Infections_Final.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "Infections Final.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4pmvxdX2O_pl"
      },
      "source": [
        "Mounting Google Drive For Data Access"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zOlTwazRgZmb",
        "outputId": "d7e319eb-3665-4cf5-d1ac-dfa1614654f5"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/gdrive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wzQLeyTVPGa5"
      },
      "source": [
        "Creating Data Generator for Data Augmentation"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YmbS2Wtjgae3"
      },
      "source": [
        "import tensorflow as tf \n",
        "from keras.applications.resnet50 import preprocess_input\n",
        "\n",
        "datagen = tf.keras.preprocessing.image.ImageDataGenerator( \n",
        "                                                          rescale = 1./255., \n",
        "                                                          rotation_range = 8, \n",
        "                                                          width_shift_range = 0.2, \n",
        "                                                          height_shift_range = 0.2,\n",
        "                                                          shear_range = 0.2, \n",
        "                                                          zoom_range = 0.2,\n",
        "                                                          brightness_range = [0.5,1.5],\n",
        "                                                          horizontal_flip = True,\n",
        "                                                          fill_mode = 'nearest',\n",
        "                                                          dtype = None,                                                      \n",
        "                                                          validation_split = 0.15)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-oohuVRWPJbl"
      },
      "source": [
        "Validation and Training Sets"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "J2uScVN-gc1O",
        "outputId": "d9e836a1-0c3d-4f19-85fd-390f585eb3c7"
      },
      "source": [
        "dire = '/content/gdrive/MyDrive/Infections'\n",
        "\n",
        "train_generator = datagen.flow_from_directory(\n",
        "    dire,\n",
        "    subset=\"training\",\n",
        "    batch_size=32,\n",
        "    class_mode='categorical',     \n",
        "    shuffle=True,                      \n",
        "    target_size=(150,150)\n",
        ")\n",
        "\n",
        "valid_generator = datagen.flow_from_directory(\n",
        "    dire,\n",
        "    subset=\"validation\",\n",
        "    shuffle=True,\n",
        "    batch_size=32,\n",
        "    class_mode='categorical',                           \n",
        "    target_size=(150,150)\n",
        ")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 673 images belonging to 5 classes.\n",
            "Found 116 images belonging to 5 classes.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "H8Opz7nagigq",
        "outputId": "dfc29d6b-5239-425d-d364-6bc5425a0aad"
      },
      "source": [
        "for image_batch, label_batch in train_generator:\n",
        "  break\n",
        "image_batch.shape, label_batch.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((32, 150, 150, 3), (32, 5))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 40
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NwJwUUdzPOML"
      },
      "source": [
        "Extracting and Saving Labels"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_7FBIYkQgku4",
        "outputId": "94162992-e3c0-4ed0-9110-97ae083ab561"
      },
      "source": [
        "label_batch[0:5]\n",
        "print(train_generator.class_indices)\n",
        "labels = '\\n'.join(sorted(train_generator.class_indices.keys()))\n",
        "with open('labels.txt', 'w') as f:\n",
        " f.write(labels)\n",
        "!cat labels.txt"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "{'Blepharitis': 0, 'Cellulitis eye': 1, 'Conjunctivitis, Pink Eye': 2, 'Eye': 3, 'Hordeolum, stye': 4}\n",
            "Blepharitis\n",
            "Cellulitis eye\n",
            "Conjunctivitis, Pink Eye\n",
            "Eye\n",
            "Hordeolum, stye"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "994hllBwgmZF",
        "outputId": "09f03101-ac93-4e46-8709-8293fa4846d1"
      },
      "source": [
        "import os\n",
        "import urllib.request\n",
        "\n",
        "#urllib.request.urlretrieve('https://github.com/fchollet/deep-learning-models/releases/download/v0.1/vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5',\n",
        "                           #'vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5')\n",
        "\n",
        "os.listdir()\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['.config', 'gdrive', 'labels.txt', 'sample_data']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 42
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "70dmmom9PRku"
      },
      "source": [
        "Initializing Mobile Net V2 Pre-trained Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "i1cKh7cjgn_K",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "cf708cc2-9328-4943-acf4-6312a29c7ed8"
      },
      "source": [
        "from keras.applications.resnet50 import ResNet50\n",
        "from keras import layers\n",
        "from tensorflow.keras import Sequential\n",
        "from tensorflow.keras.layers import Flatten, Dense, Dropout\n",
        "from keras.applications.vgg16 import VGG16 \n",
        "from tensorflow.keras.applications.inception_v3 import InceptionV3\n",
        "\n",
        "from tensorflow.keras import layers \n",
        "from tensorflow.keras import Model \n",
        "\n",
        "from tensorflow.keras.applications import ResNet50\n",
        "from tensorflow.python.keras.models import Sequential\n",
        "from tensorflow.python.keras.layers import Dense, Flatten, GlobalAveragePooling2D\n",
        "\n",
        "from keras.applications.vgg16 import VGG16, preprocess_input\n",
        "from tensorflow.keras.applications.inception_v3 import InceptionV3\n",
        "from tensorflow.keras.applications import MobileNetV2\n",
        "\n",
        "mobileNet = MobileNetV2(\n",
        "    input_shape = (150,150,3), \n",
        "    weights = 'imagenet', \n",
        "    include_top = False,\n",
        ")\n",
        "\n",
        "\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:`input_shape` is undefined or non-square, or `rows` is not in [96, 128, 160, 192, 224]. Weights for input shape (224, 224) will be loaded as the default.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WwSsotLlPWh1"
      },
      "source": [
        "Freezing Layers in Mobile Net V2 Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "J9Ep8MZIgpjZ"
      },
      "source": [
        "for layer in mobileNet.layers:\n",
        "    layer.trainable = False"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "et3QYJwNPhHp"
      },
      "source": [
        "Creating Custom Architecture on Top of MobileNetV2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Lboouu5LgrCg"
      },
      "source": [
        "from keras import layers\n",
        "from tensorflow.keras import Sequential\n",
        "from tensorflow.keras.layers import Flatten, Dense\n",
        "from tensorflow.keras import regularizers\n",
        "\n",
        "\n",
        "model = Sequential()\n",
        "\n",
        "\n",
        "## Inception V3\n",
        "model.add(mobileNet)\n",
        "model.add(Flatten())\n",
        "#model.add(GlobalAveragePooling2D())\n",
        "model.add(Dense(32, activation='relu'))\n",
        "model.add(Dropout(rate=0.4))\n",
        "model.add(Dense(5, activation = 'softmax'))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KHlOuj_fPlW1"
      },
      "source": [
        "Adjusting Adam Optimizer and Compiling Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DqH3iveMgsmA",
        "outputId": "12958c88-5248-404d-827a-5adcb8da5556"
      },
      "source": [
        "opt = tf.keras.optimizers.Adam(learning_rate=0.00001)\n",
        "\n",
        "model.compile(optimizer=opt, loss=\"categorical_crossentropy\", metrics=[\"accuracy\"])\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "mobilenetv2_1.00_224 (Functi (None, 5, 5, 1280)        2257984   \n",
            "_________________________________________________________________\n",
            "flatten_2 (Flatten)          (None, 32000)             0         \n",
            "_________________________________________________________________\n",
            "dense_5 (Dense)              (None, 32)                1024032   \n",
            "_________________________________________________________________\n",
            "dropout_3 (Dropout)          (None, 32)                0         \n",
            "_________________________________________________________________\n",
            "dense_6 (Dense)              (None, 5)                 165       \n",
            "=================================================================\n",
            "Total params: 3,282,181\n",
            "Trainable params: 1,024,197\n",
            "Non-trainable params: 2,257,984\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KkhGCLaEPowV"
      },
      "source": [
        "Training for 10 Epochs"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6Pm5d426gwiB",
        "outputId": "a3477ce2-9e2b-408a-ead6-ab3654bdbcd3"
      },
      "source": [
        "hist= model.fit(train_generator, validation_data = valid_generator, epochs = 10)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "22/22 [==============================] - 11s 380ms/step - loss: 2.7741 - accuracy: 0.2030 - val_loss: 1.9100 - val_accuracy: 0.1897\n",
            "Epoch 2/10\n",
            "22/22 [==============================] - 7s 328ms/step - loss: 1.9401 - accuracy: 0.2118 - val_loss: 1.6664 - val_accuracy: 0.2414\n",
            "Epoch 3/10\n",
            "22/22 [==============================] - 7s 327ms/step - loss: 1.6872 - accuracy: 0.2504 - val_loss: 1.5500 - val_accuracy: 0.3276\n",
            "Epoch 4/10\n",
            "22/22 [==============================] - 7s 327ms/step - loss: 1.6407 - accuracy: 0.2587 - val_loss: 1.5721 - val_accuracy: 0.3190\n",
            "Epoch 5/10\n",
            "22/22 [==============================] - 7s 329ms/step - loss: 1.6067 - accuracy: 0.2761 - val_loss: 1.5810 - val_accuracy: 0.3534\n",
            "Epoch 6/10\n",
            "22/22 [==============================] - 7s 330ms/step - loss: 1.5613 - accuracy: 0.3225 - val_loss: 1.5046 - val_accuracy: 0.3966\n",
            "Epoch 7/10\n",
            "22/22 [==============================] - 7s 327ms/step - loss: 1.5140 - accuracy: 0.3550 - val_loss: 1.4703 - val_accuracy: 0.4310\n",
