effusion-detection-using-distributed-ml.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# Analysis of Chest X-Ray images"
      ],
      "metadata": {}
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Problem Statement\r\n",
        "\r\n",
        "Neural networks have revolutionised image processing in several different domains. Among these is the field of medical imaging. \r\n",
        "In the following notebook, we will get some hands-on experience in working with Chest X-Ray (CXR) images.\r\n",
        "The chest X-Ray images we have are divided into 2 folders, one which have effusion detected, and the other are normal X-rays.\r\n",
        "If you want to know more about what is \"Pleural effusion\", read the articles below. \r\n",
        "\r\n",
        "Here is a quick explanation, from the first link below. \r\n",
        "\r\n",
        "__Pleural effusion, sometimes referred to as “water on the lungs,” is the build-up of excess fluid between the layers of the pleura outside the lungs. \r\n",
        "The pleura are thin membranes that line the lungs and the inside of the chest cavity and act to lubricate and facilitate breathing. \r\n",
        "Normally, a small amount of fluid is present in the pleura.__\r\n",
        "\r\n",
        "The objective of this exercise is 2 fold. \r\n",
        "\r\n",
        " 1. To identify images where an \"effusion\" is present. This is a classification problem, where we will be dealing with two classes - 'effusion' and 'nofinding'. Here, the latter represents a \"normal\" X-ray image. \r\n",
        " This same methodology can be used to spot various other illnesses that can be detected via a chest x-ray. For the scope of this demonstration, we will specifically deal with \"effusion\".\r\n",
        " 2. To prepare a training script that can run in a distributed ML environment\r\n",
        "\r\n",
        " About Pleural effusion\r\n",
        " - https://my.clevelandclinic.org/health/diseases/17373-pleural-effusion-causes-signs--treatment#:~:text=Pleural%20effusion%2C%20sometimes%20referred%20to,to%20lubricate%20and%20facilitate%20breathing.\r\n",
        " - https://www.webmd.com/lung/pleural-effusion-symptoms-causes-treatments\r\n"
      ],
      "metadata": {}
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1. Data Pre-processing"
      ],
      "metadata": {}
    },
    {
      "cell_type": "markdown",
      "source": [
        "Our data is in the form of grayscale (black and white) images of chest x-rays. \n",
        "To perform our classification task effectively, we need to perform some pre-processing of the data.\n",
        "\n",
        "First, we load all the relevant libraries."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "# !pip install opencv-python\r\n",
        "# !pip install matplotlib\r\n",
        "# !pip install tensorflow==2.4.1\r\n",
        "# !pip install keras==2.3.1\r\n",
        "# !pip install 'h5py<3.0.0'\r\n",
        "# !pip install pydot"
      ],
      "outputs": [],
      "execution_count": 1,
      "metadata": {
        "gather": {
          "logged": 1621081336976
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import cv2\r\n",
        "import os\r\n",
        "import glob\r\n",
        "import numpy as np\r\n",
        "import matplotlib.pyplot as plt\r\n",
        "import warnings\r\n",
        "\r\n",
        "import azureml\r\n",
        "from azureml.core import Experiment\r\n",
        "from azureml.core import Workspace, Run\r\n",
        "from azureml.core import Environment\r\n",
        "from azureml.core.compute import ComputeTarget, AmlCompute\r\n",
        "from azureml.core.compute_target import ComputeTargetException\r\n",
        "from azureml.telemetry import set_diagnostics_collection\r\n",
        "from azureml.core.compute import ComputeTarget, AmlCompute\r\n",
        "from azureml.core.compute_target import ComputeTargetException\r\n",
        "from azureml.core.authentication import InteractiveLoginAuthentication\r\n",
        "\r\n",
        "from tensorflow.keras import Sequential\r\n",
        "from tensorflow.keras.layers import Conv2D, Flatten, Dense, MaxPooling2D, BatchNormalization, Dropout\r\n",
        "from tensorflow.keras import regularizers\r\n",
        "from tensorflow.keras import optimizers\r\n",
        "from platform import python_version\r\n",
        "from keras.utils import plot_model\r\n",
        "\r\n",
        "import random as rn\r\n",
        "import tensorflow as tf\r\n",
        "import keras as k\r\n",
        "\r\n",
        "rn.seed(30)\r\n",
        "np.random.seed(30)\r\n",
        "tf.compat.v1.random.set_random_seed(30)\r\n",
        "set_diagnostics_collection(send_diagnostics=False)\r\n",
        "print(python_version())\r\n",
        "print(tf.__version__)\r\n",
        "warnings.simplefilter('ignore')\r\n"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "3.8.1\n",
            "2.3.0\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Using TensorFlow backend.\n"
          ]
        }
      ],
      "execution_count": 2,
      "metadata": {
        "gather": {
          "logged": 1621081338721
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Login and configure Azure ML Workspace"
      ],
      "metadata": {
        "nteract": {
          "transient": {
            "deleting": false
          }
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "interactive_auth = InteractiveLoginAuthentication(tenant_id=\"72f988bf-86f1-41af-91ab-2d7cd011db47\")\r\n",
        "ws = Workspace(subscription_id=\"925a5ad8-a21a-48c0-92a2-5a8a2a4dfc46\",\r\n",
        "               resource_group=\"ml-workbench\",\r\n",
        "               workspace_name=\"ml-workbench-eu\",\r\n",
        "               auth=interactive_auth)"
      ],
      "outputs": [],
      "execution_count": 3,
      "metadata": {
        "gather": {
          "logged": 1621081342128
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# check core SDK version number\r\n",
        "print(\"Azure ML SDK Version: \", azureml.core.VERSION)"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Azure ML SDK Version:  1.27.0\n"
          ]
        }
      ],
      "execution_count": 4,
      "metadata": {
        "gather": {
          "logged": 1621081342290
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "ws = Workspace.from_config()\r\n",
        "print('Workspace name: ' + ws.name, \r\n",
        "      'Azure region: ' + ws.location, \r\n",
        "      'Subscription id: ' + ws.subscription_id, \r\n",
        "      'Resource group: ' + ws.resource_group, sep = '\\n')"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Workspace name: ml-workbench-eu\n",
            "Azure region: eastus\n",
            "Subscription id: 925a5ad8-a21a-48c0-92a2-5a8a2a4dfc46\n",
            "Resource group: ml-workbench\n"
          ]
        }
      ],
      "execution_count": 5,
      "metadata": {
        "gather": {
          "logged": 1621081343011
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Download the dataset"
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "from azureml.core.dataset import Dataset\r\n",
        "\r\n",
        "web_paths = [\r\n",
        "                'https://mlworkbencheu2397988138.blob.core.windows.net/datasets/data.zip'\r\n",
        "            ]\r\n",
        "dataset = Dataset.File.from_files(path=web_paths)"
      ],
      "outputs": [],
      "execution_count": 6,
      "metadata": {
        "gather": {
          "logged": 1621081345789
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Registering the dataset so that it can be used across workspace"
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "dataset_registered = False\r\n",
        "try:\r\n",
        "    temp = Dataset.get_by_name(workspace = ws, name = 'effusion-dataset')\r\n",
        "    dataset_registered = True\r\n",
        "except:\r\n",
        "    print(\"The dataset effusion-dataset is not registered in workspace yet.\")\r\n",
        "\r\n",
        "if not dataset_registered:\r\n",
        "    dataset = dataset.register(workspace = ws,\r\n",
        "                               name = 'effusion-dataset',\r\n",
        "                               description='training and test dataset',\r\n",
        "                               create_new_version=True)\r\n",
        "# list the files referenced by dataset\r\n",
        "dataset.to_path()"
      ],
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 7,
          "data": {
            "text/plain": "['/data.zip']"
          },
          "metadata": {}
        }
      ],
      "execution_count": 7,
      "metadata": {
        "gather": {
          "logged": 1621081346942
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Configure Compute Target\r\n",
        "\r\n",
        "Note: We can also create compute target here if it does not exist. Saving time by pre-creating the compute target. \r\n",
        "In the first few samples we will use a single node, later we will use a cluster for distributed training."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "from azureml.core.compute import ComputeTarget, AmlCompute\r\n",
        "from azureml.core.compute_target import ComputeTargetException\r\n",
        "\r\n",
        "# choose a name for your cluster\r\n",
        "cluster_name = \"gpu-instance-small\"\r\n",
        "\r\n",
        "try:\r\n",
        "    compute_target = ComputeTarget(workspace=ws, name=cluster_name)\r\n",
        "except ComputeTargetException:\r\n",
        "    print('Compute Target does not exist...')"
      ],
      "outputs": [],
      "execution_count": 8,
      "metadata": {
        "gather": {
          "logged": 1621081347305
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Downloading the data to local machine. \r\n",
        "Azure ML by default mounts the dataset as network drive which is slower than local data."
      ],
      "metadata": {
        "nteract": {
          "transient": {
            "deleting": false
          }
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "dataset.download('/tmp/data/', overwrite=True)"
      ],
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 9,
          "data": {
            "text/plain": "['/tmp/data/data.zip']"
          },
          "metadata": {}
        }
      ],
      "execution_count": 9,
      "metadata": {
        "gather": {
          "logged": 1621081350199
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!mkdir -p /var/tmp/effusion\r\n",
        "!cp /tmp/data/data.zip /var/tmp/effusion\r\n",
        "!unzip -qq -o /var/tmp/effusion/data.zip -d /var/tmp/effusion/"
      ],
      "outputs": [],
      "execution_count": 10,
      "metadata": {
        "gather": {
          "logged": 1621021695170
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "DATASET_PATH = '/var/tmp/effusion/data'\n",
        "\n",
        "# There are two classes of images that we will deal with\n",
        "disease_cls = ['effusion', 'nofinding']\n",
        "\n",
        "print('Number of effusion images..',len([name for name in os.listdir(os.path.join(DATASET_PATH, 'effusion'))]))\n",
        "print('Number of No finding images..',len([name for name in os.listdir(os.path.join(DATASET_PATH, 'nofinding'))]))"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Number of effusion images.. 107\n",
            "Number of No finding images.. 1000\n"
          ]
        }
      ],
      "execution_count": 11,
      "metadata": {
        "gather": {
          "logged": 1621081359732
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "As you may notice the dataset is heavily imbalanced which should be treated before training.\r\n",
        "\r\n",
        "Next, we read the \"effusion\" and \"nofinding\" images."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "effusion_path = os.path.join(DATASET_PATH, disease_cls[0], '*')\n",
        "effusion = glob.glob(effusion_path)\n",
        "effusion = cv2.imread(effusion[0])\n",
        "\n",
        "normal_path = os.path.join(DATASET_PATH, disease_cls[1], '*')\n",
        "normal = glob.glob(normal_path)\n",
        "normal = cv2.imread(normal[0])\n",
        "\n",
        "f, axes = plt.subplots(1, 2, sharey=True)\n",
        "f.set_figwidth(10)\n",
        "    \n",
        "axes[0].imshow(effusion, cmap='gray')\n",
        "axes[1].imshow(normal, cmap='gray')"
      ],
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 12,
          "data": {
            "text/plain": "<matplotlib.image.AxesImage at 0x7f118e4c17c0>"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 720x288 with 2 Axes>",
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAj8AAAD8CAYAAACYT8UEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOy9XYisa3bft96q7q6Pruru/XGOZzQjWSOQMQKDZEGEnZtAHCuJMcZghBIQsUYgg2NDcMAxufGVwfgiyNaFkHAEEQ52BhScCITjWCKI2BN5rCGWrIgIaSRxZkZnzszZuz+qq7qru/rNRZ/f07939VPde5+9zz5759SCorveej+ez7X+67/W87xN27axkY1sZCMb2chGNvJJkd7HXYCNbGQjG9nIRjaykVcpG/CzkY1sZCMb2chGPlGyAT8b2chGNrKRjWzkEyUb8LORjWxkIxvZyEY+UbIBPxvZyEY2spGNbOQTJRvws5GNbGQjG9nIRj5R8srBT9M0/3HTNP9v0zS/0zTN33rVz9/IRjaykY1sZCOfbGle5T4/TdP0I+K3I+I/ioivRsSXIuI/a9v2/3llhdjIRjaykY1sZCOfaHnVzM+/FxG/07btV9q2XUbEP4mIv/CKy7CRjWxkIxvZyEY+wbL1ip/3mYh4R9+/GhE/4BOapvnxiPjxD75+f6/3ZqYl3cWoNU1T/Xvf703TlE+v1ysfvvu32jX3PftZrl1Xr7Zty+/8n+vl8/i/9nH7+Xm132qfq6uraNs2VqtVrFarcszX1O5dO75O1vXZXef3+/0YDAaxv78f3/jGN6rX93q9Ulb/ltvzvv5c1+dN08TTp0/j8vLSj/hW27Zv3VfnjTybZB32cZZlIxv5JEjbtncr4Iq8avBzr7Rt+zMR8TMREf1+v51MJp3f7zMyL7Ec9x6rGea7pNfrRb/fj36/3wEuBjP+jXP7/X5sbW3F9vZ2bG1tRb/fj9FoFHt7ezEajWJnZycGg0GMRqMYDocxHA6j3++X87e3tzvP3draugWctra2Ymdnp1zLc3q9Xmxvb5fzr66uynWr1SouLy/j/Pw8er1eDIfD2NraiqdPn8ZoNIrt7e1YrVYdsHN1dRX9fr9cC0C5vLyMxWIRFxcXcXV1Vf5yDR+uuby8jMvLy1gul7FcLuP8/Lx8zs7OYj6fx2w2i5OTk3L88vLy1j18bwMl/5b7NwOL2v+1vp9MJvFDP/RD8fjx4/ipn/qpOD8/L9fRvoPBIBaLRVxdXXXGxtbWVhWMGrDiKNSAsPv713/91zP4+YN7B+9Gnlmsw5qm2bw/aCMbeQ3lVYOfr0XEt+v7Zz849trLXcAnosuK8L8NEECi3+9XmZoaMMrgh4+/G6jwyff3MziX4/wfEbeMKPUBFGRm5ezsrICUwWAQOzs7BQT5/NVqVYw598X4UgaDJJebduz1etG2bfT7/WjbttOmXOMyu360Eb/1+/3OuWZaOGYGjPqaHaqxMPl/s1xN08TBwUF8z/d8TwFa8/m8/Pbo0aP49Kc/Hb/zO78TZ2dnnefQd75/fubzMEEXFxexkY1sZCOfZHnV4OdLEfHdTdN8Lq5Bzw9HxH9+30UvwvbUjOJ9LM19QOe+Z9no1lieHLbKgCf/VgM8sDp8dnZ2OqDG98ghslwO6phZDgAH4GV7e7tTX993NBpFv9+Py8vLODg4iLOzs/LbarXqGNzVanVvWwJQmqYpDEyuW9u2sb29HRcXF7fAYr/fj52dnQ7Ioix8AF2Zibq8vCz35/luB77n8eGxlcdMr9eLy8vL+Mmf/Mn49m//9vIc5M/9uT8Xf/Ev/sX4sR/7sTg5ObkFnNaxSuuA17rfqcdGNvKmyTo78KqiAS9bnnex0Zv0EvKPuk9eRlu8UvDTtu1l0zR/LSL+t4joR8TPtm37m3dd86KNmI3Ifec+yzHuVTNwDhetC20BjvI5NebHAGgd65PDWJlx4j418JWNIwyMfwMIXF5eljCa7w+YODs7K8DBIGS1WkWv14vlchkXFxedcsAqZQPPM2t5ObQV/29tbcXl5WVpj9VqVUJuvj/gLIMfrjH4MdDhQ5l87HlyhR4/fhw7Ozvxy7/8y7fCaPv7+zEej0u9agD8RecCY/ZNUqJvumxtbcWDBw/uBanPKuuYxSwvEqK/6/e7cv6e5Vl3fSfsu729XULo6JDJZBK7u7vxuc99Lt56662O/nGo38e5Z3b0sgOY2dWcc8c1mSmuMcI4F9YhnOvwulMG5vN5CeETol8sFjGbzYqjgv7lWqcBnJ6exnK5vMUWr6vrfR/3i3VuTe5ytHM+ai5TrXz3/d7r9WKxWMQv/dIvVfvpeeSV5/y0bfuLEfGLL3ofD0b+zx1XefYzHbvreP49G/pnCW2tGyB35fw4/8YfK4i7wFb+vi5PBHGOj8Gac0+cN8M55OvkMBOsi9uVweuycA+fk0NfOTRFu1ixANK4nrLXco34zu/5k0FQzkWykls3dprmOrSVQQ9/h8NhXFxcdFgZ6v0ygA/32wCfVytN08RoNOp8r/1/n4G5ywnLch8QeR5Act/32j3zsWcBRjs7OzGZTGI4HBYm2c7kZDKJx48fx/d8z/fE22+/XRw+WO+9vb2OjlqnZ9fpRc43mKmFuc2Q51zBiCi64OLiIi4uLjog6PLyMtq2jYuLi+IIHh4exsnJSVxcXMR8Po/T09NYLBaxvb1dzuW5vp/1XL/fL/mS1lfIOnBRAxa1PjQAyna2ZsPWfV8HtNb9XVf2s7OzeBny2iU8P4usa6S7zn0eqSkiDwYMvVkXA55a5+dPBkiZ6eEZnsROYK6BHjM860DVusFeY3pgO1wnjHP+n3bZ2dkpAIlQD5OHtsogwu3rfBzn95gd8nezT24zgEwGS2ajUB45AZq/GQBlIFQDQPbUspB4/qf+1J+KL37xi50xOh6PC3N2l7HbyJsnz8PQ5Gs+LEvzvEzMfY7h8wCZ+861TgK87O7udhhlQBA6o23b2NnZiZ2dnY4zhg5zCDkDFz8/O818t5OVDb7D7j4vO0QRcUsnZkcKgLJcLmOxWMTJyUksFotYLpdxenoas9msLNg4Pz+P5XJZ9JiBlPUS5cEu2LEz+50dr2xH3Z4Z/OT/83345NSL+wBOzYbXHD7/f3Fx8VL04xsFfmqA5D5w8zxsT+05ecDkxOP72Jsa4PEgqQGlDH74HWDhFVwGQpTbbEdG3tSjNvg9uPFeADS+zjkyy+UyIqIosnxvT06kBh4or1eDRUSViub/DOQy7Q37xPkkZ2eGx2Gz2u8ZAHlVmOuRzzGL0zRN/N7v/V788A//cHzxi1+85Unt7OzE6enprWTkuxjNDwPsN/LxyDrPOZ+TZR0Auo/xqR17XkDk/1+ETUIHobf4mKkB5KBjR6NRB+Q0TRODwaDoGOtGh9YzO2wQUzO4+Rz+t06ywGDnkDfSNE1hkymbwRQ6jufeZTNgfwA+3D8zNHYQs3NuPUW5AEUuM38Hg0HUJLep/7o9KSdlXAdgch88iy7jnLRS9UPLaw9+PPmz4ljHzryorJu4ZhUyYKmxN5kNykBnHfPDMU/8zPigPGqKNNOR+ZPF55klycqE53kiwAAx6fN1LhPeCR6R2Rvu5TJxX85x0vM6AJTBD+fwG0yU2R8AF+AnA567mCADITNmOT7fNE18/etfjy984QvxzjvvxOc+97n4M3/mz8RXvvKV+OpXvxq7u7txcnJSynefbIDPmy2ZiciMhM+56/pn+e1Z2KH8/S7QcxdwyvMvO2uZnbaus54bjUZFj/Ebhjkz0sw7AIcZDMpYAyq5TnlRRc7pMSjKjE/N2aPsETdgwGWMiFJfdBK5isPhsLNC1uyM9TGOW62stQ96C8eVaxHKl9toHQDPdsW6HD1dA1q1e/l5NQLC4OcTwfysq+Q6duZ571M7z8h5Xfw4g5faKq0au1NjezxRzA5FdJOoa6Eu2oL7wFwYGHFOBmLr0DleRwZvrPbKYICl7igvrmMiWcG7nn52XmVF3fN1Dit5Uhik5mRkn2vPjX4z+Knl+nDPHG93vD/vGeR7wy4NBoP463/9r8c/+kf/KL7jO74jPv/5z0fTNLFcLmMwGMSv/MqvdCjqmqJ4UdCzAU0fn2SQwzH/zf/fd7+Il8cCPevvNcCD4Tar45SAzMxmp4fzcfBs/A2inF9nIBURHWcpG2XrgOw43ScAIYfpaw5jBgm5DK4756N3qPfOzk7RGSRx006Zocr6KQOyzNZk0DYYDAoTRDiOcywsxHBqgEGg61kjKO4ap8izAPaXRW4grz34ybIOFdbOex6h8ww0PHFrwCGHqPJ5Hrjr2J58Lxt/mBJPjuxV5ee0bVuWvWeghnhZfER0PAyzFxFxqz1QMEwUlBFL28fjcWfyAJjw4gAaLEPPzEnuE2hjvrs+Vgb5Y7rZ12agl73IWo5PXh1WC4dlJol789twOIzd3d14991345133ok/+IM/iKZp4ru+67vis5/9bDx69Ch+8zd/85biybIOFLlO912/kVcr9/XZ89zjeYzC8ziFNe9+HSjKDpHD8XbOcmiK+W0Hj49ZbbMhPANgQCKxw1zWBTgp6LHValXyhDL4ibjtFK4Dpz7Xx62vATPZQfPvZkRw1ghPud7oEdqspifNRBnYGCBmIJTD9HmhyGq1ivPz807awd/8m38zvvd7vzeWy2XM5/Oygezp6Wl861vfin/xL/5FHBwcxL/+1/+6Mz5qDFJNnnduvCwd9saBn4gPTwdnyUxEnrj3gZS7PnlC1ZieGmiyABLM/jj0lp9n6jQnOFPffr/fievmiZQVgD0FJgT5QNCm/D8ej0t+EEoKz8RKw+ExvrstPHGz4VgHYmgvJzrTVx4bgCIUEMdz/9bAT1YSmSFymAyASB0Gg0Hs7u6WJMbLy8v4Z//sn3W82OdlJ9f9lttlI6+HPGt/GLDfx9DUznuWe+f73sX6WE9mkHPXZ51e9Bxm9/gcEgMEwJjCBsFyLJfL2NnZKUnSedyTw1ebK9kx8nU5jJXbC32SmSADGzNFnGNG106b9bh112Aw6LA4djYzY+Uy+Lk5HEa5DX6su7gvS+2Xy2U0TROf+cxn4ju/8zurOvfo6CiePHkS3/u93xtf+cpXCjA9Pz9fC9SfBRS9Ct31RoKfdXKfAvDgvG+SrgM8NtLPAn5qTM+63yw1RZOVTsQN/cqgMkNUY0QAIvYkmXi5nHnSWCGwcoHVCIPBIB48eBDj8bj8xislRqNR8W4ybWsPIz/X5fTHfZnPca6PlQLHeAaJ2vxuEGKvi78c93iw8kDZokR6vV6HIt7e3o75fB6//Mu/XMBpjaq+C9TUFN4G4Lz5ch/Yyf/naz027tOB68BP7X9YF2+rUXMIs35apzc9R3PeYp7f6KnBYFBC6oPBoNyPFVOAJLcfbCttw7ylTmZmrXsNDKyPa3Nta2urswwdoX4R3cRc6y0/F4cM/bGzs9MBP+gQvvPMbDOyw2r96sUX/s3AzPqHfh4MBtG2bfzTf/pP49d+7ddia2srxuNxPHjwIIbDYfyJP/EnSvv/6T/9p+Onf/qnizP8D/7BP4gvf/nLHRuyboxah9fGYW77l6Xz/n8Bfp4F9KybnNkjqdG2/F9jge5ice5ifTwZXc4MzhzeqjE6/M8EgSnKDAmD2isAXA6Dm6yMGMAZEOGR9fv9GI/HMZlM4uzsLLa2tmI6nRb0v4729XF7mF4ej/Ds3H65zL1er9DcHEdZwsZ4zFgZcgyglMNdZoTuCntRB56LMlsul/HkyZP43Oc+Fz/wAz8Qq9UqvvSlL8Xv/d7vlTJ4Lw8rhGcFRbV5sQFIr4fU+uNZwY6dgbvufd9v9zFJGOPBYHBro8GarsyOZNaBDnvxt+ag+fmUgc0OCdMTXnMIzSDCLHYGhE66zSwO+tDnZ8YkMzr5OuuuuwTd5jrbNkXcABT0mNkZl6nG7vv6HBJzrmbO2zEgYyk5uuzf/Jt/E//yX/7LODs7i+VyWWzH3/t7fy/+5J/8k7G1tRX//J//8/jH//gfl3dM/v7v//7aMZzlPvv9ouevkzcW/NynDCJirUdSY2PWAZ51oIffI7r7Vhg0rANAfDzhIrogjfvlpGvn6jDpMcxZOWRPyvto8Dw/H4DgEExWHCjGra2tEr4hoZeXijZNE9PpNPb29so9DVwy9csx6gT9a2Vgz+VZgJs3WayxKpyHMsrAzn1G6M5J2TUGCA+uBnKHw2E0TRPf//3fH3/1r/7VsvHdX/pLfyl+8id/Mv7Vv/pX5brlcnlrmeqzGrcN0Hk9xcD/PhBUO1YDKs/63FwG/q/9NhwOYzQalTmedeM63bkuXYDvdtxqDhx6xoZ9OBx2HKwMxKwz/DfiJlfRuoZ56jC8dbb1jFmk7AjSXjlsnp3G3Naex/7d+gTQU2OzvLgiX2dHNzNAXO9k6tou9rY56LW2bcvrg7a3t2O5XBYQxG//9t/+2/ja174W3/zmN+P8/Dw+9alPxeHhYafuuWz3gaD8+8sCPJbXHvzcRZmtaxA6KgOBGvuSPZecS7NuMhtE3BUGozz5GuqUPYC8JLTmUeWJ1etdbwqWV2hxHrFxzs9Jc568Psfenf+PiM4kYkJCy9p76Pf7xdCTSFebEFbG0OymY91ergPPoA4GOuQT4MHwlx2orXxdLofiavR5zgOyNwUAAljRdoPBICaTSXz+85+Pr371q/GzP/uzsVqt4kd+5EfiR3/0R+O3fuu34vT0tChnj+38jPuA0KtQHBt5fql5wrVj667LLIT1isdoZgjW3ce/9/v92N3djdFo1HlvoBmbmvOYAU3WnXnlV2aNEOf5obP45L3V7OjgeOX24Dy3mZ2pmuNsnch5OcfGOsj3dp4h+gQnzGyR71FrT8phHUBisvfnsePotrdYV2QgVMtZtPPN86zDDEJXq1UcHBzE5eVlbG9vx+c///n4s3/2z8ZP/MRPxF/5K38lfvqnfzreeeedW+OwZsctz/r7y5DXHvwgz6LAMXj2MtaBHQMLeziZps2Ax+d4EGcFwXURt3fOzB6L75s/GYBkgw3Qs6fm+rJKAmEScE97FAys7O0ANJyYTLlRBJ5I9hZQDJTFTIYnosuWlXNuJ+7tPKIMLlE2KM3aO8VcP3tCWTFaKTm85SRo2jJ7tI7lf/d3f3c8fPgwfuInfiLeeeeduLq6ip/7uZ+Lv/N3/k788T/+x+Pf/bt/18lXqHlyPPeuvS5qAGgjH588i+d7HwhCtwEIcr4M15BzsVgs4uzsrMOm1AARwHx3d7c4UDnX8FnBD/9bn+KA+FrrJyf08hedRn1zqMw62U4J8wxBR/J/jZkx84GDmPUQYqCQHZQaS4SuyI7qOhDLnj0RUe5v/YpDSBnXOeQ1pzKzTznfEL2Dc4jesh30ext7vV5Mp9OIiPjsZz8bX/jCF+Ktt96KP/bH/lj89m//dtmnqdaONXkeUPSynLk3BvzcJxj57AVk0FJjeDyxPYmzF+H71IBVDpHlwVijPfnNuT25XC4D1xqp8zLAGnCzMojo7huUw0pMzoj6CgLXPS/npGyE1mrJhBE3LJmfGREd4JSVtUEQAMNhIc6hPojLhhL2BEbRAIz6/W4SssEQz6cNACCmi+0l+cOKlbfffjvm83kcHR0VCvrk5CTm83l8+tOfjt/93d8tSi5PcupHGVGU60DQBgC93lJjX/L/zJ/BYBDj8biTg5PHGdcwdsm/411RjHXrHzYTZJVmBj41HRnRzTVZF5JapytrOjHPe5gHl8e5jDXmyMfM/tpxov6+jwWgUmNqch95s0Kfi05AF/k8/rZt20nUpt8cjqNtKSerqGo5ixmM2inl3AyActsD/nAUAdGw5H4pNc/61V/91RgOh/GHf/iH8ef//J+Pf/gP/2H8jb/xN+L8/Dx+/ud/fi34MWgzQ2fd+yyM6IvKaw9+7kOCeAWZ7clg5j7gY+CQkXQGQjWgs+4YdcjGyOWvga78XCsMQAOeDqsgPGEAPr4Ocblc9rzKKRtx2joiynuoIqIDjGygoaRddwDIOmo+K0Uf51yAA2WuUbvUzxPJXmleiUGZAHbO8fE93GYGcW63PHZ2dnZiOBzG5eVlWS3BM6fTaUyn01itrjc1yytYstLKigNl5XZYN3cyo7WRVyfrGJ/8u8cqe0OZkck6It/PBmQ0GsV0Oo35fH4rWRVAnkGGQ1VZN9V0VT6W9V8N9BiI+Dvjs8ZwrUuQRrKD6XnK79w/94XLaQDm/kBgZbPOqfWxwRdgyMwNzvrZ2VkBP4AjruEew+Gw44zx3FrboqtrYyzrSH+nzQFZBkFnZ2dlBS+g6Bd+4RfKub/wC78Q7733Xvztv/234+LiIk5PT0s5EbYrGI/HnVeVeNyyio833Nfe4/WJZn7oYG+wVWNM8kTOk7gGlO4CMvl4jRKOuP2GdE9Gys9fFJHr4Ht4gnowcQ0rvAzcdnZ2OrF7eyc538dx7V7vZhdne3b5VRouG3UjQZkJWWOGAEX8NWthryh7NZTbfe/yeoL5fp7g+VomPUwVHyd353P90tEMlL2ajHLzP4mb7777bkRE/OAP/mDMZrOYz+fx+PHj6PV68fWvf70ohJpXWPOMPA69S+s62QCfVy/3Ke4a6CEMlfNvaqy22Q/uw/jAWO7s7BTP3WM3AxuzODhO6Kf8Lq7MNFgvUo8MPDDIBvNZd5qB8AIOsyDcx+PfDlt+NuVZBzw5D/21rq8Q6xzrPzuM1kNmkzIr1e9fh/Gtu91PfAaDQcfJyrrTutfPdb0zA5T1o+0E9tXjZLValV2hASawz0+fPo1Hjx7FYrGI1WpV8jwp7+7ubmEYc5/X2pHXb/CSV4D7y3TgXnvw44q6Q53g64mXadgahWvDlRVBDdR4QmVQlNkWD8A8GD0QswefwYX/cr49gn7/ejUE7eAYeV6pEXEzwBmsTGDaDOMJoLDSAxB5IlM2jC1LY5lAnlhuH37j/p54HM8TlGPsK2SlQD3Y/4JEPZ6dFS7PAvQA2jy2bDwM8PjNYIT7eUw4h4oy9nq9ODk5iV/5lV+JH/zBHyzbAAyHw/jFX/zFODk5KcreYS9/HBqkHLSNWSBWYtw3nzbyauSuNvdvOzs7MZlMYjQa3WI9aqxPdpAy4Pc4NjtoViQ7YDyPMZsZp/sYajsHWd9RLr57rvO/9Zg3QDSYcbkpD3MiMy0ZDOS2yp8MXNb1ocEW5Uc3whajUw1U1rFD6DFv82Fdia5A15n1zuPA9aWsNaCRQRDlt+PKdyfS0741JhqGCp3k8ZXBKuV2ORi3l5eXxZaNRqNYLBZxenra2Z/tReW1Bz9ZoM6MIP03J+vl5Zp0hFcT2Oj43Bxfj7gZIKbpIuohhvzJgzQrEhtn3yNPzn7/egUVitJL2J3jg+LLSo7JaSAD3elcHM7J9CTl9IRgs8OcW+NB7T0uPPgZ0Pm+tAGAgue7bfCU+/1+nJ+fd9rfq8VM72avyoyP25sJ7msAS05ytseX4/YZaH/xi1+M+Xwe3/d93xdt28aXv/zl+NKXvtQB6ihCt0cGXwZB9qIZ/3ll3UY+XrEjxHfEnnF2ZrIeqwEOGxHuXQPOeaWlx6d1JmMwH8uOWnYaMvjJ9TU4yeVlDvn5mWWy8xlxoz8zgOKeme3JgCj3hZ0ps6v5Lw5dbuuIuAUW8jy1zuU650kOh8Pi6HGMsgMKrUO9cINjWdxe/O4y5TpkfWHnN983J3RH3H4lEufatrq8bjvbCpx06j0cDmM+n8fx8fGtOn4Yee3BjycS4Rzv9JknLEgxAyAaz8wIx2veTER3D4gardu2baEAMzVXQ97+zqBnQmbA47rbA+j1ejGZTGJ/f78zabh/bTL4WK/XK/knGRi47mZBPFmzp+NVUpPJJAaDQfF2/BqN2sSh7i5DNu6EdGw8PPmtHMiZcT/VFHBNQdfa3auq2rYtwMaUt8uRvUaOe8VK27bxG7/xG/Hrv/7rxSj1+/0ypr3HRy63vSI8eZQt/UNf9Xq9EjvnHhv5eGSdUYm4Nix7e3tlJ2MYj3XheuYN8y87T4yB/DyPIY9bG6V1ycs19tzXRkRnLvO8DMZqoMMOn8th9poPgAiQkjczrTmLiPUk3w0Usi6oHa+Fz6l70zTldT7rHCyOeaUYLLVtnPvL+hsnFXBkXZPBrO2G+5i2pj7ZgeIe1jcZbKKT0T1+PxjnY5tyO2bblMctf3HSyb/M+WjvvffeC+u01x780EiAGseBTdHu7OzEeDwuyVQAIecFZbaIe5uSQ9yRmbHxABsOhwWMLBaLmM1mJaRhBeFJzuCgHhzn3kwUx4gp53Q6jclk0kHwEV1Un0FcTTEQQ3bibwZAtC/38GBzMnDENWV/cHAQTXMdnz45OYnz8/Ni0F3WDByRDIDMvEREYZNoN8CXlc7Ozk5Jzuv3+wUMrUsIpr5+A72ZFpehBpTyOMleew2EuS2yQTo4OCheI/UATPK6EJQCeRz5TcuARtppwwJ9vLKuzQeDQezt7RW9ZhZ6HXNto4FRsI4CUGfDb+PCMT55AUgGPgYlBj7ZyfAYz6HvDIAyMIuI6nPQj9ZF2AOXIf/vutd+M5AxQ2HwmO9j5ig7SojLXMulpB3ySk30DWCCHeFrLJrLkHN/rKOsu3P93Xc5ROq/dqbQS/SFAZLbosZE5rJ7TPMc63F+r20qe98u2s8qrz34IcRjrwjAw9/hcBjT6bTze94gCy/BaBOpgRxPNA+emjeztbUVo9Gos7oixyd9/xzD53f+GrX7GjzE7Hm4fjWPI9fZBpJn5cHKcZC3ETr9Qo5PRMTe3l7xSIbDYTx48CDm8/kt4JABELR1bl8zPngBgJscwsyhvNFoVAw+oIEEPSax25pneA8dPoChrMApK2XY3d2N6XTaaRPGltvNfWsGB2UznU5LXcbjcUn8W61WpR6AIHI5WMrsfUiyIQNIbeT1kJ2dnQJ8CHU5/F4L6VsX5bBUDZAjNVBivZBBlcGP9+nJ7Hg2thbrCztd1m++DnBh58bPZd7jcObwcq2N/Ft+fmYcrJ88VzNoyyDJ1zsEh/NmHWOm1s/KQMR77HCs5hxT55xsncFCtguIdbHtDWV3grUdQesZ7sKlhyAAACAASURBVOP2qYGqWvu7bDwTAJQZL67DmbO9+rDyWoOfpmlid3e3MDgGNaBjLwc1LVZbLnl5edlZNZAbNXsXeQK5wbnGAxtAwG6+i8Ui5vN5YTeY2KZva6jeoAoKcDqd3kLyPN/hM55jD4R75mfkODRgxHXHSEd0Y7+UzawJYO/q6qr0D31gRcI9GNwAD09qb6jINc5nyorJHgRlpe3IE2P1QGbmqHdWPPYyKK+BDYB0f3+/gB6DEP7PidPOCzAg8UotlBDAEw+I1RYsBSW8uFgsCpAyyPNYBzht5NVKZmH6/X5Mp9NbTp0/65hq96fzYgil2MBm3WJDE9ENxxt85JQCPz+DJRvW7OTUmB8+Wf/lZxlkGOB5FZqNqNngrN8ph9skMza5n/ju1ADPYYfis03gHPeFwYLtiusfcbO5ITrUe4dZ1xnc1r5nsFobk9b91oMGUwYm1l0Z2PJcj7mcwpCjGB4rBlnozTzemqYpidAvQ15r8NPr9Tp7XGBIURrj8bizMiIvfa8pEE9iGtQTh+sMAChLbbB7wmB8vFR1MpkUJqht21IXT4B8Pz93Z2cnptNpKU9NGdWSnPM988D024EpD5M6AzsGMEbVyoR7+5ncA6Vcm3x5YmSv0PfKAI1y5nt70tE2LiNeNmyQ82E8yZng3umUc7nv3t5ePHz4MEajUVEYzvlymSijFREMEwALYEV/2qty3Jvjy+UyhsNh2b+FlV6EGil7LkvTNDGfz29Pto18ZJLnNMnNdg68UnM0GsWDBw864XrvPkw/An7Ozs7KvPQy5JpjlctkliDrmBro8RiuAQkfs77Mcz3fw0CGZ0bchLsAPgYLWb/4WegLAynrFSTrD37PYM36iXtzvdvY+rfGeritHEaLiE4I32CG+znJmvzBDGgzmHS/cG2un3V61rVe2EEbUo7aWFrXdlwPi4fN5f2Q29vbMZvNbjFM6OEM+l+GvNbghwGPAkAxYMT29vY6QIcJgXcA8PEEyRM7L5f3YLAxzV6EB4mTvIyQUWyDwSCm02lnGXIecNQXweDt7++X3ymvBwKKwQwFZYYdwItgaXzNG2RyZSCQzzNbxH16vZt3ebXtTegLZWwF40mVV8rRloSc3K6eSG4jvBX3nVmss7OzWwqIjbbOz89LyIg+9rUwUt7ltNfrxYMHD0qYlTFgTwvvxLlOWelQflgcFB8T37Q553q5Ml4hK7sATOfn53FyclIAkBU2AijcyKsT+hzHyKDHq1l2d3fjwYMH5bjzF80cWyc2TVPA/HK5jPl8HrPZrDgONowefxl4ZBC0DhRxLZKP2RBmvWHx85m7dtLsiHpTPECjw/YZgGV9XQuJ0S+ZWbmr/wxK+J4d2Qzssj2xjnUkgr7N4NHPzYDJq2D9vBoYdT1chhog4hr0kX+32K7mfs9SA9A44WzzgL6mXufn53F8fFwYe8bEy5A3AvygHNiKnTcPTyaT0kl4Ad4jwhni9irsMXiQELrKZTAo8OBmcJiytNLAgI1GozJJLy8vS/6FB5UF4LS/v1++m9oFZGUvkDpgVMm54R4YWIw/7eLcJLcJdc3ggrrxf84xcPjI7Wj619QpgMn3zyE9gIcVFR8v57dCRFGui7Xjfeck9QwACZvlsQXQybtAEwJzedwuOV5OW5DcTI4P59Cf29vbnQTopmkKC0qOE+G9bERc73UKfiMfjdgznkwmt3IXcUxghGxQEIAu44ddd0ejUSc1YDwex/7+fgHVi8WikyPGGMwA6C7QUwMT2WHL59bYXj8PyTrQz0LXGfx5MQFOn53TGjvF/c2645gZ/Nghdd/VgEgN0LkO2U7U6pltTcS1DRmPx7G1tVUMPv2fwaptQC63y24xUMy6IPe1/3ef2Ha6/+3YOm+S/vEzuI4V2E1zneLivEaczt3d3VgsFnF0dBSz2SxOT09vD6wPIa89+HHuiGPkBhWAHysUQElt5U3btkUh5Ix6Tyx3sD2ffNzl9cRjcAKSTG17q/AcX+31eoXVuri46FzHhKHc7M8DoAL8ePB4WehisShG0x4VW6ybIq4NcIMflznnGJgWpm1oezMfAJWax5e9mgzMDIK4B16Bw10OK3EtAGRraysODg7i9PQ05vP5LWXXNE0xTHjZbmMbtpr37N+oLyyO28j5PQ6FwQDC5FBHwCLjC4A+m806rzupAZ0N+Hl1YrBAjqLzetBvOfcHZsi5MO5P67ymaTohV2+fwJhljuaN9zILwl/PU7PbdhZrYz7rRp8fcXvDvcw8+DnoPFa3WgdTz8wYGWjUDLfDy8wz6+iszzNThl5y23jn9+yUcR7AKjuMMOOOPrittre34/z8vLyklvmf+y4DsnUfs1AAmbyRJWXIABlBX2XGLTt1PCMzR2b0WKGN80ddYTHRgywmOj4+vhXO/7Dy2oMflIBzfwAoFxcXcXBw0GF+DFBoYPIxvG+LGREbOouNLfc0E2UlluOr7mQjYpQSdSI0haHr9Xqd5Ga/vsJ1gZVg0yfvfLxarWKxWJRwF8/3Crizs7NYLBYxmUwKA4KXgZKAJcpUK5M1rx5COWRq2MeyAjHw8XWuC9dwvpd9m+niPBgtjxcrHMetqRNvKF4sFp1JCktGKBVARkiJ9vIY8TjyknRPZsKSGKIcwjMD5fag7QeDQceg0WfknBl0Zi99I69W0E949IxL5zHmzVbRD17RZeMJaMK4EDLlecvlsjOWYJdsqDJjaT2YWdAai+D6WRh7ngvWHRhFO5VN03Sc0X6/3wnTO38H42t21IwyesDsMQbW9beOoR3tILl+BjcGXq6v53b+3eCM36zX8r0Qys8Gf7lMtI1zhWoAiHs5h4xyZcBmVp5n5fawPs8A2s6lnUPGNv1Kf15dXZV3Rdpu+v6kdUwmkzg4OKiOw+eVNwL8mO3Z2trqJJleXFzEeDzugA8MxGq1ivl83kGK2RuphbmQbMhgjJbLZZyenhY2ZzKZlFVeOQHPnhrHmNweDOSfwG6hsLIiwojC1GBAeUbTNCU04mMoOzbJYh8erp9MJiWPJOJm1QGKodYW9rCo8/b2dmfyeJLUAKIVC/eKuFFOVvy0kfetyZPMjMp8Pi8JdgBmgJgVEOXY29uLiOuXtvb7/c6btH1vQHBEdJStlSC5FzByzu3J7cd3K0Yrm+z18z8GlGPL5bIsBFitVnFychInJye3xt66kMRGPjrxrvQGPrAzsEJOBjVoIayFHjCQN9PMX4+hvHomh6Ptnd/F7OTxV9MH6BnnlPk6h0yczIxOREc4NYHfcED8XimzHQaBZvA5Zj1lx8XtaH3tD+WthfMj4tZcNuNE2XBUABjrAEReEIMdHA6H5d1ZntOABYfLDVwYax5XNX1jhj2zOrm+dvjcXlzHM+hHbKMlv6rIY81t7H6569U9zytvBPhx3Pfg4CAeP34cTdOUMEUe1BgC53Hk3z04a2jbBtlipQL4OD4+jtPT05hMJiURlg704PA9jfwZlM7D4f4RNwNtPp/HyclJoUABGig7ex3eF4fQGYb/+Pi4KENA0OXlZezu7t6it60YzdBkhM9vTEC3uQEUk8LlzCAUZcd5gAjTod7cz23pcQNIsVdnps5J1fTN/v5+5wWjlAPQ4knttomIMhYODw/L0nMrwmxgMs2dFTltRE5WHsckAFKO1WoVe3t7hUWYzWbxta99rTOuawptIx+tkDOW8/YAPTA9OZ8CY44Bj+jmrljHGCCb2cx6J6K75JvzOTeDFu7rZ+Zwkg01QtkJxWcQRf0pC/qEecs5ZkPJIfFmny57ZvPdRrALACwzuWZfADe0IQ4n+gN7kkFiBjxmUPJiBc9rb+JHmclphV23Iwj4s+6CoQYAofthiA2mKFcGUC47x7JzaFBiYJiZpojbG+0acKELz8/Py1sA0NXZQaOtDIpcnxeRNwL8sELi4OCgvNYBNsh5F/w1G1ILRYCUnaCMZMCTwUr2yo1Yr66uyuaGvV6vhJOyR1EbWNvb2+W1HEwMT6j5fF5Qf0SUweMwDedzDLDDpBkMBtG2bQGNMASnp6elLRwXz16CQZ/bxclt9uoY2NmL4h6ULefpoJCJ/56fn5ecHOcsuByAHxQg7e68L1g6xhUAyGzXYDCI/f39Mq4AlAYYrtvFxUXM5/M4OjqKJ0+edDyzmleTxcdNW2cDZaXpNvQ4ZSUk7fnw4cMYDAbx9a9/vcMCbcDPqxUME8CHcYYOw0gzB83qeMwBAuyx5xwUjlny3EMMXmpOmp2eGsjK4t+YY5TVbBAsBvrCoWiuzQYbg87qzAzkAQIZFDq0wqIBMxI2yBE3ieW7u7sd5uj09LQDgCxmO2hDjllfwdoDhAAqZrexBaQjZN1IOBAmmTxAwKKBW2aeDFL8G8/O7Avg0DmO2DfqYBbT+bW0JREK7m320s5yRNcppI3dLnmsvqi89uCH1Qtvv/122UWYjnFWPAYzxxnzBLf3ZQ8mD1yDFSuWiK7RcafakF9eXsbR0VFZoYax9cTx/3iAlJ2ly23bxtnZWZycnMTx8XHMZrOYz+dFWdpTQ+bzeQF//X6/TGQrGgYmuSOE3HJSHZOpZjRN3wJ4zKR4smUxW+NzUDYoubZty6SD8QKwWKnjEdnji4iyQubs7CwirkNae3t70TQ3O8WyYyh9QF/OZrOI6O7xRD/Sx8fHx/Hee++VlXUGjJ6kHl/8tScF2OE37pOTwvMYtac1nU47DNrV1VU8evQoBoNBPHnyJL75zW/GkydPNuDnFQpjxrk73o8sgxrPA6QGOBg/gAt+514G0Bb/nsFCLdSAAadM2ZF0eTNrTHkAe5zLnOv1esWJMWAzM2vGKxtunokBpS3QKdQLltT3N7gzWGE7FZgmWKKmaYo+iLgBW1nfmX1yPuJisShgx6EvJz7TbnbuMoNtBpo2vby8LLaQcDzPID/UeiIzPQaa3M+5qG17sxLVx8xC+1q/psJ62sCWYw67uU9sg7HPOLcvS3+91uBna2sr/ugf/aPx9ttvx3g8LlSqBy3UqRNQDWqY1I6150mcvfJaLoYNuJ/PtV5R5lCJlyDzBnZ7DwxwQAiDhHssFot48uRJvP/++3FyclImOsyDUTusDXUyi3J1dRWz2axDITI53IawaZ7QlDMPZIflaE+S8sxYuB3tieSk44jo7MAMK3V6etpRgm5/Txr61Yo007xOPp5Op/HgwYMYjUadSYWCIRcrK93Ly8s4PT2Nk5OTmM/nBThRrszaUMc8rkwh5999juP4+eMxY+OytbVV6tk0Tdkok/DFRl6NMB6tG/zaHesnzxEr/Hy99Vd23myIHJJ2iDWz4RjqiO6b0rOnbaDGM7LUDJOfyQKMfv9mDy47K5yD8UWv9Xq9ogMMPDDKrgtlcztxf8+lvOLNzg1zvNfrxe7ubpk3tJNBH/3m+wEYyM+07qTc9ItBKoCQVbmTyaTYAZhd6hERJVWBcBn9zT1JjcjsivU7OtcLMty3dmTz+M2hUOup7LBxLrak5lRTf/oyRxnWjbEPI6+1FtzZ2YlPfepTxUg4B8NGDQRsqizH1+3Z2DARwhgOh52VOW70bLSM+FFQpgoZcJwf0d392aEJQnhmTzDSTBwmIPeIuEmK8wDKg4jzqDM5QzBRGHJ7EYhziswy+HkMYPKVdnd34/j4uCgtduK0Und7ZnBEMjOGG5YLxodJk99onGP9y+UyRqNRZzww6ZwwByvG3ipZkaLs8AAp4+HhYUdx0xYoOIMLGyUrxuy5ck4tDEZ5HPKkzRweOT8/L6FWXnlB2e0Nv6wdUjdyvzCnnWtidjEz0Tgndh5sRLhnDkVlXcX8W8e+YuAziDYbYJaTe1MW672ILqPJ86l7RPeVBhxjnNrI4XyROgDYIuxi8AKIMOjwLtCcY4fPzqDnIXU2E4vjAKvKgpAcPqo5x+gSszjWXVznTWid2+rd56kD1xMSA4wAGKynWEnsuruvnVKQ2Zc8fum/DMbNWFJWj83M+DDuTBpkhx0yAGfcZV4Xgfiw8lqDHxqSvQDoaEJdGKCIawScO9mMSKZr+Z+lzAz60WhUDK4ZhXUMBgPBXpI9eseJGewwQQYhZm68ookl6+SiDAaDwq7kDb4ooxUSE6Tf75fQEctg2UkWAES7GNigLDxYecZkMunQ1kziiOgoTlat2Zi7HVerVQdMwJjl1VGLxaLEnGkf19v9OplMOgqexD9A2mQyKcv7F4tFyRFi4lE2GLblchknJycFUDBRM5A2aM594THtv66rmS3GDfekjW3oUIwGNpPJpIA66kfYD2W4kVcnDncZ9OSPdUhmcAx8DZJ7vV5hlDPbmI2zGQqeyRjKIXlAuR07jjP3qFve+NNMC/+7fH4m9+D8zJJQF8awdQYhHepqA2vnEHHb3sdaASbQhRFdNj2DHhvpHNays47eoH3NaANu8qphO7DcazKZRMSNU21dQXuORqMSquJctjgBWJgRysDV4yyD9NxuWafntqL9PI7MMLrd6TecYIMz3+9F5bUGPyBu9gUwa3J0dNShIe0teWKDvBE39s7OTllh5Q7glRL53U/+n/s4jOYBYsUT0aUbjeTtDRGagnny6ymYxCgajKaBC8CBCcbzYAPm83knxHV4eBgHBwcFAJ2fn5fNDxl0RuYAqaw4aHtAlWnLzBhlZQTw4T4oANf78vIyZrNZzGazTnzfzFemR02xcx65FvbeoH0BSKb16aeLi4s4Ojoq9XLoIStpr0DJRszjhPfSNU0Tp6encXh42Jnkbqfs7RrU076wd+wgDA2OMj05OSlj4WV5Thu5X+gnVg/yYRzwW83IcM06o8N35h0bGmbQY/bWIR8np0bcAHQDF1+bjaIZDpxS9DTnUIc8nmv5RbCu3q0eEMILotFFzEsD/xzOzd+tg23sXT8EsApQtG2xTs//O8+ntijDDE7O86NfMPzUaTqdlnnL9WxmmkEY/eLvg8EgTk9PC+ixo511mOuZw60mEWyTLBl42g7lFA2eBQHAlgRmemijzHC+DHntwQ/Ah/1vrq6u4uTk5NYSwYguxQwwwZPPSyMxQLlTGRCTySSOjo5KWQx6QK1MLiuubBgRDyxyfJyrdHV1VVYzeSLwbBQNE4MdL5n8plQ9sPv969UBETdJ1d4J+vT0tCSS56WXnGOPivJ4hReMBIqCa80EcV3EDTXvkI8Vx2KxKAmCvKeIVQNuTwMgP4Njy+Uytre3Y29vr8TtaZunT5/GYrGI8XhcXiJpxUAfEn5EwZoJMxjxGHTyoj1NFDsMHu0Ehe3VWHm80a/eEM+elBU7fcbYYPWKDeNGXp1gSL3flIGOQ1iABa8Es66xbjEw8U7wBuNmSuyMME4JYxgAZEBfqw/l8bkYLvSANxe1fjWAQKc4GTg7Mg55MK/RN8wLMwKUwXPWbAK/1xg2O8a7u7sREUV/mvnmPIMv9wnnYHs47pA959CGns/9fr8wehHRsQV2Ttn81891GAybsbOzEwcHByWqYR3vhHZ0lMckthRbQls5NEV7UF6TEhHRyfehrXLYFt1MXTjXocga4Pqw8qHBT9M03x4RPxcRfyQi2oj4mbZt/37TNA8j4n+KiO+MiN+PiB9q2/Zpc906fz8i/tOImEfEX27b9sv3PcdUopOx3DgZ0Ud0E4FROmSxe9BkwAIDMxwOYzabdTwmGw/QKs9FcdnouUycw5J2sz6LxSJms1kZZIhDIQwOBpG9BK6zd4nRph3wnBhInM9KKFgII+2aZ8izcrKg2Qq3gQc951mxRXRf8Hl2dlZ2rZ7NZp1VX77GCiQrEybU7u5u7O3tFZBrsInCv7i4iOPj42jbtoTKrJCb5vpVII6NW2G5TlkBuzz038HBQfHirGhpv8Vi0amLQbxz3nLds1LIHqkNyEZendB/eSd4h62yfvB+X/7d53FfG3oAEM/IgMbCeLFxZJ4aIPlcS40NoEyMdxhddKWZBs/BnHfiFUp2+ngmDlwG8WaUPNaZO9TXrIj1lpkKHG/60AAl4vZCGJeT7+vAXESU/jW7AXtngEyuk/sBHY/DDNtmm4Cu47r5fF7GlsNh3hjRY4s2AAxhLymDVwP7vYuZ7amNW66nP7yoxGE4xg0REF//MuRFmJ/LiPiv27b9ctM004j4taZp/veI+MsR8Utt2/7dpmn+VkT8rYj4byLiP4mI7/7g8wMR8VMf/L1TaDi/y8iGIzdwRNwa3AwsJk/btp23JJtS9n2MxLmXE748+K0IuI+RLmVgZRFAhuXSMBtOJvOEM7r3AMtGGGCGwWR1gDeUMjOGtG3bMbDE3rPC5hk1Gj7iJqeEa5wjZIPMIDcFjOIjVIlH4vg3rAoTJYd+MtPjVU9t23aMg4E0Cdbn5+clH4vyDgaDmE6ncXp62ukD9zEKjnLmUAFjjlg95xGGu7y8LEna3qXWYc5aiI++owzZMzVYRWm+TAWykfvFr63IK7wYH3jVhPU9x+wh25Ag9K+T883EZKMdcZs5PT8/7zCKDrVmAGYHwcDIBs2rjXCwMNDo3Qx8CIF7WbodC69CcnKy54QBEM4ldckG2kaa9kEv8b+ZJY7lOU870E5+ZU2+P1uYcNwpGRHRyWWknT0mnOtEeJEFIc6RAVyhL2GxWdFLRAWQRn3MUlFGdh53nhd1JbTutw64vRwVycQEO3Vbj9KH7ne2gHFU4mXIhwY/bdv+YUT84Qf/nzRN81sR8ZmI+AsR8R98cNr/EBH/R1yDn78QET/XXo+W/6tpmoOmaT79wX3Wir0PT0QGpRs1ovtuFCN2T0YQPWDGoIlzPFnsqdjYGmRlVGolQ+d6BRL3ZcdmBh4MEOwA4TG8O4ei8r1cJuenYPRdVwYpNCYTyztBQ6/ylnLEht9gxN6d29/94Zwgr17r9/vFEzFoQhkDHkajUWHEKGPEjTfEBxCBF2TgRr9nmhV2rNfrxePHj+Phw4dFQQEc2esjtyPtF3GzSsz9T/lroBnmp23b2N3dLYmmfAz6DXp9jLHhtvc4xOBswM+rFfRJfr1ABjfopZxnYcORwY8BSL/fL/k3LDawXmCuRnSTcyNur45FzE5xHXXKzCZl87u5KBNOncGEd39mrjqfb51zl0Pczi9xu7hMsCBmK7LjxF+Mf34PGn8pj0GF+9mhLYcGnciboxdZnxsMwN44FBURJT2A//v9fnH40AUALQMZ9NxisSjpF+gu21czYAYytT7nL8/LttBsIOUH9Pg1Qb6/2wn7hSP/sYMfS9M03xkR3xcRvxoRf0SA5t24DotFXAOjd3TZVz841gE/TdP8eET8eETEgwcPIqKL/mtKIIOODEYMXkDLZjI8obJXxXezPUyCXIZMHfrjlxDyO2EdBrhRMMnMVkjZYwFIgIhtUCNuqEk8KcpiVuXy8rLsocS5ptyZSN5OILMQKFaez2Tm+XnyoIz8zh0mpD0sjLUNve+Tw0EIgMYAxOX1WKEs5BW17fXO1yQgP378OLa3twsj1Ov1Sm5OZl/cPvSXFRmUdqakYeuurq6KEXQiO0rDCsliZZ6VFkASgEXfbuSjE+uw/LZ2gwyDIL/WwV6/AYidAnvRdlYcmrV+M5vkMGt2Hh3WAIxg2J2/Z/2K45aBktlWftva2irLrCmDWSaH+T1fYXXNuFjX24HLdYH1AJBwrUEl+oLfPUftwBiU2RFH33lzWcoKyLP+pt6U07rC7Yvu8PWUg7aib9CnLJVv27a8D8x6j2eS/4NeMnDz82uREfe5+4+28bUeN9bFWb/ZpnqMc19sgXX9i8gLg5+maSYR8fMR8V+1bXucgEjbNM1zJRm0bfszEfEzERHf8R3f0dLIpto+eG5ERAe9p3J1lLzjvh6seB2+nqX0THpTjn52foaf7U6EYXEoJC85dCgM2hQAlL1CT8Y8+JgsGNSIm7h+HpwRN4prZ2enxFZZ8UXZuaf6qJMbZCDgyZFZn4ibpDUnxHmbdurT7/fj4cOHt2hme0kZoFE26k5uFflPrg8eIRuZ2fhQltlsFm3bxsOHD0tCNKwh293zTOpC3WvUb22jNDNf7DWFB21FRF8hNjQZvDNevfLP7N9GPlqxDptOp62XgjvsxVzxKy7sVRsMWTIAMjvkJFc/g/NtrGuMkoEBDE12SDzHbbQNtqxjPf5Zbr1arYrDgd7LTq6dS4MGgxCX38bRBjiDQq5jQQTPAQABXBx+ymDM+tPMkUNH6J68dN/OosNNmWGLiA7rfnV1sz+PdXzWh4Sp6LvhcFjaPOvsq6ur8vJlmHWPFbcrbYjj7H2PkEwI5JWN7N6f7bUBr9sXfcWzbcdfVF5IEzZNsx3XwOd/bNv2f/7g8DeaD8JZTdN8OiLe++D41yLi23X5Zz84dqeAHHO+ghuLTjc7wLWefBHd2KyVBuEWOtoMCNf5uVmxGFiofSIiCvDhOvZeye92wdtwDNbAhv0fKF+v1+uAp4ibwUK7nJ+fx/HxcQe5Q8ObyaIclAGGIgNOgyzagvLao7RSQTjXBh+FjVGgXgcHB52NwlAY7gfHppkQfHfsmkQ/rmfp7PHxcSyXy3j8+HFE3HiXKEEYoYuLi3j77bfLb+PxOCKivMdrHdBzm0d0X2FBWWlTlJjr0zRN2TDSwNn9YSNDn+Z8DMYUr0XJxnQjH53YgcpsD4Y1b5dh4GPdZXBiQ4leMQDJrLE95gwwzDagbznG7uoYc8YjAjsbUX/tD3UjL4nQNi9UZh771TSMY+sbdOG60B260MARfWlw4UUiETfhFrM9fDcgNIChfi4HZbPe9oozM1JucwPRiJutMugH50bhDLsMhLdyOHs0GhWHlnu5vSK6782KiM4iGTOF9DPPQ8dkBiuDVqda4PThhFJ3+iD3p393aBMn7mXIi6z2aiLiv4+I32rb9r/TT/9rRPwXEfF3P/j7v+j4X2ua5p/EdaLzUfsM+T4scWdCunFpDLMIGTVjXDOy5DoSysjpWK1WZcm58y381yEw7sXfCuLUKQAAIABJREFUbFyYzIALFABKgIHu5efOHWFSP336tLBR4/G4JEdG3Bj6DHCM6hG8N28cyQC2B5YnO/Xzd09y6pWp9gxYc0gstycgjzqhGEyr+zfuRXmYPIASzsmJfySZA2S8PB0PFaV5eXkZT548iUePHhVlxEth3W8GIzzbrBDnOqTgcEL2qHu9mzwx51LQlutCI/xuJWRFtgE/r05sBAyAMjjgXIOdfB/E7EN2wAhrefmywwRe2m59yvJ6ADdMAUn+3D8bODMk+QPzjE5YLpcxm83i+Pj4VlKrdWfeuI85hHNm54v2sH43KKgZZ8I7zkWhfnZkKFetLDzX55gF4n4GPi4fest6D33iXK/z8/Py3kA7fgbCgEc2+uWVO9g1cggdZvOYMqPjvqQc6MnRaBSTyaTo0Bz+tx4i/Gad5ET4zJrbYXT5DPbzuH1ReRHm59+PiB+JiN9omub//uDYfxvXoOcLTdP8WET8QUT80Ae//WJcL3P/nbhe6v6j9z0AYGJK2N6KDUEO/WBQMh1rg5RjnBE3O4xyXyYPAy4vN86TxZ3CADQwgfUhWa1pmk6iX8TN0tDFYhFHR0fFS6L8JycnMRwO48GDB2X/FtqEwUH4zOAKA8/ywl6v17mesuOlecKheOw1eUCakrbXZiRvMEKfeMmqmRGDKt+bsri+nJfbnzKSHMiu4AC9g4ODmEwmRekYjFlJjsfjslT0wYMHpbx7e3txcnJS6OWsPLJcXl7GyclJ2Z6eFXj0jxUr488sASwY47CWvJy9U7OHEdF5tcpGPnpBF9mg5U9EdECsQZANUWa2uX8+hn4CXFtfMXa4J2MdQ3t6etoJyXCN5zFSA152IFw+GOjDw8OOY8XvOJ+E7XB2zFJz3xzCsu60fncupbfc4JnMb+pvAEX9s5GO6DJcGdi4bOjgzGrk51Fv2gGdZabbeiUDBua4t+xYLBa3QmCDwaBsP4BQBmyF74kQqiLvcTqdlnKenp4W2+TxZNYRh9U5VA5Fejx5XBmMZobsZciLrPb6PyNiHfz6DyvntxHxXz7vczAoKPvsGdkQf/CcjkF2SCl7C1zHhDDQwaja+GcmiOflicd3FA+sD2wP4QcYA8AWk5xzzs7OyvJrT6C2bWM2m5UlpChO4sOs4PKANnuFIaVeBpOUnbwnAxt7TrQdXqYVoT0A2toD2l6fgaf3KyEkiBdHXcwu5XCAvVraEYXu11LQDrA+0MZmceh/7wvEHkS7u7vlXuRHOWxIfxrMOdzWtm1Mp9POSwrdLm5fA1OvFvO+LJ4LZp5oEzsHNZC1kY9OUNhm6TznMmvH8YjbO4Rb7xk48N3gg7HoMmQnh3EdEeW1N9aFZjI45jLm8YlQT8p0cXG9mzy7pNtJNftBWMRGGD1AzmPE7V3UXX/nJFkfUF+zUXaeON960gwsesjME/fMjIy/28CbVc/nUGfnOLqM1ltcZxaJssHGHBwcFF0Je7xarQoTlMFs1gm5X/NqNV6fAxAys55ZHOrsZzjU6LazY+wx5T6sMaMfRl5rF/Dq6jojHQCxWq2Kga8pAiaTO+Lq6upWvNOTnw5hAPb7/c7GetzTgCp7CFm4jj19PMmcq8MzmJRXV1dxfHwc8/m8TBorhjzwl8tlSdhFQbBXB2wSZW+apiTpohi3trbKe8LsoRKyos7Zs7K3CiBl11Dalk2xUPYYXuLK9j6ZvFdXV51l7LBiJIaTQGhlYdYIyp8yU26ze7TdcDgsipm6AIa4ls27AJlefk6yHyA2GzJAkCc3DBT7W4zH406OVvYcbcz4i8dF/RnbBuMRUcKebXsTI3f4cyOvTpzkvA7seO5kBZ+Vva/L4yM7cxHdfDiPMwOxzEZklpd72yjmulBXjDx/nzx5EsfHx50d6A1azLSb5YJRWK2uk6OJBJgRNjChnjnc4rq7vJ4zWa87XOZzLNglC9fgsNmYo1NcLoNXHFmvaqPstIttWXYiqcPx8XFcXV3Fw4cPSx87p9CALxMEmVWy8Exvo4L+ytsrZFvFc3I0ISI6bejnGhD591rZPoy81uAnouv9ENuMiFtJcRjviJt9Vmx42KTJHWwQ5YFO4i7/R3TDWxn81DoEzxzKlYmcJwHKwi/tpDxWbJl94R7sQWNwxOs/MNIAuPF4HKenp6XuXGMAZK/UE43J4rajfNSVJdqwIFmReLKa3qcubPCF4mBlCBuf0Z5ud9qnba93sT48PIzT09MS90ZhsF397u5uJ3Hc3jGJ3n5FAOW/uLgojN329nYcHByUduBN1IxD73eRPUnGz3K5jN3d3Y5C8FjKXhCKIieq53yGDLQ9BzhvI69O6BMnkHrs1/43+KGf+d1MTA0cR3QT7Z03yPjLAN6vb8n5h2YT/clMko0shq5pmnj//ffj6OioA7YMIsyutO3NW86d+BvR3R6DZ9qBQo+wksgMkAGGdRJtYafIoNL9YqfQfZZzadDNETfgKDOu/ksSM46Jga71Jd+pL696wilEAJNeqWrwHRHFFlq/OAfT4yi3B+egCw8ODko9HN51aN2RAT/Tz+B/P5/zPT7yWH8ReSPAj1ds4cnX4uYRtyleD2onxtprYSLQSePxOCaTSZmEzlDnGRG3V3i5c73U0IOXgQ4F2TRNHB8fV1cO1ZRERPd9OBE3nh1GcrFYxNOnT2MwGJSt8v3XYTTqvVgsivICBJjlcdjQoMCCAnVbGkAZwGVxXBmvzzk6TAK3P8/k/V8wdn4OIICwI0v5p9NpeXcP7cv5lH13d7coXY4tFosYDocd4AKVjOLDCx0MBh2wltkzs1eZQcwsEgrYnjUrb2wsUYjOXUOc27SRVyPonRz6ygY2h8KQzPrw3cd9nYGKHTs7HXxgVT2+yAeszTW+Z0OWwQj3J8eH+9uJy8wtrxLKTk1mT5lj5MsRDnM5uD95L7St9TVMcf49O5m5rmZqs8NydXXV2eUYltvJ5/SdWXDu43KYFXHZzMrB8nOe9chqtSoh9ogbp4k6A4C4rx1Rt78/1l9te516sbW1Ffv7+529hAws0bu18ZjHVG282Yk0GHsZ8kaAn9pE87JGGwQaPtNvSEa9DKRMiyLZ60E8IPhL+RiQETfJh875Wa1WJTSVEwwZ3L6f6UTqbg/m4uKikxhOgjTPnM/nsVwuYzKZFE+FexhUsDKM363M3I4ehFYW9pwYrGbnGPBMbNoQ5ocynZ+fl5wmKwKvlgKwUTd+swLJkxXGhefNZrMYj8exv79fPFafAxgkcZHrTk9PYzKZxGQyKXXe2dkpq/EI37E/E2MMJb611d1s04rF7QtA8rjyUny3Y8QNM+rk+FrfbOTVyl0MTwZANX0XcZvZua8fzTo4lMpzWHiRwzQOOed5b+PtujjXhvMWi0W8//77xSjXgE/btiVh1kbXjCnObmZzADeAC1Y6cT3zyYs+zG5Y37pOWcy85eOZiaiFogzmOD+H8nk+YMgOEeXKjjagEABEn/o8M3hbW1vx6NGjDvtjPZH7u9YW+VjbtmXxDY6emWiXPdfBx9Hnfo51V7a7L0tea/DjituQ8oH5AI2jPNyBTrCjwb3/A/ewEc5hmxoC9nHKVzNaHGOFF4zPfD6PJ0+exHK5LPHYvMqCweJJD63pCbVcLmM0GnXeDeVVPaPRqEwOsyhWVkyK+Xwe+/v7HSaBMqCArOxqCN57BHGd/+dcM265D6wks4dCeMteq/Oncpm8lJw2cwL16elp7O7uxv7+fnnpH/3BuPN7cLiG93/hhXkzSZ4RcbNnD2OR8UdZ+d+UvpURCsIMXwbotBs751qp14DrRl6N2Hmphamy/vA5mVHOYxtxn9acNY8FnB88desbgwGX3/+7TBkE8Z2VXehRh7E853mdjMGIQ0veBZ16OI+Ga2ezWaxWqzg4OKjuJoysq2dtCwK3X/4/Ax+caN8zAyDqDOghzM8qLKICLgMhMb6Tp0j7eL7zvi50vNtsPp/Hu+++G4PBIPb398uYHAwGxUmmjG4vgxU735arq+s81bfeequjn8yscZ7HZQZZ2SZl0GgbVZsDH0Zea/ATsX7vC3u7Tjg1+Ii4/XZ1jIhjtavVqoRXMluUlQnPyOyM7++Bsr29HWdnZyXZeHt7O2azWXzzm9/sIGXCQxlcIbAPngx+UenBwUHs7u7GfD7vbKlPro9XM1FuJ12bjTK4qrFAeSDaq2zbmwRbh224hkFOeMleHiFBG3eDoH6/H8fHx/H06dNbStNlcb9zvTcCy4aBOpPUPJ1OC7Dxh3wgjAd5VQZ3ADnK5P7NmxDaYwSomh2yN0w83eFHWDEAGfkOGaR7zG6Yn1cvZnWsyzLLYyNU83h9Lsf8jMwO5P8ZE1nXeR5loJWNXa3MrhegBmYe/WHDxqtsEOYoOsHOJ3X13HYohGOwxQcHB51r0DeUxW8gz8xa7VMztNZ1bgs+blPOo3x+rQd6Cf0LaCFHFF2S/1q/E2rDiQLQWE9cXV3ndX7961+P8XhcWG70FeWEIbPttN6uAR+cNlbB0o+0v9vCfejx6zHt8Z+/0x8vS4e9EeCn5sXY4DpZjA7LSxztWbVt26EdOe6VBDzDz+FeHiw+L+KGhTLjwLO2trbi+Pg43n///ZJk6FUVDCRTp54ksAcsefaEhj0gES1PTiYMuT3khZAcPZ1Oizcxn89jb2+v83wYi3X5VO6TPEnoDwywmTjCdl5+6T6m7sjW1lZZIp6X8rsvaBuOZVqeMpgKJxQ2m81if3+/vNcLYNa2bckzIMRkQIUyYix64vKsvLzdRsFbOeREZoPj3Eb0ZR7r9kZrhmwjH71gPLLBtQHNoCjixvC437ID5mOZjcgAiPsz17Ihyka+ZqRqZUAfYDT9VvZaGMh7lnGOHTqO8SwzOYShYRWsi2BTKbsdURtgdKjDYNlAc4/c9m4bz0Pq6IUI6IyI612yDw8Pi/NiZ5A6w4xZ33hTQLMfhPt5hY8X86CfyDlyGWezWbz33nvxmc98phz3qlVHLHLfu74ZoERc2zi2/bBzVmvb2n2z3aiBIOu2lyGvPfhBjKI9MG1A1yUVOgTFQABZe3IyYTyJanRmxHrPDK9+a+t6Z2dWcPV6vZKIzIA2EMCQopgAHIRcSOxj3x/u4UkOgwGDQ10MqDCqxGhHo1EBTLQxS+09yPKkdQZ/9hgR99c6JY0yWrfcEWVAbg8g1TkNEVEAAPf3UljuaYDkZzhpkD6BXfq2b/u2ePDgQWfZLXF2e2r0uZeTe0yxDQET2NsI0FY5POY9iwAzNgwAH8aXPWH3CdeY2t/IqxM7XrXQCr851I7keXUXQ5Gdj6yfIm4cBrPbLmcNJOf5asATcRMSZyxatznMhS5Ev3EOc7+2JYBDhjgQDuc7TMbO95eX1+9DxNE4Pz8v93ZbOW8u19VtVnPmfB661XvZMBedQI5zQ/35nstBOMxRDTtyft5isYjJZBL7+/ulfxlbhOLd3oeHh7G3txf7+/sd3YLuB6CtqytlyOOSFAKiES5rdtgyeKk5zR7PHn95gdOLyGsNfuwN0DCeoKbKyN/xRnA0HA2GMQT4MFg9cD14jMxzZ2BcrCy8YgM606uE3n///VsrF2yo/RwmM5sIOh/IeUrb29sl8fbi4qJQmvbGrGBYLgpg3N3d7SyHb5qbHaAZyLlstGMemLQ1/ZTbLntg/s2Ut+sHoIE+x9uxV8nqKvqLVWKUDeUMELDniGLxc6n32dlZfO1rX4urq6t4/PhxUUiwb4SmsndqBYtCiYhbq7NcBr90lLbFI6NfsmdMuCviGiSyHQBtl4ES5dnIqxN7qzammQXCgDvkHBG3zlv3jJpB8nMAx3a4fL11pu+Zn2Mw4ns7WRp9w/UwFeidbMjNlsLIREQnNM69WFTg50bc7EoPuIqI4qQAgmptmlnou9gfS82Jy6xPzsmhfrwaxwsX6HcADe3IPf1MAAX6inMnk0lxatEthMG4z8XFRTx58iTG43FZLs/Oz9YTGRz7ewbBXHN+ft5h6a1/ao5ybXxRz5oTTT97Gf2LyGsNfiK6iXU174fvETeTyUvjs8FjYDLwuM4fAxuXAcORAQzHea73+Gnb6xj706dPy31NMXJ/FAaJaxxj5ZGf603LYIZ4SzzsAoMTj8ehMYwgAM171nhX6gxsakrWK0kywndfrRvQJNx5byaDMK9YGY/HpZ1yiMCMCyyZl9vniWVgCyDEyzLrdnl5Gd/61rfi6uoqHj161AGjgEDXv+ZJAkxz26GcKP9isehsMYAX5mt4HjlKVkp5TLm9bdzWGdGNfDTiEI0Bg1kgFgl4fkasZx8Qn+f+9f/WW57T+b61eZ7vw8f7FpnBYa7y8YpMzvOY5h6Me/IvmTf+LYd+YU94Nvt79Xq9ogNyNCCz1MwLt73ndf7QTtl4e25SNu5JGVerVadczGP6EX3Ix8CQkJ/DhO4TXkD74MGDAmqwecPhsAM8l8tlPH36ND71qU91GDXnzkZ0mZqa/s7HsauZhc7jLbcf/WDGCXF/RdykBrwMeSPAT24sKw4DECNvNxjUaL/f7ywtN8LnWYg7jvs6fIZY+XglFPdfLpdxeHhYjLOfayFW63AKg9VACoXAxMJ4klfEvZ0w6ARkAxu+E1bjPMeuvULL7eE2MQvDRM9AIN/DYRzKsru7G1tbWyVvwAwNS85Z2QETxL0MxLgvXp9j7VZsKJIMMvJ4IImzbdt49OhRjMfjTp84UdX3RmreDq9rwUO0d26mykrGwCznOzH+ciJ+9l5rCmYjH61kJtn9ynG2SnCIIxsM/2+nEFnnKGbdmO9p8XWesxn4GDgw/+2stG1bXivjucdf9GRmeSKio8MjuosAnOdDGTwHHHrjtQueDxmIun04j4UMbo/cJzVgZN3jcDjthj5rmqY4nt51nzbnlRHUxatI7Zihy6z32C+JPcrQQw6BXVxclBfMHhwcdEKuDr3m8eGxQXkt2DsWi+RrDA5rgCfr3hozZCb9ReW1Bz9ZTI1m1FxD7lnycuR1HWyxscdI5gRaI1Ouuby83mmTyUjnOc/Iq7FgerIn4mXcNqIYTb+FmQmc9z1COdn7YyM+ABxsEECCZfgISiF7BrmcVjAGAZSH42bChsNhuRceBPdg9UNEFIUW0Q3HZZrWzA4e4u7ubgFWBr4AG+pKGzOOAEmXl5fx9OnT6Pf7MZlMirIlmdwemYGHxxaeLM8F3Jrud9n57vIa+HnM0l60n8vicbyRVycZ6GSAAjNgsFoLc+XjjBmDbo7zXP5ad2VmsKYD19XBz+V69EleOXp6elo22syhNsCeHaC8+abnA6Hms7OzktYQ0c3V6/V6xYFEzERwD+Zf7pOI6OimvJ3Ife3E7+STWrd5CTtlJ73C18OU2SGknSJuXurqtrQdiOjmC+3u7nbYJWwT+uPk5KS8+d02wwxyrntmXXKbZIYbPeTrXD8D6Nye+bnZxr6ovPbgJ3u+HMsT/S5EbmR/XzKXJTNDCIrEQIDBY++bN4lnkEA5MMhereMcEq+C4FoYEX7Duzo5OSmAAYM9HA5jMpmUN5JTNhICmVieOIRimDQ5rGM6lXq7XbMydj9YIXjjshx+pKzb29ulfrz3BiGh3J6WJxjhOyhztzlgigRxVuLlsZTrQf+wsdfDhw8j4joPgRUWmUHy5MVo0Idulzxe2bOJ8WYjmillyp0Vk+eM67ORVysGOgYvETcOR8Ttzd9q1yL2vDOg4l55Duby8L+Prys798Lps/PpPELmR06s5h44VxE3LzLmPhhLnECOmWGF0TVrCxPCggiWgdeMq/W/5yTtVQv55LbPbRhxM0cJMbm8mSEzo+92B7jQdl74wIap7DDvPsUxdHucnp7GarWKvb29AsAiorwbkzD7bDbrhJJymf2pARGPt5pT64Rn23Ha3DrT98upJ/zGNS9DXnvwU/OA+Ftjd9x42eB6wGRaLif8IUyabKBsgJg0fqcTOyvzLAYF17MpoXf5dJkw3AxS2AkGZ9M0sb+/3/EWCKEQ8+X9Lm3bxu7ubjG8xN/xOlEcPNteESuOmISZIs5eQk0J0660sTcq4zdACLucAh7n83l5Pqs28CpdBp5Ju2Wl50nZ7/eLx3N0dFQmk8vvsdc0TcmZIhn65OQktra24sGDB0WBQ1XbuDEeUOwovzyxTaufn58XJeVxzD0dV0cMamys1vXLRl6d1MAPH7Y/qIWkaiCF8ZsdvxrIZYxglGpg2GPE3/Mz7XD6vt6ZPSI64XeDGee5MJYJwSMXFxfFaHv/LLMwsES+N86cHTbPOeuyGgj0PMtAMjPdbnczV4Au8iftjOJwUg/3tfWSQYCBEnWEsfEY8OaGbnds0NHRUUyn0+Jk0hfo3Pl8HsPhsDhbzuHKINz2cx0YxMl0yG9dGNfhTOq17v6IwfSLymsNfmpGyN5RBj+5AyJu5+6s68B8n3wsn5eNF5ObQUNIw4OG3yaTSVmdw0BhkvLag4uLizg+Pi4Uqr0FFAf3cBjHrBAAguvG43FRWEwO2gy2ySvmiN/i5bHaCkra/YCXwsTLCpryw3jYS+CYASJ7EPHWerxBKHVi5bQHioAlrr6vwSWKjTbgZae0YTY4ZoxQchFRvCsAEOAnbzLmtsljMY8rngXDx7vHHE6ojXWPWXu7tTwFg6KNfPSSgUP+S8jUOYE2vp5PEd1wUB6nWV8iGSTbAfQYymEY14Fn8x2DbP2EU4IO4hrPGTuA6ExyVZjf2Uk1m2r9xPj2PjVeOGGnZF3b5PmOPnaZzdK5/dB3bv+tra3yGqGjo6PiABtM5L5y6A/WJyI6e9WhA/1MHCVY8F6vV1IsaDcWRTx48KC02cXFRQkP4lw7V9UAN4+z2l+PpYi4E/x4vDkvNQPh2jPcji9DXmvwE3EbZORGqVH9ETfeMeJOyABm3WTwc/Lv+Ry/CdyAJNOF4/E4+v1+2SGYyWTA4U0QMwInD4eVT7PZrENrMuGZvHhmsB2UHU+Ccuzt7RWvgAnIrp1MFCs15zwBPDxxcrtZofHdihBlCmA7OjoqCtHhK+9xZOaJv0x4vEEmMErYicK0P4nRKHODoAx46C8UEbQxr7/o9W4SzFFozrO6a1y5TZzT4zHAM6wE8xygD3PI8mUpjY08v9R0TNM0BeCu82izccZI4LDQ/wY4Zmazoa6NAQMyg45cVht/ADpAgz18zFJERGG3McKMRztuzNfsTNmBOjs766z6MiDJbYfOQBd6f50caqr9Xcc+ZFCWf0NXsFoXJ85Otw0/Ooh3KrLhLI4ZOgt2kDZF96DvANAsmOGdZ17pe3h4GA8fPuw4orSRd7anbNjPbMMYB26HHCo0QPG1GWTne1jP5rbN7fwydNkbAX7uGpAR3U2SmMRmITJYWfecmtjQ+NxMjTpsxkSnk5h8LHMkRssgYKXH+++/X5TIarUqaN7KBAXE3gwO5fAsU78RUVikwWAQ0+m0tA+/RdzshYOysgfl9nE7o1ysiDgvt1nuI//u/uG9QIeHh51EPnKU5vP5LeXpxGhYNytTlIdfNOqkacDheDwumz7msnuSozgAGLPZLCaTSdlPhMREv74kKwDXu6ZIc5I7k55nM7byahff20ZhA3w+PqmNc8Yrm/LV5kPWXfS1GWXnZiDZINWenw1PjRGK6K5U4zssLOUmJ88G2aEVOzgcPzk5KcAgOyY4fujMiO4mq6PRqOgGwl2us422WY7MOmfD7v5yXklt7vp4Di82zc1Ks+l02lnq7xyXfv969TEb18KaoTOs1yKi6GU7YDzz9PQ0tre3Yzwex+7ubmH8qR/Pfvz4cWGFfD3vGqPcGVDn8ZD1Yx53lgwYGQd2Qq2/8zi0LqPeL0Nee/BTG5y5kWveL+dF3H6ZowfzOmBU84Dy+fz10mQb1ogbQ7m/v9/ZhAtEv729HUdHR/H06dPCwqAwMMooFMpDbtHJyUmnLVAgNaDYtm0cHx93XoLqfTlYxUTZYImcs2La2xM+K1hL/u7J73LjgRwfHxdl4PBer3e9nB9vk351+QaDQVGKpr7JJQLgTCaTGI/HMZvNCviD9RkOh/Ho0aPyvhrqZqWZExhJ8pxOpx0GrFZ/7lP7vQZ0/FJWKzwnvGZ2Ko9ts0QbefViNtRhWrZ2YAzWQAqCcfAiiKurqxK6NsuH1Iz1Ol23jhHik1d4UWbCJtZRzAEDJFgczocl8vyyw4BR5H7oNt5hhwOE4WaLDLOuTXOTP+cQC/mNXO92o5w5pcLtUbNJBlzWvwC4w8PDok8MBGkjzsd+AHJWq1Vh77EHg8Gg6HBATMTNi1NHo1FMp9POjtrkj85ms/LqIgMg9B8rv7w3Ua3O+VhmCnOoNTtnZpboL0Cw28jt6Ge/DHntwU9utCz+zdQtA72GFNcNYKQGfPJ19s68RDDvptm2bVk95SWQLI3+xje+EbPZrLNsH6l5hCxPn8/nhS6mDHgIDHpi3ZSbAc6EItfAeSowTkw2U+wGkVbk69ortye/ecL4GvJ8zPjQBt6/g/qYBve+OZkeBkBeXV3F0dFRnJyclB1OR6NRMSrU//LysiREs3mY8wAibiYv15HgzhuavdFmztXJ7VBrG1PaBj60mQ2hlS7f3fbrgOlGXo2sYxZ4EaQNcHaumGf5beXoE49L9302RhzLczGfmw235zzzyM6ZgQ+6ykadecJCBXQdRpX7AYysu5umKXrKK1DZjTi3K68I8jzJ+X4Ixj/iZi7bIbHhtQOxrn/NtPF87sFqvtPT0w4IBNQYYNC2vge/OxEZ/U40gTYHVE4mkxJS9etvyO/h9RboPEAhYDEibjExrm9tTN8Hom2HndtlO4LD67SVDLI+MczPXQjTzE6OAd8V27UBzffKf2vlsVIAcHA+xhqak9VBeCcg/sViEd/4xjc6E4COZ9UQLJBj36zkOj4+7mTI+yWbhK4w/DaolJXNHgEPgLazs7OYzWbx8OHD4lleXFxUd9b0gM/9k9k206iYV/+KAAAgAElEQVQ1L5XnGAQCwAA6Bhy0FaCT+3t1moGR2xNvG1Zpf38/9vb2ygQlsbvX68X+/n4nvGhPxTQ2+VOEFV1mt43Hn8dhDaQ4dGtlzP/O/8hGzdf4txpDupGPTtaBXb+6JjsB2bniL2PdjlVEPe8xz/27vHdf5zKY8eWYDTArMWtMBs9BP/JCU9/DZfU8wDBSZwAQes373RAON1BwTgzvS/QSe+smz70MFG1Xcvu6z2qAM98XQEFI3XV2ef1cnFR0YcQNkFqtVjGfz+P09LRsaYLTit7u96/3I7Mdubq6zuUkxwdnLy94iYiO05gBem4PH8u6Mus/jhn48sxMWhjo3Gebn1dee/CD5MatsQkZ2TtEYOOR75eVgH/PnlBWEE50xgh6FZeXkfP96OgovvnNb3ZeruqBj1Fl0Nt72t7eLu+u8oB2fRlMplM9aHkeKB8lRh1ms1mcnZ0VWnU+n8d0Ou3sAI1xzR5SzTuysrGxdr/4FRQkLa9jM2gLX8eeP97JFCUCHWyK1+Po6Ogozs7O4uDgIPb29gqAxYBMJpPSjzXFRXvSL+yuurOzUxR+nvhuq9qYpB2yYuF7Dehxjzw/PKbX9dFGPnqh3RlT2Yj6HIsdO0JDBuFclxnA2rNrz8i6jjJ6vDDWvCjBDA56EGPKMZwpxnJmYnmu2VKeZfBF3h735nySfTmGU8Q8Bzx5U0EzZbQh7WZH08yDjXJtbmUA6+NNc7MTP6/fQD94mw3PY9rL6Q7WuTC/gNDlchm7u7udla4XFxeF4aYNuBchQud24hQCgGqg0HXKrH/NjtZ+z04f+tiRimyva+38ovLGgJ+Ien6Ej9lA+Xw35rPcvzaQa8aLicLANLvA5CfmygR48uRJPH369NaW5f7LvQAxAKDpdBr9fj8ODw/LgGFywHBQLrcJ5TH75MFLOM4eBzksDu1QNueXZIRfYzBqFKeT6sgxYuJaKeZQJuU2iwQo9KZp1AeanOeYEeQ+9AWrUj71qU+VUKDbcTqdlkRCU7OUGeAL3dzr9QoAct/WJAN3j9kaA8R4gdWqXedxvQE8H7/Q/36NzV3Ahz5n3MIEGBhklqfWz9aVtbl61/xlvqKTcJrQawAGAJH1CO8bzKCobdtiaNFzAAE7ghHRmbcYfXSamS0cTK4BFHEvXrrpNqsxNnf1hc/NzI7bmTK4fWH0I6Kju8nrMyC0bcg2zHmAjB/ag5DiwcFBqTuMvlcTA2J5FyR6kjI6ZcK25C4belfb1RgbJ3ZH3OwMzaIRO8b3OQcfVl578GPaN+KGdaBjGCBe1uzz86Bd52Vn1Lru9+wN+bgHLrkf9iYODw9LzNX3zwaUcnivDO4DWwTzQd4KRp/zvFLAq9G8EeNisSj75VBH9sohPwilZECR2YUMTp6lbW3sAXkPHz6M9957ryTeOZmT66iTAZk9UShx2sMskJWoV68g3Ofdd9+N/f398t4bl5tXZDiviL4C/DqGj5JfN7bW/W/qF8NnA0Yf4BXn/snn8v8m7PXxCH0LiI7o7jZcmzPoOMYghivrsMyS1IxUHhP22j0mslNZS0IFzEfcbHkBEw0gn81mnQUDPMMMc7/f72zmyd9cButwM560CccdDmf5Nhv4ES7KjJafm7fvQMyo1di6bNyz80z9WYlFP1KWvb29ci+vBvZmjoBHxgIOT247lszv7+93VraiJ3xt27Yxm83KjtEOGaJv7Cznj+vM/9SVts06yP2YGRwWpQDUzArVANSLymsPfiJux6DpKCamGYs8uCO64QPnm9wFePJviA2IdwylQ0HmTdMUr6ZpmrKaC6bCuUA1Ko+yUqfJZBK9Xi+Ojo5ia2srDg4OitIx8OBaQIDp5IjovJWc5dl4AU3TFK804jo+ba/TdaO+Htz0w10eaG4vK4nt7e149OhRnJycRMTNknH2sPAKBOhZsz140xHdDdUYE6ax7Tkyjly3999/Py4uLuLbvu3bCiNFOXd3dzvAivFIuUgiB/wYyNTaJLePFa/bmPtRZ373OLSneZeXv5FXJ56fXqWT+6LGuuC4eF+wdZ511ns1o2WpOYLZwct60iCHccp4ZP6zOSuG1+w49crvZ0LHtG176/U+1MWvxGFeOV+FBHA7HRFRdGe/3+/k/jDfbfAN9NymNR3tNnb7cS+uJ7xECIo2dphnuVwWXY4TCisDkHN7GPDZPgJ+Z7NZREQnqZ5nGii1bVt2eUaPwSaj97LNrDlsNTtK2yI5RaMGIHm5r3fp91yp2ZUPK689+KkZU/6/vLzevdOJpTZyTCh7+yyJrIEf37vWqdyTv14+yeRAYYHeWcoO/ctgAoRwfn4u4TuQL7sz88Zcb/rlWDcTzwnXGHomFO/MwviTNMh3DGrEzXt7vCoh4naugUEE7ZYnTFbuXAPomk6n8dZbbxWv0WEeg16UGqsd7N0Q2kMxuv/yazXoDwMl9/NqtYrDw8OIuFEiXDcej8tLUqkvCoXVYtl78hizB1nzvP0sAzrAIGCLMjvPgfJnUOzxtZFXL7A+7qeI2+H2nOvilY75Go8VxCyLzzfA8D0sdjIBM5TNOxBzDnoG/WN2m3LZwcOwO5wecZMY3evd7FJMGziEjePCvDIzS/0cJkYnoPvoB57pNqmxGPz/LH9pc4fkARLoWDZAdCoC/wNmKCNtwTggSZ6wI+3Dcw0U+v1+eVXIgwcPOnqbd4Sx6APgNZlMboUdGY8Z5GWHynq+BtB9zGREDTTxHkeHAg2kXpYOe+3BT8TNZF7njTifwsbNywZzEt060PMsnnGtgz2xT09Piwczm83KfjyeUAAgQAXeFPf3BJ5Op7G/v1/QPLQpzBJhNtqCSeT33fBM2okNsdgQML/h3cwO5al5q+s8y7sGd2aIKO/Ozk5MJpN49OhRRERhqKgvx5wHQJ2pK//jOXgfHzOD7seIG5YJZTCZTMoGdN/61rfi7OyslIv+YuNEjzPK7DYwG1kDJOvGGPcwsCfskJMVUbIYIrd/zcPayKsT+mA4HHY2NYzoJtW6z5mjDjPV5hhiJtP6w+d6vGepjUefe3l52VndxTzxajXvm+W8FZ4/Go2i1+uVlZ2UB13G/APcG/jZQWHjQusBgA6AAiY76zN0nfWHmeFaGxuUZr22jlFl3nvuAtZwfElhgCl23wNCvHAG1gh7xv0NIhkH9MvZ2VkcHR0V3UU7Yg95zmKxKK/ncfirFinJdc/1rul9634D1Nq5OO/ehmSdPnsReSPAT0SXAaIhQe68tdyD34POjZbzbdYZ6Ij1VJ4nmgcASNUD8OTk5Fb4BUBCoi6ekXN2er3rTdDG43F5gSlInrbwSgAYCFOFvGEYr4odkWGbUDZXV1fFK2JiWKE47m+pof+7xOjde0gwEfBw2JTM+1Y474nxYO+y1+t1Ei6pA0nneFc514Gy5xBczYu6urqKx48fF+XcNNc08ng8Ls+298Z9nBhfC33lNrJk9scAaDAY3HqXkdmyzLRtgM/HI01zza4SfvGYXWdAMG5+9UNEPceCayK6juK68/KYWDdvmWvoVF7D4NAQc4itM6gTBqzX6xWnymEwG1eHjAF9PG+5XJb3JFK/0WhUHLejo6N48OBBTCaT8h5Ays197UjyXMpkB8UrTnM7Z93gNs7zylEB+rppbjYSxCnzdQY5ETcsrx025r0BBLqF7VPyStWIayB6eHhYcouwCePxuLPamH3KcsqA9co6HVIbw1nc/neNQcYb7ywD5D2L/nweeWPAT8RtBOl9MuhsswS+zucweGrnrBvouQw2oi6X83yOj487hjs/MyJK/NoGbDgclnCec338vOzRmCp3qIvn591QGYjeG8JsFjQxuxZbgbkeZjJqqNxeaPY+M7vEb963w6EkT0R7OwDgk5OTaJqmlBdvcmtrq+Tp5JizwYlXzhlEIOfn5/Huu+/Go0ePShiybdsSSswbvuUxk71Kj7Ha7x63ptNpO+/thFJ1roPP9bi5D6Ru5OXLXasZDSIwZDCRrNbJ7HfNMfNKIIPtbKzXiX/PDDlhe7PqEdf6i6XWHAe8RdwkTDu0hE7wztZssOcQEEAIHZXbKiKK0cZhglVCf1q3OjTk+pqJzqx0DWBm5scsnhPOc5gGFjvnGqGPDdpgsvv9fgl3sfO97+k+sTNkVpyVYFdX1y845VhmItnVnr2BPJ7y+Fin863brdcYC/exOJx/dXVVcp2cp/oy5Y0BP9koOCkOo8gGUo4DZ0AT0X3Jqc9zh6xTFJ4wmQ1y/Pr4+LgoDCaImRomJHW5vPz/2HuXGEm3Lb9rffmOR0Y+qurce3Rvt22pERMkJIRggIRa9AiDaAbGQmJgm5Z6YixPELYZMzAj0yOkli1kEFIbPLGFEBNQz6AHxoxAlq4sUD9u171VlZnxyMjMyoyPQdZvx+9bub/IOqfy1Kmrm1tKRWTEF99j7bXX+q/nvo3RaNRpMAVDt21brCg/C4sLYGPgQh+RiHWMNWKdUIZwsludLHsYzU0ZnYDnxeXn572BZlboCJeszHlmfoNnCCVAKSt0QQgA9Jx74DwvnhFQRHiI3/N8fha7xM0bEff89vr163jx4kW8evWqAzBpGpZdxuabzEf5fwv1LCAyH2NJu/EZAj/nlDx7fb6/gRHjnDzLqKyEWWd4U+yBiYgHSsXgx8ownz+P7LX1+VhneCzMT34G5735XA7T2CjIcpb2EshDh74c9mE9AgbxIOEZWC6XJQ8PeYlctXcj09v/2wC0HOtbq9kTkteZ1y33ALhzbmTEOrfSuV0GQXTttzHJAPCgS9wNm/uMiJJHSQNbG37Qmu0xAEjQqAam+7yQeZhvuc8sy2qeINqU5IagTzV+IcCPCQPj0E20aZpiIUVEAQ8e2TuBh6g2WTVl3SdIstBAUIGgI7qt07nXbIEPh8PSQ8aVE1Q4zefzzmeOZ6PscaPiKh4Oh8Uqu76+Lp4K3KYwFAIEGkJbFhTPDmDKc7EpdlvzZtS8a6YH53GOjoWmS8odmquBDVuQvjcSD10ezzFuC8A1qMowPc7OzmJnZye+/vrrcp3t7e3OJooZIGae7ANYNd71c2Wh4wRB+BtB06fgnsfnG8PhsCTmZ2CcvXK2dGn10McX/M7fw8+bvBbZc+GR1+zW1n21UcTacsdgwFvja3mneZdjuwVExLqRqZvv5WP4HbKBa9uIhF6W7QAMAx8auuLBNg1qRhpruk/B19ZkzdDmuTDOAHP25DjfCUDoZ4UXkBl4wyLWXmp7x/EyOjGaZ766uipNazne/ALgwCPEM6NnPNcZTNZ41cfbS+bf1QAmNMa7zV+f3v424xcC/HjYKwAxQftMUC3B1AS21VU7f1aimwQEYMyeJPpbRKxRf/6dOxMbwODeRemy2adziZqm6SQfsrBgrO3t7ZhOp0WIRkTxBiAY5vN5HB0dFaBTs8xsFXgB2P3shEP+Zw4MaKBFjZ58Z6DFvTjH6fb2tljD3B/3FRGFXtAgJzf7Pbk6CJ62bUu4zDtFc13nB/Cc8/k83r59Gz/4wQ86cXas1hrP1PgtH1OzHs2L+RXBRxKow5r5t8/j846maR54fXLYxsYRSs1hJs4TUS9Ft7cnr70+2VVT3jZE/LnzdKz4yEdCvq5Wq1Iyjad2tVp1QjjX19dxcXHR4U1ADgOwYgDF/dhL4Wcj1LW3t1eAmeUZHm6qrUx70yXTFZpzvHMfawYNr1zXvb/8ent7WzZ4hX4Yt9PptNAckGj5TtSDHknoDPSiDTGuaeCGbuB7wBiycLlcxng87uSFmscyT/m7zGvmK4OfGvCpGck4Eg4PD4tH6pcW/IBUSWIFEduiAF17ojII8qTUPjPz+5hsGQF+EAK4LvkO5Wm3bsS99wHlS55S7rWwt7dXukFHRImtR6wrocjxsEXEQuF7jrfQuL6+jul0WsJKACCHUAxq/DwMC0sAQU1w1DwyWTlbgJrpeRbc4xHddvTQ7u7uLobDYadqzeDLFo7d6OPxuFwr8xD3iTUN+HRce7FYlKRLx/BNo+zVqVlHpo0FRA0I1YQPc8q9Wdn6Gs8A6PMOFJo9cRHdEAsDHq2VtvfNW/buOLemTxnVeMrncP5LNoh4JowvyxTaO5gPrdTPz8/LRsER6waJXs/OKYKHTSsUtZvvrVarGI1GMR6PizcXA5eNP7N8yvT3c9foxKh5kjN9Odf19XVnKyIndru0PyJKniP37WouJ21bPuKpx4tvI5WecqPRKC4vLzs9yVarVcxms5LmwF6G3gPx8vKygA3oZ9ltY/8xow6a2PNZ47+anESu0czXfZo+dXzx4CcrBkCNlaHDBxHrzPmagjVz5mtk8OPERN+DfwcTc27uKU+QE3Qzo/EMLIqmue97c319Hefn5yWEBsOzkHOCq92S7s68Wq138+U+l8vlg/vyM5pu9k7lMFemYxasNevU81JD+whQwn3k/HjDUVulWFNOaMTCI6THb7CaqFq5u7uL0WgUk8nkQdt+84mr6XK4imaQp6en5XkMgEzXvoXbJ0Bq1qT51fPkahKEoRVLFi7P47sfWOQOIWQZwjFU2lBZ6LyQLAcZ/s7AapM17Wv7O+7Xxo7BFIrfHc45zo1dfU5SDNgWhhw9QBP3yvvsYWI9R0QHDHirHhQ7IBNjZ2dnp5RL87/lG3mUBjMGbTUPTzaoPYeWv+yPeHNz08nXcT6ewQoghJ5A0BJ5xL1h3DmBm9YqPD/Ai88AVswh8hLvf8S6Zx76giaVfj7uwc/fJ0/6gGONrjVeNW/hZdzd3S2J7U8xPrnXfdM0203T/NOmaf6nD///uaZp/qBpmp80TfMPmqbZ+/D5/of/f/Lh+z/7Mec3sznXB7dwbdhFmBWVGSkrFt4zjHb9lz0K/AZUL9o8uK9cicRiwetChdXu7m68fv063r59W2LDCEmY10nRRtX2/Pg3EdHxBDVNU7qOUgaP5cQCJQwEE2agUQOGH+NlqH0HmFkulzGbzToVINCK+8ohL+cjLJfLAp4Qki7fRygDQnkudkdmPtzc0aXyBl8IsOVyWfbKMf949C3uDBqhRT4+K838u6ZZV820bfvAYntsTp7H0w+vm5rSRIbQqsM9cPK81YAP5/GoWdb5/ab75dXGCMDk/fv3JYEZRWsPAiCP7y4vL2M6nZYdyJFzVsSAApd0Y+C6X5A9I8PhMCKiyCw8sHhpXS3nUKDzg5BzBgbQqc9r5vszXXmW9+/flyati8UiptNpiVI4V8nJu4C5HJ6yYQzoQXcg69CF5+fnMZ1Oy5onFwjDbzwed8L5q9V9lRwFMORSurktCdLwqHVfjadqxi/3b9ld42l702sAnXnGoH0KI+4pNvr56xHx/+j//zIi/k7btr8WEWcR8VsfPv+tiDj78Pnf+XDcx92k4uFYRrY8MsFgriw8mHRbFTWBlP9n8p1DEhEd4ANatpXNvXMs1VxOajawM2M5lwZBwCLmD08QwoMF70WMcuZ7wmlbW1sP2gLc3NzExcVFnJ+fl/MSSuoDkx41gWEa+jNbpyxiKkcI7xloAeYQiM41sFBBkbDwLy8v4927d3FxcVG8PSQEYimRkAjAPDo6islkUixNN5P0s2VlNJ1OyxxhuT8GYDaNGi/mz2tKEYszW7sZjD2P734g9GtzzdwRyomIjjc7/+YxQyPLntpvs5z0OS2TfF4DHTYJZs0RYqYnF2twe3s7zs7Oyrpz+BpvDMMhGUD7aDQq3lvnJDbyMuUw/2q1isFgEKenpzEajQr4Mf1ZA1wz65Ha/HmuMlAyjfDUEOrKfXcMJO31IZUBEOncGrxC3skAWgPsuDfkt4tBeFanBQDgVqtV2Qw1Yp0aAb25f+hQ87jUZFGNnqZ11gs1Ps+8mcHaU4xPOkvTND+OiH8nIv7uh/+biPi3IuIffjjk70fEv//h/W9++D8+fP8bzUfANw7BMsCaf2xkN3MGOn1I06OGQs1U+Ry1BcE9ELOE+WEkx57n83m8e/cuVqtVTKfTeP/+fQyHw6KEES5YAPR1oNw5t2tvmqaEa1DeWSEDOKAJi4h8Gy/YDC5yQnHf4Bntxs/WEgsPgOM5zOfnOHveHEu28PDnWKG3t+vuzA7FIVDx0J2cnBQLE2FlevAsToim0VoWtjxDVlx9dPOxfYCpxp8WFA7/fsw8PY+nH5lf8sBKp1w7yzYrgm8CnPktr9kqz+fp4zNkAF4fPDER3bwy1jgGKqCHNchaRnHbC8F1UI7ICqqzsiG4Wq3i/Py80wiPdgIvX76MyWRS+pNZDlhe2+tjcJcN1z4FnenLPLsaFeDAd1zDHaY5x+HhYWkd4CRnaMsr9GMe7I1pmvtQGf3OXNHL96PRqNNMEU+VgRJGEwa9HQB9Ixtq2SjLuT6P0bXGqxS8GLB+yvjUnJ//KiL+s4g4/PD/i4g4b9uW1P0/iogffXj/o4j4w4iItm1vm6a5+HD8G5+waZrfjojfjoh4+fJlITp7mpD/4pEFA595c7bsBfJC+HDdj7KYLCzMvFbGWXENBoPi0uYz96m4urqKd+/eRcTakprP5yXx+ejoKGazWenHQ96Tw4CEt7K3hwXEYgO84DokfAazQ0svPIREvm5Nsft/0zDPAdewEIpYL5LsGcq05V5dlg+4Q1BubW2VLtfkCyH4KFU/ODiI0WjUyXNAWBEK29vbKzlSEfXSWO4l4r5Sx5VXudrPYKsG7jL/ZSuI39n9ntcAbmKqNh67zvN4umEZdnp6+uB75sJrES9lllNeL3kODQbMv86ny/KrT94hzzyc+B8RRc7kQoq2bctzUJ3K/omEoy1v4VtCODwL8giPBIZN0zRld3YnAuPJ5vuvv/46jo+P4/379zEej+PVq1cllwU5wXNFrDdLpTcW92fwk+esZkj7Lyec22iKiCJP7u7uO/bTVwevGvS0MUixBUnmBjwu4rDcOjs7i9PT0073eYy6g4ODMpcYu8vlstOnzPqB+888uAnAmK+Q9RnI1zxFNbBvmUcS+VOMbw1+mqb5dyPiZ23b/pOmaX79Se4mItq2/d2I+N2IiF/7tV9rQcp4RqxEszDI4IU9YHTuDiJFUNQUWbqnznsDH/4cp7fyxTKxB8aKfrFYdDbxdMUEAAPh0jRN8dTA7F6gZP5HRMclyiKkFBTBE7FOkiYRmEWBJwlBgbcEa89C1/SEPl70Fhj+PytvFj0AKy84PvfxADoWqWP6CGruzQmQuI2vrq5iMBjEZDLphPfsUUKwO//InVR5VmiIEGPOfH8WXPBu5i8+z4KW7+GJTRUXbXvvhfL+Y48JqufxNMMy7M/8mT/Tsk7Ms/CtS9sxUrIsqs1tRD05t/b/Jh7LSsavOaQCmLcBFxHFi4PHmOdlneT7t6eIaiwAkgsKWLNu/4GMZGDAuAM9oJL8vePj4xLmBjjZsAJomD59xkIGPn7leMv4PIfe1Z0tJsivweDlOXhvD5XPbdBGVAAa3tzcxPn5eZFLDkuORqNomqbT5mOxWBT60DYjNxbsM277dKflvaMHNRlUoyXntiwkJzQf923Gp3h+/o2I+PeapvnzEXEQEZOI+J2IOG6aZqe99/78OCL++MPxfxwRvxIRf9Q0zU5EHEXE28cuAiOTkJutotriZThOnBUvIwOnvsnJk20XYAYCHIPSRPEgCHZ27neWn81mcXV1Vao8UGaz2SyWy2WJ8TomvrOzU6qSuF4WrDTlc3wUxsGNHBHFKxbR3Vtmf3+/LBDvlzOZTDr0ykCS594ksE3DfCw0QCicn58Xb04GndASARERxZ18d3fXWSBUmSBoAT4sdOh9eHgYx8fH0TTNg20Ftra2SkWYQ2MW1CgB3NaDwaAD2DhPBnw1vst8mvkwA8fMoxFRBAXADp55Hp935PURsa4CgwdJ5mRsUr59n9cs9Mfm2+sX3mS9uSrJxRP7+/sFrCEbFotFnJ+fFxmztbXV6RzvkDaGAesLDw734x43GFIYXXxHGoErYBeLRcfYaNv1ZqrkPCIfHCJ3vqXlm73sNS9FngfTES8YvwPU4t0bDAYlL+ni4qLsT+n0BmjkEBc0Q97gJTdA5Fp4uE9OTmI0GhUDHI82v4tYb1x7fHxceIYQJp4+rs28ZF1semR6PQbI+U0NpJu22ej7lPGtwU/btn8rIv7Whxv79Yj4T9u2/Y+apvkfI+IvRMTvRcRfioh/9OEn//jD///7h+//t/Yj4BuLwl2OdQ/VBe+R81xqwCf/38fotetyDRaT+zcMBoNOqSLfLxaLODs7K4qU/B4sGm/kRzKuPSUIqewyxrXbNPexXZiJP5ifYxA6AJyIKMDp/fv3sVgsCu1YzN75vY+OHzu4L2hnr8p4PI7RaFTi0QgVLCHuwdbycDjsCAsWKb9DuBA6df4WYPFnP/tZubbzELjfyWRScmoMPrK1l6tMXMniyh9+XwPlfZ/x3Jn2NeFxeXkZo9GoPHtf8u3z+G6HZUr2+rDe7dWM6K8O9JrmfBkMO68wG3++hq/j+3N1kr0KGA8R66KMq6urUiThEJwLM3JuEModeQlvQxeeKeJeXuEV4Rqr1SoODw9jf3+/eE3oZk9YiapZZCMAiud0iDwbb5btNjjymswGoFML3FWfKjnCdXihyWOhYaGN5bZt4+LiouOpshENXQ3kuE+elZwejD/uhfsDjAF2yTO1rgGsZk9/5usan3JsrmTLhtvHeHIAXk81vos+P38jIn6vaZr/IiL+aUT8vQ+f/72I+O+apvlJRLyLiP/wsRPZS8KC6wMmm5SE47lm3j7glBdCPmd+hVFgFpL18iIhqdFdikHoNzc3BXUDaiw0aoDD5aUwJuGs8Xgcw+GweMtYJLivOZbFZuvCuSp4SLjPWqjlY5XpJo8bQgIrYzAYxNHRUbk3BAuvFu67u7slkY97zBu2YvXhRXO4ihwmLMaLi4u4vr6Ok5OTkjPEXMFPk8mkVL7kZEqsE7rOtm1bwOOmYZo4n/HwrtUAACAASURBVMP0q1k/mwQH/bCgz8cUCzyPpx9eK67CozrIo88bmL+zYs6yjDXs6/fJvnx/Eetyea95epPBn7TIyEmoyKNcyBCx3jCZcJZ79CBjbMSRgItc4tkI/3tDYbxSjL29vRgMBiXnxwaR8+DysDxH3vj/LNe9Tm3sEIJHbg0Ggzg8PCxzgcePykxeCc27YSPAEdrY+HWuI0aOUy/u7u7i7du38fLly5IGwrMcHh7GYrEony2XywKUOC+GuUGeeSjzWObViHXT3D7PTpZ1/n7TNT5lPAn4adv29yPi9z+8/+cR8a9VjrmKiP/gm5wX4tLEyagzM2ONkVFWH67ve3kwkf6Nj/OxfG9Fx6Lku6ZpYjKZFKXDZzs7O6XayD0dWPhbW1ulsstWBNd1TpHzbiidd+kmgsWLnvJR947gz8m5BnHZQxHRZWLTMNPSijwLXB/jBevf4Ra+ubkpXrL9/f3SlAvvDc9ooONFS84VXiFc5NDWLn5bptfX1/Hu3bt4+fJlHB4eduLVWK2TySQWi0UBOFn5QGvoXaOJeTMrs018md3IPm9WmvP5vJS5Poe9Pv+oeX0IadjrY0DTp0Rqn+d1FfFwR3H/JvOHZZplW8Ta64Mcge+QQRhHBuvO74hYe2dXq1XJJXF5t3+TS6v53Xg8jogougDaMYgQ+P4AWQ4fGcwYZNVoy/N8zJrJc5zBFc9+fHzc8cDbc21wgJzCQHW4iSIR3mdaWR7byDs7O4uvvvqq7PtFlfBgMCjVZTakPd9ZNvnPfOhhIO3f18BjTYfXZOQXB36+y0F+S9/IiiMvXguBnHDlyai9z9dg2BJAaHEcLk0WGoqYHgxYO9yrOwdHrF3g3IMBCNeyJ4P7YbEBtFj0KOaIKO3B6faMEB6NRuVzFp0tVFzHHq6uMp1qHp1MR15RBCw6cn28RQXvacf+6tWrODg4iPl83olXR0Sp8jg5OYl3794V0MOCWS6XnUZe5o+2bYvl5TmezWbRtvdJk1lB7e7uxunp6YPSd15d6poFhUffYva5DP4NnPtoy3kRkOyOXEtAfR7f7cjAh3AXCfcGzA7D1oCKAXbNmGDUcizyefNvMo9iECCr4GeU9XQ67ezozm8xxCLWoMTXxtOFbMJQZF3hraaEnfy71WpVtm3A+wM/YziSR3l6etrJyfOwfEVeZwPXo08J21Ph39rLhMzY3d0tOUoYqRhu3mPL3jD3ZENOQi97cdFD8JiNOudMrlb3BTbOl/TWFvAf4Irwofmlph8zrTLf1UJVPq7m8akBohpA+pTxxYMfJ9bVhgVERqYw2TdBkNk1VzvWng+f08nZBhc3NzdFibKoERoR0WHi+XweFxcXD9ynDq+wuCKiw1heeHyHBwjgg4BBsJHxj2cIS4ONPzmvKyy4V36XhXSfoq+BSIYBBc/Ib/Cy4BImMZnOqbSQxz07m81ia+u+qynXcJIztEQw18rkERDeO+709LR42ng+7/juZ6l5Z/p42LSq0Sx/n0NhvPYptIgoJbUGfs/j8wwbJbbi3Yupb95rn/UBIt7784h6GD+PmrfJwAc5iiIlTzGi24qC41wJRrUh52SdIX84zg1MMRaRoxHrXCQMI8Jc+/v7MR6PC4ByYq8NXufEZAXtpGdeM/1qcs705z4dAmO9su4IZ61W9+XkGFVv374t5+f+cgsQr3V0IvrF+sH6AnlH1SdgmxQL7tUAjNJ3vOU2+ExH89Qm4JidDqZlTe6ZzrVjngoAfdGSsG3bonz8GaNmxWTCOCGszy3H+8cQrS242oDpsNJZ9Ofn5x1rw5YRitPndJKZE1TtyckK0MnAfjYWJUKIMmyDFic1sx8YAMgJ3ByXFe7HWpKeLz7zc7szM4uNPXHG43FpdZCtGTY1dIOw2g7AABULa65Pzg/0cJuBpmlKUufx8XERCtAA+mZLMFt/DmnWAHamW+3/bHFGPMxRqF2DRMdcSv08vvuBwnBoAm+HvcjZus5yLysPzu3P/FvWycfItcxzGEcoSIMfttvheDy3ud2EvcIR8eD5XWLO5+70HhGd9U4l6sHBQfH0kF8JHalopLLTSf6158zKmbmqhfbznHreAB451yb/njBfRJQ8GtIhPH9uWOs8K4owHN7LhR4kNtuQc+4WRt3h4WEBn66ei4gyP4Aie/lz+knmJ7+ax/JnOZXgsXNkef6p44sGP8SA+xZ9BgA1QkZESRqzAKkBoj60H9Gd8HwvMOb2dreR4e7ubrx9+7YIBhZituDs5eBeI7r5NYAeW1QsVCcVWhGjeGHsiCheJ+6XtumcGzesQ14sIOetmPkzs3Jffe7MTGuex8KLhToej0u/CoMxvDj8Bj45ODiI09PTTkgM76GtFyckAkCZOwARYDZinTh8dnYWbduWjqwOU9Q2EDT4yRZT5qEMcvpApcGOj/H5/BvOhTX3DH4+74Df3ISU6qVNIMc8kPMMa0aF+Sdb/jkdoMZv/hw5VTNG3DTPv+Pce3t7Jb/QOUEAKOeAAIQ4Dk9YbvbH5+QMAX44lvANtEaG5cIB6GAvTaZJpofpXwNBPhbAkc/Btdz/i0qsN2/elOcklIcsQY+gP0xjql55Rsr6eU6AE3Kdz00L72Vo+dW2944H5pEqtSz/azQxHWvyyTyQf+fXvuPz777t+KLBT40ItUWff2NhEBFl8vMCrwEZW2JcI4886Sg+VwVtb2+XvZ4cEoFpWdAuV49Yl8pjsY1GozLxFki2Jun9w307VMb1+S2VSoS1CA0hqCw8nJwJ8LAlaQuzNmp04tWJyTnJOwPc4XDY6VmEQIAWtHvnXAjJq6ursneNrR+uzat5yN4hbxAKQKJ9/P7+fhwfH3di4t4xHtr3JQp+DH+ZxubTTG/zu93fWTBfXl52qmGex+cZ8CX8i2fDvJ7l2MeMzCd5WGb6GjWwBQ9xT4B2F0gAoO1dttx0vx6UJQqV7vz0o2nb+9wfN9qzR5xKMnidHMednZ04OTmJ4+PjaNu27GWFnCI0PhwOyz2xdrPi76N5XjfZyDBNLUP65gDZ7XxPwM9sNivdrAGT2ZMG8HOCtucNUE0uGb/xPnEGRQZebIEREcXTwzxwv25CyXAi9WO8a/Bnmpg/fZ5NIGiTvvmm44sHPyaE3YgfIzhQApsSrrJg4NyuErPSyvFgwAYgAwHC9gLZkjGDe+8blOd4PC5JaeSXkHxGPNaWHOc34/Md5fZcn4XCnlXZu4B3h5BXRDeBzp/znUf2DGX6mo45QZgFalc5cwz4sRBo27bkSrEwsfywTPEiIZDdfI3fA6Q8twgr6MwcW4ihwMbjccdVjIettlAN+jbxb+Z7W58ZrPncFka1cwJwn8fnHawfPCrsbl7zxtjL0HeuTUZFPsbXycZelmsck3NNWCMUGVhm2QNkzyyyDqBE+4q7u/ttHQjDcn6HdFhvyBzu5ebmpuT70cYDwwYPdkSU3DYrVXvRswznubMO8JzYuKgd32ekuIqNNQhYI8SH/gDA0YeHXEaOB9hhoHE/9v5wLMUjACBoagMZADocDst9kfPJfV9fX3fatvQZcv6/5lTIdKnpAL/PayK/f4rxRYOfiC5qhLCusIp4GCLIv889UTJKz8StMfOmkUsMI6JTjeReFGYiK36Xgk4mk05CMq3JWRygeitrlDwxYCwse7zonXNwcFBcqfaGeEFBC/8+W00WInk+eM2Lw8f6OtCAYw0O8P4AcgBzzJmtu8vLy2JhIgQMNtq2LYvZ4T9fK+c2MRAyCIfZbFYSxXPnZyeD5nN8zKhdf5PVU7NKa4KErrPP4/OMpmlKknPbth3ezArDALzPq5O/q33ed67a9XwcgzUAqMED5JC+160TuZG3ll2ch/B0RHQMGHsqnTNJ7o5beUTcGzvsHzabzWI8Hnf2+KJww7LL295E9CePZ5maZbZHzn3JHle8sDae+B8aQTvovFqtCuDBK2Y5i3zDu0XekCvxuF/mBG8T4BC5GBEljEmOJDqI+2d+oJmBmuW3aWqerH3mkfVGDUTW3j/F+IUBPzWl2Odd8HuYsi/Hh9dsUdtC55ia0DFjNs29Z+Xi4qJ4Jaww8SasVuueOjAunpnRaFTySlgQ9MWg54YTBx0SYzGQ7Ifn6e7urmzM6fJP3M7cn60lV3rwZxBpANQ3oFe2ovgc68JCNmK9Sexqta7IGo1GnaaCdpUj9LCksIBGo1EcHBx02gfwSgVZ27bFEudenYBuJYX16hyp6XQaTdOUbq3wEdZZ5tWa1dnHl7XvbYn2gU6Os5Lif+/K/Tw+z0AhXV9fl071GXxEdMFtBiisjb755jvzVraw83G1974P1ljT3CfluhcWChxF6lBNxDphFm/1bDZ7kACMJ8NhNEAMOY7eK8zd5+2Fmk6nEREFLBFuc4jIcso6gWtbR2SDmc9qxl42evnO7S+stzyXu7u7JTwPnehQDUjmt7mqF1nIeQDU0MFFLKRFOPeV58WIhmbMOzqHOXWelp+zj6cY0MG07ZNhBp6eJ6+LTQbgNx1fNPgxYUxYW/I1RZEZE2b8GEVd+6z2HcNgjPwTFqwnFzR9dnYWEWulSgIcuT0o8vl8XkI8HOuEL/JyDNAI6TRNU/ZCAyQhoOibhJcou99diuv9dmyRWYBlmm+ibWb2mlsYoOVKEPcsOjs7K4uYPAI2NnQ1G+CV3+HloeTfz7C7u1ta41uB2IqqAQsE8XQ6LfPh30K/GtCpCYoavfKoKb9NCjGDetP1eXz3w0AfAJGroCIeuvcNYmpyq08m9Smh2m+9BrIstTFAh33fX9u2ZV0SkuZ7FDVKl9+z3n1+ruEwfsS68tN5KxFRwAKKmc/ZC9HGIcAh04rPnRDt5zYora3dPm+HlbVzAfHiREQMh8NikFouIdN9HZ7P1+K50AtUwGL0QivnayHj+kKgABy8SOY7dIZTOjIwz6Nm3DEy7WpAsgasTNtNuuabjC8a/DC8WPusoxoBzZSO+XoS87CVbMSdXZyMnARGq/c8YSh5FiaW0sHBQQwGg1gul2VzTToHcxwKnefmnvByIDyw6mF0FDC5PCx23KmAGnumqACzoIIuzk2ykMhMXGPQPIc+J38ICO7T1gbX9T1z/dVqFWdnZyUcZgt1Op2W53MCJrxydXUVu7u7cXR0VFr1czxzyrVysqFf5/N5aRXAs+a9iwyaTBPTzEA0C4Ka4K3dC8P02wTgn8d3N5gTcgDzushrxYZd7Vw1j43PlT0cffy2iQ8NZCKiE7JHruAtsCeRa8NzFxcXRTnbg9A0TecaGGeM7e3tkvfCNVh/AB8DSfcvI6/P3nFkiZ/ZBmNWrF5XDu/kYU8I/+ffA3SQUxifhLuc58MzIX+yR9qGsI16G1nsLuDGwKaR54/nA2QOh8OS00lIjvBl9vrn/FvTqMbj+f8aAKp50Pz3Swd+soD3533fW2Ey+hjYXo2aYMlhNg+QOkwEgOEzu/FcLkgVA/stObma/gvHx8fx9u3bstszIIg4LIKO57eXwn0hSHxDoDgJeD6fF6HCeUD6KHoWo3NsMv1qTF8DCTWLCvd4rpByGSvAFWFogTUcDkvFBFuInJ2dxdbWVhwfH3dyAcbjccxms5Krw/yQBEz/EICYlUMGkeYJ5mQ6ncaLFy8KcKpZc7zWrO0an/UJjpryy8dlD88zAPr8o23bothc6dT3x2+yccHnDAPhGuCJWINfG3ooqr7z+F62traKRzSHvMh782am8PpyuYzz8/OSu8L1bdDYC+GcSXL0vHkqXmE3THSRgY28/f39DkCAJjYqbNBkgzrrAdOI93ku8lrk2XgG5BgAGA80+Zv26Nvbvb+/X+Q1IXfmwbIIAHN5eRlXV1dxeHjY4UF7oe/u7ooXjtxHDHH6o6FvLGvRKRmkmCbQwnK/BlhqnqEs0zatj6caXzz48cggpwZW+tCnwU8+R568rJj6LGfnBeH1yYLF1pBdt3gg7OZdre53C6fz58XFRQEETha0QsWicjXXzc1N2UgTRQ3z39zclF3Lt7a2yjV4HgQ15eKHh4cPNuirgdBMP95bqPJ79/twTJj7cHI3wAuAx0Z8/M+uzltbW/Gzn/2sABeHE3km5wzR+RoBYrcu+90g2AFPDj/6uXjWy8vLGA6Hnc7Y5pPHFnCNz2reHtO2do5M+75zP4/PM1B6EZvDVTWDruYJZNSUdN95fe7s9fH18FSw7vBW4QFomm4/K9YEip498ZwzGLFummiD0TzqnCbkFQYY4AHwg+HYNE0sFotyT/a6+hoGf9kI82umT01fZC+H6c/w9zZ80QHOZ+Lzm5ub2N/f78ghnhU5leUpIUFkO3KSndmRj24LYvmL7gF8Xl1dxXK5LLrBegPjNANL07PGgxkMbTrGPM//lpsGiU8xvnjwk5kyg5LMoA6JeXwbgvUpFzOSLRqUZW0BIUy2t7fj6OioMBtM6ITc+Xwe4/E4Dg4O4vXr1w9cthxndyQADBqQ+9O26xLnnZ2dGI/HpccPFgOhOnKE8IS4BxEeET+Pn8/u+hqyh541JH97e9tp0R6x3pqD58LSibivpGPhIzQWi0XpQUIlhJPEV6v7xnJUg3BeLzQAZEQUGjkhEosLT58BLYJ9sViUijo/I6CpxrvfhB9rCpLzWaFkZffUVtPz+LhhJWejJYNX/1/z+nBMxEPe2aRcasC55hHhevbK0D/LMs2N7iwDyJujISmAxtdiLfE/NOEYvETu2Ny26+Z9yACHn11ezzm4V4d9vV65nsM/fUYGo+bpyN6eHPYC+PhzQlcADcAlMt0AA69O27adZpGAqYi1V8neMfoHHR4elvuA9wiBOdrgcCSNUO3ddkpF5reaTMr82kfPTHPo6HNx/9YNTyXLvnjw42FhD1FYJM6zyeDDC4mRGbjPasrI1ujbYCz3xuDVCm93d7ckvK1Wq87uxng5lstl/PznPy9ehK+//jrevXtXzuGEan6X7x9PBgvNHgvaw/MdIAwrIWLdPZnfuPsqC9pCa5PHAprl77KANpD0/NodjtCIWIOjm5ubUuk1GAwKbbyTM/dLddb19XVMJpOyUzQ5B94ZGiCE1UVVhGPs3CdCGxBMKappBK1rz/8YHT02eTDzZ+b72rmfx3c/UFyblEUN0NZkTm1smtM+QFUDT7XjMBz43w1AfX1APxVdEeuk24ju1jGUvkfEA2VOgUdWtvzevXwIfxEeJ7xPQQPhM29Tw7Nyb/b8WO5k5e7jeOZ8Tl5pDAhw5BwGKwBivIGAvTdv3nQStQl70Y+MObEXDMOZeeM9YJTNjF3pBU1s2PHKfWPsApz8DMxvzfDKcv1jRjaInZ9r4NOnT77t+OLBTyZoRFQFiUEGv/N7L9iPESReADXl5Dh7RHQ6d9YsAJLJ3P8FJo64t1hms1ksl8siIHZ2duLFixdxcHAQb968KZ+xdwv36QoK7oGY8cHBQfE0edEjGPb29uLg4KBjhbk/Rg2UZOCZadFH46ysDU4RVnhYssB3zhENzhAuxMWZB2hCLLtpmo4XC+8awGk0GsXh4WEnKZMyXZ6T8FkGEtAxW7B4pCxAspVt5Zb/zzSrWVGbPq/NR20tPY/vduT8liy3/Bmyw+stewuzQcfnm+RkbWQgZH5mvXBfBhkoJoe1UOTIFyfbAmxQ4F6vOf+HvMQM1FmL5O0hHwFnbmJKdWct184gxjkzeWSaZdrm+8se7+VyWbx90BQvPWF63/9oNCpGFZ4bpxpAKwpf7BHiXqGB84CoJsOLTe5WRHSM34h1KgURAHJ/DEDhjcfkSAbtm8BKBj5ZTxsA/dKBn4iHHpisJEy07E7jvWOuPufHXpeRhQ/fWWFjucD8WCN2yUasGfXu7q70BqIC4Pb2tlQQ0Ujv7OyseG8iomNd1ATG9fV1aV9Oz5/ZbFbyVwA+5NCA/CPW5eVW2BEPW5RnIJoVdwalnjsWqRcy9LNHDyEA0KA6jsosV9DZUmFebPlkYctij4jSLI2cIy/+u7u74jFjbl3xkHMZbDUzaiCxxnObeDPTOh+bw7tWnH59Hp9nbAKi/t6C3kZUXl/Z49w3sgKqHWueANDwv5sKRqw9FnwGEHIuHaDfm5Y6nO/hMmyMFAMAQA2f0YID44zf85krK3nu7OWyIcdflpt9c2aa9emELH+RAW279v4hxwjz85vt7e0SbsITz9Y80DaH6l0FB80i1ikIGHPX19fx4sWLsmMAQAJ65TAr4JLjmQuHLXnu7AU0Lczb2SmQaZ3XhIFjBj9PNb548JMZ1yNbxbaacsJdFiyPXTMDrL5rM6lOBHMeyMHBQdkvxfcLUqe8mt/Z8rm8vCyxWzw47969K02xyEmhPw0AzPFdzu2kt+VyWdyzVDjRy8GWBQApNw3LIwPSzPSZfqYpQtQCnrlzno9DTgBJQlze2BRAFxHFs9W297kBJBoyVwaqnI99g5qm6SReI1C4Ph46K6Nswef2CllgIgh5vscWdxYUfUIk82fm2efx+Uc2CvyZ5VfNukV2ZYMrnz/Lw2/CU1ZkKFsrSNaJf7dcLovRZo8titJgJCJKibevu7W1VQy+iCjvbczu7u6WLStcwm4vU/bg9PVTcoiL8/v1Y0Cl163pwTxxDpfz+7okZ7dtW3q1IWcAQPQz4zzIEmRQRHRCX05+Zt7pIXR7extv3ryJr776quT72Ph02IzrXV1dlbmo8dYmOmV56PnuA+U1o47PDX6eUob9QoCfx1A2n0d0d2u30OBcPm9+3WQdbSK8QRdKGg+NNzzlXjjnfD6PxWLR8SQQfuIcd3d3cX5+HuPxOIbDYbx69Squrq7i4uKiXC/H4s0o79+/j7dv35b9cLDgcDGTMEiCI54SPjfwMb1qFlCNbvm9/4dmbj2P0OX+nfwHiImIEucHyJEoyX0igF0pwTMBdsgbiogSorq4uIiIiJOTk7JvFx4nXL8ILQPODLQRItlVjOCu8XVfaLaPb31uK0eP2nWex/czLKPy+xogYl4jHoZWasq6T5nkfJ98DzkkZ68MgMN7wjVNU7pVZ/7GmHGYhOFQHp5Xy2uULVZ/XqMOJ7MhMzmQ3vi0luNpr0kOGW2aq75joGVtLde+t/EJkCNUeHd3F8PhMCaTSdzd3e99xiBciFxzuIoIgfujubWADb2bm5t49+5dvHjxopPDs729XbYQsQxmywsMyWzc+RnzqAEbf+fPM+A3zxv0uArulwb8MPLizUIgh2MgkEMgNcXg8+eJ5Zh8XL4HL2aUHq3Lvfme3awXFxed6jAWCAuaxmI0pyKJjlLq3d3dePv2bWFgK1eX0S+Xy/I9C4zGio55I3xYnHis8J5449S+uemj66Y5zQ0YnQDpOfU2HAhAtryIuLemRqNRDAaDmM1mRajQLNL5THjXbJmR5EergDdv3sR4PI7JZBIR6wRrwONqtYrxeFxi95lnakmCfG5BA+0+BoxnnmTusyKzQM7nfx5fzshCv2Yd5+9rYMev9mr0/T6PbBzaE4uR5P/v7u5iOp12QsnwGevOAIT/kSN4KyK6vW0AXQ6ZRdyvzcViUWTh9vb9thiEcs7Pz0uLEAOcmgxHzuBd5vmdD+T14jXktZT1RD6W8xqE2KPtTaJtQM1msyLHMbh879AOPYHuyQA2h+Tb9j6cNZ/P4/T0tMwZeaeU2XPvyDUblJnnah60PDaBo9r/vDrchTyzPniK8cWDn8xQNeSZielJrBHT3hdGnkzOX/s/u5+5hsugSW5mQTmPZDqdFsY1cCF8A3iyACAvaD6fFy/O3t5e/PznPy8enLZtO4uKHKKIKNVN4/G4WB5YS1z77u4uRqNRAREsPC/GTJcMeiw0sgDx/NkSc9Id3hksEM7DnOJxGQ6HcXx8XHZXb5qmPA+fz+fziIgCWnZ2doqnDUsMfsBSZVAuend3F0dHRyV5MLv2s3cvW9MWbNnK92t2o5uvavS1lWSAn8+b/57H5x3Mk70ezrXw57Xf2aOXFYaNqZp8jKgnzPcZklacEdEJ93rfp+l02gFIKHQULOubZ+U4jK6rq6uSJ5Q9QO6txb3B49vb20U+sdYHg0GhBSFw0gC8YSigh1J9gzWDhxrAqc2L5RgDGQMIswG0Wq2rZjGMKURBrkC70WjUmWO+d66T905zM1h7zswT8NBisYidnZ04OTnphPJtyPPHNQaDQYcG2XDbRKMaT246ziODH5fnPwUA+oUBP3lkQcCrwQ6/N5GzMMnK21ZA7RifMysZvBjskBvR7V66Wq1KngoLw4mFXB+hw4QDTOjCOZ1O4+joKI6OjuKHP/xh8SIBAHiG8XhcFj59gxBkbiLm5F1fF8CTlbXRf6bLpv8/Zp5d/mrBaSA2Go2KK/7ly5dxd3cXr1+/jqZZb/jHHmoAT8JXLGq+g97Ob7Lb//b2Ns7Pz+Po6KjT+JDvEfD0FTLggw9zeXx295v/Mng0H9oC8vsssGvgh+ftE0TP47sbm8BInsMsr2oKorb+8pwzahZ2n/Fiix7jgD40rMvZbFa2aLCnBs+oFVbEPc+xXr3ekGvZY4+Cy/lweINoIArQYr27RNzeC+4dg8Wy1bmN2UgwGDIw7TPkkFHkFuJhBkAQsiMa0DRNSSjGgwZgGo/HnfJ+rkW6A/PId1zH9++8RtMj4j7dwmG229vb4pUj34s/80gf79X4zqMGbjIvZh3NH3OYZd9TjC8e/Hj0KYZMKB9r4dKXOFVDklmJ8VkfMGIRAIC8LQULbzqddvrUwGgoY3s5uGeHfzzxV1dXpVPz4eFhTKfTsgDxdPzwhz8s9zgYDEpuECCJ57EgYiETZrK3xwuqRpuIhxVNfQsiW145d4A4NPlQPBuD0te2bePrr7+Otm2LRYobnHMRBsQ1TqgPAGghCMBEsHCftPk/PDzsVFYwL1TkYb2Zr2yxRKxj86ZdzZIxrR0erfE2NMwJnXkOnoHP5x1ZKdQMCM8hIysDK+HaGuwDtRls5ftwUrCNMfMouab1DgAAIABJREFUPMUWHchaKybAiRWTDTHuDyOLPjIO/TMsyx2i4r4wYGjcSp7gcDgs3fMBPA6pN826EzX3ZtDRJ9vsYfOcZiAEsMMbQ/4TScbIXe5/NpsVUOTiDl+TOWI7DDaSJQeL9Ig8p24Imb1ayDPar+RcT3gN3cOzZT6yHOI3ffzm0Qfusz7OXp9furAXIwMfPsuE4pgMjJwLYYvC5/RfzYuRP8vgJyJiNBqVPBOYZjAYxHw+j+VyWXZAduKrf28QwflZ4G7EZ9Q/Ho/LxpwALUJXFxcXZTGsVquSvGy6uUHj1tZWAT7+Hd9lq8ijJiQ20S6fz2EAFi5CEOZfrVYdTw1birChnxeLQaX5gT4gJPvZI4SAMijEQmPujo+Pi5VkPkLgAY4ycOnzzvQt5hrwybxnaxlFtkkRPo/vZ2Qrl1f+7D3gO4d/ar/LvJNzd7L3opYPZG8Pa8Dy0vInJz1nWcwrnlBkCesLDxEgBrCAl8FrFnq48MMGCd/T7wt5kfMfbWS1bfsA0OX75/weWY55vjI9yJ9BPvB8PLdLzQGQGHn22mDckYNIzifGLi0G6Avk7VMAWBFRrrG/vx+Hh4ed5HE63lPsAc/Z+829QwMDnEyrmizr89LU1oA/t/cnexN/acBPRpf+LGJtJWTwkxVxtpxrIGeTl6K2SDID5OQ0Xi8vL+Pi4qJjHbEhX7bq/FycG2FEkrJBDCWJhLTm83lZZMvlsuQCXV1dlXNZcSNAfB8GJAizXLaa56LmIeujU43eDvc5TwYAhNVoF7FDjG17v6Mx20tEREkcxxsDiImIAoAQBvCD8wMMyq6vr4sHaT6fx+HhYWlOBq8hZMiz6rNsoG/2COXXmhVlC6zPYjJ9Layfx/c3ah6abKxlBcP3yInsIe1bV/mzTd4fA3Jvc4CcglcB/oAM7/kEf2FMOHRtb4S95vYoNE1T2le40jUiiieCPadYixFrQy1Xi7JOkSVcO+uQbDxvMupML4fkDIw4B9VoTdOUfd3YzoICGHJ/aFwIjSKi7ALAc7obtPWdw2KE0+zhst5ardZNEtFTy+UyptNpTCaTqgcLGeX+dFlPZc+cjbsM0h/jSc+F3/dFdT51fPHgJ6LuDTEhMir0BHhSPXF9QCgrno/xAHFvVCIdHBx0NgQk8dbIfrValSRovotYNw+LiCJscE/yP4xP7Hw2m5UunqvVKhaLxYOd4xFGNNByOWlElHP5HnMyHwu/pkh5VgNP09JCwqEuu939e47lObFAcr4WNG/btlg2WEhZMLqjs3MUsvs7Yu2FQvDnPY7m83nJZ+BY7h3PVOY1z3PNE4ZwqQFs/yYD95pRUJufPt59Ht/9yLS3/HLI0rxUk1F9Mqs2r32f+558H863QfGioHPXZtYmypS1iOcFI6ttu3mN19fXZU9DcoccnrGXxN4ZDDBXH3EfFB5EPNxSiGcjtGRdYsVtT31WzFlOmX61wTnwdrmQZbFYxHQ6jb29vRgOhx06zmazmM1mMRqNYjwed0AH6QyLxaIAG4xePG146AxUGavVqlP5GxFlHtAdlrvoG0cGDH4y3/UBcX9e8wJlmVVzUtjz03eebzN+YcBPzftiwhjc5GNh6iw0GNkC8jXz9zXF5MXjY7e3t+Pi4qJT9WXrjfvOCXcwXkR0ugq3bVsy82190JXYIS28EPR9YBHi9XHJ9mAwKGXxCDCHvewNyY0ObQ35M89DTQhz/gxaOTaDh3xujsPSgp40frRFSuMw+mrwvOQF2TI0wHRFnq/LMfQXGY/HDxY51/AzGLxn5danxPJ5MziqKcp8rhp/P4/vZ/TJsZpssZckexiy0qkpn2x05ONyHgfAC2u/aZrigXCDQ98zQL9t1yElwAqe3KZpitfDzUEd6kJOAbQcErM8slfV1U14B9y3ywDKTRvzWuIe8zrP85DXWUQ3x9GfmSY59HR3d1cMKrYgor8Ysufdu3cRsa64Q07yHNyrvXEYes4ZBVhBKwZyDeN5b2+vyCxo4P0Rc6WVPdePAaBsCOeRacqcOKfM759qfPHgx0TLAiJizQxewNnrw8iEM+NnYZPDZtxDn3La29srSBov0OXlZScB1vfsRWlvC8znUnOYvVZ9xX3TH8KNqWAaYupW9M5TwXtB6MxAxy5sb3dhJc79mGablG0fHWvglf9zJZavx73e3t6WfcouLy+LgEAoYB1yTha7XdFc1+E25pPzYA3t7u6W/iPD4bDQF9r39cfooxXH5WfPVlCmOUrMVlNN4OTX5/F5Rs2gqsmsDHQYeW0YAGXlm5W0r1+7L4Ny5B4ya3t7u4AR56k4oTYDLOQMCckoY8rbva0D95eNj4gugGDPLgocADbIN54dJQ9A4jwMQESOINiz43XE8+VRW18ZECE7+R8Z3DTrnEKqwpA90BRAgkwmveHu7i7G43EHDHjjZkrsc7Jz9txZrrXtfUoEmzE7Edy86rYjfmbzY41ONX4zr9aATw0IGajm837b8UWDHy/IjMQjuh0grRysoPP5aoTLFpePrVlOtUkFlFgZ41nIMWMEAgLG908rdxaBY7eEgK6vr+Pg4KDcE4vMJYtYa05+A1A1zXqjT6wxXKd4UQyCcizc3isLiI9VqqZDzTvhP1sxVg6e+62trVIdxqADLfRxDyZi38zB6elpETaZD+yVyxYiwJIqMpf02jrLvGRabaJZBoG+L+4nC4vM/5ussOfxeUZWGBHdpqyZv2pzbr73ujFPbrr+Y4AIAOA1tVrdt+ZgraDADexZB/B7BgauTHL/LntiDO4IleH5oTcQ5d9uy2HvMVW05NxZXpjGlu9UOGV5k+WSX/18+Xx+RQfwLICQ1WqdVE7lKPdICBE6+Hr7+/sxnU7j/Py8gECH5jkOD5A92Pze6RQOAd7e3sZisYjDw8M4Pj7u5FwyH8x75lk/s+lkkJL1a9ax/C7rcR9jWfdU44sGP3lkxnMSFN9nZB/RZdxNyLGPoTcpKBa94+DEV62MfS85gZHnGI1GJdktIjoNuSKi0+WYkkoESEbiVAIQb+Y7wjlYZrihT09PyznpVeHSeO49e9bsfdkEhrKLPS+YGtjJtPN8AA792fb2dgFwLPY3b97EYrEo++bwfBFRBCYbwLbtfROw7M7l2gY0CHoA6mKxKOf3b8ilqgFm02kTAK/xY5+lhWXu8/OdeeR5fN6R11CfkZWHvSv2/ng95bXlUZv3bFDyGYZBRJT3hDwwNpAleAj4HK81gISCC+9E7q1p2JQYQOMkYGTaYDAoLTrYzsJgyQYk+ZPHx8elAqrPcLVhl9fdN1kbNWONQX4SnjQ8zHiZ8RofHh52Kk85jnl3lAC9cnl5WTrw8wwYPABN6yRkELRDj1gHtW3bqRb2lhnQGjnWZ8iZxjWa8H1+zYAnG6H2+vzSeH4iHhLLn9cIUUOVj4GdbJ37tTbZ+XxeRFQQ0YOCZFy7je2upcwQ96a7vzppzQl5gB1n5HNdzs8im81msbu7G5PJpCh+rn19fV3CQ/zGAsYKtOYCrtHnMWbvmyPT0jTle4cisxBzPJyE7sFgEKenp7FYLMrmixcXF3F6ehovX76sWqF4xmiQyL5gXJc8BocDnXvFfjjZyu6zyh/jq4i6x7FmGWVFVgM+Pt/z+Hwjg9wsXyLWc+qwKeva4Voripqn0iMbc/lYXnP4yoo0Gxx4ZPjf9zwcDjtN/SK6BRzuIXR0dBRN05S9+QA9gBladVA67lAMz07C9Gq1KjLOXhEbmbXwoGWc56j2meeyb93m45AZNKalEIX7oWiCpHJkPKXrq9Uqjo+PY3t7u3SHRt4DKulrlA3HLBO2trZK/zfrK/QIAHK5XBZdwXMSZnSKiWXfJvBT4zvTMeuCx5wTv1TgJ6LuNrYCYPS52mrHGC3nc2bPESNb03kgEHBreoddfkuJKFsi7OzslITZtm2L29LxcveDiIiy1QKCghAX1UwRUSyJ1WpV8oCM9rEwWKC4Y11NgVDy83kecuw8W5b5d/l9RvO2Ynh2D2gEPXwvds0jYKj2oow2IkrsnJbtNzc3JR6OQDk9PY2trXXFS0R3uwr+z7lRWGaHh4cdUO2QKM9RE6j5u0zvLKANDC2wn0o4PI+nGX1yoyb8I9atFQA/rh7sUy41ZePr+5X3Xqf2SCFTbLgR9gCI4M0woImIstEowAn5EvHQSMSr4bVogGVwxW9dqk2HZ/Y8NKBx/p7lOs/qHJtccVujW996zWsuG43k4czn8+KpIpxI3k/EGhgi2wC/0BE6kFuKvMY7ROWY+cF5U4TRnNAMb9mIpvfPYDAoegzZlpPea3ydAco3MYozfXNOVw30f8r44sFPjXFhkprl1IcuYUqIl9FvnshvokAMBtyfIWJdKurQB8rw8PCw44lx/N65Obgpm6YpFsD79+87oS0WDMocIUE5pJ+Laih2bPdnDl/xuePKDHuovgnNMmjNFpbdsJnGJA7WLDUEg5Mzh8NhWczT6bRYkfP5vIQOKfPEJe1GhU3TlDAAPOPnRvgAHqEhIJLntXLIz+T3WWhsAjJZ4EaslYvzHD52Xp7H5xk1j102xNzZnDw+tnWwgeEwgT2lmwy0Gl9Y2fAdis7XwTuDV9nnI0mXvbV8DpQsihoZ5SpSqp6oisLjxXrGy0EFq6u6CNFdX193qs/cMDEDQO7Jnl2+R9bk8Zgiz+dBjo/H446xmyMDpDusVqvSuJXr39zcxO7ubmlsyDnJJ4Vflstlp2DF8wp4OTg4KPLJTgBALPpruVyWvSM5zg1neWaOz7R4DJRnuVUD89DI4TbPwVOMXwjwU/s/o78MiGpI3ZNtt17tmn2KJ6N+/2Z7e7uzrULEOvENzwQCZjweF+CD8sRzwLkBHt5UEKuB480kq9V9Xx+7sakocJiHex2NRtE0967n5XJZhBsMlwGiX2vo+zHGrJ2L9y5l9MLM9La3xcDY4MIt81+8eFFc8QhdBBDeOTxyduMiiNjPy+We2etjIUqlmRPSERQ1cN3nFajxGTQHBObjvy14fx7f7agpzSzwzfN4WZhjeM/en4i1Fd8XCsjyLXsl8v3YK+yNes3reCIMlra3t2M2m3U2K23bbhk8f7TSoJCD/BKavrrDtTckzs/FdVD8NALM4a4spxw+8zrOBlV+3STbsq4xgCLX8NWrV6XizYYq72nRsbOzE/P5vHijkYfs6YW8ARga3BA649r5+TEQMZw5v4EzfHdzc1M2jbWcNGDZ5NHs85T1eXy8Fpxc73wznvWpZNsXD34YtYVqQmaF7cW5WnV3F4agGSHb61ADRp4gfpOPgWm94A20qOYC3HAsjIxLGQHhe/c+V1zXfSQASk3TlIoxABBM76aGWGy4UvmtY8OEvWydPjZP9tLlYcHbJ1BM22yF5dg9n/GcLGYEy3g8jpcvX0bTNHF+fh63t7cxnU6L1w0PEG5oV5JwL+PxuNDOAMi9NRwWQ5i7ci4LVPNI36gBcP/G35lmVg5PaSk9j283MoivgZVsnDGHEWtlkEFAzu/IMnLTWvNvfG+EvJCRfGbDIMuB6XTa2caCYygG4BoAH4fb8WJj3OGR/uEPf1jWFhVRgCtyUN6/f19CbIRvUNrOnQKgMZw3mRV4/myTcWLa9/2WPMvJZFLan0SsUxcI0x8fHxcaY+ySyoDxRQoDXjJkNb9DdiHLAZMugweUEpGwh8qhS3YHqPEcz8ZrlkPwV21kfs28C/jPm+RaLz/V+CTw0zTNcUT83Yj4lyKijYj/OCL+WUT8g4j4sxHx/0bEX2zb9qy5p8rvRMSfj4jLiPjLbdv+n5vObyFeE/oWJvaq1IRCtrLwqtSI6d9mCzxPrhneyjEf50oqK+22ve+zgEAh6RZ3p5Wp45+cE4ACmDs8PCzfU/VEPNgVEBHRaeJ3dHQUJycnnbCXwccmZZqFR9+oLYyalQUt8kLjt3zPHPI9rnIDgZ2dnZhMJsXjRbO1ra2tYiniHncDNFsZGUByD9DHoQzuBTd0vp/MU5toWRMMmS/Nhx4G/8/j+x85PGxw48/8B38j38x3NfnEeSI2e5vy+ywzASwMt8tww9W2XW8m7NJpyykUMhWm9OqyzEQ2OfSCgsYIw6hEWSMb8Vaz5xWyPW8+HLE2knKfIs5bAz01mZ/fZ2PNw+Dj1atXxUOG4WQAx3EvX76M169fR0S3qeHt7f1+XPaQIaOQWfAJNMBTxHd4hPg9c+f2HQ59jUajMtcuLvGz14bpUeNBgxl/j77D+5mPfap8n4hP9/z8TkT8L23b/oWmafYiYhgR/3lE/K9t2/7tpmn+ZkT8zYj4GxHxb0fEv/Dh71+PiP/6w+vG4Yd2TkOOndubU1vgBj64S1nUNcv8MYbnGB+fvTIIq62trZhMJtG2bYmHI3BgSIMM7hMhZG+Nk4Gdr4NbnGs6kY178bXt1m7btuxV5YXgiqYaOHlMcDw2+ixT/t8ETiMeVoZBN0AMCcy7u7txdHRUKrnIAZrNZkU4n5yclP14yCPwNbguwgBwyufmUxQEVmiNXzKP5u/yZ+Z5jsmgy2DR58/nfR7f3+gDtLz3nCL0UQT2Vudz5b8sg/IaZWQgFdHdVd0eYtYCBQTT6bTcJ+EwlCnyiBJ4lDiyijULMMIgGQ6HcXh4WAwH5+Dx7Hh0Dw8P46c//WnxiIzH43I9y1VogDwzqLIMyYCnRjMDxmwYZ0Dkaw8Ggzg+Po5Xr17Fn/zJn8TW1lantH06ncaLFy9KQ8fZbBZv3rwp8thdtn0dEtAJ9Tu5nNfhcFgMbKcM0OQV/UJ4C48d+6l5rrI+NPDr4/VN/JrBDPcDEM48+pTy61uDn6ZpjiLi34yIvxwR0bbtTUTcNE3zmxHx6x8O+/sR8ftxD35+MyL+2/b+7v+PpmmOm6b5um3bn266jh/YwMDfY70b/NSEjC3hmoVQY+hN95SBkt2+fO6qK/85yZlj6bvjRLKcbGwByGabMExWwm5KNp/PYz6fF2CwWCzKwrq9vY3RaBSj0ahc0yDIQCg/8yY69X2fLSXT3ACwpsQNimzJ2VtjwW2LleTv2WxW9snhHG/evCkgEGDoRm6eb29pYfBsYYpQY75Mtywkeb78fxY4tXBJzR1sMGTe6RNSz+O7HTWwk+fRRh0gAgDh6h9XXxoo+bV2/YiHoYqaQnM5MwDBhRDImsvLy4432omv9iawTr3FT9M0xevB7yi+4DvfI95TzkvIerlcxvn5eWkx4X5ngB8bMcgKjoM2m0BMHlkW1IzkLKd4pslkEj/+8Y87vXoWi0Upeee5mqaJo6OjuL29jYuLiwJg2rbteMfcc4mEeOf6NM19Pufd3V0BMQBSDGDn8qBHm2a9l5jDjBHdTXb9rJnXajyY/7KzAB4nIsGoeYCeYnyK5+fPRcTPI+K/aZrmX46IfxIRfz0ifiBA86cR8YMP738UEX+o3//Rh8864Kdpmt+OiN+OiDg5OSkPbMWbQRCWB4KilhRVs7TIbcnM3Gedb8p5MQMhrED8eHFYBGyDYcCGQPEO66B6t4qH8V2OTtw8IkpfCY7FakDhZ3CWLTw/j2PFgJ/8zLX3mz7zd75HM7cFX+03tWsb9HpBQ5+IiNFoVHpu8GzkC7RtW6pVjo+PO91Zm2bdth8a41nro6mVVs5ZsqJ4DHhnUJMHQsRhC498/RqAeh5POyzDjo+PHxhgnoM+heBNdwkDoKzgdXtjGN8E6NaAd94DC8+ocyZZJxgJ9jxjjBGmcl5lriqiOpO1xB+5e2y5kMuzydNDSU4mk3j16lWpdrIRwm9yGL8GBLOnOdOvBpCy4vcf9+FGp8PhMH7wgx/ET3/60wLEKIhhr0C3ACA5uW3bzvYV0NXFLMwBPOKWAwAoaAl/bG1tlQpYdJFDYldXV2X7I543p5Vkz0yfLMprgD9HdFzp6Dn5LkJfnwJ+diLiX4mIv9a27R80TfM7cR/iKqNt27Zpmm8kadu2/d2I+N2IiF/5lV9pvcCzVWyBgevOihSUWssb4jyuouhz3dWEBJ/znfNiOKd3xo1YJ995B3Xye6jysrVCMpybUG1tbZWeDRyHS7Vp7svnI6KAJd5Dj+vr65hMJp2QDZUDi8UiDg4OOrlEFiY113CmW5/A7fNw2LNk8Oi5ql3Tn0MLg2PmyL2W7u7uYjKZxGw2i8ViUTxd9AOKuM+DOj8/j9VqVXY69nNjObl7qvOGzG8s5kyXTcCuBnbyos/KE5pRNZiFOO+fc4A+z7AM+/GPf9ya/nle/JktXCd8AiaQAa7CyYZCVuSWVQbztfVk8EVoq23XuR7knLj02ErKwB7Pgo0P59AhI/HYIOedrIxn3gYacuvq6irevn0bd3d3pau7dQV0yjkuluf22Na8NZvWStYFpn8eAELo++rVq+Lx4R7b9j6UTjPEyWQS+/v7MRgMSkVbBi3mIUBh9qIwb+zoTl4UABeaHRwcFJAF7d1sMYdbPUgjqcm1fHw2APwd4Jj7Nt96bTzV+BTw80cR8Udt2/7Bh///YdyDn9fNh3BW0zRfR8TPPnz/xxHxK/r9jz98tnH4YT2xZjZ7RjICrYXBzNx2vWVm/ybM79eI6LinDWi4Hq5ErKiIdQLvarUqJZH8nzf2BIFDD+eacE0z9/7+fjnn1dVVERhUBRASGgwGnQ08a8KhDwDV/rIF6jn0/wALM3ue+zxHCDb37IhYbwOSK9awUI+OjoowaZr7TWE9n+/fvy/7grkqBIDBdbg/rFXi8rlcN/MJ39X+Mn9lIdH3OyuZ2vHQ63l8/rEJvPoY5pCQKTwXEaUa0cZPjV82/Z/lmcEI14LH7fXh/inBNvDP13IYK2K9hQyGHjlzHLezc7+nINvKcI9sd2HATm4LIOH169dxfX1dQtXZgGjbthR7+Hoe2TNqemTvkM9fy0cxWAX4cSyGCU0Nf/jDH8ZPfvKTuLu7Kz3ByDV8//59vHv3ruQLUh1MpRjHkCflfmX0OyLMNpvNynPTAZ+yep4JWtOhm+eKiPIcWSdYThuIma6mqQFM7T00g7/gBVft1eTkp4xvDX7atv3Tpmn+sGmaf7Ft238WEb8REf/3h7+/FBF/+8PrP/rwk38cEf9J0zS/F/eJzhftR+T7mPmzEIm4JwohHUoEjRJzzkN+7VM6m7wbtXP4expx2QryonReC0jXHiKsnIODg07CIEoeRI7iJy58eHhYugtHrBtkUaoNg+NJQmBQKZErkzJQyYLhMcDYByo3WVYwvHOdvFjsBuXesDg9FwgcnrH9YMEeHBzEyclJ3N3dxdnZWXEZ0yPDFo6T/ra37/siubW/5xR3MQKPZ8mKjufui1/7GUx3g8WawHV3b//G9/mUguN5fPzIYN7z4M/YsJiQknNuPJfkx9Q8SbVr962z7LWw0nFoNiLKNhT2+thbi8EBsGH9IgvxOrhsPWJdDYUx6DCO17w7R3OvbCZ8dHTUeVZkiLehsQ7x+/xdBkNe45l+mc5Zv9Avx0CR+zk5OYmvvvoq3rx5U/KfhsNh6St2eXkZs9msgN3JZBKDwSCm02mhHzINHUmLAjdihd7X19dxdXVV9pwklIXx6zBazh9113o/ZwaaWSZlD1U+jj/0G/LWYXzrcv89xfjUaq+/FhH/fXNf6fXPI+KvRMRWRPwPTdP8VkT8fxHxFz8c+z/HfZn7T+K+1P2vfMwFTFB7ahjks1gp1lBinjSfv5ZZzuhT7nlB8DsYzxn0BhQIDHZLHo1GsVgsyvfO5eF/wiyLxaKcCyFlcEgPofPz87Iw+MOKwA3qHkN4lUynDDyzp8Y0MJDhObKQMZ0ywMoAk88AiRkA+/qmR9uuK2MAlG27Lt3lXNvb22XPnIuLi9je3o7FYlFCC35G6E4lnHOj7KnCRR8R5XqbgEYN5GTa5rno42meC8+h+a4PvD+PzzOywZYtYn9OfxYUgI2k1WpVPC9t28bx8fGDa/RdP6Lu9cvKy6FbFCC5I/TNapqmhHEMvG10WEmx1xbeJJqMkgtpbxbGHQob0IRHAA85wGw8HsdkMinJvFS/MryWbQTUUikc7uqjYZZBpl8NGPBsyBXowH396Ec/Kjk9hDSHw2GcnJyUbszz+byc6+joKA4ODuL8/Lz8BhpyHtOhaZoSNgNAE7rE4wYAZN6hv1vBLBaL0mGbZ3SoK/OhZVsfb/pz9zty48c8b9btTzE+Cfy0bft/RcS/WvnqNyrHthHxV7/pNYwcLSx4T98WlDufu5wyI/IaoPJCjOh6fmo5Rzmm7vCLw1c+HhcjiXosYu89ZfcsXUuhA6ErlPD79+8LgrdnAo8PfX5IdsNtzD3v7u7GyclJSfg1UIvoCmoWGgu5Rqc+YJhpl5W4j6ktpto18PY4DGehixvVFgTCHHrTUh6ghHDNQhR6kzuFsGXOXVoLnfPz1ixFXmvWZfYMZWunBo4cNvBxNcD6PD7f6AtzeX5R8K4SzAD57u6+/wqVmm6j8NjIPJC9GzYM3FMnIory417xTPkYKycAUkSUcm6DDstlKlx3d3djNBqV5G7u6fLysiRAR9zLR28Ienh4WBKBWXeuaFqtVp1cOCvkDP54nppRltdrbS79e1/f3gteAXanp6fx85//vNAMY5R7xqODV3l/fz+Ojo5KGw+u7fngOu6Vk++JaAH6gfvGowbNmberq6vSWNG8nD1lmSaZP2oGHOkC6HKfw57Ip/T6RHzhHZ4z09kigRDz+bxj7Wf3mJVDzv+JiM6icpVYTWHUJtTHO5HZoA2vAOgWIcBnhKMARi4tdXiKsBSWFovEjbTatn1gEXkvHq65Wq1K3k9OBsw5LSh458/0CdM+D0MWGjWroOalyN4gf5avabDJ99CfzxAq5ENBd1zDs9ms0AbPG0mZ8NH19XUBObZQ4MHBYFDc3dkjlulhL1gNpGc+q4Eegz2XF2cr7Hl8P6NP6Pv7q6urkl+RG7w5nIQn8vXr1/Grv/qrVSMin998yHeG3dADAAAgAElEQVQ2CFAu5lXCIHgT7FnNJcoYQ5YD9KtxhRp8SgUSCtzXAthQmbRcLktYiy7X3DvFGcgs5KV7A9kQMT1MiwwE87F+1k3GhUGTFbfpY9nOfe/t7cX5+XlJc4DGpD5QGk+OJhW98/m803CSPwxoyz3Lde57Op2WTZyhG/IEuiLf7M3OXsxNsr3G9+ZN8pCWy2XMZrNOyC1iXc1soPelhL2+08FiM+P4we/u7oqy4liXZBoEeUJyhQLnsiehNqHZs5GFjtuX21UHaqc9e8Q6vm7Q5gZTjnsSD7+7u+vE0EejUWnshcKm/wxAihwf3JRk7wPELOhYBF6wPGfbrqstsjA1fWqWE9/7fbaUGBb0PjaDCINXC0SeIQsCW1J2PbdtW9y5Ed32/uQGUWVHDsZqtXoAUDMAxN2crSLzTAba/t/egpoQyQLciskhYPMhBsDz+H6GQYfnE68P1m9ONHZYE0Awn89LJajPXwO7NeDrXMiI6FwTRUm4Cx43KMNYtAxlTbFWMBqQyfP5vOSqAOQ4d0SUHdpZq3ghqFQ6ODgoSb6EYXx9jDOuuSmRuWZU2Ztf8/zUDJlstHjOoIMH9EMGkZP57t27+NnPfhaTySQODw+L5x8QR2oHdCDJ2zmjyDb0oAtezAtOXp7P53F8fFxkhfUf+oi5v729LQnbfXz1mJGVZRl5Pm/fvi15bgbmgB48Ww6Jfer44sGPczcy+iMxzKEHVyxkRrRCiFgzewZKrpqpjZqVziI28o+Ich/L5bKEu0Dr7vNDAqN3Kea5HZLb3d2Nr776KobDYXHvAkqo1BoMBrFYLEpcGHrc3NzEfD4vydJ0FeXeORcL1MLM1kvN0unzxvh9Bj28ZkFsL51d5RmcMligzFtWCLX7pcePqwlypQG/MdjhO3aBp8qC+7IiMBDL95y9Pixovzdv1oCSacw9ukGcLftN/Pw8vtuRLV9/Dt8Z/NSsZBss8ClgxKPm4fHnmWcyQGDdU01kA8Ke34jorB1+520XWA+ETfIawmhD6XG+m5ubuLq6KuApIor3yOEtgx7LKZ4z0wK50me85TWVv+/7Td/wM2Z6W+7f3t7G6elpvH37tnSdp9s81WAkiNsjhNd6Op2WufEzIi/5TeYrZM3l5WXZWyxX3DLfGO80U8zGXgaLHOPrZVrhBVwsFnF+fl5yU2v6lehHzvn9lPHFgx+7wewKpiU4i8QeFKyKDG6YIAMhezesRPsWQW0SI7qhmPzeoTkn1RqtIwSYXBC+LS0WE1UB+dog9MvLy5IcDcAaj8fFZc6CvLq6isViEaenp0X4uFze584AxszvY2sCxMrf77OXhs+yB6QmuHNOFd4we2N8bZ4ZJRKxzrVq2/sQ4uHh4YPQ6Wg0Kp9ZWdkLl3PO+MP1z+c1XjJd8ne8GgRny4mBhwrwbEvOoblnAPR5h+VPxMPcN5JRXULeNOv+UfwP2AGQmI8zoMpznNdPNlC4hr2teKByFaW9Ccgk8kYyn2HsuVKMa7oxInLHIbXFYhER0enojOyjH5nzLA2GkBuW71mH9K23Gmjq8/jw6mfPv7NRzHxFrA02KvfYf/Ds7Cxubm46Jer0+1mtVgXo0DYA2eQcIIxhG5DQgvwoeAwevLy8jMlk0pGbeNXRT+4dZ/mSaVczuLLMatu2AJ+Li4tYLpedUKXpCK8zr78U4Ceim+hp5XV1dRUXFxcPqnv4q8XM+5je1pXDZLVRm1xb8UbOTdMU17GBDc/kOLitJwsV9zCitJPQDPF0fof3iEZZTdPEcDiMnZ2dEk+F+SPuF+Dl5WXHMqEUEg9CHx387Pn9plHzAvH7/MoCNR9YCHvucYXbAwSYMuiFfwyWsEhHo1Hs7+/HbDYrjQ4BOQgCh0ftuscas8KygPHzWgnVhKZfefYcBqvNBTxGwrv5/nl8f6NP5gAQAD6ExeFVKwHPLcOenzzP2dDoGxmMcS3nHkWsPTkGSavVqoCQGm8bRLHmHEaxp9b7LJo+VKJSbfn69esi17gX1ibXR+H7fiwTsuGbPaSP0apveD3bE4XS5nN/5/Ysx8fHRU4jXxaLRQyHwxgOh/Hu3bsS0gLcnJ6exmAwiDdv3nS2HEHu4K3hmQGa0NmtB3Z2dkqzXJ7HBqV7A5kmNdlvo64GgAjfvn37tlQ740nPW05FREn6rnk7v+34osEPqBQCoHguLy+L1wLGsdfH1rtDJTBnVj5GsA6zcA8MgxwLjezt4XuACxPmuL2TACOiY/VgLWA57e/vlwaElJ4S0rq6uipt4HFdGgyg2LE6JpNJ+T3C1vt8OWfE77NgzYmEfe/9f00Qe0EYfPo47sFAxrFhzym/Y+Hyve8bMINbl/ckf5N3AH0Ag/bc5dAo+RfkD1nA+Tn9TAho85ZpZ3r4GWq0g4fc5DILpY8FqM/jaYeBiOfS3hzy8TBoKG7AemduAQ+c1zyUr9k3cg4Iaw+eJS/QRmE20Nq2LSEK1hG8RVifcBdAh9w51uHW1n1vMuQTXp6IKEUFk8kkxuNx7O/vx+XlZUlz8ObD9orbuLS8sHzO3uHaqHmbM81qBrB1BNfznOe8KIwWcqRevXoVf/qnf1rCm8gl6zPA0fv372M8HpdNXaE9RqxlpO8bHouIAnq2traKTiBvlIo9y1v6EZkWNXlvGtRA0NXVVbx58yYuLi4KDwOIvFbgVarAfmnCXhHdPjnEHcn1YeLMEE70QlBwDhMz4mHyFQuHHJmI7mKxEOP3EQ/dxywqu3uxdjIIQyigtLgHGB/GpkHYdDotwsnXc6WAFR8eHPIJSKRjgZD7wx5E3F9G7vlZGY+FUh5TtjXPRPaS5CQ3QJppuOk6FrgsNJL5ELqAw52d9Qaol5eX5Rjfj71GfgYqMpz3g3Xl563da83zFfGwm3kfz6JsEKS2rHkPgHsen3dkwc984WFh/bPR7tHRUclxiFjn+yATUEp8VwuHPmaMmJ8NfjDYbEy60AS5anlTy/3xPXNPrK3d3d3SYZ0w+/v37ztepP39/dK6Y2dnpzTp8/NxXV6RoTbkagCpltrwMZ6fGg1Nayt9g0X/Bto4P4oKXHJ9hsNhvH37tgPyZrNZAQYYY8gc8jytZwCgeBUJeXmeKKQxTy4Wi+JpM/9xnwBYeBldZgCYeS7zJh6td+/edeRR27aFLzgvFYNUHbqZ5qeOLxr8QCgUDgsAr48T6FBivLfFjAejJiS4jr0HLHDcbzVhkpnax+zu7pbsdXsgXC3ka+ONsLu5ae539qVcnfuCqQEFxIwJhRHywj1O+Cvi4a7z0IuKORS6t3TgWsyFqxcyut9EI46pjXxsDQBw33a9ej698LLQcViPY4mle/EBjnd2dmI8Hnf6Ttzc3BTBbUuF++d+5/N5nJyclAUcEQ8Adxa0fh7+h95cow8AZe+Q9ynLAPYZ+Hw/I89dxLr7L8Dn6uoqZrNZUTyACuYN+RfRLVv2vPZ5KWrvbegA/PH6EF7LcopqH8qt+Z3zggAzyLTRaFSeE4VvAyMiOs+FF+zo6KizTgCJrFHns6D0TV/nF+VQtCMBjynRmhzrO87Hc27LqzwXvg+eaX9/P05OTuLs7KzM68XFRbTtuioVDxhheqIAFLzgUfR+YOhBh+YBqshDQBKhNkcmuD/npfUZch5Zlkfc8/67d+86IVy+f//+fQF0yDT63fken2J88eCHBeY+GNfX12UXWo6xtwdh4Tiy48me/JpSiViX1OXE2U3DsVFXY2QwFrG25kjYpsU5DAnwsdUF4MsIHKHCjuWOs19eXsZ4PC6WGkm67vjJsblnhl9tKTnklS3Ix2hU+13+s4cmHwftDGyhWfbGQBt7iTyXGdwyT8zF0dFRKTF1boE7qFq4tG1bAPrJyUmHTjVL8THa+Dnzb2rWFAqSkIkVQAZTz+PzjixnLMvYVHg+n5cSZuYT/gUkebNhz23NoLNy6wNG8AmKzm02LB+wwllfllVO6CXH0V4cFDO8d3BwUEDSzc1Np+pya2ur9B/jHlerVSfHEQAWsZajGC2bEnDtLcoeinxc7ZiPHTVQ6WHdAl295geDQRweHhYvDHQFdOKZu76+jq+++qrIHvQcRtre3l7s7e2V9BDoY4cBgNWyExBMGoDvLcsonrcGEPs+I6E78yzDYdfpdFp6PyGXN4Urv8n4osFPRHTcd3YTE/u1UM9WsP9s/daAD69WTng6ssLKE88rTE9fCnsn+B05JXhxItZbYbRtG6PRKA4ODjohiqZpOhud4hInlDEYDIrnCEYZj8elbwbH2FpjEeAWtZJ10nYNfGRm/aZIvOYRymAiu1OzJ8dJlzmmbZrn3CQ/z9bWunQXyxT+8vxSeYLlg/Dm/BbE5KSRf7UJFObP/D/n8rNkz1ANsF9dXXW8hX35a8/j840sZ1j7DuO7eIO5B9jbE5OVtL1/fQAor2OGjYmIKEqV33qfOpei8z28hRxxjyKMv6ZpOmG6XHXJWiPRleMx6PAGeY8stzHx2rJHzHLMa8FGXAY6NQCUR5ZF+fc2uqiSsiyysQJwdNUVMv7o6CjOz8/LMwMS8Srj5cUIpvuzc2JMR1cJmo+YS3u2I+717sXFRZycnJTUCVer1fioD+zkP5oz1o4z4CYP6fz8vNzbeDzupBF8yviiwY8ZAvBDrxrX/ecqLwuPGhDKLmMrqBooyolznmTn18AgTBxMGxEdIZDdnSxyGhbm0BL3SK8Hnw/GZFfk5XJZ4sYcRydUPE2gfs4HXXGregHVFndW6DXBWvs/z63pZ+HtYwyG/Op5yODGVmlOOq55lfzeApeqCqwm5tdbfAA4AOJc6+bmJkajUZUO5q8aT0X07+aewXSmGcrUis3dqZ/H5x82glBarD/yFy8vL8vay/ILKx+esxLP3p983T7g77UEH9pjDRBr27YDXvi9r08aAuccj8dFZrvCy81EMW4iohRzcI/I59VqvQEqFWGstSxDs5eaa3C//G9jIINJr8c+kJPfbzIokO0u2olYA0DfV9PcV/fSrmI0GsVoNCp8wbFXV1fFq4PcRjbZ2HLIanv7flNZgw4DWWiGR5tngkedG9THZ5keNdnPbzEmOa7GpzSzxEggl/Hw8PDJjLhfCPDDJGNduBzQgh1B4SRW59nkvwx2Ih42/sL7k63nPpeoES3WEgsasIbgAPE7vskzOIYNHdq27fTZANlPp9PyWUQUS4xqJQQugIy8KY5FyBweHnYWhXNacr+azKwf4/3JAiQLl7zAaiArLxaHNB3icpgKwcDvM4CAjzzfo9GoLEJCEfAMPGkh5nltmvsS38lkUi3b7KMBow/41Cz7HE4FADNf8N0z8Pl+Rk2+IM+Wy2UJVbO+kV8ujvBGp/BXXm+blFNEPd/H64rrG3g1TfOgU7LDNRH3IQoScQEl3IO938gSDALAnMPxfjYDxYiI4XAYBwcHnXYdvqY9vDnczDl5Fq+7bIiYRn41DWv0qxknNmydvmDw6M+QQRi1p6en8fr163J9AA/hL87Lb+kI7fQLh+kBU3mLDV+buWb+AT9U9iHr+uR9zdCzPCPUm+mZXwH85p3ZbBZnZ2fx4sWL6rW/6fiiwQ+xXtyeFhgmvgFC9gQBkgwmsos4Ez9/j2DIVoavH7GO5VoAeIM4LwrOR2w7Lw5fG9Bij0bbtiX0583g3r9/X+KkgKUsPJ1P4HwjJ1lyn47pWxBYofcBwTyygMnCzufoA1n5XLb4GDUQ5N9xPe+LE7H2+kAr6MN9shg5rgbeXF56c3PzwKLNYC7Th4VfC2WYf3Mo19/T+sD3WvOsPY/PM0x3gDO5YezbZNmD14Rhb0f2XnD+2vVssJnv7Nngc2+9E7EO56JEWWeWBS7Nh2eRZ16/9rijaAkfGyiREsDvItadfblPtn6IWFdmYgRaNvPnIg7fe/Y29K3JTfInvzc9TWsbVwaZln824ujWP5lM4uzsrGPMYFhRjQXNPNcYZm3bdlpzEE4jNxBa1ICMZdFsNisJ6vac14y3TJdMG0KrfUDdnzslA7k8nU47BuWnjC8e/LDAiHk66RdBwSTVgE9uTGdLmMUBM2Zh4fePERshwXv/fnd3t7PnF0opYg16LACsKPljQdgT5s6wrvgC2ecFz7W2trZKnhANsHAtujMoAMAbBOaFvekznj97cPKiMfhhPmtgq2ZV+PfQEQGSEzctDHlGAyBfg7kgx4B7g4d8v6aNEypZuNn71zdqQMchyBrQqQGixWIR4/G4KK+I7iaHzwDo8w7PEevMXh+HYfH64PFgzeaQl3kwX4vXmnLKCo/7IqfIoS6XZUesPRm5HD4iSiVYxLp6i3XHc3n3dYMBnoXz2ROeQVTEfZiMHCG8ILW15fAg/9c8YA6PbVqfte+ch+f79T34fQY+5AV5PqjY2t/fj/F4HNPptDMXeHcPDw/LPRm4wmfIHRt1EdHZKzJ3ErenBb4ll5S8UbxCADY/c80bZllmOcRxfvXnBkvwDWDsKcYXDX7u7u5KVZfzfSzsWWDeIoLEVBaqk+xgNL8iEHLIK6Lfm5EXTURUwQ8Mj+cmIorF40RnW1ooPDwNEd29pxaLRUHvOzs7sVgsYn9/vzAKVtXW1lYBQ/RE8pYLjrvD5PamUJaavV19YCcDpAx4Mv1sJdkDZLBpC5X75N6hVW3BeTF74fDbiCj8wW8JXSGYsa5ubm5iPB4XHgRMOxR4d3f3YF8tV9H1WZRZcPhZHc7y89W8Qh54/5x0/ez5+X6HDQxC0uyz53l1zg3Wb87p8Drj3H3e7Ii6hW7w5Aovg/tcHBFxX3FKn7AsK9wwNIN37jXvtO5mhW27bpxIWJ7zHhwclDJt9+rx/WFAZi9QjV6+J+ff1ORVzfDKc8sc8Ftfy/dnw8tyKXvcyf+hvxv9kgCO/E9lINckbE8okSgJOpD7YGsMyz17ppEZri6+vr6OyWTyYE430cdyGx7znGQa5t+yFlyZa4PwU8YXD36m02mxorGWsOpNdCYwJ+0Z+DickRUsI39WQ7W1YWGRfxcRHasH4cBzbW1tld3XATkIyq2trcLoCB6aExLS4lkIDQJa8ERw/O7ubtkMDwanKZ+fzcIsV0zYw9FHq2x91ka+no83COLzbGHk8FO2EBEU9vpxHvdfAjRT6bVarUo5LcmHCI2jo6PS+wd6mKf4jOs79GXrMwvJ/Fx+phr48fsMiODzi4uLMrfQ4Rn8fH8je33m8/mDcl8Le3t/GPaa1LwUrJ/H1l3+zuCHqtG8ppqmKTvJu9cYnmHk7OXlZadEnzWBAuM5yPGJuPeyYpyxDpFX8LC35TFYilhvZJxBTA4PmgZ+zTSsrZEaTWtrMV+DwRq0Nwr6Md88J/RFXk8mk3jz5k25BwxSqgaZ893d3SKbmqYpdCRvh/vCm40s8ma6lkmeR7xNmcf6aFijj/Wev++bA4MdA7inkmFfNPhp2/s6/729vbi7uysVTXa5sWhrCiBXgdki9wIy4v4mhDWj1NA/CoeJGwwGZe8sWx/2sLAQEDScA6HkmDiLnb1fVqtV2euFvXBwcfqVDTzv7u6KcLF3gj4a2bJ0SCyXPdaEDL/92GFgYEDmuTOIqAkdFsumc3Aeh8L4nlDVcDjs9PNxcy28bNmbGLFOmDbdssCoCdhslWZLHtr4z8+T+Z9cEisccuKeAdDnHQY+eK+n02mnlNzzjlIjBGTvqD0bHrV10PfegJt7i4jO+V0VhOx1kUTbth3gExEd7wTH4SHd39+PyWQSe3t7pWebw2IR6/4/bbsOu2UDyGE47tkeDO6ZfKCcdJ2NuD4jONOrNmo07gOf1gO+lgFPjjyQaHx4eFiaHnIdZJPlFnrEn3NcRHSMWcbOzk4Mh8MHeWems/WsN3H2cxgIZYOOufEc+bts1HnwHLn68SnGFw9+ZrNZ8VZcXl4+8Ax4YnjPxo5ZWVhJcP6MRjOa3WQl1NzJBjx8j9UCgCAcZRfu3d1dqcKiggKQNhgMigDBi2CPTNPcJwIeHBzE8fFxp0KDvXPwQiCMtra2Sonpzs5OTCaTUlFh4eLwoF2stqpqtHtMwWahn+nKn0OJ+XcGmtwr92fPYAYeFiLMlcGVu9ECTn1Nuo3y2+ypMhhEwNvyyoCZ4TDox1iZBkj5PY1AUZ7OgXoen3egnGhmeHFxUXIA83zyRxKrQY89jDnEnL3XmX/8P6CC9WXPqN8jB8i7tEyzImfdmAe5552dnTg4OCj7c5HobT5kjY3H4+LpdqEK68lGDesBj8Sm0Lzpko2r7MXo+12fBykr7Zrcy8aP8zDt4eJ5eF4a0h4fH8e7d+86RSw23mzsR0TxZnM+2pxA5wx+t7buc0Cdq5VBMvxAFWvWmfl5/d5yLsts06fvt/zOOvwpxhcPfugfgRJxgrInncViCzmDo0zMGsNmBJsVel5gtXvGAnJjMDN6RJTeGShpQlZsRzGZTEqzQ8ARw5YVz3pyclJKQfnu9va27H4csU50M7I/OjoqgsnWVt+Ct8L39/nYPgvI58k0z2DFw2DC3hYnw9WALsIGWnsuoanj53av2oVssNM0TSeWjgXvPYV4Lu7Fz2oa+H7zyLyY6Vfz+sBPq9UqLi4uSjiVar6nEhzP4+MH+XlXV1dxcXFRtpOpgR6AUtM0Zb0a/GQl36csajkVEfXWHBmQwy9bW1ud/B7uj3vxb7h/GwooXoez8Hxn3nYvMow+5+BZ/ltJ5/M8BlIyPWo08W/8202GnWWX56LP0M66BQ8QBio0fv/+fiPRwWAQs9msNCMEQOFJQwfi8acvDucmUdneRD8LuVcA09wWwNVjDsdmnZhpmIfBT40/DZaycWgnx1OMLx78kGluiyKHBLwwUFwsIiuFDIAiHoYbmJS+CfXvapMXEZ1F6yQvPBIwLcfe3d0VRXtyclLybexeBNTkFu4wObv6uhPs3t5eDIfDTg4QFgOLjZ3iDVRscfGsOdegDyRuskYz/Ri26voUvn9vy8nzaSvIrnwrDHs/UDTMAwI7C1uuxyteRf4/ODgopaMAJQsO5rwmID36FFRNyeQ1kAFQRJTmeTSIe/b8fP7Rtm1pZrhYLOL8/PxBhaEFOgrGjQVt7GS5tMnqro0c2gfkmHcw0uinxn068TrLNOSJ5d3+/n5Rqs458xogMdeGmQEUr4B3ZBfnNR0tt2tyfJM83/R933qsHZfBUp/RAqDgGAw0V8FBCzobv3nzphi1PBddsX2f0MIJ4hFRWsZERNm7y8nWOzs7DzxA3Bvyg/Ct52aTgWY6cb78eXZC+Pv8e+jyFOOLBj//f3tfG2vbepX1vPv74/Sc0xZ7c3t7EYjEhBgilWAJxhBAqI2x/ECFGL1iSRONUfGHaeOPxo8YMESUxACNoIUoBSqRBjW1Fox/tAKi0PLhvZR628Nt7+c5+5y999qf0x9rPXM/61ljvHPu77XPnSNZWWvN+c73c7xjPGO8430ncHLKJyekCn8NZiVTEDFHsT6uMJR8AvEa0G/ScGA9cp/5qPJhkBu9F3R76nt7GIvDuqvlR1DFCUSvDeN2qIBXVlbawEJ9Lw+f0xgeBR8qQHhNBVMkhJmuy9rsA4hqeUTAiyCGaTnWek/BB+9rDAyf5xiqYtL6sA94VokuL9JqYh9qwGDGN2xj1BeRJanCU3/rt24vfvDgQXsWCpctuxTkQBdHTXPyvrfXXnutja2IxpL/ucMGmJ07GXjuO/c4DygjWa4utwBogYYHynN+UL7wRaQERIw90dgW8qLG5Pg5Y/T66NZ+ysuDg4P2AD+PAwKmQSGVtcb6uJc6AiV9PUYRKXD1Z7XPfS7rQZZalsZ/cjv/G97whva1NexrjWsajUYopUwZO1wOVP3B8IpSxqES3FZPntDDE3WZUU+q1pACB7NKrmv7pvWPkhq256W5Bj+clBqDEVm+eo0TkeDHLStlVB8MR+1MR4omgzI1iZ4AnewkPd+H0fy6fq5r+6o0OYk1ncYE8RAqThYyLAGUvxNHBYS2QQGSCgRONgcFLnAzxvbfXdYYBZeTu/x9nBQE6HixrQqI+SwFPctVMMg6qDLicxxLptnc3Gw9QBQoGpiYtTkDOSpgtP/840u8rP/R0fhMjDe84Q1tvIUGqg90+dQ0TXsq/f3798OlSuVd91BnvOIf3iOpQo3SAidnP6mSo3JRkKweT8pXymLdNcQ4JXojKLP9gD/g5BRjntPD+2w7436YNmqPLsepR8E93RyHCAhGv72/s3GIgJGDgoj0WddDKoe5YYVLX7dv357iEeZFo1fBooIqL4deRZ7d431Cb9NoNGrjFV3m8NymqN+8Xx08u3PA+68GfpqmeX2AH+BkWej4+HhGgbsXSLe1K5Mo8HGhEllIpGhSZINM0vMUyIQUDABaRE7Qsra21j7HbxdirKfGo+i6MJeuKCTIXPoCON+yrq5lnXAK1LyN+jz7M4otyASDT3q97u10JaATgYBO661CnOloqfA3eccFjAMi1kUBEIUMPXi0lKkE+Ozm5ubUklk0gTOhqP3sQjvqB/1WT6jy/mg0wsOHD1vww/imga6Gjo/HAfIvvfTS1HJXZsC5p0VJd3q5d5ZUU7pKagio1/fwcPw+JwJ33V3K3+R5viyTIIYeBoIe1plzjN9Mo7F4bP/S0lL7Cgb2H9tFHaDLbiq/OXdVPvA5lxn+rcvYUV/x2xW5y7HoWjRf+e2GjgaRs04bGxsYjUa4c+dO2zcEIE3TtMtbPP9HATX7SknPNmuapj3wln3FMeNb3fVoGR79oUuzbijX5IvrEB1bl2/ajsuguQY/btXSMlDBwWUvBzxuCSsDuCVNqlnnXYPQNM3U5NFJpTsqtre3sbCwgM3NzXbdmopVy2SwIPNSC8eFIL0LumzGdNnuLG+PepwcBAAnr37w5yNh2/eakgsyFXAkTnDNj22lcNO2q4DwJVL3kLBcWk4ESso/LNsBoQJGF76aXtvq1GUJaRoHQzonHPQfHV+/WM4AACAASURBVI3P/Ll169YAfq6Bjo+P8corr2B3d7f9HwH7yKIHTsafCsmDnnUOdHkwNN6HQEeXz0uZfoULDS1VqFovnUc0LmhQaB3UA655k/QlwlovBS+Uh2wDPwRVlIkuR13WZYrajY2IIiPG+93vZ89F+erYc7zUu3/nzh08ePCgPfuuacZLqjwigO3a2dmZit2i3iRY4bEF9Nxwd7TuBCNgpGdof39/Sie4HPHYKv/t4FR5MTKIHQh5n18EzTX4ATDlpqc7lpNBl7iUUVyYuKCpoXNSTRnpMy4MlDS2hxOfy1O6pVyFBpWxBsNp/rrTAjjx1tAqUgtC68P/DhrYlsjbpPei9un9muCNrrlgiBQB68Ly2QaOaWThRfVQr4jGOPGYfl+G0Gd00lJI+7izDboMRmCrE12tMBe6NXJhTtL+chCkyoEnCXNH4QB+ro649KjjEwHWzOtDvmF8nhsx/I6uOW8pHwKYKpNKUGN+eLIyDTgaWLocT9mlxpN6LciDunTFOaSbUwC0O7nUY0uDhFvmSR7sTHCkck4pUrDaX5FsyyiSXX7P06vc1TnL59XA1TozP57Hc/fu3XblAED7vkuVN9QjGuepPKLeJT2biWebkSfUsOaH9dddadFmlagvWAcFap7ODbqsTy+C5hr8OJMQ8GRn+PjHrWZXWkAeSJtNkgi1a74OxnSHwvLyMjY3N9sAP/UckBnUk6FCSoWgThJ1H0egRS0s/U8PB4UQ76sVqN4e3nfA0eXdiPoxExY63hRoLhxZd/YxnyPI9MMXGfDNXQokxgWwH3RHBM9kATC1/V/7TJdfFWTxHvvPrdwaubUTCYUI8GhfRUspPDRTD9gc6GqIMsDHDpiOT/NdLMoDykcKgBwIOQACps+z0bmqPAOgPUCWZVBulTLeNUTvD+sNTM9HPZJDAZbKCSpSGhUMvt3b22tPJ+Zby2nc6Jlkq6urM4ahts9jGLUOLsdO40GopXNjWJ9xIOvXIsDKPFW2EFTyXYw80kR3pRK0kF/UC6ZhAYwJVZnaNE271KnGtZ4krWNCntbt9D7m3jfaB6rHHOgDmNHdfcbhLDTX4AeYdYVpUFzk1fFAZ8/LAVGtzEyh87cqPiXWS19a2jRNu/tKmdLX7hX8aFkRGNH0/tu/1YNCBa4WXMS47L8oX153AaNWTNa/NaXuaXRCLCwstKDEd32osKCg1WVCj9Vhep5pwuf43iVaRJqWoJN8qNYn0+vrMOiN8wC/TPBGQjlKF1m1vmyicwQYBzcOnp+rpwi06nUfN94jcblL42R8ycsBTmZx62+PP+KhqrTMVZlRWfGEavVG0BhVIEIFCZzE0ugrY1gP9fJwU4vKEDXCGLTPnb8KCqJ+yYzazOPg8svHMJNpeo/5uIESjYV6NZiGnjAFphwHxnRy1WB7exsAWm8YZZK3m6DG4z+1v7k0RtnHMjTQmQBUgbOWxzS1PtUxc++P9qf+jvr3okDQXIOfSEjTItED5SgA1OXq3oNMwfq1TBnzXgQConqr16KU0h4kqMIqAhXKHLq8o+5cKnVfBquBH29TZBkpAzsAc+qynmrAp0v56n1dAqObloLRt9TSgmQeKmDcUlYAQ6JVtL6+3k5ujgF37xGAqmAqpUx5hDw2g2U4WDtN3+mzfM4tJFesukTMM38YpDrQ1ZEqQ/dM65KSP0PFrzyl81znacZLPpeZVuvBOaXGDo0HXdLgcoqeLMzt1qwzcBLjw/apt7qUsTeWgbuUB5RzpYxjfhjHwnJVeevmBdbN4y31o+kykJjJzex+ZKh1yTY3pHy8KSv07DHqNwI8N5511x2AKd1AY5F9z5ddA5h60TbBLeVc0zTtC1XpFXdd42BLQUpNt6l+i7w/2h/upLgo0EOaa/ADzCoCV0oZWHH0HaWrgZvoflS3yGLQeJXj45P3bamyztY91fLipHbhppaDCkTWVwUcMH1cuVLmdnWkrXXU65HLXUn7xYFXF0OrxeEgTSe6TiYVLKpsVPB6f2p6tSLdUuc2dt1p4aCG5WgwNS1fbZe2u0tYRmncS+BgRxUq68PYn8Hzc/UULcXr2GRAmHxNueGGU2Rp6+/onhpQR0dHLT9wLqkcYR3UqOR8U/mkBgf5joYpA2n10D59ZyGf8/PD6GlYWFhot3trX6pR4bKI5KBHDZ9IV2hfqVzIgJDLNKUoP7/vMp39SVmlfcr+YejE9vb2lLGjO+f0qAGvl76LjWPNuinYZPyXBq8r4FUZ07W8qL85DsrbrjvcmKsZ2eehuQc/wDTzkBlU0Nc+GeiJAJOWR8rSRsyuEwYYB2jzFRU66JoHGZADrAo12j3G355fNvlUWCgDRxYR0+vEiRgvE6z62wVHlIeCNRVMJLd0WTd9lhNEYw5U0AOYsl61fMYQ6KtRNF8PqlQlwB0VDJzW8gBMLYtpoLXzspPzqN/LgLzyvC776n/G/Qzg52rJgY7LqCi9AnEFBtHyjs4Rn7N+jfxH+anb131OAZh64TH5jQpRvUnArNFHDxJwEkB9fHzcvo6B803nLRU8FS89HZxDfL0ClXBX21lv7QvtZ+/3yBCMxqdG2s9dQIk6inJcg8AV3OqBh7du3cLW1tbU+7p0CYs6EkAbsK6rDnpI6+7u7hRvsr8JUG/dutWOhwe6k48i3RTpAeVV1sF3mUXGso/NRdGNAD/AtLLk0hcnk0ad+xZlDTgkdSH/rjr4YFOgaR48yfXu3bst2lbUzefcY0Pm10O6dALpb1XYPuG0Pr5sFiliposUKxlWy9L6OCBxxvc+9P/eTi2n5rEBpg9fBKZf6qeTycfL+69mweg1BaT7+/vY2dmZOiBRx1fHwUFPJIT9fsSj/Diw17gRNwz4/+HDhwP4uWJS4OOgJ1vuUt7UeJZsaSdS/pGBogYGQbuWp94VVar6WgTmw00bVIgaQKuxO01zcg6NfiiDGDTL5TFfnqZcp1zkay64xK3LgRHoUy+KksvV086JyEhh32hekcEW1Uc9+QQHGr+q7SNAUa8Jg5jptdElJXr4SjkJDYgC50sp7S5k8q1uied4aB9SD7MvotWMaFw8lk3HsGYcXiTdGPCjRMHh25eB2R1DrhiiDo0UeEQKVjLrAkAbGHj79u2pt7C7xaQACDhhLHc1OnpmH5BpfYK7paNKOWq7/lbmVQ8R6+TPRstd2rYutO5IPxoz7Wv1xrDPaBFSEFNwOLhUMKSCioJDr3n9vX95ANjCwkJ7ci93QDDoWZWA9onmo/XyfvF7/lvjRiKvj95rmvFJw+ouH+hqKBqf2juKlLdVQZA/ozi/CPRk91m+ekR9HtOo8PhDbjlfW1trwYsaWrpbjO1VGcU5qa+GIfBRT5cDd+AkBoZKV8vzJUC2y4EJkAc99yWVVy7/XZ5GOsUNzshY4zhTNtFLx+NeGPRMGaSrC/S2qQza39/H3t7e1On/lKdMt76+Hhq4+pzvNnZwF/Gg9j+/6fXh2Pnyl8tE79eLoLkHPxli1jN/fJmLg+PBznqOhlvREULvgzydeY+Pj1tm4Rt2fb2ebXHLTdfTmbf2g9ZJ84smfQSYonQOnCIw01XvLH8nrZNPsOgeJ6KCRfa3pvH+d48OyUGyCmM9I0nve987QFpbW2vf68WYIJ4HpM9SwGg+3mYH7aRM0Gqb1dOpfK1zYH9/v32j+EBXRy6fqMQzappmCvi4hV7z9HQpHtaHnhz+Zzl+thW3uEcHCfqSB8vSYyUItMiz2u5STs78UQNNt7MD0wescr7ruxGjJfoIdPC6A0G915c8j8jQy8qP7rm80rbwngJCen/29/exubk51WfKawqGfAcZcCITuSzGZxWY8Cw0vgxV5ZW3JeO/SNd5ML+udkQy8KJAD+lc4KeU8r0AvgdAA+A3AHw3gCcBfBjAmwH8KoC/1DTNfillFcBPAPhjAF4B8BeapvnsWcum2zY680d3ufi298iKjgSHeySCtk/9V2UDoH1pHAfbY3PUEmmaZgb5uhswmmBu7bjCjiaS10EBgYIBTavChWVqffSb9Yr6LBI0UR21v/yeWxeRkldB4EsLzIcTm3nytwbTR3Xxfjs+Pp7y7o1GI4xGI6yurrZbQ7XeyotRv2ST3MGRAyXNU3lcgdHCwgK