fBIwLrUaCYAnwD4uo5qJhB9s6q6g5Inkq70U/X9rcGH+Ujrc7lJKVtCfvhAIk70TZSl+3BNqSOqWv8kz6JNQkCQ4mgbf0+eT8vgD/kydPIApcQu0SeeC3HcLhF0qPZuF/20hp73syAhrg78nofbiy9+kLKCo9IgbiElWCmkMXNZWeDzf8vdbSL3qtJq2ifjECl2/cQ2LOPvhurEQmpTnrKdOpHR1HoD3gFyCkHd0fgQ8F+PV6+QLUJYlkpp5/L2dJAtglybUAZ72cp/6dgHGxpOvudZb99mBbrsXiL+8rqZjzqi2d2o9IeLbPI20JeNeXkbMkaVvq0NctfdX3R55Jv8QhWJJc6/PJ+X0B/kuXLmHjxo30+4hAaWkprl279kbjggb4u/83oZV89xGQ36G5uTnGjR1HP7Gydy+o5fygIyA6/IdOX0cmY/zkVh3HWVHp0UR6Pugc9EZjGuDvjVHsR3U0kNIj6jylu04zGh5jC3ArTjs6joAG+DuOR0/uPgTg1wNiAdh6EC/X+ufvchagL2UFkL9L/ndpSxYAXanzXdp9X4D/0KFDmD9/Pv71X/+V0qWfMxbBDLi6uirwf/36de4wdlz4aoC/J38VWtl3HYEbN24owD9h/ARUVFS8azEt3wcegUa9Dn8oLfzRJdhFC3+rhvg/8Cz0vDkN8Pd8DPtVDRev30VcVi0815chvfgQLfzvL8pqv/rwLnRGA/xdGKy3ZP2QgP9dAPOHyiMLArHcy85D58WHPBPgrpP11O08vGu/3gb4hY4jAH3nzp2MxFukALtY7H8ulZeXq3dhYWH44osv8Itf/OJl+vu//3uqBX2FpUuXIiAgAIWFhTjDYFviCN744CHKiws1Ss9bfv/a6+6NgIgl1NbWIjMzE7Nnz8a3336LoKAgFWn73Llz3atUK/XeRuDBI9IZ95yFb3wVgsnh3y8c/lbNxP/eBvw9VawB/vc0sH1V7dWb95UnfVBSFXIqjjKKr0bp6TwXGuDvPCLdv+/3gF92AjrtBohVvqrmFQ3oFRjXUYJ0PP2O/P5XeZ4qkC+UHeHk55OLr+fh66k9IsWZknqW3P/Tyl9A6EHSZvs6fu76bYBfLPG7du2CpaUl5s2bh8WLF78xLVmyRL2fMGECfvWrX+Gv/uqvXgL+9uD/H/7hH9QOQFpaGm7euoX7jRrg7/5fhVbybSNQV1eHRYsWYejQofjHf/xH/Nf/+l+VVPKcOXOwefPmtxXX3n/gERCVnrpj16j8Jyo9R3Dq0m0t8NYHnoPeaE4D/L0xiv2ojob7zdh54CLyybPbd/QKHjR1L/BWP/qkXu+KBvh7b0j7A+AXoC2AWizueoqNssCrZy8UQJd3unytShpzS/plBcYFeOuf6/0E5Kyr5/W0IQH066m0Y2uXBJHazNx2U1n5S8pbkJJ2XsluLqMMqCQLqw2QYF2yONAr8fwc2JfnbwP8Mt5Hjx5VXHwXFxd4eHi8MXl5ean3hoaG+PWvf90B7P/1X/81xMr/+9//HuPHj4e7uzsqKyvRwOBsGuDvvb8RraafjsC+ffswYsSIDr9HWYD+5je/YbyXTT8toD3p0xFoJI7IpwiIc2QJPDeU63T4W5/3aZ+0xrs+Ahrg7/qY9esSx8/dggN5dvMskxAQV4lLdOLVjo4joAH+juPRk7u+BvwlVMwRNR5RvRFJzuzcegL6B3S4bUEu1XjSM69ia8ZVKuLcfWlhj4zehSXLXAjWswjWbylwX0xAnsWovBKgSwB8AeU0O1v6ZREhajuhETuUnv+v/8+f8MMPSxFPuU4B8xlZN2Bs4o+RjA8wnUG4Zs+xUJr+U6auUko+lTt+Sv/pDP7fBvhlrmTMHz9+rOJHSAyJN6WWlhb1vrq6mnEMJuDv/u7v8B//43/Ev/zLv1CidBRk0VBVVYVbtOpLXqlbNuo1Dn9P/iq0sm8bAYn47ebmhr/5m7/pAPonT54M8TfRjv41Ak2Pn6Dm8EXEZO5F3LZ9OEqcoUXa7V9z9C690QD/u4zSAMpzjbKcGQyMEZVag4IdJ3Hn/qMB1PsP01UN8PfeOPcl4BeQLcA8Ifk4vH3z4OC4CW7u6ZTLPILklDNwcdsKM/NIauaHw90zE1vSLyEz+ybWmgqffQS5w+aIiNqFLC4SotfXMnBWAt+Fw8o6ltKfO7h4uP2Sny9gv4yymxnZN2jBT6G2vyO++MtIjGRQr40JR5WiTywj746fsBwLF9krWU+J9LuEAcAkIrCnVy4XEE/U4qIzyG9//y6Avzuzd+rUKaVfL467dnZ22Lp1K44dO4ampqbXqvVogL87o6yVedcREPnpnJwc/PGPf3wJ+GUhamNjoxaf71qPlu/DjMCjlmfYf/K6ovOIUs+Zy3c0Ss+HGfpebUUD/L06nH1f2ZWb97Apbz8C4rcjm3+YtzUO/08mRQP8PxmSbj/oS8Av1BvRv3clyB89Zi6t7QtgaUWwHlpFoB/NWAsGMDb2h8maIFrcTVSgLgmeZWjky0i3X2HmzLUIj6phtNxSzJ1nhRmz1jFvMCPr2mE63/n4F6Kw+JGi/OhBudB5hJsfHlnD+leqKL5xBPzyfmP8YUyYuIIg30XRhWQ3wIBtffnlaHgzcJjkeRut530BftHfP3z4MA4ePKgkOcWa/6bjfQH+LS4LsWrWKGyICiVtqPFNXXjtOy3w1muHZUA+lMBps2bNwn/4D/9BgX6hlsXGxqrdqwH5QR9xp+9TpSe36gTsw4rgGlWGXVQCfKZRegbcjGuAf8BN2Zs7fIohr72o0GPgvBURqTtx9db9Nxf4d/hWA/y9N+l9C/hBGk8DzC3X49/++DXmzLVAQtIJpKZdgJtHJtaZRTHYVjn8A0swdtwSzKJFP4YBuOzsk/Dtd+NhZhGNpJRTWGnghcFDpmKdeRS5+fu5U7AZw0fMIO3HFalbLlJT/4UC6wr0V+p09IW+M3OmKUZ+P0vtKAiQz86pV3WOGj0Xk6YYcJGxmtF6p2LBQntsohOvLFDKWF6/eHjd+X0B/q4G9Ot9wG+IoiADpDgvgNGMkQT8YaQNPejyD1ED/F0esn5bQGhk4jfy3/7bf1PO5EI5E/Wprv5W++0HfkQde0wL/5GzN5UGf0H1KVy4dpc6/JpKz0CbYg3wD7QZe0t/rxHgZzHgVszWXSjaqUXafd1waYD/daPSvWd9DfizcxuUY+znnw+HrX2icnqVSLvBodvVQsDKJg7mBPZDh07BxIkGiI7ZS3pNDi3zs+FGmk8Ko+hOJcf+T38eBuNVAaT+ZGHduigMGz4Ts+ZYIi7xmJLX1INz5QxM4J6eeV3x9PWAX56LD4C5VQwXAbNVkh2HQYMnKvpP0qZTyrov+fR1ve78vgB/V2e3twF/cZARsr2XwstwAozmT0F2ZgaePO16lGsN8Hd1Jvsuv/iW1NfXK4v960C8+KFkZGTgT5/+CX/7t38LMzMziC6/HIIl7929jwvnL+H0ybM4e/o8Ll64hJs3b/4sDa3vvvTjb1lUevYcuYLE3H1IyT+AE+frNcA/AKddA/wDcNLe1OXL1+8hmX+UfvGVyCo7qkXafc1gaYD/NYPSzUf9AfBbknP/9Tc/wMllCyPkvkAsqTXLV7grWo7JmhCYrovEkCFTSMExRMz6OqrWZNOxdiapQFuRmHxCPf/006FYusyNi4MYtTNgsjoErnTq3Zp5Tan/6MG5WPIFlGdm1yvH3e/J4RcwL0A+iouJqdNWKXUeUfGJoy/BalKEZLHh4LRZOfyKlV9f1+vOHxvgLwwwQGmIMbYR7HuuGIsVs8YiNMgf586f79YvTgP83Rq2PikkFnyh7YiD7usAvzzbv38/Jk6YiH/+539GZGSkchqXzl65fAWbt2zChqRwbMpcj83ZsQgI9Yanpxeqd1TjaTcWi30yCB9Jo43U4S9hME/vjRUISKyiROdVzWl3AM6tBvgH4KS9qcsnL9TDnZHwljumImxTNYTTrx0dR0AD/B3Hoyd37wfwOyEyejd27MIbwXEF9e2zc+/Qwr+eAaRGK8AvwNvBeTMB/mQsI+hPTD5J4H2UlJ6lGDd+GevdQxpBFoYPn8b8qRDH2lmzLTB40GS4umWQenOetKDjqv1EAnlR/BHZTj04LynXqQIl0Sl4In0EhgydTIffOqUQ5ONbQOWbeVi7NgxFVA4SBSEX163k8H+P1WtDVJ5KLhb0db3urAP8l7hbEc7AWiVoftQ3cTR6y8JfHGiIHJ/lCDSZjBUzf0SAnzdOnj7d7Z+cBvi7PXQfvGAjfTSuXb2mVKJeB/ilQxJEztrGWgXfErUoOa5cvor1GyNh6b4M0bnWyKjzRliGBaYvG4Eh33+N0LBQWvk1MQo1WB/oP4+fPoNgi9LdZ1Cx9ywu37ivWfg/0Nj3ZjMa4O/N0ewHdTU2teDwmeuoOXiB22638IhyWtrRcQQ0wN9xPHpy15uA//nz5zhEp1JTUzuERVS/M+C3stmI78jJF1Ue4cg7E2QPJyVnyVJXhIXvVKD722/HY+iw6fCi86ybWzqt7tNofQ9BEhcE68wilZSmkbEfImRB4LWNOwTUt/fKVs63Ou69joqTnXOHde6AUIW+GzSRVKAhWEMw7xdQjIDAUkyZakIHYBuEst3ImFqsWOmuFh8ubFOoRu9i4U/hosPaJgRlZZVoIe2hL47eAPxFQYbI9loGH+NJWLVgEsJDg1F/+7aS/ezuN2mAv7sj9+HLCcjXp59rXWg/ubm5yM7OxvVr17lAuM6//SCssJiCzbsdcaA5Fvmn/WDuNwfTjQZhqcUEBAYF4AEDw2nHhxuBB8QVxbvOwCu2An4JtPAfv6JZ+D/c8PdaSxrg77Wh7B8VXbzagPVbd8M5vBCb8/fjZkPXHeP6x5e8v15ogL/3xrY3Ab/06uLFC7CwsCXPPltZwt8UoVbeif6+f2Cxkr8MJRCXAFvxiVTioTrP9BmmWEGH3FUmgQTh1pg02VDx+n39CpVVf+5cKwL1IgLz3QTtYdTNtyIdx52ymo505PVGBJV4ikpFpUcXtEss8qlpF6kEtJ55zTHmh0X0BZiHGWxnzdpQxMYdpLZ/JhYtccLiJS6KIiTqPyL1mZJ6rm2n4Oc5/LqdhGf0M6jlGHhj7959lL7rm+A2PQX8RYEGyrLvbTQBBnPGIzjAF+cvXOjxD08D/D0ewn5Vgfy+hfojtJ8TJ04iLDIIVp5LkbrHEQdbYrH9WijswhZijslImAVNglPUEgSHBKLxvvb/tQ85kQ9I6Snfe07ReUJTduLAyWuaLOeHnIBeaksD/L00kP2lmuv1jdhWcQSx6XtQUnMKDY3a1mfnudEAf+cR6f59bwN+0YXfuDGeXPogbIw7rhxmhTcvVJ2fplYUl7WoQFmJm06qoFuyCCgqaVY6/AFB5ZAkyjsinRkaXk0O/z6lxx9FGk4g3yWQwy+LhtStl6i9X80FQAmVfSq5aDjGwF33ftKmyIDGJRyhFXInwckuXeL1BoL9PL7bRu3+9RsPwI/WfqlLZD/T0q+ghP3USXK+7jt0z4TOszXzBlWCkuhnEIhLly51f2J6WLIngF8s+8Ld9zEaj6XTRsHfxwvnu8nZ7/wZGuDvPCIfx/3t2w38mwrGavt5BPvOOP4iCduvhsHUcxZmGg2DRfA0BOUug0fcCgQHE/Df0wD/h5z55jaVnhxKc+bv1Kn0/BxN60P2S2urayOgAf6ujVe/z33n3iNU7z+P3MpjqD1yGRISWzs6joAG+DuOR0/uehvwy/9ELtAS7OsbRkt6FIE3/wdDKoxYvwUw/zQ9VzQZ4caLdV8Av9Bm9NQZibhbzueVO6Dq0AW/knqeq2BZci95pXwZ6xdgLpZ8yV/BJPk6ttnWHt9JGZUkb1ubqhyvpd3isseq7Pa29x3refUt8m1FpY+56LhMR+F0An5/bN9ehSdP3qyV35N5e1vZ7gD+QvL1BexvE87+mikwmjsBIaRfXLhw8W3NvfN7DfC/81ANiIwvXgCXLl5CUKgv1jnPx6adDjj8LB4l5wNgH7oQ0w2GwixgKiKKjRFZZgS3jcs1wN8HMytU4fzqk3COKoHnhnIq9lzWKD19MA89bVID/D0dwX5W/iL1cTcy/LUHHXfTig7ilkbp+ckMaYD/J0PS7Qe9Dfj1HamtraUFPpqOtevJqc+Ct18JA2GV/iT58pmvfxl8A5h4ljzqGe/9Ask3ZfINKFfv/QLLed92zWfqvbxrK697L2V0efT1tW/3ZV4p3y6pNqQeVVdb3dK+1M++tK+j87W3bzE8fXLpg5AIH99o5OfnMzjWVdy5c4egv298cLoD+IuDddKbngbjYTB7HGk8frhMp83ePDTA35uj2bd1yeL+9KnTisaz2nEWknZa4xRSUHw2AGbeszDLaDgsQqYjvMgY6yvXILzEQAP8fTRlDx/RELLnDHziKhGcvAP7jotKD1dr2jGgRkAD/ANqut7eWaH0SITd2PTdOkrPfY3S03nUNMDfeUS6f/++AL8A3dOnzyBtaxZ8fCLg5BzGFP5RJkfnUHh4hiM+LgUHDhyERMW9ST3ys2fPKs3x7s9O90vqAX9aoCUuptvjWZUvXtQEYt9GMzgt/RFhZtOR47tCWfQLydcvCTYmZ38ZfI0nYtmMHxDg66363/0evL6kBvhfPy4D8enZ02cRGOILC3L2sw564CSSUXwuEFb+czFj5VCYB05DSC5jZ5SZIKrURAP8fTjJTc2iw38ZCTnU4S88qBR7nmuAvw9npHtNa4C/e+PWb0vde/AY+6iRW7nnLI6cuo4mrsy1o+MIaIC/43j05O59AX5dn16g6eFDpckt4P8UrYEnP8Ik33Xx4kXcu3ePqibPlbKJBCV6wEi0ra2tPZmebpfVA/5Uf3OcfwvgFxqPSG8GrZ6CVfMmIMi/o4Nu6/NWPGOS83N+X08ODfD3ZPT6R1lx1D1/7jzCooKwxmk+UnY700E3DqWX6LAbQLC/YghsImYiNN9QAf3IklU8r9IAfx9O393Gx9hSdBjG7pkw881BZe1ZPGvt2d9yH37Ov9umNcD/kU396Uv18I4th6FrOqJSa3D15v2P7At7/jka4O/5GOpreL+AX9+KRN58JfH3MV6/+tJXV/KdfXU0PmxCUUEeolyMcTTZEk+3+wJi4Y/raOEvIY0ny2spHXQnKrAfRs7+1TYaz6PmRzh95gwOMLjSQe5ciOTqoYOHcJBn0V/vjgKRBvj76hfRO+3Kb/rkyVMICPGGqetcpO93U2C/4JQfLP3nYO7qkbAKnYFwcvajaNkXsK8B/t4Z+57U8qj5CfYcvoz47Fok5+3HiXO3NJWengxoH5XVAH8fDfz7avY2nXYra88hi7SePYcvovGh5rTbeaw1wN95RLp//6EAf/d7qJXszgg8ZGCj3JxtcDCZj+rI1Xhc4Q3sCnoJ+MNJ6cnzW4ECqvF4rhyHJdPGkMZDNZ5z51RzrbTiFhYWwt3GCR5rbOFv6QJfC2dYrlyD2dNnIiExUQVk6mrfNMDf1RHrX/nPcAEYGOoHG99lyDzsjiPPElBxmeo8btMxbflgpcYjlv3ocqHx6MC+Bvj7fg4fEvBX1p1DaHI1orbswiHKcsrfuHYMrBHQAP/Amq+39vYKLfopBQcRlFSFbZVHcftu30TqfGtH+zCDBvh7b/A1wN97Y9mfanpAC39hfg6C7QxwMMEcT7b7vLTwOy/7ERHmM6jGswyhplNhMGssfL29cK5NevMZaUiV27fDaZ01ApZYo9gmBqV26xG93AmjP/0W//QPv4Sbi6vyVejqN2uAv6sj1n/ynzt7SiuSngAAIABJREFUjs7rnlhLNZ60OhccRxLKLgbD0m8uZhkOoxrPNB2Np51lX7Pw94/5u0uqcFrJIazxzIJlQB6qCP41Sk//mJuu9EID/F0ZrQGQV6/S4x5dSs6dqPRoEQk7T5sG+DuPSPfvNcDf/bHrzyV1HP58pAaY40Ibh18oPfvjzOGyfCz8TSbD03A81jCCrtB4LlBaUY6HTQ+VypDzWmtEGjjhpO82PNiwG3vdUuA2zRizvhiFYb/9AkFefirYUlfHQAP8XR2xvs//nFxvofH4h3jB3HMRUvc64vDTeOSf8FWc/dnGw2EdNgMRRato2V/9ksajB/uahb/v51CcdncduoQ4ofTk7sPx87c0lZ6+n5Yu90AD/F0esv5doJ4Av3T3GYL9Q9ix/wLuP2ju3x3ug95pgL/3Bl0D/L03lv2pJr3TbnqQJa5tcyKdJwCoDcGxZCtYzx+JpeO/VjQeiaB78dJl1fVmOhrn5+XDzmgdwpba4bR/Lp4k7kOdxxa4TjeG4dBp8BhvAMNh0xEdEIaGuw1d/mQN8Hd5yPq8wElG0A0M8YGtP9V4jrjhDKU38477YpXLNEpvDoNlKKU3C40QU/F6sK8B/j6fQjQ1t6D6wAVEpu1GLGW/j5y5qQH+vp+WLvdAA/xdHrL+XeDS9btI2LYPXrEVSC85DFkAaEfHEdAAf8fx6MmdBvh7Mnr9t6wAfnHajXE2ws5IY5zfYo3LGfbI812K2aM+x9jh38LH2/NlBN3HLQwyVl4Bx9VWWG/oirPc9n8UV4t9nlvgQrC/bPBkhE43Q/ICJwL/6Qj1CcDthjtdHgAN8Hd5yPqswLNnrThBsB8Q7ANz98XIPOiOo60JKDoXAHOfOZi0eBAddPlbEDWe19B4NAt/n03dTxqWAJ65VcdhH1YIV8b4EWt/q6bS85Nx6u8PNMDf32eoi/07d/UOwlN2wjooj8C/FjduayHIOw+hBvg7j0j37zXA3/2x688lxVpfvWM7fB3MEGi1hMDfALFuRnAxmYP5MybCy9MT59ocdJtbmlFaWgo3C3tF4zkbmIcHsbtx0GsrnEnjWfzdBPhNWY2s5d7YvMgFS7+dgAB3b9y6Xd/lIdAAf5eHrM8KCI3H0dke0xaOhkf8SmQe8sCWPS4w85mNOSYjYBs5C2EFr6Q32wP8zteaLGefTaNquOVJK05fuq3kOKv2nVPqf5oMf9/OSXda1wB/d0atH5dpYKAt2XrL4Wp877HLeMCVuXZ0HAEN8Hccj57caYC/J6PXf8uKZObt27dRt3cPKsuKVdpeXoLS4gJs316Jq9d0EXRbGCAtNzcH9qTxRJOzfyYgF42xu3DAMw1OBPtGpPGETFurwH6+QQA2LXTG8kGTEUwO/+2G210eAA3wd3nI+qTAs6fPkJKyCT9MHoJ5a0bAKng2HKIXYJnNWMwwGEoH3amIoPSmBNXqDO5fd68B/j6ZxpeNNja1IH/HCThHMCr4hjIqAF7SnHZfjs7AudAA/8CZq3fq6ZnLd5RCzzo6y20k104i72pHxxHQAH/H8ejJXW8DftHpljrba7S3tj5DCykjjx83o7n540uP+U1Pnz7tyTS8t7JPORfNj1vwqPkxU4u6FhUeOWSOhMbjtMYK4UvtccYvFw827sEBWvYdpxpi5ZCpCJy6BjkrfJBvGIA8Az8F+IXDH+4XhDsah/+9zVtfV/zo0WMkJMfC2GkKkuvMkXnKDmlHbeC4YQ5WOo2Fa8ICxNBBt7305uuAvv6ZBvj7dkb1Trsbs/YiSZx2qcPfqpn4+3ZSutG6Bvi7MWj9uYjIcorDbhhpPXlVJ3CHuvza0XEENMDfcTx6ctfbgF+A7wEGaZLATI8eNfF6P7Zs2YqYmARERzFFf2wpUX1TYmIKLefbUV9/qyfT8cHKyuKroqIcLma2iFzpiDN00G2OJ2ffIw2OUwyw9LtJ8J28CtkrvJFv4I/clb7IbQf4I/yCNcD/wWbrwzf08OEjpGYkwW3jCuxu8cdhhGA/ghFdYQKzoGlwT16kouhqgP/Dz013Wnzc8gzHzt5SVv6inSdx4dpdRs3uu8CA3fkGrQygAf6P7FcgnP3c7Scon1WHsj1ncbdRU+npPMUa4O88It2/723A30rr8fHjx1FbW4uszFz4+MTAxTUOLi6beN4MV7fUNyY3ty1wc6cqzFvy9Z/3m/ldKUyJcPeIQXzCZhw7dowOcc+6PynvueSDh5TezMuDi6lIbzpSjScHLYl12Ou+mZx9IxjSsu8/eTUyydnPN/RHjoB9DfC/51npX9U3MXCbAH6XDctR88gPBwn4654HIrJ8FdaRzuOWtFAD/P1ryt7YmwePnqCceCIgsQqhm6qx/4QWeOuNA9ZPX2qAv59OTHe7dZaUnuDkHTDzy9EoPT8ziBrg/5mB6cbj3gb80oX79+9DLN6m67zh7VuA1K1XkV/4AAXFTa9PRXwuqfgRikoeo6j0sbpWz9Tznyn3c/V94Of5RQ+RlUNn+8jdsLQOQ0BgBE4zIml7WlM3pua9FGl50kIH3TLYGZshcrkj1Xhy8TihDrUE+47TDJUaT9gMMwXyxbIvYF8D/O9lKvp1pS8B//rl2NnkiwO07te2EvCXaYC/X0/cz3Su6fFT1B67gpT8A0grOoxTF+v75b9PP9N97XHbCGiA/yP7KVy9dR9bGREvfPNOFNDJpuG+Fmm38xRrgL/ziHT/vrcBv1j4q6qqYW3tC3evXAXwyypbUVb57K2pqLQZW9IvITXtgloAlG1vRfl28syZKqpeqKS7b+1w/6p+Xf6KKrzM++rds9fUo6tb37dX7UDlbV9Wn+fnzuXsq+SPiqmDmUUAEhI3dSswVfdn8u0lxX+iansVPK0cEbXSCadoVGjauFsH9snZXzJoEjn74qDrQ77+K7CvAf63j+3HlkMD/B/XjLY8eYYzl+6gqvYcqhnf5xpxxguNwz/gJlkD/ANuyt7c4Zt3uN1OoJ+Qsw8VezVKz+tGSwP8rxuV7j3rbcB/m1KNbm7ecHBKRCat3gLUBSSXVjz92aQH0RnZN2BhFYPVa0KwafN5VU4HpJ+hpPwJistb2hYNrbr7MlqrWa++vABuuS8ue8ykz6tfaOgWHSWsQ1Jp2wLkVflX76W8PNe3LddvS9IH+db8ogfwCyyGrZ03qT3HIU7M/eEQAFdUVATndTY6Go+o8WzYhf3k7DtPNcLywVMQOGUNMpd5Ka6+3rKvP2uUnv4wix+uDxrg/3Bj/SFauv/gMbaVH4MNJXedqMW/k0qArc+ef4imtTZ6cQQ0wN+Lg9kfqupA6aFHvabS89NZ0QD/T8eku096E/C/ePEcR48cxdq1tggKrUBVDd4K9tsDbrHsz5lrgcmTDbFh4yHkFd5HWsZlZfFfH7sfIeE7kLTpFGRhIO9DeR+fdJQgu1FZ5CV/QvJxBIdW8l01UracQ2HJI/VOQPyW9MsIi9iJiKhd2JR6FpvTzrP+qy8XD6lpF0nL2YWgkAokJp9QuxOlFW9erHRcCDzD9mogKeU0rKwDsL1yB56QQtPXhzhSFxUVM6iWJWIMnCm9mYf7Ir3plQaHyYycO2QGwhhUK5ucfXHM1YP89mcN8Pf1LH7Y9jXA/2HH+3231vK0FReu3sVuBtzae+Syiu9DkS7tGGAjoAH+ATZhb+tufUMTSvecQRqj7O48eBH3H2pOu53HTAP8nUek+/e9CfjFUbWutg6mpo6IjN6N6t14q2W8M+CfOdMUEyasQMyG/YiNO4RVqwKwbJkrFi6yx8RJKzFnjgVWrw3FkqUumDLVGHPnUVIyajdy8u/C2y8f8xfaYvqMtZg+fQ2WLXdHzMYDCvDHJRyDsUkApkwxwuzZZm11OsDRaQuyc+uxfuN+LF/pganTVqs8ixc70VJfgtyCey93EDqC+9db/at2QtGSbO3CCLJL0Eylor48ZH5rdtbA3cIBYcsovUk1nocb96LOPVWnxkMajz+DauWu9FNqPO1BfvtrDfD35Sx++LY1wP/hx/x9tiiynHsOX0bitjry+PfjxHnh8PeP3cf3+d0fW90a4P/IZlRkOTcXHEQgHXdzKo9RlvPhR/aFPf8cDfD3fAz1NfQ64K8jh93MWVnRd+zqOuAXMD55spEC/P6kxgwdMgU/jFkAc8v1WLTYAZ9/PgyjR8+ncyw1wlf5Y+jQKeTMUx0n8RgsrTZg8VJn2DkkYw0XBSNGzlLl0jOvqHffj5oDk9XBsHfYhGnTTPDJp9+xTida/atZzgkTJq7EOvNIKu5sxdy5llwYmLMf+xQ9SOg97wL4xcKfuvUibO3DUUirel8C/sYHD1BcUgwnSm+GLLfDIc+tuLdhJ9V4UuE0xRArqMYj0ptZy71eqvG0B/ntrzXAr/+L+fdx1gD/xzXPD5ueMMruOYQkVyEytQYHT16nkphm4h9os6wB/oE2Y2/p70Xq425gwC3nqFKkFh7ErTsP3lLi399rDfD33pz3Z8DvQ4v9oEETFChPS7+CwOByDBkyGRMnGmBLxhVF3xk3filWGnrz+gQCgsoJtKkd7pEFU7NwfPb5CO4M0LJNGs+8+TYYN365ovWUVz4l5SYWf/nye8zkAkMWD3/5yyjuIBjAyydPgfz5C+zY9kQ4uaQpXr7w87sK+Iv60MIvztN1dfuwculyTPxqBFyorb/J2BORS+xhPn4xwf4UBE8zpc6+BNXq6KDbHujrrzXA33t/cwOhJg3wD4RZevc+Pn7SihMX6lGy6xQqyCC4fONev/Evevev0HJqgP8j+w3cZqCtitrzyKCDzW5uwd1/2Pcc4P42xBrg770Z6b+Afx98/Qsw8vuZWLsuXFF2QsJ24IcfFmIuwbtQeISjP5kcdAMDb3L6D8DJNY0A35E0HxtMpiPqb3/3hQL6Pn6FmDXLHLNmW2BbXgNq9oDSmaUYNXqeovAYG/vj97/7GiNHzsYS0ocMeT9t+lr1zpMLAHHEFQWfgQT4HzU/Ql5OLgwmzcO6EbPhOdFQWfNXDZuBKX8eDrcJBsqyX/D/s/eeUVlla7pod5/7q8+5P7p/9K/+1feOM8643aPD7t7n7Nq7au/uqh1q166wK2tVqWXOOWdBxUAQTEhSDIgCKiJKDmbFhIqKYs4RvhyB9z7Pu76FSFkKCALW+sb4+BZrzTnfOOd65pzvnBMn6Jqg/kW/FuDvuDrXE0qyAH9PsFLreXR5/XII8fvrtlcgrOeEnLty3wrpab36uk1KC/B3G1N0DCMM6UnfU4kDMg5ITtk5eVRrnbTbUrMW4G+pkfb/3/0A/yT57LOROsoeFZPfBPhzduN8ipV7Afj7AfDPUOCevrVGAf9QxOpPm5YiH/zpe+n1zSSM0u/BLkFb5Fdvf4SY/tkSvawQ4TlTNbaf5RDwR0btkd/9rrfeGzcuTv71X38jn2DtwMxZG3CA1k4NC1oQkS0bsUi4EOcCvGynIbMz0DykpytH+O0upxTtyZd1ExbJsSnr5Pb8HLm3cLcUjl4u0//4vcRhdJ+hPC233/wx0G8B/vbXsZ6Y0wL8PdFqP86z3eWTgoMXJSKxWKLXlcuxqltWSM+Pq6vbPrEAf7c1TfsYuwnAv3n3KYlZv1+yMcr/qLZrF/21T4rOzWUB/o7Tb3cD/N8AsD8F/HmIw+8t4yeuAsAn4C+XDz7oJ99hoS5H6tOx085nGOHnAttBgxfIf737JWLwEyUj66Ysxcj8L37xB+mFRcDcdWfgwHny4UcDZMOm85KL2QFu/fkfP38PMwAzELazVd597yvpDdqJycexiPcRwoNKNJwnbUtNaIvPtsfwdy3gd0hJXoFsmRojNXOyxBNVLv7Y/XJ0ylqZ++FgDefZYQH+jqtIb1hJnQX4I1KHyvLlsWKvs0JVX6fL+BDSc+n6I5y2W6N78XNg0Vqz+zot0DG0LMDfMXrsNqU8tnnk0Kkb2Iu/Wo6fuy12K6TnB7axAP8PVNLuG90J8Gdk3dDdeLhYdu06zHIBdDNGfyoW4+7a8wSx+Ifkc+yyMxDgniE9WzKu6cLa0djJhzH5H388SEfyZyOOnwt8fw5Az1H/+QszsXg3Sd7/oK8MH74YswFr5ZNPhsg//uP/0fUBa1MrdTEvOxOjRy+TOfM2S9/vZ8qQYREY4a/WbTvNPf7Nkfwf++0+I/wG4E+fEi2XZmeKO7JMfMv2y5HJa2XOh4N0hN8C/O2uNm98xs4A/PFFI2Xh2iESF7fMAvyv2YNq7V7ZmndKRmFDgolLs/WMn6C1aPc1W+HVyVmA/9V12K1KuIq9chMyj8icVQUY6T8p961Fuz+wjwX4f6CSdt/oDoDfAM9B2YVRe26FuTQqT7btuCNpWy7J/IjtOL32GPbTd+l+/BFLcnRxbl6hS0f9GZqzes0hAPMLEo4FtsOGL9FtO2fO3iBTsHvPKMTjhy/MkuiYQhmOWH8+5/1+WAfAGYABiPnfueshDvq6rDv5jBgRKSNGRuouP+xwcCbB5O/HQH7z+90R8F+cnSGuyFLxLtsnhyenWIC/3bXlp5OxowF/ctk4ido+QMYv7CWxsTFit1m7z71Ob3K6/bL/xBXdoSd5+1E5U3PPiuF/nQboIFoW4O8gRXaXYu48dMj2kipJyDos+Qer5bHNiuFvaZv169cjdvtzSUpKwsFG/paPrf/boIHuA/gDGEn34bAtu+Tiy4OyCordug9+XqEToNsvBSUePYyLB23x5F0jvQ0HZCFeHf/vzH2M7TnPy7r1Z2VHzn0N3UnfegVhPBckCgt3+/SdIQMHhcvkKQlYlDtG3nvvS8wMpGq53IUnG8A/FXm5FSdnG4r0tF6G8rT+8K3OAPw8rbeurk7Onz8vt2/fRuxt/UstbHc9HeG3AP9L1WUlaKGBjgL8CcVjJLF0rERtGywDpv9ehoz5VsrLytBuB1pQtP7tTA24PFi0i3N9UgD2N2Av/qrL1qLdztR3Z5VtAf7O0mwXlfsAB28VHa6RrQWnZd/Jq1KHI7Gtz7MasAD/s/p4lf8I+NPS0gB+35MFCxaI09n+kTc9eKsd+/A3HyHnfvfcEYeLZPk1ro34efP/p3viP31uPivbJ0LwXravQX9x2K2UALCvSTqq4UKf47Cub7+bjAW6Q2TUqGicuHslRMugy7zMw/ytDeNpzn9nAH7a99KlSzJnzhyEHI2WVatWSXl5uVy9elUc2Gu/AScct/xYgL+lRqz/26KBjgD8CcVjJQkj+zHbh8iwOR9J/xFfy+bNaeL1WIdJtsUWHZG2Fjgiq/CMjF+SLVNx+F45dgK0Qno6QrOvtwwL8L9efXc6tcs3sThx0wGZFJ2r+/FzxN/6PKsBC/A/q49X+Y+Af/PmzQr4w8PDuwDwE9jX/+DbllH15oDbyGeOyBsdAgL4gmKPLtidh7CfqYj3X7Rkl2Ruu6ng/llazGvmf/5pus/SezZNawB/IBBQoM5Re5vN9tIv05WUlMgf/vAH+au/+iv5m7/5G5xH8F8ybtw4SUhIkKLCIrlw4YLU1tYK7cmPBfhfpVZYeV8V8BPsJ5YC7GcPlRFhn8igUb0lIzNDfD5rm+mu8C4nTtrdf/KaxGccluRtFXLm0j2pb/jhQEFX8GbRbL0GLMDfel31iJT3Hjkkd+8FSd15XEqOXpZauzUa0tJwFuBvqZH2/8/wkO3btyvgnz17tgLR9pbWnhF+YySfgN8Y0TdG6jnabtx7Ebhu3TN/Uww+w34KitwK/hkOZNJsXTnPAvsfy/MywE9AXl1dLfThmJgYLGCMe+F3xYoV+nzSpEnyT//0T/KXf/mX8hd/8Rf6/e///b/LP/zDP8iv3/k1diEaKEx75MgRcblc3Rbw37x5U2cq+vfvL1VVVe11NStfJ2vgVQC/gv2SsRKzY4gMn/+RDBrzjWxK2yiPHj3uZK6t4n9MA2ZIT/K2o5KaHQrpQaig9elZGrAAf8+y10u55bacadiHP3rDPmNbzjprW86WSrMAf0uNtP9/xoeXIab297//vRBUcpS4vZ+2Af6gxt2bu+2kpV8S85ux7QZi92v1+VNgbXYInt8RaN5xMEf58wodsh7bcG5IO6+7+mRgRH/T5kt6kBZH/Y3QILNcjOqj0/GUXqCpQ8CyWzvq/zLAzxHOwsJC+fLLL+U//uM/5K233pJf/vKXP/r91a9+pc/+5V/+Rf72b/+2CeyboJ+/f/3Xfy1///d/L7/59W8kLCxMgXSdwyal+YWSPiVGulMM/40bN2TWrFlYLD1A1yS019esfJ2rgXYDfsTsJwDsRyNmf8S8T6T3gE8ke2e2uN3WWrTOtdiLS6/DwGFGQaWMXrhdJkbmWLv0vFhd3fapBfi7rWnaxxhDeuI27ZdJUbtk7Y4KufvQ3r6C3uBcFuDvWONWVlZiS8uPZdiwYXLr1q12F94WwF+6t1HyCpwSg0Oxvuo1Hvvpj9JDs3r1nqwn5U5H2A0PvSrQQ68A8jHib8TmNzbF+Jvg3JgVMGL2mYZlE8xvybyuB28NGDAXC3lPC0/cnYZy09JrpFxj/Rs03dNyny7QZZjR0/tGutbE9L8M8HNGhaA3NzdX105s2bJFXvTNyMjQ5xEREdpBMEf4/9v/9d/k7/7u7+RnP/sZzi34TKZMmSLp6elSdbbKCBFy2nUffm7L2Z0A/5UrV2T69OnYHnW4XL58ud2+ZmXsXA20F/BrzD5G9keFfypDx30nW7ami8Pe/nVBnSvlT6d0jzcgJy/ckYzC09gU5KzU3Hhs7dLTA81vAf4eaLQXsXzvkVNy9zGk54QV0vMjirIA/48opp23uSC0d+/e+j116lQ7SxHsHhOU461ctMsRdu6tP2Vaovz8f/9OPsP++qPHxsqYsXHaAfjjH/vq4VhZO+4KQXRuXp0krz0hK1btlU3pF6UIu/aYgJzAPyPzhqxcvV8X53LWYB/ypGGrzd+//528/34fSU45qdt3Llq8U/fv5xag3NUnK/uuJCQfQ74K3Z6ToJ708oqc2PEHu2UlHtUde3jwF3ngzkFmR+N5vy8D/O1RLmdhGKrz/vvv6yj/v/3bv0mfPn1kwfwFsm3bNjl27JjcuXMHL/CnMbndNYafYTwTJkyQiRMnKs/t0YeVp/M10B7Az603l2UPkaFz/iR9hnwum9PTsLuUrfOZtSi8VAMON2YWD1+SxSmlONRzL874wUm7zdqLlxZgJegWGrAAf7cwQ8cxcQ378Cchzm5efJGk46CM+4+t0ZGW2rUAf0uNvNr/9+7dUxD24Ycfyq5du9pdWFsBfw5OtJ08dY389rdfy4KIHbot5u78OgD3Awh5GSMffjhQktaelPSMKzhBdw220hwrX341Sk/HXYT9+HlibhG274yKKcC++rP0UK6veo2TEaMidWvNVGzR+cmfh8mf/zxUEgHoFy3OkYmT4gHwKyR8wTYZhn35R46Olm++nYxtXkfi8K04dCYu6XkA4QsyQWuM9Oo1Qb7vPweHcC1UHrOy7+gMw/PAPu91BuCnQTgrkJKconH/2dnZcubMGXnw4MGPbtHZXQE/w8e+//57DT169OhRu33Nyti5GmgL4NetN0Mx+wNnvC8DRvaS7OwdVsx+55qoTaU7XAbgXwLAHw3Af8wC/G3SX3dJbAH+7mKJDuLj4rWHwko5HIAkfushuYOYfuvzrAYswP+sPl71P6/Xq1s9vo9dYFauXCkcUW7Ppz2AfypG+N/HKPzSyD2Sj/30S/cGJHVDFcJ6puOU3P4SE1ssEyaulN/9vjeA+VSM+i8HgB+iJ+guX1EmKetOyhfYavPd977S2YGBg8Lkl2//CSflzpIlkXny6afD5fPPhsvyFaXYgz9M3vvtVwD7GfJtn2nyz//8SwD9UTIKJ/V+9dVYeeuXf5RZszfKgoXb5A+YFfjd77+R4TiE688o4//7x3+Xr3tP0DCjMsxOvE7AT3swFKixofV26SzAv2XAfBnx7leyKjJWHj1p+yLMtE1p8uGfPpT4+Hix261wxfbUs9eRp7WAPxFAPym0z/5I7MbTb+iXGlr2Oni0aLReA15fUPfe343ogTyc73Ptbi228219e9J6SlbKztSABfg7U7tdUDZ36dmFXXrWYZeeYmuXnudawAL8z1XLK90swkLSTz7+RKZOnSqedu6T3VbAz7CayRi5/z+/+D0A+FSZMTNVZs3dKP0HzpX3/9hHhgyNAGjPlQ8/Gijf9ZmO2PvLuo3m0qjd8tvf9QLAj0VnYJWO4M+YuV72HxSc0Htb+gHs//a3vWX8xNXy8SeDAepHAPCXyaDB4ZpvfsQ2lDdNfv3rP8vCRTt1Qe4ahO68+94XoLkQMwnj5Df/+bnMxwK3giIPTvrFs3c/R8jTRCz6vRjayvP5u/Z01gh/W43b0YB/94gY2T08RlL7zJKBv/lU4mNXis3RdsAeuTRS3nv3PcndlWsdmtdWo77G9K0B/Ikl4xTsc5/9ITP/pIdqZWRutcJ4XqOdWkvKgZN2iSeiUsslFmsET5zHAX5tGEBoLR0rXedqwAL8navf1156zY1HGmM3bslO7Jd7xFq0+xwLWID/OUp5xVvXrl6TIYOHaGw4F1a259NWwM+QnEmT4+V//a+fyb/9268Bqr/U029//4feMmz4YgDtCiy0zZP//C8C8UWSgw7CoaMiqxMOyR8/6IfQnlnaEWD4D8N6jhxjrL9Nxk9YJf8JwM4y2Fn4/IuRAPzlMnjIfJ0pUMCPGYQ/YQZh9Zoj2lHg6by/+30vzAzM0M7Ge5gxWL5yn+xDJyIj65qezNuv/2zZmFYtr3uEvz226EjAnzs8Wgj4tw5YIGEfI0b7iz6Sk5ODg3tefuJvc97v37svI4ePlI8++khOnDjR7pmk5mVa152jgZcB/kVp/QH2jZj9EWEfS5/Bn8vGTRvFZsXsd45BXrFUJ2L4yyquSOzGA7Iy/ZCcwgJeK4b/FZXaBdktwN8FSu9MktfvGDFYWGjOAAAgAElEQVT8YRrDX4kYfuvgrZb6tgB/S428+v/BQFDDebjjC0/e9fv9bS60rYA/B4tmp0xNwCFSnyGWP0mS151COM9ZLJY9Jzt23pc8hPjELi/FaDsBfwQW2T6SA4dFVsUfAODvi5H82Qj9maHx/tyBRwH/njoZN2ElOgmfYSeYxQCXzwP8WRoy9NHHg9GpOKYj9inrzhiAv58B+H+L0J/lK/caMfmZTwH/BgB+Ljh+nSE9bTYEMnQk4N89IlrS+8+XuZ8MkUFffCebEZbTnu1bed7DBx98oOtFHj582B6xrDyvSQMvAvyTY78UAv5ojOyPmv9n6Tv0c8kvyMMZHtZ6s9dknjaT8flx/gfChYsO10hpxWW5ea/O6nC3WYtdn8EC/F1vgw7lgCE9OWXnJWX7Mayqr5Endmv/4pYKtgB/S410zP+HDh2Svn37ytChQ6U9o/yMM+eOMZMmhUt8wmHZD3D+Y8CYoJkAfurURIy099PR9L0YTecI/v5Dxrd0b4OsXX9G/vTRIOkVCqcpw72Fi3YgNOdrAMeVCAlKlE8ARKdMS5LyA426o863WITL+PtJk9fg2dOQHmOEv5fMjzAB/yDswlMRAvyVCvgHDQnXNQG/+c9PQyE9blm5ah9mDD6T3t9M0hj+FwJ+yLA1C3vNz1ktBQXF4nF3zTkaHQX494yMAdgPl3kfDZGx3w3WzuDjdsTu87Rghou99957smHDhh9dbNwxnmyV8qoaeBHgn7bia5m+6msZPvcT3XozbfNG+Ll1QOSr6rwz89uxaDf/4EVZmFQsSxHWU1Fl7dLTmfrurLItwN9Zmu2icrkP/8rNB2Tqst2yPue43EUHwPo8qwEL8D+rj476j4soFyxYoAc9paamtjmWn4tLT5/G4S6jp2IU/qAcrngJ4N/1EOE3ywGmP5XFWLRbWOLDot3GUCchqLvhMIxn6owUAPze2G9+pAwYGIZR4n4aprN6zSFJxdaZvb+ZjFmAz7CbzjztGLzzzkdYbLtUorHHP7fk/ACzActiixACNF1+9fZHMidsM3bgGSu/wzoAlkEAn5h8HM/+JAT8M2dvwAzC95gl+FL6o8xPPx0mP0PIEdcR8OCuF4X07ENnJR1bhE6dFiulpXvF5/d2lHnaVE5HAP68kbGSOWChzPhDPxny2XeSkY74bHvbNxHgdqEVFRXyBywK/+7b79rVmWyT8FbiV9aAAfixgH3tMDnqj5GzslIqZYWk7McuWAv+JH8e8JaG8Wzbnil2W9vXcrwyg1YBbdKA2xeQyuq7sr34jOwsOyc1wBkNVgx/m3TYHRJbgL87WKEDeSDA34GDMVZtOSg55eflYW3XjBB2oEgdXpQF+DtcpU0FlpaWYsT9T/LFF1/I/v37n9nbvSnRCy7u3r0rs2eHS/j8LbIn36Z75Rsn1T67yJWj93vy7bqvPg/DSkUYT1GpH6fb8vArI62ZjyfvMuTnC8Tqc9cdbpMZs6xIdiM/8yxD2M/3/ediN50R8sVX47C95nItL2vHHYy0b5LZc9IwMl8ti5fu0nISU07oTjwzZ63XLT/Jy1aA9CmYbYiOLcTpvOd0MfAf/9hPvu41Ub79boqGHQ0cNF8P7foxwM9zAXhQGGP/p01bgs7PGeivbXHuL1Btmx69CuA3YvajhTvyzPrTQBn2ZV9JXbcOYTx1beLBTMzR/WUxy6DD/5LVq1abt63fbqwBN0bs0zM3ypSoPrK9epbk3QyX3dfDZEFaP/lDr5/LV9/+GWdAZInT6dQ2gr88RfplH+4Ixu1k27sxwMvKt54/XwMe7NJz5tI9Bfu5e8/LlVtPLMD/fFV167sW4O/W5mk7c5x6O3r2pmSXVsm+E1flic0K6WmpRQvwt9RIx/1vs9lkzZo18u5/vStjxozR/d7bUjrDevbszgPgjZKo6CIAYJ8wVIdgmKffPvtt0Bh5hvBwlJ3A+9nnTG+kKUVHYEfOAwDz67IHJ/RyNxwjT6NeFxR7JHP7LdmefU/KMEtgPmfZHHVnWvLB//mMi3ENugZfBPFcH3AAzzOybsrc8C1YC7AcW3hmydywdJwEPEL38OfaAuZ/ls8GhBNRRsG5AZUY3V8uiYnr5PHjrttnvr2APxfx+lyguwVhPHM+HiQjeg+QTVjTYXe0f6axpKQE6yy+lLFjx8r58+fb4k5W2i7SQABregoK82TO4vGyIn2GxGfNkdWZc2Ta4mEycFg/SU/b0sQZbcoToStPVb40LpwDAhxU4G97t/9tImxdtFoDDpdf9uyvljkr82Xe6gI5cOo6wuqeHtTX6oKshF2qAQvwd6n6O574jbt1kpBxWKbH7tGQntsP2j6F3vFcda8SLcDfufa4fv26LFq0CFtb/lYXWPKwJB7O5Xa3rvP55MkTSU3djJjtZRK3vAxbZd7DAly3jn4XFHkl3/xiNJwdgsISv/7m4/+mZ89Jw3Tmt0DzID3LQCiQed/8Zblm2bz3omvS5POiUswslNbr9p889fdjrA3g3v8c5e/bb6bExJViVsGhtEw+OaJP2XJy6yQRp/nOmpMk0dGr5ezZsxIMBjrXUC8ovT2A3wD7GNnHAt1ZHwyQMd8MlEwAOdsr7JdPMDh8+HCdNdq6datwhNf69AwN3L51S8rKSiUvPw+LcvP1u2fPHjl6tEJqn9SqEFzcv2XLFoTbfSZr16596YwgR/e5Voi/FuB/fX7gxLacR8/ekh3FZ7E28KLcvF9n7cP/+tTfYZQswN9hquweBfEI7LM192Tvscty/PwteYwRfqthfNY2FuB/Vh8d/R/9raamRsLDwuStt95CjP1/ysSJE6W4uLjViy1v3Lipo9zDhk9FGM0qxMoflHXrT+t3LX6bf3m/+f/Pu163AXk3nMEuPmf0t3keLde8r79Py2tO88euTXpmOUkpx2XJ0hwsCE7QUX3+cqEv4/2bl8F8/D8RJ/eGhW/GTkIzZfHiWA3l4faEDoyKB4PBjjZPq8prK+DPHR5ljOwPxG48Hw2W/h99JZvWY5vFdsbsEwjyNOAZM2bompCw8DA9Lbg1zNP/9KAx62Cg1qir09Jw7YXP6xO3y9P05eJc7uhlftiBS0pKkrfffltiY2NfCvhpVw4c8Nf6vD4N3MIBnut2HpMonLLL/fjtLqvj/fq033GULMDfcbrsFiXxZXft9hM9ZTcsvlCKjlwSt7frRgq7hVJaMGEB/hYK6YR/6Ye3b9+WdYjdHjZsmAwePFgyMzNbDWCZnyEtBw8ekpSUjbJsWYLExCTgNxHx9937uyw2SfmMilotS5eulMjIVeAZvIfu87pJBsgUE7NGVq9Okfz8Qrlz547GMnPP+cuXL4vL1TVrcNoK+PcgjGezhvEMllEY2d+Sli6cqWnP5+rVq5KQkCDf9O6ts0QLFi5UXRBAtuZDQMjOwqNHj6zBjtYorAvTtBXwdyGrP1nSPE/31n27pO0+KTPj8iQ1+5jcsSIHeqQ/WIC/R5rtxUyfv/JAIpJKZFFKiZQdu2oB/hbqsgB/C4V04r9cXHfhwgUA94M66t9a0NacpYcP78ulSxdQTpVUV/eE7zm5ePE8eK6GzNX6e/HiBfB+7gf8U6aLF8/JrVs3JBB4enYBgT7XQwQCXdNZbwvgzxu5TMN4Zn84UMZ8O0iyMjLbNbJv2vzIkSPYmnUSdjjqLytWrGj1yL6Znx2NXbt2aSeBHUfr0301QMCfnJws77zzjsTFxb10hL/7SvLmcmZzenHo1mXZmleJXXrOyvFzt6wR/h5qbgvw91DDvYjtx3VuOVR5XXbvv4D9+Ctk/4lr4rFG+ZtUZgH+JlVYF5YGnquB1gL+/FFxCvan/K6PDP7sW2y9uUW4PeurfG7evInY7zI9k4E79LT1wxH+qqoqa4S/rYrrgvQW4O8CpbeR5KM6l6zeckjGLd4pSduOyh1rq+82arD7JLcAf/exRYdywgO4UhFzN35JtmzcdUIePnFZi2xCGrYAf4e6mlXYG6iBlwH+FV9PlJ3DImXrgAUyGwebjfi6v6zH2QvtAeidoT5rZL8ztNrxZVqAv+N12pElcpe/Y1isuyHnhMxdXSiJWUfk1oNX69B3JH9WWW3TgAX426avHpP62p1a2QSgv3b7UckqOqNbdN57bB1dTgNagL/HuLHFaBdp4EWAfx4W5a74epJs6DtX5iCMZ1zfobJ92zZxuqz2pYvM1WPJWoC/e5vuzMV7snzTfkQKHJFDp2/IfWCIQNBaMN29rfbj3FmA/8d106OfsFI+xlRcBXrnYdg3l/vnHkaF9Qfq5ace1WoB/h7t2hbzr0EDJuDfMiVGauZkiSeqXPyxB+TolHUy78MhsvDj4TLl/X4y8qvvdetNp9sC+6/BLG8cCQvwd0+T+oEfHiAq4Pj527Iibb9MWJothQcvigcn7lqfnqsBC/D3XNu9lHOC/j37L8qYRdkSta5c9uyrlv0nr4kDe+r+lD8W4P8pW9+SvTUaIOAv2pMvaZMi5fyMdLEvLhJXVJkcmpgsE977Rvq+/aF8/0kv2bxxk9TajD3VW1Mu03DhtrWtYmu19WanswB/97MvBwQZIZCy46gkZB2WvcevSMnRGrmK3f+C1mFb3c9gbeDIAvxtUFZPS9qAHSpu3rNJOSrsFqywn4HDuLhVZ/W1R1Lr8Opof0+TqSP4tQB/R2jRKuNN1oDT7ZKCPXmyZPAUWd9njuQOi5b8kXES98UE+fN/vCt9v/wGI/uZUtvGRbWMrb+FA5muXbvWZVuOvsl262myWYC/e1nsic0jR87clGPYiWdN5mEZhzWAO0rOYKc/PzrpDT/56IDuZa22c2MB/rbrrMfl8Po50l8t89cUytodFTgp75Jswp66l2+1b5/sHqeAFgxv2LBBvvjic90OritPM23BlvWvpYFuo4H6+qCcxV72a+PiZc3cpZISHiOpC+Jk2bT5Mm/aLMnO3iF2p6Md/DbKlSuXsT1ptbicVhhQOxT4RmXx+3zPbMspFqTsEvsy1PfhE7ecOHdb5q7Kl7hN+2Rn2TnJKjwtR8/c6BKeLKIdrwEL8He8TrtdiRzpv3zzsVRduS8nLtyWyNRymR63WyqqbsqZS/fkwtWH4vT4hcdns+Izfo9Td+zR628DTq7EN4D/G3gdum9cN2oaXgeDRp56TNkH6o0ygsyn9410eq3PkQb3mY+/pGOUUQ9aKAdfvQfe6+uZpv6ZtMoP+cSX/PA57wWR1uD5Kf+U36TVgOfr1qXKpzjKfU1Corhw8iM/zBMEz6ZsTXySBp4ZMhm/Bu2QLkJ8GrSfymHKxLwGb8Yz6taUjbyr/KSNcprr0NCL8Zzysxx9zrT4Nsmv1waNJj5D5fJ/854pl6kL/q96pZ2o75COVU7KTD6Zn7ah/UL21XAM1bMpP3gL2dGk9yyfhp1N+vQvU5fUC/PQr8wyTDuaZTCtygo+zXtqpxAPhp2e+pjqVvml7D/kzSjLeGbKStkNWSkzdR0qDzRYfnMapj71Hvg2eeKvKWPz66c0SPOpnXht6MLgsUnH1AXLDfmsys98+DZdt9C34QtGPqYxfMyoNya/zXnjPaUR0qGm0Xshm+OavNlx0u+lmitSCeBvfk9Vnsa5Ajhp0+HEoVbSZEuWzy/tR95NGuSH9jLl8+GU1dpaG7bMfIITU71NOvOrzaBv5cPQq3ltysdfk4ZRrlG2Xod8yUxLG6rMLJdlgg/azLSXyqy8Gbow8pn2b2EnlK3PQ7xpec35DMnMNEa5oG3Soy5AV/XBe9Cr6gL8mXWBz0wf02v8Tz55z9CbyaPhAwYNXhtpNG0TbwY9ozyzXCO/ec/U3Q/tZNIxyjX1Z+Yz+WziDTSf6s302VA9payhOmuWw7bDrPOqK5yj5nS59aTdd955R2KxD78f507Qds3t9INr6IUyaBm8Dsmuvqc6MdoVk66hr5BMtBHyGm0Ifpk+pG+mM3X79Jr5aH/jmaGL5jp+qiuTT/KlvLFc2J/XZr7m8vParNf0i6dyGPIpD5BPy20uM8plO8X7LNeka/qV0ZYb+uFz3mdZykMoL+/xfzv22Hd7AsKze5K3VQDgn5GFicUyOXoXBgkviMNlYIJuB2oshtqlAQvwt0ttPS3T02W6VzCqvybjiB6isWvveVmQWCTrseUWY/uzS8/Kyeo7GrPHxTrsIFRgaq/q8n25gvg9rtK/frdOztbcQ8fhDmL6auUcGgp2HBg6xKlApq1B5+Lo2ZtyCmWdxf+lOLTjIsKIjiHdvhNX0cF4gOtbOCvgBk4FrpV9x68izOih3HmIMk5flyu3HssVlFGBNPceOaWq5r4cAH837tRp/pPn78ilG49wDTBSfVvT7j12RXk8i7QnziNk4M4T5Z+HhHAbsYM4l+A8OjaXbz6RsIhl8s6778vUOYuxEOkCZLIrP4dB+xxlBl3qQ2mcvGp0iq49gEw3UMYD3H+surjz0CEXcJ9yUi+n0Xmi/m7hFEKeg0A+T2OXg5KKGtUL0xUfqZGLNx7KSXS8DldCpru1upj67OV7cgs6rMBoygXwcB0xlJT5zkO7nLv8QI6Dp6uwXSXKoJ5v4qhzxlaeqbmrPBw8dU0qL96Ri9cfQp9XkL8O/99Vfi5Dn6dgr+PQy11s10q9X4DdaLsDyEd7nr10X8qhQ8p3CHyxPL4EKmDH4+fuwCZPQPeGln/p5iPlgfxcuv5I09A2pEd934a+6T/8/wLsui+kQ+q28NBFvUf7k89q6I9+cQQ83X/skMOgS70zXpR2uARd0bfIJ8tlPvrbRaShj5En8saDYcgvfYv3aBf6FK9pG+Y7Dnsw7bGq27DZXamBXvbhjArqgXkPQt+8PnnhLu5fhayU84bakjzR/ythzxraH/xS17RTRdUNnDwJO0G+k5D7NmxDn2Ve6r4M8a/sWJ+BPsrBJ+3EelIEXdyE/SvgV9QbeaNfaBkojzyQLuXgFDvLoj7pyzcgE/VDPas/4h7tS97oFyyf38Pggc9Pg/bhM0a+SsjO/1nGwVPXQ/6NcqFj8sjnrLOXrj+Wk5D3RPU9uXrXJqdrHshp+OLdxy7Idkv1UwP+yBvrPOvsIZRBH2MZlchHfmgv+hH9hbOLDClkXWGdPQffO3PprtqPccMcTWRe8kabkU/an3Xv+r06Q37wzHIpO+tsDfzxIHyW+qAOD5y8jvLvN9mf+mM7xDpNHk7AD6hjtjPknXpjO8T6RvufAU+8ZlvC6zLwydnQQ+CNNC6iPP6eQBmss/RN1nnS5r3rsCnLPgJ/efjEiTp9U++zbrGOnQBN8mTWhbPQBelRl7Q//foGZD2CX6al3elLLJM+y2vKz4Easw09gs0ZyBPrIzdnYBtxEWnYtp67CjmgY633qBfkg8+5dTPlZz017GTIdAl2ZH2iP9MvD0DfzMO2Zz90ewMyMz992qhPN7XeUQfUBXk36zVPZD2PcujjvM/nlJF+XVZxVY6fvSZxy1fL2++8IwsXR6Hc60pP6yxoU59Xb9VqG0udsT2kz+oAFmx6FHzwPmWgDpmWfkzfYF05Cr3wXcX2Ym+oPrH+0R8pM/2C+lJ9Q6/0S95nvaG8bANYn+jf5EXrJO5RL6p7+KzqB35Dv6cNyDt9mdfUMWUlz7zHNu84RtFpO/oCfZztSAneC2y7ToJXlluNe6dhM9Z15qO+2abxPUw+WLfps3zP8F3HtNygg3WLbdsB3KfP7kf7Rl7ZxvMe30HUF/NRh3wfleJbcOgSFubmyPK0AwgBviq7911Q+Z8ih56GeSx+n6cBC/A/Tytv8D0ewHUfIJrba63DEdmzlu/R7TujN+yTEfO3yYZdx2VqTK5Er9+LffyPy5K1ZVi8c0xyys7LPOzDy4YhCYd5xaJh2I1FwJompVSK8YJYkFSsZW4rOSuLkktlFQ7rWIu8YxbtkK0FlSirVKZg5CAt96REotx5qwp1IfHU6FysMTilYGEBwo52l5+XXaXnZWlKmTZiPMp71vI8fUGSt+Ub9+lWo7xeteUgeDsnkyNztJFKxsEgMRvKJQ87CqwD/5FYrMwX9uxVBdqxyS6tkl6Dpsv/+09vyQe9xsqoBVmys7xaliId1zdQnsWQZzrWO2RgtGMafhOzjkranpOQqUQ2gvcsnDY4D2nZaPL+SvBAIMOO1IrNaDDxYpkLerHYzoz3RkN+6pDpRizYhnJPY6uzfRIeXyBFyBeOshhqxZfsIuiQ26kWHLgosyHzIbwo1oOnaPC3Gx00HoCyIKFYZZoMvSVtO4Ijz0/IbOzCFL/1kGzNr1Qd50P+1fif/GSXn5MV6Qdk2ca9eLndBO9Fekw6ZZ2xIg8jO0dVRjb4lG8udnWaifI24HoRdEHfYOdwPkZ+tsCOWcVnlIdc2IkjQotxonMh/CJ+62GJSd2LF+s12GCvrNp6EPo5JdOW7ZaEzCO6l/PQsExJ31Opi8inxuxCx/MkrktlPnTATk04aGcVVsIvLsD+pTqlXHr0sswEn/uh1yXwCX4zocMI2GMh9EXexi3egfKPq1wLwOdm2GVL/int0BYcqoHvlYOnffCVC+oTK9MPyg74wjQeFQ89bIDOSWM9ylgGu9G3thaABsrnLle0eRh8Mx723MZr8JsBPgsOVav8B9CpoZ1oV4KsdbA37ZRdek5GR+yQBORLxHcCYmLjoRf69LDwTCk9ekmWri2RZfBZ2ndF2j6Jg3/vO34NesuFnGcgRyXqY7n6+WbUkzmYci/CS3oR9EA9c9td2ozX21H3WMdY79aCRtiaIjyvUtuErTH8bVX6IfjlYS1j5op8w7+xxofXCbhvHLKzAzo8JYtRTxfi1O5M0KAf8Eva9O8M6Icnb5K3HRgsYAjA3JUFCnTjt6Ac0NlWdFbrUyTq/p69F2QY7J+We0JpTIzcKevh90mIFR4P+9HnIxKK8Oww6lMN7FSmeiuruKJ1j3TTweds0KCcy6CncPjyNvA2eyXWG4DXHPjkbMqBOpsO+9Nn06E/6mN63B7JgY5jkM8MWWD7EIO2Lu9AtcxEfaP9kzDSOQtlJG87Jom4nrB0p2RDNuqS7QTlXgpasSiHvNEeBIW7kIb3WKdZNyhLJTqPrNOxqHts9yahnYrDNdu7IfMyUBdOSTJ4nYl2mH4VB/uznSnGIAH9mzJSr/PhS2zn6LPzUS71moawTLYd3IxhYVKp8sS6GQ4+1WYob0rMbpWJ+6fPQfu9CfWCe6rTNwkil0LHbCdIm/qLhf9moexxaLOoC/oU2xaGgNIvpqO+FKM+LkwskSSUuQttSxTKSEe59Ps5sAMXeG7B/9G4z070BtSLJWgjCGr5nLu+ZBWdluFh22TNlr0yf3GM/Ortt2XyzPmyOLlYZoAG62wk/Ip1dtfealmF+sp6QDA/E2WwHU/ZfkxPlE+HLlej7ZmJfEzLHeloY/oK2y++q9iWT47K1V++l/h+oszJeJdRh8XgmTpORDvF+7wmTdp6UtQuxLFXab2cFrtb6yT1MhfvL74PSW8TfJodsaXw86LDFyUf/kQfYceCbSTTFBys1jodgxl20xfo46QxeuF2fZewveZ7lrKuAS98b+1BvTDbNAJx+sQ6nLHDMF3WQ7aVrM+UifmWg+9Z0NFO1Df6PHnlAZyzwMMKPKO+psPfCjDgwHaQdYjvSPLJNpSDReYMglifN0oDFuB/o8zZemE4xacjFnhRceSLjcIsAJ4cNOBTATT4YmOHQAE/Ggu+dObgRZt/EIAfAJGgKBcNDdMQILPB5GzBOrzA2WBGACSwMVmLRnl0xHYA0dNo9EvR6OYooCWYZyPIBmwyGlS++DgKwnUGfFnmKOAvVcDPMmciBKkAIHZKNF6Y4JWNE6cdVwLIkjeCB/LDlxAByB40uOSNjW7p0SsqG4EdX5hfhQD/H3uNQScnE/mrVU42nsyzEC+nqQCpWwvOqC4IVjfhZU2ZNuw6BfBzFi/5AuWX91eikSbgJVBZjpcZR0hmQ5dsmAmCR6ExZ4NLXoeHZ6kuCAwIJJlvHn7XQqcEtpxO3YgXcj52V5qFxpo6Sc2uUFCcC7CyGjqljkrRQE+E3pLIG8EqGvDV6FCwwZ8EHbPDwxdbHPjZUVYlyzfvh43LMatwUw9QId+UlS+EJAAOvlzGLdkJ+U7iJZMvMwh+cB0Bm0UBEBFIhQNoEDwRhM4HqNiFe7xeBH2xgxEPetEhwM8O4yrISzoEzwS863eeUJCzefcpBRq0JYFCJAB/GDo/xzHyPm91vmSiU8GXGEFCJl6GpUcuAwTsUUCxOJmAv1RfkhEAB/Q58jYGPrYeL0HSI59poEGfIvgpOEjAX6ad2J3ouBLk8MW6DS/xKXiBUw+c5ZoOmekjMRv3q29Rl+xQEExnFqKTB5sTWBAQ8R47V5SbAI2dkfXZx0Ng/Sr8vkLttKPknIyE/ROQj9/xS3ZATwckFT49NCwD4OgSeCsB6CwX2pcdBgX86DSxQ5QBHyRwiQJI2InOA0EewW0hwBNtQ1+nDQiCqXPWPdqf/sa6FwbwkFlYpb5AngsARvnSZ+ewAKCaMhMMboKuZsCHeJ/1lp100qJsBD9b88+Ah73whXIFOfRv1mkCT/JGAE5QzDrNke14gH36qnYIUQYB/24AkKEAuTwIkM+41d86nBOSBLAybvF2BUkLYDt2SAj4WHfXQGesv+x4EUjTd9hOUU6CQcrHDg8HLtgO0b4cHGCdZYeFvrAZ+iM4JFjbCR1Ho/1g/SO/BPzR4H83wBN9jPZnB5/gnyEO1A0PL9wB3ZMW20VtyyAP63cpAD/tUX7sKtqsc+i4AfCjTu8CHwtBm6PEEfBRduhYTyei8xAL+vTNwXO2qjxss2agfaNfEXAvgl8Xw+fZ3lBGdhrpx7T/9uIq7XhSZoJM1hsOvCwAAOeACn2BHVMCRAW5GBSgTAnoVBE8skPPjinr72EA/iWoC6vRASXtxchPvdDXx6CTynz0Keqb+QiAqcMiBS1dCHsAACAASURBVMdFAMeHtRNCkEu7cPSbfkFgn557SqJgP7Zf61G/WJc5msznK/DuyEK9HjpvG/xkr4QvCgH+GeHaOeJgC98zDD2NXo9OevkFra8cKDgAwD9jxR5tx2kfdmw4qMC2jptS5MDHyC87prQ/2ya+q9iWT4w06hP9m50pyszBjnkceAFv1DH1xM49O1VsvwzAn6OAn7rhu5E6Zr1jR4j1fzb42YiBsqPw+6WgV4hBgDz4EwcsOLORCdqUm50A1uno1DK13QS8sxTwox6NxEAQO1V8f/DdwjrPwSKCdPpmeEKh1sdcyEefWIe2Jhf3yQPv8T1DmQjm49DB4fsgG+3itFhj8C4FumKbzs4A6/UYtPV8Z7PjRLk4C/Cozq2zFD5/feuBhJWyR2nAAvw9ylwdyyxjSPmtc3oxdX5Hp5IZSkGwzOlGNgLs+XNqkiEPBKfVmM7klDM7CZyW5JQk03DKkyM7On14GWE8AK8HT93Q52xUGNrCdGycOKXLkbECNOosg0CA0/jXMBVafAT3QIv0mIbT+ccREsDGktOyu/edV3DFKWoCfL5QOPXMkTuWxZAADRUJTbtSDk47czSDU/mcCp0+L0p++Zvfydip8wFIKjHF+UTlpXyUh+EMbGQ5Zc1RFMrLqVbKxHAPTo8yLafuGcLDaVeGTXBq+ACnlaFD0tuPUV++VNnJYLnklSOiDFPgMwIYhk1wav8YaHCKlS9LdsQ4NUuZSYM6ZfgDdcIR/2Lku4LQpF3QG0fpqE+mPYTwBk5tEzhWQ1cMQSE/DM06gJAHvpA5Hc/ZCPLNqeU8jB5paAzocxaH098ctco/cEmvqQtOhdMmfGEzpOUMwiXIJ/mhHTRkC/o+jKlopmWoF2mRV9KhLskn5dqGlyWnlcvQ0aT9CIhoI+ri1n276pXT6rQlZzzOgh7DsPLBJ6frCbD4pT+RrsnHTnRqWD7tRD5Pwp9OQRe8po5JgzypX4E2p+E5nU/7HoMeuLaFumAoAH1beQvJRn0x5Il24lQ/Qz14j3ZkfSAg5fQ6wy80bATXtDd9uQo6ov05+sdvDvg8DH/hbAZHpllvWC/2opPIKX3m30cfwjQ9O8M8+OYUQmP4nKER1Gc+RgoZdkE9UN+0B21G+RhKxDrG6X3yQL2Sd/pCIUYeWYfor5SDIR95qIO0DW3CUW7ePwg7shOtPk8dw+/JB3kgDZbBzvdp6Jh1gaPbDEk7B33y/g2E/9AX+GW4CutN+bHL8KGH2gGgrhm2wNFhnuJ5BLNOOdARwxhoTz5jfaLNqDOGstDWDLVguAJpsB5TVyofZCbwYntB+7IusB1gWoKaU+CdbRjbNYamUAbmrcK1ykTds75BrwzvYX2gvzEchNccBGGoD2lxxzOzLWMdZrgN+WEbwzT70VFjCAf5oCx3H9m1XSRN1lN2knlNv2dniHpnuBJ1z/Ah8kVZ2Rbwl+FwbN/oxyyTchfBr6h3lleETtuFq4/URvQj6pvgVW2GNPRj+jRDX9hm0X9YT8gb21bKw3aL5TE/ZaK/sHPBfOSVujBDTahDhlBRr7QNfZJ2ohza7iEt7VSJesg2i6PFbL9M+7EstklsW3egvdl/vEZilq+StzHCH75wifJFn6T96Vekz7aAbSffESyPttZwU9i0BDpiXafPUIfn4WOkwXzazsAnqFPSY72gH9O/yQ9lpn0LoU/WQ+r4CGzBdsHwQ4aG3ZNdeOcY/oZQF8S1s6y9CCdlG89wHM5M0RYMCyoHPYZVMWyG4Xs3cY/p6SMMD2LnjHlPIhSV7TRlZDuyHYMPtM1+dGhYLttx1kX6JMvSNg06ZZgS5dGwIPgs6z31w/cM2wMNbYKuqAP6I9s36oL+T72x7pMO2yT67KNal/7WYdc+6/Pma8AC/G++jVslIRcM6QIf/HIBrxuhP15fUFxY0OPBrw8dAxfu85eHb/A5rzUNrtlx4HP+zxECXjN8iP87XD7x+lEW/ucZAHqNckmHZXCxMO9xsaFJw6AXUJ68oGcuKOav8ob0LMvgg9e+ZrwZC41Im8+5OIq0SMOHb2LyWvnkz5/KylXxOJzMCboNms7kn7+kw7T8pbwqZ7MymIb6YpluPKf8lLc5PV4zr8oPXvjcHtIFn7EM1VuIxjM6DOmFvFP/XFhFvZr5uMiL8itvIT75jPw4oYvm/HChJHnkc/KsciIdZTLlM/g0dMR7/LIspmU+kzfqxLQvbcT/TTlYhkmDv2p/PDdpkJ4p//N8gbyZdA37GzYgbd6nneiP/Jq8kTbT0v6819xOZhrmp/7IE3VB2nod4o15mFZpU9d43uSnXr/KZ8pMGc1r5nnqp0ZdYF7TZk28webMZ+jYuDZ1YegVMilvhn3NMsgP9Us6yi94Vz5NvzFlgl+YdZa8MR99xtQFy6AtmtuJvKheURavTfmb+KQ+UQbzqBwhHsiHmU/l1/sGn6RDPmhHLYcy4d7TMrAQuJn/s16QT/LGa9Pm/F9thl+WQ326YAfeM+WnnOTF4M2Qn37BMgz5n/oI85CO6pN1AfkMfT7Vt+YD75SZNE0d8tpsW1QO5GV5pKW6CLUtbLtYht7DrykLZ1KZj/eZz2yzeG3qwqTnDfFmyspfg09DhyzT1Cd/yaszJLOhY1PfIb2gPK0XIZmYhnRNu9JOLJ/0tVxTL9Ct2WbxOXXBPGxDeE07sCzNF+KZ5bJNMp9revDWvC6o3yB/c13U2Z2SmJQsBPzLlsUqXfoQ03IrSKbltdmWsDw+Zz2mTakj0uZz0jb936TRVO+Rxnhu2Jf5jDKe1osmmSA/rykvy6bNqB9DF/BTvUYbqXwa5ar89FPITH75ZRnkl+lNmciDyRvbafPa9AWlgXwqE/WtNIxyqVPTx8xrQxfGO9lsFw0+DflMeur/KIs2U5lQ38ib9flpacAC/D8te1vSQgObNm2UL7/8QtauTRHsdWDpxNKApQFLA5YGukAD9djFZm1qqrz963dk+cqVXcCBRdLSwE9HAxbg/+nY+gWSNkqjyyUNT55IA7biawwGpAH/N7rd0oh9kp+5tuO51yuNHq80YB/tRh+vPUaaAPIxD/5v9IfyMS3LQNoGpsX/Rj7cQ9oGl1GGPmd55nPm4zX5QH4tg2WDhpZPGuZ16P4z/JAe02DbTZXJ75dGJ8pCeevXrpXPsS1nYkKC+PA/TgAyaDSnR/q8Txr8RTmmHEqfaakr0sBXaZMGr8kP5Ff9ecmnH89DejNpQH9GPgf0YuhKZSVd5vewDOajflCWed+8Jk+wk5EWNPy8Bm3yg2vy1sA82PZO+dT8sLHdjvKMco20uGZZfE6eaQ/KoXmNa5WHPJFP6jAkn/LGdCo/7od40/JYBg5vUjlIj3xQh5QpZHPNZ9LWMqAjpQG6qu+Q/VUm8ozyWK76XEg/yifuB4OGnZg2VIbKh+eGzIY9VBd8Th7In/Jj8kn5cD8khyGnoQvyr3oFv1puUxm4Txr0BeajHCFdNV2zXqicBm+GXniNfPQLyoM8ph20XpAPU1eUyRvyEcqrdoLf0ub0I9KD/MoXr02fpe+RT7vN4JE0WH+Zh3KQZ/pmM5m0LKVBfRs8NMlB+ehjlIW88bnyRjlC16THa/6aMjEf/Zm80TeZnzwwDa/JJ3XYlC/EG/nAPVOfBg3cow8gH4aUtQwtl9dmGSp/qP6DjspJ2nyu/g9+kF/LIw3ySdqmTLxHnkN5TH0rXfol75vPyT/lMNsC0qBvUS/0Ba1PlD+Uj/IzP78mveY6NPlheeY1aTAtt0JlWtLmPeWT95GW/q/psc0w9YNrpUP7Umbmw9fgE7R5HdJFA+qB2sGUyeRN6YIW5WNd4XOWG5JZeQjd572ndgrxxrTUK/kiPc0H3lg+rrEyVDyPHksygP6v3vqFxC5dKvVIJ6SnvBn5m9PT8mBf/SU//Ib4NfWpfsF7/LKsULuvdUj1GtIF6x79kLyFfE/9wpSP+ck79WneU3ohXSCfcZ/PQz6k9EJ2MuVnOpMP1lmTN/Mdwf/pFyGbsx4/40+8b/pKiA+jPTXzgR7a2Gd0YfJJesyjuqA+UVZILpPeC4CB9egN04AF+N8wg7ZFnODtW1J/84bU37gm3vw94lwVK57sLAnevC7e4gLx7d8rgTOn9ZmvrFT8uPZsyxT/yePiP3JIPLt3GdcHD4inIE/81RfEW1ok3r1Ie/qUeFCm7/AhCZyuFM+uneI7ViE+5sO1/8QJPM9DGTnix31P7k7x5OwQ31E837ldfIcOir/iqPH8eIX4K1FeYb4Er18DX/v0G7hYbVyXlwmvPTnZTbx5UZ4PfPkO7Feeg+DNl7dbfEUFsi46Wj776GNZExUlHuSr3wM5zp0FT8fEi+eBqjN67SsulODlGvGCrm/fXvBQKe6d4PHgfvAGnrdDF5DHd2Cf8sl8vn3l4ikuEv/ZM9BhofgPH5TgJfC2Iwt5jkAu6Co3R3nzlpfqfd9J6CIvVzx7clU/pOctK5HAWegb9PzHj4n/6GGjPPDgP3FcvJBFdYEy/NQx5CCP5C1Yc0ncoOeDPL7SEth0O8qA7suLwXOGBC+c1/zkJVhdrTLzefD8WdzPleDFC5qeNAJMCzm8ZcWQ/yRsB97JG21K3igTaHp25ypfal/I58dzb8Fu+EKZ2s5bWCBe2Ml/rkrtTB/ylkNX9KEjkA3PVEfgmXb0loAe/QL8eEtxXXVW/SYIOWknH3xMZaZceF5/+7bQXqpv0PDmw9bQi/pQ5lb4biX4hJ3gFwHY2luUJ76SQgmcPwf7wk/3lxv+nZstAVPHfI6ylA/wT9rKG3j1wSYe+I23BLamn7JewC7qC/R76oV+gWvawwNdkjeVm89Rt7Rc+gXkpIzqF/AnL+iqzWh/1A/6u199BDKBFv1M7Y9ymZd1M3jrJuqOQTN49SpkI19HUH/hs+kbDfvvg46ztym/lJ31jDJ5C/aIF/VEbVaUr75GH6GsWm9A30ufpt+Rzz3wC/i31mPISN+kHen7TOPBoVz0WfqNys3r8hKtRwH4m6GLw+I/Bb+Hz9N/1W/opydOwjb56r/+Y0e17fFA9/4zOO0T/PjgK8FLF7Ue11+9YtgVZbMdU90W5knwGuSH77Etoa9Ql7694A303FmGL7Ceq+9pvUAdoR+qTKDBtoc8oy1jmQH4nhd8al2ArrVewx7eItQL+A5t496eqT6scuzeKV7al/UC+qYNfKgHaifKD9r8X+s39Ubd04dgD9LlM/d2+AVo89qLtpV1kvZgXaP/+kCbsilP8On6O7eNvKBbX3MRfgHdoZ4Gr9SIOzNd89GPyA/90ndgr7i3blbeqW9tv2lf1Du2/SyXvkudaNsLH6GcrFNqM7TH1CG/2i7Avz2sf2zX2b7B19neaxsBm2m7iLrMd4gXbRLzNcJna3dsl9VDBslb/88/SOSA78ULGwVPnTR4Q3ujbSFoB/A+8YFPL+pD8PpV/O5CO3ZObeNDGxhEPQ6eB8/gg+8zH3ydNAPkE3JQJq1DWRmGv0EHbuoCtqZvuPHuUf8BLQ9lhW71HQjfZhvAMvzgi+8A2kPfi6Cr17wHH1Ufov/zPUb7w7f0XcZr6gXvveClS1qHyLP/fJXal3WV7bDqCvnZdnjZhkJvWgb8l3ZkPdW0fLdAh2yztG2B/dmuav2F/QLV5yUAXqljPte2l20oZKL/+1Bf9D3Ea7Q3gZoafRa8fp0Ha7QFQlhpe5gG/qKH8Wux+6oa4KiD3SYNdbXiWpckzrho8WRsFtvUcfLw5/8otqHfi//QfnHMnibO6MXi2bJJ7JPHimPBXPGkb5K6gd9pPteKGLGNGyGulESkWyq2KRPwkt4m9jnTxY60rg3rxDYJ+WKjxJ22QepGDBbXmpWgFyNPhg8SV3Ki2CaMEduYYeJcs0rqRg2RuuEDxLk8RuqG9hdH1FJxrl6J+0ORb5W416O8aRO1IXUsXiiOpRHaQDqW4HpBGBrTbLENHyjutYniQsfFNnaoOGNQRvQSqRvwnXjRCDsmjRbHnGmSMnOGfPr++xI/e6Y4AfzcY4cBsG1F3iSxz5oqnswt4lmXKI650/UlY58xSUjHvSFV6gZ/r2W64ldKLXRBfh1LF6ocLujHsShc6mZMAchKQ1nQYWy0Nt5P+iPtqjhxRC6WutFDxQG+7GGzpXZAH+giSerGj5C6scOhr2iV0z53prg3b5Qng/uJM2G1uFYsE/vMqeLemKoy2iaOwQttrzgWhuFZjMpPHl2xkQA7e8BnX3GnrIFu5kAvA3AdL86IuVI3qI/4ABzsU8bBFlH6ArXPRrnJ8eJFZ8A+cZT4oEt30mrYb4yCJdK1h8+GDVLENnKIOMLniHsTbDqkvyFTFGSiLyQnQN4oqRsJW8P+dchvh23c69eKHbazz4cPZWxBPvCzNhn/z4MuYP/lkC0cvMFOrqR4+MFAsc2ZIc74VWIbP1Jss6cDAKTDhwaB353wuQRxRITpS598O+fPlgAAE21LXXgBvuizTly7ElZIXf9v1I+dy5bAZ4eLF7Z2zJos9rkztANknzZBnEsWwGabxDYKvEPH9Gs76oAHL0VXAviYPkm80MsT2J9yONashr1GqL87V6+QJ/16iytxtTgjFwlt41yJeOQlEXpNsGgbN0p1Q7+vmzhW7IsXQL6V8oQyQ07yQr9g/aA9amF32r9uDHwC5bjXJas9XPEroMN02Gm0AgCVf+E8CRw5KG48o7/70RlyhM8Sd8JK1Fn40Fcfq32dSxdAt/1xf5U45oEe9OwFCLFPHa9+6d60XujrTviaF3WZdVbrDeqbHTagTmzkc/xo9e+6EfAF8OZcuQz1bQRsMl99gHXEBZ+l31AXroR42Bd+OGWigta6CaOVhmtdCvxmpMrsYL2HP7lTUAfBA/2XeidvrPce1CfyS736ABwds6ZoR4/yOSMM+T0bUtSugf1l4kA+V3wc9Aaekc+ovylS27+3doDsyK/tF/zNMX+O4b+bYX+0TQ62G4ngGbK6oHf1WfghOw6ss5TLhXbGNm0yvhOgl2TU4+8Mv6AcY4aKPSJcnKgXtaj37rSNqIfztN54UffqpqE9QTtB+5IHJ9oE2tWGds+VjDrLNgT1lLQdi+Zr3eEARt2oQWjTlsBnM7VtcqP99qKDaodPB06iY4P67oyM0PrvToSNIZcPQLe271fwvUjYeBF0jPYrapG2WU+++RxyUN+TpW4q6kDSGvjzDLHhf3a+2X67t6Sh3TPacqZ1Los09Ib2WO2Lts6VCjvCprYZ0AvsxzbbuQw6RLtdC5mcK1AXIIeNfgMd28NmIe8UaUD7dR/1P+7f/1l+8T/+Whb//F/FCVpelFf77edoY1LVzvZp48WTtt64RjvpB9C1T0A7hTadbZa2BQDvXtjHNhntIsC8c3G4OBbPFw99Gm2dtnGpyfKkby/1NyfeH3WD+qo8bEPrhoFP2hRpbZPRNtL+0IMD/ubGgEEd7rnQjrH9Vb1QDvq03od+0D7Y0H67Wb/R3trRPrEtqx3Uz6gXqDtan+C79nkz1R/YSaStHfAz9Qv4Keu5C7Zju+devRzlrdH3ITvoDtZN1GtvZjpsOxu2Xig+DGzY+P6CHvietaNee3K2w2ZrYQ/UEXRU7DPR1oXNhG0SpPa7r9Qv+P6p7ddL7PjVuoe2lbYOXrkiQXSkOfpvfd48DfzFmyeSJdGLNMDRZjYMHB3Qho4vQ4B+G4Duo7f/HYBusI5EOubN0peEB42LHY2ZNp4Ea+gQuNCYuNAYseFjY0+Qoi8JAkk0ggQzLtDgi1obTDQkBEcuvFCcq5ZLLUAM87GTYOPLPykBIGKkAXhQLsG/IzYGQBfgb/wofQkSYJJHjnA5otBI4aXMkSQ2XA7Q42irDYCJoNQFoFMH4Eowzg4NX3IEigQBDrxsUmbNks/e/6OsmTtHnGhI3Ujr2ZFpAFM0xh6Mril4QKPKERWCHdJRAAZAQFBLWWqhCyf4deAlbAMAc6Oj4MALl2DVnbEVjS90uDJORwdrAWbYcXEsAyCGTNS5HS+a2qEDoAuC47HgeQz0FadyElQS5NaSHl4iBARmJ0CBN17QHBlUkA8eKD95ZMeAL/m6EQDVeMHxhWIbPQQvErys8QLki42jR7QpgRpHwu1hBDYADzvRYYOOfBg9Y8dJgRYBIToffIERENoA0BwAM3w58AXmJHhEJ4X8E6AQkNWNxcsKIKcOYMi+BB0lvnRhOzte+gQrdSNpJwCGRQCgAAEEzHxGOxFUEPja4EcEIHyhEmQReNehc8iRbRflQieLo24qF3gLYDSRtqVfqhzTJ6ptXMnwIYBHdmpd9HPY2gtbszNHX+WIJ8EHO7duAGkbAIVr80b1TwITAh92ZGwA/17MYtRCZvJKmyhwJ58At7UDAdDY+SUgmgraAPOOUEeYo4C2SQAR0I2CXDy3AwxSPvUL/NoJOuEXrB+kS91yBLYOnWKWo6CL4BIgwrMNHTPYiSN4Kv+S+ejwHNZOGjtxBEQOgGCCR8/WNKn95jMjHWSsG40ODUEl6UGfnEGh/PRLdnjYYXfC19i5YZ3liCr9ijbQOgu/rJsy3gBBsLMdvDG9du6XLtI05J0dG/oNdcEOLQGwbSY6UOj8MD9t7mIHGvZlHeKXHUV3KjqH6LgR3BOgkDfWe3Zy2DnkoAE7tOywcJaHdmVHJnAUswwEiLBr4OA+cQDkEDi50QmwAYA7aN+0VHSk+qheCeZt6BTS3xRg83rrFqQdrh1QBcLQETuu6rPgkyO0rLOUi+CPHVHy5lqPcgFs1S8oB8CoHbpgZ5/1m/WYbRTrDUem62ZN03aC96k3J3TvAsBkZ1pBPtoQgkbSJkhn3eHMW904do4xQIN6SsBHUOdFvbcD5AZOYWQZ9YJAm/XfDd1RLs5+1Q74VjtYDvgmdcw0bLNq+36tbZ6DgzTQM2kriEX75cnZgfo/Sds01mVekx+Cd9Ub7ql9kda1cb3a1DYHemFbRj9GO+SEn9WOGmLUBchhmwy/gR0NHc6QhgMH5D7Sxv3vn8kv/u//IYt/8e/ixHvDA1Bd2w+8oR7SzuyccbCJwJgdeT8AvR0+xE4Q67q2BfBX/m9DvefotXZq0LFRnwZQJuAlWGc9ZR3jYFDdMHQgwSfbULY5tClBPsG7kx1PyOZAPeZAVh3u8Z2mnTTqBbazo4Nrmw69URdoHzhA4oL/siw7OhT0IbbftKF2ilifCvN1QIz+wFkFvi9YF2lH1kfqmH7PjgcHEFi/6zAIwVlVHYRD514HLFC/2blneXz3sB1jm2zHAA9nMd3oIPE9RF+gr9sXYhCOHSl0QNn5cuCdXYt2ke2zfTY68ehgeiCT+j3aJs7g4QhfDQ98EZ6wnvUsDViAv2fZq13cMqaX09zBG5j2PnLIAOJoDPnCZcPPxotTjAQVHBHQ8AiMvjAkIIjpQZ1+ZPiHhoJgahuNgU7vAwzrlO9xhotgWvVyjYJQDaXgdCXuMZ2GTRRjurcS4Sg6FY4pY4Qr+Mox7YoyAqdPC8MY2Hjp80KEB2Fqm1PJGjJxutKYxt5bDt5ugS9MER8Fb5cvK48axnHlsviQj+VqGaQHvnRqG40peQug0eSo+Lply+Tzjz+WxBiMjF+pkQaEqzC0IYCwAT+e8zpI/vBi4TSnhvMcxdQuZCJ4Y5k6zQ3wwmlgTserHMjH8AZOnQeqL2h+yl8P3pqmhBm+AZl91BmmvXlfQwY4lU59Ij2novnSYhlK74whk4ZYnatSPjkty5AshkOQhyBokDZlroetCeQZSsCwAgIkhuD4DoIedVFzCS8QTPFCDvJGgMiQFw0FgI7qoVdNj+sgpns5/a/8gLYZNqOhBNQFp7lhKw27Mf0C4R8BTK8zZIkhGJwS57Q5/SJ48SLsTPsjhOowQnkocyhUgKBGp/xhR4ZjcYqaLzrKoLooQkjXVfBLfcCPORKlMh7aL/X37unsD6et2alV+XAdqMTUNqf/MeXvP4aQF7wAGR7BKX8/7KThIQiboO44FU75AucQKqD2N3xB+aBf0I7wUQ35AG8Mf2AHVP2UvgA/ZqgXgSjv+eGjGsbBEBPIEWBdgI6oFz7TKXjKjHrBclgHVBfoyKkvgx/VD+qhhotRFyhXw2xgP9qZoQ7+wwek4e4dgL4TWh/qGd6jfJ2E3PBZgAFNB50xLEB5g297oWfqU/mFX7KeamgT7KGhANQV6of6N3hiXacfMJwgAB1qaAt4Y6gMgaX6H+WDjmgjDR1BHVd7McRDdXhFvAwPBA31e+ZjHUIZGh5BGvB/03/Zhmg+6gLlUsZ6+gDS1N+4rmEdrP/1d+5IsIp5y9BG3BQ/6SKt2pEyH0XowjnwBv+nb7Ke+2A/w39QL3jNOotOlLZZbEdYB+Gn6rOsC2zfDiH8CH6jukCoBv06UAUda8gW6yLkYP1Gm6DhGKgjgQsX1NfNdsoL2gzXUHqgy3rI0BMf06pvo16gXVAasBllop0YesO09G+CXspWfxl1GfTq79+TIHySbbbhF/AphKvUX78mGjbJthC+qe0JfQ/pOKpMf6TNNeSJ16h3lE/bCOiePBr1F/pEaJP6J3zP8BWEBkEHGvICm2q7B32Z7ZiG/cHH/JXwS7aLrAOwI/XKfII6a4Ocq0eNlF/+z/8p0UMAbJG/nvoEgFe6bAtRn4Kok7S5ygr/9sNv6tGOqS4wu0V+ta0CH/qOID3WaeajD5FP8M9OO+shfdqLkFJ9z7Ados+Ctr6TQj7LPAxXClxCWAzbCLa9rCO4NmwK28D+fviTttlsyyg/20XQZpvqKUC5rL+wFesLw+0oP30hAJ41HeoqR9bVvrCBhi9pvQBttjPw33q+h0CXYYn6TmZ9hA+xDdTQU/iE8ob71IPqDvwz/IlysN7SnzTsCLTZTmoIKdJzdty9eYP6nG3KeLFHLtL6rCFEyMf1A9bnzdCABfjfDDu+UIqGujqM2GIki6O047YGywAAIABJREFU2zO1p+9YEWuAB7w0CZaaFi5hUZpguy5drOTzGQvjuHDJh28AC4Cw+E4CWCCERkDT6LWxmEsXjjEd8mlac8GT5sPCJebRfKDBe1iIyDKMa2MhlT73NE9Lusin6ZG2HovT/MhHGkHzGml4zXxMRxpaRiidG4ub8FzID77r162Tzz//TJISE8WPZ1wk1pSPacwyeN1MF3qf5TI9abBc/vJ/TQse9DpUXiidmLwxLb9Ma8qA8prk1/ssD8+Zl3w0l4nPm/NGXYRoGLonX7CZ0gvxxjyhBcCqb/LMMryGHBJAGSG6EuR96Jj5TT75vCU/zXlrLpPJGxfCKY1mcpCG5kN55nP+H6Ktz8gnbW3qmNdN+UK6MHlj3qZrlF3/1GefkY/8qZ2Qn3lMfbJcfE2ZDd4MGrSH2sT0eZZh0nuGN+oQ5fC5Wa6PaXkvRI90yCdpIZ3WHd5jeSqr6UO4R5uYtMzyzLS8z2v+huzXZCeUxQWQZt6m+qtpkd6FRXvNy9VrlEUaId500Z9pM81Hnvk8JIfJL3/JB+8zP/2Y6UP+pLyZvJtpWS6v+TXpIY8uOm6SCWVoPpZl8Pasrgx6qkNTn/w1yzblZ17Wi2fKhY5J29S3KSd41gWuIT6b6iF5C5Wrv2YdIT3al+0Wy6KdQUfTkHeWo7rAff5SN+TD5NGUn7/8Kh9GPuUDZahO+MzUBa9ZBmmbNEz+yadpB1N+pH/GL5hP/Z80wZNZBsvVtoDyGHI8w6cpM2kp/yE5QvIZumKZIfm1jGbvC/Km+qaPUFchObQs45ojyF6bTZLj47Fo95cSGxkl9eSfaZ7HW4hPo30LtZvNeQvxrO8I0uOXz5vLpzoO8UNdmLzxunla1l/m43PqkNd83pxec5la0GjyhRCNJh9RHzJ5oz/xGmWb9oXOTF9Qe5E+eTZ5MPngL/Oa9uev8k+ZUa4pi5nP1AX54XWoXMqlIb6PHwkHPzxZGdpZ5Wy0DbMvXsxw6oJmpJfGxhfiDOth99eABfi7v43azSFfwhzN5KiRm9N5QxFPiOlcHe2+eBG7JjjbXXZPzrh+wwYA/s8lKTkZe7ujI2B9LA1YGrA0YGngtWvAh45KSiq25XznHYlbsUKsJaOv3QRNBNkR4O5I3AGI6y80lBIDhFzH48fMAjth1qdna8AC/D3bfi/kvsFh11g+J+L0GPfsWpugMcns1f+UP+vXrzcAf1ISBlMwcmF9LA1YGrA0YGngtWvAB5CZkpIiPHgrLi4Og/4W5H/tRmhJEDZgmKGukcNaKa6N82AdQwNmY6xPz9aABfh7tv2ezz2m3jhlV//gvi7wqxvcTxdYMabV6qWLWID/+W5j3bU0YGnA0sDr1IAF+F+ntltPq7GhUdeVOLij0yIsNsdaC655CSIEmKFY1qdnasAC/D3Tbi/mGhWy/hoWAXEXg0LsaY0t7bhwiwu7+EFdxlHkInW2RnlSi++TRqnFr8PRyHD2bvUJBBrl9p0GuXylXmrrcC5uB7Q1FuDvVia2mLE0YGngJ6oBC/B3X8PzwLKG+w+wULxCdw7jjmJckMz1BtanZ2rAAvw9024v5ppTctj5wjYTW4xhn2cemsPTCk20TLB/pioo23YGZP3mgGzAd+MWv2Rl+6V8X1CuXG3oNsC/trZBNmf6JTLOJ0eOBbme6ZU/FuB/ZRVaBVgasDRgaeCVNWAB/ldWYacXwF2DdItthPdwxyGeqGx9eqYGLMDfM+32Qq4bsfCGC3V5yh8Pr+FWaM0/dbYG2ZHrlzFTPNJ3qEeGjvPKqEleGTHRo7/zInxSWBwUG2YAWn7qMT3A7499GhFOVF+PkfgXrOhvYBn1L17036gj+Y3Yua1eFkV7pf8Ij+QWBHRmojntBtJ6AT9My1mB5jMDFuBvrkHr2tKApQFLA12jAQvwd43e20K1ofYJtiUtFxfOU3HgHACe0WAdzNUWDXaftBbg7z626BBOuB0X9+XlwSQ8fIT7BDc4nz017wlGzTdl+OW7wW6AfY+kYoS/qDQoaRkBmRHmk9793TJ+mkf2HsSIOkJq+GHoT+WZeilEOn7PVNUjJOhpfA1DgW7capSDR4KypzAo5fuDUnMF2655n3YO6uoa5Oy5eikpRxqA9wNH6uXqtQbhrmP8MKXd3iBV5+ultDwg+w4F5eDhgMxZ4JU+6Jjk5BmAvx5kHzxslOOn6qUAHZOSsqCcr64Xp8ugZbM3ysWaBqm+WC/Vl+rl8NF6OXO2XsyBCQvwq7qtP5YGLA1YGuhSDViAv0vV3yrijRgt477+ekDZ0O/1EDQOKlqfnqcBC/D3PJu9kGNWTh5KxJP/eGInD3/hvtzNP08AvBkm8/1wt4Qv9krNZeO5x9MohyvqZfZ8r3z1vVtWJPjk/oNGeYwY/+zcgERE+WT+Yp+E4bs4xgdgH1DQ7/M3yqnKeolP8UvYIi++PpkNkB69AmE4FUGEBzUKQXhufkDCl3j12XzQnQU6casZqlOvaQjYSxFStBB0psz2ytJYn6xY45Vh6JQMHO2R3aDnxGzijVsN2jkJXwJeQGsevpFIy3Akh7NRLgDor072SUSkV2LAA3+3Zvnl8WNDCxbgb+4N1rWlAUsDlga6RgMW4O8avbeFKs844aFe3JtfT3LGusCGR49ePEXfFgJW2temAQvwvzZVvx5CDbW1ejKqe+M6PcacJ/Fp/Ewz8s0B/7wIr1zA6Lj5sTnQGcDof79hbgXmx0/WSxmA9MQZHoBwj2zd5pct2wIycaZHps/z6og+F9SuWOOTASM9smSZT3L2BCQu3icjJnhkVaJPLl+t13RT5nhlyFi3pGzwy+78oAJ65mHn4QrScLR/AcD+t4M9shQx+7l5QUlI8UnfIW4F/PklAXmEzsf2nIAMH++RGeFe2YbrlI1+hCJ5tPNyDrMDh48GZcJ0j3zZ1y0zwzCDsckvh3DPGTp2wAL8prWtX0sDlgYsDXSdBizA33W6bwvlegB8T2aGOObNEsfSRTgh+MgPcEVbyrPSdo0GLMDfNXrvNKrBs6fFPm2CPPnuS+FpukEcQ97y0xLwn7/wFPD7MVqfVxSQwWPcQoC+Y1dAwXwfgG4C8wOHg5gFCGKU3wsg7pH4ZL/kFQRlNNYADBnrQShPvfYvLiOch6PqGTv8CL0Jypq1fuk7zKOj/vfuYRchkDyFEKHxM7zoXHg0zOcoRvqHjPHIMJTDkCEuAziNNATv/dExIOC/fgOdgqVe+RozEMvRyTh1ul7ywe/kWV4ZiDTsSJTtRQcFHRJ2WrgQmWsRmi8psAB/S4+w/rc0YGnA0sDr14AF+F+/zttDsf7hA3GlJErtwD5SN2aYeEsK9fTf9pRl5ek6DViAv+t03ymU6zGi70pYLXb0xN1b0qQBe/G3/LQE/M1H+HkmV87uoAwYhRH0MK9syQrIwkifjpaPxKJehtssAfAfN5WLfT0Si5CcTdjhZyDSj53qlROVT+P6Ceq5g9cNgHSG3PTBmoH1m/0avkOebiI0JwwhPr2wZoCj9iUA6t8DpE9CJ+AuOgX83LlbL/OR5nss2mVHhLH5U9ER+fQbY50By52P0B7yNhKj/DsRelSEjsGkWfgf9/YeCGKRsRbV9McC/E2qsC4sDVgasDTQZRqwAH+Xqb5NhLlI17u3TJyJ8eJK24CNQKqf3QmjTaVZibtKAxbg7yrNdxLd+ps3xb1pgzgil+ipug3ombf8vAjwP3jYIAkYjf9mkDGivxsLZRcs9UmvAW5dPJuGcJ9sjPpnZQc0dKccgDpjR0AGAJCPm+aVk80Av0n3FvbRZzw+Af+GdH/TQt4bNxs0pp+AfxvKK8ZiXo7KK+C/GwL8d5oB/mIA/ksNGt/PcJ2IKK/ywc4C+eEagXOYrWAs/wSEII1Bp4SLiFssYbAO3jINY/1aGrA0YGmgCzVgAf4uVH4bSNc/eYzzfDZJHU7dtU0cI77SYms//jbor7sktQB/d7FEB/ERuHBOHFisW/t9b3HGRUnw5o0flFwbWrTbD4t2Z4Z7pOI4Frvi0C0ecLVzd0Dj4fsOdUsmQPRFLICNxwJYhvTErvIJw38eYCHvhYsNGvt/83aDxviPmOCVQRjlLymvx84+Itdu1GvHgGD8+MmgJK71ST/stBOF2PzbyMP99CtO1GsnYcBIt5bBxbtDEePPkJ5KhOoEkIZ5x2LHII7w5wPwc7ZgIRbhspOQlOrHLAH5bpQq7P7DHXl4iNghgPzxCANqAvwtzgmxRvh/4BI9+kYQ00jXrl2XR4/wUsJWTLdv3Zbjx09KaUmZFBWVSHFx6U/+Sz2UlpbLqVOVcvfuXfFzWy3rY2mgizVgAf4uNkAryTfYbeLdnSv28Dlij5gvviMHEcPf4sXayrKsZF2nAQvwd53uO4Vy8NpVcSWt0R163Fs2Sf3zQnqwLSdH2r8GaObIPXfqWZXk1xj9kRO9WHzrlmUA99xWkzvslO0LYM9+N2LpvdjdJyC7MOrP9OuwGJYdgJrLDbobDhfXMvwnGyFBcYivJ3hnyM05gPH92OJzEuLsB2G3HQL1nUizCKFBjM3nAt1r1+t1ca8Z4sN77HxwjUBv8PgNZgdI99GjBsnCgWGDUc40hPZszwnq/7FYJJyONQOcNeCo/ogJ2HIU9PeBbsuDAS3A3ymu12WFEvBXVVVJRUWF7M7Nl1WrN8rSyPUSsWidLIxYKwv5i+/iJRtkSeQm/K5vusf7EYtT9f6SpZv02kzfXX8jFq8Hv2myZOlGyJj6jCw/yjP0wLRR0Rskfs1GKUIniB0k62NpoCs1YAH+rtR+62k3YMcL3/694sS5Pk7E8vtxzo80PF371/qSrJRdqQEL8Hel9juBdn0T4J+tMfzPA/wczS/GPvcLo30ydS63ycQ2mgu5laYPh1z5sOUlgHw19t7FSD0/9x80aNgOd/SZOd/YcnP2Qq+kpvnlCjoFLmyneexEUFYAdM9E3D+32+QOOgTtB7CXvhvPa+sadR99LvadiWdzkJ+/yxP8OtLPjgW31CzCnvrcZnM8Ohfkj52DJeg08MvFwtw69Op1dFiwFoB0WMZM/LI87iB0B7H/3JOfW4pypyCO/HMtQfOPBfiba+PNuL59+5akpm6UmbOWyfyITFm+8qAkpVRK8roz+k3C78r4w7JsebmsXHNEktaebnq2JumExK7YJ3Er9kti8incPyvJqcY3Bb/8muU0/Ybut/kZ+XlO3h+lA16anoX4WAE5Zs/dIgsidkp8YkUTf03pQjI38WrqIOW0xK08IOELtsqMmVGSkbENdTxUyd8MN7Ck6GEasAB/zzBYQ12tuLdniG3KOLHNmKIHcTW2HEnrGaL8pLm0AP8bZv7g+Sos2J0hT/r2Emds5HNDelhPub3l5asIy0FozvnQlwtiuZCWe+Y3B8kNCKd/Uosdc3B4VT4OusorCuoOOwyf4am6/BCw30KoDgH+LmynyYO3WD4Buvmxo1zukc+Dt3Yj3p4A/tb/z957gFeVZNfCk2ds//bzex5/DuNn+4099szY47FnPMkznTM0TTfd5CbnrCyUEwgFUEARZRRBQoCQQDmQM93knFMThXK8Wv9edTkggQAJJJTqfN/R1T2hqs4+dc9ZtWvtta+0gDr+XPiXxzA5V57UwYBbgnvq7p8Vyc5yGTSwLU1NLbh5qwX7JfEWk3wxERjPMfZzAHJBPP3nZa2SQUTLwyaoejTgV2boN3+amhqxY8cOuLkFitc7R5K6VSpqWclmE7iWbgGKZCAZFrEdCx3jELFilySaa5T+BemnEDWpS3DzWAVPrzXYkHNHbdssM9bGfh5TJseWbuFsV4var7Zxe4f2mc/lsW3KvP99c6tyeIy5HvOxrY/fIm3i/tTVZzB5qgfsFsZg3YYb2LqjbbnGOSXSVsMGxid/ezmb7sHDcw3sFy4R6tM++Q3rqfl+82PoYxeiAX/fuGEmSbRVm7sRFYvcUeGzGPW7d4L6/HrpWxbQgL9v3a9ntrZZPJ01q9NQuTwItTnZMN2WBBldtBA4E3BzfdJiHPMoyG59vFHGE4/pQD0sz6iL5T2xrNYV3/9fA/52jNKHN924cQNeXkvg6r4Sa7NuKoBPYGusBLv5RbWwtA7Dq699Alu7KBlQ1ijwTXAcvmInPhg8EcOGzUFK2mlszL0n8Sc3sWbtZSSnnVRrVs4tNUgwD1YrZJBwASuTj8jxp9SxhSUNapDBctMzL2Nl0lEkp55UgJz7CkskP8XGO8iScjOk3MSU40hLP4/sjXeRue4aktT3szIQrpDBSaMatJjbcEWVkyRl8XzWH5dwEB8Mmiig31OVxQHC+uwbSEo9gcTk41i7/hoKiusU2Dds0PqTx6/bcFMoPquwdGmAUHtu9uG7r5vely2gAX/fuHsttTVoOLAPtemrULt2DRpPndSUnr5x69q0UgP+Nubo+1+ULKdkxKtwcVBqPaavrvf9i+riK9CAv4sN2sPFnThxHNbWDlgWmHffs9+sQDOBMz37DCTPK6zG/AVB+N3/DIKVdThyC6ofeOpDw7fhnXdH4+NPZgmYPiTlFMLKJgILLIIwc9YSAdbuMphIFSB/SWRfb6v98xYEYKpsnzbdC86uyViVcUESxd0T2tA2zLdYjslT3NQ+B6d4pK46g3VZ1+Hjt0G88tGwto3AJNk/RQC7i2sKFjrEYeo0T0yfuVgoR6VSxy2h6uwSD3ystCFYtWHSFHeh8aQiQ9oQG/+FAP4JUoeHalPa6rPK2886eRwHNtGx+5Avgw9eu9kOtIV5JfjnbMGK6B1wcHDF6dOnevgO6uoHqgU04O8bd775zh2VyLN8/myU21igrrRY6/D3jVvXppUa8LcxR9//0nT8mJp2uztlgnj5A4XS83jirb5/lS92BRrwv5j9etPZLTK1c+jQIdjYOAtvv1SSrkF5wQ1w2xrwW1guxx9eGSKAO7IN4CfV5733P8fw4QsEBO/FDAHev/rV2xg50hJuAvQJxod9Ogdei9cicsVuTJzkqgYBXl4ZmDN3KUaNthJv+XoFsgnkPx+3UChCaZg3PwBDh06HnX0MomP2Y9wEJ/z+lY8wc7afAv1vvjVCZhw+xiz57uSSiE+Hz8MkGUTExh+Ag2MsXnllKEaPtoHHonQZPCzCkKHT4OqRirDIbfhwyGRMm7ZIyt0HG7tI+T5FzVw4uyWr4yZMckb8ysMK8BPgP7SHGfSTBhSXsFcAv7sE1R/pTbdUt2UAWUAD/r5xs1tqa9F49DDqhNZTl5+HpnPnnz7V3zcua8C1UgP+fnbLFaUnYxWqQoOF0rMBzbe1Esejt1gD/kct0ne/G4Df1tYVy0NKUVLW8mzAb7eiDeAPj9yhAP+IERaiYLNLAfb/+s9XZSYgQig4t0TZJh9DP5mJSZNdxUufgyEfTcNIAfnBUt+K6D1YHrZFUYHS0s/AzTNNee25jQOL117/BJ8IVSggsEgGDXPl+6cICC5GfOIhfCqDiN/9zweiJLROZgBuiId/icwyzJAZhHyZJQjE7373gfLyb8q/i4CgYgz+cBJGjrIUJaH18v9kmWGQGQKXJAz9eKbUMUtiFHZI0PE+dcwfZGDBmQdSmcxe/ragnx7+2Pg9cHT01IC/73b/Pt9yDfj7xi1sYeIt8epXLQ9AVXiI0HsOaA5/37h1bVqpAX8bc/T9L6Tw1OdvQm1aIuq3lMJUXt73L6qLr0AD/i42aA8W92zALxmcFaWnBvTw//73H8Ja6DoEwuTvc6WH/913xygPf3jkTkyY6II33xwO36W5EoQOJKw8ojztw0csQGTUbqH7hItX3xqWNmHwX5Yrij8HhId/RWg95+C3dKMEBscLUI/D+IlO+M1v35fBxDj4L83D8JEW+GzEfEX/yRSe/eTJ7hgkXPxIUdrZIu2wlgEGgbuPXzYsrEJk3wQEBZdg914Ix/+kAvIffDAezjIbMHjwJKH0uKsZhjffGomPh82GhwQdey5aI7MBi6Wc2UJDSpOBTZUG/D3YP3XVT7eABvxPt09v2UuVntr1mbhnb417zg4i0blFJ97qLTenE+3QgL8TxuoLh5pEd7++IE8Cd1OUbq4G/I/fNQ34H7dJX93SUcBfUFwvNJl4/FYAOCk55Ntv3w3FZSdI/h/h9o+b4IyYuC9kv4tQbYaKvn2W5I+A2jZ8uIUAbmuRwBRpy5AyuIuqD739js4JmDLNQ9FqPLxWq1mAeRIr4Ou3EdNneOMXv3gD7703DssCChTg/1QAf1r6OaSLMtDEic6Kix8hNCFSkSwsQzF02EzR2BfAbxmiBiH+MujYKYA/IfGImhH4UDz7bh4paqAwabIbFsgg5u23RwmNZ7q0ZxOWBRWJ5OZKRRkKF4+/2cPf1rtPeo/28PfVHt+/2q0Bf9+4n83iOKzLXo8KV0dJvuWG+u0iF9ak1b36xt172EoN+B/aol/813z9Omo35aA6KUFNwZnuip6mXtpYQAP+Nubo0186BvjNij2kvDAwl3z9xeTcx+1HSPhWjB5rgzffGg7PxWuwSsD4BPHM//znvxduva+o3hyDqwTLkjYze85SBAodZ6EE03rLYCBZVHE8PVfj08/mKG+7tXj+P5HybYTKkyD8eVJ6fv3rd/CaKANxgMA4gBGjLB4A/kmT3DBIAHyEzBoQ8FtahaljOEvAYN3f/u59iQMIVEG/7h6r5dhJmDZjEZYG5GHIkKkqKNjbJwtjx9lj2Gdzhdu/Q3j7RwTwx8sAJEIF9xLct8fh7yjgp32rq6vVygRnetEW6EoLaMDfldbsvrJIDa5JTcK9udNQbjEL9UX5aNE5PLrP4N1Usgb83WTYniqWclmVS5fg3vyZknE3As1Xr/RUU3ptvRrw99pb0+mGdQTwE/SWbJacDaLPvyyoEKPH2AglZrKA89kYLMGuH4l33NF5peSPuCtyldcV4P/3f/+t4swT9I8Szz5nBcJEgYcBtQyyHfu5g9BpAjBJtk8Sak6AKPtQ7WeigHgG2s6dt0yCfb1koDBDFIDGCCc/COMnOGK+ZTDWrLsqnP2vYCvBvFOFfpOQdETa1yJ0nEx1Xmj4FgH/oUIHelfFC3CgwdmFz8c7qjYkiuQnZw8cJKcAZws40zBc4g8myIwB4wBGCHWIgbyrMs63S+d51MN/7NjRx+xeU1ODSxLwz/wGiYmJyM3NxW0dD/SYnfSGF7NAQ0MD4uLi8Ic//AHBwcEi+fwUzecXq0qf/QIWaKmqRsOuHahJjEd1SqIE8Eqgv9bhfwGL9sypGvD3jN27rVbKZ9Xv2Kai6Rv274OpqrLb6uqrBT8R8DNB1yNJurrlGqUOAtXuWFguPbH0nDUPgAdyRwG/2dPdjI35FYgXWg49/E4S8OoqnnNKaa7f8BW27oTo+F9TVJs/iELO3HkB8PHNNs8GxB4Qjfx7ao1beUhA9ka4e2ZgieyPFhpQ9qa7sparoNlFEoTL7cyCS7qOt88GURDarPj/SamnVMAw9fqZQIt6/TyPXvjVay6KN/+sgPizEjAcpig9FgL8fYWqw/ZGidQmdfo35lbIIOGoOja/qE70+W+rXAKLlqwVJaFM4f0XI231OUmSV/uYOg/t8DTAX1VZhS+/+BKxsbGYP3++0JHew09/+lOJL7DG+fPnu6PL6jIHsAWY6dkA/EFBQRrw99K+YJKg3XoJ2q0OC0JVZCgavtivg3Z76b16WrM04H+adfrgvqazZ1AdEYIKZzvR4U9A8/VrffAqurfJrQE/Xzhc+Hnw0EHs2b1HMvZ2T6Azwemli5ewbds2nDl9plsusq6uDjt37lRe2WPHjnVLHb2p0M4Afkp0MgkXwXVuQaVk1b2NHAHxTIzF7ZSqpPedlJ633xmlQHuOgHGCbCbO4nmGrj3BPxNjEayb95kzSDNWgOCfSbL4P8vemFepZheYDMsA3E/6ZBAxk2gR8H/44RRR7ClQ5+cIyDfqaX0u28TZC+5jvcwTYCjzcF/rY1v/z2ulSo+9vat48bfjzJkzWLNmDexs7aTeD/HjH/8Yf/7nf45vfvOb+NrXvobPPvsMW7ZswbVr13D58mXl/ecMgF61DZ63D7Avsd/5+fkJ9e3XCA7SHv7e9Gxt3RYTdfjTklA+dzruWcxBXVGBpvS0NlAf+V8D/j5yozrazOaLF2TaLQ5Vvl6ozUxHs2Qh1UtbC7QH+KuqqhATGwN/P3+cPXO27QnyrVZ0iC9evAiC6OsSJ8GpaNIeuHIaml51lsHj+J0DCO7jccZCj3tJcQk8PDyQk5PzmJef4JVlHz16FBcuXFBl0VNfKd4VoxyWzTpYNsu7Iw/iEydO4Ny5c8qrz2OTkpIwffp0FBUVGVX328/OAX4j8VSzZONtuZ94q0WB/SIBx2VbWxRQ91qUoQB3vHjyy7ZAHUdQbQBmgn7jfH623seBA7eZt5tU2a2/GwMGlvVwAGEG5vxOwM/BRIBQj6jfz8y53FYq7WhdD8t5COjNAxlzPXLcFvOgxmhve5/bZDYjJm6PUJJmwtXVRfWXn/3sZ/jud7+rAD5Bfuv1X/7lX4TaNApz5szBzJkzMWPGDL1qG7xQH5g9e7ZIy04VRaw3RT3r90hLTXvsmdhvH1x97MJaJI6nYe8e1K5ORZ3IfjdKskNN6eljN1GaqwF/37tnT22xUukpzJUfZrKo9JRpWc52rNUe4L937x78/P1gb2eP40fkYdZqoce/sLAQK1asQGRkJBITVioven5+PgoKCnDr5i1FdyDQLi0tRUVFBY4cOYLi4mK13SjK1GzC+nXrwRddWlpam+lrgvcvDnwBti00NFRRKsibPiB6x5mZmfjyyy9VMTdv3kRZWZla9+/fr8oJDw9HRIRITUp7Tp48qabIJ0+erNpm1N1fP58H8LcHgA0ATi/8OqH3MKPtRlHyKZZZAc4MGOcQZBsA/iHgNg8kjGOe55NlcbDAld76DZJtN2PtFTW78Dy+StbWAAAgAElEQVTlPescA/BPnjwbXl6eir5D4PWP//iP+M53vtMG7BP4/+QnP8GECRNgZWWFBQsWqONJ+dGrtsGL9IF58+bBwsJC8fePH2/73O2vz6y+eF0t8k6rLy1CjVB6yCAgXVir9PS9O6kBf9+7Z09tcfPZU6gODUDFQkvUrIyB6drVpx4/EHe2B/gJ0gMCA+Dk6IQTx048MAs96tu3b5cERy7w8fHB6tWr4eXpicWLFsHb2xuLFy/Gzh07sWnjJpFK/EDAkxfOnzuPhPgESai0WAF2lkFg2tTYhOwN2QowpaenPwD83Hf50mV4L/aGk5OTouOwLoIrcqktFlggLjYOHJTs3rNb1REQEIBFUpeNjQ1WRK6Ar4+vqMcsRHJyshqUaA//84Nws5e+fS85wTiDf7mSsvMsYN2R/QT7huefFCCeUywzCmxHVw4qWrfFoPQsXOiO/fLyvibPiV27dqm+Q+89gyh/8IMf4I/+6I8U+B8xYgR2796tZpQ46NSrtkFX9IEbMgPNcjg7OhBijh68WPrYP1T7q12bIbjCChVOdqgvKyEPto9dhW6uBvz9rA80XziH2thIVHq5iJc/BUzEpZe2FugM4Cd9Jj4+HnZ2dsp7T28/wfqsWbOUJ97f3x8xMTEIWR6Ct956UwYMjmoA4O7mLgMDL/X/nj17lIeelKC1a9eKxrqFKoMDAS4cCHAGYezYsXBzc8OO7TuUx57UCc4qkALEAQABGeu2tbVV9Y0cORKWlpbYmLMRyUnJKrDSUwYjgYGBmtIjoLk1wO38/w+9+uZzzd/JkQ8QrXsG4ZKnT3lLevwNig1pNwZ4N+rkd/P2R8tkG83bVmVcQEjYNjAgmAG93N5dYJ/tai3L2VqlhzEgpJVRnYf9iDNFr732mtB+XBVfv+0vSX/TFtAWGBAWkOdC08njAvSLUL+5FE2XLkA8VgPi0vvTRWrA35/uplxLy+2baBQqT132OjTu2YmWyop+doUvfjlPA/zOTs44deLUg0rIlQ8JCYGDg4NSLyFXn1SbadOmISMjA9HR0cpjb2VppQYBi8TzHxUVhXlz5yE0JFSBcc4CBAUGqQHDqlWrhB9upQIkjUoaGxqRvjpdpCIHK04raUOcPeBKyhAHCQRc9OpzFsDd3V1ReQYNGoQxY8Zg6dKlWLZsmRoUsD2Ut2P7NIf/RUH/w/MNQE9lHWbQpSxmYsoJAeYNEiT7MBjX8NAbQN4A/dz+cJ+5XLNnv0lJdFKD//XXPxPJz6XIyr71YABhnN/Vn60B/1FK7D2ycNaJg10GVnLAevjwYeWFfeQw/VVbQFtgAFiACTzrsjIl8ZY9Kt2d0LClTGfa7YP3XQP+PnjTntbk5hNHUbXYHfemjxcJrUA0X7n0tMMH5L4nAX6CZnt7exw6eEgF3XKKmVQfAnjSZXbt2KUSEK1bt07x8MnRJ+j/6KOPFL+Znn56/Am2SYsgWP/qq6/wxRdf4OiRo7hy+YoC7+S8pqSkPAjwZWKjDVkblIef0nSnTp1SMwJ79+5V3lYG5dLzT48+gyYTEhLU4IFgn1QjUi3If6Wc4v59+xG1IgpTpkxRswb9/QYTmB46dEhmPVyxPKQUJWX0trfnSX8I3p8HPBuAP2XVaYwaYy0DLVu4uqeC2XV9l+aIDOZZUci5qTLiMnkXYwFIyckrrEZy2gnZflj2E8g/bBv354s3f1lAviQDG4sf/fN/YNw4RxkAXO9xwP9ovzFoaY9u19+1BbQF+r8FWijLuaUU1VFhqImJROMXB3TQbh+87Rrw98Gb9rQmm25cR/39oN0G8fS3lOtMu4/aqz3AT3UbetAJqulBp5ed/HkGwjJolh52euk3bNgAevFJnaFiD0H9O++8o6gPDKKNjIjEm2+8qQYI5y+cb1M1QRPVeUaOGIkpk6YgIlyyoUodLJODB84icDaAZWaKRGKSJDyiBCIHHQzkJbWC3HzWc1400dkOcvjZPpbB2YPS4lJER0Vj6pSpGvC/MK3n4SDBDPibRTv/rCTussWrr3wkCa4WKAnPTz6dhSlTvRAkA475kiHX1i4K6aKpT548j585a4lk3Y1Q+vqk/5jpOmbqT+qqMzJbsFjiP8YLb36ISty1thcC/jYdWX/RFtAWGFAWMMlsX8Phg6jbtAH1eRtB+W9N6el7XUAD/r53z57aYtOVy6hbuxo1kctRlyfSjzo75mP2ag34DblLyl9u2bxF0WHoZSfAJn+e3Hoq3/CTSjhqIBATqwJ5GWh29uxZResh6CYw37d3n9KTLikpUTKZj1bOQQJVdZYsXoKw0DBVXlZWlipn69atagDAelj3WimTAW2caaBaT1hYmAL3hkwngT8HDEa7qKPO2YldO4XrLxQhtru/Ly/Xwy+AXxJjjRhpKeB8MKxswlV23SlTPfA/vx+MBZYhknXXVTL0TlbZb0u3NMLHP0cGap/AwipEqf+UbYUC/PzMzi2Hi2uyJPpykbiOEIwabYVJU9wl+Vfv8/D3936kr09bQFvgyRYwVdwTPLFR5L4XoWqpDxp2C11Y3kt66VsW0IC/b92vZ7a26dQJVC3zkeQYM1ETHY7mq1eeec5AO6A9wE/g2FDXgLuiRsCgRXKX+ckgXQJuBjMS3B88eBBXr15Vuvu0G/eR58/9XKi/T0BuJPRSG1v94XaWSaqPUcdtGZQxNoAzACybEpyk9XBAYSyP1sPtbPOtW7cUv5qUHqr4sAwjB8CT2mCU2R8+Xy7gbxLAfxojR1lhxIgFir6zcw+EkpOHV179CBMnu8PRKUEl7Zoy3Quk/8yZ6y/B3COU979k80PpTfL+Q8O3YuzndnB1S5aMvDvFu+8qM0XukvX3Rrcq9JDS9CwOf3/oG/oatAW0BbrGAi3yLmqQBH3M8cOEnk1HDmvA3zWmfamlaMD/Us3d/ZU13/gKdYV5qF2VogJrTJrS85jR2wP8jx2kN/QJC5gB/0GhNrmIck6ZAGW04ck/D1+/vXNac/hJ6Zk4yU24+aexXRJYBYeU4I03P1PbQsO3Y8QoSwweMhl2C6MxfPh85cFPSz+vEmiRw0/ufnzSEYwf74i33h6BhQ4x8FqUjg8GTZR4kOmIjNqjZD9b8/3ba9OLbCPdKC5hr9DIPCTR2+NBu33i5utGagtoC7wUC7QIpadRQD7pPPX5m9B87qym9LwUy3dtJRrwd609e7w0k3C+a9NTUR0SgPrsLJhu3+rxNvW2BmjA39vuyIu0p0ViKY7C2sYRgcEFAvZNKuD1RcBwe+caHH4VtDvaUrz8lkpCk8DZz3+TePiHYMZMH0nYdRkubikS1zESr78xTOg9k7DEN1sBeNJ4CPap5U9pz3ffGysZRj/E0KHT8aEc9+Mf/xI///krQhWKUOUw+y/rba89L7KNAwl6+KNidgrgdxPq18O8Ey9yJ/S52gLaAv3TAibJ6F6bnoYKq7mosF2A+uICrdLTB2+1Bvx98KY9rclNZ86IOk8wKhxtUJsUJzr81552+IDc11HAT5oNuf2k61CikJ/8zoy5euk9FmDSKHc3D3h6pYoSzl0QWHe1d9wA/JTlZLDuq699DK/FmViz7qri3//udx8I0E8Wic5aUKXns8/mSdbaf8Unn85BUupJAfpGIq1mJc+ZJsG63kvWwdk1UZ03d94yvCKBwG8K/cfXf6PS+Oc1dDXgZ5kMHN6wsVzqWSdSrr5CL9O5OnpPb9Yt0RbohRaorUPj8WOoKypAfUkhmilIYdIc/l54p57aJA34n2qevrfTJHzwhm1bUL9xAxr37dE6/O3cwo4AfgbgMoNuamoq4iXxFqUwuTKj7pUrVxR/vp2in7iJgwdy98n3Jw1FL11ngfr6OmRnZ8PRcRmWh20Vj/hDcG0A9Rf9pGeea3rmJcwQ1Z133x2DcULJmTnbT3noR4ywFE7/UUXbIbifNz8Q//7vv8XseQECrm/fH4SYZx8IugslS29uQRU25t1Ta2zCQUybvlh0+P2Qsfay8uqzvhdt96Pns8yC4gb4LyuCrb2fCkZvaKjvupuhS9IW0BbodxZQspybRZYzMkSkOcPR8OUBzeHvg3dZA/4+eNOe1uSmc2dQE7Eclc52qElZiWYJPNVLWwt0BPCfkZkSSmRaWFioBFhMZkXZTkpfXrhwQYF2gncjLXzrGhg4y0Bc7jNUgC5euIjVq1Zj27ZtKqiXx7cG/vzf+M7zWXalDDqM7RwsMECXMwzGwkBeBhkzPT2PH8jLnTu3kbgyDfYLgySItlBA81UB0lVCpSGofvF1E8uSdW3WDYRF7IC3T47o8EsSNdtI2C6MQeDyMjW7UChgelN+JeYvCMIbbwzHIu8NWC+zDnkFtffbUak+c/OrRYO/ToFvAvB12bcRvmIPIqP3InvTPSmjWtW3Mc98/Iteg2GH9DVX4Lc0V2RC/REXnyz99M5A7jb62rUFtAU6YAGTCELUbcpBpbcHKv0Wo36XcBnFiaWXvmUBDfj71v16ZmubLl6QSPp4VPp7o3bdGjTfvPHMcwbaAR0B/EzmxARZPj4+YAKsc+fOgYMAqugQeO/etRtMtLXUf6lKzLVv3z6ltEPdfCbjWh6yHAGBAWCSLibOylqfhRHDR2Du3LlqG7Pg8hzShC5duqR09FkntyWsXKmkO6mtf/r0aWzatElJcjIxGAcc3EaVH+6nfCi3sy08dyAo8zypv9L2q1evhYdHMJycI+DmHg93j3i4dcUqZZnLS4CH50p4LkqCk0sUbOyC4OAUDnfPBKHmxIk0ZyDGTXDEm28OwzvvDsMCC/8H57Xbjvvlcp+7lOsh5ag2t9re7nmdvCajTCfncAkQDsH69TmqP9NmHDByBkov2gLaAtoC7VmgRRxKDQf2ozYzHbXrM9EoaoAiCdfeoXpbL7aABvy9+OY8T9Oa6QXeXIraDetQL1q51M/VS1sLdATwHz58GJaWlli+fLmSyKQnnSulL5k519XFFbY2tvBe7K2y6jJZF6U0d+3apbLf8ruNtY3KeEud/NSUVAnM/FAl6AoWnX9HR0eln88yt2/fjqlTpyI5ORkhISEYMmQIZs6ciURJvMVtzK7LpFzMqjtr1iwF8qndP3bsWPXd19dXzUIwFwCz9g7kpbKyQuy5TQZAsZIPYbkkSwtWeRGCg7rwM3i55GtYjuXBIQhZHip9JFR99/Xzx9x58zDs02HqfpIfz2NUOzpS//1yu7St9+tl0jiWGx8Xhz17dstMEeNR6tTgkRS1gTxQHMi/F33t2gIdsYCi9BQXomr5MlRJjGDD/r06025HDNfLjtGAv5fdkBdtTtNp0W8PWYYKewtUJ8TAJAGNemlrgY4A/iNHjmCegDcDYBN4k8+/fv16lWiLYJsAqlgy5NrZ2SnwzX2G55+Jr5hJ95OPP1HZcJnUy8rSSgF4evfd3NwUcDcA/5QpU5CUlKRoQ5MmTcLatWtBbX1m+GX2XybiIqDnDMHs2bOxInIFxo0bp8phXAETfXEgYuQDaHvFA+8bPdbV1VWKFkVqVNeulVLe4+tdUbI4LYO+vfv2qqRnHBzWimesRgZhzM3w7DYYZXZ1e83lVYuWdnPzQ08+6WK0E6lhBp1s4PUUfcXaAtoCz7KASRxTZAxQDKTC1QH1W8sGpEoPn5Ok3HbmeWmcw/N6etGAv6fvQBfX3ywAn1Nu1TGRqCvIA+W09NLWAh0B/EePHlWAf/LkyViyRLLiSpZbeuoZuOvv74+PP/4YC+YvgL94dUn94WwAM90yeNTZ2Vl54zkD8Mbrb6j9zKJra2urQDv/9/DwULQdgsIdO3Y88PAzy6+NjY2i55w/f1559999911VPmMKWJe3tzc2l21WWXZZJgclLI+gn2pCeukdFuADngGxTZJsTS/aAtoC2gJ91QKk9DQe+hJ1G9ajTgRBms6e6XFKD+PjSEckxZWOltOnToPvTCORZVfYms9wUm5Z9h3BUqyPjjg66p4F+o39TKDJNpI+acT0dUXbnqcMDfifx2q9+ByTcPYbivJRl5GGhu1bYLpX3otb2zNN6wjgJ6WHAbsE4CdPnlQPET5I+ONPSUlRNB6C+2tXr6lMtwb/nvQaAn6C+OKiYoweNVp55cvKymBtbS0c89XYsmUL3MTDz7LJxS8oKFAzBGwXt3HGgN76SxcvYeHChcqjX1hYqB4YrIezD3x48MFz7NgxNRvAWQGCfz7w9NI7LMDg7pycHEXZGuhB1b3jjuhWaAtoCzyPBZrl3cfYwPKp41E+ayrq8nN7PGiX78Dw8HDw3cdZb87I01kWGBio3r9dQW9lGYyP8/LyUoIb27dtR2RkJPgeNgB9e/ZkrB9BPs+/ePEi+G7Pzc1VSn3tHf+ytmnA/7Is/ZLqaTp1EtWBfrgnCTKqY1eg+eqVl1Rz36mmI4D/4MGDypseGxv7QFXHuEKCeSsrK4SFhimvOgcA5NQTpBteeHr7ORvwzjvvKEpOWlqa8sQvWrRIBe06OTmp8tPT08FtPC5e5D+pBGQp1J89e/ag/G652kY+Px86+fn5ila0MWcjtm7ZqmYcSCMi/YcPPdJ9GFisl95hAc64kKrl6empPEO9o1W6FdoC2gLaAp2zgElio+ok2VZVoL9QhgPRILRF4QJ2rpAuPppOOb7zPvjgAzXDTWcZ36Wff/75g1l1Olo4i06hDTrDqJzHhWp3/M4yOHBoLVpAiuN1UTekM40z/Xynsx5ScU+eOKmkjAniudBjz//phLt27Zoqh4p6xBh+fn7YvXu3kvHevHmzwgeG44fH0GHH8m/euIkWU4sqj/spCc79Z8+eVfvpFGSbumLRgL8rrNiLyjBdv4a6nCzUJMRKNrxCmO5qSs+jt6cjgJ8/XoJpBuE+6ingw4OjdfLoSfWhUg7BOM+h956Bt+Tvk/fPhwW9A9miqEPQTkCfJ8dydoCeCNKEoqOj1TlU2eH5G7Ky5SFySf3ICeA5oGAdXFcmrFQeYz6oGNTL+nk+Vz6Q+LDQS++wAO/xoEGDlOeJHh+9aAtoC2gL9EULtCjAL0G7gUtRFRqkgnZbmroGhD6vPehgI8V1wYIFoIOOYJ7gmxLan332mVKv27lzp3pHciBAIQ2+I8+eOaveqY4OjrC3t4eLs4tywhFkkxLLYxg7x9l1I4aOwhmlpaWKasv3MZX37oikMZ/xHu4eSlSD1F8q6jHWbsSIEfjVr38toh6LsU0ovJniANwq73bWwcEBxUBYN8U4lngvAWm+pP4QbxA/EBvQUcQZC7adSoFdEQOgAf/z9rZeep7p8iXUrkqVaPoA1GVnwSTgVC9tLdARwM8fFwNgOYJvb+qOqiakbHz55ZdqJG4MCgzPwZHDR5SEJ0fnhpwnvQYc0V+VgQF/3IaHgQMFBnXSy8D66mrr0NzqYcpjSSviNOL1a9cfBFqSU0gvBNvAMrrKC9DWWvrb81qAD/9PPvlEUbl4f/SiLaAtoC3QFy1guiOUntQklM+eivIFM1FXmIeWxocCAD1xTV98+YV6thKwt36+cnZ9zJgx8PLwUup4gwcPxqeffqoGAJw5p2LexIkTFaDmLLy1lbWaiSW1dv/+/SpubtSoUYpeSy/922+/jQkTJijPPiW3lbe/uEjRNTnYcFjooAYVavAgSnpxsXEYM3oMXnv9dSXZXZCXJ8csVP/TUUexD9ZP5bTQkFCl3McYQM4GEJuwvvETxiMyIlI5C4d8OARBwUEo7wJ6tgb8PdFTu7HOZpkmqvT2xN0Zk5R8VrMMAPTS1gIdAfxtz9DftAU6bwEN+DtvM32GtoC2QO+zAOW96/I3qaRbVct80bBnl2Ta7VnAT6/+3DlzMfyz4QpM07NOJT0C8qlTpmJNxhpQLY+zrPSS00HGGXAXVxd89NFHagafTjTOkr/22mtKaIP8fA4QOEtAJ9yNr26oWQRSZjkgIDWXFFt69imTTcomef101nGwkCb1HxVn3yKvRbCSmD3y+I8IyGd8HWfoGdNlscBCcqF4iYJbrVo5O8864+PjVUwCKUrLApYphyNpPdOnT4ezizOYvPNFFw34X9SCvex8UnpqN2ajWgJs6kuKhNJzt5e1sOebowF/z9+DgdACDfgHwl3W16gt0P8tYBLwW1cilJ6gpagODUbjgX29gsNPMPxroc7QA8/AXYLxObPnIDkpWQlsMLZu2LBhiJP8I5x9v3D+ghoQvPLqKwqwU2SDYJy5b6h0RwouwTcHDsZsO4+hgAdpu0x8SVU8DiZmzJiB6dOm4/Chw6oDcNafA4p75fdUWfPF+0+e/iEZmND7ryi54sFncDGDjcki4DmkDrNOxiCQykPJbw4sOIggRYnHc3Bx/tz5F+5oGvC/sAl7VwFNx4+hYrE77kyboLh2TdrD/9gN0oD/MZPoDd1gAQ34u8GoukhtAW2Bl24BpdKTshLlMyehfO501Bfkig5/z8oNk8pKSs2smbNQWlaqePXk1p87d07F3TU0NigxjeHDh6scNw31DTh35pySuB49ZrTKVE/vfGlJKfJy8/DFgS+Ut58edgLuB4BfEii2BvwcVJDaM2P6DCWnzVkCLqT+8pyqyioVL8D4ArbnsOw3AH/iykTMmyuAP6wt4GdSTgX4AwIxftx4NaAwaMMsh4k+z50998L3XQP+FzZh7yqg6bQk3lIqPfPMKj3XdLDgo3dIA/5HLaK/d4cFBgrg54vOWLvDjizTJCoWjGtpur/yf8asmAPZzAoX3VW3LldbYKBboKW6Bg1796AmLRk16eJ9PnG8xz38DNqlWp6zk3MbDr9xr+ghp/ecAbxMasmYPEprMxiW+XUokmFo5BP0nzl9RslmkwJE8E1xDsbpkSLE4+nhNyg9Odk5KqCXXPzCokI1wCDwzxO+PpV83N3c1XmM2eN2g9KzceNG1WZPD0/VFgb+UuBj6NChSEwyi3DQw5+xOkN5/1m/8vBLnIIG/Mad1Z8PLNB85TJqVqWgMngZaiVJhg7afWCaB/9owG82Bac4T0nCEgYSdUfCrhuSEyI/Pw+paalYnbH6wZqxJkNJiZIHyQci29Efl4EC+JmEhuoSfIHSw9V6YTA7A8v50nue+8yBxJ7de7AyPhGRYVGIioxRa7R8xsbEyUsyCSmpyUhbnfagf7GvsY+R08sXPvsYX/560RbQFng+C5iEplKXswGVHi6oXOyBhh3bejzTLp83pPEwweXVK487Nukhz1yTicGDBqsklZS85HOA3nkG9VIhJyoqStFlSJnZt3efeoZNnTpV0YAYsEvu/2vC7x89erRS4OHAYYLQhzgjsEGU9zi7wAScSk7bwhJukl/nwIEDKjnmoA8GqaDcdSKbPVdUfhgXwFkJflI6lKo+S5cuxYRxE2BjbaOUeEj1+eSTYaIilKoAP2crSFuyt7NX6kLPd/cenqU9/A9t0S/+azp5QgXWlM+ZhurIUHAAoJe2FtCA32wPgnyqFlAvvqvlPMk9DA8Nh5f9IixzDkSwR6isIQj1CsMy10DMHD8L7737vnrgMXthf1wGCuBncBul5gjs+ZJtvbBfbd++HcXFxY/J27Y+rr3/q6tqpG8Ww95FgtxC5iNyjSsiM90QtdYd0evd4R1ugc+Gf4I5k+ZLnwrAck9zHwvzCkeAWxCmj5uB1155HYEyTc426kVbQFvg+SxgEi372k05qFjkjkqfRWjYtaPHAT9V7+h157Olvd83Z/9IqWHgLoG28WyiRDKdAf5+/iKb6a3eQcxlY6jllRSXgLx9gn0q5TCols4pqu3xOcf/z50/p2YAWD9BO2cNCPoZ2MuZgS1bt8DL00t57ynLSc8/pTWZE4C8fgboent7w3uJN1asWIG9e/aq2QYOFqgydOTIUTV7SRlPvkf4jmZenhddNOB/UQv2svObz55Bdfhy3HO0RU1SPJoliFcvbS2gAb/ZHvR28KFJcP483te2VjV/M7WYcOr0KURFRMFhtiPWB+fg6LqTOLLuFI5nncGB9ENIWJSEcYMn4D9//F+ws7FT06btldXXtw0UwP+0+8SXrNHHHvX+P+28GqEQFMrL08HNAs7LJ6HwwjKcQLKsKThsSsCm077wjpoDy+kLsMpvLQ5lSpKc9adwQvrYF+mSo8I7BaMHjcU//PX/g72NvdLMflp9ep+2gLbAky1gEu94/f59qBY6T01mBppEJhrNpief0Af2MOEVBwQE6I8uBNdU2CHgftZChxVlsznTaSx8t3JgQUoOqUSPLnwWch8HEe3tf/T4rvquAX9XWbKXlNMsnP3adWtRFRWB2rxcNIt+rl7aWkAD/rb26MpvpAgF+EvSEKtFKIndjMpddWihiMERoGZfI0oTtsF9thfmjVqAsYM/h+9iP5y/cL4rm9BrytKA//luRV1dPQrEc+bsaY3F0XOw+WoYjgvQP4lU+UxGzglfOPpOht0cK2xZuRPVexphYtyc9LGqPXUoYx+b44WZw+fg3d++Dx8vH+HLPv5Sf77W6bO0BQaeBUwyU1eXtxGVSzxR5b8EDbt3AhJHo5e+ZQEN+PvW/XpmaxuPH0WlpyvuTvlcJd9quvzi2q3PrLSPHaABf/fcsCtXr2BF+Ao4z3HFloQdKN9ejaYv5L1wEGj8ognbknbCZqIdLD+3QZjdCiwYawWfRb5qerR7WtSzpWrA33n7U0mjuEhoPK4LsGjFbBSeC8KRZrNnn2C/4FwAPJbPhNMCewH729WAsln6l0nWmr2NKInfDPspDqqPBVuHYtKQKfCVPnbr9s3ON0afoS2gLaAsYAb8m1Dh7YVKfx/U79ndKUoPY3EYZE/Pd29Z2R5jfbRNxvaOtPlJxxrbHy2b3419xmd7x3THNg34+9kPmh7+mnWZZg9/fi5MEpWul7YW0IC/rT1e9Bsf5kwKEhIcAi/x7O9JO4CqXfUC8gn0xeu6ux6bE7Zi4VRHzB0xDyvdkrHeLwdW42zh7bVEA/4XvQH95HwG+BaJ4oWN49DW4hcAACAASURBVFy4hU3B5mthAvaTcIw0nqaVKL4YDOel02A73QbbE3cJ2K9H05fmPla7r0l59u2nOsD6c1usdE3Gau9MTPlomvLw39SAv5/0En0ZPWGBlrpaNB47IqyBTagryEeT0FEg9M2OLsw4v3nzZiXWQG18vfaMDTTg72iP7SPHNV+6pFJgV0qCjLqsdaLSoz1bj946DfgftciLfSd/MSQoFK4L3FAYVYK6feI5uQ/EqvbUoyx+G5xmuMJqrDXiXVaiYHkxcpblihfWWgP+FzN9vzm7rrZOgX3XRbZYEjtHwH2Q0HdShbOfikNNicg+4gfHpVPhZLEQpdHbUb27wQz2pZ9V7q5DafwWOEsfs5Q+FuecgPzgImT6rMekD6dowN9veom+kJ6ygOleuVL9u+eyEBUerqjfvrVTmXYZWPvxxx/jn/7pn/DP//zPeu0hG2jA31O/oG6qt0kCJisD/VFuKTr8MSvQLDQLvbS1gAb8be3xIt9uCTfaz8cfrvPdURyzGeU7hMYjIIwUi0b5LIgugeU4G8XZj3GKV2C/MKQEWf45AvitNOB/EeP3k3OprZ+fXwBbx3nwipypPPknkabAPgF/6ZXlcFCcfWuUxm1D7V6hBUjfIlWMcSGFMSWwmSxUMYJ9GVDmBhWofpYuHn4N+PtJJ9GX0aMWaJGg3YYD+1Erkrc1a9eg8dQJech3nMNPFZyf/OQn+NrXvqbXnrRBj/YiXXmXW0AF7a4XSk90JGpJ6dFBu4/ZWAP+x0zyXBsuSLBtTFSMovGUxm6V4Emz15VAjMGT25N3w3aSPeYMn4soh1jkBhciT9b85UUa8D+XxfvfSZWSlZKSc3YuAvZXzED+2WU41pKsvPtHW5KQe2opXJfNEM6+ePZVH2sEOfvNEqRbvbdBZo+EKiac/QVjLBHjHI/8kCLVv9jPNODvf/1FX1HPWKClphoN+5h4K0USb61C48njwujpOKWHmvW//OUvNdjvSbDPunum++hau8sC9OjXZq5GVUSI0s1t1uoUj5laA/7HTNKpDdQ3PilyZqHBoXCY46SAff1+k/K4EoxV7Ko1Uyymuyiva4JrIopCJX25gDB6X0m30B7+Tpm8Xx5cLdKbeVTj8bKWAN0Z2Hw9VJR4zJ79Q8LZ33DEB24BM2A91RplsTtAnj77F73793bVSB/bCoepzkIVsxGqWCIKQopVH2M/04C/X3YZfVE9ZAFT+V3x7Kej3HYBykXyu35LWaeCdjXg7yUzGz3Uf3S13WSBJtHhrwoNxr2FNqheGad1+Nuxswb87RilE5uoHRwo0pvuFp4oiinDvZ01yuPaIl5X8qlJ46HX1UrUeOIF7BsAjGBfA/5OGLofH0rpzdxNuXBwt4Rv/HxsvhKqPPuk8HDNPu4LB58pcBHO/g6ZKaq6z9mnZ58DyoLoYthNXgjrcXaIdRJd/sD8BwNKo79pD38/7kD60l6qBUwSUF8vybaqV8ajOiVJAniPyjRbxyk9GvBrwP9SO+xAqcwkiSIaJKNbfUkRGiSdvamqaqBceoevUwP+DpuqzYFU47lz5w5iV8TCcZazyCLuELBfbVbjEa9rg0hvUirRcpw1Zg+fhxinODPAFwqPAfY14G9j0gH5hcm4SiSIz3rhHLgET0HRhSAcNSUprf2jQucpvRoCz9DZsJlpJepO25Xik1LjkT5G2ljpfcUnCwaBy4CSfYog3+hjGvAPyG6lL7obLUAcUb+5DFXhoYou3HBIpto6weHXgF8D/m7sngO36KYzp0V/PxD37CxRnRALcvr10tYCGvC3tUdHvhHsM/NgUEAwvKwXY0fSbuXNV0BMpDfpdS1L2AKn6VTjsUGCa5LyupK+YwAx41NTejpi8f55DDn7TD/v5EnPPpNqheCIgH1Kbx5qXImNJ/xgu2gS7GfaYevKXdKv6lQuB8q7EuwXx22GwxRH2Iy3Q4KbcPzvU8SMvsVPDfj7Z9/RV9VzFjBJFtmazHTcs7XAPaH01G0u0ZSenubjP0/9PdeFdM3dYYHmSxdRk5QgyTGWoHZtBppvalnOR+2sAf+jFnn2d9J4lgcuh/M8VxRFlz7gU1OJh9KbBSLHaTdloaLxEOxTepPBua2BmPH/QAH8eXl5GDZsGGxsbMD06wN9qa2tRW5uLpw8JOFa3ByUXg7Gqfuc/cMC+nOO+cHZfwqcFzhgc8xO1EgGXUPetXJ3LfKjJCEXqWLi2Y91jm8X7GvAP9B7mb7+7rBAi6L0SFZrofTUpJLSI2mtNaWn7wUhd0fn0GX2nAVMkuCiLnej8OwSUVdWCgbb6KWtBVJTUjF8+HDEx8e33aG/tWuBW7duITgwGG4WHgL2yyThUa0K0DUJn7pufzOKYstgNd4ac0bOEz61KKWIV781xcIA+sZnfwb8DGhm9sSGhgZkZ2cr7WlLS0tcvHgRKtvkAE1HXy+c/XwJ0LW0mw3P8BmKxnNcvPonFWc/BYUXAmC/ZCIWzrWRLM07UX9AgsBlMEnAzwy6pPHYiOLTgtFWKkDXAPZGn2r9qT387f6M9UZtgee2gNLhz1mPCncnVCzyQP2ObRI939Th8jSlR1N6OtxZ9IEdt0DTubOoigzDPVcH1AjoN311veMn9+MjyRu+LVmHr1y+gqCgIAwaNAh+fn44d+4cbt+6rQBaP77857o0gtfTQhGLjYoT6c3FEqC7WYIn68xKKULhpPTm1qSdcJjmJEDMUklvPgvsE5j1Z8BPoE/PvqubK0aOHIkf//jH+M1vfoN58+apAeapk6ee61705ZMqK6tRXFwE18VmNZ68MyK9iWQVnHu4WZJqHfOBa6BIb86zlwGl9DHJoKukN6WPMQi8OHYznKe7yeyRrXj2E8A8DuxDrUF+6/814O/LvUW3vTdaoFkch7Xr1+Cegw0qXBxQJyo9GvD3EhDfGWpPb+xcuk3Pb4Fm0UavTohBhben0ss13fjq+QvrR2cy2DQlJQWOjo4K7BOIvffee7C3t0daahquX9cDo9a3uwUtOC5aywFLA+A42wXbhE9NigU19qmUUr6jCvkrirBwqiNsx9tjpVvyA1nE1uCrvf/7M+DnwNLb2xt/+qd/+th0LweZO3bsaG3mfv9/VUUVcnJyRHrTCv6J87DjVoQC+vTsU2d//WFvOPpMheMcUeNJ2vsgqRY9++USEF4oik9KjWesrfLs5wlN7GmzR+xvGvD3+26lL/AlW8BULUG72yTXiuT3qY6LRsNhHbTbJ5OIveR+o6vrZguYBLjWbsxGlXDt6kqKYLp7p5tr7BvFfyVUJ1Ir/vZv/xbf/e538e1vfxvf+c538Fd/9VeKY03KhV4eWuC8DBzDQsJEetNDlFK2qaBcUiyYQVdx9qOLYDHOCvMl4RE5+wTxT/O6tgb+/Rnw04LFRcV45Q+vtAH8HABwcHnlysDJfF1f34BNOZtg6zQX3jHk7Ic8kN6khz/nuB+chLPvZGGP3av2q6BcxoTQu88+pjLoTrRTAbrx9zPoPgvsa8D/8Des/9MW6CoLmMRhVpOWjPJ501FuOQd1RQUStNvY4eI1paeXzAZ0+I7pA/uEBZqOH0OFlzvuTp0gevxBaLp8qU+0u7sbWS1BR/7+/vjBD37QBoh9//vfx7Jly1AucqZ6geKZM8A0NioWLhKgWyw6++U7RHpTgBhXcvYpx6mym462UBQLaqA/KUC3NdA3/u/vgP/evXuws7PD17/+9Qd9jTNKDBZn4OpAWOrq6lQGXQc3C7iHTRPOfjCONJulN6nKQ1qPk99UJb1ZGrfloeKT9LEqUeNhH3Oe4WZO3CZqPHlB5gzNRh962qf28A+EHqav8WVawCTPtNr1a3HPyRb33JzE279Fq/R0hkrTW459mZ1G19X9FjCJB7FmTbrSy63btBHNwlvXixnIFov292uvvfYAhHFK7r//+7+VTCD56gN9MTLoksZD6c0tooHOpFqURCTYJ5+aAbqO05xhO85eefYJvDrq2TdAWn8H/OxHiYmJ+Lu/+zvV1771rW9hyJAh2LZtmxpQ9fd+ViWa3fn5EsfgbYNlyQtEV385jgh9h0G6h5sSkSUZdBcumQJ3SxdsT9yDip215j6mpDfrkSdUMduJ9qqPxd337FP1yeg/z/rUgL+/9zB9fS/bAg90+CNClA5/40F5IYg4QUcX7eHXHv6O9hV9XCcsoIJ2V4Sh4n7QbrMO2n1gvcuXLmPu3Ln4kz/5EwXEvve972HUqFE4ceLEg2MG8j+kNYUGhcB1vptS46E6CukVBPsM1s2PKlJqPBZC4yFnn8C9M559A6gNBMBPcP/OO++ofsb+Zm1tjQsXLvT77lVTXaM4+3ZO87EkdjbKBOyfVNKbaSpQN/fUUtj7TJIMuo7YkbhXZoxaVFwI6WLmAWWpgH07zB9lgTgJ0GWf6eyAUgP+ft/N9AW+ZAs0CzW4Ji0F5XOE0mMhlJ7iAh2021u89p1px0vuN7q6brZAkwRaVvosQvnsqaiS0XjzlcvdXGPfKZ4KKlFRUfjRj36kgBj5+56enqDs5EBfrly9gojQcPG6ivRmTKkA/PoH0pu1+5tQKgmP7EQWce7I+YiWDLoFIcXPDJ40AP6jnwMB8J89exbz580HB5V//dd/jZiYGJDm0p8X0pWKCgth7zIfHhEzUHAuQHn1GaCrOPvHfGHvPQl2c61RFr8ddfvM0psMBK/e2yCxIltVEDh19snZ52Cys2CffU0D/v7cy/S19YQFTBWVwtsvROUyf1RKYs+G/Xu1h78zQLu3HNsTnUfX2X0WaBYvbXXySlQu9UXtukyYdOKtNsbevWs3Bg8arPjVpPNwqpHKKgN1IY2HSbXio+PhY+8rQEyUGMSzj6NikcP0utaiNJ4ZdF1gM05kEV0SFAgjqHoUyHf0+0AA/BUVFQgJCcEPf/hDJcvZ39V5qkR6s7SsGEuCnBC0ygpbvwrDWazGOaQL6JcAXUmq5RkwF55WbqL4tBv1+4VCd0z6mOTvqdhVI4nbikV60wXW4+wU2Kf05vP2MQ34B+rTTF93d1nAVCWAf3MZqiNCFaWn4aDw7zSlRzkO+5RaT3d1EF1uz1ig+ewZ5dkvd7JHjQB/Telpex8oz8nMp9//y+9j4sSJCuy2PWLgfGsxteDEyRPwXeKLBZMskBtegKslN3B7WwVubS3HleKvsCkiD1YTbATs2yGBwZMdkEV8FvAfCICfSbYoR/nhhx9iwYIFuHbtWr/tWNVC49mwfgNmzZsK2yXjkLl/MTZfDUXZ5VD5DEPaDgm+9RiLeZPmCC2sBFeKvsL1slu4JuulomvIjSzAgnGWKkCXnH0DsD+rHz1pv3F+uncmJn04BT5ePrh5W2cc77cdUF9Yt1ug+c5tSea5EndnTkb53BmoK8gTSk/HHWWaw685/N3eSQdiBU0njooGvxvuTtMqPe3f/xasWBGJt99+SyXg6u80i/ZtYN5aU1ODeMk2PGzwpyoQd5VfBjaEbMT64GysX56NCJcozBk9X2g884Sz3znpzSeBMW4fCICfFv7yyy/h6+uLTZs2gbbur8vBLw9i3vxZeH3wLzHHYxjcoybDM3YqPGOmwiVsIsbMehfvy+9t4RQnJCxKAftZis8qpPquRtDCEMwZMx8LxliZB5Qyc/Q8NJ7W/U0D/v7a0/R19ZQFTKJyV7drO6oSYoVBkIDGozI1Z9JBu33Ku09aUU91IF1v91iAmXXrcrNRnRSPepliN0mGvN640ANaX1//gE7T3NyIKpk2LJf2dtd67145KisrsW7tWtiLbOJa+WT2XdIvuqtOlst6yW/ubUpAt8VrExYYjokfTkKIbRhWeiQh3m0l4t1XIsY5Do6TnTFt2EyE2UVgY0AeOqOU0hqAPfr/QAH8d+/exaFDh1SMCPt7f11Kiksw334yFvgORWjuHEQUzENY3lxEFs6HZ8I4fD51CKYPm4Vo53ikLl6NZK80ta70SIbLVHdM+Xg6wu1XoLPyro/2K+N7TwL+pqamB8803u/m5iY12OOzrT+t1ZKIqba2Rl0fr7MnlmahlHDlwt8XY7TYrsftzG2d2d5b75VxHe1dS/ttNu5TUye88e3dS5O8I2vFq1/hvxiVgX6o37NbZDk14NeAv73eore9NAtQh7/S2wPl080e/uZeqsNPz/rRo0dx5vQZXLp0CYWFBQgMCoePT4h4RcPg69c9q//SCEmA5Inp063g4LBY1eXnH95t9S3xDYWfv4Dplck4cGC/vPyrX1pfeFZFN2/dRHpcBqId4nAq4ywuZl3BuXWXcEE+T2WcQ4rXajhNdkWyZ5ri6xNIGaDqRT57EvATGDTKy4+DTQOE19fX4Y4Mfm7evIFbEvPSVesNSfZ27drVLi+X7WN7e4umf2FhEdyXzkdUyTxsueeLrRV+2Cyf26v8kbzPBu7ecxHvnIRzmZdwNecGLmVdxZXs6zi/9hJWL86EwyRnpEgfM4D6i/QtnmuU0xOUHj7Ljh0/rgK0mb27uLhIArYTsHx5pMR0rOhXa2xMPMrKStXv5VnPmq7ez98u7cvgeNLljh07hoyMTISFRT1m69DQKHB91P5P2v7ocb3le2hotFxfdLvX8rQ2xsh9ys8XuqYIMzzvYpI8NXXZ65X6X4Wnq8q623J/sNWRMjWlR1N6OtJP9DGdtECzPACrI0NR4bKwV3P46Qm7cP4CdmzfgeioBDg6LZM1BouEd7vEd0O3rj5+2fD1z1HrEr/urWuJ1OW5KB0LHZbDw3MZNm7MFY//vU7e1e45/JYA/uzUHGzw24jazcLH3CWztDukLvms32JCUVgZvGZ6I9kjTXlf+wPgNzWblMedgcpfyWzYgQMHsHr1GqF5xSM8IhYRXbhGRsaLKlQC+BkREdelZa9YEYeU5FXYvHkLbkqitJ5ciiSz8KJgKyTttsYXCMZBhMjncon5DsX6085YFmKPLH/pY2XmPtYsfaxF+ljDVhNKQrfAc+YipHikdtmgsicBP0HoiRPHcfjwYUTKPXJ08pPffiicXWL72RqjrsvRyR/xcUk4ffq0DKBfXi4TAn4mCKQTJT1d3hlLQrBwYaDYOwrOrmZbu7jGKZs7OkfDSVbeAxcX8zYe4+Qc82B7b74/6jqkrdY2wbCyChS7R8q13L+ODvQrB8dw2Nn7qefPsWNHn+s+tQglseGLA6hdm4G6rLVoPH1KXhYdv98DAfB/vRNKPN+WY7/39a/hO1/7Ojpz3gvPKPTki0LX3fUWaJbEW7VrVqMqbDnqNmbD1IsTb3311Q15WaTKAywYSwMKkZF5FZvyq5BXWNuta36RaMoXSYIf+ez+uuqRk3sPsfGH4C5eTCenpSgqKhXvbM97+gn4s5I3YI33OpQXVaF+a7MC/vysKK5Frnj0PWYsQpKAMdIt+gPgZ6DyHflN7Ni+HcnJaZJ9ORpeixPgvWQ1Fi9Z1WdW7yVpWOydBK9FEUhKSpfB8/muf5h0sEQCfq8gSyRst8TepkDsNwVjb3OQAv2Zxx3hF2SDNUvW415htQL5HFyyj1WVyO8wuBju0z1lUNk/AD8pJl9+8QUCloUL+IyEf0ABEpNPIH3NFaxec7kfrZewUq7Lf1meXGcYwsJjVD4TOnJe1sJZ4pycXHmmLoObe6oMsPYiLf0C0jPNts5YexWR0Xvh4ia/lSVZSE49rd4xfM8kp54Sp1IOnFySERV7QJ2zZt01rFl7DZnrriNz/Vfgd6MsfvI7tz953w3zPinDuN+si2Wqcu+fmyHf2bbMdeay+H+69A3WwX1t23BdfU9KOSW5PCIxa7Y/gkO2mK9DlWG09bra1l4fS5RzlwUUy32KkFmOaBwXFkBn7xMpPQzUrVy6RCg9/qjfu1vr8AtoJ1gncP/f3/g6/o+s37u/rT1gzmO/K+v/kgzsf/PNr+Pvv/UN/OU3voFvybb2ju+WbS/rx6nreTkWaDp1ElUBvrhnOQvV0REwvcA0Xne2mHSE/PwiODgGYnnYVhSWNKJsK1Bc1tzP1iaUbDahdAuwYeNdcHZh6dJIHPzywANKSXfa+WllD0TAT3vcvn0LKSmr4egYBL9lm+RFe1kNNDflVaIvreuybiA0fDscnEJlhiJOURsMmtLT7ntX79OA/6FFb9y4jjjxeFvbBCEp5SSKSpvUc61EZsz619pifl7L9cXEfSlOmxCZzYoX+trLUUNiPz937pw8S6Pg7rEaWTl3pT0tKJV3CO3M5+2W7UBE1C58/rnkD5kfgKzsW2obt6/PvgELy+UYIQneouP2q/tUuqUFRSUN8iyoFCdNOQqK69U18r3EtaC4Dtmb7jxj3z15lzWo49mGotJGtRaU1GNj3j1ZK+T91qTWTfL/pnzzd9bN9wT38ZxNBVWqHpbF9rLeceMdJGP3VISEb1Xb2KbcgmrZd1ccVyKLK9fNbaqfSVlGfyulXeR7YtJxsVWs6p+cierMYpJZ6bqcDaj0dBZREHfU75CKOjG4648e/m/eB+//V8D7v377G2r9gYD+P3oEwBPof0fWPxeg/3cC8H8kQP/fvv1N/Lus/yD/c8DQLeD+kXaoOjpz0/Wxvd8CzRcvoCYpDpUSXMPpt96qw39VuM3uHr5w90rFug031YPKeDj2x09jILNh4x2h+MQgIz1D8ch7skcNRMBPoPDlFwfh7LwMYRHb5IVbLS/algcvYeNl3Bc+SzY3q99NxIpdsLH1kSD0daiW4MSXvWjAb7Y4+1ZRUQFcXJchOna/An4KRAoo7o/PNA5m+NvhjGl45DYZQPuJMtXBl+LIoIc6NXWVUFsisVKALIEun7FsE23N3y+3BQQV4YMPJuDzcQ7iKb+itnF7hsSTTJzogrfeHonloWUy83JRDQ4CAguFVrpOqD6JWBZYgHVZX8l9rFfnBkpZLq5JQhlKlpmNfFVGfmGNnHsBSznTIZmhXdxSECAZyNdn31SDh8TkowiSnBKBss3DKwPOkqE8JHwLIlbslBm6tfIOTMOK6D3I2VSO9RtuiB13qGO9fbOkniSZ+c7H2vVXsTbrOkaPscH7H4xDsLSXQD8u4ZC0da3YIFFmMNaKHY4o4G88ux72OblPMhgoKG6Qmeb9EicXoe6TqRMqOxKUgibJSF8vMSkUA2m6dHFAU3rolf8/AuD/+ZvfwM8E7P/sPoD/sQD4v74P+hVtR477X9/4Gv5WBgWtgT6P56oB/8t+W/Wz+kwycq/blCP8/QTUl/ZelR5yqKdOs0bg8jL1kCZ4efiAMntF+tv34jI+dOtkKnk1EhNXC5e/ZxWUBiLgp7JIfn6JvCSj1RT6Q6DQ9/qcAWrWZd0S8LBaAtCX4pIM+F/2ogG/2eIM/l6VliagMEzN5rX22Pa3Z1nr6+FvaFX6eQHJoRKkXCaOjNpu74IMul+0yEd+xwnYmFuhvNut22T8Np4F+N95dwyClhcrYP3Jp7OVB32iCBUM+2wOPvlkljhn1iBt9VmpJwkjR1uKl30hxoyxxajR1gpsp646pYD5iJEWGDPWDsNHzMeQoVPVoIEA3tE5XkD65wLWrTF+vBMGDZ6AwUMmY8JER0yZ6oGPPpqKCROcZJbkACKjduOzz+Zh8OBJmCRt+HT4XNUeB6d4AfNHMfZzOzl/ohpQhMuAYbyc99nweTKYsVftnS7xVhxoFpU2PJgpMGxiHgSYkLbqDAID07Br1140SH/t6MKg3dqsdaiQ/D4V7k5CydssKj0dp2/1ZQ8/PfRc6dGnp/5PxCP/lwLqf6TAvhm401tvrP8qoP8fZN/fyfr3shLo/7TVfh5nHiDIMVIOBwbaw9/RnqiPa2OBphOi0rPEE+WSIKM6fDmar1xus7+3fDl96rR4WCzFq7L3vnemv3rBHgJJeqDoLfJfloWk5DW4e/d2j96OgQj4KyruSnblAhl0rRNQVq4oAMZLsS9+ElTmFdQIr7cQ7u4+OC5qJS970YDfbHFKICYnpYjXNvo+deOhx7kv9q2Otll5zMV77uoWhdzcApHGrOz2LsjBlbOLF1zdU9QMg9m73/pZ+9DDP2jQRAHqjo95+CdNcsV7731u9r57rsZvf/ueAPYFCngzNmHEKEsMF0DtKwHn9K6PlO/01NMLT2GJ4JAyAeLHFOC3WxiF8BU7lMf+vffGqONj4r7AnLlL8d///QZmz5agWQHpVtah+I3UM3mqO6Jj9sHCajmGfDRFYojWq0HHH/7wEYZ+PAOhMvvIGYZPZeAxWGST/ZbmykDDBh8NnQYXuebZc/3xzjujxAaJWCEDBbuF0fhwyBRY2YQ/6HsE+cY9NAP+ZqzKOI/g5WuwY8ducdp3fGDWIjr8DSLFWZOahNrVKWjkc0YEEDq6EPD/4pe/fHnAtgtA9DekjO+KF//PZP2+rH8j4JwA/ocC4H98H7T/xwPw/hDwE9CTsmOsDwH+w2N43r9IOf9bBg+sRwP+jvYkfVwbC5iEn8zRd112FhoksKalsqLN/t7yhYB/0iQrAfz7BhzgXyqAPzklUwP+z63g7bUE586fe2ndsrz8tgD+fJmC34QNm+4prq/xUuyLnwT8+RLkHhJWKoDfH0cOS0IctLw0e7IiDfjN5iadKjk5VUBfjArUb00x6Yt9q6NtJuBnHAwB/8aNAvgl50B3LwT8Lq4C+D2eDfg/GDThCYDf7T7gF3ECAfyvvz4MC4TXz+teJ975OXOX4c23hovYQrrMRi/CW2+NwKQpbli0ZJ2i4WRKEG9Wzi2Jo9kqsWjxsJVcEhNFYvbXv34b773/uWzfJkG2vnjllY8kuD5TBiY1MptQgjffHA7bhTHKE+/jvwEfC8B3dUuVeKJcAfGjBcwvVYP4rJzbmDXHT51PihFnET6WvCjzLYJkUDAdv/nNe7C0CoW3z3rMnbcMr7z6McZPdFb0H/a99gH/OQH8GZ0G/Caq9OzfpwRBatetQdPJk51OvPWrPgL46c3/ngD8vxCAT8oNPfSkd5ytSAAAIABJREFU6vxUqDsGeCdgN8C+QefhPmM1KDtP2vfTb31TBe6+1IBdDiy6+4epy3+5FmiWIN2atelKmrMubyNMotfdG5enAX6zN+Khd6KjL56H5z37XMPjQcBEOlHrh2NH6+vscYaHXwP+IpFpzIFljwL+jWYPf6vAutb3k/2B/ULRMl5S/2hdf0f/Vx5+4RGHhJYI4F+KI0c04O8pWU4N+HsZ4N9m5vC//8F4ob0sFJWdK9gs27iuWXtZcfjfFUoPvfaUTn5fQLqDU4I4AVrk2XAHluKNf+ONT5USUWz8l5JNOlBoPjMxVFbSaXx8s9VsAIN/SeeZM28pJkxywf/87n28J1z7EBGjIOB/SwC+79JNKBQ653KRoWVMAbn+fMb4iojDsGGzlcqQvyQ3pLffjoMBEbGguhsB/SuvDIX9wliZbZivAP+sOb74YNB4/OpX72DyFE8ZAARjtgwMJgtFiIOTDRtvP/Y+M953qyS/ynMB/gdBu66S58dTgna3dYrSkyUe/o/efgs//Mvv4y//7M9enke7k95z0mv+XLzupOP8i4ByAvbW4L41yH8aqDeAf3ufPI+zBH8qg4qX5tk37NAbwaBu0/NboOm86PDHRqLCywU1q1LQfOOr5y+sG89sD/ATEJfJw7ZMQJgCWq2mJJ8GgPgw40Oa51KNgN+fdjzLpoKCAuDCd+SxVDigmkF3xhJowF+gtNZ7MvHWQw9/+4Bf9UHxWLI/FEu/oEqG0b/Yb57Wr7pzH9vA+s1teNi/+T1PA/42CeF6F+A3DxrpBSfQ5Kf5GfPwHj6p3/A4qs6U3T9P9cmnPNuMfso6Hp7DgPTujY9ifb3Swy92i4zao7zhHwpXPkGCWnkPqHoTJXQaKt7Q+8+gdy/h6r8u4N7SKkx+403i4f8Ks+b6iVd/uKLWMEZhpQTg8lgeQw8+BwhTprjj7XdGYtr0RUgTugyBPWcCOMigghYB/+tvDMMS/2wUFNWKh79UzSo4SUwAufZLfLLUIMLNPU15+N99b6x46wNk1q5GZg9uY8YsX7z66lAVKDxC6Eb07HMW4hMZJLzyyscSCJyuePsMBCb1iLEAfB48+g7kd/aD5wX8itKzawdqkxJQm5qIxqNH0JnEWxtzcrBw4gTMf+dtjP397/HLH/4Qf/Gnf4pviGrNSwe9Bvh95JPqOn8lXn0G4v5bq0Dc1qC9Nchv/X/rY571P6k+/1cAP+lCL/3auxHT6aJ7wAKmu3eEyrMLdcUFaDj4BUzCK+2Ny+OAv1nJoFFJISp6n6gJfCkejrvqIWUGO2298Hx48YXIT4IyTq3GJxxUD7z18rB+0kuUx1N2LUGCoPhw5gPTa/Fa8cZskZfWJRVUa5zLYzmQ4OCAwMp4iBoPT/VCVvuNgYL5ocrzDLUIlmW0n2WQw689/L3Tw2/0JXoCw2Q6frFM3dNjxhcp+xbl79T9FEDAe2oeJJpl8Iy+wfvOctg3eL/N/cesAvSwX5nBO/e17lfmvtK2zxWpF3WT6uPMUcG+ay7HDBh5vgb85oGkkaG39wB+8wCN92etqL3w+cJATgI/1V+eCt6bFVWLai/pkpF47frrDxRY2CeMvmR8sjz2O8o5Zsqx6WsuiUf7qgS03lN1sT8ax3b1Z28E/LxGOnBoN3rAXxMwP3deoKLUBIduxpRpnkKBGYpZAqjJa6cyz6+EivPpZ3NFl3+/gPxCFYBL3v5SyUHi4BgHN49VSF19TgF5cvrp/Wcg7XsC0i2sQrA644IK7v39Hz7Ea0IP8vHLkXq88MZbn8FHZjR538n7p9IOlX4I+H0kFmDYsFlCA1ulAP8fXhmivPxmtZ5ifCyBwwwCJp9/tAQLD5GBi7vQj3gtb741UqhAKUhMOa5mIaysw9WAokByzDx8HpnjGoz+8byA36Q4/LtQuyoZNRlpaBQtf3Qi027upk2Sb2Ma/EaOQMSUyfAZPQoTXnsVP/+Hf8Qff++7Lx/4tgL7isIj3/9GBh+k7hDIPwm0twb5rf9/0vHtbSfg5wyCBvy9EZn2sTY1X76kfpRVwctQK6mwTZJcqTcubQC/eF2oXpOUckL0qyMkSGmyTJk6yoPsmPKI8cFoeFrND66HL7x8eYimpZ9TwGzkKCv1UAyP2KFefo8+9PjSoxYy5dXGSfn0prwtsmzkTX4o3p65MmXLQCu2hTJmPJ8v61x5iRKo83yjTPV5fz81kPmCKTZeqnz5tpKHM58jnmLZpgF/gSRb6n2A3zxIa0RK6knR5w6R4Ljp4rkbpWT72E/4gqcEHvW5DS9tgcgR5goAV/1FDQhbgSrpA9Txpnwe+yj7gHk1D1J5DvsVP1m3GiTcHyiyP6k+J+ezz3HAmSngjfzf5WGbVRuM9mrA3xbsE/T3BsBPvXUO/gjYAyW5mK0EVRJ42jvEixzsTjV4NAaExr1k/+Agkts5sCPoWyjc8AXSH23to9Rza53IM5oB+8NnIM9nP+JzkINTcslJBbG2jRCqSoZSeOFg0eh/ajaUswCysi5uV88vKccYoHJf6/08jmvrwYaxjV7z3ubhN1+PeWAdt/KQqN64yfP+c0XHGTpsBt55dxSoxhO38qAKciUv/ze/eUd4/J8oD/vkKR74TBR3OOgnoGYw7JjP7eVehAq3PwCjRaXHWrb5idNogvDmR4kKj5XYfLokKuSg4f33x0lugjih+SxT8QPk+RdIMkOq6LAtDMJlG+loIh2H8ptLBdS/8upH+P0fBsvsgr/QddxVAK+j80qkiBqQpQwqqOyzImav9I2dMtiwFyqRrXpvjRXK0sRJbtK3tj94V5r7SRcBfklWyIDde/OmodxyNuoK80WHXzJnd3DZkJ2NEYMG4dNf/gJ+Y8cgbe4cJM2eBYePh+Jn//D3PQb4yaH//8TT/rcC9v9VKDztAfTW21qD/Nb/tz7mWf/zvH+UgcUfaw9/B3uPPuyJFjDr8Mejyt/7vg7/jSce25M7WgN+lVhEOJNUKqAH479+8ZqaMuXDkfs4vboy+bgKRmKCLoIoZkpMXX1GMhNegr88POlpYYDV2++MUdkt+RI0XqTGw5+ffPBSeWGQTOWSK0k1BmZhnCCKDXwh2NhGKk1kPpzTpPwAARDePhskuHi3mkXgdmoiJ6edVlPEfJkv8duA+MRDCoitEi8PH85MqmK8TJnkZWXyEXU+NauXBmzosqBdan83NzWrtbP3s6+r9LSYWtDY2CgZ3juuFvEkSg+BDoPwOOB89bWPMVZ0uz0l0I6ePlv7aNUvPxMObZIMCDiwIwCnfjcBQYiAcM5McWDKAWVS6gnpD4cVbYBePsr+ZcqsgbkfNittb76YmYSN/ZH9iUCKajsc9PpL8N4S0eGOXLEb2cLHJXAMCCqUafxZ8oK3EVWOPSoAkIMB9rGu9vCrPiXeO/arjiz9OWiXtqDmu6n52X3M4PB7eMYoTztBPwMqGWj58SczFO/7o6EzxHNsobzEuQLCeW9bP5+Kyxrl3taJl3mf4p1TwnHcBGfxAs8RCsoULFq8Tj1n2F+N/sTPlLRTCqhStpFKM9PEs0zpSHK9J0x0UlQUDkDZX8yDTUkyd9+RYQB+OlXokDAGo3mFdHSYKY/cbvzP+riq46VMtqWrAL/R9/i7ftryrKBdo428Xg64+btd4pMtA644GUTFqYFRijzDeRwH116L10gw7WcS9Ook/2ea9fTFMcTgXf6+UuRdwBlhUnEYYEunEeMAqJ8fFbNfvP9p4qVfrX7PpAstkjL4bqDUZrTQbPgbp72yZAAYK4nKMiTImXVzBoJKP8wDwOPfe3+sckZx1tnNfZXy7HPGkfciMfmYmvkm1YcJuzgLrnIGSJvojKCziu0x7qdhA34ag73n9fBDFH2ajh9FfVE+6ksK0XThvGgDmFTOhabGZ/8+srKy8EsJ2v2LP/4jTBTPfsKMGVizYD7iZkzDmN//j6L3vEx6CyU2SeH5vlB4/p+A7588xavfGsC3Bvmt/299zLP+53lU6GFmXs4uvMzr1kG7T3uq9MF9LeV30XhgHxq2lKLp6GGIRlqvvIrWgJ8eojx56KauOi2e+lVKhozTniskLfr2XVAPMsqb0dORuf4aYmQgMFU4k07OSWqanFOy5EbOnOUjGsYyXei3SYF9A/CbH3bkZd4AtYo5MLB3iFXeN7MXtkoNKPz8N8nDe6+ArDsqKdO4CQ4CsmYqvuSQj6apFwW5nARq9K7QkzNSlBOo5TxcPhfLgMVJ2jhJArc4WGHb6anzkofxBFFviJaXOBVVlgXmdBngrxH1hE0yXRoXFyfayrski+xtNDY8/WVpdIi+DvgvXryIlQnSJzIzcVJUI6qlrz8L/D8G+MWTyX5SIiv70IcfTlXeO3WvBCiwf5COERq2TQZqBYomQS7wjJlLlIb2sE9nyezQZKWbTbC+MvGoUvLgNs5SEaST62tpHSaD08tqoGBrF4WhAvyGie43FTrokU2X5D306I4TTx1l+ajH/YmARFeZ+k/mrIN49370o5/jF794VQalESpTqOGB7WrAzyzYW7duley9K7B7927cuX1HnHlP1tzuz4D/0qVLkuApFakpqTh69CiqRImn+QlUhoeAP1b97jMyr2D+ggAZPNorGcfVGRfh6JQoz58RMrCMlAHjZjVoTJHnHmcUmeSJg0f2AxdJ0EQPL4M7U0Q7fZEAQHqfJ0ifIrgnR5/PNWaXZYApHRWkmEwUxwXPT5Vz+Ayyl5kFylIS5JL2yD5Ir/+MWUuUZCQDVrmdIJ/9mpmnvX3WSYBqgFqpNMMkVUwOxUFqjtDaWC+dLpSZDA4pxQYBoBky4HVzFznSF1TpYd8rKipCcHAwtm/fjhs3bqhBvfHMMj4bGuqfqtLTFuyaZ2b5LCalKkscMHzf8HdPOxLwE6Dzd2rnEKMcTFk5N+UePpwVYXkcwPH9Q/DOsoznBgcUtCEzX6uZOQH22QK8jQy4zLL7YLAk+/hMMb7zk/bke4F0H0pw2titUPUQ2LM84/1lHozVq4ECt/Fc1sH2MBOvuaxWs4zSZsMO5jJegMMvOvz16zNR5WKPKndHNIoSIJer164hKTEJGRkZOH78+BOfwZTl/I///E8Fbn/1T/+ERcOHI3XObKTPn4fA8Z/jjX/7abfx+QmqKX35DQnG/ZasDJYlfeeHQquhV58UG4LwZwF17m8N8lv/35FzHz2GtB62RwN+41etPzttgaazp0WhJ+T/b+89wOuqrvTv+WeSkAzJpEwmmRQIIR1Shky+SWDSmCQklGQgVFNs04xx771J7r13y71j3HsD2xiDC+ACppfgALaxdPu9qut7f/vqmGtZkq9kdR89z9E599S9115773etvYqF+nSVc80cK/ygYim0K/zBSj6QCvgBLjhIPvlUkdOOosVEUzVFk8xTe7RcLK05ERBuk+YKrWtLLXfeKi09mnnMH3geW9Vevec7IDVk2Do3GHuA3zPFmTP3iADdA+7dM2Y9f9qJDg0VE643+DKhPtJ8oN1wU1MH4NGusoQLEMtUtkSiN7D821TAfszYJ+SYNUw2m7c4gEfEhl/9+iYXm3nHrgKbNeewErZ0cEIDqwCkbh8xquoAfywas4EDB9pPf/pT+x85Q7Vs2dKysrIcYHvrrbccQCmrieoC4B+osJxvVzJZ1ItHXrRbbr7FfvzjH6tdb7KMjAxbsWKFHTx40AGF3Nzcs6peGuD3tIA9FdMaxz3saQmht+tpnHfx4Sh2fFTbsVrTqfNUaU6bCsyNtfGTd0rQHOzanIkaofG3v/2bYm//XiEDFzrhEGc7BEYmdDSJJNtBOB02Yp21bjtawoL4SOY6aOngqT5y4OOYGNxsrCQQNvCaq29wTnvE6AbkU262swH/WdWu0Ak02lOmTLFva2KGp1q3bm2zZs6yPU/vsbffViZvCZmpf/UC8J+qnGkjIKZJ4yZ2xQ9/KED4ZwlgfWzp0qV24MCBs3gsKoHTC8sJOFy3Ub5CWukBSLNSw/gySE6a2Hx30arR8FEbnKAI3yAMDNW45ZmOMW4g6LG6uXl73AFrzA8J+/gR4C9wY9hcCZrXy0mUuOyjx2xz4yG8y7gIuB2vyDAkbpojDXErAfkblfyJMfYujUvwFwIAGvruSvB07f/eKvOUDhrLerp33iUzSXi2U5cpLunUTPmyUI95i15xKxCYpyDY4C/Qt9+M8wb8CJajR4+2b3zjG3b11VfLxv5RN54B/t98883T4xmCfb9+g8oMy+kBXW/PXEA/hiZsXp/GhIq+jjBDaEuEfvoUSqjkPd4KTGnPJ695Znc4WHt9kmfZvN8AbsrCnnOpvykP8wLzQ1uZfTHfbNG9Jcvgreh5zya/m6xT8lulg33vu9Cg0hp+CWJFG9ZabmZvyxuUYSY/Qf5ee+01rZjfbldeeaX45frTY/Bzzz1n7ysJKGMJfyil/uvnP3fg9rOf+rT9WfPVqHvvcYD/sbatre31f7ZvfOnfqgT8AvAx1SFJ1qf/KRlHn0RZ/6FstzjLfkebB/I90F4SjJf127u/5L6s+8s6/1MJD99WOS7yAb/jD/9fJSlAWM74iqUWnTbREpvXWVE9Csu5U5MUk1wjTTge4MekB2Fg8jRlFtRyKxoQshV2kYb+MS11MtBxD9kW0Z799a/NBLLOBvwsnU/TigH2+mhd0YCRFt4bPBkMGWB5F5o3l+1Q2jI0OspR4qI4EE4NQEaUBJZ/SaG+74AZPgNMomRZxMTjj9c1chM5Tlw4BROloYM0vKsIlyYzjBEjzzbpYSn7H//4hzFQolXdu3dvudu+ffuMDW1YkyZN7OMf/7gbLC+66CL7ocDJjTfeaG3atHGasg0bNtjLL7+szL6BM9Le1wXAP3TQMAfQAVbnqnPqdeg0Z84c+/FPfnx6kvjSl75kv/zlL5Wg5m4HzObNm6d4008bKwGJeML1qLIA//rNIQdqrlW7slwPyIK3PMDAZA5/MGE2kt0sQGjOvKN24GAy4gd2/pj8ENbv2mtvcz4hq9ad1PO5zpEPnu3dZ4E9IIDE6tUURftgokfjOkpmQ7O1MoBmtqM0tWhjsb/GHAOt7Ug5DY6TsMBqAAl30OrxrAcePMCfkZEMy5mTk+200Xv27KkQTaHv/v37befOnQ7ke5qnT33qUwK8V9j//fX/rF27djZx4kTbvHmzW1UJBUO2ZfNW6z+2g83e3d725Y+2A4VjbV/BGHvextnjR3vYsDGdbJlWvwISznN3FVpsR55yhRRYWDTeJDOGfs0ybb6EI8/+3nO+rey+pA3/0AFD7X0pPj744IO0+5dHiwULFtivJPR4tPjiF79oP/+vn2tl707r3bu3zZ4923bu2OkEoQ9PnrQFCzDtSMbhx/fiKeEiNLyYBjJ2XC8+IJ46GnPMEVFkMG4QzQV7bMJHYpqxU7xGGxMVhlXJvypcI6tG2OhjzpGM2CMNtTT8ZF29WqEgAfAIA/ApvMFGGXZLabJD78LkxHM0nTpjrxMyb1aISey/XUSZlsPdClKHThPFi3tc2QgJSRk7yNQNgQITEvgPU7Xf/PYW2cH3tXXSdgP4+/T9CPCzEsKYAx3TGc88ekPLR2Tu4dH7k5/8pBvPEOgRPNH8b9++zZ5//jnZzHe1vpmKw1/sB+P11YrsoRGaczT0rLSwyuIB5NT3ANI9mqeeP99jrw/zHoJUYKZKGTzl0/m+33s+yQ9nA35WRV999dVzttOzErj2rFhuu7Nm2O5ZWk3esN4JvYuUWfqqn/3sdHsxBv/iF7/Q+HiX9ezZ062OsfI8efJk+4kUM167EqHnjmuuton3N7Xl7dra6PvutT9qLP/kJz5x+h7v3nT2APyLtJEF9wvS4BNt52sC+CTKAlj/oFiTT1hNwDb7ssB4eedLAn3vd3nPlHaNMmDWc3FN2/FXElf6j9VRChRqUktsXu8cd3N3PWFFgew6WdKzNPwC3Ds1MTFh3X1PR2nX73d2rABwNjSrbaQN/d73r3JxiUcJKKAFYSDm+npp0zDxwVQChygvsg4DHpoQAD/aKTRhZCRE88aEmnSSS2pveBf3Dxm61oVWa6XEK0S8YALFqYrnSGXeTs5ZaGyZaJWh3Dle3aTyNlK5WUZvJnMj7G2x98QxDK0ZtpxJe938UgE/ZgKYpjzwwAOa3P9qt956a7nbbbfdJrOP21SOv7gJkYmx5MAI+P/mN7+pifqP0ob1MwBgqtb7pBy6V81fbcs0iedsDTsQ5oGxoHwVNoxVQho5oc0TGFsv0AmQqiwIS33OOe2OUBz++zra6BFjbI0cunr16uXq+7e//a3cent0adSokYSu/7Uvf/nLZ9UbOnzhC19wqx733HOPTZs2zf7+zt9dPygL8BMGj+Q51/7uVhso7Tq+GvCVp1XzwD+aUnjgoYcy5ePxpu3Za05zerOESPxP8AH4g4TKRnLiYzkeQaJPv4UOrHWR4yWgDGESO15P64h2b42y/mK20Uia1wce7Cvnv4HOnpckPcNlZoHWD8D3sM6zhA9/e2DBA/z9+49ycfgPHNhvXbp0cUKfR69092jrbr75ZvvJT35i/6+UyejTn/60Xf6ty09r8/bv22/r122wgeM71VnAP2LwCLeKtFahARGOqV869Ljjjjtc3/nG179eKi0+97nPOc0m/RBAc+TQYZn+LNZqjAC/fDkcSNQYg7YfoH7Hne1dZBhCKrrxoFire6/8RX78k2uc3xLKBgDsDvEeUV3Gjd+hMvdy4xErA4NlTugBfk/Di+PnL355nd3buLscd99y45UH9CgD4xf+Ajh8/lYgnZVRN2bqHLHl/3ZrS2nxV8ipeLxTtGCDjh8BDqmMXWi+MWfDD+GhZv0VseqQVllHOsEVAYCyLlsO4J9u65VpNyYQiQDPmHPLLbfI/yC9Pu3x3o9+9KNS+zTj2aWXXur4umuXzhr/GskEdGExOC5bu+3Rorw9dGI+8OaAM+/NdSvJ9GdMn8689pHZTEXP0//5Hv2YZ/lddhkq/53ku5NRnDwN/549yUy7gP2hQ4c6LX257cR8IyH3tkZ32213a9MxQi9zy1e+8pVS24v+cdVVVzkFDPcyF52eozS2XPrv/27Nfv+/NlPRexbIvKednHq/+9Wv2j/L3Ob0fRrLSzt2WXB1DYfbf9P2FYF7TGQu0/ZdgXts8pNa/I+djqNfMoZ+aUD8XOc8gF9yf67nSl73AH+Nx+Kvk2jQL1SlKYDdfrhPN8u593YLjxpi+ZU0l6h0AdJ88GzAn+8mjnmKiMCSMpqwKdOfdZMVYGu2zHHuF3j+b2UXZNLqLdtmbBy5xoSGbWmPnnOKNfxr3SDKAOoNpAyqOFlhV40GNqP/EmcKQex+rmHbOXnaM4apz2BNiGhsmz0yyGn4WR4nlfpNAmpo4TDpIYoLGlkAH3a3N97UxIE8nHYxz7heWlwm6D/Jvp9nvOguaG5KC8vJEjV2ji1atDCABkD1XNu9997rBmomSE/Dz+BIbOPPSoPydQEVBlwmXcx+0LTUNcA/avhoW7d+nfXv39/VF+38uerNdYAbS8glAT+CD1rY73znO4pf/RsB84ecSQCmKPwFAqdE581qg+I4/G6ST9r44vz9Jzluw2c49MFbbABqIkaRNIelf0L1EZkD4fRZrfBkSZC86S/3u/B6OII7wN+oUxLwSxuLHTbCIiYT9zbu5oDSVPE2K0zY/+IsPEFJetDgIzTA68u1soRmlWgfAH5srQH82F6jjQQYlAT8mZlJwP/CC8877TOgNh1apt5z3333Oe3cz1K0dvDUP//zPzuewtQC5zvA2eDBg7XKBOBfbwPHdbQ5T3eocxr+B2560DzAv3HjRgnjzWSe18iBkNR6l3YMj7FSRp1TQccnpIVEoMTk6Ve/+pU1btzYpk+fbodkSrZg/iKZ/WU5wO9su9ecdH1/yzY5dCvCE86df/jDXS4SFOY6OHGSxfXyy38sfm7q+A4H7rUbc9yYRiSoTVoZIYILq4o45KK4SCo0kgARgE50F1YuiUgDmOc6vMv4h+/SwiWvuXGIsW/4yE1uzERwIPoMz/Xtt0RlmujMFrM0Bq6XjT4rpQD+Nm3HaOX1ZcWa76/+cbc0t3Oc/xJmP/MWvGKszGLDnwT8WywWjTjtPmNOurSG/h7v/WexrbdHcwTPz3zmM/a1r33NuOZWV6QguP32u0XPqgH80MlzoE4eJwE2fQyz0SmaG1D6sGqLgJME5slxI6kUSCqVvGeTfTN53TuXukfRhIMtjrcoiVavKw75q++l3ldVx5SH+fAjwL/X4vGovf76686ECh4+59hLv7njTrvnzrvsHh030TP0j5KAP7V//O53v5MD+cNuHLrsssvO6Ecf15jyX9/6lvW99RZ7XFr+8U2b2vVX/af9SynKK48X0OSjxcfZ9hKBfLLgXlmsvfdANWDcO2ZfEpx7v1PvSffYe7bkPt3nvfs8wP951aNGHXfTxGf+bfWEAoUfvGeJtassNnuGJbZtNuLy18W/JOBvL5C9PwnYNRkShaB3n3maZG4TsP+jdZZtKyHR5ipCD3auLH1jP03UCTSlIzUAk1UQ85wRylJ4nzRl2DkDihigMflhwPQGOyZgbGj/fL3CswlcDVamRBKVIFh0FVBD84ZJxRg5opE58TY55Y4VyENr1kLL3X8R4MLUAsBPwpaJsqUG8I+boGQqf5bWQ5rfZQJvq2XKcfsdre1bl33ffqO06hlyRGYyYZJg+bg0wE8bJRIJCwaDliMHKcxvytu4jw07/fbt29vnP/95QwN2ySWXOKDLSgGAbOXKlc7OErvrks6GtarhL860O7j/YHvt9dc0+cTLrW9JWuDc9+STT9o111zjwCiAADtSlv6hB1r9p5566rSdNQIVf6Vp+GkbwNFCgXzAF2YMtDcOi4ArNJ2AG0wxRmuVgzB6mEJgvoW2vbeidlwrs63mzYc40y9AFcDMafgFqvDQzplOAAAgAElEQVQNQADEvAv+waSnh0x/iBwCD8N7mHwROeq229s6e2s0wH+TffXV1/zZgUTieJPM525Fo0LQSPJ00h7Y0/B7mXbz8/OcnX06fFSSrizxY1qWmZnphEhvlei3v/2tkgw96LSBJNHBnpo2gKfqvA3/gCF2/ORxZ08cCoXS4jP6lue8fMP1NzhaXHzxxW41DUET8xL8HDB/gl4I0qFQ0NnwZ2RmnXbwxP6dcYaoTjjx4wzO2HO/VnEWyA4e8zEESHyAGFMQDjHzIpHSYEUHQ/BE4TBfbY5ASKhYxic0wwBNwOeSx946vXrQo9dcZ3LDNfhilgQA/AVwxL2vSTc5/v7N+YggbOLsiy/Jbbe30vXVLuTkX/7vYSk9nnMOqwQn+O21NzvnXRQqGf0XuxUKhFJ4Fb4mGhkCBlFn+vQtzrQbCRn9DV4qyV/l/eb+Y8eOOVMQAB7CO+PZr3/9awVCuN8GDRpkRHthzCM4Qc9eGeobMumRUOQpd6oKIHvzBnREsYTvw61aCaG/MobTt5P3JPPAbJHpHn0y9fvch4In9RzHvJM9iiFi8DPX4F+Req3kM+f725sDPcD/tOxUAfy0E3MD/F5m2+hajiwHPtz7jB1fudyOr15l2VIs0j8wmWRsQOFE/7jiiivcGNy2bVubMnmKu35Spm6sXqMo8IC7t//K5/7VbvvlLwT2m9jYJo3trwrbebHmMe+6twcUn46qIy3+DwXyPfDsgW/vd8m9d73kvuR96fwu+Q7vdzrPpt7DasMPJaz8hwA/QoxXz2rfu1nQ/9dgKFBw7B2LL55nkTHDLb56udXtOPztBcz3O8dcBjxiGP/yl3/S0t8PpJ2+XKnD/yAThgEuwkkjaUxbKVIPYIfwZaQjb65IKSx5d1LUk9/Loe3KK/9bE8R33PL4LYqeQnjDpC1kUtPCoAdAw0b699KyoRUjBfp1is7ARIhjLunQuQdwhunFLRrk79K3r5NmC3MKHIm7dJtuZG7EpCep4d/ptK/kAcAECPDYTkLBD37wM+ckjMaWCWm7zpcH+CvDhEySAPtrf3etmxQnTJjg7KtfeuklO3HihBMiynpvXQD8gwT4K+u0e+TIEadB/NOf/uRMWLC3xmwJbT7ADr+Ikn9nAX5pQplMaR9MKLCZBqxfffX1AvG3uwgn1yqq05+kaUcAXfr4OxICdzjhDrOzRvd0dQD+/5QcB+dIhARAGfzqAf4+/Ra4rJ4AKLR5jZv0cu+7Q/xCzgns+llBul9RWX4lAfEuxfxnVej6G+6z/7zq1w4QdJO50V8Fxv7nV39x8b1XSnOM/TbAryTgL1nniv4GvKKxxh8CkI/N/pYtW5zWlsnbc8Tz3lvnAX//IXbiwxNecSu0xzSl2cPNZFLzeydI4hdCBKM33njDgSRPkOSlMQEnnHYzpOEnwgvaYIA9WVkx9Zo8dbe06OPsd5jJCEAS5ebBhzNkt9/dOWnjn4EfCEmWiPaFYPCIhMiJk3dJQJwlB+obnVnNjJkHTq/wMNZgrkP4T6L//P4Pd7gIZKw+Dhi83PHOddfdI2C82Hr1necUEyhMCGuMUIng2kS+BZMVArZd+0lK8tTMrXJS/iTgv8WtNG3dnu/CS+Jb9aMf/UIrqRr/ZOdPZnNWWKsqLCdKj3HjxjmTMrTOY8aMOWM888J1FhYWaFViYNpOu5UBzMwXHuAnMANmm8TeHyRBH4d8Yvjjz8DchcPtOq3KcD9jPAIewhYhNfHj8VYPksA734XfbSIlAnPEjQoOgd8G95VuUnR+5jzJ8e1MDX8S8MfS7wvhkBUsWWix1s0s1qGlnEK2uWdffe01OXM3tuuuu846duwoH5bkGIxCACHCG4Nx2v15sdNuKqj9uMD717/4BbtRQP8uhef8oVal0fyn3oP5zmel1b9E936/WHtfEaDt3VtynwrA0z0u+Q7vd7rPe/d5z10u0P8Z1a3GtPzpt7h/Z32gQP7LRy08bIAFWj7kovUUHnu3Tha7pIYfbRMJRYYJpLOUjB09tviE5iRixfSs52yxEsvgTElEHgZItK+PLXtXg+0RTZJbXZxknMmwl8UWlgk3OdB7mpekIxtL3BOV9hxtGBMvy9o8x2oCDmjEwsZnAPDfo8dsF/IObS5aNsKxoSUhG2PStKLA3Tt95n5XHpZ/GfSXKMwijpaYCBHCjgEfUFnVgJ8JkEgJOFsCQtDSlARkZTFAXQH8b771ZllFLPc8dX3hhRec0y9aVjRV3gRT1oNlAX4P9DvzHU3gaGXJXNlaae7xxcCUZ/mq95wAubaY/7DNx5kRMIVZBSYShHQFSM0Uz9LWhOAjxjZAn9WfTYq1jwDIM4BAIvJkiZd5bo6iuRCOkSRLQ/R9zHwwXcMMhPePGbvdrXBherR67SkJlmVF6Tlb0CmLHqWdh38IR7lXDpcesC25OpT6XEMG/ICWw4cPy0n0eXv33Xcdj6WC/FQ6eGE5+2XKeVXjBP1+rBQStyrE6jX/c5M01Tc7R1eCD2B3j1IBR16SPjmzESkT+I1CgYhOhB7+lVabfvMbnrtZpnkoMTYKWCb9Az4Ccvkai0641UtCBfMdMr2yaoAZGKsGZOtlTMOsDAfyGyRoImyizCCyz4pVJxwvEgqUcdArO6taCJusCBB0oItCyP5BgkXbdnLIFqhl1ZKtqgA/vIcGH8Edc5Ps7OxSx7N04/Anx//KAWYP8DM34etwzTU3yGb9Tq2mtBXt7ndKH0xD23UY50KvYnaF8MMc0VNzC75cE10SyOS844H9Favf17gxUcJWMyma1K4K/cxKM8qpOgv4pc3PO/S8xdessPj61Vbw5uumwdZFTSIq2qFDh9xKV1ljMOaqpWn4AfaA3c9++lP2xc981i4q4bTLtc/JpOtyme9ckWKe4wHndPYeuC65T+fZkveUfIf3u+R95/rNc2j5f6CNCELkBUgVcqrtOHXA8o/rPwUK3pdJz7pVFp2T5RJk1AeTHiYMwmIyWBKGk6VvNo45x3WiVnAMaEarxDIyvxkguc6yNxExvOe47tmwpg76DLrcz7PEPkabj1kQ3+cZ3ufdw/OE0yNqAtpfV87i73n38m6e4fuc41nOcS/RMbBv9TLwVgfgPx+Ore+AvzJ1Lw/we23p8Y6Lca0VGxJiOT50/If5T5JXmdhJnAM4or05zwYvYGqR1Ool+c17J7zrtLIyCYD3WAXA3pooKkltbciFYsT8jPNcR7Dk9xYJD8Q8h1+x7YbX4L2zNfznB/grSteGDPgrQovTgJ8oPTJ1oT1pM5QDgGayriI8EqYT0I4JI0CPbLzwF4mw0BSjGYb3MPlCIGwloROzL5Qf8Bpt7mmM4dnTfKBrJHkiUhTfIrsvig+chr0xD55DudFOGWO5jnAKH29VQkPymXjfRljlXgRVIsfwDQApMfnR7C9a8oZzYGVM491VBfjTpXdNA36ysP/85//rwD6mOLTnf//iOudk3/xR5XaRc/2Dys5L+2CSRbhTEp+hqPLmFPiBfswKDgEeCELRTH4R5OOAznUZ8BdpJTlPJj2xRfMsvnShS8JVJHO+dP/KA/xlgVvi5n9OYBgteGWBNcDbe7bk/lygvLTrJd/h/S7t3vLOec+xYuED/nS5yL/vLArkv/2mRbOmWqh/H4stXWAFxz846566cCJVw8+EkQTKZ2pCvHNMaqkT3Ee/Pc29Z6d/5vPJ+0rX7rhJUuDMATX3/aRdZeoz3MMgzT0lJ9lzlce9x4H/1DJWvYb/fNrSB/w5DlSntrl37ARLAW7HH+KTJC9+xEupvAHIP/P6R23uvS+VX5LPFvNeyrOp70wFCfAgQiMrR8nyfBROlvt8wL/5jOhRJcNyDjkPk56K9K8zAL9WgeAh2iepmJDJmCK80IbJ38nzCIe0L3zCvQiGbBx7bc1zhCZGiETT/hFPfcSPqc/zDYAl3/cETe8ZhEhCzALoWcmkLO77GqtIQoWCgm9zP+ed8KrzyeeT96BY4TmP5ylnQwb8q7Sa1kJRiTCZYlVm3/Nmj0sYIoMyZnb9Mpcq8WJb59eQNfsFmfXt1ErAXda+g0Ixy/TO9V8UVaInzs2PthjiwD4rfQhn+FCcNvsspr3XXlW1p63ghzNt+NM36SnUamp87WoLDehjoUEZlnj6KSvSaky6fxUF/Gj2MXW57OP/z9nqA5DLA9DlXfPAdcl9ec+Uda3kO7zfZd1f1nmeIywnuQEuVq4A36QnXU7y7zuDAvmvvmLhkUMsp+2jFpkx2fLfO3bG9bryo3TAf+YEVlWDXV15DwNuVZv0nE97+oC/bMBfV3jmXOUAnPmAv+4Cfk/QS4L/sqO2lNXOADUAPM+XdU/J82c+c7bw6YG/5DvPvl7yfef6XVuAv1fvzGIbfoSX9Olzrvp41z064rTbouUImUc95EyqiMxF9m2SpBF9C7MtzLBuuLGxC+jw6KNDFW3rQRupULoIVbQfYJ9VG0z58AdrKX80fH5IXEbmd4JJ4BidFLbOv028Onh7r82TgH+ZnGmTYTnTnT8KleAx97kDFlv+mMWUcRecYfKjSPevooAfB12i8Fxxntp9ALcHykvuywLj5Z0v+Q7vd3nPlLzmHHZVLhx2PyGhBh+FslY5qvx8ug3m31c/KIAJT+4zeyy2eZPlKjlJoZxt6uLfa6+9LifTDnIW2+80WQxI3uDUUPdMSj7g36xkS1tttbLMdrivo+G0+8abb9QYi35k0rNepjENCPBP3K6450q8dfiIaOmb9KDlf0xOq/crLGeNafgj4dOZdvHzqCpAXdfHw9oC/BkZg+SfME8mSZ6ZU9UqjDzAjwkdgB9n/MwBy+zZ/eQdOOYcnsnHMlX+NUln/B4uyeKv5XhPJC8c/L2VEPaYQd3XpId8Mm5xYZ0x+SFh2ve+91O7R9G3CH0Kz1SX8LL1iQJFnHtDPhuL5SPxrCJLpa/hx6Qnd89ui86dabEFcyzvyGGzgmTks3QG74oA/k8IAAOGf5AC9gHWJYFzur89UF5yn+7zqfeVfIf3O/Wecx3zzGWq27/IN6HKAf25hId0Gsu/p/5QIP/1Vy0yYbQFu3Ww6OwsK5RNf138A/A3vb+dAP++CwzwJ+Pwz5u/TA5pH9Zq09SWhn/zuG22UmEmWzVqY/379bfKOu1WhnjE4V+1epMSGK2UbbwAv8wR6jqgKq98SQ1/TJGDtgjwD7WXXnyxMmQ5r2fqog3/pnFbbcmgx+0+ZcceJGB4spJReipCmIgA/7x5C1ymXWz0LyTAn4zDr7Cc6zcpHGf1K5ny8ogkNVPOsdPkY/B2tcwhHwF+mfQI8P/8//u9i6pE8AhyGRDl7Z57uypD7+tOe48j7hVX/lyRtX5VHL4zV+NLsbmWTPfw1cB/gtDSBARo1368TH8auahyaPyXyrl/25OYB1b1akXS1wfAj4/I0GEzlHH6eYXUzUubvQsVKjq2arkFe3ezYEZvSzy1s1pMeghR+UWB/e8p7GYqmOb4XEC6rOvee0ruy7q/vPMl3+H9Lu+Zktcw5fnSx5TbRNp9H/CnzYL+jaVRoEBxjKOPLbbwxLEWW7faCj88WdpttX6OCAwPPNhey6E73QDHsmd5wKYhXGMg37Q1oQgsS23u3CWKCV+7ORJqA/B79tVzFKryzhsa2cABA+yDGvQzicXCAiVbFOZwmuxpk1FGPLOL+sdjSYdgViqGDH1cccqH2jvvvFXjfbuuAX5WkMjsPEWOrrf+8XYbMXSEQgQGqp0usVhUibcWKFzkFOeUC3irv7yVvrYc7fVSRUvr1Xuybdq0VfHZI9VOa6JxERkms/9EgegNbu5Imj6d6ceVNGWp3LmkKU6hA/MkH/v1r//ikp/deOMDSrh2s0Lk3qQMxcts49aoM8UhktvV11yvML33SJv/uvN/YMz3ykDELhIwkuRxpZKukaQRoH+jQp3iXL1RQSSS9v6VK6/3ndL2+G6sVqKv0eM2KtHWZBe6uEIrgbkJwz8wd69WBg7ss8IP3pdJT9Vq+F20HoFgz0kX0xcPLAOsveOK7j1QXnJf0fdwf8l3eL/TfZcXe/+T5WjioUO1mflUe8/0P1CjFChUgoz4pvUWXTjXEjufMCTzuvh3/PhxGzZsnKJQzNXSZ30HXueeHJMDP+E639UkNc0lxEJLVZt/NQ34NwqIrVNYwWk9Zljbxh2sX+8M239gv3mxtWuCFoWyOz148JCSBI22sYq2sW5DMoLJNtnCe21UX/aAAwDleMX+795jlK1Zs0ZhI6sfbJVsp7oE+B3YVxSULDlDtmvc3vr36W/PK/Nwfn769sYl65fub0KXbtm8UYB/tM2cdVDme0lH3PonSJ57PPPqxArTZikxiPbTs+dIe/65F85K8Jcu/Sp6H5F6li5dId4f5wAzYW/Xb1Q4VJn4kJvg/De9R1F3yIhLmF1yvpAMrWOnKdam3XgjPC6mPThIr9E9hOklDCrhdgn9jPkm+QxOl0PvIjIcwJ+IPAB8MhVDuxWrjrtv8b3T96c+W9ljRX6ibPMXvaGkgEu0+jRKeSSeckmzKkLvIuGI+OoVFurT3YKZfSx3tyQIJfhL9y9dkx7i7ZM9F7DfkAA/ggFCweXKJUCm4LI0+4TnJAzp13Qf0Xv+Rfd+UttFcuz9tLZ/Sdku0nnur5DDb7oN5t9XPyhQ8MZrFpF2P9ijk7O3K3ivbpr0AHZ37dqtwXqESxvPIJiMTpG0YawvoCudcpIgCTtXJqReyiQ8Zuw0e/XVo7XOUDUJ+DHjWS8gNlVa13YC+3169HUx9GuDCGh7ly1bKS3/GAlfy5TM7XVN2nE3CW8mdn4d3xxYEMggMsrwUZutW7exNmvWPPtAwn5t/NUVwO94TPSYrARprWUuNqDvAAl3L9QoSY4p78mUKTOVCG6iy9BNVJ2PtLYNa2yjXpjEzZn/spLBTbJJk7Ls+AfHa5Te77zzti1atEzJsKZK2z5f/XmJIuk8Jlv7Kth4j7b+stsnCdkIZdll30+Ot+TGGCrzHJxtu/ecYw83Gyzn27tcYrzmctrNUOSeAQOXl1GOZe68915yzwwctCJ5b1WVPaX+AP2M/jOV4X2KlAKrXZ4DEuhVRNHiovSsX6MoPX0tNGSAJWTPLyk67bZOF/BfJAD7TYFdwHFDBPzfVd2+KKCOn0JpoB+Tpq9+7GN2pUyaWA2AFl+X8/LXtb9U22UpG07N/y7hAJqV9q5Sz6XdYv6N9YICBf845rzoIzOmWHzzBis8Vbt24uUR7dSpUxqsH1MylxE2cPAqLXG+K41ITJoPUpfHG8S2aVuurZEWeer055QBc7omxsG2UxqWgoL0B8vyaHg+12oK8APE2CYrI3LzOx+1Pt37KqmTloUrsCR8PvUs7dn3JAgvWfK47N5HW+fO41xcbWKk9+w9t+5vveZZ916zlaBrohOY589faq+88rIdkzkfiYrKS5JVGi3O91xdAPzwF9r9SdK+PnJbc+vRuYfslJ+rcVrI0sQl6crMGCEgOMNmzDqs+PpRtxKDxr8hbRukoZ6W9YJyBExWtu+xzi483aR/58tzHz1f5ATdnUpkMWfuYhs7doaNGj1dZiszqnjLSr5vTJa+MUurg7Pk/Jplo0bNsMzMkUrM1dkeeriF7PN7aOV6krvG9fTKke59lavT5EmztRKy3A4rMzkZyBn7GCvi8fhHZDzHkXPafXaPRefNttjC+Zb3knyFqiEOP6YsX5JmmyRbDQnweyY/VwrEf1ugHafkfxVYR3uPlp56A/b/Vb8B9T8q9mHguSuLN449EyJvxeD7uu/Lelfa/gDnaGf/cj2jQOGJ40q4JTvKx5c6KRzJvO7+FSk73