            "Epoch 8/10\n",
            "22/22 [==============================] - 7s 328ms/step - loss: 1.4392 - accuracy: 0.4348 - val_loss: 1.4250 - val_accuracy: 0.4310\n",
            "Epoch 9/10\n",
            "22/22 [==============================] - 7s 334ms/step - loss: 1.4621 - accuracy: 0.3287 - val_loss: 1.4072 - val_accuracy: 0.4914\n",
            "Epoch 10/10\n",
            "22/22 [==============================] - 7s 328ms/step - loss: 1.4062 - accuracy: 0.4381 - val_loss: 1.4096 - val_accuracy: 0.4310\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CAxCoPoFPqp7"
      },
      "source": [
        "Training for Additional 50 Epochs"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9p5Aif5nmanU",
        "outputId": "ea753e26-0de2-42af-bc04-79c87881863b"
      },
      "source": [
        "hist= model.fit(train_generator, validation_data = valid_generator, epochs = 50)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "22/22 [==============================] - 8s 350ms/step - loss: 1.4208 - accuracy: 0.4131 - val_loss: 1.4122 - val_accuracy: 0.4310\n",
            "Epoch 2/50\n",
            "22/22 [==============================] - 7s 334ms/step - loss: 1.4075 - accuracy: 0.4101 - val_loss: 1.3118 - val_accuracy: 0.4828\n",
            "Epoch 3/50\n",
            "22/22 [==============================] - 7s 332ms/step - loss: 1.3910 - accuracy: 0.4042 - val_loss: 1.3510 - val_accuracy: 0.4741\n",
            "Epoch 4/50\n",
            "22/22 [==============================] - 7s 329ms/step - loss: 1.4088 - accuracy: 0.4354 - val_loss: 1.3628 - val_accuracy: 0.4310\n",
            "Epoch 5/50\n",
            "22/22 [==============================] - 7s 326ms/step - loss: 1.3299 - accuracy: 0.4146 - val_loss: 1.2796 - val_accuracy: 0.5000\n",
            "Epoch 6/50\n",
            "22/22 [==============================] - 7s 328ms/step - loss: 1.3717 - accuracy: 0.4398 - val_loss: 1.2759 - val_accuracy: 0.5086\n",
            "Epoch 7/50\n",
            "22/22 [==============================] - 7s 324ms/step - loss: 1.3270 - accuracy: 0.4398 - val_loss: 1.2277 - val_accuracy: 0.5259\n",
            "Epoch 8/50\n",
            "22/22 [==============================] - 7s 328ms/step - loss: 1.2796 - accuracy: 0.4651 - val_loss: 1.2650 - val_accuracy: 0.4828\n",
            "Epoch 9/50\n",
            "22/22 [==============================] - 7s 327ms/step - loss: 1.3328 - accuracy: 0.4487 - val_loss: 1.2320 - val_accuracy: 0.5086\n",
            "Epoch 10/50\n",
            "22/22 [==============================] - 7s 329ms/step - loss: 1.3205 - accuracy: 0.4279 - val_loss: 1.2142 - val_accuracy: 0.5603\n",
            "Epoch 11/50\n",
            "22/22 [==============================] - 7s 337ms/step - loss: 1.2939 - accuracy: 0.4547 - val_loss: 1.2394 - val_accuracy: 0.5259\n",
            "Epoch 12/50\n",
            "22/22 [==============================] - 7s 334ms/step - loss: 1.2765 - accuracy: 0.4695 - val_loss: 1.1412 - val_accuracy: 0.5259\n",
            "Epoch 13/50\n",
            "22/22 [==============================] - 7s 310ms/step - loss: 1.2651 - accuracy: 0.4933 - val_loss: 1.1526 - val_accuracy: 0.5603\n",
            "Epoch 14/50\n",
            "22/22 [==============================] - 7s 328ms/step - loss: 1.2319 - accuracy: 0.5082 - val_loss: 1.1201 - val_accuracy: 0.5776\n",
            "Epoch 15/50\n",
            "22/22 [==============================] - 7s 309ms/step - loss: 1.2119 - accuracy: 0.5007 - val_loss: 1.1522 - val_accuracy: 0.5690\n",
            "Epoch 16/50\n",
            "22/22 [==============================] - 7s 312ms/step - loss: 1.2194 - accuracy: 0.5171 - val_loss: 1.0812 - val_accuracy: 0.5948\n",
            "Epoch 17/50\n",
            "22/22 [==============================] - 7s 329ms/step - loss: 1.2640 - accuracy: 0.4785 - val_loss: 1.0897 - val_accuracy: 0.5776\n",
            "Epoch 18/50\n",
            "22/22 [==============================] - 7s 328ms/step - loss: 1.2137 - accuracy: 0.4933 - val_loss: 1.1366 - val_accuracy: 0.5431\n",
            "Epoch 19/50\n",
            "22/22 [==============================] - 7s 328ms/step - loss: 1.1936 - accuracy: 0.5111 - val_loss: 1.0643 - val_accuracy: 0.6034\n",
            "Epoch 20/50\n",
            "22/22 [==============================] - 7s 341ms/step - loss: 1.1628 - accuracy: 0.5290 - val_loss: 1.1820 - val_accuracy: 0.5431\n",
            "Epoch 21/50\n",
            "22/22 [==============================] - 7s 331ms/step - loss: 1.1687 - accuracy: 0.5156 - val_loss: 1.1738 - val_accuracy: 0.5172\n",
            "Epoch 22/50\n",
            "22/22 [==============================] - 7s 330ms/step - loss: 1.1722 - accuracy: 0.5215 - val_loss: 1.1705 - val_accuracy: 0.6379\n",
            "Epoch 23/50\n",
            "22/22 [==============================] - 7s 329ms/step - loss: 1.1642 - accuracy: 0.5379 - val_loss: 1.1306 - val_accuracy: 0.5431\n",
            "Epoch 24/50\n",
            "22/22 [==============================] - 7s 322ms/step - loss: 1.1319 - accuracy: 0.5379 - val_loss: 1.0257 - val_accuracy: 0.6121\n",
            "Epoch 25/50\n",
            "22/22 [==============================] - 7s 329ms/step - loss: 1.1187 - accuracy: 0.5453 - val_loss: 1.0325 - val_accuracy: 0.5603\n",
            "Epoch 26/50\n",
            "22/22 [==============================] - 7s 328ms/step - loss: 1.1397 - accuracy: 0.5260 - val_loss: 1.0204 - val_accuracy: 0.6121\n",
            "Epoch 27/50\n",
            "22/22 [==============================] - 7s 325ms/step - loss: 1.1016 - accuracy: 0.5423 - val_loss: 1.0774 - val_accuracy: 0.6466\n",
            "Epoch 28/50\n",
            "22/22 [==============================] - 7s 330ms/step - loss: 1.0923 - accuracy: 0.5706 - val_loss: 1.1048 - val_accuracy: 0.6121\n",
            "Epoch 29/50\n",
            "22/22 [==============================] - 7s 329ms/step - loss: 1.1121 - accuracy: 0.5379 - val_loss: 1.0665 - val_accuracy: 0.5690\n",
            "Epoch 30/50\n",
            "22/22 [==============================] - 7s 329ms/step - loss: 1.1200 - accuracy: 0.5468 - val_loss: 1.0213 - val_accuracy: 0.5948\n",
            "Epoch 31/50\n",
            "22/22 [==============================] - 7s 321ms/step - loss: 1.0868 - accuracy: 0.5706 - val_loss: 1.0321 - val_accuracy: 0.6034\n",
            "Epoch 32/50\n",
            "22/22 [==============================] - 7s 329ms/step - loss: 1.0930 - accuracy: 0.5646 - val_loss: 1.0298 - val_accuracy: 0.6034\n",
            "Epoch 33/50\n",
            "22/22 [==============================] - 7s 330ms/step - loss: 1.0639 - accuracy: 0.5884 - val_loss: 0.9795 - val_accuracy: 0.6379\n",
            "Epoch 34/50\n",
            "22/22 [==============================] - 7s 327ms/step - loss: 1.0554 - accuracy: 0.5780 - val_loss: 0.9486 - val_accuracy: 0.6552\n",
            "Epoch 35/50\n",
            "22/22 [==============================] - 7s 331ms/step - loss: 1.0812 - accuracy: 0.5349 - val_loss: 1.0152 - val_accuracy: 0.6121\n",
            "Epoch 36/50\n",
            "22/22 [==============================] - 7s 338ms/step - loss: 1.0926 - accuracy: 0.5646 - val_loss: 0.9678 - val_accuracy: 0.5948\n",
            "Epoch 37/50\n",
            "22/22 [==============================] - 7s 327ms/step - loss: 1.0534 - accuracy: 0.5736 - val_loss: 0.9802 - val_accuracy: 0.6379\n",
            "Epoch 38/50\n",
            "22/22 [==============================] - 7s 326ms/step - loss: 1.0625 - accuracy: 0.5765 - val_loss: 0.9352 - val_accuracy: 0.6897\n",
            "Epoch 39/50\n",
            "22/22 [==============================] - 7s 330ms/step - loss: 1.0650 - accuracy: 0.5423 - val_loss: 1.0062 - val_accuracy: 0.5862\n",
            "Epoch 40/50\n",
            "22/22 [==============================] - 7s 331ms/step - loss: 1.0481 - accuracy: 0.5750 - val_loss: 1.0073 - val_accuracy: 0.6034\n",
            "Epoch 41/50\n",
            "22/22 [==============================] - 7s 329ms/step - loss: 1.0401 - accuracy: 0.5914 - val_loss: 1.0243 - val_accuracy: 0.6121\n",
            "Epoch 42/50\n",
            "22/22 [==============================] - 7s 310ms/step - loss: 1.0902 - accuracy: 0.5676 - val_loss: 0.9753 - val_accuracy: 0.5948\n",
            "Epoch 43/50\n",
            "22/22 [==============================] - 7s 332ms/step - loss: 1.0443 - accuracy: 0.5884 - val_loss: 0.9577 - val_accuracy: 0.6552\n",
            "Epoch 44/50\n",
            "22/22 [==============================] - 7s 331ms/step - loss: 0.9950 - accuracy: 0.6033 - val_loss: 1.0085 - val_accuracy: 0.6207\n",
            "Epoch 45/50\n",
            "22/22 [==============================] - 7s 326ms/step - loss: 1.0114 - accuracy: 0.6033 - val_loss: 0.9931 - val_accuracy: 0.6466\n",
            "Epoch 46/50\n",
            "22/22 [==============================] - 7s 321ms/step - loss: 1.0433 - accuracy: 0.5958 - val_loss: 0.9807 - val_accuracy: 0.5776\n",
            "Epoch 47/50\n",
            "22/22 [==============================] - 7s 330ms/step - loss: 0.9886 - accuracy: 0.6062 - val_loss: 0.9430 - val_accuracy: 0.6293\n",
            "Epoch 48/50\n",
            "22/22 [==============================] - 7s 335ms/step - loss: 1.0218 - accuracy: 0.5944 - val_loss: 0.9972 - val_accuracy: 0.6293\n",
            "Epoch 49/50\n",
            "22/22 [==============================] - 7s 337ms/step - loss: 1.0426 - accuracy: 0.5750 - val_loss: 0.9475 - val_accuracy: 0.6379\n",
            "Epoch 50/50\n",
            "22/22 [==============================] - 7s 329ms/step - loss: 1.0335 - accuracy: 0.5869 - val_loss: 0.9995 - val_accuracy: 0.6121\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "e6w1kfkoPtsI"
      },
      "source": [
        "Unfreezing Layers in MobileNetV2, adjusting Optimizer and Compiling Model again to ensure that the model can be trained on custom dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JBC9qdGL3eV7",
        "outputId": "563ffe44-4265-4fad-9ac3-9bdffdc9f04f"
      },
      "source": [
        "for layer in mobileNet.layers:\n",
        "    layer.trainable = True\n",
        "\n",
        "opt = tf.keras.optimizers.Adam(learning_rate=0.00001)\n",
        "\n",
        "model.compile(optimizer=opt, loss=\"categorical_crossentropy\", metrics=[\"accuracy\"])\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "mobilenetv2_1.00_224 (Functi (None, 5, 5, 1280)        2257984   \n",
            "_________________________________________________________________\n",
            "flatten_2 (Flatten)          (None, 32000)             0         \n",
            "_________________________________________________________________\n",
            "dense_5 (Dense)              (None, 32)                1024032   \n",