2traqinegiyflTS7ddBlW6uHR+J8+wEeXjRz4cE7wbe4KeNS7yznsO4J0TikPchmG7+TS8rUflBdZNz+7hwqRRE+QziWWHQGfSP64t90Vs8uVCNRFFOmNSE5H8zoyuHld81Z+oK44OjrCxsbG1Hk7Gh+mh+vydHHmyWWwxcVFbG5utjJTnQBeTxp1PPGZICuTVy6jFbSzLeQZ5W33vF8m9T/W0qiU8hSAvwnga5um+SMAFgF8J4DvB/CDTdP8IQCvAXjP5JH3AHhtcv0HJ+n6ljVzjZase3cid7+vuavS0OdcIbtS8muqzOjxIVPyfVuR8tQ8KVR0wiqjKDhwUMIP6x5Zhdl6uFuRDqSiPFjnLs9B5LXw+w4+XUC7petCjlYK3e6uYBTAkLRP9bwo55do8vN5/lflROtqY2OjfVEt3/CudcrAqfdHBri17zMg5GXyOSodBkoOdDXk/EjgExkj+oxaxe79ieZLpLT9PoDWm0NSgEVlxG8NYF5dXW0PDCXvejwKt67rkkpkoGj8IWWXb+UHpncmkZfdM67eDjc02HZfFouWV3Qe+fWMXF5FZTt4iZ7X8lVPqfeL+Wqf8pgBhlnwWXpRVldX27OBHIDRO0cjTY1BpiGw0jGk7GB/umyMeDLiW/aH8puCt66+vwg6M/iZ0BKA9VLKEoANAC8A+CYAH5nc/xCAb5/8fvfkPyb3v7lkcLoncXCipS9gNthQGYvkSiJSQD4IPokoCCgMbt++PbVGnwkpnbxkzAyE+H+3GCIBGC1L6W9/TvtE6+j9EVHWR553jbzt0YRRwKdLPApMfKu5gzHf/aXXtGw9X0jbqNYMFRQDQDc3N7GxsdECLOcjBz8RmI4+ei8CeerpzIyBg4MDbG1tYW9vr9d4DHR+0nHSOBsgBj38jrw9fCb7RPeBWa8H39bt6XSpTZdHeOAdZS2AmS36Go+jnieW0zRN623S+ep9ocsx3ibmRaDm7dI+1OeyfmH6LqMjowgg1eqi39pWNy4dACkY0LgofkoZe3NofPs5S2tra+1RK9p/PBUaOIkzdUCp8gdAWwbHWQFtplt8PNSbqbu+3PsTLatdJJ152atpmnullB8A8DyAXQD/GeNlrvtN09C0/DyApya/nwLwucmzh6WUBxgvjb2s+ZZS3gvgvZPfen2mEyjQ3asTWdzu9XGlmLSx18RompMTfom0GV0fAQ4OsK5x8rrnG7lMeV1Bk/cX/2saBTWah5MyKb+VGSNQ5c+fhiIhof/VK+YWJwW15qOKhpaQ8xLTuZuXHqVSTt7YDqBdYtWAa6ZdWlqaOupdQWPkJYt4KeI17WOtp/Oy8z/zcHCkFuJAl0cqw3TXTuRxcwNGPRy+zNVnySuSOQoS1GB0pctyeb+UcUwbwbwqVX6Tr9QTwGePj4/bnUBUeFSU7BvlaQbt8jmX4aqo19bWpnb7+vxWmRF5NTLgyd9ZOi3DKTJYvL/8eU0TgT0FA7xH8NI0zdRhv+wj32gBoF2a506w9fX11vvNPJkPeUTBCIEKl8x0Sd/Bj7ZBx0P5WPNUD6fyq/fjRdOZwU8p5Y0Ye3O+HMB9AD8L4J3nrVDTNB8E8EEAWFxcTNEJBzZ607v/1wkUKR0AU4pDr/PZzGUJnARfMwqfwEbTqcUTndaq+XYNNpmHglKZLAM6XudIWPIZdz+zzKy+F0HM271y/NZxjNIwD3/HFwUIJ27UDlpUzIMAh/2sW4Az4Kd15+m0x8fjraJUHlpP9ncEevy/7+Dye1Hgf8T7arkPdHmkMmxtba0hP0bGUyRzyIN+UrgvfZ0W/JCXlS94n3NjYWGhPeWXZ5TxGvMg35HnPb6O9wh89MwaNwC1jjx7hnOVhoUagQDaJTg1Tng/iovKZFdtDPR/F9CJ8nJQw752QAScyBXXUZQRLufZt9pf7HuOr3vGFhcXcefOndb40Xr5obH6jIYb8LoCIAUoBE8O4vS3gh8+w2tReZdJ5wl4/hYAv9c0zUsAUEr5OQDfAOBuKWWpGXt/3gbg3iT9PQBPA/h8GS+T3cE48LkXZV4Xnvmjrn9XAtF9B0T8714GLT8CEQDaQ7t08MiQTK8D7AIso+ieA5oobSQAFaU7A0bWh+flgMhd8dpPfdrhaSNLiP8ji0rTawApd1zRqnRQy7Hx/nAlwJ0UFCQKGqL2qHdMz6Di+7T0PTk6ft62CPwoKQ8r6FaApJ4vXfLd29tL488GuhxqmqaVUcCsgcI0Os66UULnWeQFIkWyhMqFvwFMeZ+c98nDuvsm8ugsLCxMLeEp6PC6qDL0umgbCHIUMB0dHYUeAS4zu/HicYHeTyo7o/5yw7EvRbLY5Ze22+Wle5+ZRpeVVA7y3vr6Ovb399txouxivysPLS8vt0tdem7T4eEhNjY2sLS01L74WHe+UidqP9KDd3R0NCUbM56M+kfBlcbG6v+MLgoYnQf8PA/gHaWUDYyXvb4ZwK8A+CUA34Hxjq9nAPz8JP1HJ///++T+LzYZojkFqUXrAAeYPWPDreIMvUeKSL9Jx8fH7VZAMg0tFR3ETDjoRIwmgE/EaOIqg6m3Jpro7pGqgQ7/kNyF7L9r16L2Rf8d4GR143++mI9xASoI1BrVs6EUqGodVHDSsuXSFvPzvmUZKthZnlrb0Xgq+I68OwBmrivI4Sda/uU3l7y4W2OgqyHKJyVVZBExADXyELshknmCSAo41Gvqc00PCtUlDRoXwAnAH41GbbCsziHWl0smvlQbLc8CaBUx5xo9TfQkACfv3QMwc/YR+8rnLvtf56bec+qjVB20kvS3xvqpsegGtusYLd8NcZZLmUKQypeSajwP5RbBifaTglweGKneIva1hxboKcysj8c0sp3elkiXqKfHl8F0bC8AIqR0npifT5ZSPgLgfwE4BPBrGLt6/wOAD5dS/tHk2o9NHvkxAD9ZSnkOwKsY7wy7EKLgdwVCBtB1bk0zacdUB+t1fnNSRp4H374JnAw23bIcUN5zgBIper2vCjMDIz4Ra4Aq+u/CwikCbFo/7bMaueL3aw502K9qhShY4CTmRNYdKvzms7yn76RRoKSCWpUMMLs1XnlBrTFdIm2apg0qpBUb9YEDoOh6xLfkbV3q08PlFPhrHMdAV0c6pn2IFr1vc88Aj5MrGP2voENBuy650OPD2B/yFHmcQOj27dvY3NyciuMhgKIiLaVMnUTsxiQ/BFM0InlOls5x5k95yrbTo+Ft9TnK9qvn87yKNQIv3j7fHZUZcjr+qqcccKgcKKVgY2MDe3t7ofzSXcSUcyxb+5AebjXYtM4rKytT4Fi9Sx4/qLKPdfZ26jgp4CG/81VF7Avtb/0+L53rnJ+maT4A4AN2+TMAvi5IOwLw585alqNA/e9LX1QM+pv/fYuzT8iIgfk7qpMGjUWMTOHF//psBHoitKsTOirHlXAGerrARlSO94O2QUFQNjZOtesO9LyOFOhc7mFa9j/7mkKOlgwtG05m5Q+Sj00EBgleVBmogPJ1bAUqPPhSBbK2PeJJ5z0FQcrP6vnRvJiO3ie2bQBA80HRPFleXp7aeuxgpuYBimSKXvfYEcovVbTAtOKix4RKitunubuRadU4oEdCy3aQoMYMD+kDxnNrZ2cHo9EIu7u72Nvbm4rfo+JeWVlp4350ucvnkir/TMZmMtCNyug5BRwRoHGjRsdMjaxIz0QywPOmjlH5rMuFLMfj/Tg+i4vjs8k4dtzYoWBIgY/yJeWP85R/VN5FfBsddsixzTYJXATN/QnPfYixFYyrUOZXC9iDo6MlBlc8vKdpHHjxN5UuB123jUbI1ykbVAcrivIzUKbKOwMTWZkRA3ta7wel0yDzKI9oDKK8HXgsLi5idXV16i3IvKfgI8pLy2Q73T3O64wHiix6VQjMl+XTos/amtVN/9eAj3qCFPzzf+YtGOhqqTY/9LUpwGx8hC4R1OapygBVQsD0zitdpmJ63UHJgwv1DCtdSvHy1HPkoMPBV+QlKKVgc3MT6+vruHv37pRHl3P41q1b7ZKb9gvJ56XyfSRvMqNL6xSl9XuqTxTcKLmc8T5i32s6zUv7kMtTGxsbLbiJNuZQBhA4+jI/n1daWVnB8fHxlNdHdY/q1oODA2xsbMz0t/ePXo/ifHR5jR7Dy6QbBX4yZQugnSRu9ZKoGHx3DClC2EqRcmyaZiaegwOsS13KkL4m6gLErRWvk09k7x+m0WU6778oX/9f+9QEheaTUfa81jEaB59IwMlJ35zc+pZglqXxP7yuCobj4/zgOw/IUxQiLN+XybTubCtjbviSR+3LzBPpy7QqMD2OLTvjR0H/Ve6kGGiWaiCf9+hZUW9O5vFxeQDUY/VUgZEIVvgs75GPVldXsbm52b7+gMsS+rwuMavsiZbddI7pvFKDQ40UDf6mstS5HMlKgjqtYxZEW9Mpns7HztsWGTH0xGgd+d89Vio/VC6wPAAt6NO637p1C7dv357Sd8xDQz7oRVH5w980GvW1IQQh2gccY22PHtxJ3afnSPFZHWPlFX91iurQy6QbBX5qpOBGlYHv7tLffg2YVr7KyA4iAMwwFF3Eiq51gurzkcBSygCJ3iO5hcf6RMKv5gaOGFUFYla+5h/V3ammfKN+5vWovjqRCIDU3Uugov2hL1F0oanjFQFH/dDLQ6Gj1q4LLm4PjTw8Dvi0HxzwKH9rTEYE/NU7pBbpQPNJVEIEJGrIOBiKQBHz4HdmKDlI59Zyj7PhGT+urJTPI/Cj89P52JeXeV3jhbSenM80VlRes65KNcMs8khEMj4zuKJ6RzJT81Vwo/JX5YaWzz705UZuK2d+1HdsPwEVwYOOg8cBkkaj0RQg2djYmFo6Vx3m7SRAYjsODw+xuro642H25/y/85XyexY/dFF0o8GPTi5d13QU7mAoOjcFmN1542Up+fM6qA58dAI4E0Rt0omZMY9e04nm9dU+isqM2quTx+tR65PzUFRfL1cFKCcggBkLgUJD83SAooCOPMN7HjTt9dNrui3UhSi/CT5UWEWgpzY+CqjU6+Mg3z8UTMy/xnsDXS51GQULCwvtspce/OYBoQ76dVwdFNXq4vOEfH98fIxbt261pwJHIKtWvssc/e99wbwcjDjg8rno3tDIYNG+7uoL/Z3JUiedwy7no7zUwGI/u+wA0C416Zir8aSGDXVZ05zsKiSY9d2sKm9WV1exvb2N7e1tjEajtvyNjY2pA3sd2Op/BScHBwdYW1ubSut94LzpIF6Xv/TVJtlYnZduHPhxRtH/WdCnW8TZfZ8oqsT0w/tqUZPJ1JXpCDgTBJk17pNH8631j377b5+MfRVvVIbWsas+URqvl5fL35Gr3AVxtKPEd6LwmvaDgj8+r+vqusXTgSnTrK2tzSx/aj3cSnJPmveTeyOjpSw940cBlXqGdMlL+2qgq6fMgABO5hCFPnlZlb4qH/9E4COSA1qePqvbwAl8Njc3Z7baqzLyPLra5vWhF8fryN8aNqDGDoEA+yUCVbW6kLzsSKf0UbSRbFJyGavyRcFQBPZ0mVCN+0hms93uXVZ+4HEBrMfm5iZWVlbw6NGj9mBLyrxoNcHzjYAz7ylfsC2eJ9OoZ1MB/2XLqhsHfoBc4ar7H8jPRom8QzXlq9d1Yui5LsvLy1hdXZ2qI78jpalplGoeHN81lnlIPO/MI5HVwe9FwqsmYDOB6/XR/w4WOD48HdZ3v2h6Agu3Duk+phXDsjwmgOnpIVHQ54GbuoTpgkWFE9Poej/BiXuftEwnBT3MT93YzF/7TQGQ7kyrKYaBro5cfilY93ccKd/zWVcuzC9TLjrmmhf5UDdlMNjYPT7R2Stq4bNcbZ8bDGyr1k+f03xprHi7eF89tV3Axevjhmwky2rzxI1ib5cCAgVZkZ4hcYec148yhRssFEyQdM5rfVQ/HR+fHCFw69atNvaQrwnhDjseELuystLKGY+hzOSxGlkez8Q+UM+lgj7mp/E/l003Evw4aef7lnf9uHLwZS+lTBlxIioq5y4jDqBP+j5KJ0PzWm4kyJTh9DlluqyNTi44dDLXwE4fitqn1zPwCZxYFRkI40ThZHe3KgGHCqJSyowA9WA/r5ted+DCiUuhw+v8zfocHh62QCyyXiMwroCd/z2WzXmaZbtFNoCf+SW1eD0GIlMcmaHh98ij5AHyPXmH/Lu+vj71agSPv3DQkhlqbji5wQDkZ9/o3IqWu7Qs/62xfV6OllEzHL0va5TJLwcjvK7/dbz5oRzSAwQpkyg/FGB43JAGH7MvVG7wMNiNjQ3cuXMHCwsL7Xivrq5iZ2cH+/v7U8tt7Aee/pwBcPIR5WqkR6JlL+V15U8PFPd+Pi/dSPBT6wAeeBdte3emjGIkIguBZfpE5n+eeRFN+kxAaFu6gIULs8iqiQCQ19mFUFanqPw+llFNkER1ro0jJ62DDD1QUi0gTac7FHSyqVfQJ66DBl1G0/rrxAWmvVRUIFlMmbc760tNoy5yzVOXvtyNzo8e+nia8R7oekiDVclnuvQEdMfaRB8HE1R4etinvixUFZGCL8/T56+D+UxeeXrm6XNFf0d5qryJgJO3342nmgzKwFH02w0Wj0fS9Orl8LYTuPg5ZiyDQJXt5DWtJ+Wc5kvvucqPnZ0dAOMXnm5sbLSxZvQC7e7utuOu4ItlO08R/ES8omm8DxTYO+j37fcXTTcS/PhEcVefWsl6XxnelYYzcJY/82AZipxdMSoDZso/AztaXgSiNJ9IyGj7gTh4N3tW6+7AS+9F/ZsBs0iw+XMuiDwvBSiHh4dTO08ij1vUHgoWTjhdNorKdCGp3r3IpU1+xk+DNQAAIABJREFU4Gm12ocKmFwIK7lQdv52AK/8qLztuztcEQ50ddQHcKoC0EPflH9q4AeYBRTZfFxcXGyXedXj4ztumKfOL1/ucF7OlsKAWKZlaaK8Wb7Ld5chmj6SCZERGF3v8jxE7dKx1PFQrwiAqfnK3Zuuv3RcVQ5rGq3zwsLJK3xKmd55zDFk+sPDQzx69AgA2vHXQGMuk1GWKD96bCsw/R43Hyem9YB9/e2xbuq19v6+CLqR4Ccjos9o6Qs4YRh1KTroAXIw5MzXNE164Bd/R3V0htB7Sip8ann6dW2Ley/6gIyMusqOvBwRuYDS3y4U9eRX9fowHwJQDWTO8tQzJHRiEoxwaWxhYaENDKSw0knJeql1pn2r7wHTfmc/8RBGjo/W2/kwAjl+rIPyugrU7HTUixQgA10sacyD/lZlGi176f8MGDGdKi3GffjuoiivaLkr+3byueh1iQxB9/q6vNDrEYBy75Tn3wWQMlJAojKIZTH2M/KOKfDhHPbdoPy4DGV+kbeX6bg0pjpJj02g/NC89UXQa2trU/VnADX/R+A68vhRnkbL7hGQ9zg31ZNZ/5+XHivwA8zGQ/CjzBLd573ah6ReAH1lgQsjRfluzWj6zArJBFiURj1JUT7R/8y60XZqXbU8Txc9o/3q5dXaof1CIcHlTFoI2m86OXW8tV28t7i4iPX1dTRNM+UF0mUGXfLiPSojjqvzj/YLAZC/+oTtpHsYmH3Ttfaj9mXk2XH+jgCSj9kAfOaDIuNAlxeiOIhIaTCvCFA4uCA/0etTSmlfVaEgy8vTOgOxFzki99zoMxEoy35HAEbblfVv1B9ZGdHSmOfj1xTcuNEV1UXnqZ/N5UvlalhpORrf6PX35TT1NOmYav0pjygvGNezurqKUgoePnw446lWYKU84u2P5I7WIQI92q/Z2F4UPXbgB0BoFUfBoBHAqYEiEn/zHARfWnBF6BNPwQSfcW9F5O2JhEmNQRxs1dJFeeqE0Tqflin71EH7WN2obv1RsSuo1JgbX9aJvDPaNt+RocIjO19ErSHtDyoVlreystKm1fI1SFGtR8/TQbvysJ/x4c9ofJPWf1j2mm/y5RJgVoH38QQ7MR2tcfK3vkRSd9loOS7DMuWelalt0OvRcxHQidJGBkf0fAZ+onpmoKsPcNL5q3XScdQzuTLg4waO7sTTAwR9SclfYso0fHWJUrQTCzjxRukp+byvS1oebqA84v3i8jYD8w6CMoPcx/g8dGPBT6ZMVTF4TI/+zpYLPH+37vlbkSrTZ5MkAhVRe0gR03g+EfhRhtZnaowUUa08JwdtkZDSPLIx8/IdgAInW3M9Pf+7t8fBhIMbYBbAcgIyaFrBMLebuqDzsdNYCj7HOjjIjtzIGa+q4NRlN+dnXdp1qimBgS6HfF7UhLcua0Wgx+8BsXL2+8r/zEPfj6XX+YzPFdavBkq0fTXA0KefXMEreX96W70foj7K5HEXsNTxW1g4iU3x+av/9QDe2ioEML1jLTIASylTstABEcvkJhx9Vg3M5eXlKa8Oy26aps1/bW0NpZQ2QJrlUD6Sbzyf6J4DnwwAlTK72SUan/PSjQU/NcqCmT0Q1L1B/lu3g0ZKLlJYSvpMpiw1rd6PhIUDja48amCrdj2zvrK61PLT9GdB7JzoOlYqqB0odIFQfc7XwlmWrq0TaJxGyKuHjC8f9Lo4gPJ2RH3n1qHf0zRZsOAAfK6e+va5GlXRElcGiph/Boy8HqXEr+CJ8lKKPE61tpGvIyXW1SdZ/SPjRv9HczGT0/yfGWZdRqPLW/2tQCZaZYgMb5VvWicFgCr/fKu78gjlmb7aR2N9uIyvMpP5KMDSpdDRaNTypsZaRv2eyeEI+GhsG69RPl8mPRbgxxlX3f4ObKKlMAdLJAU+mrcvJzji9oHPJlHGHDqBIve3BtidFoz0oRoocqHUF9Bont7Hfo1t5G9OYiB+Z1kmuEiR0uBzHqzuMUZ6+i3z8n7XcXJhUsp4+UsFj/OVrnEr+Ik8RdFhZp7Wz/dRGpa95pPcWFEgHinxCMh2XaOyzGKJsrN8akZFHxDjoMXbq9f6/O4CJF7nzJOe5eftrZXnBozKaD+ENMvLl828zZFsJH8As6dMq0eH6fxkefaJjjnL5Hu6VNbxUEStA2VnxB8ZrzrPRctewEkQ9mXSYwF+nJqmOyZCEbkDH/12JUWKJlFUjz4CitddMWUTU79r5UfUN10EdLK6RM925anpIuCiH05S9QDV2tPHK6bXKQi43XQ0Gs0IIwJOBZ5ReZGrPnr3mAojFSLuFmfZ/FYA5ICev7N4nz7CfKDrIcosXbrt4/0hdckZ5udK0POKgELGOy4XagZYBIL8uSzfjFyWRMDnLHlpnpp3H2I66peojpQfmRGrmzB4LZJHJDe8FES4PFJDTeum4667alnm4uIi1tbWsL+/34KfTB9G4FI/EfDW65RnerL0ZdCNBj8cTL8GIPT8qFLxuCBfd1XmVABEplBkzXr0nSTuHdB6+29tp0/sWnkRyDgNI0XAJ6qX5t83z9p9t3J0knicDMvOQJWnccHuVtvx8cn7c7T8PqCNSoUCw+vL5a8IiDvoVWGngMeBu6flbz+76DRjMNDlkc9dEv8fHh5iNBpNXff558tT0Xhm1xTgqFFBHu2KJ8qUW1b2aeqWXYvAg5LLthoA8vxrciszbmrzx8GD9wWNNzeqtf8V5HCcGENIw0bHUUGOGvLKK6yPnmDvbXHZGB30yud4VEcGZrXNNUCk3h496Zp62YOuL5puNPjJSJVAFiOh6XwCRd4iZxQHMK5Ma5Qh4yiND34mQLP8tYxa/Wrekuh6H4EQCfiu8jWdC4EIGDGd369Zof6f4IKCJTozKOoDYPoFiTo2XmcXLgpYVOipRaftcg+Pg54IGEV1zpTBQJdPkbGgfHN0dISdnZ0ZPvKAZP9oXpmxogqHddD8stcRaL3PItdYVmZEnIYXVRG7fIy+PW7Jy47yceoj0/VZNbI972hc+YzmwVPi2WYN0WCd2BfR9nBthwIKPqv9w/qxPrqcRTmlY7mystKm8X7NdGHEV7rs5vE+evJ4NC4XJb/mGvycB/H1CW6OlAdJB98FD+um96J6RxNS89dy9JqTP+fLLlE6py5B4wyrirlLEHaNU59ntEy9z751EKtt0jSah060KCZH81IvDYWF7uLw/uB1BS7KE748xrRqxWk73WPjwIcUeSu1fD00rTYWA80fHR8fY2dnpz0gs2t5IFPsEQhRg035PMovmq+R4dCXlzL55te1flF6fyYCMF311uuRseTP6Xzskp9dZQLTGx102ZuGl8oMznWNi/HnlTdUVlBGuJdH0yjI4W9/sajKMubFnV6RnK3pOv6OlnK1zNFodC7d35fmGvycVVBzALKXnCoo8gnkyoekyxlRWbX6ZhMnA0PZ/a5Ylz73uqhrEkfW23nKyawtbavGQUT/gdltlfrRmJqoHZGVwusOpvRb83HQpuXzvwopt8QioOO/fQlMeULBT83z4/w70OVTpAycb4AxX29vb2N3d3fmgEoHKc5XzvMRONKYClV8nl8EhPy3UqTwaunZJ1n+tXnWN/+ojlrXWjq972VoPdhPUX4aa0O9QdnkckrHJeILghI3ykh6IrKPhQISlSEA0t1+BEDaPr0HjF994f0TyZYIJKss5G89O4hno0VntJ1Htzk9tpIwWvaKXJKa3hWYptNBchSdgZ/aJIustz7Ulb5rctfyckFUA3P63Sf/LoGjaX3pR8m9cDo5PI1u04zKiYCKCo3oHKfISlZQoTym1hLTc4Ir+In62gGRLmn5O7vUs+len0y5DTR/1DQNtre3cf/+/db7o3LH+Un5kRSBGFeEmm8NNGUgqPa/pqgy+VJ71uue5Z+VVysnStc1P6J72fNRbI3eiwwe94ioPFfQoOSeoFLKlCdF89VxJyjzAwY1/ibiE4IjB1nR6gi/szH2cvf29trz1LIXm16U/Jprz08fUgXolnK0RMB7voSieUUegkjQ6P1IuUeTDpidFDUB44ykSj9rfxdzaDuz/usSBH0AVi2N38s8IA4ugBMAQYosYj7rabNxcrcxy/G19pqQ9dgdAi/fEurt5fNRbJp7BjT/6Eh8Bz+R4Bnoauk0fb63t4fXXnsNDx48wFNPPRV6aFxxZPFApAgw1Za8anKpBnqiNM73Xfl05Z2l8fKidC5Lu/KsyVOvG9vo288pV3xJ3p/16zru1GMZiIoAVZ/xiHhBX+PjvOQeIH17vBqLXobzlvMqgdrS0hJ2d3fb9urp+JdBNx781MgtZBI7U5WJAwJg1tXqx4R7MFgNgETM2ociQOb3MgGUTbSofn69r3CKQFRX3aO2RKAnE1IZCHGLSIV+1G63qCKPUwa0vM7OA/w4z2R9FJUfeX8c2Hi8kb/MtM84DnT1lIGCw8NDbG9v4969e/jSL/1SvPnNb54BN2otR8ol+68KOgJTfKZWZ//fh5+60vS573JZn8tkhs91n6M1kKfponGKjApe0/jCqC0aQpHFKSpI8Oc1D/ds68nP+qwGKbtnms8tLCy0r8ngfbbD26JhBnpulLchqrv2k5899ejRo1aO8WR8z+Oi6LECP878HtfDNJECdKXC/6pAnRkVjXcNePS71g5PcxpQou2J2hqVV7POInDneeq3C6UofVQ217Q5GWvLX1k7FABFa9AZIPTy9T53X7il7eDZyzk+Pp4KDnQBp/Vx0KTPOPihl0ev6ZJYbZz7KqyBrocY9Ly1tYVnn30WTz/99NQyhAKfzBMN1D0iteWuCBTUDKysrKhOvHcWwBSV2+deVNfMWKt5GbqMTP2mLKnFaEYyldcIbKJ0tXpkz+s4Oohm+ihMgPmUMuuVLmV8mCLP/dF4xiiWVu95zJO+SWFnZwdHR0fY29ubWfa6aLn1WIEfJ18GcEbwQGhXUCRlFM8jQsQ1xJ9N1IsaWJ/U2WRxUqDQ5XGJAKOX1UVZP7tVE4FSb5uDHs3L650Bs0wB6FZMAiAqoyiY3vM9Pj6e2roaeYQy8BSBH41Z812M0eGGTgPwmU/SebW7u4vd3V3cu3cPn/nMZ/DVX/3VU3xI/tNn+46r8nkUd9gX5JzGGIsUfHS/qw0ONDzvLH9ezwCQPhfJheh6VgYwvUNKwYOmqbVZZWD0n7KApHLQ68JvLmX52+KVF6KTv7UOfn6RvvhZl7yi/laApIc4stzj42Ps7e1hNBq1Hw/697adlx4L8OMTgOSWeQRAgNgbwN+KTqPy/MwDzyPLt3ZPEXoEyE4zQfswSld99H8fIaDP1rwQEalV4MDNgQDrE9W/jxCOLNIMvPq6uFrRKojUwqEA1PaSH/leHbZB00T/FehEwfl0FWfgp8b/A10vuTI+OjrCo0ePsLq6ik9/+tN4y1vegre+9a0zL6rMZJXnzW/fCenpsmd9zmRAotaurN3R7646dD0b5dVVR29XBo6yvLyeLsfUuxHl53XN2qX1dE9M1gfUJ55/5gHMApcJXPQ/5Vh06GFUH+pS9rnKy93dXezt7bXgvxaychH0WICfiNihvl0u2vbuAxZ5gaLYDf24kosUa99663NRPlEaF0hdVGNQn2B98jsrKPJxAWbfn8X8fZlRhVYkQLoAkF9XSwaYPbDM+zcTji7gVOlkoEnbFrnLI9CjPMxXc3SNQ7QUOND1kM83XmuaplUAr7zyCj772c+2b9c+PDzE5ubm1IF0fK42TzV/52V/tqZIo999y615f7L+yOrUVR7z6UrbpayzJag+5bsc49zz4OWorlo3vefySZ+N9JBf1/J9N5gDFD6relLjlABM8aHWt8YnBFgkyq7t7W08evQIo9EIu7u7M+04jX7rQ48d+HFm0YPqonSKzhXwKOl6JskVrw+8D3SXgPG6ZcCnz/M16hI8WZoMYPgzXc+eJp9MIDhwytrjEzHLj/f0Gf73I+kd9BKU+JhrubrzS8uqAWYHOdGOD00bnf/RJXQGmg+K5ge9P8vLy3j++efxxBNPYGVlBQsLC9jc3Ay9itF8UN7qs1TWJR/6KP4uQFF7titd17O1+kTzK5MNes9lwmnqwWd1NxSAcEnH9Q/vRQBMvUlR3bLNHtRlnr/v+qrpHy9HZVvXsrsGe/P3wcEBdnZ2cP/+fezs7ODRo0dTwc5Z2eelxw78KFGIZBaAKjAgtxpcaLjHp6a8o1iUqB56zz1VWb0caGVlnHXSZvdqz9XyyayZ6Lorgwycsv+zMc5AFMuI+krTR7sx3BrS+mq9NE9fclDeyASrtzk6noHk4CfjiRpgHOhyKOLBPul3d3exurqKL37xi3j55ZexsbGBO3fuhBs59Nmu8e0DfE57/yzg5SwgpybLMtmape/yRPUpKzLkXHZmRrO3S5+JlqlcPmZGpAKLDFCRFPQ4OPK+0brpiomving/ad7qbGBg/6NHj/DSSy/h1Vdfxc7ODra3t089zmehxx78ZHESXR4EZQTfTlibNKrUMkurD9hwQFCzKrqYIQNoWue+YCcCL6dN01VONE7u4mUeEZDx8ewCXVpuVI5e93o5H3l5HnOmZavQqXkgo37wNmfHOiipoBvo6iibe11z8vh4fOLz0tISfv/3fx9vectbpmRan7mvFMmYLsMperZW7z516JJntbwvSiH6/Kqly4DOaWUygYIbt5k+iYCH1kXlpHugXV6oTnK9p7quFruTAdxoqc2f07YeHBy0S1tbW1u4f/8+XnnlFdy/fx/b29s4ODiYqpt/Looea/ADTMdKZCi4pshLOTnSWy1wpcwCc0WTTYpIefI7Qv0ZE0Z56PXapIzIvRM1a6dWtqar9bdP6OxaVG5tcujzfcr3s3M8vfaJkuYd7c7weuruhxqpcNOytM41D6c+U7PSBrp8cpDeZUTs7e1he3sbL774Ivb29sKDLLuUQpeRlD1fky19ldBp5E3XvbOkr8mD0+Z/FoClZenOKO3LPsdTML9oTLgEqvwU5afGlhtkutsrMxQd8FCm1TbneH/oCfXUrTs7O3jw4EG77NUF6i8KAL0uwI+6iTMFmgELXavlPd+NFAmxmlXlg+tpHWT0Gew+kzQDRFoWKaojv3WZMAMUXV6fLmurTzrNL1PqNQtG7ynI8ZPB/VkXNtn4eT0ODw9nAp0XFxfD9+TwGfVcOpD2vGttVcB1kdbTQP0pAwI1y5+7YO7fv4/d3d2poPausc7uZ2n6yo0uOg8wuqxyTgN4usrK5HMmtxwAEfAAJ/GAkefWjcnMyObHn89iBCkH9Hl/HYa2IwLPkd7r4qmo/svLyzg4OGjBT5cheJH0WICfGmN3gZ/sebfQNT9gerdOxGCOsiOBFSnOLD9l7lqckV5zb40L2pqVoHWM8uzr7dGyszbyvnvosgmm/cnTQWuCOxNKnhfrooIoet6VTDa2nubo6AgrKytTPJUJvlofaB+p56cPDeDneijid71Xs/yPjo7w8OFD7O/v4+DgYOrMqD5lRfciuePpMjlUe8b/RwDhvGCkS5ZE97T8Pv3m13z+9ZlDOj+9LhqTwznpOiozQPV/ZDDVPDAZ8OHZUVn/RaQ6KQJffcaZ+T948ADb29vV9mWg/qx048FPV+f6+Si+jBHFeAAng+LKwtfaM+DDPGuDlTFzZllEzNQ3/8yK6BIEPumjbehKGZDMyohAQg20RULIgUKf8msCoqstLsQiCyvqP4//UdDnY6wf8i3P8dG+oNenD/ihoBtovigC1873fOHpaDTCwcFBasz1MTL8v/N0Vz4R9ZnjtfxPk2ff+32fPYvh1leOZmOkAIg6yHdvZUDRZR2/l5aWZsI8vG5+tg/LzYKca/Jc+TR6Nvqv9dKz0La2tmZifWr5XAR17nstpfx4KeXFUsqn5NqbSikfL6U8O/l+4+R6KaX8UCnluVLKr5dS3i7PPDNJ/2wp5ZkLb8mEfLD0rJ9oIGuMFikLblHUMxsi5uLzEePqd9cgq6L0PC8CBXseWr8oViqKgalZUxlp/l6XyDuhkyISplr3rj7pa/k5QPJJTv5Q4aF1j7w1PN00UgIuVP0TtYP92Bf8DJ6fq6dIDpAyj4IrgYODA3zhC1/A1tZWC3a7+MPzj+rVJUP6yJjM4Og7H09DXXX2/iNlHphorjrVjL3sejQmLrd0M00kByKgUmu/G4G+i8vzYfkMco7ak/Unf2e7mTMAE8lwAHj06FG1ny+D+hz68a8BvNOuvQ/AJ5qm+UoAn5j8B4A/DeArJ5/3AvhhYAyWAHwAwB8H8HUAPkDAdF7KBoVEhRM9QwZgukjR6RkIwPSBdREj1pTzaaimoPsKgK58a4rXPWJdArJWtte3Bp50okaTv1Zvz6urvtnkjM7CyQSKvwVZ66HPqpfH2561z4FN5MVkAGxtG7zz+wB+rpcyBVOjpmnw8ssv46WXXmoDn5UPmCbK0/m9r9zI5kffNH3aFcnSTLb1qU9XuX2NzqyuzKPmzXLPbFQHl3fepqgvMgCkddf57ef3+PPc3dV1Dl0kbyNAFaXXvnC9yt+UYRldlrzqBD9N0/w3AK/a5XcD+NDk94cAfLtc/4lmTP8DwN1SypMAvg3Ax5umebVpmtcAfByzgOpcVAMDHpWu6SOGcOXWNCcHSul6e8ZUmRJkmq42RCAsY8A+QoP1iNygGaNGdapZlzVhG9UjEwrePp3A+u4ZTlwtv9b+rN5err7cT4USy9f3K0UWUw0sRtYnecvrxN9+wKLnpcthfWg45PB6SOe1Ux+w3jQNHj16hBdeeAEPHjxoA5+VDyJZo8Db72X/u2RaTSb1odPO1S65dhrqkn1Z+V11y+QAf0dyj79rR1B0yflIb0V6zHWfyrLaONb4oKtveD/yeKl8rcnvywI+wNljfp5omuaFye8vAHhi8vspAJ+TdJ+fXMuuz1Ap5b0Ye41mFFAXo9fcZqpA9RN5Fxw9856eFh0pu4x59TurY9fzkYegD+lztbIja6WP0Iv+Z2kzN7CWlwEj5qPbMR3QRVsuu9oYEYOSo75wwRPdj4Kl+d/P9WC9NZ0H2Ovb2lXZ8bpvfc5IBc9Al0cqw7JYio7nw7m6vb2NV199FS+++CKefPJJ3Lp1q7q00qecaC73qWckF7rS1+RzBg77KsM+QCYzKrvq1VWWytcI3ER1cNKg5yyGMDKsvNzMENX/bkh6Xt4nXeAvqmNEKr9Ur0Ven7PoutPSuQOem6ZpSimng9/1/D4I4IMAsLi42Lil2uV9cMXC30tLS1hZWWk/q6urWF1dxcrKCpaXl7G8vNxa9GrZExkvLS2Fhy+RUTMgxfJrdBphE03WiGH1udrk5vNZmWdpR2Zd1Z6LrCMHmTWvSlc9ugCPp+eb2D3AOKqr5x2BHq83hRDBj7ZXy9OlLQc/GgjdF/wMAc+XTyrDlpeXqwOTKe1Ikezt7WFra6s9CXd/fx9ra2u9AU82l8+qZPoqxT7pazIjk2F95EGX7NP8L4I8ry4wpNfVOPEzvdxwcwDn+UWym7ImWurKgE9UR29v9t/lWCS/Dg8Pe72T8DLorODni6WUJ5umeaGMl7VenFy/B+BpSfe2ybV7AL7Rrv/XrkIyYV1jokghLSwsYGVlBevr69jY2Gg/6+vrWFtbw/Ly8hQw4md5ebkFQR6H0uVWzJBrjWlrAqOGxLsQep/7kadEy+1C4tGk7FNOlk9mNWld1AOU9UmUVySQ3HrSmBr3ykRLR1H/sV7+2gnmr0HSUZ7ZLkU9vkG9QjVin9WOBRjoesnnnH8fHh5ia2sL29vbGI1G2N/fT3er9i3PZUMNFEWyKEvbp+zTpDsNmMqM4C6wFBldWb7Rc12khlMtDz9EtQbgFNREp377bwc+meyP6ngWuaH56AGHlF18i3smwy6C1zI6K/j5KIBnAHzf5Pvn5frfKKV8GOPg5gcTgPQxAP+4nAQ5fyuA93cVUkrB8vIygHhXRE2JqVBYXl7G2toaNjc328/GxgbW1tamPuoVojeI3h9fMlOw4+BILfmoTX4vQt9AfCbLZVopWn5Ubq1sb4NeI9UmcQZSfJmIsVdapz5WStTObNKrx8W3jRK0+MGXfcZF0zIfCi5vi1tIfMZPn6b11IdqsQUDXR5lS/ddSsbnYtOM34e0s7OD4+PjqTN/srng17sMs+j3ae7VyOVbLc1p8j0N1eRpF1CK5FMme/TZ2phHAIcywdOpNygyhql/NI0fxqr6KdupFenXDPhGx3dE5EYt+Xd/fx9bW1vpbtXLlFWd4KeU8lMYe22+pJTyeYx3bX0fgJ8ppbwHwP8D8Ocnyf8jgHcBeA7ADoDvBoCmaV4tpfxDAL88SfcPmqbxIOqQ3POjlr4zKjC7FVu9P/oMlcb+/j4WFhaws7PTItODg4MW+Ozv77cA6ODgAAsLC1heXm6ZiAzmCqyG1Pntwk1BkzIu6xsdQnVRgsLBowvMSIBnk/s05ZGiIwkc/LBeDj5rZeqkq/GOfqtXxXdScax1h5dPegXeUbvYlysrK9je3sbR0RG2trbaM1z29vbaF/zx5X/b29vY3d1tLaXd3V2MRiPs7e11LntpPdRbMNDVUya3nLJ7+/v7ePnll3H//n2sra2hlILd3V2srKxMLdm7ocayOZcjQBTVtdaOrrRdQOc05Z2WXAdEnhWXX13ejkjWRnlEcsvr1rdML1tlSy2dyxr+jvSUy9dIf0b/vV/5n0cx6PfBwcHUZ2dnBw8fPsSLL76Ie/fuYX9/f8rY1F3V7BPK5r7GXheVLmv1OqmU8hDA71x3PXrSlwB4+bor0YNuSj2Bm1PXm1JPIK7rH2ya5g9cR2Ued7pBMuym8/A80k2pJ3Bz6nph8mveT3j+naZpvva6K9GHSim/chPqelPqCdycut6UegI3q66PCd0IGXaT+OKm1PWm1BO4OXW9yHoOh34MNNBAAw000ECvKxrAz0ADDTTQQAMN9LqieQc5vOvcAAAFfElEQVQ/H7zuCpyCbkpdb0o9gZtT15tST+Bm1fVxoJvS3zelnsDNqetNqSdwc+p6YfWc64DngQYaaKCBBhpooIumeff8DDTQQAMNNNBAA10oDeBnoIEGGmiggQZ6XdHcgp9SyjtLKb9TSnmulPK+a67L06WUXyql/GYp5dOllL81uf6mUsrHSynPTr7fOLleSik/NKn7r5dS3n7F9V0spfxaKeUXJv+/vJTyyUl9frqUsjK5vjr5/9zk/pddcT3vllI+Ukr57VLKb5VSvn6O+/R7J2P/qVLKT5VS1uahX0spP15KebGU8im5duo+LKU8M0n/bCnlmcuq7+uF5kl+TeozyLDLqeeNkGHzKr8m5V2PDItOwL3uD4BFAL8L4CsArAD4PwC+6hrr8ySAt09+vwHA/wXwVQD+CYD3Ta6/D8D3T36/C8B/AlAAvAPAJ6+4vn8HwL8F8AuT/z8D4Dsnv38EwF+b/P7rAH5k8vs7Afz0FdfzQwC+Z/J7BcDdeexTAE8B+D0A69Kff2Ue+hXAnwTwdgCfkmun6kMAbwLwmcn3Gye/33iVvPA4feZNfk3qNMiwy6nn3MuweZZfkzKuRYZdGZOcsjO+HsDH5P/7Abz/uusl9fl5AH8K45Nbn5xcexLjA80A4EcBfJekb9NdQd3eBuATAL4JwC9MmORlAEvetwA+BuDrJ7+XJunKFdXzzmRCFrs+j336FIDPTSbW0qRfv21e+hXAl5ngOFUfAvguAD8q16fSDZ9Tj8dcy69JnQYZdv563ggZNu/ya1LOlcuweV324mCRPj+5du00cQF+DYBPAniiaZoXJre+AOCJye/rrP8/A/B3AfAFTm8GcL9pGr4QRevS1nNy/8Ek/VXQlwN4CcC/mri3/2UpZRNz2KdN09wD8AMAngfwAsb99KuYz34FTt+HczvfbijNdX8OMuzC6EbIsBsov4ArkGHzCn7mkkoptwD8OwB/u2maLb3XjOHmtZ4bUEr5MwBebJrmV6+zHj1pCWNX5w83TfM1ALYxdm+2NA99CgCT9eZ3Yyzs3gpgE8A7r7VSPWle+nCg+aBBhl0o3QgZdpPlF3B5fTiv4OcegKfl/9sm166NSinLGAuNf9M0zc9NLn+xlPLk5P6TAF6cXL+u+n8DgD9bSvksgA9j7Db+5wDullL4HjetS1vPyf07AF65gnoCY2T++aZpPjn5/xGMBcm89SkAfAuA32ua5qWmaQ4A/BzGfT2P/Qqcvg/nbr7dcJrL/hxk2IXTTZFhN01+AVcgw+YV/PwygK+cRKOvYBx09dHrqkwppQD4MQC/1TTNP5VbHwXAqPJnMF5H5/W/PIlMfweAB+LCuzRqmub9TdO8rWmaL8O4z36xaZq/COCXAHxHUk/W/zsm6a/ESmma5gsAPldK+cOTS98M4DcxZ306oecBvKOUsjHhBdZ17vo1KL9PH34MwLeWUt44sRK/dXJtoLPRXMkvYJBhl1TXmyLDbpr88jpcjgy7zCCmcwZAvQvjHQm/C+DvXXNd/gTGbrdfB/C/J593YbwO+gkAzwL4LwDeNElfAPyLSd1/A8DXXkOdvxEnOyW+AsD/BPAcgJ8FsDq5vjb5/9zk/ldccR3/KIBfmfTrv8c4Sn8u+xTA3wfw2wA+BeAnAazOQ78C+CmM1/EPMLZE33OWPgTwVyf1fQ7Ad181vz5un3mSX5P6DDLscup4I2TYvMqvSXnXIsOG11sMNNBAAw000ECvK5rXZa+BBhpooIEGGmigS6EB/Aw00EADDTTQQK8rGsDPQAMNNNBAAw30uqIB/Aw00EADDTTQQK8rGsDPQAMNNNBAAw30uqIB/Aw00EADDTTQQK8rGsDPQAMNNNBAAw30uqL/D//SUDo1ByWPAAAAAElFTkSuQmCC\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "execution_count": 12,
      "metadata": {
        "gather": {
          "logged": 1621081361012
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print('Shape of images..')\r\n",
        "effusion.shape, normal.shape"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Shape of images..\n"
          ]
        },
        {
          "output_type": "execute_result",
          "execution_count": 13,
          "data": {
            "text/plain": "((1024, 1024, 3), (1024, 1024, 3))"
          },
          "metadata": {}
        }
      ],
      "execution_count": 13,
      "metadata": {
        "gather": {
          "logged": 1621081361201
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Next we run some image transformations to improve the training quality and make the model more robust.\r\n",
        "\r\n",
        "The below transformations will be applied to trianing samples."
      ],
      "metadata": {
        "nteract": {
          "transient": {
            "deleting": false
          }
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Resizing the image to 256,256\r\n",
        "resized_img = cv2.resize(effusion, (256,256), interpolation = cv2.INTER_NEAREST)\r\n",
        "plt.imshow(cv2.cvtColor(resized_img, cv2.COLOR_BGR2RGB))\r\n",
        "print(resized_img.shape)"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(256, 256, 3)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 432x288 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "execution_count": 14,
      "metadata": {
        "collapsed": true,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "gather": {
          "logged": 1621081361678
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Applying morphological transformations. \r\n",
        "# There are 3 types of morphological transformations\r\n",
        "# 1. Erosion - Erosion shrinks bright regions and enlarges dark regions\r\n",
        "# 2. Dilation - Dilation shrinks dark regions and enlarges bright regions.\r\n",
        "# 3. Opening - Erosion followed by dilation. This is used to remove small bright spots and connect small dark spots. \r\n",
        "# helps in opening up dark spots between bright regions\r\n",
        "# 4. Closing - Dilation followed by erosion. This removes dark spots between white regions. \r\n",
        "\r\n",
        "# Since we are more focused on the white regions near the lungs, I'm using Opening. \r\n",
        "# Also if you notice the images below the problem is more noticable when opened. \r\n",
        "kernel = np.ones((5,5),np.uint8)\r\n",
        "opened = cv2.morphologyEx(resized_img, cv2.MORPH_OPEN, kernel)\r\n",
        "plt.imshow(opened)"
      ],
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 15,
          "data": {
            "text/plain": "<matplotlib.image.AxesImage at 0x7f113f5588e0>"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 432x288 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "execution_count": 15,
      "metadata": {
        "collapsed": true,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "gather": {
          "logged": 1621081361971
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "kernel = np.ones((5,5),np.uint8)\r\n",
        "opened = cv2.morphologyEx(resized_img, cv2.MORPH_CLOSE, kernel)\r\n",
        "plt.imshow(opened)"
      ],
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 16,
          "data": {
            "text/plain": "<matplotlib.image.AxesImage at 0x7f113efb2190>"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 432x288 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "execution_count": 16,
      "metadata": {
        "collapsed": true,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "gather": {
          "logged": 1621081362503
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Normalizing the image\r\n",
        "normalizedImg = np.zeros((256, 256))\r\n",
        "plt.imshow(cv2.normalize(opened, normalizedImg, 0, 255, cv2.NORM_MINMAX))"
      ],
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 17,
          "data": {
            "text/plain": "<matplotlib.image.AxesImage at 0x7f113eefc850>"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 432x288 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "execution_count": 17,
      "metadata": {
        "collapsed": true,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "gather": {
          "logged": 1621081362918
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Few more techniques we can experiment with are. \r\n",
        "- Cropping the image. \r\n",
        "- Rotating the image <10 degree"
      ],
      "metadata": {
        "nteract": {
          "transient": {
            "deleting": false
          }
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Data Augmentation ###\n",
        "\n",
        "Now that we have read the images, the next step is data augmentation.\n",
        "\n",
        "In this re-usable method we will also apply slight rotation for training data. "
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "\n",
        "datagen = ImageDataGenerator(\n",
        "    featurewise_center=True,\n",
        "    featurewise_std_normalization=True,\n",
        "    rotation_range=10,\n",
        "    width_shift_range=0,\n",
        "    height_shift_range=0,\n",
        "    vertical_flip=False,)\n",
        "\n",
        "def preprocess_img(img, mode):\n",
        "    normalizedImg = np.zeros((256, 256))\n",
        "    resized_img = cv2.resize(img, (256,256))\n",
        "    normalized_img = cv2.normalize(resized_img, None, alpha=0, beta=1, norm_type=cv2.NORM_MINMAX, dtype=cv2.CV_32F)[:,:,np.newaxis]\n",
        "    if mode == 'train':\n",
        "        if np.random.randn() > 0:\n",
        "            kernel = np.ones((5,5),np.uint8)\n",
        "            opened = cv2.morphologyEx(normalized_img, cv2.MORPH_OPEN, kernel)[:,:,np.newaxis]\n",
        "            normalized_img = datagen.random_transform(opened)\n",
        "    return normalized_img"
      ],
      "outputs": [],
      "execution_count": 18,
      "metadata": {
        "gather": {
          "logged": 1621081363081
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "## Output of the final pre-processing message, with few random transformation\r\n",
        "\r\n",
        "img_path = os.path.join(DATASET_PATH, disease_cls[0], '*')\r\n",
        "imgs = glob.glob(img_path)\r\n",
        "image_id = 8\r\n",
        "print('Before pre processing image....')\r\n",
        "print(imgs[image_id])\r\n",
        "img = cv2.imread(imgs[image_id], cv2.IMREAD_GRAYSCALE)\r\n",
        "print('shape..', img.shape)\r\n",
        "plt.imshow(img,cmap='gray')"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Before pre processing image....\n",
            "/var/tmp/effusion/data/effusion/00007633_000.png\n",
            "shape.. (1024, 1024)\n"
          ]
        },
        {
          "output_type": "execute_result",
          "execution_count": 19,
          "data": {
            "text/plain": "<matplotlib.image.AxesImage at 0x7f113eed29a0>"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 432x288 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "execution_count": 19,
      "metadata": {
        "collapsed": true,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "gather": {
          "logged": 1621081363664
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "processed_img = preprocess_img(img[:, :, np.newaxis], 'train')\r\n",
        "print('After pre processing image....')\r\n",
        "print('shape..',processed_img[:,:,0].shape)\r\n",
        "plt.imshow(processed_img[:,:,0],cmap='gray')"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "After pre processing image....\n",
            "shape.. (256, 256)\n"
          ]
        },
        {
          "output_type": "execute_result",
          "execution_count": 20,
          "data": {
            "text/plain": "<matplotlib.image.AxesImage at 0x7f113ee9dfd0>"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 432x288 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "execution_count": 20,
      "metadata": {
        "collapsed": true,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "gather": {
          "logged": 1621081363976
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 2. Model building"
      ],
      "metadata": {}
    },
    {
      "cell_type": "markdown",
      "source": [
        "Let us build a CNN model."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.metrics import AUC\n",
        "\n",
        "img_channels = 1\n",
        "img_rows = 256\n",
        "img_cols = 256\n",
        "nb_classes = 1\n",
        "\n",
        "cnn = Sequential()\n",
        "cnn.add(Conv2D(32, (3, 3), padding='same', activation='relu', input_shape=(img_rows,img_cols,img_channels)))\n",
        "cnn.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "\n",
        "cnn.add(Conv2D(64, (3, 3), padding='same', activation='relu',))\n",
        "cnn.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "\n",
        "cnn.add(Conv2D(128, (3, 3), padding='same', activation='relu',))\n",
        "cnn.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "\n",
        "cnn.add(Flatten())\n",
        "cnn.add(Dense(512))\n",
        "cnn.add(Dense(512))\n",
        "cnn.add(Dense(nb_classes,activation='sigmoid'))\n",
        "cnn.compile(optimizer='sgd', loss='binary_crossentropy', metrics=['accuracy', AUC(name='auc')])\n",
        "cnn.summary()"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d (Conv2D)              (None, 256, 256, 32)      320       \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 128, 128, 32)      0         \n",
            "_________________________________________________________________\n",
            "conv2d_1 (Conv2D)            (None, 128, 128, 64)      18496     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 64, 64, 64)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_2 (Conv2D)            (None, 64, 64, 128)       73856     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_2 (MaxPooling2 (None, 32, 32, 128)       0         \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 131072)            0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 512)               67109376  \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 512)               262656    \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 1)                 513       \n",
            "=================================================================\n",
            "Total params: 67,465,217\n",
            "Trainable params: 67,465,217\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "execution_count": 21,
      "metadata": {
        "gather": {
          "logged": 1621081364204
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "\n",
        "class AugmentedDataGenerator(tf.keras.utils.Sequence):\n",
        "    'Generates data for Keras'\n",
        "    def __init__(self, mode='train', ablation=None, disease_cls = ['nofinding', 'effusion'], \n",
        "                 batch_size=32, dim=(256, 256), n_channels=1, shuffle=True):\n",
        "        'Initialization'\n",
        "        self.dim = dim\n",
        "        self.batch_size = batch_size\n",
        "        self.labels = {}\n",
        "        self.list_IDs = []\n",
        "        self.mode = mode\n",
        "        \n",