zEX2aJHjyHS0ixQRtFdSkazW1EPnmoQW9bsPcq2uEXL/JOl9RlmTz65wzBnysEJSqHOavOvJgA/IGz96E02Qxk+OzTpZN06d7Nnnn1GESJq15wJuocVwerA/gM2efIMOf/1t149B1ivXgPr/qZy9u4z0MaMnmDbtj3heAmH0VdeecVlu6yI5g46nO9fbQN+eGzjmK3isZnW9r72Avs9bd/+faVmaD3fuqbzPKuXzz67Vz4Vo6VZzVJc9m0a0/ZoPGNcayDbzKdtwuRtMmOZYiOGj7d9+/YbChyyX9c0/9EmCLl8+30FqSDrdk1tx479Q34z79gb8kl788037di7xxygrqnvp/OdE8IEnqkfvg+JRMJtFVG4EP0vJt/AQPuWFujSzuJPbKsWp1200p8W6L1cQBbA7wFbb+8B54rsU0Fy6nFF3uHdm/p86rF3Pa296oYGH9MlNPcAdjT1X9H2LZ1D2EnrPcX04T3/Lm3/p0S3c5r3pDOA+ffUHwrkH33RQpm9Lfv+uy08boTl//2dOl94tN2HFU5wyZKllpU1U9EXZjSYbdrU6TZnzhwBs22ajN5XNvIiO3HihOG0nH0qu1bbproBP0Bs7agNNrnbNAGxdjYoc7AdPXpUQSNrNmzkuYiMXT+ayfq00WeSmvyPaAlvsdX0X20CfpzA143aqNWj6dauSUcbmDlIY8nhWqFDKt0LBUDffvttW7VqtY0bN8mGDx9rI0Y0pE128+Mm2rp169y4li977uMyKQP4RhTC0f9rWBQoDAblG7jRQsMGWmjUUEso625RBfxi0jXpAfCj7cZ+/fufSGbYBfinavvLAsNlAfDU86nHZb2nvPOpz6cel/dMade8Z71r3u/UvXctnf0PRaNLBfrxgSgX9DcstvRrU4CGf9sWSeNLlA1vlxUG6qbTbsmWQtuAlpLINadkhtRwtpPSPOVIo504XWWAJVqWml/+Pl0Ed1CdgN9p9sdssikyY2p5d2vr17OfM7GoDUB6Zq39X1VNgdoC/N7q0ZRu0+3BWx623t17O7CfFISqupaVex9j2nvv/UPg/y176603G8xGfdCme5pjqMOYBtiv7XGtci3lP1UeBQplCpR4cruFx4+x8OTxLsdPdcXhB/TjkIrZC+E5AbwA4XSAb2n3pILo1OPS7j3XudTnU4/P9VzJ66nPcuwJNd6qBudKPlPeb+7H9Oeb0vbjFFwm6C+vkf1r9Y8CBW+9YdHpU6Tl72XRRfOt4LjCZ/l/PgVKoUB1AX60roRFnNVnrrW4q6X17NLDnnnmGR8IlNIGDeFUbQB+4uxvUCbTLPkytLqrtTPj2auQgRUxU2gItPfr4FOgJihQoPDe0TkzZTlwr2U3a2rxjevM8tL3Q6uIht9zNsUZ1UWsEfAnVOf3SzF3AeSSoOs72r4pDTdabiLh/KBYUAAklwTX3u/yAHRZ17xnS+7Lur+s897z3nXvd+reu5a6966nnvOOuQY9LlH9MYvy6HjGviaYxf9GzVEg//XXLKzEWzlKvBWZQ+Ktf9Tcx/0v1SsKVAfgB4hhYoHWtWPTTjZs4DA5MT5XJ2z261Xj1KPC1jTgR7O/Ab+QHrOsfZMOLhoP4V3z64AjfD1qNr+oPgXSpkCh/J3iT0jDPwEN/wRp+A9Ui9PuGeBUgB9NNRFtyEqLkyu26l7Emm9g/67tCzp/sbZP6T42Qldyjqg2P5RZUKrW3APM7D2gXJF96vOpxxV5B/d6z3rPeb9T99615D4pxLDiwQaw597Ue7xnMe/BJApn4JL0/Ke0W9y/sV5QwEXpWbbEwpPGWXzDGiXeqrtReuoFQRtwIasa8BMpBbA/XlkkMeMZ0GeAvVgLmV8bcJPVyarVJODfJB4jvOv4DhOtmaLxDOw30A4eesEKi9JPAlQniegXyqdAHaaAs+FX1L/Q8EGy4R/uEnAVVUOUnrMAagnQCoglaRUabMx+yktQRQx7hALAcWmgPxUsp3vsgeqS+3Sf9+7zni/52zvPnmvsAfBE7sGx9/PavqiNehGhh3u8unn3s3da/hK0c7StwzzmF60SFMh/Qxr+8aPkSd/eorNnWF2Nw1+JqvmPVDEFqhLwnw6LqBjojzZqZb1kT00EnLpkT13F5PNfV0yBmgL88BjReKZ0nS5TsVbWuX1nl8vB9wvxWdGnQPVSoEBRejARzmn5sOW0bWHxLZutKK/qE2+dC/BX5DqrA58ROCaWvacRBxx7Wyq4PtcxINoD1KXd611Pd++9w7vf+5265xoRewDvn1U9PI099fqUVjH+TVp8bPa/50ydPnJw5h2YNVH3s+hVvWziv72mKZD/ylHnSZ+tjhmeMsHylQjG//MpUBoFqgrwk0F3rTLoTu2WZR0f6GKZ/QYog+6BWgnPV1o9/XPVS4GaAPzOL0Sa/WndZ1j7ph0tUxl0nzvwnGz2qz+DbvVSz3+7T4G6T4HCcNgSO5600MTxFp462XKVvbo64vCfBVBL01JX4BwrAJ8VMCZzLVpyfAEIY4ndP2AaQYCNY+z+2fAVYPtB8XmuA75TwXjJYw+4p7v3nvfu936n7j2wf7HKX5oTLnUjFOfntX1d93xH9bui2NyHumIGdRY96z6r+SWsCAUKFTkhvnaVhWfNsPjWLVaoRDz+n0+B0ihQFYAfe+o1yhQ5vvNka3lPaxs6aJi9Kj8SX+taGsUb5rnqBvzwmAvvqsRtre5pYxl9M+3QoUO+E3jDZCe/VnWQAkT7i61dbYF+vS04MEMRAHdbbZj0nAVg0wT/nikQNv6Ervy3jykKkPwB0J6zIRB8WRlusX3n2pe0fVnXv6rzaNG/K18AAHgqIE899oB7unvvWe9+7zd7ViA4f5m+e7HKWxrYL0kH/Bw+p3p9TeVnRQPzn1Kz+dZB3vKLdB4UyD/6kiL09HHe9OGxoxSH/+/n8Tb/0YZMgfMF/F5YxIldp1rzRi2tb69+9jyanzoWZ78ht2FdqFt1An54bN3ojU6z//Dfmlnvbr3t4OGDfjSeutDwfhkuGAqQwDO2ZpUF+vSwQP++ltj9VLUl3ioJZqv6NwAa7TiZbfEHYEMgYON86sY9TosuAQDBAA06YDwVoHPsAfd0997z3uoO90MAACAASURBVP3eb/YAflYaED4oS7r1p17Ov0H7UsE+77pgOPYCqWjBW29ZZMY0B/qjSxZagZKh+H8+BUqjwPkAfoDYBplYzFBYxDZN2ll/zHj271ektvTtOksrk3+u/lGgugC/x2OE3mzfuKP17d7X9u7zQ2/WPw7xS1zfKVAUi1li/16LLl5gUQUFyXvlZZPUnXa1KhOWM12gWxP3AbxxFP4PrQIQAvQKbYD10vwBPBBf3t4D+N493m/2PxLYJ8IQqxFVXre0W8y/sV5QIF/ZHaMy5wkNynSpsAuPH68X5fYLWfMUqCzgx3lyrcx4xikaTxtl0B09fLQdefGI76Bb801YJ75YHYCf8K7rFXpzStdp1lJO4AMzBtpzCu/qO4HXiSb3C3GBUcCZ9KxZacE+3S0oDX+uknoWKYFkun/1HfB7wBvN+WcE/L8o7ftXtV1SbEJzuUA64TKJqOMJAmeA+OLzJQF+6m+EB575toSJf9U30jHl8cqV9j7dBvPvqx8UyH/1ZRc6K6fVI3KumWCF//CddutHy9V8KSsD+F3oTZlYjO0wwR6581Hr27OvvfyytD3+3wVLgaoG/B7Yn9h5irW4s6X17tHbXjjoh968YBnMr3itU8CF5dy43kKD+1toxFCF5XzmggT8qcAacx+Sg6H5/4yA/xe0kR34UgF2VgFcojAJAamgPlUI8AA+59xKge79rjai73gReVK/VyXHtc5JfgGqlAIFf3/HovPnqFMOsejyx6zwxIkqfb//soZDgYoAfsx3tozfrrCIW2ySbPbbNG6nSCkZisazXxnW/UgpDYcrKl6TqgT8m8dvc1map3WbYW21etSzS89iv5CKl8t/wqeAT4GqoUBRJGq5+2TSs2iBxZbKpOflo/UiSk+VgOQ0TWvQyCMEfFoOs2QI/pI2BIBvyPmXDMCXafuWtsuLt29rn7pdqnuJsf8JhdystnJXDTv4b6krFChQGM7YkkUWGj/GedX7ibfqSsvUvXKkC/jnZywywiKSVGtS56nW8cHONnzoCGfG42c3rXvtWtMlqhLAn7nQ8RgC5VTF2e/QpKMNUFKtAwrv6vNYTbeo/z2fAmdSgGh/0eXLLNC1gwV7dbP4zicveA3/uUA5dv9o6p1zsFYBcP5lNeBiVgRK2XC4rYiT7rm+X+r1M5vV/1XfKZAvyTs0MMOymzWx8MSxlv+ub9JT39u0usqfLuBfmLnYVo9YZ+M7TbKHb3/EhgwYYq+/8boferO6GqaevbcqAP8C8Rh+ITN6zLRmdzS3nt172eHDh32wX894wS9uw6RAYURx+BWKMzxzmkXmzrS8I4cvKKfdUsGzAHq9O98w2fPCrVXB++87zX54dpbFtykO/6lTFy4x/JqXS4FzA/6t1r/5IJvbd4FN6DLZWt/XVjHQ+7vspuW+2L94QVGgVMBfOMaet3H2+NEeNmxMJ1s2eKUFtkQsd1ehxXbkWWJXgYW3x22THMAzHhngeGxajyxrc09b66fwrs+9QFKt9KOAXFAE9yvrU6CGKVAYjVjimT0WmTPLIgvmWt5LR3yTHh/w1zAX+p87iwL5b74hZ92JipfbzaLqmAgA/p9PgdIokAT8a2z54FUW3B61vN1FDoixDz+RsM2y2e/7cKYNajHE2jZpbyOGjLAXDh30ta6lEfMCPgfgHzC2vc19poM9Z2PtBRuv/Tg7ZBNsxas9bfi4zrZ8yGoLbYtZ/m47zWPRJ3Nt64Qnrc/DGTa45VAX8SmjV6YL71pY5IP9C5il/KrXMQq4KD2rV1igZzcL9utliad2+iY9PuCvY1x6ARYnX6YW4UnjLNC9s0Xnzhbgf+8CpIJf5XQocOLkCVs2e5nN6TXf3l31vn24MWDH12fbSe3fW3vSHhv4uN39p3vtrpsa2egRo+0VRYCqygy6+QrrFovGLDc3N53i+vfUUQps2bzFegxqYSNXPmjr/t7PNhzLtPXvZtim9/rbtB2trFdmc5vbe4EdW/mBnSrmMXjt/TUnbYUEgSY3PWB33HCn9dfqEaE3K2KzT5jObNkX5+TklJp5Nx6PWzgctrw8Pz9EHWUfv1j1gAJF0vDn7n3WafejixdanhJ8Xkhx+Oud6U5Zwkg94DW/iBWgAAA/pni5kZkzLL7VN+mpAOkuuFtPZZ+yxbMXW2aLAbZ02HJbNWatrRi9xlaNXmfLR65y52/589+sa9euAvuvVDl9Tsnc7KWXXrJjx45V+bv9F9YcBfY+u9cyhvSwzkOb2ogF7WzUog42cmEHG72ko/Ua95C1aNXYMlr0t2XDV9jqsWfy2MBWg+3WG26zjh07umg8RRXM0hyNRm3Xrl22d+9eB+xL1vq1115z1z7wExCWJI3/26dA2hQoisikRzb8EeX4QZHobPglbKf711Di8Nd74J9ug/n31Q8KFCjxViRrmgUH9LXo0oVW6E909aPhaqGUsXjMnt71tE2bOMPmTptn87MWaFtoC7TNm77AJo2ZYlnTs+zIkSPVYk998uRJO3jwoL3zzju1UHv/k1VFgeNK7rd56yabPW+mzV88xxYsnuu2hUvm6lyWTZ400aaMnyqemu94Cx5z2wzx2OhJNmXyFOegWxmbfQD/7t27bd++fRYRKCn5d/ToUScQ+EJlScr4v30KpE8BTHqiK5dboEcXJd/qIbO8HRUy6Vm1apX97Gc/q39OrmVpyuvr+fSb3L+zPlAg/623LDxtirOziyyabwU+4K8PzVYrZcQ8J5FIWCAQKHMDUFUGiKVTIcwxMLnwzS3SoVbdvQf+gI8A3JHwmRvmNMFgsEz+whSHeyrLYx4P8/3SzM04Dw9jPub/+RTwKVA5ChTFZHp56JCsB1ZZbP1ay5OvYEVMegD8V111lQ/4a1tQqFzz+0/VVQoQlSf36actpqx4uc/tt8JQqK4W1S+XTwGfAj4FfAr4FPApUMcpUCilUHzDegsOyrTgsCGWeLZimXa3bNliN954o33lK1+xSy65xC699FJ/qwUa/FMd5zO/eBWkQP7rr1l47EgLdGore7vplv/ePyr4Bv92nwI+BXwK+BTwKeBTwKdAkgKFWqWLb1pvocEDLDR8qOUqRGdRfvo2/JhtrlixwiZNmmTTpk3zt1qigQ/4G1iPLnjvPYvLaTcqsJ/YutmPw9/A2tevjk8BnwI+BXwK+BSoSQoUyuwuseMJJfMcZ+Epky33heetqAJOu575pjP7w/TP32qFBj7gr8leUwPfwmY/vnGdRRbOs8ST242U2P6fTwGfAj4FfAr4FPAp4FOgMhQAR8SWLbVAxzYW6NpR2GJbhZx2K/NN/5mqp4AP+KueprX6xvw3ZNIzYazl9OhskXmz/MRbtdoa/sd9CvgU8CngU8CnQP2mQGF2jsUUpSfYu7uFBvSz3D27fcBfD5vUB/z1sNHKK7KLw79iuZJvjXce9YUnTpR3u3/Np4BPAZ8CPgV8CvgU8ClQJgWcDf+2bRZSUs/I7JmW99KLFYrSU+aL/Qs1SgEf8Ncouav/Y5j0kAkvmNHbwllTLf/tt6r/o/4XfAr4FPAp4FPAp4BPgQZJgYLjH1hE4b6zG99h2Y80ldnwelM85QZZ14ZcKR/wN7DWzVd83NDIIZbTrKmFRg+z/NeUIVXx1v0/nwI+BXwK+BTwKeBTwKdARSkA4I+teNzCo4a76H95Lx42q4DTbkW/599fPRTwAX/10LXW3kpGvNx9z6hzPmaR+XMUi/+AFSn5jP/nU8CngE8BnwI+BXwK+BSoCAWKlLSOnD6hoQNkOdDLhecsUsI8/6/+UcAH/PWvzc5R4iLLf/UVCw7MsJyWzSwyc5oVfPjhOZ7xL/sU8CngU8CngE8BnwI+Bc6kAOE3sdkPjxvlMEVswVxDsej/1T8K+IC//rXZOUucd+ig5RA+q31LOdjMsLzDh/yMu+ekmn+DTwGfAj4FfAr4FPApkEoB/AJja1dZZNECi86ZZXkk3YpGUm/xj+sJBXzAX08aqiLFLAoGZNaz12LLZdaDV/20yZb3imz5/T+fAj4FfAr4FPAp4FPAp0CaFMh/9VULdO9kOZ3bWe6uHb7tfpp0q4u3+YC/LrZKFZTJedVLu5/z0H0W7NfLCQBF0agceKvg5f4rfAr4FPAp4FPAp4BPgQZLgcJAQIrCly33qZ1OcZjTvpXF1q+xoni8wda5oVfMB/wNtIWJm5vY/ZSW4la7EFrRhfMtvm2LFSmltR+1p4E2ul8tnwI+BXwK+BTwKXCeFCiKxVzAD5x0c1o9YrENay3/9ddlGhz04++fJ21r83Ef8Ncm9avz2wWFRqctOH7cIuNHW84j91t03myF6XzVyMbrS+nVSXz/3T4FfAr4FPAp4FOg/lGgSPH1E09st9hjSyw8dZLldO1gkbmzrTA7u/5Vxi/xGRTwAf8Z5Gh4PwjJGVuy0CLquOzDoxVHd8ZUK3j/g4ZXWb9GPgV8CvgU8CngU8CnQIUpgIIw/+/vOmzgcvm0bmZRYYbEnqeT4b2xDvD/6jUFfMBfr5svjcIrpFbhyRPS6r9hkSkTLOfhxhaePN5iq1dabOUKLdNJ2x8KWZGkd8x9Ck+dkvY/5sJuFSrWbmE4bAW6hnRfyPUPT+r+oHvG3SsH4UI9X3jqQyuU5z52f2x48RdxzL2sNGTrXZGw7omaey/vKn431905hQ/lnHtX8bdPf4/nKGMi7t5JWDBWKQqDKrvudedVdp4vcu/QMfUKF5fZfUPl4b18W+WmfK6u2bq3uDyFOcV0oKzQori8BV55+Abf4xsnTybrK/OpQsrO9zimnNGYO3a0opzuGV2PUX9dp5yp31BZCorLX6DrjuZ8m29AQ76n865+0FnfK4on9H19T990NKQdqBc04jp117OOnmF9m+u8z9VJ7+Je0cW1E3T0aMy+mIaOJ/gG7c85aOVomKSPawfeC63Ye8e8A1rAA6KF+5ZXNve8+IN6FNPS0UK85doBOvAdlZn3unfTftQJ+tFWPMc3OHZ8kVA91LbQgvPQimOuw7OitSsf7+Mc5YIneS/PuXLyXHHbco+7V/X06u3Kq/dyT/F3C08l29J9Q/VzdVbZXXu553Sd9uBacV9w5eIYvtA33LuoKzSGfpRF7eTqynX3vSTt4S9XP/W3oojK631DdHH107PuvdSJtocWxbzp7v2QsukblId3Q0++x3Pcy15lcPxEn1Y93XuhMd+iTaEL99JGAc5Tdh2rnI62fM/RW/c6XhcteZZ65MKzuk+/3Tf4NseptOA69Yd/9UwBdYX2lMcrJ7SibMXt4N6l99Bv3DHfoqxcVxu5+lJm3sX3uMY7OIbmlJ/r8A1003lX3tP05rrexTX4huv0C2jBN6gzdXXv0t77NjwL/4tGyfFG9ece+kJurtq2mN7c4/ENz1J/3gt9XP2jyfGBenl15h3wAO1I2UOUWeVXOV1biY6OrtTNlUfXqTPPUX73DWihb3jtpHHAtSP9opg+rt7FdXPf9upMO3n38F3ax7WTvkcd3fimc9DYKzPjja7RV6l/si/otxvLVA74hrqcOK7nVA/V19GW8vItxgjXjtSp+Bte2YrPu/anvtCO73LMexyNi9vIvU90ox+6cup9lN27lzLzDOVMoYUrG+XgGfj+NC2K61/8LtfWfJv21zu9d9AejkfgY+pMOWhTKeZc21F/zvFeysKx2smVje9y3rVzMe299oem3CvedHu+zXv5Hu/he8XnTvch3iXeKHj3XSNhZ97zz1lIibWii+dbZM4MC/TtpuMF7jkr9J3/0kBbdf4WH/DX+SaqmgIyeMSXP2bRGZMtvmGdBTq2tpwOrQX6l1t8/RqLb9lseQf2W2zpIst78YjF1q12QkHiqV0uwx4Rf5D0o0rmFd+yyfkDxDQYxDdvSC7/LZxniWee1nNr3Ja7f598BzZYfOsmyzty2L0jsWe3JZTAI7p6he7dY4ldOy32+GOW+/zzFpevQXThXON70cULLbZK98gHge8lntiqew64qEN5MklKPLld715n+S8ftcSOJyyxdo0zVYotW2y5em+uvhNbtkTv3+F+RxU3OO/IEX1D9dR3iGAU36Sy6R25L6ps3KuyJ3arrvou76CusaWLLe+F55PC0epVFtf16JJFSbqo7IQoi69f5+jh+Ugk5CcRfXypnnvBElu3qOzb5Ph0NFnnXU9aLu8TzV3ZKKfonbt/vysP5YjLQSom+pA4LVcDcMzVf5vlPv2UxVctV9mkbTmwzz1XICEusW3raRondooW27c4WkQXzbfcZ58Rrba5lZ2E3pvYvlVmXaIn7axvQIM8aL9+rcVUbr4XXaaya5/Q96hrLnTYsN7VybWNlnld+bmuMG34iNAeMR3z/viO7a79HI23bNSzorHqDJ/wPTI/8+3Es3ssTzzC865OShbnyrZXSePUjvEVy1QGtYfKE3tMZTp8SPwxT/uDopfaT3xG2+Tuf9bxFhNWQnxM/Sk/tIqLNrnP6voCJaCjfSmn2hTegzejEnjhPcpAORMqD+fgS+6JLhVfwLOqq+MRrqsfuPrDp5RNW2LnDovMEy/AU7Q//KRzjobLl1me8mJwjbagHvGtm5PlFP9SHvoK7RNf+bh7Lg7d5otnVdfYYyovfLqPeoom1O/QC65NSYYDz8ZFT84RQYOU9zja0aYsy+eqrVmOd/1EZaLN4BPKQz+jzeMb1P6iJ3zF8/Htek6hfaOL6NPqC+LhmOqR2PGkxTZv1Pn5ltir8qieri3VB107wdfiOdop8bTotl3Pef0Jvlab5L/5ur6hsuk3sb3jep87fvVlV9e8fSqD2oI+myseiSscYJQx6XnxrLSNlJuww2T9dDxEOXQMXfmmOz76ksXWrFL/E51VHsrm6A+/0S/oS/quazu1I3wbe2xRsl+ovIxv7l3uWOPiQdF7sa7redrOfVs8Ele/iKmPJJ5RnVVeN2bBYxrT4ptUr73ioQXiWcaC3eIp9YF8tU1i55P6xkYreOcd8exa1275GiPhecen9CH1e/qQVzZXBvFCXPVK8sLjrk7cAyhzPLRTYxI8pP6RK1ow1uWJFvBxfJtoAc+6sokXxSuR2Vmqp/hH5YrxnMZb+gz1o/1jGsfcuOF4T9dVN97LuO/GR/iGeYH+IVpFl0obrG/TL+KiKWMqZeMdufRr1Y92yqX+erbg2LuOR3LFlyieoAnvpV9Hs6a69k0wn6j9oR987viCMZSxUmWlf8ZWi+9UtjwlmYSn6Qu0oevTlF08HVu1Mjn30J/UL+ifmKxAl1zRxY1fKiP8n6SF6MK4sGala3/GdMenzF/UWd+Ni+9dX2c+Wa/5a7H4XuejtJPmhVzxJuVlTqB9GdPJkZMrmidUV+rM+EpfLXjrreLxfbfCXmrMdTTU/ME4O3eWeFO8qzZyPMu8uUZloxzF7R9XX0vsSs7ViR0ae9SPGKfyXnxR39p8ujy0Uwyeo2xLNH7rnuisLK38D1O5lxihvHO05WpcyXvpiKOl5fqJO6sGhdX+W3zAX/ttUDMlKJRNf0BagePva2BU1ryB/dzAxIAc6NXFIlnTJBAss5w2zd2ES2SfQK9uzoYv0KW9BTq3UxKv6Zb9YGMLDhtkoZHDLLvFwxYc3N/CY0ZZ9sNNlXI7S1oBPdevtyb9BRYc1N9Cwwc7cBbo0sHC+kZEk3ZOD31vjgYZrTgEOrXVIPaYhfr2cH4Gzmaw9SMW6NlF90+17Ie0IjFupCYfDUZd2jngEpk41sKDMy1X4CcyaayF+vd2kwhCTEx+ClFNZMFObSw6fZL7ndP8fksIROCAxBaTA3NoiLIGDurnBr+ABB8mv8j0KRagbLNnaZtpOW1bOJoEenZN0mLaJMtuqbL16mqhSRMsu+k9FszsYyHRI7v5gxYaMdRpSHLa69vLH3c0Co0d6QZ86hyeNN4ijy9TiLPOqvt4C4ueOW0ftYgmqdDQgebKoW/wfmjOpJbTrImFx492CdSCokls7kxNVAst2K6F5Wlyiej94ZFD3AQa0TsjY0c4WuS0eMiBhvB4JUtp84jMuKZYaMwIR08AFN8IigZMzsH+fSw4cqgDKjkdVHadi4j2OWoH8jgEMvu6vA6ubUSTQFe1pczCch59yIJ9elho/BjL1jHaodAEHbdSwjcBVujCu5mYgqpfSLzCxEzbRjSJOd7r2zN5rImR1SeyQwc6tLKgvsGkHxR/BJRTIiEBNKfFgxLuVjnAQZs7UL9kvgW6dXRCXkjfov5xTazBHqKxaIPfSk4z8eacmaLjGMtpp9wUEtSCgzJdOzCBBlSHQP++DgwHuqudJk+UMJds/8is6c4Ezj0nsJjdvrXq31H3yLZVtEJoDqldsx+4172T1PPcG544Tu07zYWyA2CExK+0IyAiJFpjWgd4zFE7AmQRxIMKfUefCI0YbAHxLPfC06EhmarzfF3Xd8fqOYGQgPgmKsGIdgqqLaNrVij87iTH1wjIwQF9XX0BpKfuv1e8pz6jMuW0fNhC6j/BIf3Fy2ongUHaKKeP2kFjAXQJqT8DBLKbP+DoFhozUvVob+EJek5tmP3og67NqDv9hfaHFoFunST4zLbsR+4XzabpPSOT/UltTdkI7QeQCA7KUF+bJAC8UePDIPXTyU4Yoq4IJJQhR9+IzZ9tocyejtfj9H/VGZ4C9PKusAQZAFWO2j++dqXrI/BCQmAv0Ke7ozN8Rjkpd4T2E5+GZ4pmQwcJ2LTW+DY52WfF14BFeBNehb/5RpDxC1q0Ul9YpH6n3zmMZVMmJpMb0rfEK9Q5SB8XLRjT+F5U5c9upnYUgI6ojqFhAwVwtyT7vt6TJ3AcGiCenTZR49Na174xCY6RKWqnR5o6voio/9If4gK1gLFgvx4Cp/Mcz9N/I9RfbQrdKBP21ghkUfEtfTMhwSO7TQvXv8NTxLMdVTY9B8+eatzIwtOmOF7I6dzBQgrfTB+mfghn0A1TUPqyq7P4F7pka2yJzJohvumbnBfEg9At0La5+za8EBo9wtGFMSSstkeAYUxLCBgzToUGZlj+wRcsomROUV3PFaAPi+8ZuxF8s++51dGPfh5o38LRL9i7qzGWI3y6MUJ9iP4ZUF8PqWxRCc/MSbQX7RTUHAd9AvoWYzhzT3Yb8ZD4OzxZq92KOsO4B114DsEY3oSGjHUBjU3MgQhAOWqD8FSNC8xfzIXQV/0hp4PGE80nwUw5t7YUXYiMx/il/se3s9WOrq+rfQNd21suwtzs6RYeNUR1ltCheTA0sK/lC4AzZvE7Kh5g3AtrvCZxZvaD97jvMj4wFtImtC1zMgIl7R8UvUO0r9ouNGGc44cc9UeE0tAw0UJt5dpR5wKas8LqF7Qj85dzzNX4RV9BwKGvIijy5+v1awae1dRX/qmmPuR/p+5QgKXD/KMvWsF777nBJSBwHF8urYw0MDj3MqECLh1IZ1AD5AFMNSFk33enm+gAqIBxB3gFdLOb3K1BWQO/Jj0G1+gcDZ4ZAsMDMwXMVlgAEMTkwQClwQygEhY4DEjAAPCENNjlAEwAJSoDkybg55S+FwaMSntEGdB4RcZqwNaAmti6SRPGSAtp6dEBSd6lATeqQTLYToBIE2dME0jOA3cLEKhOvTsLHHVxIDCk54MZPQ3TpgDAVhNZRJMA32WQdYBXIB4NKwO8o4XKduqhJjpuZ6Gxo92kFOzdTZN2hgN8oUGaSAUksls84rQlCBWh4QLj0koy0ZKpMKJJifpT97DoCbBhYkAAc+VQ3gQnYGlyhy45Te9y2hcAUlDfZSJ34E+TS540Z5FRmrCU8twBnokSDPQ7V5rknIfvUxtIqBo1zE2SmHMhfGU3vtNpYJl8mJSYwIJ9ujkQgLY5p5UmbYHEiCa3HIEVJwT17ObOu7ZReQEegOKcBxXyVQArNGqEndIxYIyJPvvhJpoQNZGJLrwbASM4oJ8EM6VlF8h1wuOsaZqsBAJ6CCRK8IlKkMlpcqcmTLVXKwlVmkijKkNQbR4QqALwEGKWyZxMj7Q5K1axhZrQBRQJHRfq293VPy4gHOwsGosW0CunaSPXpuExAjOAFQmeTHJMngBigB0AkRUGx3fjxzphI0ftD89HAFICimjxs1s0Ey9r8tc9lBNAExI/Zt97h3snkyv3MlkDEBDo0DSHBAiY+NGwkaI+LKDr+ELlYQUiKhCEkApfIBwEHrzXCTTwBKAwtkD9qZMA6kgA6OMWkPCDwEQ7BdWWUfVdcm4AHtBaBvt1VxmGu7Y+dfftjvfCAHcJVSEJRYD87GYSsAArqntOt84OEAQzegsoD5Hwt1Q8fY+jVWjYEAmNLVR+PZfZWwDkPtdmrp3UN2hzdyweBwhm36/nRDMnuLi+rv4EXwtAolmG9pHJ45ymPzQ4w9UdIEw7Awz5Bm3twFBvAfBWD1mc/t9SQrWEATSdAWjh8Y0AVmyl6i+g4sDxRq1gwpeiM22d06qZq68TRlR22iUo3qTM4YkCn8VtFp0vIN2pneNVAGFAwDWo8QuhMBuhkbIBCBnLBKqC/Xom+5a+kX3/3UmeU7sHWj/q6hCFFhoXEdwjkwG5/Zzm2fV9CVysYITUTlEJY/Rf2p/xKzJBQRY0ZsHbkdFDRZdmhtAcANj16iQlxizH8w7EIZiqvyHYhnlOfZNjhKigFAuslmQ/8oDr3whsOa2bO/AXGjfaTjW61SkhnMDXpqUEQWghQaitxh6tBLhM7RrzXJ/UOYRZ+j1KnwiCgvgmoLkBHnRj1qMPOJDu+sTwwUnaaywPq+0ZiwISbuMabyPwuHiAlbgImmXxAlr/sGgSlTDGXJR9x181JmgMUT8PAKRFv2C3DqKLwK/G5EBrtakEKASpgPp6SH0LQYc5ibIjNCIcoVgKaM+YDeCFFihmEL6pX3Cw+qLaMlvjeiRLIF7lSrav5iaBYwRdNN/cG5ag7OYvCQ2MC0G1dU4rBAUB9N4SRlBYoAiB36TAcuOF+CI8Qn1dZ6xxOwAAAHdJREFU7cu4CZ9HpXwJq+y5O59w9Q1l9LC8Z59W+7fVuKNxUQJBsF1z19fh6ez7blc/nuv4w80RCOwS/p3Q79q/WAlH+4o3EbQZyxDYWL1kDqatXDuisNCcFVZd6cfM446GCP8S0gtOnEia/yR8rX7dQWxVV5L/H9HfJLh2Ucc1AAAAAElFTkSuQmCC)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jjlFHBL20GZa",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#from resnet import ResNet18_OS16, ResNet34_OS16, ResNet50_OS16, ResNet101_OS16, ResNet152_OS16, ResNet18_OS8, ResNet34_OS8\n",
        "#from aspp import ASPP, ASPP_Bottleneck\n",
        "\n",
        "class DeepLabV3(nn.Module):\n",
        "    def __init__(self, model_id, project_dir):\n",
        "        super(DeepLabV3, self).__init__()\n",
        "\n",
        "        self.num_classes = 20\n",
        "\n",
        "        self.model_id = model_id\n",
        "        self.project_dir = project_dir\n",
        "        self.create_model_dirs()\n",
        "\n",
        "        #self.resnet = ResNet18_OS8() # NOTE! specify the type of ResNet here\n",
        "        self.resnet = ResNet50_OS16() # NOTE! specify the type of ResNet here\n",
        "        #self.resnet = ResNet101_OS16()\n",
        "        #self.aspp = ASPP(num_classes=self.num_classes) # NOTE! if you use ResNet50-152, \n",
        "        #set self.aspp = ASPP_Bottleneck(num_classes=self.num_classes) #instead\n",
        "        self.aspp = ASPP_Bottleneck(num_classes=self.num_classes)\n",
        "    def forward(self, x):\n",
        "        # (x has shape (batch_size, 3, h, w))\n",
        "\n",
        "        h = x.size()[2]\n",
        "        w = x.size()[3]\n",
        "\n",
        "        feature_map = self.resnet(x) # (shape: (batch_size, 512, h/16, w/16)) (assuming self.resnet \n",
        "        #is ResNet18_OS16 or ResNet34_OS16. If self.resnet is ResNet18_OS8 or ResNet34_OS8,\n",
        "        # it will be (batch_size, 512, h/8, w/8). If self.resnet is ResNet50-152, it will be (batch_size, 4*512, h/16, w/16))\n",
        "\n",
        "        output = self.aspp(feature_map) # (shape: (batch_size, num_classes, h/16, w/16))\n",
        "\n",
        "        output = F.upsample(output, size=(h, w), mode=\"bilinear\") # (shape: (batch_size, num_classes, h, w))\n",
        "\n",
        "        return output\n",
        "\n",
        "    def create_model_dirs(self):\n",
        "        self.logs_dir = self.project_dir + \"/training_logs\"\n",
        "        self.model_dir = self.logs_dir + \"/model_%s\" % self.model_id\n",
        "        self.checkpoints_dir = self.model_dir + \"/checkpoints\"\n",
        "        if not os.path.exists(self.logs_dir):\n",
        "            os.makedirs(self.logs_dir)\n",
        "        if not os.path.exists(self.model_dir):\n",
        "            os.makedirs(self.model_dir)\n",
        "            os.makedirs(self.checkpoints_dir)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BZmj3OkL6brT",
        "colab_type": "text"
      },
      "source": [
        "###Train"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hDSBaVoj6dpp",
        "colab_type": "code",
        "outputId": "d12131cb-b3c7-4e60-cd33-ea6d4589313f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "\n",
        "\n",
        "# NOTE! NOTE! change this to not overwrite all log data when you train the model:\n",
        "model_id = \"18_B4_LR_00001_W2\"\n",
        "\n",
        "num_epochs = 150\n",
        "batch_size = 4\n",
        "learning_rate = 0.00001\n",
        "workerNum = 2\n",
        "#network = DeepLabV3(model_id, project_dir=\"/root/deeplabv3\").cuda()\n",
        "network = DeepLabV3(model_id, project_dir=datasetPath).cuda()\n",
        "\n",
        "#network.load_state_dict(\n",
        "    #torch.load(\"/content/drive/My Drive/U_Net_Datasets/training_logs/model_101_b3_w2/checkpoints/model_101_b3_w2_epoch_141.pth\"))\n",
        "#network.load_state_dict(\n",
        "    #torch.load(\"/content/drive/My Drive/U_Net_Datasets/training_logs/model_101_b9_w1_from_148_LR_00001/checkpoints/model_101_b9_w1_from_148_LR_00001_epoch_22.pth\"))\n",
        "\n",
        "train_dataset = DatasetTrain(cityscapes_data_path=\"/content/drive/My Drive/U_Net_Datasets\",\n",
        "                             cityscapes_meta_path=\"/content/drive/My Drive/U_Net_Datasets/meta\")\n",
        "val_dataset = DatasetVal(cityscapes_data_path=\"/content/drive/My Drive/U_Net_Datasets\",\n",
        "                         cityscapes_meta_path=\"/content/drive/My Drive/U_Net_Datasets/meta\")\n",
        "\n",
        "num_train_batches = int(len(train_dataset)/batch_size)\n",
        "num_val_batches = int(len(val_dataset)/batch_size)\n",
        "print (\"num_train_batches:\", num_train_batches)\n",
        "print (\"num_val_batches:\", num_val_batches)\n",
        "\n",
        "train_loader = torch.utils.data.DataLoader(dataset=train_dataset,\n",
        "                                           batch_size=batch_size, shuffle=True,\n",
        "                                           num_workers=workerNum)\n",
        "val_loader = torch.utils.data.DataLoader(dataset=val_dataset,\n",
        "                                         batch_size=batch_size, shuffle=False,\n",
        "                                         num_workers=workerNum)\n",
        "\n",
        "params = add_weight_decay(network, l2_value=0.00001)\n",
        "optimizer = torch.optim.Adam(params, lr=learning_rate)\n",
        "\n",
        "with open(\"/content/drive/My Drive/U_Net_Datasets/meta/class_weights.pkl\", \"rb\") as file: # (needed for python3)\n",
        "    class_weights = np.array(pickle.load(file))\n",
        "class_weights = torch.from_numpy(class_weights)\n",
        "class_weights = Variable(class_weights.type(torch.FloatTensor)).cuda()\n",
        "\n",
        "# loss function\n",
        "loss_fn = nn.CrossEntropyLoss(weight=class_weights)\n",
        "mIOu=[]\n",
        "epoch_losses_train = []\n",
        "epoch_losses_val = []\n",
        "for epoch in range(num_epochs):\n",
        "    print (\"###########################\")\n",
        "    print (\"######## NEW EPOCH ########\")\n",
        "    print (\"###########################\")\n",
        "    print (\"epoch: %d/%d\" % (epoch+1, num_epochs))\n",
        "\n",
        "    ############################################################################\n",
        "    # train:\n",
        "    ############################################################################\n",
        "    network.train() # (set in training mode, this affects BatchNorm and dropout)\n",
        "    batch_losses = []\n",
        "    for step, (imgs, label_imgs) in enumerate(train_loader):\n",
        "        #current_time = time.time()\n",
        "\n",
        "        imgs = Variable(imgs).cuda() # (shape: (batch_size, 3, img_h, img_w))\n",
        "        label_imgs = Variable(label_imgs.type(torch.LongTensor)).cuda() # (shape: (batch_size, img_h, img_w))\n",
        "\n",
        "        outputs = network(imgs) # (shape: (batch_size, num_classes, img_h, img_w))\n",
        "\n",
        "        # compute the loss:\n",
        "        loss = loss_fn(outputs, label_imgs)\n",
        "        loss_value = loss.data.cpu().numpy()\n",
        "        batch_losses.append(loss_value)\n",
        "       \n",
        "\n",
        "        # optimization step:\n",
        "        optimizer.zero_grad() # (reset gradients)\n",
        "        loss.backward() # (compute gradients)\n",
        "        optimizer.step() # (perform optimization step)\n",
        "\n",
        "        #print (time.time() - current_time)\n",
        "\n",
        "    epoch_loss = np.mean(batch_losses)\n",
        "    epoch_losses_train.append(epoch_loss)\n",
        "    with open(\"%s/epoch_losses_train.pkl\" % network.model_dir, \"wb\") as file:\n",
        "        pickle.dump(epoch_losses_train, file)\n",
        "    print (\"train loss: %g\" % epoch_loss)\n",
        "    plt.figure(1)\n",
        "    plt.plot(epoch_losses_train, \"k^\")\n",
        "    plt.plot(epoch_losses_train, \"k\")\n",
        "    plt.ylabel(\"loss\")\n",
        "    plt.xlabel(\"epoch\")\n",
        "    plt.title(\"train loss per epoch\")\n",
        "    plt.savefig(\"%s/epoch_losses_train.png\" % network.model_dir)\n",
        "    plt.close(1)\n",
        "\n",
        "    print (\"####\")\n",
        "\n",
        "    ############################################################################\n",
        "    # val:\n",
        "    ############################################################################\n",
        "    network.eval() # (set in evaluation mode, this affects BatchNorm and dropout)\n",
        "    batch_losses = []\n",
        "    for step, (imgs, label_imgs, img_ids) in enumerate(val_loader):\n",
        "        with torch.no_grad(): # (corresponds to setting volatile=True in all variables, this is done during inference to reduce memory consumption)\n",