            "_________________________________________________________________\n",
            "dropout_3 (Dropout)          (None, 32)                0         \n",
            "_________________________________________________________________\n",
            "dense_6 (Dense)              (None, 5)                 165       \n",
            "=================================================================\n",
            "Total params: 3,282,181\n",
            "Trainable params: 3,248,069\n",
            "Non-trainable params: 34,112\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dBtrQxwuP3ky"
      },
      "source": [
        "Training unfreezed model for 50 Epochs"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xr7XJ1lJj67-",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c5488faa-4fd3-4c6f-8ad5-451867f9007d"
      },
      "source": [
        "hist= model.fit(train_generator, validation_data = valid_generator, epochs = 50)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "22/22 [==============================] - 13s 394ms/step - loss: 1.6983 - accuracy: 0.3298 - val_loss: 1.0477 - val_accuracy: 0.5776\n",
            "Epoch 2/50\n",
            "22/22 [==============================] - 8s 352ms/step - loss: 1.5530 - accuracy: 0.3757 - val_loss: 0.9314 - val_accuracy: 0.5862\n",
            "Epoch 3/50\n",
            "22/22 [==============================] - 8s 357ms/step - loss: 1.4065 - accuracy: 0.4089 - val_loss: 0.9928 - val_accuracy: 0.5690\n",
            "Epoch 4/50\n",
            "22/22 [==============================] - 8s 352ms/step - loss: 1.3786 - accuracy: 0.4607 - val_loss: 1.0624 - val_accuracy: 0.5690\n",
            "Epoch 5/50\n",
            "22/22 [==============================] - 8s 356ms/step - loss: 1.3420 - accuracy: 0.4203 - val_loss: 0.9820 - val_accuracy: 0.6034\n",
            "Epoch 6/50\n",
            "22/22 [==============================] - 8s 360ms/step - loss: 1.3281 - accuracy: 0.4628 - val_loss: 1.0229 - val_accuracy: 0.6207\n",
            "Epoch 7/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 1.3957 - accuracy: 0.4070 - val_loss: 1.0468 - val_accuracy: 0.5948\n",
            "Epoch 8/50\n",
            "22/22 [==============================] - 8s 349ms/step - loss: 1.3197 - accuracy: 0.4556 - val_loss: 1.0645 - val_accuracy: 0.5690\n",
            "Epoch 9/50\n",
            "22/22 [==============================] - 8s 356ms/step - loss: 1.2673 - accuracy: 0.4742 - val_loss: 1.0682 - val_accuracy: 0.5862\n",
            "Epoch 10/50\n",
            "22/22 [==============================] - 8s 355ms/step - loss: 1.2452 - accuracy: 0.4838 - val_loss: 1.0547 - val_accuracy: 0.5603\n",
            "Epoch 11/50\n",
            "22/22 [==============================] - 8s 359ms/step - loss: 1.2345 - accuracy: 0.4948 - val_loss: 0.9995 - val_accuracy: 0.6121\n",
            "Epoch 12/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 1.1906 - accuracy: 0.5196 - val_loss: 1.0122 - val_accuracy: 0.5517\n",
            "Epoch 13/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 1.2688 - accuracy: 0.4813 - val_loss: 1.0448 - val_accuracy: 0.6034\n",
            "Epoch 14/50\n",
            "22/22 [==============================] - 8s 355ms/step - loss: 1.2340 - accuracy: 0.4865 - val_loss: 0.9520 - val_accuracy: 0.6034\n",
            "Epoch 15/50\n",
            "22/22 [==============================] - 8s 365ms/step - loss: 1.2067 - accuracy: 0.4668 - val_loss: 1.0563 - val_accuracy: 0.5862\n",
            "Epoch 16/50\n",
            "22/22 [==============================] - 8s 351ms/step - loss: 1.2007 - accuracy: 0.5183 - val_loss: 1.0003 - val_accuracy: 0.6379\n",
            "Epoch 17/50\n",
            "22/22 [==============================] - 8s 363ms/step - loss: 1.1450 - accuracy: 0.5322 - val_loss: 0.9410 - val_accuracy: 0.6552\n",
            "Epoch 18/50\n",
            "22/22 [==============================] - 8s 357ms/step - loss: 1.2482 - accuracy: 0.4703 - val_loss: 1.0293 - val_accuracy: 0.6034\n",
            "Epoch 19/50\n",
            "22/22 [==============================] - 8s 346ms/step - loss: 1.1464 - accuracy: 0.5245 - val_loss: 0.9981 - val_accuracy: 0.5948\n",
            "Epoch 20/50\n",
            "22/22 [==============================] - 8s 361ms/step - loss: 1.1335 - accuracy: 0.5476 - val_loss: 0.9885 - val_accuracy: 0.5948\n",
            "Epoch 21/50\n",
            "22/22 [==============================] - 8s 357ms/step - loss: 1.1457 - accuracy: 0.5363 - val_loss: 0.9586 - val_accuracy: 0.6034\n",
            "Epoch 22/50\n",
            "22/22 [==============================] - 8s 366ms/step - loss: 1.1863 - accuracy: 0.5126 - val_loss: 1.0018 - val_accuracy: 0.6207\n",
            "Epoch 23/50\n",
            "22/22 [==============================] - 8s 360ms/step - loss: 1.1402 - accuracy: 0.5456 - val_loss: 0.9709 - val_accuracy: 0.6379\n",
            "Epoch 24/50\n",
            "22/22 [==============================] - 8s 353ms/step - loss: 1.1558 - accuracy: 0.5303 - val_loss: 0.8859 - val_accuracy: 0.6466\n",
            "Epoch 25/50\n",
            "22/22 [==============================] - 8s 351ms/step - loss: 1.0882 - accuracy: 0.5454 - val_loss: 0.9271 - val_accuracy: 0.6379\n",
            "Epoch 26/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 1.0740 - accuracy: 0.5792 - val_loss: 0.8543 - val_accuracy: 0.6810\n",
            "Epoch 27/50\n",
            "22/22 [==============================] - 8s 361ms/step - loss: 1.0910 - accuracy: 0.5646 - val_loss: 0.9315 - val_accuracy: 0.5948\n",
            "Epoch 28/50\n",
            "22/22 [==============================] - 8s 357ms/step - loss: 0.9797 - accuracy: 0.6293 - val_loss: 0.9064 - val_accuracy: 0.6121\n",
            "Epoch 29/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 1.0538 - accuracy: 0.5598 - val_loss: 0.8895 - val_accuracy: 0.6466\n",
            "Epoch 30/50\n",
            "22/22 [==============================] - 8s 351ms/step - loss: 1.1106 - accuracy: 0.5617 - val_loss: 0.9755 - val_accuracy: 0.6293\n",
            "Epoch 31/50\n",
            "22/22 [==============================] - 8s 370ms/step - loss: 1.0557 - accuracy: 0.5391 - val_loss: 0.9269 - val_accuracy: 0.6207\n",
            "Epoch 32/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 1.0794 - accuracy: 0.5725 - val_loss: 0.9718 - val_accuracy: 0.5948\n",
            "Epoch 33/50\n",
            "22/22 [==============================] - 8s 353ms/step - loss: 1.0017 - accuracy: 0.5921 - val_loss: 0.9706 - val_accuracy: 0.6121\n",
            "Epoch 34/50\n",
            "22/22 [==============================] - 8s 355ms/step - loss: 1.0154 - accuracy: 0.6022 - val_loss: 0.9820 - val_accuracy: 0.5862\n",
            "Epoch 35/50\n",
            "22/22 [==============================] - 8s 357ms/step - loss: 1.0512 - accuracy: 0.6279 - val_loss: 0.9309 - val_accuracy: 0.6638\n",
            "Epoch 36/50\n",
            "22/22 [==============================] - 8s 353ms/step - loss: 1.0171 - accuracy: 0.6263 - val_loss: 0.9101 - val_accuracy: 0.6207\n",
            "Epoch 37/50\n",
            "22/22 [==============================] - 8s 357ms/step - loss: 1.0425 - accuracy: 0.5881 - val_loss: 0.9053 - val_accuracy: 0.6121\n",
            "Epoch 38/50\n",
            "22/22 [==============================] - 8s 360ms/step - loss: 0.9856 - accuracy: 0.6150 - val_loss: 0.9342 - val_accuracy: 0.6379\n",
            "Epoch 39/50\n",
            "22/22 [==============================] - 8s 352ms/step - loss: 0.9021 - accuracy: 0.6656 - val_loss: 0.9174 - val_accuracy: 0.6810\n",
            "Epoch 40/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 0.9138 - accuracy: 0.6288 - val_loss: 1.0130 - val_accuracy: 0.6207\n",
            "Epoch 41/50\n",
            "22/22 [==============================] - 8s 352ms/step - loss: 0.9924 - accuracy: 0.6333 - val_loss: 0.9235 - val_accuracy: 0.6207\n",
            "Epoch 42/50\n",
            "22/22 [==============================] - 8s 353ms/step - loss: 0.9979 - accuracy: 0.6027 - val_loss: 0.8329 - val_accuracy: 0.6379\n",
            "Epoch 43/50\n",
            "22/22 [==============================] - 8s 354ms/step - loss: 0.9910 - accuracy: 0.5911 - val_loss: 0.8695 - val_accuracy: 0.6552\n",
            "Epoch 44/50\n",
            "22/22 [==============================] - 8s 352ms/step - loss: 0.9400 - accuracy: 0.6156 - val_loss: 0.8637 - val_accuracy: 0.6638\n",
            "Epoch 45/50\n",
            "22/22 [==============================] - 8s 356ms/step - loss: 0.9345 - accuracy: 0.6265 - val_loss: 0.9218 - val_accuracy: 0.6121\n",
            "Epoch 46/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 0.9781 - accuracy: 0.6136 - val_loss: 0.9231 - val_accuracy: 0.6810\n",
            "Epoch 47/50\n",
            "22/22 [==============================] - 8s 355ms/step - loss: 0.9794 - accuracy: 0.6094 - val_loss: 0.8838 - val_accuracy: 0.6724\n",
            "Epoch 48/50\n",
            "22/22 [==============================] - 8s 355ms/step - loss: 0.9611 - accuracy: 0.6056 - val_loss: 0.8481 - val_accuracy: 0.6552\n",
            "Epoch 49/50\n",
            "22/22 [==============================] - 8s 359ms/step - loss: 0.9402 - accuracy: 0.6436 - val_loss: 0.8695 - val_accuracy: 0.6724\n",
            "Epoch 50/50\n",
            "22/22 [==============================] - 8s 355ms/step - loss: 0.9954 - accuracy: 0.6063 - val_loss: 0.8738 - val_accuracy: 0.6724\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jqjpHJB2P8ft"
      },
      "source": [
        "Adjusting Optimizer again for slower learning rate"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lEx1q2MR6VAa",
        "outputId": "9c864284-2662-4aaa-8002-2f4a4bed442e"
      },
      "source": [
        "for layer in mobileNet.layers:\n",
        "    layer.trainable = True\n",
        "\n",
        "opt = tf.keras.optimizers.Adam(learning_rate=0.000001)\n",
        "\n",
        "model.compile(optimizer=opt, loss=\"categorical_crossentropy\", metrics=[\"accuracy\"])\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "mobilenetv2_1.00_224 (Functi (None, 5, 5, 1280)        2257984   \n",
            "_________________________________________________________________\n",
            "flatten_2 (Flatten)          (None, 32000)             0         \n",
            "_________________________________________________________________\n",
            "dense_5 (Dense)              (None, 32)                1024032   \n",
            "_________________________________________________________________\n",
            "dropout_3 (Dropout)          (None, 32)                0         \n",
            "_________________________________________________________________\n",