        "        for i, cls in enumerate(disease_cls):\n",
        "            paths = glob.glob(os.path.join(DATASET_PATH, cls, '*'))\n",
        "            brk_point = int(len(paths)*0.8)\n",
        "            if self.mode == 'train':\n",
        "                paths = paths[:brk_point]\n",
        "            else:\n",
        "                paths = paths[brk_point:]\n",
        "            if ablation is not None:\n",
        "                paths = paths[:int(len(paths)*ablation/100)]\n",
        "            self.list_IDs += paths\n",
        "            self.labels.update({p:i for p in paths})\n",
        "        \n",
        "            \n",
        "        self.n_channels = n_channels\n",
        "        self.n_classes = len(disease_cls)\n",
        "        self.shuffle = shuffle\n",
        "        self.on_epoch_end()\n",
        "\n",
        "    def __len__(self):\n",
        "        'Denotes the number of batches per epoch'\n",
        "        return int(np.floor(len(self.list_IDs) / self.batch_size))\n",
        "\n",
        "    def __getitem__(self, index):\n",
        "        'Generate one batch of data'\n",
        "        indexes = self.indexes[index*self.batch_size:(index+1)*self.batch_size]\n",
        "        list_IDs_temp = [self.list_IDs[k] for k in indexes]\n",
        "        X, y = self.__data_generation(list_IDs_temp)\n",
        "        return X, y\n",
        "\n",
        "    def on_epoch_end(self):\n",
        "        'Updates indexes after each epoch'\n",
        "        self.indexes = np.arange(len(self.list_IDs))\n",
        "        if self.shuffle == True:\n",
        "            np.random.shuffle(self.indexes)\n",
        "\n",
        "    def __data_generation(self, list_IDs_temp):\n",
        "        'Generates data containing batch_size samples' # X : (n_samples, *dim, n_channels)\n",
        "        # Initialization\n",
        "        X = np.empty((self.batch_size, *self.dim, self.n_channels))\n",
        "        y = np.empty((self.batch_size), dtype=int)\n",
        "        \n",
        "        delete_rows = []\n",
        "        # Generate data\n",
        "        for i, ID in enumerate(list_IDs_temp):\n",
        "            img = cv2.imread(ID, cv2.IMREAD_GRAYSCALE)\n",
        "            img = img[:, :, np.newaxis]\n",
        "            if img.shape[:3] == (1024, 1024,1):\n",
        "                img = preprocess_img(img, self.mode)\n",
        "                X[i,] = img\n",
        "                y[i] = self.labels[ID]\n",
        "            else:\n",
        "                delete_rows.append(i)\n",
        "                continue\n",
        "        X = np.delete(X, delete_rows, axis=0)\n",
        "        y = np.delete(y, delete_rows, axis=0)\n",
        "        # return y since we have only 2 categories, for multi-class this needs to be converted into\n",
        "        # one hot encoded vector.\n",
        "        return X, y"
      ],
      "outputs": [],
      "execution_count": 22,
      "metadata": {
        "gather": {
          "logged": 1621081364365
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Ablation Run"
      ],
      "metadata": {}
    },
    {
      "cell_type": "markdown",
      "source": [
        "An ablation run is when you systematically modify certain parts of the input, in order to observe the equivalent change in the input.\n",
        "\n",
        "For the following section, we'll be using the Data Generator concept that you previously worked on."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "training_generator = AugmentedDataGenerator('train', ablation=50)\n",
        "validation_generator = AugmentedDataGenerator('val', ablation=50)\n",
        "\n",
        "cnn.fit(training_generator, epochs=1, validation_data=validation_generator)"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            " 2/13 [===>..........................] - ETA: 0s - loss: 0.5548 - accuracy: 0.4844 - auc: 0.4966WARNING:tensorflow:Callbacks method `on_train_batch_end` is slow compared to the batch time (batch time: 0.0501s vs `on_train_batch_end` time: 0.0886s). Check your callbacks.\n",
            "13/13 [==============================] - 8s 596ms/step - loss: 0.3514 - accuracy: 0.8462 - auc: 0.4623 - val_loss: 0.2870 - val_accuracy: 0.9167 - val_auc: 0.5497\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n"
          ]
        },
        {
          "output_type": "execute_result",
          "execution_count": 23,
          "data": {
            "text/plain": "<tensorflow.python.keras.callbacks.History at 0x7f113ee3a9d0>"
          },
          "metadata": {}
        }
      ],
      "execution_count": 23,
      "metadata": {
        "gather": {
          "logged": 1621081373395
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "training_generator = AugmentedDataGenerator('train', ablation=5)\n",
        "\n",
        "cnn.fit(training_generator, epochs=5, validation_data=None)"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/5\n",
            "1/1 [==============================] - 0s 86ms/step - loss: 0.3006 - accuracy: 0.9062 - auc: 0.6954\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Epoch 2/5\n",
            "1/1 [==============================] - 0s 85ms/step - loss: 0.3001 - accuracy: 0.9062 - auc: 0.7299\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Epoch 3/5\n",
            "1/1 [==============================] - 0s 85ms/step - loss: 0.3093 - accuracy: 0.9062 - auc: 0.5977\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Epoch 4/5\n",
            "1/1 [==============================] - 0s 85ms/step - loss: 0.3199 - accuracy: 0.9062 - auc: 0.4253\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Epoch 5/5\n",
            "1/1 [==============================] - 0s 86ms/step - loss: 0.3150 - accuracy: 0.9062 - auc: 0.5460\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n"
          ]
        },
        {
          "output_type": "execute_result",
          "execution_count": 24,
          "data": {
            "text/plain": "<tensorflow.python.keras.callbacks.History at 0x7f118e2676a0>"
          },
          "metadata": {}
        }
      ],
      "execution_count": 24,
      "metadata": {
        "gather": {
          "logged": 1621081377145
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.metrics import roc_auc_score\n",
        "from tensorflow.keras import optimizers\n",
        "from tensorflow.keras.callbacks import *\n",
        "\n",
        "class roc_callback(Callback):\n",
        "    \"\"\" ROC callback : Calculates the validation AUC and adds it to the dictionary \"\"\"\n",
        "    def on_train_begin(self, logs={}):\n",
        "        logs['val_auc'] = 0\n",
        "\n",
        "    def on_epoch_end(self, epoch, logs={}):\n",
        "        y_p = []\n",
        "        y_v = []\n",
        "        for i in range(len(validation_generator)):\n",
        "            x_val, y_val = validation_generator[i]\n",
        "            y_pred = self.model.predict(x_val)\n",
        "            y_p.append(y_pred)\n",
        "            y_v.append(y_val)\n",
        "        y_p = np.concatenate(y_p)\n",
        "        y_v = np.concatenate(y_v)\n",
        "        roc_auc = roc_auc_score(y_v, y_p)\n",
        "        print ('\\nVal AUC for epoch{}: {}'.format(epoch, roc_auc))\n",
        "        logs['val_auc'] = roc_auc"
      ],
      "outputs": [],
      "execution_count": 25,
      "metadata": {
        "gather": {
          "logged": 1621081377397
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Validating AUC callback\n",
        "training_generator = AugmentedDataGenerator('train', ablation=20)\n",
        "validation_generator = AugmentedDataGenerator('val', ablation=20)\n",
        "auc_logger = roc_callback()\n",
        "history = cnn.fit(training_generator, epochs=5, validation_data=validation_generator, callbacks=[auc_logger])"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/5\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3192 - accuracy: 0.9000 - auc: 0.6302\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Val AUC for epoch0: 0.5517241379310345\n",
            "5/5 [==============================] - 4s 711ms/step - loss: 0.3192 - accuracy: 0.9000 - auc: 0.6302 - val_loss: 0.2546 - val_accuracy: 0.9375 - val_auc: 0.4500\n",
            "Epoch 2/5\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3027 - accuracy: 0.9062 - auc: 0.6179\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Val AUC for epoch1: 0.45977011494252873\n",
            "5/5 [==============================] - 3s 593ms/step - loss: 0.3027 - accuracy: 0.9062 - auc: 0.6179 - val_loss: 0.3173 - val_accuracy: 0.9062 - val_auc: 0.5345\n",
            "Epoch 3/5\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3177 - accuracy: 0.9000 - auc: 0.6721\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Val AUC for epoch2: 0.5172413793103448\n",
            "5/5 [==============================] - 3s 591ms/step - loss: 0.3177 - accuracy: 0.9000 - auc: 0.6721 - val_loss: 0.3425 - val_accuracy: 0.9062 - val_auc: 0.4770\n",
            "Epoch 4/5\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3137 - accuracy: 0.9000 - auc: 0.6549\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Val AUC for epoch3: 0.5333333333333333\n",
            "5/5 [==============================] - 3s 641ms/step - loss: 0.3137 - accuracy: 0.9000 - auc: 0.6549 - val_loss: 0.3183 - val_accuracy: 0.9062 - val_auc: 0.5115\n",
            "Epoch 5/5\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3035 - accuracy: 0.9062 - auc: 0.6320\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Val AUC for epoch4: 0.6\n",
            "5/5 [==============================] - 3s 598ms/step - loss: 0.3035 - accuracy: 0.9062 - auc: 0.6320 - val_loss: 0.2503 - val_accuracy: 0.9375 - val_auc: 0.4833\n"
          ]
        }
      ],
      "execution_count": 26,
      "metadata": {
        "gather": {
          "logged": 1621081397034
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Final Run"
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\r\n",
        "\r\n",
        "def plot(history):\r\n",
        "    \r\n",
        "    \"\"\" plot: plot helps in plotting, various metrics against the iterations using run history \"\"\"\r\n",
        "    \r\n",
        "    fig, axes = plt.subplots(nrows=1, ncols=3, figsize=(15,4))\r\n",
        "    axes[0].plot(history.history['loss'])   \r\n",
        "    axes[0].plot(history.history['val_loss'])\r\n",
        "    axes[0].legend(['loss','val_loss'])\r\n",
        "    \r\n",
        "    axes[1].plot(history.history['accuracy'])   \r\n",
        "    axes[1].plot(history.history['val_accuracy'])\r\n",
        "    axes[1].legend(['accuracy','val_accuracy'])\r\n",
        "\r\n",
        "    axes[2].plot(history.history['auc'])   \r\n",
        "    axes[2].plot(history.history['val_auc'])\r\n",
        "    axes[2].legend(['auc','val_auc'])"
      ],
      "outputs": [],
      "execution_count": 27,
      "metadata": {
        "collapsed": true,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "gather": {
          "logged": 1621081397418
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "\r\n",
        "After deeply examining our data and building some preliminary models, we are finally ready to build a model that will perform our prediction task."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "from keras import backend as kb\n",
        "class DecayLR(tf.keras.callbacks.Callback):\n",
        "    \"\"\" Decay: Decay helps in applying decaying Learning rate \"\"\"\n",
        "    def __init__(self, base_lr=0.01, decay_epoch=1):\n",
        "        super(DecayLR, self).__init__()\n",
        "        self.base_lr = base_lr\n",
        "        self.decay_epoch = decay_epoch \n",
        "        self.lr_history = []\n",
        "        \n",
        "    def on_train_begin(self, logs={}):\n",
        "        kb.set_value(self.model.optimizer.lr, self.base_lr)\n",
        "\n",
        "    def on_epoch_end(self, epoch, logs={}):\n",
        "        new_lr = self.base_lr * (0.5 ** (epoch // self.decay_epoch))\n",
        "        self.lr_history.append(kb.get_value(self.model.optimizer.lr))\n",
        "        kb.set_value(self.model.optimizer.lr, new_lr)"
      ],
      "outputs": [],
      "execution_count": 28,
      "metadata": {
        "gather": {
          "logged": 1621081397687
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "training_generator = AugmentedDataGenerator('train', ablation=None)\n",
        "validation_generator = AugmentedDataGenerator('val', ablation=None)\n",
        "\n",
        "# Applying class weights to treat imbalanced data, every image of class 1 is treated as 10 images of class 0\n",
        "class_weight = {0: 1., 1: 10.}\n",
        "\n",
        "# Defining callbacks\n",
        "auc_logger = roc_callback()\n",
        "os.makedirs('models', exist_ok=True)\n",
        "filepath = 'models/best_model.hdf5'\n",
        "checkpoint = ModelCheckpoint(filepath, monitor='val_auc', verbose=1, save_best_only=True, mode='max')\n",
        "ES = tf.keras.callbacks.EarlyStopping(monitor='val_auc', patience=5, verbose=1, mode='max')\n",
        "decay = DecayLR()\n",
        "\n",
        "# Fit the model\n",
        "history = cnn.fit(training_generator, epochs=10, validation_data=validation_generator, callbacks=[auc_logger, decay, ES, checkpoint], class_weight=class_weight)"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "27/27 [==============================] - ETA: 0s - loss: 1.3503 - accuracy: 0.3935 - auc: 0.5425\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Val AUC for epoch0: 0.6526162790697675\n",
            "\n",
            "Epoch 00001: val_auc improved from -inf to 0.59435, saving model to models/best_model.hdf5\n",
            "27/27 [==============================] - 22s 797ms/step - loss: 1.3503 - accuracy: 0.3935 - auc: 0.5425 - val_loss: 0.7256 - val_accuracy: 0.1719 - val_auc: 0.5943\n",
            "Epoch 2/10\n",
            "27/27 [==============================] - ETA: 0s - loss: 1.2537 - accuracy: 0.5347 - auc: 0.6679\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Val AUC for epoch1: 0.6487394957983194\n",
            "\n",
            "Epoch 00002: val_auc improved from 0.59435 to 0.68808, saving model to models/best_model.hdf5\n",
            "27/27 [==============================] - 23s 838ms/step - loss: 1.2537 - accuracy: 0.5347 - auc: 0.6679 - val_loss: 0.6291 - val_accuracy: 0.8021 - val_auc: 0.6881\n",
            "Epoch 3/10\n",
            "27/27 [==============================] - ETA: 0s - loss: 1.2350 - accuracy: 0.5845 - auc: 0.6763\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Val AUC for epoch2: 0.6947637292464879\n",
            "\n",
            "Epoch 00003: val_auc did not improve from 0.68808\n",
            "27/27 [==============================] - 18s 682ms/step - loss: 1.2350 - accuracy: 0.5845 - auc: 0.6763 - val_loss: 0.7061 - val_accuracy: 0.5156 - val_auc: 0.6531\n",
            "Epoch 4/10\n",
            "27/27 [==============================] - ETA: 0s - loss: 1.1980 - accuracy: 0.5579 - auc: 0.7215\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Val AUC for epoch3: 0.6136653895274584\n",
            "\n",
            "Epoch 00004: val_auc improved from 0.68808 to 0.69764, saving model to models/best_model.hdf5\n",
            "27/27 [==============================] - 21s 760ms/step - loss: 1.1980 - accuracy: 0.5579 - auc: 0.7215 - val_loss: 0.6810 - val_accuracy: 0.5938 - val_auc: 0.6976\n",
            "Epoch 5/10\n",
            "27/27 [==============================] - ETA: 0s - loss: 1.1878 - accuracy: 0.5880 - auc: 0.7392\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Val AUC for epoch4: 0.6615767045454546\n",
            "\n",
            "Epoch 00005: val_auc did not improve from 0.69764\n",
            "27/27 [==============================] - 19s 688ms/step - loss: 1.1878 - accuracy: 0.5880 - auc: 0.7392 - val_loss: 0.6946 - val_accuracy: 0.5677 - val_auc: 0.6125\n",
            "Epoch 6/10\n",
            "27/27 [==============================] - ETA: 0s - loss: 1.1752 - accuracy: 0.5822 - auc: 0.7373\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Val AUC for epoch5: 0.6625159642401022\n",
            "\n",
            "Epoch 00006: val_auc did not improve from 0.69764\n",
            "27/27 [==============================] - 18s 680ms/step - loss: 1.1752 - accuracy: 0.5822 - auc: 0.7373 - val_loss: 0.6598 - val_accuracy: 0.6510 - val_auc: 0.6639\n",
            "Epoch 7/10\n",
            "27/27 [==============================] - ETA: 0s - loss: 1.1622 - accuracy: 0.6146 - auc: 0.7428\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Val AUC for epoch6: 0.6588235294117648\n",
            "\n",
            "Epoch 00007: val_auc did not improve from 0.69764\n",
            "27/27 [==============================] - 19s 690ms/step - loss: 1.1622 - accuracy: 0.6146 - auc: 0.7428 - val_loss: 0.6513 - val_accuracy: 0.6667 - val_auc: 0.6627\n",
            "Epoch 8/10\n",
            "27/27 [==============================] - ETA: 0s - loss: 1.1680 - accuracy: 0.6308 - auc: 0.7569\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
            "Val AUC for epoch7: 0.6924247033769394\n",
            "\n",
            "Epoch 00008: val_auc did not improve from 0.69764\n",
            "27/27 [==============================] - 18s 681ms/step - loss: 1.1680 - accuracy: 0.6308 - auc: 0.7569 - val_loss: 0.6575 - val_accuracy: 0.6562 - val_auc: 0.6600\n",
            "Epoch 00008: early stopping\n"
          ]
        }
      ],
      "execution_count": 29,
      "metadata": {
        "gather": {
          "logged": 1621081560341
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "plot(history)"
      ],
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 1080x288 with 3 Axes>",
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA2oAAAD4CAYAAACdbRXeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOzdd3hUVfrA8e9NJj0hvRBCIBBqCCFUAelSdFFEpYmNtay9rX0t7IqrP127rIIuVhQRxIIIgoAoSBeBQAghtABJJgkJmYSUmTm/P+4QQiAFmMlMkvfzPHmm3PZOuNzMe88579GUUgghhBBCCCGEcB1uzg5ACCGEEEIIIcSZJFETQgghhBBCCBcjiZoQQgghhBBCuBhJ1IQQQgghhBDCxUiiJoQQQgghhBAuxuCsA4eFham2bds66/BCCAfYsmVLrlIq3NlxXCy5PgnR9DSF65Ncm4Roemq7NjktUWvbti2bN2921uGFEA6gadpBZ8dgD3J9EqLpaQrXJ7k2CdH01HZtkq6PQgghhBBCCOFiJFETQgghhBBCCBcjiZoQQgghhBBCuBinjVETwhVVVFSQmZlJaWmps0Nxad7e3sTExODh4eHsUIRwOrluuJbmdn2S869uze2cEE2HJGpCVJGZmUlAQABt27ZF0zRnh+OSlFLk5eWRmZlJXFycs8MRwunkuuE6muP1Sc6/2jXHc0I0HdL1UYgqSktLCQ0NlT92tdA0jdDQULl7K4SNXDdcR3O8Psn5V7vmeE6IpkMSNSGqkT92dXOV35GmaWM0TdujaVq6pmlPnGN5rKZpqzRN+0PTtO2apl3hjDhF0+cq/ydE8/y3aI6f+XzI70c0Vi6fqOUUlfL84l2UVlicHYoQwoVomuYOzAQuB7oCUzRN61pttaeB+UqpZGAy8N+GjdLBdn0HBYecHYUQQgghgNIKC19vzeTLTfb52+zyidqvabn877f9TJ69npwiabYWTZ+/v7+zQ2gs+gLpSqkMpVQ5MA8YV20dBbSwPQ8EjjZgfI5VXgJf3QzLn3V2JEIIIUSzlmE0MWPxLi558Wcenv8nC7ccsct+Xb6YyLW9YvDzcuehL/9k3Dtref+m3nRrFejssIQQztcKOFzldSbQr9o604GfNE27D/ADLqtpZ5qm3QHcARAbG2vXQB0iNw2UFfb8CKUnwLtF3dsIcRHMZjMGg8t/bRBCiAZRbrayfFc2czccZN2+PAxuGqMTopjaL5b+7UPtcgyXb1EDGNOtJQvu6o8GXPfeOpbsOObskIRwOKUUjz76KN26dSMxMZEvv/wSgGPHjjF48GB69OhBt27d+PXXX7FYLNxyyy2V677++utOjt5lTAE+UkrFAFcAn2qads7rnlJqtlKqt1Kqd3h4eIMGeUFy0/RHcymkLnZuLMLprr76anr16kVCQgKzZ88GYOnSpfTs2ZOkpCRGjBgBgMlkYtq0aSQmJtK9e3cWLlwInNmSv2DBAm655RYAbrnlFu6880769evHY489xsaNG+nfvz/JyckMGDCAPXv2AGCxWHjkkUfo1q0b3bt35+2332blypVcffXVlftdvnw548ePb4hfh3CCc52DNZ1X2dnZjB8/nqSkJJKSkli3bp0zQhbighzOL+GVZakMeGkl93y+lYN5JTw6uhPrnhzOzKk9GRAfZrdxkY3m1lhCdCDf3nspf/t0M3fP3cpDl3Xk/hHxMkBUOMw/v09h19ETdt1n1+gWPHdlQr3W/frrr9m2bRt//vknubm59OnTh8GDB/P5558zevRo/vGPf2CxWCgpKWHbtm0cOXKEnTt3AlBQUGDXuF3UEaB1ldcxtvequhUYA6CU+l3TNG8gDMhpkAgdyZgKmju0aAXb50OP650dkcB51405c+YQEhLCyZMn6dOnD+PGjeP2229nzZo1xMXFkZ+fD8Dzzz9PYGAgO3bsAOD48eN1Hj8zM5N169bh7u7OiRMn+PXXXzEYDKxYsYKnnnqKhQsXMnv2bA4cOMC2bdswGAzk5+cTHBzM3XffjdFoJDw8nA8//JC//vWvF/8LcSGapo0B3gTcgQ+UUi9VW/46MMz20heIUEoF2ZZZgB22ZYeUUlddbDzO/LtV/Ry89tpra1z3/vvvZ8iQISxatAiLxYLJZLJnyELYncWqWJWaw9wNB1mdZkQDhneOYGq/NgzuGI67m2PykUaTqAGEB3jxxR2X8OTXO3h9RRppOUX857okfDzdnR2aEHb322+/MWXKFNzd3YmMjGTIkCFs2rSJPn368Ne//pWKigquvvpqevToQbt27cjIyOC+++7jL3/5C6NGjXJ2+A1hE9BB07Q49ARtMlA9WzkEjAA+0jStC+ANGBs0Skcx7oGQdpBwNfz6KhRlQUCUs6MSTvLWW2+xaNEiAA4fPszs2bMZPHhw5bxRISEhAKxYsYJ58+ZVbhccHFznvidMmIC7u/53trCwkJtvvpm9e/eiaRoVFRWV+73zzjsru0aeOt6NN97IZ599xrRp0/j999/55JNP7PSJna9KQaOR6F2vN2ma9p1SatepdZRSD1VZ/z4gucouTiqlejRUvI5W/Rzcu3dvjeuuXLmy8lxwd3cnMFCGtAjXlH2ilC83HWbexkMcLSwlIsCL+4bFM6lvLK2CfBx+/EaVqAF4Gdx5dUISnaMCePHHVA7mFfP+Tb1pGej4X5ZoXurb8tXQBg8ezJo1a/jhhx+45ZZbePjhh7npppv4888/WbZsGe+99x7z589nzpw5zg7VoZRSZk3T7gWWod/NnqOUStE07V/AZqXUd8Dfgfc1TXsIvbDILUop5byo7Sg3DcI7QeJEWPMK7FwI/e9xdlTNnjOuG6tXr2bFihX8/vvv+Pr6MnToUHr06EFqamq991G1d0r1+ab8/Pwqnz/zzDMMGzaMRYsWceDAAYYOHVrrfqdNm8aVV16Jt7c3EyZMaGpj3CoLGgFomnaqoNGuGtafAjznyICc9XfrXOdgaWlpreeVEK7KalWs3ZfL3PWHWL47G4tVMahDGM9e2ZURXSLxcG+4kWONYoxadZqmccfg9nxwU28O5JZw1Ttr+eNQ3d03hGhMBg0axJdffonFYsFoNLJmzRr69u3LwYMHiYyM5Pbbb+e2225j69at5ObmYrVaufbaa5kxYwZbt251dvgNQim1RCnVUSnVXin1gu29Z21JGkqpXUqpgUqpJKVUD6XUT86N2E7M5ZC3T0/UwjtCyyS9+6NolgoLCwkODsbX15fU1FTWr19PaWkpa9asYf/+/QCVXR9HjhzJzJkzK7c91fUxMjKS3bt3Y7VaK1tFajpWq1atAPjoo48q3x85ciSzZs3CbDafcbzo6Giio6OZMWMG06ZNs9+Hdg3nKmjU6lwraprWBogDVlZ521vTtM2apq3XNO3qc23XWJzrHISaz6sRI0bw7rvvAvr4xsLCQqfELURVeaYyZv2yj2GvrubG/21kw/48brs0jtWPDOXTW/sxplvLBk3SoJEmaqeM6BLJ13cPwNvDjUmz1/PNH/YphSmEKxg/fjzdu3cnKSmJ4cOH8/LLLxMVFcXq1atJSkoiOTmZL7/8kgceeIAjR45U3kW/4YYbePHFF50dvnCk/AxQFgjrpL9OnAjHtkFuzV2NRNM1ZswYzGYzXbp04YknnuCSSy4hPDyc2bNnc80115CUlMSkSZMAePrppzl+/DjdunUjKSmJVatWAfDSSy8xduxYBgwYQMuWLWs81mOPPcaTTz5JcnJyZVIGcNtttxEbG1t5zfr8888rl02dOpXWrVvTpUsXB/0GGoXJwAKlVNVJYdsopXqjd9l+Q9O09ufaUNO0O2wJ3Waj0TV7bp/rHISaz6s333yTVatWkZiYSK9evdi1q6ZGSCEcSynFxv35PDDvD/q/uJIXf0wlMsCbNyf3YP1TI3jyii60DfOre0cOojmrF1Dv3r3V5s2b7bKv/OJy7vpsCxv253P30PY8MqoTbg4a1Ceatt27dzf3LxP1dq7flaZpW2xfPBo1e16fHCLlG30OtTt+gegecOIYvNYFBj8Kw//h7OiaHblu1O7ee+8lOTmZW2+9tcGO2RDXJ03T+gPTlVKjba+fBFBKnXWnTNO0P4B7lFLnLG+oadpHwGKl1ILajnmua5Ocf/UjvydRVeHJChZtzWTuhkPszTER4G3g2p4xXN8vlo6RAQ0aS23XpibRWTzEz5NPb+3Hc9/t5L+r97E3x8Trk3rg79UkPp4QQpzpVGn+sA76Y4uWEDcYdsyHYU+BVMMVLqJXr174+fnx6quvOjsUR6hPQSM0TesMBAO/V3kvGChRSpVpmhYGDARebpCohWimlFJszyxk7oaDfPfnUUorrCTFBPLytd0Zm9QSX0/XyxtcL6IL5Glw49/jE+kUGcC/Fu/iunfX8f5NvWkd4uvs0IQQwr6MeyAwFjyrdMfoPhG+vQcyN0PrPs6LTYgqtmzZ4uwQHKaeBY1AT+DmVStk1AWYpWmaFX0YyktVq0UKIeynuMzMd38eZe6Gg+w8cgJfT3fGJ7fi+r5tSIxx7YqjdSZqmqbNAcYCOUqpbudYPg54HrACZuBBpdRv9g60PjRN45aBcbQL9+eez7cybuZaZt3Yiz5tQ5wRjhBCOIZxj15IpKouV8Lih/VWNUnUhGgQSqklwJJq7z1b7fX0c2y3Dkh0aHBCNHOpWSeYu/4Qi/44gqnMTOeoAJ4fl8C45Fa08PZwdnj1Up8WtY+Ad4CaJj/5GfhOKaU0TesOzAc62ye8CzO4Yzjf3DOQ2z/ezPXvr+eFqxOZ2Kd13RsKIYSrs1ogby+0G3Lm+96B0Oly2Pk1jP43uDeOP0JCCCHExbJaFUcKTpKeYyI9x8TSlCy2HDyOp8GNsYktmXpJLD1jg8+YMqIxqDNRU0qt0TStbS3Lq04n74c+V5HTtQ/3Z9HdA7n3i608tnA7adlFPHlFF4fNHC6EEA2i4CCYS89uUQO9++OubyBjNXQY2eChCSGEEI5UWmFhf24x6Tkm9hlNtsdiMowmyszWyvXahfnx9F+6cG3PGIL9PJ0Y8cWxyxg1TdPGAy8CEcBfalnvDuAOgNjYWHsculaBvh58eEsfZvywmw9+28/eHBNvX5/caJo7hRDiLMZThUTOkajFjwTvIH1ONUnUhBBCNFLHi8tJN5rYl2OqTMr2GYs5fLyEU6M9NQ1ign2ID/dnYPtQ4iP8aR/hT/twf0IacXJWlV0SNaXUImCRpmmD0cerXVbDerOB2aCXmLXHseticHdj+lUJdIwM4NlvdzJ+5lr+d3Mfp86JIIQQFyx3j/4Y3vHsZQZPSLhaT9TKTODl37CxCSGEcLqScjNHC07i52XA38uAn6fBJaetquyuaEvI9hlN7MspJt1oIr+4vHI9L4Mb7cL96R4TyPjkVnpCFu5Pu3A/vD3cnfgJHM+uVR9t3STbaZoWppTKtee+L9b1/WKJC/PjrrlbGDdzLe9O7cmA+DBnhyXERfH398dkMp1z2YEDBxg7diw7d+5s4KiEQxn3gH8k+ASfe3niRNjyEexZoneFFKKa2q4bQjianH+OU1Ju5uN1B5m1Zh8FJRVnLPPzdMffW0/c/L098Pdy15972Z572557G2zLPPD3MhDgbahM+AK8DXgZ3M57nNep7opVuyqm55jYn2uitOJ0d8VgXw/iI/wZ1TWyMhmLj/AnOsin2Q5duuhETdO0eGCfrZhIT8ALyLvoyBygf/tQvrvnUm79eBM3ztnI9KsSuPGSNs4OSwgh6s+4B8LO0Zp2Smx/aBGjt6pJoiZcmNlsxmBoMrMECeE0J8stzN1wkHdX7yOvuJyhncIZ1yOa0gorxWVmikrNmMrMmErNmMptj2VmcotK9PdtPxZr3Z3dDG7aGYnbqef+3gYCvE6/Lik3VyZk1bsrtgryIT7CnwGnuivaErKm0l3RnupTnv8LYCgQpmlaJvAc4AGglHoPuBa4SdO0CuAkMKnaXCEuJTbUl6/vHsAD87bxzDc7Scsq4tkru+Lh7ubs0ISr+fEJyNph331GJcLlL9W4+IknnqB169bcc889AEyfPh2DwcCqVas4fvw4FRUVzJgxg3Hjxp3XYUtLS7nrrrvYvHkzBoOB1157jWHDhpGSksK0adMoLy/HarWycOFCoqOjmThxIpmZmVgsFp555hkmTZp0UR9b2IlSeqKWNLnmddzcIPE6WPc2mIzgH95w8YlGf90wmUyMGzfunNt98skn/Oc//0HTNLp3786nn35KdnY2d955JxkZGQC8++67REdHn9Ga/5///AeTycT06dMZOnQoPXr04LfffmPKlCl07NiRGTNmUF5eTmhoKHPnziUyMhKTycR9993H5s2b0TSN5557jsLCQrZv384bb7wBwPvvv8+uXbt4/fXXL+rX26Q00fOveg+RqudUeno6d955J0ajEXd3d7766ivat29vn8/u4korLHyx8RD/Xb0PY1EZl8aH8dDIDvRqc/7TUimlKK2wUlRWQXGZBVOpmaKyCkylZoptyV2RLdkrLjv93FRmpqCknMPHSypfl5Rb8DS40S7Mj0Rbd8X2Ef7Eh/sTF+aHj2fT7q5oT/Wp+jiljuX/B/yf3SJqAAHeHrx/U29eXprKrDUZZOSamHl9T4J8JZMXzjVp0iQefPDByj948+fPZ9myZdx///20aNGC3NxcLrnkEq666qrz6nowc+ZMNE1jx44dpKamMmrUKNLS0njvvfd44IEHmDp1KuXl5VgsFpYsWUJ0dDQ//PADAIWFhQ75rOICFB2D8qJzV3ysqvtEWPsGpCyCfnc0TGzCaex53fD29mbRokVnbbdr1y5mzJjBunXrCAsLIz8/H4D777+fIUOGsGjRIiwWCyaTiePHj9d6jPLycjZv3gzA8ePHWb9+PZqm8cEHH/Dyyy/z6quv8vzzzxMYGMiOHTsq1/Pw8OCFF17glVdewcPDgw8//JBZs2Zd7K9PXKSGOP9qM3XqVJ544gnGjx9PaWkpVqu11vWbgjKzhfmbM5m5Mp2sE6X0iwvhnSnJ9GsXesH71DQNH093PYkKuLj4TrXMNdfuivbUbPscuLtpPHlFFzpEBvDU1zu4euZaPri5D/ERMvhe2NRyB9FRkpOTycnJ4ejRoxiNRoKDg4mKiuKhhx5izZo1uLm5ceTIEbKzs4mKiqr3fn/77Tfuu+8+ADp37kybNm1IS0ujf//+vPDCC2RmZnLNNdfQoUMHEhMT+fvf/87jjz/O2LFjGTRokKM+rjhfRlshkdq6PgJEJkBEgj75tSRqDauRXzeUUjz11FNnbbdy5UomTJhAWJg+tjskRL9jv3LlSj75RJ9m1d3dncDAwDoTtaot9JmZmUyaNIljx45RXl5OXFwcACtWrGDevHmV6wUH62Myhw8fzuLFi+nSpQsVFRUkJsqc0WdooudfTYqKijhy5Ajjx48H9ESvKauwWFm4JZO3V6ZzpOAkvdoE89rEJPq3D3Wp+cEkQbOfZpuonXJdrxjiwnz526dbGP/ftbw9JZmhnSKcHZZoxiZMmMCCBQvIyspi0qRJzJ07F6PRyJYtW/Dw8KBt27aUlpba5VjXX389/fr144cffuCKK65g1qxZDB8+nK1bt7JkyRKefvppRowYwbPPPmuX44mLdCpRC+9c97rdJ8CK6ZCfASHtHBqWcD57XTfscb0xGAxntGpU397P73TV5fvuu4+HH36Yq666itWrVzN9+vRa933bbbfx73//m86dOzNt2rTziks4jqPPv7rOqabObLGy6I8jvLVyL4fzT5LUOogXr0lkUIcwl0rQhP3JwCygV5sQvr33UmKCffnrR5v432/7ceFhdqKJmzRpEvPmzWPBggVMmDCBwsJCIiIi8PDwYNWqVRw8ePC89zlo0CDmzp0LQFpaGocOHaJTp05kZGTQrl077r//fsaNG8f27ds5evQovr6+3HDDDTz66KNs3brV3h9RXKjcPeAdCP71uJnU7Tr9cccCx8YkXIK9rhs1bTd8+HC++uor8vL0WmGnuj6OGDGCd999FwCLxUJhYSGRkZHk5OSQl5dHWVkZixcvrvV4rVq1AuDjjz+ufH/kyJHMnDmz8vWpVrp+/fpx+PBhPv/8c6ZMqXVkhmhAjj7/ajqnAgICiImJ4ZtvvgGgrKyMkpISx3xIJ7BYFd/8cYSRr6/h0QXbCfTxYM4tvfnm7gEM7hguSVozIImaTasgHxbc2Z+RXSN5fvEunvx6B+Xmpt/PWbiehIQEioqKaNWqFS1btmTq1Kls3ryZxMREPvnkEzp3rkdrSjV33303VquVxMREJk2axEcffYSXlxfz58+nW7du9OjRg507d3LTTTexY8cO+vbtS48ePfjnP//J008/7YBPKS6IcY/emlafP85BraHNQL36o9x4avLsdd2oabuEhAT+8Y9/MGTIEJKSknj44YcBePPNN1m1ahWJiYn06tWLXbt24eHhwbPPPkvfvn0ZOXJkrceePn06EyZMoFevXpXdKgGefvppjh8/Trdu3UhKSmLVqlWVyyZOnMjAgQMru0MK53P0+VfbOfXpp5/y1ltv0b17dwYMGEBWVpZDPmNDsloVi7cfZfQba3jwy214GdyYdWMvvr/3UoZ3jpQErRnRnNVy1Lt3b3VqMLErsVoVb6xI462V6fRtG8K7N/Qk1N/L2WGJBrJ79266dOni7DAahXP9rjRN26KU6u2kkOzGVa9PvNweOl0O496p3/pbPoLvH4A7VkN0sgMDa97kutGwxo4dy0MPPcSIESNqXKepXp/OdW2S869+GsPvSSnFspQsXl++lz3ZRXSI8OehkR0ZkxDlkhNWC/uo7dokLWrVuLlpPDyqE29NSebPzALGzVxLatYJZ4clhGjuivOgJLfuio9VdR0H7p6w/SvHxSVEAykoKKBjx474+PjUmqQJ0dgopVixK5uxb//GnZ9tpcJi5c3JPVj64GCuSGwpSVoz1uyLidTkqqRo2oT4cvsnm/nLW7/RJtS3ckK+eNtj+wh//L3kVyica8eOHdx4441nvOfl5cWGDRucFJFwiNzzKCRyik8wdBgFOxfAqOfBTeauEbrGeN0ICgoiLS3N2WEIO2iM558jKKX4Jc3I68vT+DOzkDahvrw2MYmrkqIxyPy+AknUapXUOojv77uUuesPkpZtIt1oYlVqDuYqM7e3DPSunFW9fZUkLszfU/oQN1JKqUb1b5eYmMi2bdsa9JhSbMcJ6luav7rECZC6GPb/Au2H2z8uAch1w5U0x+uTnH+1c7VzQinF2vQ8Xlu+h62HCogJ9uHla7szvmcrPCRBE1VIolaHyBbePDzqdFejCouVg3klpOeY2Gc0VT7O33yYknJL5XqBPh6VrW/tI/xszwOICfaRJmwX5u3tTV5eHqGhrjUniStRSpGXl9fk56txOcY94OELga3Pb7uOY8Crhd79URI1h5DrhutojtcnOf9q52rnxPqMPF5bnsbG/flEB3rz7/GJXNcrBk+DJGjibJKonScPdzc96ao2MbZSimOFpaTn6Mlbui2J+zk1my83l1eu52Vwo12VLpSnkri4MD+8DNItydliYmLIzMzEaDQ6OxSX5u3tTUxMjLPDQNO0McCbgDvwgVLqpWrLXweG2V76AhFKqaCGjdJOcvdAWAdwO88/5h7e0OUq2PUtjH0NPHwcE18zJtcN1+Iq16eGIudf3VzhnNhyMJ9Xf0pj3b48IgK8+Ne4BCb1aS3f/UStJFGzE03TiA7yITrIh8Edw89Ydry4vLL17VQS98eh4yzefrSyarabBrEhvpVj3+JPdaWM8KeFt4cTPlHz5OHhQVxcnLPDEPWgaZo7MBMYCWQCmzRN+04ptevUOkqph6qsfx/QeEsfGtOg7cAL27b7BNj2Gez5EbpdY9+4hFw3hFPJ+efath0u4LXlaaxJMxLm78kzY7sytV8s3h6SoIm6SaLWAIL9POntF0LvtiFnvH+y3EJGrq37ZJVWuF/SjFRYTvenbhXkw/DOEYxOiKJfuxDpvyyEri+QrpTKANA0bR4wDthVw/pTgOcaKDb7KiuCE5nnPz7tlLaDwD8KdnwliZoQQjSAnUcKeX15Gj+n5hDs68GTl3fmxv5t8PWUr96i/uRscSIfT3cSogNJiA48432zxcrh4ycrW+D+OHScr7Yc5tP1Bwn08WBElwjGJEQxuGO43JERzVkr4HCV15lAv3OtqGlaGyAOWFnTzjRNuwO4AyA2NtZ+UdpDrq3S3fmU5q/KzR0Sr4MNs6AkH3xD6t5GCCHEBXllWSozV+0j0MeDR0d34uYBbaVKuLggcta4IIO7G3FhfsSF+TGyaySgt779kmbkp5QsVuzK5uutR/DxcGdIx3DGdItiWOcIAn2ki6QQNZgMLFBKWWpaQSk1G5gN+qSyDRVYvRgvoDR/dYkT4Pd3YNc30Puv9olLCCHEGZbsOMbMVfu4tmcMz13VVYaviIsiiVoj4ePpzphuUYzpFkWFxcr6jDyWpWTxU0o2S1Oy8HDX6N8+jNEJkYzsGklEgGtUNxLCgY4AVUsgxtjeO5fJwD0Oj8hRjHvAzQOCL2IcSsskvevk9q8kURNCCAfIMJp4bMF2kmODePGaRKnkKC6aJGqNkIe7G4M6hDOoQzj/uqobfxwu4KeULJamZPGPRTt5+pud9IoNZnRCFKMToogN9XV2yEI4wiagg6ZpcegJ2mTg+uoraZrWGQgGfm/Y8OwoNw1C48H9Ii7ZmgaJE2HVDCg4BEEu1r1TCCEasZPlFu6euxUPd42Z1/eUJE3YhZxFjZybm0avNsE8eUUXVj8ylKUPDuLBER0pLrfwwpLdDH5lFZe/+StvrthLatYJl5v0UYgLpZQyA/cCy4DdwHylVIqmaf/SNO2qKqtOBuapxnzyG1Mh/AILiVSVeJ3+uGPBxe9LCCEEoE/R9I9vdrAnu4g3JicTHSTToAj7kBa1JkTTNDpHtaBzVAseuKwDh/JKWJaSxbKULN74OY3XV6TRJtSXMQlRjEqIIrl1kEy+LRo1pdQSYEm1956t9np6Q8ZkdxWlcPwAdLvu4vcVEget+8H2+XDpQ3ormxBCiIvy5abDfL31CMDBxlIAACAASURBVA+M6MCQalM0CXEx6kzUNE2bA4wFcpRS3c6xfCrwOKABRcBdSqk/7R2oOH+xob7cPrgdtw9uR05RKct3ZbMsJZs5a/cza00GEQFejEqIZHRCFJe0C5Wy/0K4orx0UNYLr/hYXeIEWPIIZO+EqET77FMIIZqpnUcKefa7FAZ1COP+ER2cHY5oYurTovYR8A7wSQ3L9wNDlFLHNU27HL1q2jlLZAvniQjwZmq/Nkzt14bCkxWsSs1hWUoWC7cc4bP1h2jhbeCyLpGMSohiSMdwfDyl7L8QLiH3VMVHOyVqCdfA0if0VjVJ1IQQ4oIVnqzg7rlbCfXz5I1JPXCXXkrCzupM1JRSazRNa1vL8nVVXq5Hr7wmXFigjwdXJ7fi6uRWlFZYWJNmZFlKNit2Z/P1H0fw9nCrLPs/vHOklP0XwpmMaaC56cVE7MEvFNqPgJ0L4bJ/gpu0pAshxPlSSvH3+X9ytOAkX/6tP6H+Xs4OSTRB9h6jdivwY00LXXpC2WbK28OdUbYxaxUWKxv351eW/V+Wko3BTSM5NoiWgT6E+nsS5u9FmL8noX5ehAV4EeqnvyctcEI4iDEVgtqAhx0Hp3efCAuXwcG1EDfIfvsVzYvVok+mLkQzNHtNBit2Z/Ps2K70ahPs7HBEE2W3RE3TtGHoidqlNa3j0hPKCjzc3RgYH8bA+DCmX5nAn5kFLEvJZvOBfLZnFpBnKqeozHzObf083Qn19zo7mfP3rHw/3N+LUH8vgnw8pIiJEPWVm2a/bo+ndLocPPxgx3xJ1ETtlIKiY5C7F/L26o+nnhccBk8/CIiCgJa2n6gzH1u0BP8o8JC5PUXTsSEjj5eX7eEviS2ZNrCts8MRTZhdEjVN07oDHwCXK6Xy7LFP4VxubhrJscEkx555l6i0wkJecTl5pjJyTWXkmsrJNZWRZzr1XjmH80vYdriA/OJyLNaz83F3N40QP09C/TwJt7XKhfp7EVaZ6HnanuvLvD3kjq1opixm/Utxh5H23a+nH3QZC7u+hSv+AwbpstPslZdA/j79xkBu+umkLC8dyk2n1/Pwg7B4iOkL3SdBebGeyBVlweEN+qOl7Oz9+wRXS+TO8dw/Atylq71wbTlFpdz7xR+0CfHlpWsT0aR6rnCgi07UNE2LBb4GblRKpV18SMKVeXu40yrIh1b1mCPEalUUnKywJXR6MleZ1BWXYSzSHw/kFZNnKqek3HLO/QR4G4hq4U1UoDdRLbxpGehNZKDtsYU3LQN9CPb1kIulaHqOHwBrBYTZuUUN9Mmvt38Je3+CLlfaf//C9SgFJ47qyVheuq11zPa88HCVFTUIbA1hHSD2En18ZFgHCOuoJ1W1XWuVgpPH9YTtVAJXdNT2aHvPuEd/rqpf8zU9WaveKhcQBQHRp1/7hsrYSuEUZouV+7/4g6LSCj69tS8B3nJjQThWfcrzfwEMBcI0TcsEngM8AJRS7wHPAqHAf21flM1Kqd6OClg0Hm62lrMQP086RgbUuX5JufmMZC7XVEZecTk5J0rJOlFK1oky0rKNGIvKqN5Q52lwOzuZq5bUhft7YZApCERjYkzVH8M723/f7YaCX7he/VEStaalvPh0IpaXbmsl2wt5+6Ci+PR6ngF661hsfwi7SU/GQjtAaPsLHxOpaeAbov9Edq15PasFinOrJHPVkroTR+DIFig2nr2tm0HvThkQBcOehPjLLixWIc7Ta8vTWJ+Rz6sTkugc1cLZ4YhmoD5VH6fUsfw24Da7RSSaLV9PA74hBlqH+Na6ntlixWgqI6uwlKzCUo4VlpJ9Qn/MOlHKtsMFLE0ppdxsPWM7Nw3CA7wqE7qWgT6nkznbY1Sgt3S1FK7jVGn+MAfMzeNu0Ev1b/kISgvBO9D+xxCOZcrR58OrOm4sNx1OZFZZSYOgWP0canvp6dax0A56ouOsnghu7hAQqf/UxlwOxTlw4tg5krpj4C7ddkXD+Hl3Nv9dvY8pfWO5tpcUOBcNw95VH4VwOIO7Gy0DfWgZWPMdX6UUx0sqOFZ4sjKJyy48ncxlGItZty+PotKzi6ME+Xqc0ToXFehNdKCP/hjkTVSgD/5e8l9HNABjmt7ly9tBd267T4SNs2DXd9DzRsccQzjGpv/Bj4+B1XYN82pxOhkLi9cTsbCOENKucRfyMHhCYIz+I4STHM4v4aEvt5EQ3YLnrqylpVgIO5Nvm6JJ0rTT3S4TomtuKTCVmcmq0iKnP54kq7CMrBMn2XnkBLmmswfGB3gbaGlrlTvjMci78rWfJHPiYhlT7V/xsapWvSA4Th+rJola42Auh6WPw+Y50GEUDHxAT8r8I5zXOiZEE1ZaYeGuuVtQwLtTe0mvG9Gg5JukaNb8vQzER/gTH+Ff4zrlZmtlInes8KT+WGB7LCwl5WjNydwZLXEtfGgZdGbrnK+nc/4LWq2KkgoLxWVm24+F4nL9uanMTEm5pfJ9gAcuc0DXO1E7q1XvzubIBErT9Mp9v/yfXmSiRbTjjiUuXnEuzL9Jn/9u4IMw4lmZx0wIB3t+8S52HjnB+zf1Jja09qEZQtibJGpC1MHT4EbrEN9ax86VmS3knCjjaMFJsk6UcrSglKzCkxy1jaNLOVpIrqn8rO1aeBuIDvKpHDPXMvB0i9yp1jlfTwPlZisl5aeTKFOZmZIy/bG4zGxbZjm9TpkFU7mZEluyZaq2Tk0VNs8l1M9TEjVnOJGpF35wZIsa6N0ff3kJdiyAgfc79ljiwmXtgC+u18drXfMBdJ/g7IiEjaZpY4A3AXfgA6XUS9WWvw4Ms730BSKUUkG2ZTcDT9uWzVBKfdwwUYv6WPRHJnM3HOLOIe0Z2bWO8ZRCOIAkakLYgZfBvV7JXHZh2elWuaotdIUn2Xnk3Mmch7tGhaV+88O7aeDnacDPy4Cvlzv+XgZ8Pd0rW+/8vAz4e7nj62nQl1WuY8CvyvNTy/w8DbjL5OTOYbTNduKI0vxVhbaH6J765NeSqLmmXd/CojvBOwim/Qitejo7ImGjaZo7MBMYCWQCmzRN+04ptevUOkqph6qsfx+QbHsegl5JuzeggC22bY834EcQNUjLLuKpr3fSNy6ER0Z1dHY4opmSRE2IBuJlcCc21LfWrhOlFbaWucKTZBWWcrTwJEWlZvw8TydX1ZOwqsmVt4ebzCfXVJyq+OjoFjXQW9WWPgE5qRDhgKkAxIWxWvVuqb+8BDF9YNJneqVG4Ur6AulKqQwATdPmAeOAXTWsPwU9OQMYDSxXSuXbtl0OjAG+cGjEok6mMjN3frYFPy8D70xJlql9hNNIoiaEC/H2qDuZE82EMVWf2NcvzPHHSrgGlj2lt6qNeNbxxxN1KzPBor9B6mLoMRXGvg6G5luK3mJVHM4vYZ/RRHqO/rPPaOKRUZ0YEN8A/0dq1gqoOlt4JtDvXCtqmtYGiANW1rJtq3NsdwdwB0BsbOzFRyxqpZTi8YXbOZBbzNzbLiGiRSOumioaPUnUhBDCFRnTHN/t8ZSASH0C7B1fwfBnpHqgsx0/oI9HM+6G0S/CJXc1m3+T0goLGcZi0m0J2T5bQpaRW3zG3Jhh/l60D/ejfp3CXcZkYIFSqv6DhAGl1GxgNkDv3r0b2UdufD75/SA/bD/G42M60799qLPDEc2cJGpCCOFqlNJb1BLGN9wxEyfCN3fC4Q0Qe0nDHVecaf8amH8zKAtMXQDxI5wdkUMUlJSf0TKWnmMi3Wgi8/hJlC0V0TRoHexLfIQ/gzuG0z7cj/gIf9qH+xPk6+ncD3DaEaB1ldcxtvfOZTJwT7Vth1bbdrUdYxPnaeuh48z4YReXdYngb4PbOTscISRRE0IIl1NshNKChhmfdkqXsbDYB7bPl0TNGZSCTR/Aj49DaDxM+UIv9NKIKaU4Wlh6VkK2L8dEXvHpwkmeBjfahfmRFBPEtT1jaB+uT5kSF+bXGOas2gR00DQtDj3xmgxcX30lTdM6A8HA71XeXgb8W9O0YNvrUcCTjg23mbCYoewE+IbUe5P84nLunbuVyBbevDqhB25SSEu4AEnUhBDC1RhthUTCGrDSmFcAdLocUhbB5f8H7h4Nd+zmzlwOPz4KWz6CjmPgmvfBu4Wzo6q3crOVg3nFZ4wfSzeayDAWnzEVSKCPB/ER/lzWJZL2EXrrWHx4AK2CfRptdVmllFnTtHvRky53YI5SKkXTtH8Bm5VS39lWnQzMU0qpKtvma5r2PHqyB/CvU4VFxEWwmOHTq/V5KB/YBh4+dW5itSoe/HIbuaZyFt41gEBfuf4J1yCJmhBCuBpjqv4Y3sAVGLtPhJSvIf1n6DSmYY/dXJmM+iTWh9bBpQ/D8KddfhLr/OJyVuzOZlVqDnuyiziYV4LFenroVHSgN+0j/JnUJ6SydSw+wp9QP88mWZVWKbUEWFLtvWervZ5ew7ZzgDkOC645Wv0iHPhVf75jAfS8sc5N3lmVzpo0I/8en0hiTKCDAxRN3skCsFrA7+LHOEqiJoQQriY3DTwDoEV0wx63/QjwCYbtX0qi1hCObYd51+tdXa/9HyRe5+yIanS04CQ/pWSxLCWbjQfysVgVLQO9SYoJ4vJuUZWtY+3C/fDzkq8WwknSV8Cvr0KPG+DoH7BxFiTfUGsxnl/3Gnl9RRrXJLdiSt/WNa4nRL2U5MMn48DdE25dDm4XN7WDXE2FEI2WpmljgDfRuxx9oJR66RzrTASmo08o+6dS6qzxIy7HmArhHRu+0p/BUy9gsu0LKCvSu0MKx0hZBN/crSfGf10K0cnOjugs+4wmlqVksWxnFn9mFgIQH+HPnUPaMTohisRWgU2yhUw0UoVH4Os7IKILXPGKPt3I9w/Aod+hzYBzbnKs8CQPzNtGhwh/ZozvJuezuDimHPjkashLh8lzLzpJA0nUhBCNlKZp7sBMYCT6/EObNE37Tim1q8o6HdAH5w9USh3XNC3COdGeJ2MatB/unGN3nwSb58DuxdBjinNiaMqsVr1r1pqXoXU/mPipPj2CC1BKsfPICZalZLE0JYv0HBMASTGBPDq6E6MT9JYzIVyOxQwLb4WKUpjwMXj66pVslz8HG2adM1GrsFi5Z+5WyiosvHtDL3w95SuxuAgnjsEnV0HBYZg6X5/yxg7krBRCNFZ9gXSlVAaApmnzgHHArirr3A7MVEodB1BK5TR4lOfrZAGYshq24mNVrftBUKx+N1oSNfsqK4JFd+qTWCffAH95zemTWFusis0H8lmaksVPKdkcKTiJmwb94kK5oV8soxKiiA6quxiDEE618nm95eya9/XeCKAnaz1vgt9nQmEmBMacsclLP6ay9VAB71yfTPtwuQEhLkLBYfj4Sr0b+w0Loe1Au+1aEjUhRGPVCjhc5XUm0K/aOh0BNE1bi949crpSamnDhHeBctP0R2clapoGiRPgt9ehKNtlWnsavfz98MUU/d93zP9Bv785bRLrMrOFdel5LN2ZxYrd2eQVl+NpcGNQfBgPXNaBy7pEEuLnMvOUCVG7tGWw9g3oebNeEKmqPrfB7+/ovQRGnK7vsmTHMf73235uGdCWsd0beCywaFqOH9CTtJOFcOM30LqPXXcviZoQoikzAB3QJ5WNAdZompaolCqovqKmaXcAdwDExsY2ZIxnckZp/uoSJ+oD8lO+hkvucl4cdSktBIO301ul6pTxC3x1sz5X2g0Lof2wBg/BVGZm9Z4clqXo1RpNZWb8vQwM6xzBmIQohnQKx1+KgIjGpjATFv0NIhP1aUWqC24Dna7Qp74Y/Bh4eJNhNPHYgu30aB3EU1d0afCQRROSm653d6wogZu/dchYY7kqCyEaqyNA1RJdMbb3qsoENiilKoD9mqaloSdum6qth1JqNjAboHfv3qr68gZjTAV3Lwhu67QQiOgMUYn65NeumKgpBRtnw7J/gJsBYvtB20EQN0T/Q+nuIn/alIKN78PSJyCsA0z+vEEnsc4vLmfFrmyWpWTxa3ou5WYroX6ejO3ektHdohjQPhQvg2tPBSBEjSwV8NU0/XHCRzXPl9bvb3p3450LOZkwmbvnbsXDXWPm1J54Gi6+2INopnJS9STNaoGbF0NUN4ccps6/ZpqmzQHGAjlKqbOi0DStM/Ah0BP4h1LqP3aPUgghzrYJ6KBpWhx6gjYZqF7R8RtgCvChpmlh6F0hMxo0yvOVm6Z/qXf2XFqJE2H5M5C3r0GTizqVFcF39+lVEzuMhpB2sH+NPkaF58HTXy8c0HYQxA2CqO4O+V0WlVaw6UA+FisEeBv0Hy8PArwN+Hsb8FBmWPJ32PoJdLwcrpndIJNYnyqjvzQli43787EqaBXkww392jCmWxS92gQ32smlhTjDz/+EzI1w3RwIi695vbaDIKIrasN7PJ2WwJ7sIj6a1pdWMvZSXKisHXoJfjcPmLbEoUMV6nPb8SPgHeCTGpbnA/cDV9spJiGEqJNSyqxp2r3AMvTxZ3OUUimapv0L2KyU+s62bJSmabsAC/CoUirPeVHXg3EPxPR2dhT6nF7Ln9Vb1YY9ec5V0nNMBPp4EB7QQF0Ps3fpk0Pn74PL/gkD7j9d/rg4V5/kdv+v+uPen/T3vQOhzaUQN1hP3MK7XHDJ5P25xaxMzWFlajYb9+dTYTl3w2sohczyeoPe2h4+87iOr/Juxv+zXfh7GQjw9rAldfpzf1uSd8Yy22s/TwNu9Uiq0nP0Mvo/pZwuo98x0p97hsUzOiGKhOgWUnZcNC17foR1b0PvW6HbtbWvq2nQ93a0xQ9x4OBK7h8+liEdwxsmTtH0HNkKn44HTz+4+XuH38isM1FTSq3RNK1tLctzgBxN0/5ix7iEEKJOSqklwJJq7z1b5bkCHrb9uL7yEig4BD2mOjsSfbLttpfq1R+HPlFZ+EIpxfqMfN5euZd1+/Sct1NkAAPiQxnYPoy+7UJo4e1h/3j+nAffP6jP7Xbz93psVfmF6XPAJYzXX584Bgd+g/2/6Inbnh/0931DT7e2xQ2B0Pgai3qUm61sPpDPz6k5rErNISO3GNCToL9eGsfQjhH4exkoKq2gqMxMUakZj5wdDN82He+KAj6P+SfrvQcTbFuWW1RSua6pzIyqo4OtpqEncDUkdR7uGuv25VWW0e/ROojHx3RmdEIk7aSKnWiqCg7p1VOjusPof9drk5Swy4lRT/No0Gr6jHjIwQGKJuvwRvjsWvAJ0v8ONcAQhQbtyO8yg/WFEMIV5e0F1Ony0s7WfaLezfDIVlSrnvy6N5e3V+5l04HjhAd48cTlnVEK1u3L5YuNh/hw7QHc3TQSWwUy0Ja49WwTjLfHRXQ9rCjVx3ht+VBvGbvufxAQVfd2LVpC9wn6D+hf7k61tu1fA7u+0d/3j7IlbYOh7SByPVqyOi2XlanZ/JqWS1GZGU+DG/3bhXLLwLYM6xRB6xDfcx9z50LYco+eDN78E9dH9zirL+4pVquipMKiJ26lZttPBSZbUmeyvS6q+rqsguPF5RzKK+FEqZmT5WZ6xAZxU/82jOoaRVSg93n/eoVoVMzl8NUt+rigCR+BR93nfOHJCu6an8qdhhFMKf0BzXRMvxElxPk48BvMnahXQr75+7Ome3CUBk3UXGawvhBCuKJTFR/DOzs3jlO6XIX64e9k/vIh9xWUsu1wAS0DvfnnVQlM6tO6MgG7a2h7yswWth4sYN2+XNam5/LeLxnMXLUPL4MbvdsGM6B9GAPjw+gW3QKDez27Hh4/oHd1PPYnXPoQDHv6wguFBMVC8lT9RynIz4ADv6L2r8GcvhqPHV8BUKrCUJauRHglcX2XofTu3oOB8aG1T4ZrtcKqF+DX/0DrS2DSp+Bf+9zqbm4a/l5661jLwAv7SEI0Oyumw5Et+qTW9ehyppTika/+5GjBSRKvfxRtwfd6qf7hTzs+VtF07FulT68SFAs3f1e/m4V24iKlsYQQQmDcA5o7hDi/eIfVqvhp30kC3HrTMe1bjvuM49/jE7m2V6tzVgr0MrjTv30o/duH8vdRnSqLbaxNz2Ntei6vLNvDK8v2EOBtoF9cqN7iFh9Ghwj/c4+f2vOjXnYbYMo86HS53T5bSYWFtdn+rDzYk1Vprck6cQ0d3I5wXch+RnilMt70BxPK1kDq25DTDvYNtnWXHHx2AlZ6Qo9zzxJIvhH+8qrrTxcgRGO0ezGsnwl974CE+pVFmL0mg+W7snl2bFcSu8XBjsth84cw6JF6tcYJQdoy+PJGvcjXjd+Af8OOb5RETQghXEXuHr2KocF5kw1brIofdhxj5sp09mQXcVPQYAZq6/h5vMLQqf5d1gO8PRjeOZLhnfUJs3NNZfy+L8/W4pbHit3ZAIQHeDGgfajtJ4zWgZ56Bce1b0DLJP3OeUjcRX+uw/klrNqTw8+7c/g9I49ysxV/LwODO4YxvHMnhnYaSZi/LcGyWiEnRe8iuf9X2Pm1Pg8T6K2dp5K2oNb6WJncvXD5y/oXSCnaIYT9HT8A39ytT78xaka9NtmQkcfLy/ZwRWIU0wa21d/sZ7upkrIIekxxWLiiidj9vT4FRGQC3LgIfEMaPIT6lOf/An2y2DBN0zKB5wAPAKXUe5qmRQGbgRaAVdO0B4GuSqkTDotaCCGaIuMeh5b5rY3ZYuXbbUeZuTqdDGMxHSL8eXNyD/7SZRi8/i6GlAXQadQF7z/M34srk6K5MkkfG3I4v4Tf9+Wx1pa4fbvtKOEUMNt3JsnWFA60nUTA1a8QGnRh/QLNFitbDxXwc6o+wXNatl5wo12YHzde0oYRnSPo3Tbk3PMoubnp88hFJUL/e8Bihqw/9aRt/xrY9jlsel9f1ycYbvwa2g29oDiFEHUwl+nj0kAfl1aPFusjBSe574s/iA3x5f+u7X661T5uiH6zZcN7kDRZbqyImu1cCAtvh1Y9YeoCvYCIE9Sn6mOttxyUUlnoE80KIYS4UOZyfdxU57ENethys5Wvt2by39X7OJRfQpeWLfjv1J6MSYg6XRq+6zjYsRDKi/WSxHbQOsSX1iG+TOzTGqUUmduWE/rjs7hXFPGkupcvUgfAS7/ROSrANr4tlH7tQvH3qvnPVkFJOb+kGfl5dw6/pBkpPFmBwU2jX7sQJvWJZXjnCOLCLiB+dwO06qX/XPqg/m91dKs+dq7jaOdOTi5EU/fTM3D0D5j0WY3/17IKS9l4IJ+N+/PYuD+ftGwT3h5ufHJrXwKqVqHVNL3l+4eHIXMTtO7bMJ9BNC7bvoBv79bHHE+dr1cbdhLp+iiEEK4gPwOs5gYrJFJaYeGrzYd575cMjhScpHtMIM+M7c1lXSLOHjOWOFGfuHnPj/r8avZktaKte5PWP/9L7/Y58XueD+vMxCOFrLN1lfxsw0HmrN2Pu5tGUkwgA+PDGNA+jJ5tgjiQW1LZarbl4HGsCsL8PRnZNZIRnSO4tEPYmV/U7MHgCbGX6D9CCMdJ+QY2zoJL7oYuVwJ6gZDD+SfZYEvKNh7I52BeCQB+nu70ahvCuB6tGNU1kg6R5/iC3X0SrPin3qomiZqobstH+lQwcYNhyhd2uzl5oSRRE0IIV5B7quKjY0vznyy38PnGQ8xes4/sE2X0ahPMC+O7MaRjeM2TIrcZCC1a6ZNf2zNRO3lcH3eyZ4k+/9lVb4NXAAYgOTaY5Nhg7hkWT2mFha0Hj1d2k5y5Kp23V6bj7qZhseoFhLu1asG9wzswvHME3VsF1muiaCGEC8vPgO/uQ0X3Ym/3R9iw/iAb9+ezaX8+WSdKAQjy9aBP2xBuvKQNfeNC6NqyHlVlvfwh+QY9ATxxTJ/KQwiADbPhx0ehwyiY+Al4+Dg7IknUhBDCJRjT9McwxyRqpjIzn60/yAe/ZpBrKueSdiG8PrEH/duH1pygneLmBt2uhfX/heI88Au9+ICObtNL7584WmchDm8PdwbEhzEgPoxHR8OJ0go2ZOSz+WA+caF+DOscQWQLqeAmRFNgtlhJPWwkauH1+FRYmXDsr+x6az0AEQFe9GsXSt+4EPrFhRAf7n9hN2X63qZfzzbPgeH/sPMnEI3S2rdg+TP68IPr5rhM9V5J1IQQwhUYUyEw1u7dLE6UVvDx2gP8b+1+CkoqGNQhjPtHdKBP2/OsXtV9Iqx7C1K+hr63X3hASuldS358HPzCYdqP0LrPee2ihbcHI7tGMrJr5IXHIYRwCWVmCzsyC9mwP5+N+/PZcvA4j1ne5ybDbp70fIquHRK5xZaYxYb41n1jqT5C2unjS7d8CIMfcZkv5cJJfnkFVs3Qe3Zc8z6427m7/EWQRE0IIVxB7h67dnssKClnzm/7+XDdAYpKzYzoHMG9w+NJjg2+sB1GdoPwLrDjqwtP1MqLYfHDsH0etB8O13xgn9Y5IUSjUVJuZuvBgsriH38cKqDMbAWgY6Q/T7fZzeRDyzH1upMXr3zccYH0+xt8Ol4v1Z802XHHEa5LKVg5A379D3SfDONm6sWjXIhrRSOEEM2R1aLPxRU35KJ3lWsq44Nf9/Pp7wcoLrcwJiGKe4fH063VhZW5r6Rp0H0C/PwvyN9//nOb5e7Vuzrm7IahT+l3sd3OnjhbCNG0FJZUsPmg3lq2YX8+O48UYrYq3DRIiA7kBtv4sj5tQwgpPQyzpkJMX/yvqN98aRes3TC9q/mG9/QCI1Kqv3lRSu/quO5t6HkTjH3DJf8mSaImhBDOVnAIzKUXNT4t50Qps9ZkMHfDQcrMVsZ2j+beYfF0irJjWeFEW6K2YwEMebT+2+38Gr67T+9edOPXemuaEKLJyios5b1f9rFhfz6pWSdQCjzcNZJigrhjcDv6xoXQq03wmRVZK07CJzfrLRoTPnR897NTpfqXPAKZm8+7C7ZoxKxWWPo4bJwNfW7Xx0m71VGExkkkURNCCGfLtRUSuYDS/EcLTvLeL/uYt+kwFqtiXI9o7hkWT/twfzsHCQTFQmx/2DFfbxGrgH7gkAAAIABJREFU6w60uVy/Y7nhPYjpq3/5CpRpN4Vo6l5YspulO4/RNy6EB0d0pG9cCMmxQXh71NJisfRJyN4B189vuOtE0hT95tPGWZKoNRdWKyx+ELZ+DP3vhVEzXLo1VRI1IYRwNmOq/ngeY9SsVsWMH3bz6foDAFzbM4a7h8YTG+rrgACrSJygTxZ77E+I7lHzeoWZ8NUt+qSyl9wNl/1Tn39MCNGkWayKX/cauSqpFa9OTKrfRjsW6IU9Bj6gF/loKJWl+t/Xv7AHRDXcsUXDs1rg23vgzy9g0CMw/GmXTtIAXLOdTwghmhNjGvhFgE/9C33MWbufOWv3Mz65FasfHcZL13Z3fJIGelUsN4NeVKQm6SvgvUGQkwoTPoYxL0qSJkQzseNIIQUlFQzuGFa/DXL3wvcPQOtLYPgzjg3uXPrcBlYzbP6w4Y8tGo6lAhbepidpw56GEc+4fJIGkqgJIYTzGVMhvFO9V995pJD/W/r/7d15fNTluffxz5UQthDWAAJhJ7hWRSlYF8AdbYu2WkVbq20VPKfWttrN1mNb2/Mc23Oe9mlPbSsutZviUrVYcauiwWop4IKCJAFcCItM2MIihCTX88c9gSEkZICZ+c0k3/fr9XvNzG9+M/eVADdzzX3f172Uc47qz08uOpZBPTO4KWfX3mEz0DcfDt9OJmqohzn/BX+6GIoGwLQX4OgLMxebiESurCKGGZxW2rf1i2u3w4NXhvWrF98TTVn0PiNDn7bgnjBdW9qeup1hhsfiR+DsWw9sjXXENPVRRCRK7mGN2rGXJHX59to6rp/5Gr0LO/KTi45NzZ5CB+ojn4Hy2fDuXBgxKZzbVg2PXAPLnw/rPj7+M+iYgRE+EckqZRUxjhnYg96FSYyiP/ktWLcYPvsX6DEo/cG1ZPw0+NNFsOSxpPviNquhAep3huSmvrbJ7c6QzCbe1tcmnEu4tkOn8IVd94Hh6HZYNDMrdu2AB6+AymdC0ZDx0zMfwyFQoiYiEqUta2FnTdKFRH44awnvVG/jz1ePp1cyH4TS4fDzoGMRLHooJGrvzwvfVm5fD5/8ZSh1nANTSkQktWp27OK1lZu4duKI1i9+Yya89kc47UYoPSv9we3PiDOgzyiYd0fbSNRWvABvPBCqCbeYZNU2n3g11KUvrsK+8eRtEHQfAEUDw233gXvud+qeuv8/arfDzMtgxYuh/P7YL6TmfTNIiZqISJQaC4kkUZr/iUVreGDBSv590khOHpnk+o90KOgCR34S3p4VCqA8d2uo0nb1szAgyeIBItLmvLxsPfUNzoTWpj3GyuFvX4ehp4Z9FaOWlwfjpsOT34SqhVByYtQRHbzqZXD/5WEaaWEx5HcKI1mNt52KwmhXfsc9t4n3m71t8h75nVp/j10fwpY1ULMGtqwOtzWrwrnNVbByHny4Yd/4Cwrjo3CJidygeIIXv1/Yt/U9z3ZugfsuhfdfgQt/Dcdfnp7fd5opURMRidLu0vz7X6NWtXE733lkEccP7snXzz74/dZS5tjPwBv3wbO3wOEfD/8RdukZdVQiEqGyyhjdOnXghKH7KYxUuy2sSyvoChfdFfZNywbHJ5TqL5kRdTQHp64W/vKlkExd+49op5MWdAlrmvsf3fI1u3aExG3LGqhZHY7E++/9IzxuOspn+aFCZ+LUysT7XfvArOth1UL49J3wkYvT+7OmUZb86xARaadiS6FzD+jWv8VL6uob+PoDr+MOv5w6hoL8LKgDNXxiqAA56MSwF42mOoq0a+5OWUWMj43ss/8+avY3Q793xSNhhCRbdCqCMZ+F+XfD2T+Copb75Kz1/K2w5nW49M/RJmnJKugMvYeHoyUNDbC9OozG7TU6tzrcj5WHqZ47a/Z+XV4BXPL7MPsjhylRExGJUqwCig/fb6LzqznLmP/uRn5+6XGZKcGfjLx8+My9UUchIlnineptVG38kOkT9rM+7bU/w+t/hgnfgpFnZC64ZH30Gpj3W1h4L0z6dtTRHJhlf4eX/xfGfhGO/ETU0aROXh506xeOgWNavm7nlr0TuX5H7n+vzxyRBV/LiogcHDObbGblZrbMzL7TzPNXmVnMzF6PH1dHEed+tVKaf/67G/jlc5V8aswgPjWmJIOBiYgkb25lNQATRrewPu2DJfDEjTDsNJi0T3edHYpHwaizYcHduVWqf2sMHv036HsknPt/oo4mGp2KwprpEZPCNNY2kKRBEomamd1jZuvM7K0Wnjcz+2X8g9IiMzsh9WGKiOzNzPKB24HzgKOAy8zsqGYufcDdj48fd2U0yNZs3xCmdLSQqG3evouvzXydkl5dufWC/czzFxGJWFlFjKF9ujK0T+G+T+7cCg9dGT5MX3R364UgojR+Omz9IBRLygUNDfDYtbBjM1x8d1gbJm1GMiNq9wKT9/P8eUBp/JgG/ObQwxIRadU4YJm7r3D3WmAmcEHEMR2YWHm4Ld43UXN3vvvYm3xQs4NfXjaGos4RbAQrIpKE2roGXlmxvvlqj+7wxA2wflkoHpLta79Gngm9R4YpkLlg3m/CtMdz/3P/hTskJ7WaqLl7GdBM/czdLgD+4ME/gZ5mlkWrQ0WkjRoErEx4XBU/19RF8dH+h81scEtvZmbTzGyBmS2IxWKpjrV5jaX5mxlRe2hBFU8sWsMN54zm+MGqpigiB2D167D4UVg5PxRdaKhPa3ML3tvA9tp6TittZtuQ1/4Iix6Aid+BERPTGkdK5OXBuGlQNT9UDcxma96AZ78Ph58PH82+mf1y6FJRTKSlD0trml5oZtMIo24MGTIkBU2LiOzX48D97r7TzKYDvweaXcHu7jOAGQBjx471jERXXRFKVPfYO39cHtvK92ct5mMj+jB9wsiMhCIiB87MJgO/APKBu9z9tmauuQT4AeDAG+5+efx8PfBm/LL33X1KygJb9AD889cJQeSH8uU9BsVLmA8Kex92HwjdS8L5wn4hSTkIcyur6ZBnfGxkn72fWPtWqPI4YhJM+MZB/zgZd/zl8PyPYN4M+PQdUUfTvNpt8PCXwl5pU36lyrttVEarPkbyQUhE2qpVQGKGUxI/t5u7r094eBfw0wzElbzYUigu3evD0c66eq6//zU6F+Tx80uPJz9P//mKZKOEdbJnE76knm9ms9x9ScI1pcBNwCnuvtHM+iW8xYfunp6KB5O+E5KNmtVhc+GaVXvur1kE5U9C3Y69X5PXIb7B8MB4Qhc/Eu8X9m02mSuriHHC0F57T9HeuSWsS+vcEz59V3avS2uqc/fw+1t4L5zzY+jWygbeUXjy22E66ZWzoLBP69dLTkpFotbqhyURkTSYD5Sa2XBCnzMVuDzxAjMb4O6No/tTgLczG2IrYhUw9OS9Tv33U+UsXl3DjCtO5LAenSMKTESSsHudLICZNa6TXZJwzTXA7e6+EcDd12Ukss494LCPhKM57qGYUc2qcGyuim8yHE/oVr0Kb/8N6nfu/br8jvGNhfckcFs69Wfg2hhnnXRCqD5YGJ/++PjXYMMKuPLx7Ex0WjNuGvxrRkjWJn4z6mj29tYjYUrpqTfA8AlRRyNplIpEbRZwXbyDGg9sTvhgJCKSFu5eZ2bXAU8Tph3d4+6LzexWYIG7zwKuN7MpQB1hre1VkQXc1M4tUFO11/q0Fyti3PXSO1xx0lDOOfqwCIMTkSQ0t/RjfJNrRgOY2T8I/dQP3P2p+HOdzWwBoX+6zd0fa66RtCwbMQujMIV9YMCxzV/jDtvX753EJd6vmg81f6WovpY7OwKvxo/8TmHPq80r4YybYdipqYk504pLQ2GRBXfDqV+D/Cwp6LTxvZAEDxoLp3836mgkzVpN1MzsfmASUGxmVcD3gQIAd/8tMBs4H1gGbAe+kK5gRUQSuftsQh+UeO6WhPs3EaYdZZ/qinAbT9Sqt+7kxgffYHT/bnzv40dGGJiIpFAHQlXsSYQZR2Vm9hF33wQMdfdVZjYCeN7M3nT35U3fILJlI2ZhdKywuOU9qRoa+I/7X2DF8gr+ePEg8rasCV9AbV4FH/kMnHpjxsJNi/HT4b5LQqn+Yy6KOhqor4NHrgFvCBU0syV5lLRpNVFz98taed6BL6csIhGR9iAWT9SKD6ehwfnGQ29Qs2MXf7p6HJ0Lcmgth0j7lczSjypgnrvvAt4xswpC4jbf3VcBuPsKM3sBGAPsk6hlswaMJ99p4OTS8eQdNSbqcFJv1NnQazjMuyM7ErWyn8LKeWHNX+/hUUcjGXBw5X1EROTQxJZCXgH0Hs7vXn6XF8pj3PzxIznisO5RRyYiydm9TtbMOhLWyTbdJfkxwmgaZlZMmAq5wsx6mVmnhPOnsPfatpzw9toaqrfuZMLoHFyDlozGUv0r58Hq16KN5d1/QNl/w3GXw7GfiTYWyRglaiIiUaiugD4jWfzBdn7y5FLOOrI/V5w0NOqoRCRJ7l4HNK6TfRt4sHGdbHxtLPHn1pvZEmAO8M14NdojgQVm9kb8/G2J1SJzRVlFNQATmts/ra0Y81koKAyl+qOyfUOY8thrGJyfXcWLJb0yWp5fRETiYuXU9Tua6+9/jZ5dC/jpxcdi2gdHJKcksU7WgRviR+I1LwMtlGTMHWUVMY44rIh+3dtwhdrOPeD4y+DVP8DZt2a+gqU7PH49bF0HX3oGOhVltn2JlEbUREQybdcO2PgOL27oxYrqbfz80uPpXdgx6qhERJK2vbaOBe9taLvTHhONmwb1tfDqvZlve+G98PbjcOZ/wKATMt++REqJmohIpm1YDt7AX6uKmD5hJKeMasPThkSkTfrnivXsqncmlLaDRK3v4TDidJh/D9Tvyly765bCUzeFtj/2lcy1K1lDiZqISIZteHcRANbvcG48Z3TE0YiIHLiyimo6F+QxdlivqEPJjPHTYcvqMLqVCbt2wMNfhI6F8Kk7QmETaXf0py4ikkH1Dc5zZXNpcOOGqedTkK9uWERyT1lFjPHD+7Sf7URKzwnFPP6VoaIiz94C6xbDhb+Bov6ZaVOyjj4hiIhk0O1zltGlZjnbC0sYepimPIpI7lm5YTsrqre1j/VpjfLy4aPXwPuvwJo30ttW+ZPwrztg/L/B6HPS25ZkNSVqIiIZsvC9DfziuUpO6PoBhSVHRR2OiMhBmVsZyvJPHN3Ovmwa8zko6JreUv01a+Cxf4fDPgJn/zB97UhOUKImIpIBmz/cxfX3v87gHgUMqFuFFR8edUgiIgelrCLGwB6dGdm3W9ShZFaXnnDcVHjzIdi2PvXv39AAj06Huh1w0T3QoVPq25CcokRNRCTN3J3vPfoma2t2cPv5vbH6Wuh7RNRhiYgcsLr6Bv6xvJoJo/u2z70fx02H+p3pKdX/8i/gnRdh8m3QV4WmRImaiEjaPbywir8tWsMNZ4/m6IK14WRfjaiJSO55feUmtuyo47T2UJa/Of2OgOETYf7dUF+XuvetWgjP/xiOuhBO+Hzq3ldymhI1EZE0WhHbyvdnLeakEb25duJIiJWHJ4r1bamI5J6yihh5Bqe25/0fx18LNatg6d9S8347auAvX4SiAfDJX0B7HKmUZilRExFJk9q6Br4683U6dsjj55ceT36ehUStaCB07h51eCIiB6yssprjBvekR9eCqEOJzuhzoeeQ1JXqn/0N2PQ+XHRXWAcnEqdETUQkTf7nmXLeXLWZn1x0LAN6dAknq8u19kBEctKm7bUsqtrEhPY67bFRY6n+9/4Ba988tPd64wFY9ABM/DYMOSk18UmboURNRCQNyipizChbwWfHD+Hcow8LJxsaIFahQiIikpNeWlZNg8OE9laWvzknXBEv1X/Hwb/HhhXwxA0w5GQ47Rupi03aDCVqIiIpVr11Jzc8+Aal/bpx88cT9kurWQW7tml9mojkpLKKGEWdO3Bciabn0aUXHHtJKNW/fcOBv76uFh7+Uhid+/QMyO+Q+hgl5ylRExFJIXfnWw8vombHLn552Ri6dMzf82R1vJCIRtREJMe4O3Mrqzl1VDEd8vXxEQil+ut2wKu/P/DXzvlPWP0qTPlf6Dk49bFJm6B/aSIiKXTvy+/y/NJ1fPe8IzhyQJOCIY0VH1WaX0RyzLJ1W1mzeQcTRrfz9WmJ+h8Fw06Df911YKX6l8+Bf/wCTrwKjrogbeFJ7ksqUTOzyWZWbmbLzOw7zTw/1MyeM7NFZvaCmZWkPlQRkey2ZHUN/zV7KWce0Y8rTx627wWxcujSGwq1vkNEcsuLFTEATitV/7WX8ddCTRWUz07u+m3V8Oj0MAX+3P9Kb2yS81pN1MwsH7gdOA84CrjMzI5qctn/AH9w92OBWwH9zRORduXD2nq+cv+r9OhawE8vPhZrbh+cWLmmPYpITiqrrGZE30JKenWNOpTscvh50GNIckVF3OGxf4cPN8LFd0NH/S5l/5IZURsHLHP3Fe5eC8wEmo7THgU8H78/p5nnRURSrrXR/oTrLjIzN7Ox6YrlR08sYXlsGz+75Dj6dOu07wXuKs0vIjlpx6565q1Yr7L8zcnLh3FXw3svwdq39n/tv2ZA5dNw9o/gsI9kJj7JackkaoOAlQmPq+LnEr0BfDp+/1NAkZn1afpGZjbNzBaY2YJYLHYw8YqIAEmP9mNmRcBXgXnpiuWpt9Zw37z3mT5hBKe19EFmW3X4FlUjaiKSY+a/u4GddQ1M1Pq05o25Ajp02f8G2GvfhGf+A0rPhfHTMxeb5LRUFRP5BjDRzF4DJgKrgPqmF7n7DHcf6+5j+/bVP3YROSTJjPYD/Aj4CbAjHUGs3vQh3/7Lmxxb0oMbz9lPkZDY0nCr0vwikmPKKmJ0zM9j/IjeUYeSnbr2DqX6Fz3YfKn+2u3w8BdDSf8Lfw3NTY0XaUYyidoqILFuaEn83G7uvtrdP+3uY4Dvxc9tSlmUIiL7anW038xOAAa7+xOtvdnBjPjXNzhff+B1dtU38IupY+jYYT9darUqPopIbiqrqGbssF507ai9vlo0fjrUfQiv/XHf556+Caor4VO/VTEpOSDJJGrzgVIzG25mHYGpwKzEC8ys2Mwa3+sm4J7UhikicmDifdLPgBuTuf5gRvx/PWcZ897ZwA+nHM3w4sL9Xxwrh47doHvTmeMiItlr7eYdlH+wRWX5W9P/aBh6aijV35AwqWzJLFh4L5xyPYw8PbLwJDe1mqi5ex1wHfA08DbwoLsvNrNbzWxK/LJJQLmZVQD9gf9MU7wiIo1aG+0vAo4BXjCzd4GTgFmpKiiyefsuZsxdwSePG8jFJyaxI0msPEx71JQXEckhcyvDDAMVEknC+Omw+X0ofzI83rQSZl0HA0+A02+ONjbJSUmNYbv7bGB2k3O3JNx/GHg4taGJiOzX7tF+QoI2Fbi88Ul33wzsnmNiZi8A33D3BalovEfXAv765VMoLurUfCn+pqorYIS+TRWR3FJWWU1xt04cOaAo6lCy3+HnQ/cSmPfbULb/kWlhdO3iu6FDx6ijkxyUqmIiIiIZleRof1qN6NuN7p0LWr9wx2bYskal+UUkp9Q3OC9VxphQWpzcF1LtXX6HUKr/3bnw6LXw/svw8f8LvUdEHZnkKK0KFZGc1dpof5PzkzIRU7NiFeFWpflFJIe8tWozG7fv0vq0A3HClfDCbfDmg3DspXDc1KgjkhymETURkXRTaX4RyUFlFWF92qmlqlSYtK69Ydw06H8MnP8/UUcjOU6JmohIulWXQ34n6DUs6khERJI2t7KaYwZ1p7hbp6hDyS1n3wrT50Ln7lFHIjlOiZqISLrFKqC4FPLyo45ERCQpW3bs4tX3N6ra48Ewgzx9xJZDp79FIiLpFluqaY8iklNeXr6eugbnNCVqIpFRoiYikk6122HT+yokIiI5pawiRmHHfE4c2ivqUETaLSVqIiLptL4ScJXmF5Gc4e6UVcb42Mg+dOygj4oiUdG/PhGRdGoszV98eLRxiIgk6b3121m54UOV5ReJmBI1EZF0qi4Hy4c+I6OOREQkKWWVoSy/1qeJREuJmohIOsWWQu/h0EHlrUUkN5RVxBjcuwvD+nSNOhSRdk2JmohIOsUqVEhERHJGbV0Dryxfz4TSvphZ1OGItGtK1ERE0qV+F2xYrtL8Im2YmU02s3IzW2Zm32nhmkvMbImZLTaz+xLOX2lmlfHjysxF3bJX39/Ittp6rU8TyQIdog5ARKTN2rACGuqgrwqJiLRFZpYP3A6cDVQB881slrsvSbimFLgJOMXdN5pZv/j53sD3gbGAAwvjr92Y6Z8jUVlFjPw84+SRfaIMQ0TQiJqISPrEloZbJWoibdU4YJm7r3D3WmAmcEGTa64Bbm9MwNx9Xfz8ucCz7r4h/tyzwOQMxd2issoYJwzpSVHngqhDEWn3lKiJiKTL7tL8mvoo0kYNAlYmPK6Kn0s0GhhtZv8ws3+a2eQDeC1mNs3MFpjZglgslsLQ91W9dSdvraphgqo9imQFJWoiIulSXQ49hkDHwqgjEZHodABKgUnAZcCdZtYz2Re7+wx3H+vuY/v2TW8C9VJlNYDWp4lkCSVqIiLpElsKfTWaJtKGrQIGJzwuiZ9LVAXMcvdd7v4OUEFI3JJ5bUaVVcbo1bWAYwb1iDIMEYlLKlFrraKRmQ0xszlm9pqZLTKz81MfqohIDmmoh+pKKNb6NJE2bD5QambDzawjMBWY1eSaxwijaZhZMWEq5ArgaeAcM+tlZr2Ac+LnIuHuzK2s5pRRxeTnqSy/SDZotepjMhWNgJuBB939N2Z2FDAbGJaGeEVEcsOm96FuhwqJiLRh7l5nZtcREqx84B53X2xmtwIL3H0WexKyJUA98E13Xw9gZj8iJHsAt7r7hsz/FMHba7YQ27JT0x5Fskgy5fl3VzQCMLPGikaJiZoD3eP3ewCrUxmkiEjOqY4XElGiJtKmuftswhfUieduSbjvwA3xo+lr7wHuSXeMySirDIVKVEhEJHskM/UxmapEPwA+Z2ZVhM7qK829USYrF4mIRCpWHm5V8VFEckBZRYzD+xdxWI/OUYciInGpKiZyGXCvu5cA5wN/NLN93juTlYtERCIVK4fCftC1d9SRiIjs1/baOha8u5HTSoujDkVEEiSTqCVTlehLwIMA7v4K0BnQv3YRab+qyzXtUURywrwVG6itb9D6NJEsk0yilkxFo/eBMwHM7EhCoqa5jSLSPrmHETUlaiKSA16siNGpQx7jhmsGgEg2aTVRc/c6oLGi0duE6o6LzexWM5sSv+xG4BozewO4H7gqvnhWRCRtktg65Foze9PMXjezl+JVadNvy1rYWaPS/CKSE8oqY4wf0YfOBflRhyIiCZKp+phMRaMlwCmpDU1EpGVJbh1yn7v/Nn79FOBnwOS0B1cdLySiETURyXJVG7ezIraNy8cNiToUEWkiVcVEREQybffWIe5eCzRuHbKbu9ckPCwkbCWSfjElaiKSG+ZWVgNofZpIFkpqRE1EJAs1t3XI+KYXmdmXCfsXdQTOaOnNzGwaMA1gyJBD/GY5Vg6dekC3/of2PiIiaVZWEeOw7p0p7dct6lBEpAmNqIlIm+but7v7SODbwM37uS5124c0FhIxO7T3ERFJo7r6Bl5aVs2E0cWY+iuRrKNETURyVTJbhySaCVyY1ogaVZdDX210LSLZ7Y2qTWzZUadpjyJZSomaiOSqVrcOMbPShIcfByrTHtX2DbAtBn2PSHtTIiKHoqyiGjM4ZaS2vhXJRlqjJiI5yd3rzKxx65B84J7GrUOABe4+C7jOzM4CdgEbgSvTHlhjIRGV5heRLFdWGePYkp70KuwYdSgi0gwlaiKSs5LYOuSrGQ9KpflFJAds3r6LN1Zu4rrTR0Udioi0QFMfRURSKVYOBV2hx+DWrxURichLy6ppcJXlF8lmStRERFIpVg59RkGeulcRyV5zK2MUderA8YN7Rh2KiLRAnyRERFKpukKFREQkq7k7ZRUxTh7Vhw75+igokq30r1NEJFV2boXNK1WaX0Sy2vLYVlZv3qFpjyJZTomaiEiqVFeEW42oiUgWe7GiGoAJpUrURLKZEjURkVRRaX4RyQFlFTFGFBcyuHfXqEMRkf3I/kStrhaW/DXqKEREWlddDnkdoPfwqCMREWnWjl31zHtnPaeVapNrkWyX/Ynawt/Bg5+HJ26E+l1RRyMi0rJYRaj4mF8QdSQiIs1a8O5Gduxq0Po0kRyQ/Rtef/TqsDj/5f8N6z8+83vo2jvqqERE9hVbCv2PjjoKEZEWlVXGKMg3ThrRJ+pQRKQV2T+ilpcP5/wYLvwNvP9PuPOMPetARESyRd1O2PiOComISFYrq4gxdmhvCjtl/3f1Iu1d9idqjY6/HK78G9RuhbvOgspno45IRGSP9cvAG6CvComISHb6oGYHS9du0bRHkRyRVKJmZpPNrNzMlpnZd5p5/udm9nr8qDCzTakPFRgyHq6ZA72Gwn2XwMu/Ave0NCUickB2V3zUHmoikp3mVoay/CokIpIbWh33NrN84HbgbKAKmG9ms9x9SeM17v71hOu/AoxJQ6xBz8Hwxafh0Wvhme/BuiXwiZ9Dh05pazIyuz6EOf8Jix8L06lKxoZj0InQpVfU0YlIouoKwKC4NOpIRESaVVYRo7hbR44a0D3qUEQkCclMUB4HLHP3FQBmNhO4AFjSwvWXAd9PTXgt6FgYioq8eBu8+JMw5ejSP0O3NjSU/94r8Ncvw4blMOos2FwFy/4OxEcQ+4yCQWP3JG/9j1GlOZEoxZaG0f6CLlFHIiKyj4YG56Vl1Uwc3Ze8PIs6HBFJQjKJ2iBgZcLjKmB8cxea2VBgOPB8C89PA6YBDBky5IAC3UdeHpz+3TDS9Ni/w52nw2X3w2EfObT3jVrtNnjuVph3Rxg9/PxfYcSk8NyOGlj9KlQtgFULYflzsGhmeK5DZxhwHJR8NIy4lYyFHoPB1BmLZESsQoU55ReCAAARjElEQVRERCRrvbV6Mxu21TJhtKY9iuSKVJf8mQo87O71zT3p7jOAGQBjx45NzeKyYz4dNpe9/3K4+xz49Aw48pMpeeuMe2cuzLoONr4L46bBmd+HTt32PN+5e0jaRkwKj91h0/uwagFULQy3/7oT6n8Vni/sFxK3khPD6NugE6BTUUZ/pKQ11MOWtWHksKYq3DYe9btg5OkwejL0GRl1pCL7qq+D9ZUw6syoIxERaVbj+rRTR7Wh2UcibVwyidoqYHDC45L4ueZMBb58qEEdsIFjYNocmHk5PPA5OP1mmPCN3BlN2rkF/v4DmH8X9BoOVz0Bw05t/XVmYapVr6FwzEXhXF0tfPBWGHGrWgBV86H8icYXQL8j94y4DRobHuflp+sn22PHZti8Kp58rdw7EdtcBVtWQ0Pd3q/p1AN6lITzT383HH1KYfS5cPh5MHi8pntKdtj0HtTXquKjiGStFytiHDWgO32L2uCafpE2KplEbT5QambDCQnaVODypheZ2RFAL+CVlEaYrKLD4KrZ8Pj1MOfHEHsbLrg9+9eLLJ8Ds64PyctJX4YzboaOXQ/+/Tp0DCNng06AcdeEc9s3wKpX4yNv82Hp3+C1P4bnOnYLiW5j8lby0fC7PBD1u2DLmoTEq5lEbGfN3q/J6wDdB4bpmUM/FhKyHiXhcY8S6D4ojCA22vguVDwDFU/Cv2bAK7+Czj3C+r3R54WRDG2ELlFprPioqY8ikoW27NjFq+9t5OrTRkQdiogcgFYTNXevM7PrgKeBfOAed19sZrcCC9x9VvzSqcBM9wjr5Rd0hk/dEUaJ/v5D2LACpt4XEoJss2MzPHMzvPqHMEr0xafD9gPp0LU3lJ4VDghTJjes2DPitmpBSHwaR7S6l+yZLlny0TDdsHFaYnOJ2Na1Yf+oRF16h4Sr13AYdtq+iVi3fgc2ktdrGIyfFo6dW0KCW/E0VD4Nb/0FLB+GnBRG20afFyrv5cqIquS+2NJwq4qPIpKFXlm+nroG1/o0kRyT1Bo1d58NzG5y7pYmj3+QurAOgRmc+vXwzfZfroYZp4dkreTEqCPbo/JZePyrYRTqlK/CpJsyO/JnFpKvPiPhuEvDuV07YO2ivZO3JX9t/vX5naDHoJBwjTw9IQmLJ2LdB4bKnOnSqQiOmhKOhoZQYKXiKSh/Cp69JRy9hoc1bYdPhiEnh5FGkXSproCigWGUV0Qky5RVxuhSkM+JQ7W1j0guSXUxkexx+HnwpWfh/kvhd+eFaZDHfibamD7cCE9/D17/c0gkL/lDmG6YDQo6w+Bx4Wi0dV1Y67bpfSgasCcZ61ocqm5mg7y8PVsUnHFzGOGreCqMti24B+b9Bjp1h5FnhMSt9Bwo7BN11NLWxJZCX210LSLZaW5lNR8b2YdOHTKwJl1EUqbtJmoA/Y+Ca+bAg5+HR64Om2Of8R/RJBnlT8LjX4NtMTjtRpj47ezfpLtbv5Dw5pIeJfDRq8NRuw1WvLgncVvyGGAhGR09ORz9jtQUSTk07lBdCcd/NupIRET28d76bby3fjtfOHlY1KGIyAFq24kaQGExXPEYzL4RXvpZWPT/6TsyV6Z++wZ48lvw5kNhU+rLH4CBx2em7fauYyEccX44Ghpg7RthemTFU/DcD8PRc0g8aTs3rKXL9uRZ9mJmk4FfENbP3uXutzV5/gbgaqAOiAFfdPf3UhpEzSqo3aqKjyKSlcoqYgBMGK2y/CK5pu0nahDWJ33yl9DvaHj6Jrj73LA5dq+h6W13yV/hiRvDlMdJN8GpN2itVFTy8kJ1y4Fj4PSboGY1VD4TErdX/xgqSRYUhjV3h58Xpkh26xd11LIfZpYP3A6cDVQB881slrsvSbjsNWCsu283s38DfgpcmtJAGguJKFETkSz0YkU1Jb26MLw4jWvHRSQt2keiBmF620nXhqpsD30B7jwdLv0TDD059W1tjcHsb4SpdocdC1c8Cod9JPXtyMHrPhBOvCocuz6Ed8r2TJFc+jfAwpYFoyfD8NOgsG+ontmpR/asz5NxwDJ3XwFgZjOBC4DdiZq7z0m4/p/A51IeRawi3Ko0v0i7k8So/lXAf7Nn/9lfuftd8efqgTfj59939ympjm9XfQOvLK9myvGDME3zF8k57SdRazTqTLjmObh/Kvx+CnziZ3DC51Pz3u6w+BGY/c1QQv6Mm+GUr2lT5mxX0CVe1v/c8Ge49s2QsFU8GfbkS/yob3nQuSd06RUSty69wlYEu+/32ve5Lr3CVNts+U/SPSSntVvD39PGo6XHlgeT/yvqqJszCFiZ8LgK2N8eF18CnmzpSTObBkwDGDJkSPJRxJaGP+dClb0WaU+SHNUHeMDdr2vmLT5097SuhXj1vY1sq61nosryi+Sk9peoQRhVu/rv8PAXYdZXYN3bcPaPIP8Qfh1bPoAnbgijMQNPgAt/HQpVSG4xgwHHhmPiN8Of65rXw/TV7RvC7Ycb9jze+kH4oL59I9Ruafl98wpaSOJ6xh/3bj7Ba9z83B3qdsDOrWHz8N1JVfy2dksrjxuvrwnnvD6J30UedCwKU0CzM1FLmpl9DhgLTGzpGnefAcwAGDt2bPL7QVZXaNqjSPvU6qh+1MoqY+TnGSePUqImkovaZ6IG4UPw5Q+FTaf/+etQZOTie8IH5wPhDosegCe/HUYpzvohfOy6Q0v6JHsU9Yeic5O7tn5Xywld4+PG+5vehzVvhMd1H7b8nh06h6N2654NyffLwuhdpyLo2C1+v1v4OToW7Xm8+/nuTc4lXFPQNXtGAZu3Chic8LiEPdOLdjOzs4DvARPdfWdKI3APifpRF6T0bUUkJyQ7qn+RmU0AKoCvu3vjazqb2QJCsaPb3P2x5ho56NF+oKyimjGDe9K9s2b2iOSi9p1N5HeA826DfkeEoh93nQWXzYTiUcm9vmY1/O3rYW1TybiwV5v2Umq/8gvC6NOBFiHZ9eF+ErwNULczIelqJgnr1H3P44Ku7WkN3Xyg1MyGExK0qcDliReY2RjgDmCyu69LeQTbqsOfVbFG1ESkWY8D97v7TjObDvweOCP+3FB3X2VmI4DnzexNd1/e9A0OdrR//dadvLV6M187U59LRHJV+07UGp14FfQphQevgLvOgM/cGzZIbol72LT6qe9CfS2c+39g/LWQp40k5SAUdAlH94FRR5JT3L3OzK4DniYs5L/H3Reb2a3AAnefRVjE3w14KL6QPrUL9qvLw62mPoq0R62O6rv7+oSHdxEqzzY+typ+u8LMXgDGAPskagfrpWXVuMMErU8TyVlK1BoNOwWueR7uvwz+dDFMvg3GXbPv1K9NK+Hxr8Ly52DoKTDlf6HPyGhiFmnn3H02MLvJuVsS7p+V1gBUml+kPUtmVH+Au6+JP5wCvB0/3wvYHh9pKwZOISGJS4Wyimp6di3g2JIDXNIhIllDiVqiXsPgS8/AX66BJ78J6xbDef8d9j5zh4W/g2duAW8I5z96dXuaZiYiTcUqwrTT7oOijkREMizJUf3rzWwKYR3aBuCq+MuPBO4wswYgj7BGLWVFSNyduZUxThlVTH5eVq81FpH9UKLWVKcimHofPH8rvPRzqF4G5/wI/v4DeOdFGD4hjKL1GhZ1pCIStdhSKB6d7UVXRCRNkhjVvwm4qZnXvQykbYPVpWu3sG7LTiaW9k1XEyKSAUrUmpOXB2f9APoeGcr333l6qIb3iZ/DiV/QhzIRCaorYMSkqKMQEdlLWUUMgNO0Pk0kpylR25/jLoU+o2DRTDj5eug5uPXXiEj70FAPI06HES1uzSYiEon+3Ttz4fEDGdCjS9ShiMghUKLWmpITwyEikigvHz71m6ijEBHZx4VjBnHhGK2dFcl1qoQhIiIiIiKSZZSoiYiIiIiIZBklaiIiIiIiIlkmqUTNzCabWbmZLTOz77RwzSVmtsTMFpvZfakNU0REREREpP1otZiImeUDtwNnA1XAfDOblbgxo5mVEvYJOcXdN5pZv3QFLCIiIiIi0tYlM6I2Dljm7ivcvRaYCVzQ5JprgNvdfSOAu69LbZgiIiIiIiLtRzKJ2iBgZcLjqvi5RKOB0Wb2DzP7p5lNbu6NzGyamS0wswWxWOzgIhYREREREWnjUlVMpANQCkwCLgPuNLOeTS9y9xnuPtbdx/bt2zdFTYuIiIiIiLQtyWx4vQoYnPC4JH4uURUwz913Ae+YWQUhcZvf0psuXLiw2szeSzLOYqA6yWvTQe2rfbWfnKHpDCRT1D+pfbXfJtvP+f5JfZPaV/ttsv0W+yZz9/2+0sw6ABXAmYQEbT5wubsvTrhmMnCZu19pZsXAa8Dx7r4+yQBbi2GBu49NxXupfbWv9nOr/WwX9e9H7at9ta/+qTlR/27UvtpX+4fefqtTH929DrgOeBp4G3jQ3Reb2a1mNiV+2dPAejNbAswBvpmqJE1ERERERKS9SWbqI+4+G5jd5NwtCfcduCF+iIiIiIiIyCFIVTGRdJuh9tW+2m+37We7qH8/al/tq31pTtS/G7Wv9tX+IWp1jZqIiIiIiIhkVq6MqImIiIiIiLQbStRERERERESyTNYnamY22czKzWyZmX0nw23fY2brzOytTLab0P5gM5tjZkvMbLGZfTXD7Xc2s3+Z2Rvx9n+YyfYT4sg3s9fM7G8RtP2umb1pZq+b2YII2u9pZg+b2VIze9vMPpbBtg+P/9yNR42ZfS1T7We7KPumePuR9U/qm3bHob5JfVNW0men9t0/Rdk3xdtX/5Si/imr16iZWT5hD7ezCZtqzyfs17YkQ+1PALYCf3D3YzLRZpP2BwAD3P1VMysCFgIXZvDnN6DQ3beaWQHwEvBVd/9nJtpPiOMGYCzQ3d0/keG23wXGunskmyaa2e+Bue5+l5l1BLq6+6YI4sgn7KM43t2T3Wy1zYq6b4rHEFn/pL5pdxzqm9Q3ZZ2o+yd9doq+f4qyb4q3/y7qn1LSP2X7iNo4YJm7r3D3WmAmcEGmGnf3MmBDptprpv017v5q/P4Wwj52gzLYvrv71vjDgviR0czezEqAjwN3ZbLdbGBmPYAJwN0A7l4bRUcTdyawXB+Edou0b4Jo+yf1TeqbUN+UzfTZqR33T+25b4K21z9le6I2CFiZ8LiKDP5jyyZmNgwYA8zLcLv5ZvY6sA541t0z2j7w/4BvAQ0ZbreRA8+Y2UIzm5bhtocDMeB38SkMd5lZYYZjaDQVuD+itrOR+qY49U3qm9Q3ZR31T3HttH+Kum8C9U+NDrl/yvZETQAz6wb8Bfiau9dksm13r3f344ESYJyZZWwag5l9Aljn7gsz1WYzTnX3E4DzgC/Hp3RkSgfgBOA37j4G2AZEsRaqIzAFeCjTbUt2U9+kvkl9k2Sr9tg/ZUnfBOqfUtY/ZXuitgoYnPC4JH6u3YjPb/4L8Gd3fySqOOLDxnOAyRls9hRgSnyu80zgDDP7Uwbbx91XxW/XAY8SppRkShVQlfBN3MOEzifTzgNedfcPImg7W6lvUt+kvkl9U7ZS/9R++6fI+yZQ/xSXkv4p2xO1+UCpmQ2PZ6ZTgVkRx5Qx8QWpdwNvu/vPImi/r5n1jN/vQliYvDRT7bv7Te5e4u7DCH/2z7v75zLVvpkVxhciEx82PwfIWBUrd18LrDSzw+OnzgQyVqwiwWVoalFT6pvUN6lvUt+UrdQ/tdP+Keq+CdQ/JUhJ/9QhBYGkjbvXmdl1wNNAPnCPuy/OVPtmdj8wCSg2syrg++5+d6baJ3wzcgXwZnyuM8B33X12htofAPw+XrUmD3jQ3SMp9RqR/sCjoc+nA3Cfuz+V4Ri+Avw5/p/tCuALmWw83smeDUzPZLvZLuq+CSLvn9Q3RUt9k/qmFkXdP+mzk/on1D+lrH/K6vL8IiIiIiIi7VG2T30UERERERFpd5SoiYiIiIiIZBklaiIiIiIiIllGiZqIiIiIiEiWUaImIiIiIiKSZZSoiYiIiIiIZBklaiIiIiIiIlnm/wPygLuVSwX0bQAAAABJRU5ErkJggg==\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "execution_count": 30,
      "metadata": {
        "collapsed": true,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "gather": {
          "logged": 1621081561143
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Summary\r\n",
        "\r\n",
        "1. The validation loss and training loss have come down gradually with every epoch\r\n",
        "2. The accuracy is low but there is less underfitting observed. The accuracy can be improved further with more training samples or layers. However accuracy is not a right\r\n",
        "measure for imbalanced data, we should try to improve AUC instead. \r\n",
        "3. Validation AUC also looks Ok, it can be better. \r\n",
        "\r\n",
        "With this model let us try predicting.\r\n",
        "\r\n",
        "Please note the intention of this Notebook is not to achieve high accuracy, it is to teach how to run a CNN model using distributed ML techquies.\r\n",
        "For practical purposes this accuracy will note be sufficient."
      ],
      "metadata": {
        "nteract": {
          "transient": {
            "deleting": false
          }
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 3. Making a Prediction"
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.models import load_model\n",
        "\n",
        "# Loading the weights from the best model.\n",
        "cnn.load_weights('models/best_model.hdf5')"
      ],
      "outputs": [],
      "execution_count": 31,
      "metadata": {
        "gather": {
          "logged": 1621081568099
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "validation_path = os.path.join(DATASET_PATH, disease_cls[0], '*')\n",
        "validations = glob.glob(validation_path)\n",
        "image_id = np.random.randint(0, len(validations))\n",
        "print(validations[image_id])\n",
        "validation = cv2.imread(validations[image_id], cv2.IMREAD_GRAYSCALE)\n",
        "plt.imshow(validation,cmap='gray')\n",
        "img = preprocess_img(validation[:, :, np.newaxis], 'validation')\n",
        "print('Effusion detected, % of confidence: ', round(cnn.predict(img[np.newaxis,:])[0][0] * 100,2))"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/var/tmp/effusion/data/effusion/00007520_000.png\n",
            "Effusion detected, % of confidence:  64.47\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 432x288 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "execution_count": 32,
      "metadata": {
        "gather": {
          "logged": 1621081569241
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 4. Submitting Experiments using ScriptRunConfig\r\n",
        "\r\n",
        "From the above experiements we find out the best model and create a training script\r\n",
        "\r\n",
        "__What is ScriptRunConfig?__\r\n",
        "- A ScriptRunConfig object is used to configure the information necessary for submitting a training run as part of an Experiment.\r\n",
        "- A ScriptRunConfig packages together the configuration information needed to submit a run in Azure ML, including the script, compute target, environment, and any distributed job-specific configs.\r\n",
        "\r\n",
        "More details: https://docs.microsoft.com/en-us/python/api/azureml-core/azureml.core.scriptrunconfig?view=azure-ml-py\r\n"
      ],
      "metadata": {
        "nteract": {
          "transient": {
            "deleting": false
          }
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import os\r\n",
        "from azureml.core import ScriptRunConfig\r\n",
        "import shutil\r\n",
        "from azureml.core import Experiment\r\n",
        "from azureml.core import Environment\r\n",
        "from azureml.core.conda_dependencies import CondaDependencies\r\n",
        "\r\n",
        "## From the experiments above, the training script is extracted and stored in effusion_detector.py\r\n",
        "## This is a manual process at the moment, although the notebook helps convert the code in cells to python scripts\r\n",
        "project_folder = './effusion_detector'\r\n",
        "os.makedirs(project_folder, exist_ok=True)\r\n",
        "shutil.copy('effusion_detector.py', project_folder)\r\n",
        "\r\n",
        "## Giving the experiment a name\r\n",
        "experiment_name = 'tf-effusion-detector'\r\n",
        "experiment = Experiment(ws, name=experiment_name)\r\n",
        "\r\n",
        "## Choosing the environment the training script depends on, ML creates a docker image with the \r\n",
        "## dependencies declared below\r\n",
        "tf_env = Environment(\"tensorflow-2.3-gpu\")\r\n",
        "packages = CondaDependencies.create(conda_packages=['scikit-learn','pip'],\r\n",
        "                                    pip_packages=['azureml-defaults', 'opencv-python==3.3.0.9', 'matplotlib', 'keras<=2.3.1', 'azureml-sdk', 'azureml-widgets', 'tensorflow'])\r\n",
        "tf_env.python.conda_dependencies = packages\r\n",
        "\r\n",
        "## defining arguments\r\n",
        "args = ['--data-folder', dataset.as_mount(), '--epochs', 20]\r\n",
        "\r\n",
        "## choosing compute target, in this example we will choose single machine in the next one which is for distributed training we will choose a cluster with 4 nodes\r\n",