        "            imgs = Variable(imgs).cuda() # (shape: (batch_size, 3, img_h, img_w))\n",
        "            label_imgs = Variable(label_imgs.type(torch.LongTensor)).cuda() # (shape: (batch_size, img_h, img_w))\n",
        "\n",
        "            outputs = network(imgs) # (shape: (batch_size, num_classes, img_h, img_w))\n",
        "\n",
        "            # compute the loss:\n",
        "            loss = loss_fn(outputs, label_imgs)\n",
        "            loss_value = loss.data.cpu().numpy()\n",
        "            batch_losses.append(loss_value)\n",
        "\n",
        "    epoch_loss = np.mean(batch_losses)\n",
        "    epoch_losses_val.append(epoch_loss)\n",
        "    with open(\"%s/epoch_losses_val.pkl\" % network.model_dir, \"wb\") as file:\n",
        "        pickle.dump(epoch_losses_val, file)\n",
        "    print (\"val loss: %g\" % epoch_loss)\n",
        "    plt.figure(1)\n",
        "    plt.plot(epoch_losses_val, \"k^\")\n",
        "    plt.plot(epoch_losses_val, \"k\")\n",
        "    plt.ylabel(\"loss\")\n",
        "    plt.xlabel(\"epoch\")\n",
        "    plt.title(\"val loss per epoch\")\n",
        "    plt.savefig(\"%s/epoch_losses_val.png\" % network.model_dir)\n",
        "    plt.close(1)\n",
        "\n",
        "    # save the model weights to disk:\n",
        "    checkpoint_path = network.checkpoints_dir + \"/model_\" + model_id +\"_epoch_\" + str(epoch+1) + \".pth\"\n",
        "    torch.save(network.state_dict(), checkpoint_path)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "pretrained resnet, 18\n",
            "num_train_batches: 743\n",
            "num_val_batches: 125\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 1/150\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/torch/nn/functional.py:2404: UserWarning: nn.functional.upsample is deprecated. Use nn.functional.interpolate instead.\n",
            "  warnings.warn(\"nn.functional.upsample is deprecated. Use nn.functional.interpolate instead.\")\n",
            "/usr/local/lib/python3.6/dist-packages/torch/nn/functional.py:2494: UserWarning: Default upsampling behavior when mode=bilinear is changed to align_corners=False since 0.4.0. Please specify align_corners=True if the old behavior is desired. See the documentation of nn.Upsample for details.\n",
            "  \"See the documentation of nn.Upsample for details.\".format(mode))\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "train loss: 2.16469\n",
            "####\n",
            "val loss: 1.69487\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 2/150\n",
            "train loss: 1.65215\n",
            "####\n",
            "val loss: 1.39005\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 3/150\n",
            "train loss: 1.47825\n",
            "####\n",
            "val loss: 1.20474\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 4/150\n",
            "train loss: 1.3565\n",
            "####\n",
            "val loss: 1.16247\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 5/150\n",
            "train loss: 1.29846\n",
            "####\n",
            "val loss: 1.0277\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 6/150\n",
            "train loss: 1.22213\n",
            "####\n",
            "val loss: 1.02045\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 7/150\n",
            "train loss: 1.17305\n",
            "####\n",
            "val loss: 0.963467\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 8/150\n",
            "train loss: 1.17454\n",
            "####\n",
            "val loss: 0.957368\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 9/150\n",
            "train loss: 1.10925\n",
            "####\n",
            "val loss: 0.963633\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 10/150\n",
            "train loss: 1.07758\n",
            "####\n",
            "val loss: 0.922322\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 11/150\n",
            "train loss: 1.04545\n",
            "####\n",
            "val loss: 0.885564\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 12/150\n",
            "train loss: 1.02528\n",
            "####\n",
            "val loss: 0.892433\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 13/150\n",
            "train loss: 0.991554\n",
            "####\n",
            "val loss: 0.840224\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 14/150\n",
            "train loss: 0.977075\n",
            "####\n",
            "val loss: 0.847475\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 15/150\n",
            "train loss: 0.977318\n",
            "####\n",
            "val loss: 0.832883\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 16/150\n",
            "train loss: 0.953989\n",
            "####\n",
            "val loss: 0.841769\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 17/150\n",
            "train loss: 0.929672\n",
            "####\n",
            "val loss: 0.803534\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 18/150\n",
            "train loss: 0.926482\n",
            "####\n",
            "val loss: 0.785622\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 19/150\n",
            "train loss: 0.914631\n",
            "####\n",
            "val loss: 0.769448\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 20/150\n",
            "train loss: 0.90946\n",
            "####\n",
            "val loss: 0.790204\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 21/150\n",
            "train loss: 0.904336\n",
            "####\n",
            "val loss: 0.77844\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 22/150\n",
            "train loss: 0.893664\n",
            "####\n",
            "val loss: 0.769413\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 23/150\n",
            "train loss: 0.901306\n",
            "####\n",
            "val loss: 0.744824\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 24/150\n",
            "train loss: 0.868098\n",
            "####\n",
            "val loss: 0.780674\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 25/150\n",
            "train loss: 0.83556\n",
            "####\n",
            "val loss: 0.730849\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 26/150\n",
            "train loss: 0.847022\n",
            "####\n",
            "val loss: 0.741201\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 27/150\n",
            "train loss: 0.837395\n",
            "####\n",
            "val loss: 0.730265\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 28/150\n",
            "train loss: 0.82836\n",
            "####\n",
            "val loss: 0.744311\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 29/150\n",
            "train loss: 0.824203\n",
            "####\n",
            "val loss: 0.774123\n",
            "###########################\n",
            "######## NEW EPOCH ########\n",
            "###########################\n",
            "epoch: 30/150\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7LnKitgx4dCi",
        "colab_type": "text"
      },
      "source": [
        "###Evaluation"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0sV3FTGsbz_G",
        "colab_type": "text"
      },
      "source": [
        "###### Network Summary"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Njg6oGbTa66I",
        "colab_type": "code",
        "outputId": "83709f9b-d5cc-45d9-ef67-6097a65a28cc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "from torchsummary import summary\n",
        "network = DeepLabV3(\"eval_val\", project_dir=datasetPath).cuda()\n",
        "#network.load_state_dict(torch.load(datasetPath+\"/pretrained_models/model_13_2_2_2_epoch_580.pth\"))\n",
        "#network.load_state_dict(\n",
        "    #torch.load(\"/content/drive/My Drive/U_Net_Datasets/training_logs/model_50_b3_w2/checkpoints/model_50_b3_w2_epoch_128.pth\"))\n",
        "summary(network, input_size=(3, 1024, 2048))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "pretrained resnet, 50\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/torch/nn/functional.py:2404: UserWarning: nn.functional.upsample is deprecated. Use nn.functional.interpolate instead.\n",
            "  warnings.warn(\"nn.functional.upsample is deprecated. Use nn.functional.interpolate instead.\")\n",
            "/usr/local/lib/python3.6/dist-packages/torch/nn/functional.py:2494: UserWarning: Default upsampling behavior when mode=bilinear is changed to align_corners=False since 0.4.0. Please specify align_corners=True if the old behavior is desired. See the documentation of nn.Upsample for details.\n",
            "  \"See the documentation of nn.Upsample for details.\".format(mode))\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "----------------------------------------------------------------\n",
            "        Layer (type)               Output Shape         Param #\n",
            "================================================================\n",
            "            Conv2d-1        [-1, 64, 512, 1024]           9,408\n",
            "       BatchNorm2d-2        [-1, 64, 512, 1024]             128\n",
            "              ReLU-3        [-1, 64, 512, 1024]               0\n",
            "         MaxPool2d-4         [-1, 64, 256, 512]               0\n",
            "            Conv2d-5         [-1, 64, 256, 512]           4,096\n",
            "       BatchNorm2d-6         [-1, 64, 256, 512]             128\n",
            "              ReLU-7         [-1, 64, 256, 512]               0\n",
            "            Conv2d-8         [-1, 64, 256, 512]          36,864\n",
            "       BatchNorm2d-9         [-1, 64, 256, 512]             128\n",
            "             ReLU-10         [-1, 64, 256, 512]               0\n",
            "           Conv2d-11        [-1, 256, 256, 512]          16,384\n",
            "      BatchNorm2d-12        [-1, 256, 256, 512]             512\n",
            "           Conv2d-13        [-1, 256, 256, 512]          16,384\n",
            "      BatchNorm2d-14        [-1, 256, 256, 512]             512\n",
            "             ReLU-15        [-1, 256, 256, 512]               0\n",
            "       Bottleneck-16        [-1, 256, 256, 512]               0\n",
            "           Conv2d-17         [-1, 64, 256, 512]          16,384\n",
            "      BatchNorm2d-18         [-1, 64, 256, 512]             128\n",
            "             ReLU-19         [-1, 64, 256, 512]               0\n",
            "           Conv2d-20         [-1, 64, 256, 512]          36,864\n",
            "      BatchNorm2d-21         [-1, 64, 256, 512]             128\n",
            "             ReLU-22         [-1, 64, 256, 512]               0\n",
            "           Conv2d-23        [-1, 256, 256, 512]          16,384\n",
            "      BatchNorm2d-24        [-1, 256, 256, 512]             512\n",
            "             ReLU-25        [-1, 256, 256, 512]               0\n",
            "       Bottleneck-26        [-1, 256, 256, 512]               0\n",
            "           Conv2d-27         [-1, 64, 256, 512]          16,384\n",
            "      BatchNorm2d-28         [-1, 64, 256, 512]             128\n",
            "             ReLU-29         [-1, 64, 256, 512]               0\n",
            "           Conv2d-30         [-1, 64, 256, 512]          36,864\n",
            "      BatchNorm2d-31         [-1, 64, 256, 512]             128\n",
            "             ReLU-32         [-1, 64, 256, 512]               0\n",
            "           Conv2d-33        [-1, 256, 256, 512]          16,384\n",
            "      BatchNorm2d-34        [-1, 256, 256, 512]             512\n",
            "             ReLU-35        [-1, 256, 256, 512]               0\n",
            "       Bottleneck-36        [-1, 256, 256, 512]               0\n",
            "           Conv2d-37        [-1, 128, 256, 512]          32,768\n",
            "      BatchNorm2d-38        [-1, 128, 256, 512]             256\n",
            "             ReLU-39        [-1, 128, 256, 512]               0\n",
            "           Conv2d-40        [-1, 128, 128, 256]         147,456\n",
            "      BatchNorm2d-41        [-1, 128, 128, 256]             256\n",
            "             ReLU-42        [-1, 128, 128, 256]               0\n",
            "           Conv2d-43        [-1, 512, 128, 256]          65,536\n",
            "      BatchNorm2d-44        [-1, 512, 128, 256]           1,024\n",
            "           Conv2d-45        [-1, 512, 128, 256]         131,072\n",
            "      BatchNorm2d-46        [-1, 512, 128, 256]           1,024\n",
            "             ReLU-47        [-1, 512, 128, 256]               0\n",
            "       Bottleneck-48        [-1, 512, 128, 256]               0\n",
            "           Conv2d-49        [-1, 128, 128, 256]          65,536\n",
            "      BatchNorm2d-50        [-1, 128, 128, 256]             256\n",
            "             ReLU-51        [-1, 128, 128, 256]               0\n",
            "           Conv2d-52        [-1, 128, 128, 256]         147,456\n",
            "      BatchNorm2d-53        [-1, 128, 128, 256]             256\n",
            "             ReLU-54        [-1, 128, 128, 256]               0\n",
            "           Conv2d-55        [-1, 512, 128, 256]          65,536\n",
            "      BatchNorm2d-56        [-1, 512, 128, 256]           1,024\n",
            "             ReLU-57        [-1, 512, 128, 256]               0\n",
            "       Bottleneck-58        [-1, 512, 128, 256]               0\n",
            "           Conv2d-59        [-1, 128, 128, 256]          65,536\n",
            "      BatchNorm2d-60        [-1, 128, 128, 256]             256\n",
            "             ReLU-61        [-1, 128, 128, 256]               0\n",
            "           Conv2d-62        [-1, 128, 128, 256]         147,456\n",
            "      BatchNorm2d-63        [-1, 128, 128, 256]             256\n",
            "             ReLU-64        [-1, 128, 128, 256]               0\n",
            "           Conv2d-65        [-1, 512, 128, 256]          65,536\n",
            "      BatchNorm2d-66        [-1, 512, 128, 256]           1,024\n",
            "             ReLU-67        [-1, 512, 128, 256]               0\n",
            "       Bottleneck-68        [-1, 512, 128, 256]               0\n",
            "           Conv2d-69        [-1, 128, 128, 256]          65,536\n",
            "      BatchNorm2d-70        [-1, 128, 128, 256]             256\n",
            "             ReLU-71        [-1, 128, 128, 256]               0\n",
            "           Conv2d-72        [-1, 128, 128, 256]         147,456\n",
            "      BatchNorm2d-73        [-1, 128, 128, 256]             256\n",
            "             ReLU-74        [-1, 128, 128, 256]               0\n",
            "           Conv2d-75        [-1, 512, 128, 256]          65,536\n",
            "      BatchNorm2d-76        [-1, 512, 128, 256]           1,024\n",
            "             ReLU-77        [-1, 512, 128, 256]               0\n",
            "       Bottleneck-78        [-1, 512, 128, 256]               0\n",
            "           Conv2d-79        [-1, 256, 128, 256]         131,072\n",
            "      BatchNorm2d-80        [-1, 256, 128, 256]             512\n",
            "             ReLU-81        [-1, 256, 128, 256]               0\n",
            "           Conv2d-82         [-1, 256, 64, 128]         589,824\n",
            "      BatchNorm2d-83         [-1, 256, 64, 128]             512\n",
            "             ReLU-84         [-1, 256, 64, 128]               0\n",
            "           Conv2d-85        [-1, 1024, 64, 128]         262,144\n",
            "      BatchNorm2d-86        [-1, 1024, 64, 128]           2,048\n",
            "           Conv2d-87        [-1, 1024, 64, 128]         524,288\n",
            "      BatchNorm2d-88        [-1, 1024, 64, 128]           2,048\n",
            "             ReLU-89        [-1, 1024, 64, 128]               0\n",
            "       Bottleneck-90        [-1, 1024, 64, 128]               0\n",
            "           Conv2d-91         [-1, 256, 64, 128]         262,144\n",
            "      BatchNorm2d-92         [-1, 256, 64, 128]             512\n",
            "             ReLU-93         [-1, 256, 64, 128]               0\n",
            "           Conv2d-94         [-1, 256, 64, 128]         589,824\n",
            "      BatchNorm2d-95         [-1, 256, 64, 128]             512\n",
            "             ReLU-96         [-1, 256, 64, 128]               0\n",
            "           Conv2d-97        [-1, 1024, 64, 128]         262,144\n",
            "      BatchNorm2d-98        [-1, 1024, 64, 128]           2,048\n",
            "             ReLU-99        [-1, 1024, 64, 128]               0\n",
            "      Bottleneck-100        [-1, 1024, 64, 128]               0\n",
            "          Conv2d-101         [-1, 256, 64, 128]         262,144\n",
            "     BatchNorm2d-102         [-1, 256, 64, 128]             512\n",
            "            ReLU-103         [-1, 256, 64, 128]               0\n",
            "          Conv2d-104         [-1, 256, 64, 128]         589,824\n",
            "     BatchNorm2d-105         [-1, 256, 64, 128]             512\n",
            "            ReLU-106         [-1, 256, 64, 128]               0\n",
            "          Conv2d-107        [-1, 1024, 64, 128]         262,144\n",
            "     BatchNorm2d-108        [-1, 1024, 64, 128]           2,048\n",
            "            ReLU-109        [-1, 1024, 64, 128]               0\n",
            "      Bottleneck-110        [-1, 1024, 64, 128]               0\n",
            "          Conv2d-111         [-1, 256, 64, 128]         262,144\n",
            "     BatchNorm2d-112         [-1, 256, 64, 128]             512\n",
            "            ReLU-113         [-1, 256, 64, 128]               0\n",
            "          Conv2d-114         [-1, 256, 64, 128]         589,824\n",
            "     BatchNorm2d-115         [-1, 256, 64, 128]             512\n",
            "            ReLU-116         [-1, 256, 64, 128]               0\n",
            "          Conv2d-117        [-1, 1024, 64, 128]         262,144\n",
            "     BatchNorm2d-118        [-1, 1024, 64, 128]           2,048\n",
            "            ReLU-119        [-1, 1024, 64, 128]               0\n",
            "      Bottleneck-120        [-1, 1024, 64, 128]               0\n",
            "          Conv2d-121         [-1, 256, 64, 128]         262,144\n",
            "     BatchNorm2d-122         [-1, 256, 64, 128]             512\n",
            "            ReLU-123         [-1, 256, 64, 128]               0\n",
            "          Conv2d-124         [-1, 256, 64, 128]         589,824\n",
            "     BatchNorm2d-125         [-1, 256, 64, 128]             512\n",
            "            ReLU-126         [-1, 256, 64, 128]               0\n",
            "          Conv2d-127        [-1, 1024, 64, 128]         262,144\n",
            "     BatchNorm2d-128        [-1, 1024, 64, 128]           2,048\n",
            "            ReLU-129        [-1, 1024, 64, 128]               0\n",
            "      Bottleneck-130        [-1, 1024, 64, 128]               0\n",
            "          Conv2d-131         [-1, 256, 64, 128]         262,144\n",
            "     BatchNorm2d-132         [-1, 256, 64, 128]             512\n",
            "            ReLU-133         [-1, 256, 64, 128]               0\n",
            "          Conv2d-134         [-1, 256, 64, 128]         589,824\n",
            "     BatchNorm2d-135         [-1, 256, 64, 128]             512\n",
            "            ReLU-136         [-1, 256, 64, 128]               0\n",
            "          Conv2d-137        [-1, 1024, 64, 128]         262,144\n",
            "     BatchNorm2d-138        [-1, 1024, 64, 128]           2,048\n",
            "            ReLU-139        [-1, 1024, 64, 128]               0\n",
            "      Bottleneck-140        [-1, 1024, 64, 128]               0\n",
            "          Conv2d-141         [-1, 512, 64, 128]         524,288\n",
            "     BatchNorm2d-142         [-1, 512, 64, 128]           1,024\n",
            "          Conv2d-143         [-1, 512, 64, 128]       2,359,296\n",
            "     BatchNorm2d-144         [-1, 512, 64, 128]           1,024\n",
            "          Conv2d-145        [-1, 2048, 64, 128]       1,048,576\n",
            "     BatchNorm2d-146        [-1, 2048, 64, 128]           4,096\n",
            "          Conv2d-147        [-1, 2048, 64, 128]       2,097,152\n",
            "     BatchNorm2d-148        [-1, 2048, 64, 128]           4,096\n",
            "      Bottleneck-149        [-1, 2048, 64, 128]               0\n",
            "          Conv2d-150         [-1, 512, 64, 128]       1,048,576\n",
            "     BatchNorm2d-151         [-1, 512, 64, 128]           1,024\n",
            "          Conv2d-152         [-1, 512, 64, 128]       2,359,296\n",
            "     BatchNorm2d-153         [-1, 512, 64, 128]           1,024\n",
            "          Conv2d-154        [-1, 2048, 64, 128]       1,048,576\n",
            "     BatchNorm2d-155        [-1, 2048, 64, 128]           4,096\n",
            "      Bottleneck-156        [-1, 2048, 64, 128]               0\n",
            "          Conv2d-157         [-1, 512, 64, 128]       1,048,576\n",
            "     BatchNorm2d-158         [-1, 512, 64, 128]           1,024\n",
            "          Conv2d-159         [-1, 512, 64, 128]       2,359,296\n",
            "     BatchNorm2d-160         [-1, 512, 64, 128]           1,024\n",
            "          Conv2d-161        [-1, 2048, 64, 128]       1,048,576\n",
            "     BatchNorm2d-162        [-1, 2048, 64, 128]           4,096\n",
            "      Bottleneck-163        [-1, 2048, 64, 128]               0\n",
            "ResNet_Bottleneck_OS16-164        [-1, 2048, 64, 128]               0\n",
            "          Conv2d-165         [-1, 256, 64, 128]         524,544\n",
            "     BatchNorm2d-166         [-1, 256, 64, 128]             512\n",
            "          Conv2d-167         [-1, 256, 64, 128]       4,718,848\n",
            "     BatchNorm2d-168         [-1, 256, 64, 128]             512\n",
            "          Conv2d-169         [-1, 256, 64, 128]       4,718,848\n",
            "     BatchNorm2d-170         [-1, 256, 64, 128]             512\n",
            "          Conv2d-171         [-1, 256, 64, 128]       4,718,848\n",
            "     BatchNorm2d-172         [-1, 256, 64, 128]             512\n",