            "dense_6 (Dense)              (None, 5)                 165       \n",
            "=================================================================\n",
            "Total params: 3,282,181\n",
            "Trainable params: 3,248,069\n",
            "Non-trainable params: 34,112\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qIyRausFQBdH"
      },
      "source": [
        "Training on slower learning rate for fine tuning"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "08rhHZRr1aX-",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ad68ff10-3c88-4303-c15f-68f8905fc2e0"
      },
      "source": [
        "hist= model.fit(train_generator, validation_data = valid_generator, epochs = 50)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "22/22 [==============================] - 13s 411ms/step - loss: 0.9085 - accuracy: 0.6271 - val_loss: 0.8815 - val_accuracy: 0.6897\n",
            "Epoch 2/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 0.9207 - accuracy: 0.6245 - val_loss: 0.8310 - val_accuracy: 0.6897\n",
            "Epoch 3/50\n",
            "22/22 [==============================] - 8s 359ms/step - loss: 0.9148 - accuracy: 0.6201 - val_loss: 0.9349 - val_accuracy: 0.6897\n",
            "Epoch 4/50\n",
            "22/22 [==============================] - 8s 355ms/step - loss: 0.9027 - accuracy: 0.6555 - val_loss: 0.8648 - val_accuracy: 0.6379\n",
            "Epoch 5/50\n",
            "22/22 [==============================] - 8s 352ms/step - loss: 0.9725 - accuracy: 0.5975 - val_loss: 0.8528 - val_accuracy: 0.6638\n",
            "Epoch 6/50\n",
            "22/22 [==============================] - 8s 369ms/step - loss: 0.9542 - accuracy: 0.6079 - val_loss: 0.8403 - val_accuracy: 0.6810\n",
            "Epoch 7/50\n",
            "22/22 [==============================] - 8s 370ms/step - loss: 0.9025 - accuracy: 0.6646 - val_loss: 0.8396 - val_accuracy: 0.6810\n",
            "Epoch 8/50\n",
            "22/22 [==============================] - 8s 362ms/step - loss: 0.9785 - accuracy: 0.6151 - val_loss: 0.9120 - val_accuracy: 0.6638\n",
            "Epoch 9/50\n",
            "22/22 [==============================] - 8s 364ms/step - loss: 0.9480 - accuracy: 0.6295 - val_loss: 0.8665 - val_accuracy: 0.6552\n",
            "Epoch 10/50\n",
            "22/22 [==============================] - 8s 367ms/step - loss: 0.9524 - accuracy: 0.6029 - val_loss: 0.9009 - val_accuracy: 0.7155\n",
            "Epoch 11/50\n",
            "22/22 [==============================] - 8s 366ms/step - loss: 0.9785 - accuracy: 0.6200 - val_loss: 0.8467 - val_accuracy: 0.6207\n",
            "Epoch 12/50\n",
            "22/22 [==============================] - 8s 378ms/step - loss: 0.9580 - accuracy: 0.6236 - val_loss: 0.8856 - val_accuracy: 0.6724\n",
            "Epoch 13/50\n",
            "22/22 [==============================] - 8s 352ms/step - loss: 0.9843 - accuracy: 0.6004 - val_loss: 0.8175 - val_accuracy: 0.6552\n",
            "Epoch 14/50\n",
            "22/22 [==============================] - 8s 369ms/step - loss: 0.9964 - accuracy: 0.6260 - val_loss: 0.7887 - val_accuracy: 0.6897\n",
            "Epoch 15/50\n",
            "22/22 [==============================] - 8s 366ms/step - loss: 0.8784 - accuracy: 0.6440 - val_loss: 0.8300 - val_accuracy: 0.6810\n",
            "Epoch 16/50\n",
            "22/22 [==============================] - 8s 364ms/step - loss: 0.9209 - accuracy: 0.6447 - val_loss: 0.7885 - val_accuracy: 0.7155\n",
            "Epoch 17/50\n",
            "22/22 [==============================] - 8s 367ms/step - loss: 0.8846 - accuracy: 0.6428 - val_loss: 0.8698 - val_accuracy: 0.6121\n",
            "Epoch 18/50\n",
            "22/22 [==============================] - 8s 361ms/step - loss: 0.8597 - accuracy: 0.6661 - val_loss: 0.7968 - val_accuracy: 0.6724\n",
            "Epoch 19/50\n",
            "22/22 [==============================] - 8s 361ms/step - loss: 0.9185 - accuracy: 0.6393 - val_loss: 0.8708 - val_accuracy: 0.7069\n",
            "Epoch 20/50\n",
            "22/22 [==============================] - 8s 362ms/step - loss: 0.9198 - accuracy: 0.6422 - val_loss: 0.8409 - val_accuracy: 0.6724\n",
            "Epoch 21/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 0.9887 - accuracy: 0.6069 - val_loss: 0.8029 - val_accuracy: 0.6810\n",
            "Epoch 22/50\n",
            "22/22 [==============================] - 8s 355ms/step - loss: 0.9235 - accuracy: 0.6581 - val_loss: 0.8156 - val_accuracy: 0.6983\n",
            "Epoch 23/50\n",
            "22/22 [==============================] - 8s 355ms/step - loss: 0.8802 - accuracy: 0.6688 - val_loss: 0.8410 - val_accuracy: 0.6724\n",
            "Epoch 24/50\n",
            "22/22 [==============================] - 8s 371ms/step - loss: 0.9687 - accuracy: 0.5850 - val_loss: 0.8335 - val_accuracy: 0.7155\n",
            "Epoch 25/50\n",
            "22/22 [==============================] - 8s 350ms/step - loss: 0.9075 - accuracy: 0.6540 - val_loss: 0.8404 - val_accuracy: 0.6897\n",
            "Epoch 26/50\n",
            "22/22 [==============================] - 8s 361ms/step - loss: 0.8982 - accuracy: 0.6447 - val_loss: 0.8761 - val_accuracy: 0.6121\n",
            "Epoch 27/50\n",
            "22/22 [==============================] - 8s 351ms/step - loss: 0.8499 - accuracy: 0.6558 - val_loss: 0.8958 - val_accuracy: 0.6897\n",
            "Epoch 28/50\n",
            "22/22 [==============================] - 8s 356ms/step - loss: 0.9031 - accuracy: 0.6219 - val_loss: 0.7975 - val_accuracy: 0.6897\n",
            "Epoch 29/50\n",
            "22/22 [==============================] - 8s 355ms/step - loss: 0.9410 - accuracy: 0.6175 - val_loss: 0.8776 - val_accuracy: 0.6293\n",
            "Epoch 30/50\n",
            "22/22 [==============================] - 8s 357ms/step - loss: 0.9152 - accuracy: 0.6362 - val_loss: 0.8393 - val_accuracy: 0.6379\n",
            "Epoch 31/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 0.9328 - accuracy: 0.5936 - val_loss: 0.8816 - val_accuracy: 0.6552\n",
            "Epoch 32/50\n",
            "22/22 [==============================] - 8s 359ms/step - loss: 0.8789 - accuracy: 0.6257 - val_loss: 0.7955 - val_accuracy: 0.6724\n",
            "Epoch 33/50\n",
            "22/22 [==============================] - 8s 359ms/step - loss: 0.9720 - accuracy: 0.5950 - val_loss: 0.7634 - val_accuracy: 0.6810\n",
            "Epoch 34/50\n",
            "22/22 [==============================] - 8s 365ms/step - loss: 0.9531 - accuracy: 0.6460 - val_loss: 0.8714 - val_accuracy: 0.6121\n",
            "Epoch 35/50\n",
            "22/22 [==============================] - 8s 361ms/step - loss: 0.9510 - accuracy: 0.6147 - val_loss: 0.8564 - val_accuracy: 0.6638\n",
            "Epoch 36/50\n",
            "22/22 [==============================] - 8s 363ms/step - loss: 0.9871 - accuracy: 0.6287 - val_loss: 0.8462 - val_accuracy: 0.6552\n",
            "Epoch 37/50\n",
            "22/22 [==============================] - 8s 364ms/step - loss: 0.8658 - accuracy: 0.6583 - val_loss: 0.9505 - val_accuracy: 0.5862\n",
            "Epoch 38/50\n",
            "22/22 [==============================] - 8s 359ms/step - loss: 0.9226 - accuracy: 0.6287 - val_loss: 0.8890 - val_accuracy: 0.6207\n",
            "Epoch 39/50\n",
            "22/22 [==============================] - 8s 359ms/step - loss: 0.9490 - accuracy: 0.6051 - val_loss: 0.7583 - val_accuracy: 0.6552\n",
            "Epoch 40/50\n",
            "22/22 [==============================] - 8s 360ms/step - loss: 0.9726 - accuracy: 0.6203 - val_loss: 0.8568 - val_accuracy: 0.6379\n",
            "Epoch 41/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 0.9215 - accuracy: 0.6508 - val_loss: 0.8070 - val_accuracy: 0.6983\n",
            "Epoch 42/50\n",
            "22/22 [==============================] - 8s 356ms/step - loss: 0.9137 - accuracy: 0.6555 - val_loss: 0.8066 - val_accuracy: 0.7241\n",
            "Epoch 43/50\n",
            "22/22 [==============================] - 8s 349ms/step - loss: 0.9460 - accuracy: 0.6240 - val_loss: 0.7148 - val_accuracy: 0.7759\n",
            "Epoch 44/50\n",
            "22/22 [==============================] - 8s 361ms/step - loss: 0.8746 - accuracy: 0.6925 - val_loss: 0.8155 - val_accuracy: 0.6724\n",
            "Epoch 45/50\n",
            "22/22 [==============================] - 8s 356ms/step - loss: 0.9108 - accuracy: 0.6495 - val_loss: 0.7649 - val_accuracy: 0.7241\n",
            "Epoch 46/50\n",
            "22/22 [==============================] - 8s 359ms/step - loss: 0.9240 - accuracy: 0.6404 - val_loss: 0.8448 - val_accuracy: 0.6379\n",
            "Epoch 47/50\n",
            "22/22 [==============================] - 8s 363ms/step - loss: 0.8459 - accuracy: 0.6368 - val_loss: 0.8080 - val_accuracy: 0.7241\n",
            "Epoch 48/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 0.9526 - accuracy: 0.6352 - val_loss: 0.8536 - val_accuracy: 0.6897\n",
            "Epoch 49/50\n",
            "22/22 [==============================] - 8s 359ms/step - loss: 0.8661 - accuracy: 0.6530 - val_loss: 0.7759 - val_accuracy: 0.6983\n",
            "Epoch 50/50\n",
            "22/22 [==============================] - 8s 364ms/step - loss: 0.8632 - accuracy: 0.6503 - val_loss: 0.8321 - val_accuracy: 0.6810\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-Ma8KpXM9nKa",
        "outputId": "32f76fb5-7587-446d-d38a-55961bb52ad3"
      },
      "source": [
        "opt = tf.keras.optimizers.Adam(learning_rate=0.00001)\n",
        "\n",
        "model.compile(optimizer=opt, loss=\"categorical_crossentropy\", metrics=[\"accuracy\"])\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "mobilenetv2_1.00_224 (Functi (None, 5, 5, 1280)        2257984   \n",
            "_________________________________________________________________\n",
            "flatten_2 (Flatten)          (None, 32000)             0         \n",
            "_________________________________________________________________\n",
            "dense_5 (Dense)              (None, 32)                1024032   \n",
            "_________________________________________________________________\n",
            "dropout_3 (Dropout)          (None, 32)                0         \n",
            "_________________________________________________________________\n",
            "dense_6 (Dense)              (None, 5)                 165       \n",
            "=================================================================\n",
            "Total params: 3,282,181\n",
            "Trainable params: 3,248,069\n",
            "Non-trainable params: 34,112\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wdoTdX-Z7w8_",
        "outputId": "a4491477-e126-41d2-fb7b-8ea9ae1bb3b6"
      },
      "source": [