        "compute_target = ComputeTarget(workspace=ws, name='gpu-instance-small')\r\n",
        "src = ScriptRunConfig(source_directory=project_folder,\r\n",
        "                      script='effusion_detector.py',\r\n",
        "                      arguments=args,\r\n",
        "                      compute_target=compute_target,\r\n",
        "                      environment=tf_env)\r\n",
        "\r\n",
        "run = experiment.submit(src)\r\n",
        "print(run)\r\n",
        "run.get_details()\r\n"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Run(Experiment: tf-effusion-detector,\n",
            "Id: tf-effusion-detector_1621081608_4fe7f8b6,\n",
            "Type: azureml.scriptrun,\n",
            "Status: Preparing)\n"
          ]
        },
        {
          "output_type": "execute_result",
          "execution_count": 33,
          "data": {
            "text/plain": "{'runId': 'tf-effusion-detector_1621081608_4fe7f8b6',\n 'target': 'gpu-instance-small',\n 'status': 'Preparing',\n 'properties': {'_azureml.ComputeTargetType': 'amlcompute',\n  'ContentSnapshotId': '52295cdf-7227-4333-a1c9-d560485f0190',\n  'azureml.git.repository_uri': 'https://github.com/sriksmachi/octopus.git',\n  'mlflow.source.git.repoURL': 'https://github.com/sriksmachi/octopus.git',\n  'azureml.git.branch': 'master',\n  'mlflow.source.git.branch': 'master',\n  'azureml.git.commit': '350a13c17e02d0a6b1e1cbc41d88390723744e28',\n  'mlflow.source.git.commit': '350a13c17e02d0a6b1e1cbc41d88390723744e28',\n  'azureml.git.dirty': 'False'},\n 'inputDatasets': [{'dataset': {'id': '936fc596-dc32-48e8-8ba5-dcbc3bf2b310'}, 'consumptionDetails': {'type': 'RunInput', 'inputName': 'input__b7001421', 'mechanism': 'Mount'}}],\n 'outputDatasets': [],\n 'runDefinition': {'script': 'effusion_detector.py',\n  'command': '',\n  'useAbsolutePath': False,\n  'arguments': ['--data-folder',\n   'DatasetConsumptionConfig:input__b7001421',\n   '--epochs',\n   '20'],\n  'sourceDirectoryDataStore': None,\n  'framework': 'Python',\n  'communicator': 'None',\n  'target': 'gpu-instance-small',\n  'dataReferences': {},\n  'data': {'input__b7001421': {'dataLocation': {'dataset': {'id': '936fc596-dc32-48e8-8ba5-dcbc3bf2b310',\n      'name': None,\n      'version': None},\n     'dataPath': None,\n     'uri': None},\n    'mechanism': 'Mount',\n    'environmentVariableName': 'input__b7001421',\n    'pathOnCompute': None,\n    'overwrite': False}},\n  'outputData': {},\n  'jobName': None,\n  'maxRunDurationSeconds': 2592000,\n  'nodeCount': 1,\n  'priority': None,\n  'credentialPassthrough': False,\n  'identity': None,\n  'environment': {'name': 'tensorflow-2.3-gpu',\n   'version': 'Autosave_2021-05-15T07:07:30Z_b971783c',\n   'python': {'interpreterPath': 'python',\n    'userManagedDependencies': False,\n    'condaDependencies': {'channels': ['anaconda', 'conda-forge'],\n     'dependencies': ['python=3.6.2',\n      {'pip': ['azureml-defaults~=1.27.0',\n        'opencv-python==3.3.0.9',\n        'matplotlib',\n        'keras<=2.3.1',\n        'azureml-sdk~=1.27.0',\n        'azureml-widgets~=1.27.0',\n        'tensorflow']},\n      'scikit-learn',\n      'pip'],\n     'name': 'azureml_5a170d726f14b127b6e0a0211d1ec132'},\n    'baseCondaEnvironment': None},\n   'environmentVariables': {'EXAMPLE_ENV_VAR': 'EXAMPLE_VALUE'},\n   'docker': {'baseImage': 'mcr.microsoft.com/azureml/intelmpi2018.3-ubuntu16.04:20210301.v1',\n    'platform': {'os': 'Linux', 'architecture': 'amd64'},\n    'baseDockerfile': None,\n    'dockerContext': None,\n    'baseImageRegistry': {'address': None, 'username': None, 'password': None},\n    'enabled': False,\n    'arguments': []},\n   'spark': {'repositories': [], 'packages': [], 'precachePackages': True},\n   'inferencingStackVersion': None},\n  'history': {'outputCollection': True,\n   'directoriesToWatch': ['logs'],\n   'enableMLflowTracking': True,\n   'snapshotProject': True},\n  'spark': {'configuration': {'spark.app.name': 'Azure ML Experiment',\n    'spark.yarn.maxAppAttempts': '1'}},\n  'parallelTask': {'maxRetriesPerWorker': 0,\n   'workerCountPerNode': 1,\n   'terminalExitCodes': None,\n   'configuration': {}},\n  'amlCompute': {'name': None,\n   'vmSize': None,\n   'retainCluster': False,\n   'clusterMaxNodeCount': None},\n  'aiSuperComputer': {'instanceType': None,\n   'imageVersion': None,\n   'location': None,\n   'aiSuperComputerStorageData': None,\n   'interactive': False,\n   'scalePolicy': None,\n   'virtualClusterArmId': None,\n   'tensorboardLogDirectory': None,\n   'sshPublicKey': None},\n  'tensorflow': {'workerCount': 1, 'parameterServerCount': 1},\n  'mpi': {'processCountPerNode': 1},\n  'pyTorch': {'communicationBackend': 'nccl', 'processCount': None},\n  'hdi': {'yarnDeployMode': 'Cluster'},\n  'containerInstance': {'region': None, 'cpuCores': 2.0, 'memoryGb': 3.5},\n  'exposedPorts': None,\n  'docker': {'useDocker': False,\n   'sharedVolumes': True,\n   'shmSize': '2g',\n   'arguments': []},\n  'cmk8sCompute': {'configuration': {}},\n  'commandReturnCodeConfig': {'returnCode': 'Zero',\n   'successfulReturnCodes': []},\n  'environmentVariables': {},\n  'applicationEndpoints': {}},\n 'logFiles': {},\n 'submittedBy': 'Srikanth Machiraju'}"
          },
          "metadata": {}
        }
      ],
      "execution_count": 33,
      "metadata": {
        "collapsed": true,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "gather": {
          "logged": 1621081575993
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from azureml.widgets import RunDetails\r\n",
        "\r\n",
        "# Dislay training log\r\n",
        "RunDetails(run).show()"
      ],
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "_UserRunWidget(widget_settings={'childWidgetDisplay': 'popup', 'send_telemetry': False, 'log_level': 'NOTSET',…",
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "31c59f52b54e4e2b9f56066949a13181"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/aml.mini.widget.v1": "{\"status\": \"Running\", \"workbench_run_details_uri\": \"https://ml.azure.com/runs/tf-effusion-detector_1621081608_4fe7f8b6?wsid=/subscriptions/925a5ad8-a21a-48c0-92a2-5a8a2a4dfc46/resourcegroups/ml-workbench/workspaces/ml-workbench-eu&tid=72f988bf-86f1-41af-91ab-2d7cd011db47\", \"run_id\": \"tf-effusion-detector_1621081608_4fe7f8b6\", \"run_properties\": {\"run_id\": \"tf-effusion-detector_1621081608_4fe7f8b6\", \"created_utc\": \"2021-05-15T12:26:52.066863Z\", \"properties\": {\"_azureml.ComputeTargetType\": \"amlcompute\", \"ContentSnapshotId\": \"52295cdf-7227-4333-a1c9-d560485f0190\", \"azureml.git.repository_uri\": \"https://github.com/sriksmachi/octopus.git\", \"mlflow.source.git.repoURL\": \"https://github.com/sriksmachi/octopus.git\", \"azureml.git.branch\": \"master\", \"mlflow.source.git.branch\": \"master\", \"azureml.git.commit\": \"350a13c17e02d0a6b1e1cbc41d88390723744e28\", \"mlflow.source.git.commit\": \"350a13c17e02d0a6b1e1cbc41d88390723744e28\", \"azureml.git.dirty\": \"False\", \"ProcessInfoFile\": \"azureml-logs/process_info.json\", \"ProcessStatusFile\": \"azureml-logs/process_status.json\"}, \"tags\": {\"_aml_system_ComputeTargetStatus\": \"{\\\"AllocationState\\\":\\\"steady\\\",\\\"PreparingNodeCount\\\":0,\\\"RunningNodeCount\\\":0,\\\"CurrentNodeCount\\\":1}\"}, \"script_name\": null, \"arguments\": null, \"end_time_utc\": null, \"status\": \"Running\", \"log_files\": {\"azureml-logs/55_azureml-execution-tvmps_eb7aba4ad42b8357acfe6c8579e0b4eb7788c6bd0ffa3bcc94c4ff70352d47a1_d.txt\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-effusion-detector_1621081608_4fe7f8b6/azureml-logs/55_azureml-execution-tvmps_eb7aba4ad42b8357acfe6c8579e0b4eb7788c6bd0ffa3bcc94c4ff70352d47a1_d.txt?sv=2019-02-02&sr=b&sig=zpfjEeA1I5bIAoa0zvbGEW%2B%2B8RMKKax0KDZ%2FICJEnR0%3D&st=2021-05-15T12%3A18%3A05Z&se=2021-05-15T20%3A28%3A05Z&sp=r\", \"azureml-logs/65_job_prep-tvmps_eb7aba4ad42b8357acfe6c8579e0b4eb7788c6bd0ffa3bcc94c4ff70352d47a1_d.txt\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-effusion-detector_1621081608_4fe7f8b6/azureml-logs/65_job_prep-tvmps_eb7aba4ad42b8357acfe6c8579e0b4eb7788c6bd0ffa3bcc94c4ff70352d47a1_d.txt?sv=2019-02-02&sr=b&sig=%2B%2BWojgO7wIbsmGUfYhOfyQBVGVaOQ7QF6FHHSwOv4Sw%3D&st=2021-05-15T12%3A18%3A05Z&se=2021-05-15T20%3A28%3A05Z&sp=r\", \"azureml-logs/70_driver_log.txt\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-effusion-detector_1621081608_4fe7f8b6/azureml-logs/70_driver_log.txt?sv=2019-02-02&sr=b&sig=meXCqwsOpFgUkHAyFzYJC%2F6u5buLJJ2DIu%2FnCQ0nplw%3D&st=2021-05-15T12%3A18%3A05Z&se=2021-05-15T20%3A28%3A05Z&sp=r\", \"azureml-logs/process_info.json\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-effusion-detector_1621081608_4fe7f8b6/azureml-logs/process_info.json?sv=2019-02-02&sr=b&sig=xEp1Q4CdzTwf3oDy3ZDBWKIwIOvSYESjiYzJGkmsv%2F8%3D&st=2021-05-15T12%3A18%3A05Z&se=2021-05-15T20%3A28%3A05Z&sp=r\", \"azureml-logs/process_status.json\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-effusion-detector_1621081608_4fe7f8b6/azureml-logs/process_status.json?sv=2019-02-02&sr=b&sig=xm6WPKxZkzzu2SLQiSY5Ee%2Bt971%2FhQdHb0WC%2BqiJ40Y%3D&st=2021-05-15T12%3A18%3A05Z&se=2021-05-15T20%3A28%3A05Z&sp=r\", \"logs/azureml/86_azureml.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-effusion-detector_1621081608_4fe7f8b6/logs/azureml/86_azureml.log?sv=2019-02-02&sr=b&sig=xiWy4D66RpTgNyyaIvk1%2FmGJQShNsKETuys4VdJGD%2FY%3D&st=2021-05-15T12%3A18%3A05Z&se=2021-05-15T20%3A28%3A05Z&sp=r\", \"logs/azureml/dataprep/backgroundProcess.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-effusion-detector_1621081608_4fe7f8b6/logs/azureml/dataprep/backgroundProcess.log?sv=2019-02-02&sr=b&sig=uw7zsQaWhPY2n9i5QkaopUdEVcTU8JZJdmeCP8wgwJQ%3D&st=2021-05-15T12%3A18%3A05Z&se=2021-05-15T20%3A28%3A05Z&sp=r\", \"logs/azureml/dataprep/backgroundProcess_Telemetry.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-effusion-detector_1621081608_4fe7f8b6/logs/azureml/dataprep/backgroundProcess_Telemetry.log?sv=2019-02-02&sr=b&sig=jk%2BZGPpsBsZkEHhJ%2Fulmmh4q6MM4RMqiqWoTAga%2FpHE%3D&st=2021-05-15T12%3A18%3A05Z&se=2021-05-15T20%3A28%3A05Z&sp=r\", \"logs/azureml/job_prep_azureml.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-effusion-detector_1621081608_4fe7f8b6/logs/azureml/job_prep_azureml.log?sv=2019-02-02&sr=b&sig=agwWcaaw8To1aawqXpqG0H%2B7BF0JdScbVKlODXbVbfA%3D&st=2021-05-15T12%3A18%3A05Z&se=2021-05-15T20%3A28%3A05Z&sp=r\", \"logs/azureml/sidecar/tvmps_eb7aba4ad42b8357acfe6c8579e0b4eb7788c6bd0ffa3bcc94c4ff70352d47a1_d/all.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-effusion-detector_1621081608_4fe7f8b6/logs/azureml/sidecar/tvmps_eb7aba4ad42b8357acfe6c8579e0b4eb7788c6bd0ffa3bcc94c4ff70352d47a1_d/all.log?sv=2019-02-02&sr=b&sig=8Dj5U5FSFz670TvxJMf%2BVmxPCXnHc2qAO0SXsUANEPQ%3D&st=2021-05-15T12%3A18%3A05Z&se=2021-05-15T20%3A28%3A05Z&sp=r\", \"logs/azureml/sidecar/tvmps_eb7aba4ad42b8357acfe6c8579e0b4eb7788c6bd0ffa3bcc94c4ff70352d47a1_d/task.enter_contexts.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-effusion-detector_1621081608_4fe7f8b6/logs/azureml/sidecar/tvmps_eb7aba4ad42b8357acfe6c8579e0b4eb7788c6bd0ffa3bcc94c4ff70352d47a1_d/task.enter_contexts.log?sv=2019-02-02&sr=b&sig=77waJAODiysM7lHmn6MGLxnYWOLRCymF45ZoSAH1E00%3D&st=2021-05-15T12%3A18%3A05Z&se=2021-05-15T20%3A28%3A05Z&sp=r\"}, \"log_groups\": [[\"azureml-logs/process_info.json\", \"azureml-logs/process_status.json\", \"logs/azureml/dataprep/backgroundProcess.log\", \"logs/azureml/dataprep/backgroundProcess_Telemetry.log\", \"logs/azureml/job_prep_azureml.log\"], [\"logs/azureml/sidecar/tvmps_eb7aba4ad42b8357acfe6c8579e0b4eb7788c6bd0ffa3bcc94c4ff70352d47a1_d/all.log\", \"logs/azureml/sidecar/tvmps_eb7aba4ad42b8357acfe6c8579e0b4eb7788c6bd0ffa3bcc94c4ff70352d47a1_d/task.enter_contexts.log\"], [\"azureml-logs/55_azureml-execution-tvmps_eb7aba4ad42b8357acfe6c8579e0b4eb7788c6bd0ffa3bcc94c4ff70352d47a1_d.txt\"], [\"azureml-logs/65_job_prep-tvmps_eb7aba4ad42b8357acfe6c8579e0b4eb7788c6bd0ffa3bcc94c4ff70352d47a1_d.txt\"], [\"azureml-logs/70_driver_log.txt\"], [\"logs/azureml/86_azureml.log\"]], \"run_duration\": \"0:06:09\", \"run_number\": \"1621081612\", \"run_queued_details\": {\"status\": \"Running\", \"details\": null}}, \"child_runs\": [], \"children_metrics\": {}, \"run_metrics\": [{\"name\": \"Loss\", \"run_id\": \"tf-effusion-detector_1621081608_4fe7f8b6\", \"categories\": [0, 1, 2, 3, 4], \"series\": [{\"data\": [0.3163427710533142, 0.3237936794757843, 0.32190555334091187, 0.31122857332229614, 0.3042263686656952]}]}, {\"name\": \"Accuracy\", \"run_id\": \"tf-effusion-detector_1621081608_4fe7f8b6\", \"categories\": [0, 1, 2, 3, 4], \"series\": [{\"data\": [0.9047619104385376, 0.8999999761581421, 0.8999999761581421, 0.9047619104385376, 0.9095237851142883]}]}, {\"name\": \"AUC\", \"run_id\": \"tf-effusion-detector_1621081608_4fe7f8b6\", \"categories\": [0, 1, 2, 3, 4], \"series\": [{\"data\": [0.6041824137062233, 0.6190476190476191, 0.6573684210526316, 0.6034720308624966, 0.6231578947368421]}]}], \"run_logs\": \"2021-05-15 12:27:51,028|azureml|DEBUG|Inputs:: kwargs: {'OutputCollection': True, 'EnableMLflowTracking': True, 'snapshotProject': True, 'only_in_process_features': True, 'skip_track_logs_dir': True}, track_folders: None, deny_list: None, directories_to_watch: []\\n2021-05-15 12:27:51,028|azureml.history._tracking.PythonWorkingDirectory|DEBUG|Execution target type: batchai\\n2021-05-15 12:27:51,040|azureml.history._tracking.PythonWorkingDirectory|DEBUG|Failed to import pyspark with error: No module named 'pyspark'\\n2021-05-15 12:27:51,040|azureml.history._tracking.PythonWorkingDirectory.workingdir|DEBUG|Pinning working directory for filesystems: ['pyfs']\\n2021-05-15 12:27:52,931|azureml.core._experiment_method|DEBUG|Trying to register submit_function search, on method <class 'azureml.train.hyperdrive.runconfig.HyperDriveRunConfig'>\\n2021-05-15 12:27:52,931|azureml.core._experiment_method|DEBUG|Registered submit_function search, on method <class 'azureml.train.hyperdrive.runconfig.HyperDriveRunConfig'>\\n2021-05-15 12:27:52,931|azureml.core._experiment_method|DEBUG|Trying to register submit_function search, on method <class 'azureml.train.hyperdrive.runconfig.HyperDriveConfig'>\\n2021-05-15 12:27:52,931|azureml.core._experiment_method|DEBUG|Registered submit_function search, on method <class 'azureml.train.hyperdrive.runconfig.HyperDriveConfig'>\\n2021-05-15 12:27:52,931|azureml.core.run|DEBUG|Adding new factory <function HyperDriveRun._from_run_dto at 0x7fbb5ac288c8> for run source hyperdrive\\n2021-05-15 12:27:53,964|azureml.core.run|DEBUG|Adding new factory <function AutoMLRun._from_run_dto at 0x7fbb58a0dae8> for run source automl\\n2021-05-15 12:27:53,976|azureml.core.run|DEBUG|Adding new factory <function PipelineRun._from_dto at 0x7fbb5a8a8620> for run source azureml.PipelineRun\\n2021-05-15 12:27:53,987|azureml.core.run|DEBUG|Adding new factory <function StepRun._from_reused_dto at 0x7fbb5a8aa0d0> for run source azureml.ReusedStepRun\\n2021-05-15 12:27:53,996|azureml.core.run|DEBUG|Adding new factory <function StepRun._from_dto at 0x7fbb5a8aa048> for run source azureml.StepRun\\n2021-05-15 12:27:54,007|azureml.core.run|DEBUG|Adding new factory <function ScriptRun._from_run_dto at 0x7fbb5ad982f0> for run source azureml.scriptrun\\n2021-05-15 12:27:54,008|azureml.core.run|DEBUG|Identity in use is not set. Falling back to using AMLToken\\n2021-05-15 12:27:54,009|azureml.core.run|DEBUG|Identity in use is not set. Falling back to using AMLToken\\n2021-05-15 12:27:54,009|azureml.core.run|DEBUG|Using AMLToken auth for remote run\\n2021-05-15 12:27:54,009|azureml.core.run|DEBUG|Identity in use is not set. Falling back to using AMLToken\\n2021-05-15 12:27:54,009|azureml.core.run|DEBUG|Using AMLToken auth for remote run\\n2021-05-15 12:27:54,010|azureml.core.run|DEBUG|Using AMLToken auth for remote run\\n2021-05-15 12:27:54,011|azureml.core.authentication.TokenRefresherDaemon|DEBUG|Starting daemon and triggering first instance\\n2021-05-15 12:27:54,032|azureml._restclient.clientbase|INFO|Created a worker pool for first use\\n2021-05-15 12:27:54,032|azureml.core.authentication|DEBUG|Time to expire 1814337.96772 seconds\\n2021-05-15 12:27:54,032|azureml._restclient.service_context|DEBUG|Created a static thread pool for ServiceContext class\\n2021-05-15 12:27:54,033|azureml.core.authentication|DEBUG|Time to expire 1814337.966659 seconds\\n2021-05-15 12:27:54,034|azureml.core.authentication|DEBUG|Time to expire 1814337.965749 seconds\\n2021-05-15 12:27:54,034|azureml._restclient.service_context|DEBUG|Access an existing static threadpool for ServiceContext class\\n2021-05-15 12:27:54,034|azureml._restclient.clientbase|DEBUG|ClientBase: Calling get with url None\\n2021-05-15 12:27:54,034|azureml._restclient.service_context|DEBUG|Access an existing static threadpool for ServiceContext class\\n2021-05-15 12:27:54,066|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,067|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,067|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,068|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,068|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,069|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,069|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,069|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,070|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,070|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,070|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,070|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,071|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,071|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,071|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,072|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,072|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,072|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,079|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,079|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,079|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:27:54,148|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.RunClient.get_by_exp_id-async:False|DEBUG|[START]\\n2021-05-15 12:27:54,156|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.RunClient.get_by_exp_id-async:False|DEBUG|[START]\\n2021-05-15 12:27:54,163|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.RunClient.get_by_exp_id-async:False|DEBUG|[START]\\n2021-05-15 12:27:54,163|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.RunClient|DEBUG|ClientBase: Calling get_by_exp_id with url /history/v1.0/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.MachineLearningServices/workspaces/{workspaceName}/experimentids/{experimentId}/runs/{runId}\\n2021-05-15 12:27:54,163|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.RunClient|DEBUG|ClientBase: Calling get_by_exp_id with url /history/v1.0/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.MachineLearningServices/workspaces/{workspaceName}/experimentids/{experimentId}/runs/{runId}\\n2021-05-15 12:27:54,163|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.RunClient|DEBUG|ClientBase: Calling get_by_exp_id with url /history/v1.0/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.MachineLearningServices/workspaces/{workspaceName}/experimentids/{experimentId}/runs/{runId}\\n2021-05-15 12:27:54,232|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.RunClient.get_by_exp_id-async:False|DEBUG|[STOP]\\n2021-05-15 12:27:54,233|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6|DEBUG|Constructing run from dto. type: azureml.scriptrun, source: None, props: {'_azureml.ComputeTargetType': 'amlcompute', 'ContentSnapshotId': '52295cdf-7227-4333-a1c9-d560485f0190', 'azureml.git.repository_uri': 'https://github.com/sriksmachi/octopus.git', 'mlflow.source.git.repoURL': 'https://github.com/sriksmachi/octopus.git', 'azureml.git.branch': 'master', 'mlflow.source.git.branch': 'master', 'azureml.git.commit': '350a13c17e02d0a6b1e1cbc41d88390723744e28', 'mlflow.source.git.commit': '350a13c17e02d0a6b1e1cbc41d88390723744e28', 'azureml.git.dirty': 'False', 'ProcessInfoFile': 'azureml-logs/process_info.json', 'ProcessStatusFile': 'azureml-logs/process_status.json'}\\n2021-05-15 12:27:54,233|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunContextManager|DEBUG|Valid logs dir, setting up content loader\\n2021-05-15 12:27:54,235|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.RunClient.get_by_exp_id-async:False|DEBUG|[STOP]\\n2021-05-15 12:27:54,236|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6|DEBUG|Constructing run from dto. type: azureml.scriptrun, source: None, props: {'_azureml.ComputeTargetType': 'amlcompute', 'ContentSnapshotId': '52295cdf-7227-4333-a1c9-d560485f0190', 'azureml.git.repository_uri': 'https://github.com/sriksmachi/octopus.git', 'mlflow.source.git.repoURL': 'https://github.com/sriksmachi/octopus.git', 'azureml.git.branch': 'master', 'mlflow.source.git.branch': 'master', 'azureml.git.commit': '350a13c17e02d0a6b1e1cbc41d88390723744e28', 'mlflow.source.git.commit': '350a13c17e02d0a6b1e1cbc41d88390723744e28', 'azureml.git.dirty': 'False', 'ProcessInfoFile': 'azureml-logs/process_info.json', 'ProcessStatusFile': 'azureml-logs/process_status.json'}\\n2021-05-15 12:27:54,237|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.RunClient.get_by_exp_id-async:False|DEBUG|[STOP]\\n2021-05-15 12:27:54,238|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunContextManager|DEBUG|Valid logs dir, setting up content loader\\n2021-05-15 12:27:54,238|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6|DEBUG|Constructing run from dto. type: azureml.scriptrun, source: None, props: {'_azureml.ComputeTargetType': 'amlcompute', 'ContentSnapshotId': '52295cdf-7227-4333-a1c9-d560485f0190', 'azureml.git.repository_uri': 'https://github.com/sriksmachi/octopus.git', 'mlflow.source.git.repoURL': 'https://github.com/sriksmachi/octopus.git', 'azureml.git.branch': 'master', 'mlflow.source.git.branch': 'master', 'azureml.git.commit': '350a13c17e02d0a6b1e1cbc41d88390723744e28', 'mlflow.source.git.commit': '350a13c17e02d0a6b1e1cbc41d88390723744e28', 'azureml.git.dirty': 'False', 'ProcessInfoFile': 'azureml-logs/process_info.json', 'ProcessStatusFile': 'azureml-logs/process_status.json'}\\n2021-05-15 12:27:54,239|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunContextManager|DEBUG|Valid logs dir, setting up content loader\\n2021-05-15 12:27:54,239|azureml|WARNING|Could not import azureml.mlflow or azureml.contrib.mlflow mlflow APIs will not run against AzureML services.  Add azureml-mlflow as a conda dependency for the run if this behavior is desired\\n2021-05-15 12:27:54,239|azureml.WorkerPool|DEBUG|[START]\\n2021-05-15 12:27:54,239|azureml.SendRunKillSignal|DEBUG|[START]\\n2021-05-15 12:27:54,239|azureml.RunStatusContext|DEBUG|[START]\\n2021-05-15 12:27:54,240|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunContextManager.RunStatusContext|DEBUG|[START]\\n2021-05-15 12:27:54,240|azureml.MetricsClient|DEBUG|[START]\\n2021-05-15 12:27:54,241|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient|DEBUG|[START]\\n2021-05-15 12:27:54,241|azureml.WorkingDirectoryCM|DEBUG|[START]\\n2021-05-15 12:27:54,241|azureml.history._tracking.PythonWorkingDirectory.workingdir|DEBUG|[START]\\n2021-05-15 12:27:54,241|azureml.history._tracking.PythonWorkingDirectory|INFO|Current working dir: /mnt/batch/tasks/shared/LS_root/jobs/ml-workbench-eu/azureml/tf-effusion-detector_1621081608_4fe7f8b6/mounts/workspaceblobstore/azureml/tf-effusion-detector_1621081608_4fe7f8b6\\n2021-05-15 12:27:54,241|azureml.history._tracking.PythonWorkingDirectory.workingdir|DEBUG|Calling pyfs\\n2021-05-15 12:27:54,241|azureml.history._tracking.PythonWorkingDirectory.workingdir|DEBUG|Storing working dir for pyfs as /mnt/batch/tasks/shared/LS_root/jobs/ml-workbench-eu/azureml/tf-effusion-detector_1621081608_4fe7f8b6/mounts/workspaceblobstore/azureml/tf-effusion-detector_1621081608_4fe7f8b6\\n2021-05-15 12:28:24,018|azureml.core.authentication|DEBUG|Time to expire 1814307.981246 seconds\\n2021-05-15 12:28:30,830|azureml.core.run|DEBUG|Identity in use is not set. Falling back to using AMLToken\\n2021-05-15 12:28:30,830|azureml.core.run|DEBUG|Using AMLToken auth for remote run\\n2021-05-15 12:28:30,830|azureml._restclient.service_context|DEBUG|Access an existing static threadpool for ServiceContext class\\n2021-05-15 12:28:30,831|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:28:30,831|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:28:30,832|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:28:30,832|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:28:30,832|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:28:30,832|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:28:30,832|azureml._base_sdk_common.service_discovery|DEBUG|Found history service url in environment variable AZUREML_SERVICE_ENDPOINT, history service url: https://eastus.api.azureml.ms.\\n2021-05-15 12:28:54,033|azureml.core.authentication|DEBUG|Time to expire 1814277.96615 seconds\\n2021-05-15 12:29:21,145|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient|DEBUG|Overrides: Max batch size: 50, batch cushion: 5, Interval: 1.\\n2021-05-15 12:29:21,145|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient.PostMetricsBatchV2.PostMetricsBatchV2Daemon|DEBUG|Starting daemon and triggering first instance\\n2021-05-15 12:29:21,146|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient|DEBUG|Used <class 'azureml._common.async_utils.batch_task_queue.BatchTaskQueue'> for use_batch=True.\\n2021-05-15 12:29:22,146|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|[Start]\\n2021-05-15 12:29:22,146|azureml.BatchTaskQueueAdd_1_Batches.WorkerPool|DEBUG|submitting future: _handle_batch\\n2021-05-15 12:29:22,147|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient.PostMetricsBatchV2|DEBUG|Batch size 3.\\n2021-05-15 12:29:22,147|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch|DEBUG|Using basic handler - no exception handling\\n2021-05-15 12:29:22,147|azureml._restclient.service_context.WorkerPool|DEBUG|submitting future: _log_batch_v2\\n2021-05-15 12:29:22,147|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|Adding task 0__handle_batch to queue of approximate size: 0\\n2021-05-15 12:29:22,148|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient|DEBUG|Metrics Client: _log_batch_v2 is calling post_run_metrics posting 3 values.\\n2021-05-15 12:29:22,148|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|[Stop] - waiting default timeout\\n2021-05-15 12:29:22,148|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient.PostMetricsBatchV2.0__log_batch_v2|DEBUG|Using basic handler - no exception handling\\n2021-05-15 12:29:22,148|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient._post_run_metrics_log_failed_validations-async:False|DEBUG|[START]\\n2021-05-15 12:29:22,149|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|[START]\\n2021-05-15 12:29:22,149|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient.PostMetricsBatchV2|DEBUG|Adding task 0__log_batch_v2 to queue of approximate size: 0\\n2021-05-15 12:29:22,149|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient|DEBUG|ClientBase: Calling _post_run_metrics_log_failed_validations with url None\\n2021-05-15 12:29:22,149|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|Overriding default flush timeout from None to 120\\n2021-05-15 12:29:22,156|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|Waiting 120 seconds on tasks: [AsyncTask(0__handle_batch)].\\n2021-05-15 12:29:22,158|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch.WaitingTask|DEBUG|[START]\\n2021-05-15 12:29:22,158|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch.WaitingTask|DEBUG|Awaiter is BatchTaskQueueAdd_1_Batches\\n2021-05-15 12:29:22,158|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch.WaitingTask|DEBUG|[STOP]\\n2021-05-15 12:29:22,158|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|\\n2021-05-15 12:29:22,158|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|[STOP]\\n2021-05-15 12:29:22,429|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient._post_run_metrics_log_failed_validations-async:False|DEBUG|[STOP]\\n2021-05-15 12:29:24,035|azureml.core.authentication|DEBUG|Time to expire 1814247.96449 seconds\\n2021-05-15 12:29:54,035|azureml.core.authentication|DEBUG|Time to expire 1814217.964154 seconds\\n2021-05-15 12:30:03,492|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|[Start]\\n2021-05-15 12:30:03,492|azureml.BatchTaskQueueAdd_1_Batches.WorkerPool|DEBUG|submitting future: _handle_batch\\n2021-05-15 12:30:03,493|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient.PostMetricsBatchV2|DEBUG|Batch size 3.\\n2021-05-15 12:30:03,493|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch|DEBUG|Using basic handler - no exception handling\\n2021-05-15 12:30:03,493|azureml._restclient.service_context.WorkerPool|DEBUG|submitting future: _log_batch_v2\\n2021-05-15 12:30:03,494|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|Adding task 0__handle_batch to queue of approximate size: 0\\n2021-05-15 12:30:03,495|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient|DEBUG|Metrics Client: _log_batch_v2 is calling post_run_metrics posting 3 values.\\n2021-05-15 12:30:03,495|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient.PostMetricsBatchV2.1__log_batch_v2|DEBUG|Using basic handler - no exception handling\\n2021-05-15 12:30:03,495|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|[Stop] - waiting default timeout\\n2021-05-15 12:30:03,496|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient._post_run_metrics_log_failed_validations-async:False|DEBUG|[START]\\n2021-05-15 12:30:03,496|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient.PostMetricsBatchV2|DEBUG|Adding task 1__log_batch_v2 to queue of approximate size: 1\\n2021-05-15 12:30:03,497|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|[START]\\n2021-05-15 12:30:03,497|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient|DEBUG|ClientBase: Calling _post_run_metrics_log_failed_validations with url None\\n2021-05-15 12:30:03,498|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|Overriding default flush timeout from None to 120\\n2021-05-15 12:30:03,500|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|Waiting 120 seconds on tasks: [AsyncTask(0__handle_batch)].\\n2021-05-15 12:30:03,503|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch.WaitingTask|DEBUG|[START]\\n2021-05-15 12:30:03,503|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch.WaitingTask|DEBUG|Awaiter is BatchTaskQueueAdd_1_Batches\\n2021-05-15 12:30:03,504|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch.WaitingTask|DEBUG|[STOP]\\n2021-05-15 12:30:03,504|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|\\n2021-05-15 12:30:03,504|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|[STOP]\\n2021-05-15 12:30:03,630|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient._post_run_metrics_log_failed_validations-async:False|DEBUG|[STOP]\\n2021-05-15 12:30:24,036|azureml.core.authentication|DEBUG|Time to expire 1814187.96383 seconds\\n2021-05-15 12:30:45,964|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|[Start]\\n2021-05-15 12:30:45,964|azureml.BatchTaskQueueAdd_1_Batches.WorkerPool|DEBUG|submitting future: _handle_batch\\n2021-05-15 12:30:45,964|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient.PostMetricsBatchV2|DEBUG|Batch size 3.