            "AdaptiveAvgPool2d-173           [-1, 2048, 1, 1]               0\n",
            "          Conv2d-174            [-1, 256, 1, 1]         524,544\n",
            "     BatchNorm2d-175            [-1, 256, 1, 1]             512\n",
            "          Conv2d-176         [-1, 256, 64, 128]         327,936\n",
            "     BatchNorm2d-177         [-1, 256, 64, 128]             512\n",
            "          Conv2d-178          [-1, 20, 64, 128]           5,140\n",
            " ASPP_Bottleneck-179          [-1, 20, 64, 128]               0\n",
            "================================================================\n",
            "Total params: 39,049,812\n",
            "Trainable params: 39,049,812\n",
            "Non-trainable params: 0\n",
            "----------------------------------------------------------------\n",
            "Input size (MB): 24.00\n",
            "Forward/backward pass size (MB): 13394.52\n",
            "Params size (MB): 148.96\n",
            "Estimated Total Size (MB): 13567.48\n",
            "----------------------------------------------------------------\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sIIqzmN-4jqR",
        "colab_type": "text"
      },
      "source": [
        "####Eval on Validation Data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TKDbz7DWWw_3",
        "colab_type": "text"
      },
      "source": [
        "#####Code\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hPKGInk04i9Z",
        "colab_type": "code",
        "outputId": "c4ae7949-772c-4182-95d4-13221714ae9d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "batch_size = 2\n",
        "\n",
        "network = DeepLabV3(\"eval_val\", project_dir=datasetPath).cuda()\n",
        "#network.load_state_dict(torch.load(datasetPath+\"/pretrained_models/model_13_2_2_2_epoch_580.pth\"))\n",
        "#network.load_state_dict(torch.load(\"/content/drive/My Drive/U_Net_Datasets/training_logs/model_5/checkpoints/model_5_epoch_92.pth\"))\n",
        "#network.load_state_dict(torch.load(\"/content/drive/My Drive/U_Net_Datasets/training_logs/model_5/checkpoints/model_5_epoch_92.pth\"))\n",
        "#network.load_state_dict(torch.load(\"/content/drive/My Drive/U_Net_Datasets/training_logs/model_50_b3_w2/checkpoints/model_50_b3_w2_epoch_128.pth\"))\n",
        "#network.load_state_dict(\n",
        "    #torch.load(\"/content/drive/My Drive/U_Net_Datasets/training_logs/model_101_b3_w2/checkpoints/model_101_b3_w2_epoch_141.pth\"))\n",
        "network.load_state_dict(\n",
        "    torch.load(\"/content/drive/My Drive/U_Net_Datasets/training_logs/model_101_b4_w1_from_148+22_LR_00001/checkpoints/model_101_b4_w1_from_148+22_LR_00001_epoch_22.pth\"))\n",
        "\n",
        "val_dataset = DatasetVal(cityscapes_data_path, cityscapes_meta_path)\n",
        "\n",
        "num_val_batches = int(len(val_dataset)/batch_size)\n",
        "print (\"num_val_batches:\", num_val_batches)\n",
        "\n",
        "val_loader = torch.utils.data.DataLoader(dataset=val_dataset,\n",
        "                                         batch_size=batch_size, shuffle=False,\n",
        "                                         num_workers=1)\n",
        "\n",
        "with open(cityscapes_meta_path+\"/class_weights.pkl\", \"rb\") as file: # (needed for python3)\n",
        "    class_weights = np.array(pickle.load(file))\n",
        "class_weights = torch.from_numpy(class_weights)\n",
        "class_weights = Variable(class_weights.type(torch.FloatTensor)).cuda()\n",
        "\n",
        "# loss function\n",
        "loss_fn = nn.CrossEntropyLoss(weight=class_weights)\n",
        "\n",
        "network.eval() # (set in evaluation mode, this affects BatchNorm and dropout)\n",
        "\n",
        "myMIOU = mIoULoss(n_classes=20)\n",
        "a= []\n",
        "mIOu = []\n",
        "countBatch = 0\n",
        "batch_losses = []\n",
        "for step, (imgs, label_imgs, img_ids) in enumerate(val_loader):\n",
        "    with torch.no_grad(): # (corresponds to setting volatile=True in all variables, this is done during inference to reduce memory consumption)\n",
        "        imgs = Variable(imgs).cuda() # (shape: (batch_size, 3, img_h, img_w))\n",
        "        label_imgs = Variable(label_imgs.type(torch.LongTensor)).cuda() # (shape: (batch_size, img_h, img_w))\n",
        "        \n",
        "        outputs = network(imgs) # (shape: (batch_size, num_classes, img_h, img_w))\n",
        "       \n",
        "        target_oneHot_2 = to_one_hot(label_imgs.data.cpu(),20)\n",
        "        mIOu.append(myMIOU.forward(outputs,target_oneHot_2.cuda()))\n",
        "        #print(mIOu)\n",
        "        print(\"Mean Intersection Over Union on val is = \" ,1-torch.mean(torch.stack(mIOu)).item(), \"Count batch: \", countBatch) # for debuging\n",
        "        countBatch +=1 #debug val\n",
        "\n",
        "        # compute the loss:\n",
        "        loss = loss_fn(outputs, label_imgs)\n",
        "        loss_value = loss.data.cpu().numpy()\n",
        "        batch_losses.append(loss_value)\n",
        "\n",
        "      \n",
        "\n",
        "        ########################################################################\n",
        "        # save data for visualization:\n",
        "        ########################################################################\n",
        "        outputs = outputs.data.cpu().numpy() # (shape: (batch_size, num_classes, img_h, img_w))\n",
        "        pred_label_imgs = np.argmax(outputs, axis=1) # (shape: (batch_size, img_h, img_w))\n",
        "        pred_label_imgs = pred_label_imgs.astype(np.uint8)\n",
        "\n",
        "        for i in range(pred_label_imgs.shape[0]):\n",
        "            if i == 0:\n",
        "                pred_label_img = pred_label_imgs[i] # (shape: (img_h, img_w))\n",
        "                img_id = img_ids[i]\n",
        "                img = imgs[i] # (shape: (3, img_h, img_w))\n",
        "\n",
        "                img = img.data.cpu().numpy()\n",
        "                img = np.transpose(img, (1, 2, 0)) # (shape: (img_h, img_w, 3))\n",
        "                img = img*np.array([0.229, 0.224, 0.225])\n",
        "                img = img + np.array([0.485, 0.456, 0.406])\n",
        "                img = img*255.0\n",
        "                img = img.astype(np.uint8)\n",
        "\n",
        "                pred_label_img_color = label_img_to_color(pred_label_img)\n",
        "                overlayed_img = 0.35*img + 0.65*pred_label_img_color\n",
        "                overlayed_img = overlayed_img.astype(np.uint8)\n",
        "\n",
        "                cv2.imwrite(network.model_dir + \"/\" + img_id + \"_overlayed.png\", overlayed_img)\n",
        "\n",
        "\n",
        "\n",
        "val_loss = np.mean(batch_losses)\n",
        "print (\"val loss: %g\" % val_loss)\n",
        "print(\"Mean Intersection Over Union on val is final = \" ,1-torch.mean(torch.stack(mIOu)).item())"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "pretrained resnet, 101\n",
            "num_val_batches: 250\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/torch/nn/functional.py:2404: UserWarning: nn.functional.upsample is deprecated. Use nn.functional.interpolate instead.\n",
            "  warnings.warn(\"nn.functional.upsample is deprecated. Use nn.functional.interpolate instead.\")\n",
            "/usr/local/lib/python3.6/dist-packages/torch/nn/functional.py:2494: UserWarning: Default upsampling behavior when mode=bilinear is changed to align_corners=False since 0.4.0. Please specify align_corners=True if the old behavior is desired. See the documentation of nn.Upsample for details.\n",
            "  \"See the documentation of nn.Upsample for details.\".format(mode))\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Mean Intersection Over Union on val is =  0.6888691484928131 Count batch:  0\n",
            "Mean Intersection Over Union on val is =  0.6884414553642273 Count batch:  1\n",
            "Mean Intersection Over Union on val is =  0.6978885233402252 Count batch:  2\n",
            "Mean Intersection Over Union on val is =  0.6880359649658203 Count batch:  3\n",
            "Mean Intersection Over Union on val is =  0.6895788908004761 Count batch:  4\n",
            "Mean Intersection Over Union on val is =  0.6982326209545135 Count batch:  5\n",
            "Mean Intersection Over Union on val is =  0.6964055597782135 Count batch:  6\n",
            "Mean Intersection Over Union on val is =  0.6969465613365173 Count batch:  7\n",
            "Mean Intersection Over Union on val is =  0.6970607340335846 Count batch:  8\n",
            "Mean Intersection Over Union on val is =  0.6982169449329376 Count batch:  9\n",
            "Mean Intersection Over Union on val is =  0.6974457800388336 Count batch:  10\n",
            "Mean Intersection Over Union on val is =  0.6927207708358765 Count batch:  11\n",
            "Mean Intersection Over Union on val is =  0.6938765347003937 Count batch:  12\n",
            "Mean Intersection Over Union on val is =  0.6951688230037689 Count batch:  13\n",
            "Mean Intersection Over Union on val is =  0.6955061554908752 Count batch:  14\n",
            "Mean Intersection Over Union on val is =  0.696075439453125 Count batch:  15\n",
            "Mean Intersection Over Union on val is =  0.6940321028232574 Count batch:  16\n",
            "Mean Intersection Over Union on val is =  0.6947451829910278 Count batch:  17\n",
            "Mean Intersection Over Union on val is =  0.6962894201278687 Count batch:  18\n",
            "Mean Intersection Over Union on val is =  0.6964831650257111 Count batch:  19\n",
            "Mean Intersection Over Union on val is =  0.6953250765800476 Count batch:  20\n",
            "Mean Intersection Over Union on val is =  0.6942723393440247 Count batch:  21\n",
            "Mean Intersection Over Union on val is =  0.6924067735671997 Count batch:  22\n",
            "Mean Intersection Over Union on val is =  0.6934092044830322 Count batch:  23\n",
            "Mean Intersection Over Union on val is =  0.690374106168747 Count batch:  24\n",
            "Mean Intersection Over Union on val is =  0.6897941529750824 Count batch:  25\n",
            "Mean Intersection Over Union on val is =  0.6904486119747162 Count batch:  26\n",
            "Mean Intersection Over Union on val is =  0.6895462274551392 Count batch:  27\n",
            "Mean Intersection Over Union on val is =  0.690406084060669 Count batch:  28\n",
            "Mean Intersection Over Union on val is =  0.689773827791214 Count batch:  29\n",
            "Mean Intersection Over Union on val is =  0.6881288588047028 Count batch:  30\n",
            "Mean Intersection Over Union on val is =  0.6891173124313354 Count batch:  31\n",
            "Mean Intersection Over Union on val is =  0.6882531642913818 Count batch:  32\n",
            "Mean Intersection Over Union on val is =  0.6890382766723633 Count batch:  33\n",
            "Mean Intersection Over Union on val is =  0.689509242773056 Count batch:  34\n",
            "Mean Intersection Over Union on val is =  0.689367026090622 Count batch:  35\n",
            "Mean Intersection Over Union on val is =  0.6903863549232483 Count batch:  36\n",
            "Mean Intersection Over Union on val is =  0.6910431385040283 Count batch:  37\n",
            "Mean Intersection Over Union on val is =  0.6928535401821136 Count batch:  38\n",
            "Mean Intersection Over Union on val is =  0.6939026415348053 Count batch:  39\n",
            "Mean Intersection Over Union on val is =  0.6941318213939667 Count batch:  40\n",
            "Mean Intersection Over Union on val is =  0.6953658163547516 Count batch:  41\n",
            "Mean Intersection Over Union on val is =  0.6967146694660187 Count batch:  42\n",
            "Mean Intersection Over Union on val is =  0.6971142888069153 Count batch:  43\n",
            "Mean Intersection Over Union on val is =  0.6969485282897949 Count batch:  44\n",
            "Mean Intersection Over Union on val is =  0.6972976326942444 Count batch:  45\n",
            "Mean Intersection Over Union on val is =  0.6972265541553497 Count batch:  46\n",
            "Mean Intersection Over Union on val is =  0.6975460350513458 Count batch:  47\n",
            "Mean Intersection Over Union on val is =  0.6974406540393829 Count batch:  48\n",
            "Mean Intersection Over Union on val is =  0.6969779133796692 Count batch:  49\n",
            "Mean Intersection Over Union on val is =  0.6965555250644684 Count batch:  50\n",
            "Mean Intersection Over Union on val is =  0.6964284479618073 Count batch:  51\n",
            "Mean Intersection Over Union on val is =  0.6957822144031525 Count batch:  52\n",
            "Mean Intersection Over Union on val is =  0.6948097348213196 Count batch:  53\n",
            "Mean Intersection Over Union on val is =  0.6936416625976562 Count batch:  54\n",
            "Mean Intersection Over Union on val is =  0.6935508549213409 Count batch:  55\n",
            "Mean Intersection Over Union on val is =  0.6947650909423828 Count batch:  56\n",
            "Mean Intersection Over Union on val is =  0.69484743475914 Count batch:  57\n",
            "Mean Intersection Over Union on val is =  0.6944600641727448 Count batch:  58\n",
            "Mean Intersection Over Union on val is =  0.6944220066070557 Count batch:  59\n",
            "Mean Intersection Over Union on val is =  0.6945885419845581 Count batch:  60\n",
            "Mean Intersection Over Union on val is =  0.6940398216247559 Count batch:  61\n",
            "Mean Intersection Over Union on val is =  0.6951122879981995 Count batch:  62\n",
            "Mean Intersection Over Union on val is =  0.6950470805168152 Count batch:  63\n",
            "Mean Intersection Over Union on val is =  0.6943747997283936 Count batch:  64\n",
            "Mean Intersection Over Union on val is =  0.694953203201294 Count batch:  65\n",
            "Mean Intersection Over Union on val is =  0.6957239806652069 Count batch:  66\n",
            "Mean Intersection Over Union on val is =  0.6963405013084412 Count batch:  67\n",
            "Mean Intersection Over Union on val is =  0.6970404386520386 Count batch:  68\n",
            "Mean Intersection Over Union on val is =  0.6963170170783997 Count batch:  69\n",
            "Mean Intersection Over Union on val is =  0.696321040391922 Count batch:  70\n",
            "Mean Intersection Over Union on val is =  0.6965059638023376 Count batch:  71\n",
            "Mean Intersection Over Union on val is =  0.6962072849273682 Count batch:  72\n",
            "Mean Intersection Over Union on val is =  0.6963129639625549 Count batch:  73\n",
            "Mean Intersection Over Union on val is =  0.6965030431747437 Count batch:  74\n",
            "Mean Intersection Over Union on val is =  0.6962895691394806 Count batch:  75\n",
            "Mean Intersection Over Union on val is =  0.6962243020534515 Count batch:  76\n",
            "Mean Intersection Over Union on val is =  0.6956924498081207 Count batch:  77\n",
            "Mean Intersection Over Union on val is =  0.6961216926574707 Count batch:  78\n",
            "Mean Intersection Over Union on val is =  0.6964366436004639 Count batch:  79\n",
            "Mean Intersection Over Union on val is =  0.6964095532894135 Count batch:  80\n",
            "Mean Intersection Over Union on val is =  0.6962180435657501 Count batch:  81\n",
            "Mean Intersection Over Union on val is =  0.6966167688369751 Count batch:  82\n",
            "Mean Intersection Over Union on val is =  0.696719616651535 Count batch:  83\n",
            "Mean Intersection Over Union on val is =  0.6962723731994629 Count batch:  84\n",
            "Mean Intersection Over Union on val is =  0.6966840624809265 Count batch:  85\n",
            "Mean Intersection Over Union on val is =  0.6967915594577789 Count batch:  86\n",
            "Mean Intersection Over Union on val is =  0.6972363889217377 Count batch:  87\n",
            "Mean Intersection Over Union on val is =  0.6975256204605103 Count batch:  88\n",
            "Mean Intersection Over Union on val is =  0.6978955864906311 Count batch:  89\n",
            "Mean Intersection Over Union on val is =  0.6978993117809296 Count batch:  90\n",
            "Mean Intersection Over Union on val is =  0.6974808871746063 Count batch:  91\n",
            "Mean Intersection Over Union on val is =  0.6978302597999573 Count batch:  92\n",
            "Mean Intersection Over Union on val is =  0.6976814270019531 Count batch:  93\n",
            "Mean Intersection Over Union on val is =  0.6976142227649689 Count batch:  94\n",
            "Mean Intersection Over Union on val is =  0.6982759833335876 Count batch:  95\n",
            "Mean Intersection Over Union on val is =  0.6984729170799255 Count batch:  96\n",
            "Mean Intersection Over Union on val is =  0.6981858015060425 Count batch:  97\n",
            "Mean Intersection Over Union on val is =  0.69841268658638 Count batch:  98\n",
            "Mean Intersection Over Union on val is =  0.6988495588302612 Count batch:  99\n",
            "Mean Intersection Over Union on val is =  0.6990602612495422 Count batch:  100\n",
            "Mean Intersection Over Union on val is =  0.6993702352046967 Count batch:  101\n",
            "Mean Intersection Over Union on val is =  0.6995121538639069 Count batch:  102\n",
            "Mean Intersection Over Union on val is =  0.6997779905796051 Count batch:  103\n",
            "Mean Intersection Over Union on val is =  0.6998471319675446 Count batch:  104\n",
            "Mean Intersection Over Union on val is =  0.6993259191513062 Count batch:  105\n",
            "Mean Intersection Over Union on val is =  0.6987791061401367 Count batch:  106\n",
            "Mean Intersection Over Union on val is =  0.6985377669334412 Count batch:  107\n",
            "Mean Intersection Over Union on val is =  0.6986887753009796 Count batch:  108\n",
            "Mean Intersection Over Union on val is =  0.6981383562088013 Count batch:  109\n",
            "Mean Intersection Over Union on val is =  0.6981231272220612 Count batch:  110\n",
            "Mean Intersection Over Union on val is =  0.6980409622192383 Count batch:  111\n",
            "Mean Intersection Over Union on val is =  0.6981597244739532 Count batch:  112\n",
            "Mean Intersection Over Union on val is =  0.6980392634868622 Count batch:  113\n",
            "Mean Intersection Over Union on val is =  0.6984430551528931 Count batch:  114\n",
            "Mean Intersection Over Union on val is =  0.6989656984806061 Count batch:  115\n",
            "Mean Intersection Over Union on val is =  0.6994227170944214 Count batch:  116\n",
            "Mean Intersection Over Union on val is =  0.6988169848918915 Count batch:  117\n",
            "Mean Intersection Over Union on val is =  0.698756992816925 Count batch:  118\n",
            "Mean Intersection Over Union on val is =  0.6986550390720367 Count batch:  119\n",
            "Mean Intersection Over Union on val is =  0.6981694996356964 Count batch:  120\n",
            "Mean Intersection Over Union on val is =  0.6982680261135101 Count batch:  121\n",
            "Mean Intersection Over Union on val is =  0.6979498565196991 Count batch:  122\n",
            "Mean Intersection Over Union on val is =  0.6979890167713165 Count batch:  123\n",
            "Mean Intersection Over Union on val is =  0.6981501281261444 Count batch:  124\n",
            "Mean Intersection Over Union on val is =  0.6978827714920044 Count batch:  125\n",
            "Mean Intersection Over Union on val is =  0.6977033615112305 Count batch:  126\n",
            "Mean Intersection Over Union on val is =  0.6978882849216461 Count batch:  127\n",
            "Mean Intersection Over Union on val is =  0.6979882121086121 Count batch:  128\n",
            "Mean Intersection Over Union on val is =  0.6979222297668457 Count batch:  129\n",
            "Mean Intersection Over Union on val is =  0.6978373229503632 Count batch:  130\n",
            "Mean Intersection Over Union on val is =  0.6980554461479187 Count batch:  131\n",
            "Mean Intersection Over Union on val is =  0.6981349885463715 Count batch:  132\n",
            "Mean Intersection Over Union on val is =  0.697978287935257 Count batch:  133\n",
            "Mean Intersection Over Union on val is =  0.6977770328521729 Count batch:  134\n",
            "Mean Intersection Over Union on val is =  0.697228729724884 Count batch:  135\n",
            "Mean Intersection Over Union on val is =  0.6970219016075134 Count batch:  136\n",
            "Mean Intersection Over Union on val is =  0.6966524422168732 Count batch:  137\n",
            "Mean Intersection Over Union on val is =  0.6965708136558533 Count batch:  138\n",
            "Mean Intersection Over Union on val is =  0.6962127983570099 Count batch:  139\n",
            "Mean Intersection Over Union on val is =  0.6960597634315491 Count batch:  140\n",
            "Mean Intersection Over Union on val is =  0.6959397196769714 Count batch:  141\n",
            "Mean Intersection Over Union on val is =  0.6960155665874481 Count batch:  142\n",
            "Mean Intersection Over Union on val is =  0.6958818435668945 Count batch:  143\n",
            "Mean Intersection Over Union on val is =  0.6956097781658173 Count batch:  144\n",
            "Mean Intersection Over Union on val is =  0.6953475773334503 Count batch:  145\n",
            "Mean Intersection Over Union on val is =  0.6950197219848633 Count batch:  146\n",
            "Mean Intersection Over Union on val is =  0.6948351263999939 Count batch:  147\n",
            "Mean Intersection Over Union on val is =  0.694637805223465 Count batch:  148\n",
            "Mean Intersection Over Union on val is =  0.6941806972026825 Count batch:  149\n",
            "Mean Intersection Over Union on val is =  0.6940197944641113 Count batch:  150\n",
            "Mean Intersection Over Union on val is =  0.693895548582077 Count batch:  151\n",
            "Mean Intersection Over Union on val is =  0.6935057640075684 Count batch:  152\n",
            "Mean Intersection Over Union on val is =  0.693012148141861 Count batch:  153\n",
            "Mean Intersection Over Union on val is =  0.6927888989448547 Count batch:  154\n",
            "Mean Intersection Over Union on val is =  0.6924085319042206 Count batch:  155\n",
            "Mean Intersection Over Union on val is =  0.692200094461441 Count batch:  156\n",
            "Mean Intersection Over Union on val is =  0.691729724407196 Count batch:  157\n",
            "Mean Intersection Over Union on val is =  0.6916316449642181 Count batch:  158\n",