        "hist= model.fit(train_generator, validation_data = valid_generator, epochs = 50)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "22/22 [==============================] - 13s 420ms/step - loss: 0.1545 - accuracy: 0.9483 - val_loss: 1.1307 - val_accuracy: 0.7759\n",
            "Epoch 2/50\n",
            "22/22 [==============================] - 8s 340ms/step - loss: 0.1260 - accuracy: 0.9456 - val_loss: 0.7964 - val_accuracy: 0.8276\n",
            "Epoch 3/50\n",
            "22/22 [==============================] - 8s 364ms/step - loss: 0.1555 - accuracy: 0.9461 - val_loss: 1.1572 - val_accuracy: 0.8103\n",
            "Epoch 4/50\n",
            "22/22 [==============================] - 8s 356ms/step - loss: 0.1263 - accuracy: 0.9527 - val_loss: 0.9403 - val_accuracy: 0.8017\n",
            "Epoch 5/50\n",
            "22/22 [==============================] - 8s 371ms/step - loss: 0.2452 - accuracy: 0.9173 - val_loss: 0.9349 - val_accuracy: 0.7931\n",
            "Epoch 6/50\n",
            "22/22 [==============================] - 8s 360ms/step - loss: 0.1402 - accuracy: 0.9538 - val_loss: 0.9503 - val_accuracy: 0.7931\n",
            "Epoch 7/50\n",
            "22/22 [==============================] - 8s 363ms/step - loss: 0.1224 - accuracy: 0.9578 - val_loss: 1.2996 - val_accuracy: 0.7672\n",
            "Epoch 8/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 0.2531 - accuracy: 0.9109 - val_loss: 1.0807 - val_accuracy: 0.7586\n",
            "Epoch 9/50\n",
            "22/22 [==============================] - 8s 363ms/step - loss: 0.2030 - accuracy: 0.9274 - val_loss: 0.7447 - val_accuracy: 0.8190\n",
            "Epoch 10/50\n",
            "22/22 [==============================] - 8s 354ms/step - loss: 0.1428 - accuracy: 0.9501 - val_loss: 1.1423 - val_accuracy: 0.8362\n",
            "Epoch 11/50\n",
            "22/22 [==============================] - 8s 360ms/step - loss: 0.1662 - accuracy: 0.9419 - val_loss: 0.9973 - val_accuracy: 0.7845\n",
            "Epoch 12/50\n",
            "22/22 [==============================] - 8s 372ms/step - loss: 0.3234 - accuracy: 0.9157 - val_loss: 1.1006 - val_accuracy: 0.7931\n",
            "Epoch 13/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 0.1521 - accuracy: 0.9629 - val_loss: 0.9238 - val_accuracy: 0.7672\n",
            "Epoch 14/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 0.1501 - accuracy: 0.9485 - val_loss: 1.0969 - val_accuracy: 0.7931\n",
            "Epoch 15/50\n",
            "22/22 [==============================] - 8s 359ms/step - loss: 0.1385 - accuracy: 0.9597 - val_loss: 1.0417 - val_accuracy: 0.7931\n",
            "Epoch 16/50\n",
            "22/22 [==============================] - 8s 360ms/step - loss: 0.1604 - accuracy: 0.9561 - val_loss: 0.9758 - val_accuracy: 0.8017\n",
            "Epoch 17/50\n",
            "22/22 [==============================] - 8s 341ms/step - loss: 0.1577 - accuracy: 0.9488 - val_loss: 1.2755 - val_accuracy: 0.7586\n",
            "Epoch 18/50\n",
            "22/22 [==============================] - 8s 359ms/step - loss: 0.1039 - accuracy: 0.9601 - val_loss: 1.0282 - val_accuracy: 0.7759\n",
            "Epoch 19/50\n",
            "22/22 [==============================] - 8s 352ms/step - loss: 0.1136 - accuracy: 0.9566 - val_loss: 1.0814 - val_accuracy: 0.7759\n",
            "Epoch 20/50\n",
            "22/22 [==============================] - 8s 364ms/step - loss: 0.1574 - accuracy: 0.9481 - val_loss: 0.8772 - val_accuracy: 0.7845\n",
            "Epoch 21/50\n",
            "22/22 [==============================] - 8s 361ms/step - loss: 0.1142 - accuracy: 0.9679 - val_loss: 1.0447 - val_accuracy: 0.7845\n",
            "Epoch 22/50\n",
            "22/22 [==============================] - 8s 364ms/step - loss: 0.1705 - accuracy: 0.9476 - val_loss: 1.3502 - val_accuracy: 0.7672\n",
            "Epoch 23/50\n",
            "22/22 [==============================] - 8s 360ms/step - loss: 0.1326 - accuracy: 0.9622 - val_loss: 1.2320 - val_accuracy: 0.7672\n",
            "Epoch 24/50\n",
            "22/22 [==============================] - 8s 354ms/step - loss: 0.1354 - accuracy: 0.9524 - val_loss: 1.0215 - val_accuracy: 0.7759\n",
            "Epoch 25/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 0.1307 - accuracy: 0.9500 - val_loss: 1.4021 - val_accuracy: 0.7500\n",
            "Epoch 26/50\n",
            "22/22 [==============================] - 8s 361ms/step - loss: 0.1465 - accuracy: 0.9484 - val_loss: 1.2913 - val_accuracy: 0.7241\n",
            "Epoch 27/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 0.1155 - accuracy: 0.9526 - val_loss: 0.9996 - val_accuracy: 0.7672\n",
            "Epoch 28/50\n",
            "22/22 [==============================] - 8s 331ms/step - loss: 0.1153 - accuracy: 0.9598 - val_loss: 1.0251 - val_accuracy: 0.8017\n",
            "Epoch 29/50\n",
            "22/22 [==============================] - 8s 357ms/step - loss: 0.0986 - accuracy: 0.9675 - val_loss: 1.2596 - val_accuracy: 0.7845\n",
            "Epoch 30/50\n",
            "22/22 [==============================] - 8s 359ms/step - loss: 0.1136 - accuracy: 0.9490 - val_loss: 1.1649 - val_accuracy: 0.8017\n",
            "Epoch 31/50\n",
            "22/22 [==============================] - 8s 353ms/step - loss: 0.1768 - accuracy: 0.9476 - val_loss: 0.8531 - val_accuracy: 0.8103\n",
            "Epoch 32/50\n",
            "22/22 [==============================] - 8s 360ms/step - loss: 0.1583 - accuracy: 0.9424 - val_loss: 1.2586 - val_accuracy: 0.7672\n",
            "Epoch 33/50\n",
            "22/22 [==============================] - 8s 346ms/step - loss: 0.1152 - accuracy: 0.9507 - val_loss: 0.9993 - val_accuracy: 0.7672\n",
            "Epoch 34/50\n",
            "22/22 [==============================] - 8s 360ms/step - loss: 0.1347 - accuracy: 0.9626 - val_loss: 0.9302 - val_accuracy: 0.7931\n",
            "Epoch 35/50\n",
            "22/22 [==============================] - 8s 361ms/step - loss: 0.1363 - accuracy: 0.9594 - val_loss: 0.8714 - val_accuracy: 0.7845\n",
            "Epoch 36/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 0.1067 - accuracy: 0.9680 - val_loss: 1.1498 - val_accuracy: 0.7672\n",
            "Epoch 37/50\n",
            "22/22 [==============================] - 8s 358ms/step - loss: 0.1359 - accuracy: 0.9537 - val_loss: 1.0654 - val_accuracy: 0.7759\n",
            "Epoch 38/50\n",
            "22/22 [==============================] - 8s 357ms/step - loss: 0.1428 - accuracy: 0.9526 - val_loss: 0.9927 - val_accuracy: 0.7759\n",
            "Epoch 39/50\n",
            "22/22 [==============================] - 8s 369ms/step - loss: 0.1646 - accuracy: 0.9073 - val_loss: 1.2484 - val_accuracy: 0.8190\n",
            "Epoch 40/50\n",
            "22/22 [==============================] - 8s 363ms/step - loss: 0.1027 - accuracy: 0.9700 - val_loss: 1.1297 - val_accuracy: 0.7931\n",
            "Epoch 41/50\n",
            "22/22 [==============================] - 8s 339ms/step - loss: 0.0948 - accuracy: 0.9748 - val_loss: 0.8669 - val_accuracy: 0.8103\n",
            "Epoch 42/50\n",
            "22/22 [==============================] - 8s 362ms/step - loss: 0.0986 - accuracy: 0.9673 - val_loss: 1.0787 - val_accuracy: 0.7845\n",
            "Epoch 43/50\n",
            "22/22 [==============================] - 8s 372ms/step - loss: 0.2349 - accuracy: 0.9084 - val_loss: 1.0735 - val_accuracy: 0.7759\n",
            "Epoch 44/50\n",
            "22/22 [==============================] - 8s 362ms/step - loss: 0.1847 - accuracy: 0.9229 - val_loss: 1.0052 - val_accuracy: 0.7845\n",
            "Epoch 45/50\n",
            "22/22 [==============================] - 8s 361ms/step - loss: 0.1464 - accuracy: 0.9504 - val_loss: 1.0644 - val_accuracy: 0.7845\n",
            "Epoch 46/50\n",
            "22/22 [==============================] - 8s 363ms/step - loss: 0.1222 - accuracy: 0.9563 - val_loss: 1.2146 - val_accuracy: 0.7759\n",
            "Epoch 47/50\n",
            "22/22 [==============================] - 8s 356ms/step - loss: 0.1326 - accuracy: 0.9404 - val_loss: 1.1109 - val_accuracy: 0.7845\n",
            "Epoch 48/50\n",
            "22/22 [==============================] - 8s 350ms/step - loss: 0.1405 - accuracy: 0.9435 - val_loss: 0.9316 - val_accuracy: 0.8190\n",
            "Epoch 49/50\n",
            "22/22 [==============================] - 8s 360ms/step - loss: 0.1473 - accuracy: 0.9513 - val_loss: 1.1958 - val_accuracy: 0.7586\n",
            "Epoch 50/50\n",
            "22/22 [==============================] - 8s 362ms/step - loss: 0.1735 - accuracy: 0.9467 - val_loss: 1.2116 - val_accuracy: 0.7672\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EDTPQ4zlQGpd"
      },
      "source": [
        "Final Stats"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "P1G5robigyTt",
        "outputId": "92d8757f-904b-4ba0-983c-a6181e1ac733"
      },
      "source": [
        "final_loss, final_accuracy = model.evaluate(valid_generator)\n",
        "print(\"Final Validation loss: {:.2f}\".format(final_loss))\n",
        "print(\"Final Validation accuracy: {:.2f}%\".format(final_accuracy * 100))\n",
        "\n",
        "final_loss, final_accuracy = model.evaluate(train_generator)\n",
        "print(\"Final training loss: {:.2f}\".format(final_loss))\n",
        "print(\"Final training accuracy: {:.2f}%\".format(final_accuracy * 100))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "4/4 [==============================] - 1s 305ms/step - loss: 0.7183 - accuracy: 0.8276\n",
            "Final Validation loss: 0.72\n",
            "Final Validation accuracy: 82.76%\n",
            "22/22 [==============================] - 6s 273ms/step - loss: 0.0357 - accuracy: 0.9896\n",
            "Final training loss: 0.04\n",
            "Final training accuracy: 98.96%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Vn3bIrS7QIme"
      },
      "source": [
        "Saving Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "80TCnkz-g0ym"
      },
      "source": [
        "model.save('/content/gdrive/MyDrive/Infections/infection_model_v2.5.h5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Gmvpl6sfNqgX"
      },
      "source": [
        "Prediction on Validation Set"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eHWuMSqnOAsM",
        "outputId": "253786e8-9186-4bd3-f93d-46b6d29895c4"
      },
      "source": [
        "import numpy as np\n",
        "\n",
        "val_image_batch, val_label_batch = next(iter(valid_generator))\n",
        "true_label_ids = np.argmax(val_label_batch, axis=-1)\n",
        "print(\"Validation batch shape:\", val_image_batch.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Validation batch shape: (32, 150, 150, 3)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-xx1ZeoCOLdc",