\\n2021-05-15 12:30:45,964|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch|DEBUG|Using basic handler - no exception handling\\n2021-05-15 12:30:45,965|azureml._restclient.service_context.WorkerPool|DEBUG|submitting future: _log_batch_v2\\n2021-05-15 12:30:45,965|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|Adding task 0__handle_batch to queue of approximate size: 0\\n2021-05-15 12:30:45,966|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|[Stop] - waiting default timeout\\n2021-05-15 12:30:45,966|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient|DEBUG|Metrics Client: _log_batch_v2 is calling post_run_metrics posting 3 values.\\n2021-05-15 12:30:45,966|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|[START]\\n2021-05-15 12:30:45,967|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient.PostMetricsBatchV2.2__log_batch_v2|DEBUG|Using basic handler - no exception handling\\n2021-05-15 12:30:45,967|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient._post_run_metrics_log_failed_validations-async:False|DEBUG|[START]\\n2021-05-15 12:30:45,967|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|Overriding default flush timeout from None to 120\\n2021-05-15 12:30:45,968|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient.PostMetricsBatchV2|DEBUG|Adding task 2__log_batch_v2 to queue of approximate size: 2\\n2021-05-15 12:30:45,969|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient|DEBUG|ClientBase: Calling _post_run_metrics_log_failed_validations with url None\\n2021-05-15 12:30:45,969|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|Waiting 120 seconds on tasks: [AsyncTask(0__handle_batch)].\\n2021-05-15 12:30:45,971|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch.WaitingTask|DEBUG|[START]\\n2021-05-15 12:30:45,974|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch.WaitingTask|DEBUG|Awaiter is BatchTaskQueueAdd_1_Batches\\n2021-05-15 12:30:45,974|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch.WaitingTask|DEBUG|[STOP]\\n2021-05-15 12:30:45,975|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|\\n2021-05-15 12:30:45,975|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|[STOP]\\n2021-05-15 12:30:46,097|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient._post_run_metrics_log_failed_validations-async:False|DEBUG|[STOP]\\n2021-05-15 12:30:54,045|azureml.core.authentication|DEBUG|Time to expire 1814157.954432 seconds\\n2021-05-15 12:31:24,045|azureml.core.authentication|DEBUG|Time to expire 1814127.954206 seconds\\n2021-05-15 12:31:28,415|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|[Start]\\n2021-05-15 12:31:28,415|azureml.BatchTaskQueueAdd_1_Batches.WorkerPool|DEBUG|submitting future: _handle_batch\\n2021-05-15 12:31:28,415|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient.PostMetricsBatchV2|DEBUG|Batch size 3.\\n2021-05-15 12:31:28,416|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch|DEBUG|Using basic handler - no exception handling\\n2021-05-15 12:31:28,416|azureml._restclient.service_context.WorkerPool|DEBUG|submitting future: _log_batch_v2\\n2021-05-15 12:31:28,417|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|Adding task 0__handle_batch to queue of approximate size: 0\\n2021-05-15 12:31:28,417|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient|DEBUG|Metrics Client: _log_batch_v2 is calling post_run_metrics posting 3 values.\\n2021-05-15 12:31:28,418|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient.PostMetricsBatchV2.3__log_batch_v2|DEBUG|Using basic handler - no exception handling\\n2021-05-15 12:31:28,418|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|[Stop] - waiting default timeout\\n2021-05-15 12:31:28,419|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient._post_run_metrics_log_failed_validations-async:False|DEBUG|[START]\\n2021-05-15 12:31:28,419|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient.PostMetricsBatchV2|DEBUG|Adding task 3__log_batch_v2 to queue of approximate size: 3\\n2021-05-15 12:31:28,420|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|[START]\\n2021-05-15 12:31:28,420|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient|DEBUG|ClientBase: Calling _post_run_metrics_log_failed_validations with url None\\n2021-05-15 12:31:28,422|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|Overriding default flush timeout from None to 120\\n2021-05-15 12:31:28,426|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|Waiting 120 seconds on tasks: [AsyncTask(0__handle_batch)].\\n2021-05-15 12:31:28,427|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch.WaitingTask|DEBUG|[START]\\n2021-05-15 12:31:28,427|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch.WaitingTask|DEBUG|Awaiter is BatchTaskQueueAdd_1_Batches\\n2021-05-15 12:31:28,427|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch.WaitingTask|DEBUG|[STOP]\\n2021-05-15 12:31:28,427|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|\\n2021-05-15 12:31:28,428|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|[STOP]\\n2021-05-15 12:31:28,551|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient._post_run_metrics_log_failed_validations-async:False|DEBUG|[STOP]\\n2021-05-15 12:31:54,053|azureml.core.authentication|DEBUG|Time to expire 1814097.946308 seconds\\n2021-05-15 12:32:10,637|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|[Start]\\n2021-05-15 12:32:10,637|azureml.BatchTaskQueueAdd_1_Batches.WorkerPool|DEBUG|submitting future: _handle_batch\\n2021-05-15 12:32:10,649|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient.PostMetricsBatchV2|DEBUG|Batch size 3.\\n2021-05-15 12:32:10,649|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch|DEBUG|Using basic handler - no exception handling\\n2021-05-15 12:32:10,649|azureml._restclient.service_context.WorkerPool|DEBUG|submitting future: _log_batch_v2\\n2021-05-15 12:32:10,650|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|Adding task 0__handle_batch to queue of approximate size: 0\\n2021-05-15 12:32:10,650|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient|DEBUG|Metrics Client: _log_batch_v2 is calling post_run_metrics posting 3 values.\\n2021-05-15 12:32:10,650|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|[Stop] - waiting default timeout\\n2021-05-15 12:32:10,651|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient._post_run_metrics_log_failed_validations-async:False|DEBUG|[START]\\n2021-05-15 12:32:10,651|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|[START]\\n2021-05-15 12:32:10,651|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient|DEBUG|ClientBase: Calling _post_run_metrics_log_failed_validations with url None\\n2021-05-15 12:32:10,651|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|Overriding default flush timeout from None to 120\\n2021-05-15 12:32:10,656|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|Waiting 120 seconds on tasks: [AsyncTask(0__handle_batch)].\\n2021-05-15 12:32:10,665|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient.PostMetricsBatchV2.4__log_batch_v2|DEBUG|Using basic handler - no exception handling\\n2021-05-15 12:32:10,666|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient.PostMetricsBatchV2|DEBUG|Adding task 4__log_batch_v2 to queue of approximate size: 4\\n2021-05-15 12:32:10,758|azureml._SubmittedRun#tf-effusion-detector_1621081608_4fe7f8b6.RunHistoryFacade.MetricsClient._post_run_metrics_log_failed_validations-async:False|DEBUG|[STOP]\\n2021-05-15 12:32:10,907|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch.WaitingTask|DEBUG|[START]\\n2021-05-15 12:32:10,907|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch.WaitingTask|DEBUG|Awaiter is BatchTaskQueueAdd_1_Batches\\n2021-05-15 12:32:10,907|azureml.BatchTaskQueueAdd_1_Batches.0__handle_batch.WaitingTask|DEBUG|[STOP]\\n2021-05-15 12:32:10,907|azureml.BatchTaskQueueAdd_1_Batches|DEBUG|Waiting on task: 0__handle_batch.\\n1 tasks left. Current duration of flush 0.0002105236053466797 seconds.\\n\\n2021-05-15 12:32:10,907|azureml.BatchTaskQueueAdd_1_Batches.WaitFlushSource:BatchTaskQueueAdd_1_Batches|DEBUG|[STOP]\\n2021-05-15 12:32:24,070|azureml.core.authentication|DEBUG|Time to expire 1814067.929454 seconds\\n\", \"graph\": {}, \"widget_settings\": {\"childWidgetDisplay\": \"popup\", \"send_telemetry\": false, \"log_level\": \"NOTSET\", \"sdk_version\": \"1.27.0\"}, \"loading\": false}"
          },
          "metadata": {}
        }
      ],
      "execution_count": 34,
      "metadata": {
        "collapsed": true,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "gather": {
          "logged": 1621081577629
        }
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 5. Distributed Machine Learning using Horovod\r\n",
        "\r\n",
        "The top 2 ways of training deep learning models in distributed fashion is using data-parallel and model-parallel approaches.\r\n",
        "Currently the most common scenario is data-parallel training with synchronous updates. This approach is the easiest to implement and is sufficient for most use cases.\r\n",
        "\r\n",
        "In this section we will convert the training script run on distributed ML environment created using Horovod.\r\n",
        "\r\n",
        "__What is Horovod?__\r\n",
        "\r\n",
        "Horovod is a distributed deep learning training framework for TensorFlow, Keras, PyTorch, and Apache MXNet. The goal of Horovod is to make distributed deep learning fast and easy to use.\r\n",
        "\r\n",
        "https://github.com/horovod/horovod\r\n",
        "\r\n",
        "\r\n",
        "\r\n",
        "\r\n"
      ],
      "metadata": {
        "nteract": {
          "transient": {
            "deleting": false
          }
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import os\r\n",
        "from azureml.core import ScriptRunConfig\r\n",
        "import shutil\r\n",
        "from azureml.core import Experiment\r\n",
        "from azureml.core import Environment\r\n",
        "from azureml.core.conda_dependencies import CondaDependencies\r\n",
        "from azureml.widgets import RunDetails\r\n",
        "from azureml.core.runconfig import MpiConfiguration\r\n",
        "\r\n",
        "## The training script capable of running in distributed environment is extracted to effusion_detector_distributed.py\r\n",
        "project_folder = './effusion_detector-distributed'\r\n",
        "os.makedirs(project_folder, exist_ok=True)\r\n",
        "shutil.copy('effusion_detector_distributed.py', project_folder)\r\n",
        "\r\n",
        "## choosing an experiment name\r\n",
        "experiment_name = 'tf-distributed-effusion-detector'\r\n",
        "experiment = Experiment(ws, name=experiment_name)\r\n",
        "\r\n",
        "# loading the env dependencies from conda configuration\r\n",
        "os.makedirs('./envs', exist_ok=True)\r\n",
        "tf_env = Environment.from_conda_specification(name=\"imageclassification\", file_path=\"envs/distributed-tensorflow-with-horovod/conda_dependencies.yml\")\r\n",
        "\r\n",
        "# Specify a GPU base image\r\n",
        "tf_env.docker.enabled = True\r\n",
        "tf_env.docker.base_image = 'mcr.microsoft.com/azureml/openmpi3.1.2-cuda10.1-cudnn7-ubuntu18.04'\r\n",
        "\r\n",
        "# using a cluster which can autoscale uptp 4 nodes.\r\n",
        "cluster_name = \"gpus\"\r\n",
        "compute_target = ComputeTarget(workspace=ws, name=cluster_name)\r\n",
        "\r\n",
        "# running the script\r\n",
        "args = ['--data-folder', dataset.as_mount(), '--epochs', 20]\r\n",
        "src = ScriptRunConfig(source_directory=project_folder,\r\n",
        "                      script='effusion_detector_distributed.py',\r\n",
        "                      arguments=args,\r\n",
        "                      compute_target=compute_target,\r\n",
        "                      environment=tf_env, distributed_job_config=MpiConfiguration(node_count=4))\r\n",
        "\r\n",
        "run = experiment.submit(src)\r\n",
        "print(run)\r\n",
        "run.get_details()"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "'enabled' is deprecated. Please use the azureml.core.runconfig.DockerConfiguration object with the 'use_docker' param instead.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Run(Experiment: tf-distributed-effusion-detector,\n",
            "Id: tf-distributed-effusion-detector_1621081615_407d41d4,\n",
            "Type: azureml.scriptrun,\n",
            "Status: Preparing)\n"
          ]
        },
        {
          "output_type": "execute_result",
          "execution_count": 35,
          "data": {
            "text/plain": "{'runId': 'tf-distributed-effusion-detector_1621081615_407d41d4',\n 'target': 'gpus',\n 'status': 'Preparing',\n 'properties': {'_azureml.ComputeTargetType': 'amlcompute',\n  'ContentSnapshotId': 'acc16e72-9489-423f-bd90-778e444e17f5',\n  'azureml.git.repository_uri': 'https://github.com/sriksmachi/octopus.git',\n  'mlflow.source.git.repoURL': 'https://github.com/sriksmachi/octopus.git',\n  'azureml.git.branch': 'master',\n  'mlflow.source.git.branch': 'master',\n  'azureml.git.commit': '350a13c17e02d0a6b1e1cbc41d88390723744e28',\n  'mlflow.source.git.commit': '350a13c17e02d0a6b1e1cbc41d88390723744e28',\n  'azureml.git.dirty': 'False'},\n 'inputDatasets': [{'dataset': {'id': '936fc596-dc32-48e8-8ba5-dcbc3bf2b310'}, 'consumptionDetails': {'type': 'RunInput', 'inputName': 'input__99e4f6dd', 'mechanism': 'Mount'}}],\n 'outputDatasets': [],\n 'runDefinition': {'script': 'effusion_detector_distributed.py',\n  'command': '',\n  'useAbsolutePath': False,\n  'arguments': ['--data-folder',\n   'DatasetConsumptionConfig:input__99e4f6dd',\n   '--epochs',\n   '20'],\n  'sourceDirectoryDataStore': None,\n  'framework': 'Python',\n  'communicator': 'Mpi',\n  'target': 'gpus',\n  'dataReferences': {},\n  'data': {'input__99e4f6dd': {'dataLocation': {'dataset': {'id': '936fc596-dc32-48e8-8ba5-dcbc3bf2b310',\n      'name': None,\n      'version': None},\n     'dataPath': None,\n     'uri': None},\n    'mechanism': 'Mount',\n    'environmentVariableName': 'input__99e4f6dd',\n    'pathOnCompute': None,\n    'overwrite': False}},\n  'outputData': {},\n  'jobName': None,\n  'maxRunDurationSeconds': 2592000,\n  'nodeCount': 4,\n  'priority': None,\n  'credentialPassthrough': False,\n  'identity': None,\n  'environment': {'name': 'imageclassification',\n   'version': 'Autosave_2021-05-09T18:48:30Z_cbfc4fde',\n   'python': {'interpreterPath': 'python',\n    'userManagedDependencies': False,\n    'condaDependencies': {'channels': ['conda-forge'],\n     'dependencies': ['python=3.6.2',\n      {'pip': ['azureml-core==1.25.0',\n        'azureml-defaults==1.25.0',\n        'azureml-telemetry==1.25.0',\n        'azureml-train-restclients-hyperdrive==1.25.0',\n        'azureml-train-core==1.25.0',\n        'opencv-python==3.3.0.9',\n        'azureml-sdk',\n        'azureml-widgets',\n        'keras',\n        'tensorflow-gpu==2.2.0',\n        'matplotlib',\n        'pandas',\n        'torch==1.6.0',\n        'torchvision==0.7.0',\n        'horovod==0.19.5',\n        'sklearn']}],\n     'name': 'azureml_06e6d8e54a94c28c30dca2efbf811a2d'},\n    'baseCondaEnvironment': None},\n   'environmentVariables': {'EXAMPLE_ENV_VAR': 'EXAMPLE_VALUE'},\n   'docker': {'baseImage': 'mcr.microsoft.com/azureml/openmpi3.1.2-cuda10.1-cudnn7-ubuntu18.04',\n    'platform': {'os': 'Linux', 'architecture': 'amd64'},\n    'baseDockerfile': None,\n    'dockerContext': None,\n    'baseImageRegistry': {'address': None, 'username': None, 'password': None},\n    'enabled': True,\n    'arguments': []},\n   'spark': {'repositories': [], 'packages': [], 'precachePackages': True},\n   'inferencingStackVersion': None},\n  'history': {'outputCollection': True,\n   'directoriesToWatch': ['logs'],\n   'enableMLflowTracking': True,\n   'snapshotProject': True},\n  'spark': {'configuration': {'spark.app.name': 'Azure ML Experiment',\n    'spark.yarn.maxAppAttempts': '1'}},\n  'parallelTask': {'maxRetriesPerWorker': 0,\n   'workerCountPerNode': 1,\n   'terminalExitCodes': None,\n   'configuration': {}},\n  'amlCompute': {'name': None,\n   'vmSize': None,\n   'retainCluster': False,\n   'clusterMaxNodeCount': None},\n  'aiSuperComputer': {'instanceType': None,\n   'imageVersion': None,\n   'location': None,\n   'aiSuperComputerStorageData': None,\n   'interactive': False,\n   'scalePolicy': None,\n   'virtualClusterArmId': None,\n   'tensorboardLogDirectory': None,\n   'sshPublicKey': None},\n  'tensorflow': {'workerCount': 1, 'parameterServerCount': 1},\n  'mpi': {'processCountPerNode': 1},\n  'pyTorch': {'communicationBackend': 'nccl', 'processCount': None},\n  'hdi': {'yarnDeployMode': 'Cluster'},\n  'containerInstance': {'region': None, 'cpuCores': 2.0, 'memoryGb': 3.5},\n  'exposedPorts': None,\n  'docker': {'useDocker': True,\n   'sharedVolumes': True,\n   'shmSize': '2g',\n   'arguments': []},\n  'cmk8sCompute': {'configuration': {}},\n  'commandReturnCodeConfig': {'returnCode': 'Zero',\n   'successfulReturnCodes': []},\n  'environmentVariables': {},\n  'applicationEndpoints': {}},\n 'logFiles': {},\n 'submittedBy': 'Srikanth Machiraju'}"
          },
          "metadata": {}
        }
      ],
      "execution_count": 35,
      "metadata": {
        "collapsed": true,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "gather": {
          "logged": 1621081581195
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from azureml.widgets import RunDetails\r\n",
        "RunDetails(run).show()"
      ],
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "_UserRunWidget(widget_settings={'childWidgetDisplay': 'popup', 'send_telemetry': False, 'log_level': 'NOTSET',…",
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "38fed3cb8e114370b5cd65c3630722e9"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/aml.mini.widget.v1": "{\"status\": \"Running\", \"workbench_run_details_uri\": \"https://ml.azure.com/runs/tf-distributed-effusion-detector_1621081615_407d41d4?wsid=/subscriptions/925a5ad8-a21a-48c0-92a2-5a8a2a4dfc46/resourcegroups/ml-workbench/workspaces/ml-workbench-eu&tid=72f988bf-86f1-41af-91ab-2d7cd011db47\", \"run_id\": \"tf-distributed-effusion-detector_1621081615_407d41d4\", \"run_properties\": {\"run_id\": \"tf-distributed-effusion-detector_1621081615_407d41d4\", \"created_utc\": \"2021-05-15T12:26:57.613307Z\", \"properties\": {\"_azureml.ComputeTargetType\": \"amlcompute\", \"ContentSnapshotId\": \"acc16e72-9489-423f-bd90-778e444e17f5\", \"azureml.git.repository_uri\": \"https://github.com/sriksmachi/octopus.git\", \"mlflow.source.git.repoURL\": \"https://github.com/sriksmachi/octopus.git\", \"azureml.git.branch\": \"master\", \"mlflow.source.git.branch\": \"master\", \"azureml.git.commit\": \"350a13c17e02d0a6b1e1cbc41d88390723744e28\", \"mlflow.source.git.commit\": \"350a13c17e02d0a6b1e1cbc41d88390723744e28\", \"azureml.git.dirty\": \"False\", \"ProcessInfoFile\": \"azureml-logs/process_info.json\", \"ProcessStatusFile\": \"azureml-logs/process_status.json\"}, \"tags\": {\"_aml_system_ComputeTargetStatus\": \"{\\\"AllocationState\\\":\\\"steady\\\",\\\"PreparingNodeCount\\\":0,\\\"RunningNodeCount\\\":0,\\\"CurrentNodeCount\\\":0}\"}, \"script_name\": null, \"arguments\": null, \"end_time_utc\": null, \"status\": \"Running\", \"log_files\": {\"azureml-logs/55_azureml-execution-tvmps_522962d35ec9e649ca9dfd0c74a0bc6fe2aac5a9197cf41cf503c5c81f7134d8_p.txt\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/azureml-logs/55_azureml-execution-tvmps_522962d35ec9e649ca9dfd0c74a0bc6fe2aac5a9197cf41cf503c5c81f7134d8_p.txt?sv=2019-02-02&sr=b&sig=9Gmt%2BnPXdKQSsU6EpJm7%2BUHqZB2nojVqRjPIubNHP6A%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"azureml-logs/55_azureml-execution-tvmps_c6e6d2858679b9c78371ef9b5664c6a9aeefef5840ad0d3aee9b4180a75f95b5_p.txt\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/azureml-logs/55_azureml-execution-tvmps_c6e6d2858679b9c78371ef9b5664c6a9aeefef5840ad0d3aee9b4180a75f95b5_p.txt?sv=2019-02-02&sr=b&sig=22ZTuMHuHh2N7CiahTngA5GN6COd5oeJ3IIUjQiQ9gk%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"azureml-logs/55_azureml-execution-tvmps_deca333c28ad0ee2af61010e7f808644c8cbc0760dc094cae395b58102473b22_p.txt\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/azureml-logs/55_azureml-execution-tvmps_deca333c28ad0ee2af61010e7f808644c8cbc0760dc094cae395b58102473b22_p.txt?sv=2019-02-02&sr=b&sig=1T0i1PqpBNoWJSeZSZt%2BHB%2FVEmx6FWscnqVkH4kh81k%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"azureml-logs/55_azureml-execution-tvmps_f0f5b814e7ee3abf3f56f97236c9892d5d99fbbfe5144209416f780fa80a7a1f_p.txt\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/azureml-logs/55_azureml-execution-tvmps_f0f5b814e7ee3abf3f56f97236c9892d5d99fbbfe5144209416f780fa80a7a1f_p.txt?sv=2019-02-02&sr=b&sig=rxBMmu1wXdZUub3nF50FICDGfVAL%2FNYaJ2hE14DNC8M%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"azureml-logs/65_job_prep-tvmps_522962d35ec9e649ca9dfd0c74a0bc6fe2aac5a9197cf41cf503c5c81f7134d8_p.txt\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/azureml-logs/65_job_prep-tvmps_522962d35ec9e649ca9dfd0c74a0bc6fe2aac5a9197cf41cf503c5c81f7134d8_p.txt?sv=2019-02-02&sr=b&sig=Y3I1L60JI70ktbZNackYM%2BSTPs74meixjaydjn6JBY4%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"azureml-logs/65_job_prep-tvmps_c6e6d2858679b9c78371ef9b5664c6a9aeefef5840ad0d3aee9b4180a75f95b5_p.txt\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/azureml-logs/65_job_prep-tvmps_c6e6d2858679b9c78371ef9b5664c6a9aeefef5840ad0d3aee9b4180a75f95b5_p.txt?sv=2019-02-02&sr=b&sig=Kjz3vOopfH8sMwd35H5sw%2BYbIdnY6Een%2FqvBKHBFXYg%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"azureml-logs/65_job_prep-tvmps_deca333c28ad0ee2af61010e7f808644c8cbc0760dc094cae395b58102473b22_p.txt\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/azureml-logs/65_job_prep-tvmps_deca333c28ad0ee2af61010e7f808644c8cbc0760dc094cae395b58102473b22_p.txt?sv=2019-02-02&sr=b&sig=d9XsWIwP2xhHFnt6BSWZxqhJ4tTgOx%2FzS2qmVAdWhNA%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"azureml-logs/65_job_prep-tvmps_f0f5b814e7ee3abf3f56f97236c9892d5d99fbbfe5144209416f780fa80a7a1f_p.txt\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/azureml-logs/65_job_prep-tvmps_f0f5b814e7ee3abf3f56f97236c9892d5d99fbbfe5144209416f780fa80a7a1f_p.txt?sv=2019-02-02&sr=b&sig=OK3rwknAgv%2FoQlcm%2FGdTD1nSLLYH5qnHtOQ%2FCdgeVrY%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"logs/azureml/dataprep/backgroundProcess.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/logs/azureml/dataprep/backgroundProcess.log?sv=2019-02-02&sr=b&sig=5v1gXiW%2BydDD28llVhRYgqEB9pfoDIx%2FuCtir%2FN%2FHOo%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"logs/azureml/dataprep/backgroundProcess_Telemetry.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/logs/azureml/dataprep/backgroundProcess_Telemetry.log?sv=2019-02-02&sr=b&sig=hAIpAxx17IHoHpfyH%2BiBt8%2FX2bRbU6U5kPrS%2B6UbtZc%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"logs/azureml/job_prep_azureml.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/logs/azureml/job_prep_azureml.log?sv=2019-02-02&sr=b&sig=9pBMHgdYG2y%2F1x0z4G6tWCK4HtuqaynoPDaiHWjqoNk%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"logs/azureml/sidecar/tvmps_522962d35ec9e649ca9dfd0c74a0bc6fe2aac5a9197cf41cf503c5c81f7134d8_p/all.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/logs/azureml/sidecar/tvmps_522962d35ec9e649ca9dfd0c74a0bc6fe2aac5a9197cf41cf503c5c81f7134d8_p/all.log?sv=2019-02-02&sr=b&sig=JwYtAm%2F2yZju12y%2B4IzRQvAM4%2FzcjMu7OibZmsLfols%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"logs/azureml/sidecar/tvmps_522962d35ec9e649ca9dfd0c74a0bc6fe2aac5a9197cf41cf503c5c81f7134d8_p/task.enter_contexts.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/logs/azureml/sidecar/tvmps_522962d35ec9e649ca9dfd0c74a0bc6fe2aac5a9197cf41cf503c5c81f7134d8_p/task.enter_contexts.log?sv=2019-02-02&sr=b&sig=Wz6hxH1Zl0nD3dOYrIIeYjj5R0w5Di2iEGbx8wmXQFI%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"logs/azureml/sidecar/tvmps_c6e6d2858679b9c78371ef9b5664c6a9aeefef5840ad0d3aee9b4180a75f95b5_p/all.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/logs/azureml/sidecar/tvmps_c6e6d2858679b9c78371ef9b5664c6a9aeefef5840ad0d3aee9b4180a75f95b5_p/all.log?sv=2019-02-02&sr=b&sig=%2FtU%2FQVLuECDThiBGb25Kt9b%2FFOjjyViTnDFzLoR7D%2Bk%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"logs/azureml/sidecar/tvmps_c6e6d2858679b9c78371ef9b5664c6a9aeefef5840ad0d3aee9b4180a75f95b5_p/task.enter_contexts.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/logs/azureml/sidecar/tvmps_c6e6d2858679b9c78371ef9b5664c6a9aeefef5840ad0d3aee9b4180a75f95b5_p/task.enter_contexts.log?sv=2019-02-02&sr=b&sig=o6KfdGrrJ9wpgfRgBuR6EBGe7PjMDSj7CX74ujlPkOE%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"logs/azureml/sidecar/tvmps_deca333c28ad0ee2af61010e7f808644c8cbc0760dc094cae395b58102473b22_p/all.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/logs/azureml/sidecar/tvmps_deca333c28ad0ee2af61010e7f808644c8cbc0760dc094cae395b58102473b22_p/all.log?sv=2019-02-02&sr=b&sig=ms7xVV9oUno00v4lcpYzb6oXcBghuvXyINN5MlTWFj4%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"logs/azureml/sidecar/tvmps_deca333c28ad0ee2af61010e7f808644c8cbc0760dc094cae395b58102473b22_p/task.enter_contexts.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/logs/azureml/sidecar/tvmps_deca333c28ad0ee2af61010e7f808644c8cbc0760dc094cae395b58102473b22_p/task.enter_contexts.log?sv=2019-02-02&sr=b&sig=fnDSwj%2FPpt%2Fk7QCH5hMQbf1N1cE%2BniB7n7C4Nh7inmM%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"logs/azureml/sidecar/tvmps_f0f5b814e7ee3abf3f56f97236c9892d5d99fbbfe5144209416f780fa80a7a1f_p/all.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/logs/azureml/sidecar/tvmps_f0f5b814e7ee3abf3f56f97236c9892d5d99fbbfe5144209416f780fa80a7a1f_p/all.log?sv=2019-02-02&sr=b&sig=yDOXaMS4ztm6uBt9%2Bvzui%2FeyQzy8tWgSTUhXNaBX1lk%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\", \"logs/azureml/sidecar/tvmps_f0f5b814e7ee3abf3f56f97236c9892d5d99fbbfe5144209416f780fa80a7a1f_p/task.enter_contexts.log\": \"https://mlworkbencheu2397988138.blob.core.windows.net/azureml/ExperimentRun/dcid.tf-distributed-effusion-detector_1621081615_407d41d4/logs/azureml/sidecar/tvmps_f0f5b814e7ee3abf3f56f97236c9892d5d99fbbfe5144209416f780fa80a7a1f_p/task.enter_contexts.log?sv=2019-02-02&sr=b&sig=GdsFwHenbI4XUI51VVNTdQm3p%2BM2kOgb7%2Bn2rjoZPVU%3D&st=2021-05-15T12%3A21%3A14Z&se=2021-05-15T20%3A31%3A14Z&sp=r\"}, \"log_groups\": [[\"logs/azureml/dataprep/backgroundProcess.log\", \"logs/azureml/dataprep/backgroundProcess_Telemetry.log\", \"logs/azureml/job_prep_azureml.log\"], [\"logs/azureml/sidecar/tvmps_f0f5b814e7ee3abf3f56f97236c9892d5d99fbbfe5144209416f780fa80a7a1f_p/all.log\", \"logs/azureml/sidecar/tvmps_f0f5b814e7ee3abf3f56f97236c9892d5d99fbbfe5144209416f780fa80a7a1f_p/task.enter_contexts.log\"], [\"logs/azureml/sidecar/tvmps_c6e6d2858679b9c78371ef9b5664c6a9aeefef5840ad0d3aee9b4180a75f95b5_p/all.log\", \"logs/azureml/sidecar/tvmps_c6e6d2858679b9c78371ef9b5664c6a9aeefef5840ad0d3aee9b4180a75f95b5_p/task.enter_contexts.log\"], [\"azureml-logs/55_azureml-execution-tvmps_f0f5b814e7ee3abf3f56f97236c9892d5d99fbbfe5144209416f780fa80a7a1f_p.txt\", \"azureml-logs/55_azureml-execution-tvmps_c6e6d2858679b9c78371ef9b5664c6a9aeefef5840ad0d3aee9b4180a75f95b5_p.txt\", \"azureml-logs/55_azureml-execution-tvmps_deca333c28ad0ee2af61010e7f808644c8cbc0760dc094cae395b58102473b22_p.txt\", \"azureml-logs/55_azureml-execution-tvmps_522962d35ec9e649ca9dfd0c74a0bc6fe2aac5a9197cf41cf503c5c81f7134d8_p.txt\"], [\"azureml-logs/65_job_prep-tvmps_f0f5b814e7ee3abf3f56f97236c9892d5d99fbbfe5144209416f780fa80a7a1f_p.txt\", \"azureml-logs/65_job_prep-tvmps_c6e6d2858679b9c78371ef9b5664c6a9aeefef5840ad0d3aee9b4180a75f95b5_p.txt\", \"azureml-logs/65_job_prep-tvmps_deca333c28ad0ee2af61010e7f808644c8cbc0760dc094cae395b58102473b22_p.txt\", \"azureml-logs/65_job_prep-tvmps_522962d35ec9e649ca9dfd0c74a0bc6fe2aac5a9197cf41cf503c5c81f7134d8_p.txt\"], [\"logs/azureml/sidecar/tvmps_deca333c28ad0ee2af61010e7f808644c8cbc0760dc094cae395b58102473b22_p/all.log\", \"logs/azureml/sidecar/tvmps_deca333c28ad0ee2af61010e7f808644c8cbc0760dc094cae395b58102473b22_p/task.enter_contexts.log\"], [\"logs/azureml/sidecar/tvmps_522962d35ec9e649ca9dfd0c74a0bc6fe2aac5a9197cf41cf503c5c81f7134d8_p/all.log\", \"logs/azureml/sidecar/tvmps_522962d35ec9e649ca9dfd0c74a0bc6fe2aac5a9197cf41cf503c5c81f7134d8_p/task.enter_contexts.log\"]], \"run_duration\": \"0:06:07\", \"run_number\": \"1621081617\", \"run_queued_details\": {\"status\": \"Running\", \"details\": null}}, \"child_runs\": [], \"children_metrics\": {}, \"run_metrics\": [], \"run_logs\": \"[2021-05-15T12:31:02.313722] INFO azureml.sidecar.sidecar: Received task: enter_contexts. Running on Linux at /mnt/hostfs/mnt/batch/tasks/shared/LS_root/jobs/ml-workbench-eu/azureml/tf-distributed-effusion-detector_1621081615_407d41d4/mounts/workspaceblobstore/azureml/tf-distributed-effusion-detector_1621081615_407d41d4\\n[2021-05-15T12:31:02.314467] INFO azureml.sidecar.sidecar: Invoking \\\"enter_contexts\\\" task with Context Managers: {\\\"context_managers\\\": [\\\"Dataset:context_managers.Datasets\\\"]}\\n\", \"graph\": {}, \"widget_settings\": {\"childWidgetDisplay\": \"popup\", \"send_telemetry\": false, \"log_level\": \"NOTSET\", \"sdk_version\": \"1.27.0\"}, \"loading\": false}"
          },
          "metadata": {}
        }
      ],
      "execution_count": 36,
      "metadata": {
        "collapsed": true,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "gather": {
          "logged": 1621081582032
        }
      }
    }
  ],
  "metadata": {
    "anaconda-cloud": {},
    "kernelspec": {
      "name": "python38-azureml",
      "language": "python",
      "display_name": "Python 3.8 - AzureML"
    },
    "language_info": {
      "name": "python",
      "version": "3.8.1",
      "mimetype": "text/x-python",
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "pygments_lexer": "ipython3",
      "nbconvert_exporter": "python",
      "file_extension": ".py"
    },
    "metadata": {
      "interpreter": {
        "hash": "63fd5069d213b44bf678585dea6b12cceca9941eaf7f819626cde1f2670de90d"
      }
    },
    "kernel_info": {
      "name": "python38-azureml"
    },
    "nteract": {
      "version": "nteract-front-end@1.0.0"
    },
    "microsoft": {
      "host": {
        "AzureML": {
          "notebookHasBeenCompleted": true
        }
      }
    }
  },
  "nbformat": 4,
  "nbformat_minor": 1
}