            "Mean Intersection Over Union on val is =  0.6911441385746002 Count batch:  159\n",
            "Mean Intersection Over Union on val is =  0.6909807026386261 Count batch:  160\n",
            "Mean Intersection Over Union on val is =  0.6904767453670502 Count batch:  161\n",
            "Mean Intersection Over Union on val is =  0.6900398135185242 Count batch:  162\n",
            "Mean Intersection Over Union on val is =  0.689740389585495 Count batch:  163\n",
            "Mean Intersection Over Union on val is =  0.6896829307079315 Count batch:  164\n",
            "Mean Intersection Over Union on val is =  0.6893126964569092 Count batch:  165\n",
            "Mean Intersection Over Union on val is =  0.6891690492630005 Count batch:  166\n",
            "Mean Intersection Over Union on val is =  0.6891039907932281 Count batch:  167\n",
            "Mean Intersection Over Union on val is =  0.688886821269989 Count batch:  168\n",
            "Mean Intersection Over Union on val is =  0.6887820363044739 Count batch:  169\n",
            "Mean Intersection Over Union on val is =  0.6887753009796143 Count batch:  170\n",
            "Mean Intersection Over Union on val is =  0.6889483034610748 Count batch:  171\n",
            "Mean Intersection Over Union on val is =  0.6887664496898651 Count batch:  172\n",
            "Mean Intersection Over Union on val is =  0.6882235705852509 Count batch:  173\n",
            "Mean Intersection Over Union on val is =  0.6885444819927216 Count batch:  174\n",
            "Mean Intersection Over Union on val is =  0.6882201433181763 Count batch:  175\n",
            "Mean Intersection Over Union on val is =  0.6881671845912933 Count batch:  176\n",
            "Mean Intersection Over Union on val is =  0.6881353557109833 Count batch:  177\n",
            "Mean Intersection Over Union on val is =  0.6879637539386749 Count batch:  178\n",
            "Mean Intersection Over Union on val is =  0.6879055500030518 Count batch:  179\n",
            "Mean Intersection Over Union on val is =  0.6880311667919159 Count batch:  180\n",
            "Mean Intersection Over Union on val is =  0.6878135204315186 Count batch:  181\n",
            "Mean Intersection Over Union on val is =  0.6875771284103394 Count batch:  182\n",
            "Mean Intersection Over Union on val is =  0.6877770125865936 Count batch:  183\n",
            "Mean Intersection Over Union on val is =  0.6877776086330414 Count batch:  184\n",
            "Mean Intersection Over Union on val is =  0.6874378025531769 Count batch:  185\n",
            "Mean Intersection Over Union on val is =  0.6874096691608429 Count batch:  186\n",
            "Mean Intersection Over Union on val is =  0.6871019303798676 Count batch:  187\n",
            "Mean Intersection Over Union on val is =  0.6871635913848877 Count batch:  188\n",
            "Mean Intersection Over Union on val is =  0.6870580315589905 Count batch:  189\n",
            "Mean Intersection Over Union on val is =  0.6870020627975464 Count batch:  190\n",
            "Mean Intersection Over Union on val is =  0.6868966817855835 Count batch:  191\n",
            "Mean Intersection Over Union on val is =  0.6867792010307312 Count batch:  192\n",
            "Mean Intersection Over Union on val is =  0.6866558194160461 Count batch:  193\n",
            "Mean Intersection Over Union on val is =  0.6865823268890381 Count batch:  194\n",
            "Mean Intersection Over Union on val is =  0.6864044964313507 Count batch:  195\n",
            "Mean Intersection Over Union on val is =  0.6862363517284393 Count batch:  196\n",
            "Mean Intersection Over Union on val is =  0.6861675083637238 Count batch:  197\n",
            "Mean Intersection Over Union on val is =  0.6859458982944489 Count batch:  198\n",
            "Mean Intersection Over Union on val is =  0.6859168708324432 Count batch:  199\n",
            "Mean Intersection Over Union on val is =  0.6859833002090454 Count batch:  200\n",
            "Mean Intersection Over Union on val is =  0.6859543919563293 Count batch:  201\n",
            "Mean Intersection Over Union on val is =  0.6859820783138275 Count batch:  202\n",
            "Mean Intersection Over Union on val is =  0.6859764158725739 Count batch:  203\n",
            "Mean Intersection Over Union on val is =  0.6862524151802063 Count batch:  204\n",
            "Mean Intersection Over Union on val is =  0.6867208480834961 Count batch:  205\n",
            "Mean Intersection Over Union on val is =  0.6872668862342834 Count batch:  206\n",
            "Mean Intersection Over Union on val is =  0.6873466968536377 Count batch:  207\n",
            "Mean Intersection Over Union on val is =  0.6873104870319366 Count batch:  208\n",
            "Mean Intersection Over Union on val is =  0.6870907247066498 Count batch:  209\n",
            "Mean Intersection Over Union on val is =  0.6870996356010437 Count batch:  210\n",
            "Mean Intersection Over Union on val is =  0.6870351135730743 Count batch:  211\n",
            "Mean Intersection Over Union on val is =  0.6868133544921875 Count batch:  212\n",
            "Mean Intersection Over Union on val is =  0.6866461634635925 Count batch:  213\n",
            "Mean Intersection Over Union on val is =  0.6866820454597473 Count batch:  214\n",
            "Mean Intersection Over Union on val is =  0.6868396103382111 Count batch:  215\n",
            "Mean Intersection Over Union on val is =  0.6867583692073822 Count batch:  216\n",
            "Mean Intersection Over Union on val is =  0.6866151392459869 Count batch:  217\n",
            "Mean Intersection Over Union on val is =  0.6866924166679382 Count batch:  218\n",
            "Mean Intersection Over Union on val is =  0.6865858733654022 Count batch:  219\n",
            "Mean Intersection Over Union on val is =  0.6867884993553162 Count batch:  220\n",
            "Mean Intersection Over Union on val is =  0.6869570910930634 Count batch:  221\n",
            "Mean Intersection Over Union on val is =  0.687053769826889 Count batch:  222\n",
            "Mean Intersection Over Union on val is =  0.6870793700218201 Count batch:  223\n",
            "Mean Intersection Over Union on val is =  0.687177836894989 Count batch:  224\n",
            "Mean Intersection Over Union on val is =  0.6872981488704681 Count batch:  225\n",
            "Mean Intersection Over Union on val is =  0.6871905326843262 Count batch:  226\n",
            "Mean Intersection Over Union on val is =  0.6873571872711182 Count batch:  227\n",
            "Mean Intersection Over Union on val is =  0.687435507774353 Count batch:  228\n",
            "Mean Intersection Over Union on val is =  0.6883154213428497 Count batch:  229\n",
            "Mean Intersection Over Union on val is =  0.6891728639602661 Count batch:  230\n",
            "Mean Intersection Over Union on val is =  0.689746081829071 Count batch:  231\n",
            "Mean Intersection Over Union on val is =  0.6900258958339691 Count batch:  232\n",
            "Mean Intersection Over Union on val is =  0.6904417872428894 Count batch:  233\n",
            "Mean Intersection Over Union on val is =  0.6908925473690033 Count batch:  234\n",
            "Mean Intersection Over Union on val is =  0.6914818286895752 Count batch:  235\n",
            "Mean Intersection Over Union on val is =  0.6920301914215088 Count batch:  236\n",
            "Mean Intersection Over Union on val is =  0.6925503313541412 Count batch:  237\n",
            "Mean Intersection Over Union on val is =  0.6930802464485168 Count batch:  238\n",
            "Mean Intersection Over Union on val is =  0.6933228671550751 Count batch:  239\n",
            "Mean Intersection Over Union on val is =  0.6940323114395142 Count batch:  240\n",
            "Mean Intersection Over Union on val is =  0.6942955851554871 Count batch:  241\n",
            "Mean Intersection Over Union on val is =  0.694840669631958 Count batch:  242\n",
            "Mean Intersection Over Union on val is =  0.6956247985363007 Count batch:  243\n",
            "Mean Intersection Over Union on val is =  0.6963844001293182 Count batch:  244\n",
            "Mean Intersection Over Union on val is =  0.6971959471702576 Count batch:  245\n",
            "Mean Intersection Over Union on val is =  0.6978270411491394 Count batch:  246\n",
            "Mean Intersection Over Union on val is =  0.6981569230556488 Count batch:  247\n",
            "Mean Intersection Over Union on val is =  0.6984968185424805 Count batch:  248\n",
            "Mean Intersection Over Union on val is =  0.6987471282482147 Count batch:  249\n",
            "val loss: 0.524581\n",
            "Mean Intersection Over Union on val is final =  0.6987471282482147\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wq3f2kw1Lfx-",
        "colab_type": "text"
      },
      "source": [
        "ResNet18 Epoch 92 \n",
        "\n",
        "- val loss: 0.682222\n",
        "- mIOU is : 0.723001\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "I_A4cfHMR1LF",
        "colab_type": "text"
      },
      "source": [
        "ResNet50 Epoch 128\n",
        "\n",
        "- val loss: 0.606357\n",
        "- mIOU is : 0.7202193737030029\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dpR6T39xMC8t",
        "colab_type": "text"
      },
      "source": [
        "ResNet101 Epoch 148 + 22 \n",
        "\n",
        "- val loss: 0.515101\n",
        "- mIOU is : 0.6987471282482147\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HM3GWpVd5o5w",
        "colab_type": "text"
      },
      "source": [
        "###Visualisation"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Wx4KIJoKYOJe",
        "colab_type": "text"
      },
      "source": [
        "###### Visualisazion ResNet18 --> Epoch 91"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tY3hJ7m2G0EO",
        "colab_type": "text"
      },
      "source": [
        "first try with the ResNet 18\n",
        "- Batch 3 \n",
        "- Worker 2\n",
        "\n",
        "https://drive.google.com/open?id=1fmEINKCeWl-2f0G_umkUx9ecccS3Ct2a"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jPeEKcx2W-ag",
        "colab_type": "code",
        "outputId": "af69154c-cf9f-4b79-bc0d-354bf744de49",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        }
      },
      "source": [
        "%matplotlib inline\n",
        "ResNet18_Train_losses = pickle.load(\n",
        "    open('/content/drive/My Drive/U_Net_Datasets/training_logs/model_5/epoch_losses_train.pkl','rb'))\n",
        "ResNet18_val_losses = pickle.load(\n",
        "    open('/content/drive/My Drive/U_Net_Datasets/training_logs/model_5/epoch_losses_val.pkl','rb'))\n",
        "\n",
        "\n",
        "\n",
        "plt.plot(ResNet18_Train_losses,label = \"Train\")\n",
        "plt.plot(ResNet18_val_losses,label = \"Val\")\n",
        "plt.title('ResNet18 Training to Epoch 92')\n",
        "plt.xlabel('Epoch')\n",
        "plt.yticks(np.arange(0.5, 1.7, 0.1))\n",
        "plt.legend(loc=\"upper right\")\n",
        "plt.ylabel('loss')\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nonEJBBERjfsXeNSLy/guaeMAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BRZD899HYaBe",
        "colab_type": "text"
      },
      "source": [
        "###### Visualisazion ResNet50 --> Epoch 128\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "WmIDFLhhZJpi"
      },
      "source": [
        "secound try with the ResNet 50\n",
        "- Batch 3 \n",
        "- Worker 2\n",
        "\n",
        "https://drive.google.com/open?id=1U3m5msu7esP6fTAgY2o9m0BVVHiuniAH"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ASjsFn1XW5Kj",
        "colab_type": "code",
        "outputId": "dd145a5b-2173-43f2-d56b-6bed290ebe7d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        }
      },
      "source": [
        "%matplotlib inline\n",
        "ResNet50_Train_losses = pickle.load(\n",
        "    open('/content/drive/My Drive/U_Net_Datasets/training_logs/model_50_b3_w2/epoch_losses_train.pkl','rb'))\n",
        "ResNet50_val_losses = pickle.load(\n",
        "    open('/content/drive/My Drive/U_Net_Datasets/training_logs/model_50_b3_w2/epoch_losses_val.pkl','rb'))\n",
        "\n",
        "plt.plot(ResNet50_Train_losses,label = \"Train\" )\n",
        "plt.plot(ResNet50_val_losses,label = \"Val\")\n",
        "plt.title('ResNet50 Training to Epoch 128')\n",
        "plt.yticks(np.arange(0.5, 1.7, 0.1))\n",
        "plt.legend(loc=\"upper right\")\n",
        "plt.xlabel('epoch')\n",
        "plt.ylabel('loss')\n",
        "plt.show()\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ndLpfZGQAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XB-pYckhvGLo",
        "colab_type": "text"
      },
      "source": [
        "###### Visualisazion ResNet101 --> Epoch 148\n",
        "secound try with the ResNet 101\n",
        "- Batch 3 \n",
        "- Worker 2\n",
        "\n",
        "https://drive.google.com/open?id=1nJcNiG_fSUGSGMcORFXB1RbTRdr4xH3c"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aE8F9lJWvMed",
        "colab_type": "code",
        "outputId": "17063d7f-d8e3-4a43-bb91-537eef43fe1c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        }
      },
      "source": [
        "%matplotlib inline\n",
        "ResNet101_Train_losses = pickle.load(\n",
        "    open('/content/drive/My Drive/U_Net_Datasets/training_logs/model_101_b3_w2/epoch_losses_train.pkl','rb'))\n",
        "ResNet101_val_losses = pickle.load(\n",
        "    open('/content/drive/My Drive/U_Net_Datasets/training_logs/model_101_b3_w2/epoch_losses_val.pkl','rb'))\n",
        "\n",
        "plt.plot(ResNet101_Train_losses, label = \"Train\" )\n",
        "plt.plot(ResNet101_val_losses,label = \"Val\")\n",
        "plt.title('ResNet101 Training to Epoch 148')\n",
        "plt.yticks(np.arange(0.5, 1.7, 0.1))\n",
        "plt.legend(loc=\"upper right\")\n",
        "plt.xlabel('epoch')\n",
        "plt.ylabel('loss')\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nuf8HkU5aJTHLnrIAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "x9T3lscz6gP9"
      },
      "source": [
        "###### Visualisazion ResNet101 --> Epoch 148 as pretrained model\n",
        "Try to train the model from epoch 148 further.\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "outputId": "0f2be8ba-4723-4220-ccf2-c1a34708b305",
        "id": "pU7QU1hU6h8y",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        }
      },
      "source": [
        "%matplotlib inline\n",
        "ResNet101_Train_losses = pickle.load(\n",
        "    open('/content/drive/My Drive/U_Net_Datasets/training_logs/model_101_b9_w1_from_148_LR_00001/epoch_losses_train.pkl','rb'))\n",
        "ResNet101_val_losses = pickle.load(\n",
        "    open('/content/drive/My Drive/U_Net_Datasets/training_logs/model_101_b9_w1_from_148_LR_00001/epoch_losses_val.pkl','rb'))\n",
        "\n",
        "plt.plot(ResNet101_Train_losses,label = \"Train\" )\n",
        "plt.plot(ResNet101_val_losses,label = \"Val\")\n",
        "plt.title('ResNet101 Pretrained 148+')\n",
        "plt.legend(loc=\"upper right\")\n",
        "#plt.yticks(np.arange(0.5, 1.7, 0.1))\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DQMhU27R5tS5",
        "colab_type": "text"
      },
      "source": [
        "####Run on Seq\n",
        "\n",
        "Create a video on test data from Stuttgart to present how it looks like."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uQyJ3Snf5oM_",
        "colab_type": "code",
        "outputId": "a55ae51c-f0e0-441d-9567-47fc1f1b7762",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 165
        }
      },
      "source": [
        "\n",
        "\n",
        "batch_size = 2\n",
        "\n",
        "network = DeepLabV3(\"eval_seq\", project_dir=datasetPath).cuda()\n",
        "#network.load_state_dict(torch.load(datasetPath+\"/pretrained_models/model_13_2_2_2_epoch_580.pth\"))\n",
        "#network.load_state_dict(torch.load(\"/content/drive/My Drive/U_Net_Datasets/training_logs/model_5/checkpoints/model_5_epoch_92.pth\"))\n",
        "#network.load_state_dict(\n",
        "    #torch.load(\"/content/drive/My Drive/U_Net_Datasets/training_logs/model_50_b3_w2/checkpoints/model_50_b3_w2_epoch_128.pth\"))\n",
        "network.load_state_dict(\n",
        "    torch.load(\"/content/drive/My Drive/U_Net_Datasets/training_logs/model_101_b3_w2/checkpoints/model_101_b3_w2_epoch_141.pth\"))\n",
        "for sequence in [\"00\"]:#, \"01\", \"02\"]:\n",
        "    print (sequence)\n",
        "\n",
        "    val_dataset = DatasetSeq(cityscapes_data_path, cityscapes_meta_path, sequence=sequence)\n",
        "\n",
        "    num_val_batches = int(len(val_dataset)/batch_size)\n",
        "    print (\"num_val_batches:\", num_val_batches)\n",
        "\n",
        "    val_loader = torch.utils.data.DataLoader(dataset=val_dataset,\n",
        "                                             batch_size=batch_size, shuffle=False,\n",
        "                                             num_workers=1)\n",
        "\n",
        "    network.eval() # (set in evaluation mode, this affects BatchNorm and dropout)\n",
        "    unsorted_img_ids = []\n",
        "    for step, (imgs, img_ids) in enumerate(val_loader):\n",
        "        with torch.no_grad(): # (corresponds to setting volatile=True in all variables, this is done during inference to reduce memory consumption)\n",
        "            imgs = Variable(imgs).cuda() # (shape: (batch_size, 3, img_h, img_w))\n",
        "\n",
        "            outputs = network(imgs) # (shape: (batch_size, num_classes, img_h, img_w))\n",
        "\n",
        "            ####################################################################\n",
        "            # save data for visualization:\n",
        "            ####################################################################\n",
        "            outputs = outputs.data.cpu().numpy() # (shape: (batch_size, num_classes, img_h, img_w))\n",
        "            pred_label_imgs = np.argmax(outputs, axis=1) # (shape: (batch_size, img_h, img_w))\n",
        "            pred_label_imgs = pred_label_imgs.astype(np.uint8)\n",
        "\n",
        "            for i in range(pred_label_imgs.shape[0]):\n",
        "                pred_label_img = pred_label_imgs[i] # (shape: (img_h, img_w))\n",
        "                img_id = img_ids[i]\n",
        "                img = imgs[i] # (shape: (3, img_h, img_w))\n",
        "\n",
        "                img = img.data.cpu().numpy()\n",
        "                img = np.transpose(img, (1, 2, 0)) # (shape: (img_h, img_w, 3))\n",
        "                img = img*np.array([0.229, 0.224, 0.225])\n",
        "                img = img + np.array([0.485, 0.456, 0.406])\n",
        "                img = img*255.0\n",
        "                img = img.astype(np.uint8)\n",
        "\n",
        "                pred_label_img_color = label_img_to_color(pred_label_img)\n",
        "                overlayed_img = 0.35*img + 0.65*pred_label_img_color\n",
        "                overlayed_img = overlayed_img.astype(np.uint8)\n",
        "\n",
        "                img_h = overlayed_img.shape[0]\n",
        "                img_w = overlayed_img.shape[1]\n",
        "\n",
        "                cv2.imwrite(network.model_dir + \"/\" + img_id + \".png\", img)\n",
        "                cv2.imwrite(network.model_dir + \"/\" + img_id + \"_pred.png\", pred_label_img_color)\n",
        "                cv2.imwrite(network.model_dir + \"/\" + img_id + \"_overlayed.png\", overlayed_img)\n",
        "\n",
        "                unsorted_img_ids.append(img_id)\n",
        "\n",
        "    ############################################################################\n",
        "    # create visualization video:\n",
        "    ############################################################################\n",
        "    out = cv2.VideoWriter(\"%s/stuttgart_%s_combined.avi\" % (network.model_dir, sequence), cv2.VideoWriter_fourcc(*\"MJPG\"), 20, (2*img_w, 2*img_h))\n",
        "    sorted_img_ids = sorted(unsorted_img_ids)\n",
        "    for img_id in sorted_img_ids:\n",
        "        img = cv2.imread(network.model_dir + \"/\" + img_id + \".png\", -1)\n",
        "        pred_img = cv2.imread(network.model_dir + \"/\" + img_id + \"_pred.png\", -1)\n",
        "        overlayed_img = cv2.imread(network.model_dir + \"/\" + img_id + \"_overlayed.png\", -1)\n",
        "\n",
        "        combined_img = np.zeros((2*img_h, 2*img_w, 3), dtype=np.uint8)\n",
        "\n",
        "        combined_img[0:img_h, 0:img_w] = img\n",
        "        combined_img[0:img_h, img_w:(2*img_w)] = pred_img\n",
        "        combined_img[img_h:(2*img_h), (int(img_w/2)):(img_w + int(img_w/2))] = overlayed_img\n",
        "\n",
        "        out.write(combined_img)\n",
        "\n",
        "    out.release()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "pretrained resnet, 101\n",
            "00\n",
            "num_val_batches: 299\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/torch/nn/functional.py:2404: UserWarning: nn.functional.upsample is deprecated. Use nn.functional.interpolate instead.\n",
            "  warnings.warn(\"nn.functional.upsample is deprecated. Use nn.functional.interpolate instead.\")\n",
            "/usr/local/lib/python3.6/dist-packages/torch/nn/functional.py:2494: UserWarning: Default upsampling behavior when mode=bilinear is changed to align_corners=False since 0.4.0. Please specify align_corners=True if the old behavior is desired. See the documentation of nn.Upsample for details.\n",
            "  \"See the documentation of nn.Upsample for details.\".format(mode))\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vRd6JHN2GByK",
        "colab_type": "text"
      },
      "source": [
        "#Sources \n",
        "\n",
        "- https://towardsdatascience.com/the-evolution-of-deeplab-for-semantic-segmentation-95082b025571\n",
        "- https://github.com/fregu856/deeplabv3\n",
        "- https://towardsdatascience.com/the-evolution-of-deeplab-for-semantic-segmentation-95082b025571\n",
        "- https://arxiv.org/abs/1606.00915\n",
        "- http://blog.qure.ai/notes/semantic-segmentation-deep-learning-review#deeplabv3\n",
        "- https://arxiv.org/pdf/1802.02611.pdf\n",
        "- https://towardsdatascience.com/review-deeplabv3-atrous-convolution-semantic-segmentation-6d818bfd1d74\n",
        "- https://neurohive.io/en/popular-networks/resnet/\n",
        "- https://neurohive.io/en/popular-networks/u-net/\n",
        "- https://towardsdatascience.com/review-deeplabv3-atrous-convolution-semantic-segmentation-6d818bfd1d74\n",
        "- https://towardsdatascience.com/types-of-convolutions-in-deep-learning-717013397f4d\n"
      ]
    }
  ]
}