        "outputId": "d83c3e03-5007-416b-85b4-4a596fb5e6a9"
      },
      "source": [
        "dataset_labels = sorted(train_generator.class_indices.items(), key=lambda pair:pair[1])\n",
        "dataset_labels = np.array([key.title() for key, value in dataset_labels])\n",
        "print(dataset_labels)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "['Blepharitis' 'Cellulitis Eye' 'Conjunctivitis, Pink Eye' 'Eye'\n",
            " 'Hordeolum, Stye']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YChOYzx8NqLN",
        "outputId": "288023ea-c128-4ab5-db05-92a16be39edf"
      },
      "source": [
        "tf_model_predictions = model.predict(val_image_batch)\n",
        "print(\"Prediction results shape:\", tf_model_predictions.shape)\n",
        "print(tf_model_predictions)\n",
        "predicted_ids = np.argmax(tf_model_predictions, axis=-1)\n",
        "predicted_labels = dataset_labels[predicted_ids]\n",
        "print(predicted_labels)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Prediction results shape: (32, 5)\n",
            "[[1.11890696e-02 5.54975748e-01 1.43679703e-04 1.91179069e-03\n",
            "  4.31779683e-01]\n",
            " [5.09586253e-12 1.16555066e-09 2.99375031e-08 1.00000000e+00\n",
            "  3.19376134e-08]\n",
            " [1.96978566e-12 1.00000000e+00 4.50140425e-10 1.91115635e-12\n",
            "  1.31474067e-08]\n",
            " [3.32312938e-03 6.78713404e-05 8.47166521e-04 5.85179623e-05\n",
            "  9.95703280e-01]\n",
            " [1.85604231e-03 2.04600915e-01 7.34460627e-05 1.27571897e-04\n",
            "  7.93341994e-01]\n",
            " [7.11561415e-06 2.34162304e-04 1.58641109e-04 1.09614735e-03\n",
            "  9.98503923e-01]\n",
            " [2.21225317e-04 1.14122686e-05 9.99663949e-01 4.21737823e-05\n",
            "  6.12696313e-05]\n",
            " [9.85490680e-01 3.76331968e-06 1.44295339e-02 7.19598975e-05\n",
            "  4.13697171e-06]\n",
            " [3.20747660e-13 1.00000000e+00 3.11273251e-13 4.70478564e-15\n",
            "  6.36770974e-12]\n",
            " [9.86473083e-01 1.98478345e-03 9.55302501e-04 3.95063224e-04\n",
            "  1.01917787e-02]\n",
            " [3.53542446e-21 1.32684179e-17 5.22400082e-14 1.00000000e+00\n",
            "  6.40264595e-17]\n",
            " [6.90964271e-06 9.99968052e-01 1.66723312e-05 7.18181496e-08\n",
            "  8.30196768e-06]\n",
            " [4.02595006e-08 9.99999881e-01 4.93136532e-10 1.86941254e-10\n",
            "  1.45828395e-07]\n",
            " [9.37180885e-04 9.86433864e-01 5.72646513e-06 5.53327845e-06\n",
            "  1.26175955e-02]\n",
            " [4.79372628e-02 1.48678906e-02 9.92964581e-02 1.90779015e-01\n",
            "  6.47119403e-01]\n",
            " [5.29330410e-02 3.38281244e-02 7.30811208e-02 7.77528405e-01\n",
            "  6.26292825e-02]\n",
            " [2.93042604e-11 2.64143820e-14 1.00000000e+00 4.70083889e-15\n",
            "  3.97652506e-13]\n",
            " [2.16314402e-05 5.57053818e-06 9.99968290e-01 4.19190246e-06\n",
            "  3.05333884e-07]\n",
            " [1.91939868e-15 1.12195832e-12 2.01799168e-11 1.00000000e+00\n",
            "  1.81684651e-13]\n",
            " [5.65334558e-05 4.32604924e-04 2.85311043e-02 9.66759086e-01\n",
            "  4.22073109e-03]\n",
            " [7.55517959e-08 7.32922636e-05 2.30914393e-05 1.18029936e-06\n",
            "  9.99902368e-01]\n",
            " [9.93782103e-01 3.65987391e-04 3.93352704e-04 7.22097000e-04\n",
            "  4.73646168e-03]\n",
            " [1.69063132e-04 2.07482744e-03 1.92127214e-03 2.11256696e-03\n",
            "  9.93722260e-01]\n",
            " [9.95469809e-01 6.85744362e-06 4.70561281e-05 2.09208025e-04\n",
            "  4.26711142e-03]\n",
            " [8.90140785e-25 2.19729396e-17 5.88877157e-17 1.00000000e+00\n",
            "  6.51454405e-18]\n",
            " [9.82110141e-05 5.84083963e-08 1.21838024e-08 1.29129730e-05\n",
            "  9.99888778e-01]\n",
            " [7.25778690e-08 3.39434706e-07 1.46404316e-04 9.99845862e-01\n",
            "  7.44392173e-06]\n",
            " [1.85912466e-15 5.63910585e-10 1.96297284e-10 1.00000000e+00\n",
            "  5.90804183e-10]\n",
            " [4.86326257e-10 1.61150222e-08 7.91893981e-08 9.99999881e-01\n",
            "  5.76389390e-08]\n",
            " [2.77454342e-06 9.98903394e-01 4.21123468e-06 2.49712821e-06\n",
            "  1.08723785e-03]\n",
            " [5.96715808e-02 2.47714831e-03 4.96221182e-05 1.01923099e-04\n",
            "  9.37699676e-01]\n",
            " [9.44996715e-01 5.61627885e-03 1.45076187e-02 1.38140807e-03\n",
            "  3.34979445e-02]]\n",
            "['Cellulitis Eye' 'Eye' 'Cellulitis Eye' 'Hordeolum, Stye'\n",
            " 'Hordeolum, Stye' 'Hordeolum, Stye' 'Conjunctivitis, Pink Eye'\n",
            " 'Blepharitis' 'Cellulitis Eye' 'Blepharitis' 'Eye' 'Cellulitis Eye'\n",
            " 'Cellulitis Eye' 'Cellulitis Eye' 'Hordeolum, Stye' 'Eye'\n",
            " 'Conjunctivitis, Pink Eye' 'Conjunctivitis, Pink Eye' 'Eye' 'Eye'\n",
            " 'Hordeolum, Stye' 'Blepharitis' 'Hordeolum, Stye' 'Blepharitis' 'Eye'\n",
            " 'Hordeolum, Stye' 'Eye' 'Eye' 'Eye' 'Cellulitis Eye' 'Hordeolum, Stye'\n",
            " 'Blepharitis']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 393
        },
        "id": "NQ4ZK94iNhkc",
        "outputId": "9bd0110b-b170-477d-88d1-42df544f6521"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.figure(figsize=(10,9))\n",
        "plt.subplots_adjust(hspace=0.5)\n",
        "for n in range((len(predicted_labels)-2)):\n",
        "  plt.subplot(8,6,n+1)\n",
        "  plt.imshow(val_image_batch[n])\n",
        "  color = \"green\" if predicted_ids[n] == true_label_ids[n] else \"red\"\n",
        "  plt.title(predicted_labels[n].title(), color=color)\n",
        "  plt.axis('off')\n",
        "_ = plt.suptitle(\"Model predictions (green: correct, red: incorrect)\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAngAAAF4CAYAAADHf4SnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOy9ebRlSVWg/+2IOOcOb8qxhswamaHEoVsEbFFaUcQWZdkOjcpgK0rbojTNoDb9ezzbWs6idovaoM0kiqgoODSDbTGJoiiDQAFVVFJZZFWOb7zDGSL27484972br97LoczMl3krvrXuuvdMEXH2iXPOjj3EFVUlkUgkEolEIjE5mJ1uQCKRSCQSiUTiwpIUvEQikUgkEokJIyl4iUQikUgkEhNGUvASiUQikUgkJoyk4CUSiUQikUhMGEnBSyQSiUQikZgwkoKXmFhE5CYRURFx57Dvc0Xk/ZeiXVvUfVo7ReQvReQ5D6CcG0RkTUTshW/lGet9qoj8yaWsM3HhEJFXiMgbz3HfJ4nIpy92my4lIvJ0EXnzTrcjkbjQJAUvcVkgIodEpBSRfZvW/1Oj/Ny0My279Kjq01T1dWfbr5HZU8aOu1tVp1XVX9wW3o9bgZ+9xHVOJCLyZBG5Z6fbsR2q+j5VfeROt+OBstWgT1XfDtwiIl+8g01LJC44ScFLXE7cBTxztCAijwW6O9ecB8a5WAwnBRF5HDCnqn/7AI+/4mW12WK6k+c0CfLczOZzksiFfnf9HvCDF7jMRGJHSQpe4nLiDcCzx5afA7x+fAcRmROR14vIcRH5vIi8fPSwFxErIr8oIidE5HPAv9vi2N8WkXtF5Asi8tPn4s4cG/X/oIgcaY5/8dj2V4jIH4rIG0VkBXjumeo6h3beJiI/MLb8PBH5lIisisgnReRficgbgBuAtzdu2Zdu4eo9ICJvE5FTInKHiDxvU5v/oJHlqoh8QkS+fGz7y5p2r4rIp0Xk67YRz9OA92xq/zc0xyyLyKtE5D2j82lc4R8QkVeKyEngFSLSauRxt4gcFZHfFJHOWHnfLCIfEZElEfmbcUtLY8V8sYh8rKnvzSLSPts1PZNsm/WPbq7DUiObbxk75rUi8hsi8hci0gP+bdOOl4nIx4CeiDgReULT3iUR+aiIPHmsjD0i8n+a/rQoIn8iIlPAXwIHmmu6JiIHzuEczrfum5trsioi7wL2bVv4/es6zcJ4NvmLyLc2125FRO4UkW9s1p+tb26+n24TkVtF5ANAH3iIiDxKRN7VlPFpEfnOsTI6IvJLEp8RyyLy/qZPvbfZZamR7xOb5dvYdB8mElc8qpo+6bPjH+AQ8BTg08CjAQvcA9wIKHBTs9/rgT8FZoCbgM8A399sez5wO3A9sAf46+ZY12x/K/BbwBRwFfAh4Ieabc8F3r9N225qyvm95tjHAseBpzTbXwFUwDOIg6bOWeo6WztvA36g+f0dwBeAxwECPAy4cVxmW7RzVM57gVcBbeBLmzZ/7Vibh8A3NbL+GeBvm22PBA4DB8bKfeg2snkL8JKx5X3ACvBtgAN+rJHND4zJuQZe0GzvAK8E3tbIYgZ4O/Azzf5fBhwDHt+08znNebfGZPAh4EBz/KeA54+1Zwn4qm3avqVsgQy4A/hJIAe+FlgFHtkc91pgGfg3zfVuN+34SHNNO8BB4GQjXwN8fbO8vynjz4E3A7ub+r6mWf9k4J4HcO+cT90fBH4ZaAFf3ZzbG8fK+xjw3dvUdVr7ziR/4CsaOX19046DwKPOsW9uvp9uA+4GbiH2mzliH/2+ZvnLgBPAY5oyfr055iCx33xlc743MXaPjJ3Hnmb97E4/C9MnfS7UZ8cbkD7po3qagvdyorLxjcC7moe3Ng9mC5Sjh3hz3A8BtzW//x+nv9y/YfQwB64GCqAztv2ZwF83v5/L2RW8R42t+3ngt5vfrwDeO7btbHVt285m+TY2FKJ3AD92Jplt0U5HfNl7YGZs+88Arx1r87vHtj0GGDS/H0ZUqp4CZGe5bu/adC7PBj44tizNi3hcwbt70/YeYwok8ETgrub3bwD/Y1Odn2ZDIToEfO+m6/Kb59jntpQt8CTgPsCMrfs94BXN79cCr9/iWvzHseWXAW/Yor7nANcCAdi9Rd1P5oEpeOda9w1EBXtqbNubGFPwzlLXae07k/yJA5xXblHGufTN92465jbgp8aWvwt436Z9fguYJyqFA+BLtqj7JrZW8LJm/Q3nI/v0SZ/L+TNx8RqJK543EEf3N7PJPUu0DmXA58fWfZ44SodoRTi8aduIkWXmXhEZrTOb9j8bm8t+7DbbzlbXmdq5meuBO8+jjSMOAKdUdXVTPV8+tnzf2O8+0BYRp6p3iMgLiS/aW0TkHcCLVPXIFvUsEq1u4/Wun5uqqtw/aWD83PcT4yw/PCYrISrzEGX5HBF5wdgxeVPPdudxVrdmw3ayPQAcVtUwtm68n20+h63W3Qh8h4g8fWxdRrTWXk+8Novn2M5z4VzrPgAsqmpvbNvnmzY9ULaT//XAX2yx/7n0zXOR7+NFZGlsnSM+P/YRLYPnc9+M+vDSGfdKJK4gkoKXuKxQ1c+LyF1E99L3b9p8gui6uRH4ZLPuBqKbDeBeTn9R3TD2+zDRqrZPVesH2Lzria7VUdnjCo+eR11naudmDgMP3WabbrOepm17RGRm7EU6LqszoqpvAt4kIrNEy8jPAc/aYtePAY8YW74XuG60IFFru27TMePtPkG0ttyiqlu17TBwq6reei7tPk+2k+0R4HoRMWNK3g3EcIARW8l+cx94g6o+b/NOInIt8drsUtXNCsWZrumZONe6bwR2i8jUmJJ3w7+g3jNxJvmerW+ei3zfo6pfv3kniTG5w6buj56hjHEeDRxS1ZVtticSVxwpySJxOfL9xHiccSsDGqf/+APgVhGZaV5WLwJGc3j9AfCjInKdiOwGfnzs2HuBdwK/JCKzImJE5KEi8jXn0a7/LiJdEbmFGPuz5dxZ51DXtu3cgtcALxaRfy2RhzXnDXAUeMg2bTgM/A3wMyLSlpiY8P1syGpbROSRIvK1ItIivigHRJfiVvwFMC7DPwceKyLPkJjs8Z+Ba7arq1GgXg28UkSuauo/KCJPbXZ5NfB8EXl8c/5TIvLvRGRmuzLPg+1k+3dES9RLRSRrEhSeDvz+eZT9RuDpEucItM01eLKIXNf0j78EXiUiu5s6vro57iiwV0TmRgU1x52PAnamuj8P/AOwICK5iHxVc24Xg98Gvk9Evq65Bw6KyKP+JX1zjD8DHiEiz2rkl4nI40Tk0U2f+h3glyUmc1gReWLTn48T+/Lm++ZriNckkZgYkoKXuOxQ1TtV9R+22fwCYszW54D3E+OHfqfZ9mpirNFHgX8E/njTsc8muvc+SXQt/iExHupceQ8x+P6vgF9U1XeeYd8z1XW2dq6jqm8hzjP3JmIw/J8QA8Ihxi29XGKm5Iu3OPyZxJijI8Skj3lVffdZzzIGo/8s0bp2HzFJ5Ce2ad8/Assi8vhm+QQxeeHniYH9jyEqFMUZ6nsZUa5/KzFr8t3ERA+afvA84H8R5XgHMY7vnGgyJZ+0Tdu3lK2qlkSl52lEGbwKeLaq3r5VOduUfRj4VmKixnGixeklbDxzn0W0Rt9OjHd8YXPc7cR4v8811/UA0dr7Nxew7u8mJq2cIsasbc5U/4SIfM+51neGdnyIOBB6JTHZ4j1E6zs88L45KnuVGLv6H5oy7iNamVvNLi8GPg78PfE8f44YU9knXvMPNPJ9wlh7fusBnWgicZkiqhfDMp9ITA4SJ1m+i5hw8EDduxOLiHwD8MOq+owtthliNvT3qOpfX/LGTQAi8hrgLar6jp1uyyTSxCo+S1W/86w7JxJXEEnBSyTOQlLwzo/Gvfp3RNfuS4hu2oeo6mBHG5ZIJBIPIpKLNpFIXGieSMxgPEF0dT4jKXeJRCJxaUkWvEQikUgkEokJI1nwEolEIpFIJCaMpOAlEolEIpFITBhJwUskEolEIpGYMJKCl0gkEolEIjFhJAUvkUgkEolEYsJICl4ikUgkEonEhJEUvEQikUgkEokJIyl4iUQikUgkEhNGUvASiUQikUgkJoyk4CUSiUQikUhMGEnBSyQSiUQikZgwkoKXSCQSiUQiMWEkBS+RSCQSiURiwkgKXiKRSCQSicSEkRS8RCKRSCQSiQkjKXiJRCKRSCQSE0ZS8BKJRCKRSCQmjKTgJRKJRCKRSEwYScFLJBKJRCKRmDCSgpdIJBKJRCIxYSQFL5FIJBKJRGLCSApeIpFIJBKJxISRFLxEIpFIJBKJCSMpeIlEIpFIJBITRlLwEolEIpFIJCaMpOAlEolEIpFITBhJwUskEolEIpGYMJKCl0gkEolEIjFhJAUvkUgkEolEYsJICl4ikUgkEonEhJEUvEQikUgkEokJIyl4iUQikUgkEhNGUvASiUQikUgkJoyk4CUSiUQikUhMGEnBSyQSiUQikZgwJlrBkwW5SRZEZUFcs3ybLMgPnOOx6/vKgnyPLMg7z7Dvk2RBPn1hWp1I/MtJfX9nkAV5rSzIT18u5TzoEHktcgHkdqHKeRCR+v7lh9vpBpwLsiDfDbwIeBSwCnwEuFXn9f2Xon6d198FfnesPQo8XOf1jmb7+4BHnm+5siCvBb4bKMdW36nz+iX/ogZfociCHAKuBvzY6tfqvP7IzrRo50l9/8LR9K8f0Hl999i65zbrvupi1XulIQvyk8DzgP3AEvABndfvarbdBrxR5/U1519wlD+6IX8kyh9N8l9H7i9/NMofifJHz0/+qe+fGxet7+8Ql72CJwvyIuDHgecD7yC+EL4R+FbgkrzkLjI/r/P68p1uxGXE08cfQg9mUt+/vJEFsTqv/ux7XjnIgjwHeBbwFJ3XO2VBrgG+ZYebtTUiFp0s+SMb8kf1TuTylH/q+1cGl7WCJwsyB/wU8H06r388tuntzQdZEAO8lKh17wL+Cni+zuups5T9CuBhOq/f2yzfBNwFZDqv9aZ9n0sz0pEFeW+z+qONNeP7gaNEzf66Zv+XAT8KzAJHgB/Wef2r8zz3Pwf+r87r/xxb9zFgXuf1rbIgjwL+J/CvgePAf9d5/YPzqeNKQBYkB+4Dvkbn9ePNuquAQ8CNOq/HZUG+Gfhp4Cbgk8Tr/7GdafGFIfX9nen7siCPBn4D+FLgC8BP6Ly+rdn2WmAA3Ah8DfCtsiAngd8GHg78BaCbytu2b56prk1lPJdNlpZxS2rTrj5wM/Ak4KPAvycODp5DvEbP1Hn9p3MQweOAd+i83gmg83of8L+bOm9tyn+CLMivAK8FBBjqvP7Xsba9DfhrnddXyoIcIF6rr757ll2/9nie8Quw/QBO7i8TtJGJ3F/+yJnlj9xf/mjzbDhTXaeX8Vw2Wxklyh/VO5p2nVX+6LnLH43yRzfkj2zIHzld/uiG/JEof1RfiUT5X/9fuP43/ow38Ar5GVR/bauKU9+/eH0fWANeqfNby/5icbnH4D0RaANvPcM+LwCeQex0B4BF4NcvVoN0Xr+6+fklOq/TOq9vHt8uC/JI4EeAx+m8zgBPJSoj58vrgO8dK/dLgIPAn8uCTAHvAt4EXAX8B+BVsiCPeQD1XNbovJbA7zMmC+CZwF81yt2XAb8D/BCwF/gt4G2yIK1L3tgLS+r7G+Vekr4vC5IRled3NmW/APjd5rxGfDdwKzADfAj4E+ANwB7gLcSXy6i8bfvmOdZ1Pnwn8HJgH1AAHwT+sVn+Q+CXz7GcvwWeLQvyElmQL5cFsaMNOq//DXgf8CPN9f8R4rV6ZjPYQBZkH/AU4E3NurcTX7oHr1nj6Kf28+9lQZ66Zc2ytUyQByZ/ZGv5I9I6x7rOhwsqf0RegsiXIxvyRzfkj+o0uiF/JMof2ZB/s+7twEcPz3L4V57AS4AXIveXf+r7wEXs+8DXAS/ctu9fJC53BW8vcGKzVWETzwf+m87rPTqvBfAK4NtHweU7gAdawGNkQTKd10OjEcE2vFgWZGns87pm/duAR8iCPLxZfhbw5kbh+WbgkM7r/9F5rZvRyR8B33GRzulS8ieb5PE8Nm4kafZ5FvHBAvCDwG/pvP6dzqvXeX0d8SZ/wqVv+gUl9f2L0/dP61/Aq8a2PQGYBn5W57XUef1/wJ8RBxQj/lTn9QM6r4FofciAX9F5rXRe/xD4+7F9z9Q3z6Wu8+GtOq8f1nkdEgcFQ53X1zdutDcDX3Yuhei8vpH4wn0q8B7gWGOV3W7/DwHLxBcYRIX7Np3Xo0SLyH6d15/SeS2zQP3Hv8+etVt5GyJLyPbyR7VEt5Y/qh9AT5c/qhW6tfxR/TtUPbq1/M9Q1/nwVlQ/jG7IH9XXNy7kc5Y/en/5I9vLH91a/uiG/FH9KQTe/VB+s/1yrp36Sd6W+v79uZh9X+f1c8Crm30uGZe1ixY4CeyTBXFneNHdCLxVFiSMrfPEYP1LTmM2fiHxZXuLLMg7gBfpvB7Z5pBf1C3ikHReh7Igbwa+VxZkgdj5v73ZfCPw+OYmHeHYUHquZJ6hW8TgyYL0gSfLgtwLPIyoBECUxXNkQV4wtntOtGhdyaS+f3H6/mn9a+QCahYPAIebF9iIzxNH4CMOj/0+AHxB51U37T/iTH0znENd58PRsd+DLZanz7UgbRJrGkvLM5rfH9F5fcc2h4wsru9qvn+1WX8jcGB0rQ7NMfuDT2f4zodxm87rNwEb7s/IAeBwo7yNOKv80TPLH9le/mep63y4YPJHm8Qm2ZA/Ih9BH5j8EVm64YXM/u+303/qndTA+1D9ptT378/F6vsNlmgFvGRc7greB4la/zOIptatOAz8R53XD2ze0MQWbUcP6I4tX/PAmnh/dF7fRDTTzhJN0z9HtEKcL68jvrjeD/R1Xj/YrD8MvEfn9esvRHuvEEY30n3AHzajNYiyuFXn9dYda9nFIfX9S9/3jwDXy4KYsZfPDcBnxvYZf6HdCxyUBZGxF90NwMhquW3flAV50jnUNeK06yUx+Puio/NaAW9prBhfREz00S12fSPwz40r/dFE1x3E879L5zVaYl8hh97xxk1ZtKdzBLgeETOmeJ1V/ojImJJ3P/mjWzwbJMr/LHWNOP1+kUsjfzTKv7HgnVX+yNbyR/Xhdy/IoW981ulZtJtIfX+MC973d4jLWsHTeV2WBfn/gF+XBamJPvuK6Of+tzqvLwV+E7hVFuQ5Oq+flwXZD3ylzuufnqX4jwAvkwW5gWhm/YnzaNpR4CHAHZs3NHEEB4EPAEPiCMJu3u9c0Hn9YGOd+SVOt1D8GfCzsiDPIsanQTSZr+m8fuqB1HUF8EZiPMMqpysMryZasd5NjAvpAk8G3qvzunqpG3mhSH1/R/r+3xEDtl8qC/JLwL8Bnk50t2zFB4Ea+FFZkFc1+34F8NfN9m375nnW9VGiRfRLgduJFtIHjMTAdHRen7vFtucSE1feS3y5PhW4pWkvbFz/dXRe75EF+XvidfojnddBs+lDwGrzkvy1APzxo7np2xfkcTqv4+68EesyQc5P/sj28kfOLP9zqOujwC3IhZF/k5QBen/5NxbN85I/qvcgG/JHT5c/Ii+TeaRVYxD5IqCD3k/+qe9fxL5PnAHh0UBnm75/UbjcY/DQef0l4jxgLycK/zAxkHukKf8q0V33TlmQVWKg5OPPodx3Ef3zHwM+THxxnCuvAF7XxPF856ZtLeBngRNEa9NVnPkF+lJZkLWxz4lN218PPJao4Izavgp8A9Gff6Sp5+eauq903r5JHm8F0Hk9TAycVcbM3Dqv/0DMIv1fxCSDO4DnXvJWXwRS37+0fb+J8Xs68LTmHF4FPFvn9fYz7P9txP52Cvgu4I/Htm/bN8+nLp3XzxAzqt8NfJZ/+RQ51xOV8K1YAX4SuJs4D9jPA/9JN+Zd/FVinOeiLMh4RuDriNdqXRlvYqC+maiA33X3HNe/8Yt5ETC3Zc26tUzQreXf7L+t/NEzPBvOpy7defmjp8sfkUXkzPJv4v++GfjS61a47q2/zx8Br2EL+ae+D1zEvk88zy1lfzGR08MXEpcbsiDPBn5Q02SUyIL8DnBkq7itxOSR+v6FR+K0Qx8FvrhxQ12ocr+aqIjfuCkuKzGObMi/ccFeqHLX5U96qW/Jg7HvX9Yu2gc7siBd4Ic5PdvpQUkTU/ZtnGs2WuKKJvX9i0NjPXn0hSyzCUj/MeA1l9sL7rJDL7z8m2SMHwNek5S77Xkw9v3L3kX7YEXifDnHiX7/N+1wc3YUWZD/Afwz8As6r3ftdHsSF5fU968cJE5YuwRcC/zKDjfnwYck+e8UV0LfTy7aRCKRSCQSiQkjWfASiUQikUgkJoyJj8Hbt3umsVFqnMRGoZU5plqWuW7GnpbQUU8rF1qZQfG4TDAoeTunpqLfrylLYTAUsnaXmbk5sk6LYVmyttZnOCgZFjWKQY1FbAbAYDBgWHmMy5mdncFZwTlLt9ulled0Wx3a7Tazc9MEX/DLr/4D2f5MrkxWP3qXZt0MFzLIHDVgjAMDWgdC8AQvFGtL9JaWGaws03aWqqwJ3mMkkOWOzvQUrtPCKviqosbTneqg3YzgA2ICJoCoIfgKMosXoVpbZWlpkalWjsiZxXvNk79+ouR//Z69emJ1FVHLgV2zPOzqOa7fnbFnxjA31WKq08FlLY6fXGFtUHHs5CJrawV53qbbbjM7M02eWTQMMNYwt3sfB6+/idnd+xGXY8WCOGx3mpm9+zHW4quKoiihLAnDHt5X+HoI3hPqmrou0eBjqJBqvCtV+c6f/rWJkj3Ac572RCUoRVkyLIaURcXyyirelxACxhiMMYCs/zbGYI0BEcQYnLU4l2GtxRhLq5XT7nRot1rYLMPlGVmWN98Z1lhM5rBZjnU5xsRjRQSDoArWGsSeLu7//OM/NVHy/8jv/YJC/LPQ0QcAVcp+j1AUVMUQHwK1KojBqzIsStb6A4ZVxb0nl1hcGdAvSvpFQTks6GQ5ndwxPdWilec4a7HWMhj0sSKIgGhANRBCwNcVVgSjEIJHg8caaOcZ6j3WWv7rWz40UbIH+Kc/fV3zuh15CON3HQKhqgElBI81BqNKqEsgvhtC4TFVzbHF43RFmB5WrB07Qjks6O6eJp+exszsId+3H+l2Ie9A3kYJHD25hGYdOt0Zjp5YRE3G9dddx6mlFao6sNZbY3Z2lhPHj+PEct11B/nab3rKxMl/xMQreKIeGd3eAohgCIRa8aWg1mIcCIoSQBQUxAhVXSFO8N6RZRmunaHqEBtoty3WtXDWsmr6tHOh8krpYWZmF4PBAKuWmekcMZagSjtr0W636HQ6OGfIMhv/QSd48izbUTldLAQlhADo+oNWGX/gsnFdjEFRmlfR+j4jZSDuGEBAgzblj5XD/e/TqNRN7P17RqqyQlUxziDOUVQ1So6qUvmasq5Z7ZccPX6KlX4NtsX+a6+ilefNJ6PTbdFpWawYsrzFoF+QtwdkuYI4xHgQw9qJ47h2l6mZaUKrRVnGF1sIUbFDPSHUhLrGhyoqd6qATmxceG+tx9rqKmVVsbK2hgU6maOdZYgRrLU4a/EhxH6qEFSxxuBF1ufvrb3Hh4AxPioNIVBXFVme06pb+KzG1Y7axXLzLMe1PC4LWFeDcxgTlXFiNWgQjIkK3yRi6hrVgIaAqjbfIfa7uoZQA4GgnuADVe3BWNZ6a/T7QwZlxaDXJ1QDOk6oy5qpqZy2y5id6jI7O03mXFRQrKHoODQEyrIgBI+vNcrY5tRlhTWCj7cMpplbWQRCONM/EV651GZzx4rOQiuCNM9uY038HTwSFPUetMaWFbK6SvvIPdx1++10EHa3Z8lm96K1Y3j0GNWn7yCbmyO/4Qbcrj3I9AyVCMXakMX+Kabn9hJqZWn5BDdfdzVWC9rdDlftuQYfPG27B+ccnc5kOzEnXsGz1jVWAhrlAfI8w4SANNYgMgfqAYugUR+wiliHD2Cs0GpnTM3M0B8GhuUQkQ7tVhv1ynS3QytvMywDZR3Is4yZdk5QpUbwQanrGuMyjHWICt12G2vj+KauK2SSveUa4DSLgcaXV2NRE+LLLE4qr2Na4MYE9XE4qM0xEhU+YeM7FnT/ucblLC+x5kXKWax7VyJVWWKtwVhhaXWVve1oLQohvtBW1wYcP7lG3pnlYdftY25uN7Nze6O8NMpYGuXbGYs1ivcVogEqjxpFpY6XwwhiM3xVYfIWrt0mlD2M1giK9x7va4KvogWruR8nVsMAFk+dRDUOcGY6LYwIudtQ7kzT53Mh9n0xGBEQCD4QFKqqAhQjJt4zRCWi9jUtX+O9x1iDtYJzjizLqZ3DFQUuy3GZw2XRyicSLYPWZbgsQ41t7p3J6/t1fw2IA0wgDtobe5Kop6yGBPVoCBAUVc/aWo9Bf0CvN6A3GBCqAS3xoLCnkzPd7TI3O8v01AwzM3Pr5WKgLIf4sqAY9KjrmsFgEDeGQBk8AhiihVZVCd6D6Fm9ClcqzmTc72Gsig81xph4/s1zwWqUk60VKStkZZnDH/4HTt51D+2rD9B9+CPZe9NDyQ8exO25imptBX/qOOWhuzjxsc/S5zNoy5HPzrL7wAHmOh26+RoyMwVXHSSv+lyzp4NYhzOAZMj0brB2/R00qUy8gpfn+ZiSEBU8Zy3OGHJnsEYRoxgLxjT7jJQNoChrxDiyzNLKHFiJVrdaaU9n5Fl8qFaDmtagxGU53kOWtzEuoyhLjHMMi4L+sMRlseMbE2i1W3EkBxNrxRjpYDQWipGF9H77iVnfJWoVG3a+qHDo+rr1e3K9HFlfXD80el2aF5rFWIsYg0p85IfmoxptizqB97mzGc4EBlWBKEAHZwLGg69gddBjamoXB66/md1795DnOc7kCBYRi6rgNaBioiy1wvioYNdVhRqDMYJYg9YF9bCHyQytTgfb6bCy6DF1hYTorgq+jkreyLqirFtVJhHnoqyMyTAi0ZXqMoyz0ZVqTeyTqhBAG6XaGEMINb72OOfwoaYsy7hs3fpgpM5LfLsmy7LmmSWI9MnzPH5aLfJWjnOOOs+xziHGYqoSX+dxoLuu5E0YISoQijbda3qUFJwAACAASURBVMNarBowFkIVMKoUZUFZVtRlSfA1mYV2ZrA2o64Mvva08y7TM7NMz87RnZ5lemYXBmms04ECGFYV1mWoCC1ViqIgEHAuXuOKaKE1QrSqizTejckjlBvTzN2vewXFEvB1iVTxmWDKEi0C9WqfpU//M/fddTe7b/piDj7u8bQPHiTLLWFmFtm/Hzs9RdmeZomcqasfSmvxKP4LnyMMPCfvOcJyf4W56S7Tu3fT3rOH7v79uOlpnG0R8g6m1YWsFV3DMsGGFR4ECl67015XELRxF1Z1iZGoVIlRUI8xgjEgJioagsEHwRqHmCgmMYJTZaqd0W1lZHharTYuyzB5TtXNsdZRlgGbdVCTUTtLlrcpO1MMqiG9fo+qqgkVSMfh8oy6LDEyoZeisWDISNMzo3G0gZG1VAETlWozbk0bs8hpaB7UJm4IIT6sJei6lU5GD3EDSLR6ZHlGa3qKoQ+x3EbBG1lWosvLr7t8J4npqWkG/RXqOrrp8jxjZMcYDkva7Q7XHjzI1PQUViAzYEUxoutucycZgUAdSlQrjAFf1+BBDY1FwmOCQh1fYnnepdWdInMZVaPQ1XWFr0qqqqAuh41iFxpFbzJfclNTU4gIzjqsczhrQWzsqyJgBGMc1rg49NQ47Ig9sYX3nqyu8VWBs5ZyUBDq6KYNKHVVUQyGuMzRarVw1mGMwVc1dVnh6xjHmmUZvq6xWYbNMjKX4YHagLFhIhU874sNq/9YGEAM9wgQAupr8Eo57FOWBWAIdUlmhWA2XOXBOdp5h12zu5nevYfO7Bzt1gzqK0JRIr5ABkKn06VCqYaeTKLTQgEvhko9ilIbHXM6aBODOXmY5p25PugePcxDDUYIdcCXQ0LlkRDwgwJ/6hTl8VN84c77aO2/mc7Bm2nvu5r2vqsImaJHj3Myz+n1CvZ0W1x94/VYm6PlDeC/CLEthkWP6bsPEQY9yrUVysP30v/cITp7d9O+ah/Z7C7c7v3YmT2odRM7uBwxoVrFBkUxGBvdmnXLgXUWKwFjFGNYV/CMbWJTsKiPFqNOu9OEfwW6eYbptGi5DCdC5hTnFANIyyFi8S2HOEtQgfY0GEcVhGmZYnZqiv5gQFEPqYshrTwDA/YB/WPnFYBqfG+xETN3v1tq9AAYU8A24u9Yd2lEs74gBMQ5ht6TlSMLkOAJBIlKpATFGMHlOTPOcu+x402sEzhnyZyj22mR5xmtLMe5ybsAWZ7BIMrUWkun1cI4SxUqfFCu3ruXvOUwEpBQI0FAFI9gcRhxjZUJaqkIeIKGGCgdQlS6jSF4h6oleMX7wACHVcFZSx1qtOyhZVRUQtmjKgdoaBJsQkCD32lRXRRarRbGREuNNY0sxbLeu0UQY6PrVInWNWMRwAfFGItaS20tmBJE8HWFrz34QF3XVEWBr0pCE5OXWYdptQgIlUiUc94i+AzrA7lIDAcRg3iPjNzwE0btGwvSKBSg+Q3RauyrEgmBqqxj0k+IA47cRPu+GsWHgBcl704zM7ef3fuuYWr3Hly7gzUOXyihLpA64FxGKD3ORO+QFcEZwYtQYwBPhRJECNIonBo9F5OImlFMeQyric/9xneiijbJKKKest8nLK4xOLHC2ok1uvtvwExNk3UdpiUE48HkSG+Z/K5l5h77ZdjC46shuEAoFNeeQTot2jLNrnaXuizwZcV0q025eB/HP/sZTn7mLrpTbXbdOKB9neC6MxjX3kEpXXwmXsHzdclpQ1QB53KsgVxq3GiEbU0TixUtQiEoZVVHvy1Cu53TbmVYIHOG3Cg5gcwoFo+IwzkbH57OYrIcFUsIGUEsQRy1GtpZfAj3B0oZaqwG1ICzE/iUBWTkkh35X8dHTOvu1NHoOiBiKIsC72uqosSXBcZZyrKiVVZYa7CZRQyIMwyKEhfq5kUZ3bLGGjAGbSkhc7Qyx03XHYwB0Y2F9n5a5gSO5Ixr4rwAa4R2nq0H7re6XbIsx/sK9RUuc1DXMdPPWByKI2CIt4AjUKunVo94T1V7gkIQQGpCAC1Lgh0gKHUnw2pN8CWmLpCqJJQlVb/HcLgaXVPryTOTJ3sA51wcWDJKmIBR/Ok6jZs6uu4Er7qx77oSEOPrhDbeOrzzEGpsbTHOUNdRSfFVGa2iBPLgUa0RWtSwbpHV2uOljlm0dczUnUQTnvoNq/Dp4S8arfUhhusYgVaeU1UVgsc4qGuPGMWLQNZiZvde5vZeQ3d2L+3uDK7VRkJF8JZgFbGCGkOhSiYGsQZsTBqrm+eeUW2iU5TQhOXE+N/J7PtixxTXZsQugDEZWtaEsqQuhhS9HuXKGroyZLlfML3vWtoiaOaYmZnCBk9YW8TYVhywH/4sft8+vOuwtnwcn1kGxxcxKNN75ljprXFqeQmXOeb2Xc3cvn202zexO5tiuLRI79CdnLr9M+zu9Zi+8WbYc81OieiSMPEKXszYaW4iAZGYESjq4yjLWszo4QfrLkI1cQQu1tJud+lkFXtnCopBC7BkIeZm5AbQgDOGzMRsNWNysBkYByYjGEfAUqmjqmtQxVExKKIlAyNMqH63HoPnNWA0hhqHEAjDgHpPXRX4YUlVFtT9AYP+gH5/BeuVsqypi2G0gGSOvN0id5Z2u4XLHHkZzf1VXUbXohXECkYVk8VpCEzmsNPdGKdE05j1sK8m+aZ5kU6aDa/dxJ8aG7OSrYIGTxU8s1PTSOMeJBjqYQAJGHHYLMrCqWJipCJ1iK4tkRJDjMXzoY53lsmoasUgmMzgjTI4Ba7VIlQ14mtMXUI5wBd9tBjGjHVtMhwnULkeMTozDTFEIN7moRnzCMbE8/fB430AFcTIemyerGcax7wXGgufmujnsyLYJkt/5OoO3lNRIsK6C1AbS2JlDRa/PpAVM5lZ5iH4sXGDnjZ+08b6XJQFdVlR1TERiBAt2fgqum/F0u5M0+rO0pqeozU1Q6vdxbkcaoM3JSbrgFdqW+GJyUYYNmK1g4/WcR8HSxIUCSHGepuR/CeP06zCsnEpDIbaB0JZ4atAqKLVc7Ws2LVvL532FK1msgTKgnplGS37EAJVM92SfvZOjpRDBmsFBcqpo/exZ/csu9Z2c2rxJL1eD5M5Th29j+HSYrT+WQtGyB/xSGaqkv49hyiLT7DvkZP77IEHgYLH6YO3OG1HVSM24DqNNWc9EwCMxMwaKxaxjk63TbtlyMTS7faxKvg6kNPCmWjRM9aQufhSFI0WD3ExgDpakyyVCkYFIxZCRqZtUE9RV6godjIt9VCWGAxDBpgsR7yhqgJ1WRPKGvUFVVFQDUvKfo9qMGRlZQ1qT115ynJI5WNweWeqi1NPq53TnZom6w1o5VFBN0bIM8FoiMtlSahrXDuPc4uJbayI2mTN6YYFqVEyJk3Bs9YgCnt374F+D4hWIrE2BoMjaDCAQUPM0FTj1wc4VTPBkEBMqqg9wgBjPFaFylcgDjWGKtRYa2kZiy099arHyR5EA74Y4Ptr1MMh9aBHXQ0ad82EK3fj0/uIQFBCE/sZw0njMqqEuqaqfeyetV+fMgiJccFKVPBGcWUja5BqDEUQwKiN25ps9LquYyC/KkGj9Q6UXOIgNSqTZhL1u9MseOOM8vbrul6PUdTgCb5ejwmNYRwG15pmetc+pnbtpTMzS97tkmdtZCwpywSP1jU2VDgJVL5Cq4pQ1BgP4gMSCgSPEAe5Xpt5ILFgJ7P/j/rpOo3bZNT/RUwMBHE55bCis2cPmXG02y0sgi+GQIn6grUTpzh1/ATkXWpjGRw7xe1HDrPcG7Cy1qc3HJC1LK08o9tqM9OZ4sA1e2mTceizn2Glv0SoYfeuvTzssV9C2H8NWteUJw+z8omPw7fsjIwuBROv4K2PYMfcHVYMzkLmLNZupM8DcbShilhDp90lyyxOooWu6mVcu7/P4n3XEpSYCUecQsKIYBoXoV1P2BCsiQ+D6P31EBRvBZNn1D5DQ4n1fiKzOAF8FSjqGhFHXdW0XE4VLNWwjta7ckg1HDBcWaUoSwa9HsPGkucRimJIURZYa2gVQ9rWkhcZw9KTWaGduSZr0FIbxYpircE5i2sUN7EmXqtGwRtZ79atR813vtPCusDUVUU7y2i3WpSDHplzURkwOWIdlRcya2PcIr7pwzGuyNuMEMCp0LIWW/apBn2gQm3ANNOmqBWC+iZWL2AQKIeEeojptDChRssBoVjDlzHJQkO9nkoAE+kdBzY5njXgQ5wWQps5v0JVUVU1xbCkKmp8UKRRmKWZQsMi2MZlG4+vm3nxfHSni+AyR6fTpt3OcDbG3WnjEzM+umCFCkExQxq3sd1wE06ghmdG5yWsu8hHv4FoRatrsIHax5kVVCTGNwqIaTM1O8dUt0sry8mNIVOgLuKxZR8tevjBgDAcUpcldVmgdYlWHq38eqq+eBMjh9VTaRUH9UCWm4mUPbA+vdKmtfg6zoHp6+hGKYYFNmsRCIgRynqI84KGmn5RUhY1i70+9x4/Dq0Zeu2c+1Z6HDpylNV+D2cM3ekufQMri6totYjUNQeP7+YrH/ev2HfNQfboNSyeOM7i8fv48PsWmd5/FQ+54WZm91/H6pG7dkI8l4zJV/CsYyxkH1UoqgLbzrAuBtybxt2hcZANxpC5DOccTmIQupMAZc5NN5+EQZfllV3Nw9cgGnDCerKGtdH0bozBiCE0wWGqECxkweDVkjmL9w4pFV8XOyili0dRxxeXXxsgnQ5rvSFlWeEr37ysSsreGoN+n0F/wOraMsOi4NTKKt4Iw6KgqobkLqNaGzA31cKuQafdo+Mcncwx1e3QabfIMocTJXOGLHNkjfImEpVxa0yT9NFk0Wk43Zo3YXiv5EagruMcbFajAuAy6gC2DmgWCFpTK9EdWAWMBmo1WHH0B0NWBwM6tkZ8iUogGPDW4FsCmUcyAeMw1oI02Ymhord0LE40rhXqh4S6oKqLqFCOxaaZCXVTKWY9izMExdcxUchXNeVwyOraGqrQ6kzT2TVLu93F5W3I4wTqmQhq40BRkTghb1FS9vuUgz6Dfr+ZTHnA0mKPLLfMznaYne6QZXE6FT8aPAbAK1KxnjijWk/sROB51gzX1jM44/coYCdQgcniM5+aQI2RJm5SDa12lzzPAQ/1AKkGMabae0yoMXURQw7KAq1ihrivK+pigC+H+LqirCqGdUlfA2t1xUpdslpVlD7grGHKtsgnNMkisHXilDWgCC5rUQ/WwAbUWEKoYux1VbFaeYK06deWoQ2cEktv7z4Gqty7ssrn7ruXtWLAwRuu47prrmFuZgYVOHrkOMtLSyytLHKk1+dDn7qDb/mmp3HdNVdz/bDHFw7dySc//jGOfeqT+H6fa+Z2kw3KSyyZS8vEK3jDYdFk0LI+tx0hNH8bFkfK1hmwcZY8JU4M6jKHNYozAYc2o2jD5z55Fb7MGYVaGBl9FCvxb4aiJWQjc9eIoQaEgBUhs4IEyDMh1I2LpJ68hyzA2mqP4bBCshYMakLtGQyL9fi6qioo+isUZUm/qFhdXWG1t8rK6hKVGNaGQ6r+Cu0so6g8p8TTnZ6m027T7XSZnZqhBAZlTTt3tDNHK3f4oDE12QYoK1y/Hye1bHI+Ygy7rv+eRAUvy9vx3yR8M9GqdTgrSLtNEIv3gVCV1DUxK9PHpKN2ntFqtTHGUnuLVDDorVIVq1FxsBnBQuUVWhUSKqx2cWYa7+s4B5l6pOxj1INWiNaEUFEHDyZmphtjmtj/yez7OpoKKDTxcSqESin6BYNBQXd6F93ZWVrd6XitshxMVMqss7gmIcbmGZnLY+KQRGWv7A8Y9nr011bpLS+zurjI6soaJ06sUAwK9uydpdNpR+USjckbGjC+pqoNVIagAZD7G1omgGCagf36tEujJLr4LWoxLscUQxxKKBTUxxAaE+fAUyUqbvRRt4yUwzhLU1BCMaAe9gjDkrooqKsBvhxQl0OqumBYVwxCoBeUXlFxfHWNYytrrHqPZJbZqS5ZRxAzmf9gZN1IwT7dRqnGYTvRw1UXDlvmlINlpKzoLa1RrA3xnSl23fwQrjpwE9Jqc5UqR06d5BN3fJbe4hoHbryJq67Zz9TUFCEI7U6Xbt4i7+5mdnmZzrH7OHHfEY6cOsk9x45xw0MfzuzcbqZ27Wb/1VfzyQ//PXfc8SlWpqZ5xCMetSPyuVRMvIK37ooLoATEGKY7HfLckmcx2WIUmiXNvF7WxklI16dOibmxqFgWV2bJxGCNgNHGlRKtdK6Zzd+YLM6dNxqdyShGZpQaHj8WortY+uRZdYZzuHIpBwVV5am9EIbxf0pLFcqyxpcFw94Kg8Eaa4M+/V6P1aJgaWWRxcWjLPV6tDsz7Ork7N01x5oXHnLtVSyfOsby4jKry6c4bh3XHryR2ekZpkpH2crplJZOtxWHi85gqphR6qw9fT4sNvS6CdTv4nQNxhB8wNpoYTOiGGvj/Fy+xpdgabI4g5K1YiILRgnWks/MxD6uJWI96ivqqqauohXW+gqoEWswvhOzbI02MWQBDJgsx2QtbFGSGfBG1rN7gYmcgxDieY3muyOASpx2Q3LHrpl9tDtTUTZOqOqCfjGkKkuKXp+Ty0sxASwErEC726U9NYXLc+b27mN6Zjft7gzTs7P0ZqaZmpmie2qJ5VOLDAd9Tp5aYc9eodttkZE103EIAah9QKo6ZjKPuS0nCdOe2fiHmuYfPNYVPsBKRrAFmXVA/Ds46hpjYxx1jNELiA6pzACKktncxolxA4SqxFdD6mFBXRR4X1HXFVVVUlY1wzqwXNQcWxvy8aOnuPvoSfp1zcxMh2t3TYPNsC7D5pMWGBIxWYv11NkGASS3uLzFwBfgJP7RQDGkWOyztNjHdXex+yGPYf8XfTGdvfux1rJ46hT3ffxj9Ff77N93Fa12Rt3rc/vtd3BiZZW8lbNvzy4OXHc97ekOe8Mulk8epS4Md37+EI97/BOYmZrDtFrst4avaOXs6rb46D98iI//44d2SkSXhIlX8Oom2HZkqXcSg/KDr0Ec1sSpPFRoMmBNtOBZh2kUNGmC+MWYmPUkTbamNRhn4gNcmr/eav5yiOYT5/uJ7l9BkLpe/5gQMFLQdlM88kv/fueEdBGpFPrDIYNyDa9CLYZgckrv6a2uMuitsLa6RK8YMigKllZX+MLRw6z1V5mbnYU8p9Pt0HKGpdozPTPNnj17OHn0MIsnFzl2/Bif+Od/5GGPugWd2wPW0GrnDMsaNSVYQaSNrTyE0STHsW0bM1FM5n9yGqDb7TKsa+p+japHmvjTIIqvKwKsW7dt9J9QFBUSDNKxBNfC2JKCDJUcJQalV2WFp0AIZLlDfPy/WUTBxSmDRlNzYISs08UNh2Su2pgZZcNzNpGYRp7QiAKh1WrT6XSopaaqh9SDPqvDIb1+wcragFDXTZKFYKyjqGJ2rTOncBba7ZyZ+47QmZ5jz75r2LV7F3uvvob2VJc8b5O32pw6fpyy6NHr9clyh8tiH4/advRihOY/WicVyTvASLEzzZQkzfx/YshdwPkKX/TQoGRVQGwJocJrwPuY4DVYG7LaH2Bszu6pWXZ3O8y2DFYryqIkVCWhrKl9TR0ChVcGpXKqX3PoRI9PHT7KocVlxCqtdkY3bzMzPcO+PXuYm5thaqq706K6KNh257Tl9cGcCFqC1QBVwfDUUXr3nmBtucTO7mPvQx7Bnoc+nO7efdg8/tPT6soyn/vc5zh030mmp9vU/SWO3HOUT99xN6WPfjfr4JZbbuG6m29mbnYXVizOOfqDAcOyxGYOkzlCKGlPTXPzjQ9hsLTI8vLypRfOJWTiFbzh2Bw8ooLBxCk1jAe1cWLjZg486xwuc417tSmgScwwo/mNTBNsZ6X5E3eDsRZnsxiDZOI+0fkxyhJssuO8J/gKvI9TToSACRZjj3Hs8zM7JqOLybA3YDAs6JWBohhgWl2KYo1BOWBYlCwvLVKUQ5bXVji+ssjJ5UVC8Fx7/U3s2Xs19544SZV38HP7cWYNpmbZNdvlmquv4tixY0zf8RlW+kPuvP2TmMd8Ebg97N99NYMTS1BUOCPxb8psRo40iRWsJ1iM/oQ8TKAVyZc1tQ/RFaoa5+IyFosioaCqh5QayF2OGKUMUIcSgtLSnHyqRS05VeUJdcDUJaGuqGtP0DiBq1WhFSwaFC8h/teysTFRwJj4W5Ws3cFlGbYwuCataeM/iCdTw7Ni1tO3TPNPFhoCRVFQDvss90uWVgYxTlUD1uXkU7N0ux2yLOPEyVMMilUQJQBVHRiuDVgZePLlAYNen6UTU/z/7L1pj61Zduf1W3t4hjPEeMe8mTcrswa7Btvl2eAGS1bTgOBFyzQtMUgt9cfhs/ACoQYagRASCKltyRZ2ldvuqnJV5Zx3iBsRJ845z7D3XrxYzzkRmZVZpqWqSnzMSkXGHHFjP9N/r/UfHj99g6PTE4KLhFgRY+TVi2eM44a+GwixxgcbWVoMnZv4gV/o8vxcy9cnk0LYzsHbt+21c5DHjv7GE4aeakz4HEj9hjx0pJRYb9Y8f/mKZxeXXHUjV1eJ0+WCr739hLPFjCZbjNm2G0mlZzMmPnq15sOXK95/+YqXmy2jKPMmcLKccXy85N7ZMY/Ojjk/PWY5n9HWBz6i/VRlzYzdirx6RnrxMZfff4dn77+iPn3Aw0evc/LmWyzu38fXDSIB8kg/DPRdz83lK9IqUzYvuHyxoh8Su5M4q/Liow85OTslJaw7Ok1sXPDgjTolTUupGpwLvP7GG/TDYXLfd3XwAM/ducnu7mfRQx2NG4eATsa4LjpD+hPBTik48Xuz0ltPFSjTf1kz4Cf/IwN8Ejzm92YJDEZoVvI4mHKrFKRktCSkZPqu4p1XX/0ilufnXv2Y6IbE9eqaBJQhMxZLQ1htVry6uabrN3z04mM+uF6xPD7jbLnk+PycTdexGUduujXf+2ALpfCvf1Sx/PXfJETH4viMb/7O7/Dq+Uccnx7z19//a3j7LU7PjnBi0VspF/phRNiSx7wHdT/hjfVFLdDPs1To+0SWQiXmri/Ta6+JVAZGFbYqVDHShIDmhI4jfV4x846qCozrC3RzzdCt6YYRjY64bFie3EOGAXB4xHhmUydbxSE+IN665KWMuGhRWl4NdDvnJ07kQa4+fgK5KOQ8MqaesRtZrzdcrG5YbUeyCrP5jHaxpFksidUM5xzPnz9jM3Scni3xogTnSEUpuVgahgQurte8vLggjRsePXmNk/uPWJ4fgVgc43YdyKmnpJGsAaceV/TWc3zC1YcIr+PR6cS5uTOi3Y3Lp3PVeUfIibrvLHIMZYDJU1BJKbPuR666kfevev76xYbN+5ecvHdJ6XvaylFUGMfEMgYchdVmSxHBO8dsVvNo0XA8q1nMW46PlpydHnO6nLOcz6nr2vjfh1juc6CFKp6R7uoZz//yBzz73ge4+X2q0/u0Z/dpFsdIrGxzOCWuxBA4PV5y3ASGviN6z6wJOC+kYuKwynvO799neXLK+cPXKClxcxU5f/CQqm4QmYzwfY3ULZHAyeKYL3/lK7/YdfkF18EDvLaOd3hWShMDTeUJPhO82aT4CeAZ98780lQnl36s8wegqVCcoj6QccZnErMv8CoUF3AuoM4AnhOhlMw4jqgm8jiaerMoZUoQ0DwiuUXC9otZoJ9zbYeBIQ0MKTHkxKDO7FC2azbdmtnJjL/+0+9yuVqxOH/MECoGzayuL9hsVjgROjxlSNw/O+Wji5f8yf/9Z9y/f58vP3nEkwf3yWmgjImU3+Ldd3/M6ckZ987OkaYhK/iiJAVS3vtg3b7Y+4fYREp77p1QT8IfPym7vY/0qJlE+8r86oYtmjP9mOi7NesP3qPrelzOxCLUVeT87JTHj8/xjeCCktdbqqx4dRYlFwpIMLqC86gLtrlRZyKPGEjJ4gKNe3eYKk4wysfOxDgls0O5Xq1Z3fT0yTOrA828ZX56Sr04glDhnefy4hXPXzzn9OyYeVsxbm+oYoWrlzx+48vMFsfktOW7f/onbNfKj9/7mDFZ5u/5g0ccn5ygOQOFscuknIkmGCdpATV+pByqNxPAbG6v5ZPX+f5CLwWXC95Fsos4XzF2a1JWCnbupqzcbJW/+WDNxXYkFsdxCJzOT7gqa358eUkAjtuW3nnQjmoeaevASV2zrCuWdU3b1syXc46OlhwfHdEER1NbTKI/UAPUUrrbd3a2VEBR68Btnl9x8vYvEc9fJ6WKeHZMu5gbFUrV/AW93ZdnTcvZ0RGPHt3n6tIhg1le3HSJV+uRKgaePH7It3/929x78gazxTGqyuXFnIf3z1nMF9M/ZMe/tGdzOz8hzA9zcrargwd4QrkzBhKiw/zTqmg7uGDxVbgpssxNWZE6eU+JmC1EgTJm49d5i7SRyRiz9g5VR1FHEodTBU3kDHlMaLbsw5LSZDZqMnItGUpGGPDpMPM4V+sbVtueLmfUB8aupx87VtsVF5evaHXJi4sLzh88QpqWzXbNxeBZlw3ZB/ricUcP2LQ1pV4SQ8f764HV9j2O6sD58ZKTh69Rty0nT57QHC35/l//a+a/+mu0dYVrZnfsb/zkD7bjS35qd39g1Y03eO9wLtDWGefKFIIue8V3tx3oupeE6LnejLy42vLypqPvM3WsWLQt56dnzELg8WunnDSBpgoIiZQSMVTI2FmweilIns5jgaSZJraTV1tD1bSEfoOUAee9mYl/wsbisCpW0Yx0kwHZXMDFyPGRJ1YtQ1KkrmjnS0I7x4UARbm6usI7YTlvGYcNdRVZrzdcvezpS+SXv/Et5sslrz19g6997Rv8L//j/8DNJvH8o+c4hdN7D5ktl3TdmuCVoeuQImS1+5ooaJFPc+APqsTF/TW9P8d2ghLBpijZDL+dt85yKYW6bmwSkzPOR1arFc3ymG//0hN86Xh1dcW6L8znJ4RZpKkDy6bCN091iAAAIABJREFU5y3zULNoAvOmIgZHEE9bzVgsFszncxbzOW3b4ig0dQUcZg4wmPXMvu5ibC1InFEf30P8nNnpGX1XGFPCjz2UYRI9WldGRJjP5xwtFyyaGk7PqOSIs3uJ88cd3ZiYL485v3+f80evcXL/Pi7WxFhzNG/5yltv0rTN/viqZhh60IRr5tTuMKcHuzp4gJfTdKJNqr1QR+rgCJNi1sjM1gr23kQWqn6y0FCSKokpfQaj2GVVC0pPSgiKSESk4AJkc0VgGAbylNZAMad6ckZTmTaVhVISeUykrJ/rvP53vUxRlkk4VteXlKKsuy3XqxUfvXzOxbvv2gOpnfNqsybqSFcGXt5sqB9/ieVv/0M+KIGbUamqivtnLfO8JfZb5LThg48/xCXlzbfform54enTL9F3Gy6uX3HvtQeEWpAiqPNo1dx5oukUl5UnP7zDW/+r1QuWR2fM6jktSuXMW22n3m7qmm47slp33FwMrAfYlkiRGdUCjprA/cWMs2Vk5oVlVELuGbcDLgi+iTjNFLHwLedMPSFaUIzML866ISFE65g7ywvZUVn3x+EAK9YVmgsqQj+OxFhN3peFgqOetWio8VVNiBEXAnkYyTkxn9U4Cg7jh15fr/n4cuCj589AC68/PsV5xTcV87MTqjgnBOXy4oKmqZid3mcxHHOTzaOtlEyYRC8WQCcHykuwcsF/BrfzFvBNZyoqgguBWNXQzvBO6a8viSGyPDrm9cf3ef34CQ+fvE63vqQgbLYjXTeQS0/llMZBdAXvsk1ociYXwblIrFva2ZzZrKVtW+N5S4GSKWkwTvYhlsTp9PrkMXAI/uQNjr7VMFw85+r9dyBlvIvQbRi7G/zQE2I0igfCfLHk4cNHvPPO31BrzWx+zGyx5LEWnBQWx8fEdoGrZ4QmEqpITpkHD+7z+LXX8N5gjmhGhy15u6E6OUKDY9yuf+FL84usgwd4Wsr+Rua8w7tMDN4MkJ3s7UxsPBvsqVPExkfTmDblgisTN0MKxTlz7AdGPzImJY2eOgcCgVhVjH2i7zaUYRrLZrVc3KymNJwAXhpHSjHV1iFW1/dcX1/RiWfIha7fcnV1wcuL51xcX3Jy/hh1kavNDVEKgzg0BNzpI+a/8e9x/Cu/wfaHH7P+zp9ylUe6m8f83m9/E+1v+Ith5BtuIK/eJb4rvP70Lerthl/7jV/lz/7Vn1BCIC6OyOstUMhps48j2gW579R14g/vUtCieIEgUHvj3kWEIOCDQ+qKWdvgqwVLAjfbjm4YySnRVkqdB07azPlRYB4qWjUTZO/9xKfDvNSCUIJxuvwUSaQl47yp5YoqmhO+TF6RCOREZops0sMEeC4EsmQkO6qqwgFp6Ag+Ij6SitC0La5p0ODBeVyEGALia5q2JQQhjwOhbnAyUlT48J0f0KQjqsUpf/kXf05d1SxOz3ny+hu8+50/5mZ1TTub084WbFcrQhwpwj73VKc85kMuSzD6/PbYzkJFS7GGnhZyKdaVjg0xVnhf8xV1yPI+R/cecXNVM243+IUiJaGSCAHK2JPTwDBmigXSMoyJdr6gni+JsaGuK2IdjQKYB4vico40HuZ9f0fw/EmIrVC3+HtPaE7u0Tx4nXG9Yry8IqeMH9aMH7+DnD8knNwDX+ND5I23vsz7777Dux+8T1PPifUcgqOpatrjY3xV0zYtoyaGfksVa15/7TGLdmZirpIpw5ZxdWk59Is5pejBA6BD//uoYrWXWDgRYrCunQ8O5/xeYeWCEcIRoSjk7A2UiZJEcWScFgjGXxLvzU7FBULV4Lyn7we64crk+E4Z85bUD4z9iGZwxbCd9w4ko5TJ3b6YJ9UBVi6FISf61LNNmU2/4epmxavVFc38mEdn5zy/vmINNE6oqoZHX/9Nhkdfxj18wlgiLisX/8d/y/rDH5D/wT/m6A9/j7OnD/nwez/kS7/7H/HBH/9vfPe7f8b1ZsuTR4+pQs3p2SnP3nmHo68f4ZuWcX1NiPWk8nS7JFRKShb0Ph6eo3lV1ThxBC80VdibcAdnXDz1jqat8RpogieEnrFLRCfMYkXt5iznMxbzBSEJMWOq8+jwQaAk+snrrSgI3vKGnbN8zyExui0xzBiyMtmu4YCSxlvl8oF2knzwFFXjQUqaPDYjgokvvA9U3tHOWooPDGMhO2hnM7rtmlgvUPHU8zn3cqHyShkHjpYN7WIBHtKwphTh0eMnnN57wOaNL7N673tsb66ZHZ8zmy+so6SJfbLwoc4FP1E/xd1vsukpas+EnDOCEOsWwbh54oRYNTgf6aWiihVnD16jW6/oby6RcWtJDKNlzEbvLAIQj3eRIo7QtMS6IYSKWFXEGGxEmBRNo31fPEwfPNm3QD5ZxewkJh/ZltQKoW3xRyeU7YAMGUjk9TW5ZMJsiQsVp+fn/PI3v0XKatnkoWbUQtPMqGJNkUDqLWvZZ+Htt7/EvbNzHIKkjKaevN3icFA3aB4tCnjPzzvMOniAd1eh552jqSvilE0qbkqa2AksZOeTVijq0GKBK+JMxRaDEOtAqAO+qqhnC6qqoa1bQtMgsaZo4OrikstXr9isrpBSyGMhZ8WrdTBCCIgULNap3GaiHmCNfUdKIzkp49hxvbrm8maFItw7u8/Lqwtc2tAnWGXhjS+/zX/2X//n/IvvvqB99ITv/eg5vq7pri/Ifc+LP/6feecH/4Rv/Ie/x/XHLfOTit//T/4h/+LiBT/+8fd5/d4Zrp5xfH6PbttT1TXeCXkIbLc3kEc0J0se8YKPgRgr6tnh2RW0zdxEP8UsSXY2PiJq3FOpKQhRanx0zKOgQ0PjPU2s8EznPZbVKS7YeR6EohMFoWAJDQpOzbvNOnJmdmy9bsGFGvER7wxoUqzDeKgKWgDvw9QtK5QYJ0ueScyppiLOw8Dq4hWzo2PaqmYsjocPH/KjH/2AIfXTmHfk+HhJdIk6LM3yZ7YklcJ2TJw9eMKjx68RY+Ts0Wusnn3A9ZVxx+ZHx4x9z5C6CXhP2cvA4TLwbCMNfMLJ/G5ijZZCSYmUrItc1RVJsrku5EzSQvSBRVMjvTkhhLrFeQu172+ucRS0jKiY6U8TWwSLuVQfiG1rCs+6gjCjCsLYbcw6KASzxvH+i1ien3uJfB5wLfZc7TZGm/EF3WxxLuLqGg0ZJzMDYgV0dUlyNcSGp1/6EvN2znvvvMt6dc1CAk1yUJStK6RhRGLF22++xaP7D/Fi9k2MPWm7RkqxCV1dIyWSc4FwmOu/q4MHeMGH/YUdQ6CuKmLcZdC6KbHClIUiNprde6Uh5CmvNIklUYQiUCZe1/Rww0VcbGmOjixWqHE0M8+L9xPPPvyI9boHV+FVCWJdRefAe8EVOWiAV8goyphHskI/9AzdmrP7T6hDpNuMbMeRkpU/+I//U/7xf/lHPHrykP/+L5/xwarj2994mz//q3cpvgJxpH7D9Yc/ZDn7A6oYGfqO8wfH/KN/+kd853/6l1xefMS5f8DJ8TEvxxe8unjB47eest5cM2vn+KqmiKdIICNkHCMcpBeb9xEtxgctdQDvKM4ERzohDS2FelbjnaP1ARdqAlCFgFPF5UTQyQLIByRGMoVcBCQg3pNzxsWIj9bB2OGGHacVZz6ELtR4H/E+mIp8AoKHCjS8d+Rik4LgPVpVOCeUUnBqfoDiPMOQWL98TtU2NPMl95ZHbO495OOPP6Sd1yyaSEkQ6mMg4WIgxBl9N3J8csY3fuXXWSzmNl6sAnF2xPW7H3O6vWF+9oDYVJYx7AIlT10U4GBbp2Cc55/2d075084J6j1aPPiAKlSVR8dxAl+Oto0MgI8eYosLgXq2RIdzUurJWib7H+xU1oIET6wqyIlhGGhnLVKS0RpG29TblPgwz335nBG5TKbToyrarQlVhdYn5M2VOdp49vennBNOwaul5uA9Dx7c4+z0lNW/eYfw7nPax494dnmJW3qOX3vE8vSU5XJpPOChM4rW0KEloUVtUjdtK6MK+UDXf1cHD/DqumbXx6u90AQIcQJ13iNh6uQ5t1fU7ywcUja/OrxjdIorWEyTCM4pecgkr6TiiDSImwOFKnqqWFgeR0paIAKvLnu6MRO9ZxwV7yEER5QIqgdptAt2oTtRtCTGoWfsN9y794j5bMHV1XOqIFz1ni9/8+v8V//sn/DmkwdsspJU+M2np4zUbLIg9dwUryrkq1ccV8K3n55zfb0mcMzJ2TEPf/W3+fj/+pfcvPiI5fkD6toZX9IFYtMy+jkpVngRgoPWCUGEGDzhAO0KhiERvGcEhmTdaSb1rBRw2DWAKl48lfcWy1cyXsWMuBVKyuCFUNVksczanBKSMiDgHRIqXFUjLkwdjcno2DsLUAgBHytcqMwjy3u8fDrS4rDK7UbhweO1Qp0gouQxUcZCQhG1qMNKPa5PdMNLpKp4cnrM+axlc70ieljeP6NZLnn24Ydc3Vxxsx249/B13vrqNzg+OaWUjEOJ3rFcLLjEc315zezsniWNSEWZNpNuxz8FDtaF8Cd4nbu/+c5rJoFniIASxJGHAUQJ3oyhS1ETCFUtWRRCNBPvorhmhs8ZFRPOhWKiraQJH4JtZlD8MBjgKSYuGrLRE8rE+TvEEv8517ROlmTOzrw0jGaAPj8lbVe41MGkwN1dO3kY8BIsRq4M+KLcO1pyU1+wmTnOz7/EG8cncDQDTACZN1tIA8O2s2d/0xoH33s74xUzPj7A+/7dOniAN/T9lEwh5vLurC2+A3XWmZjenxSAUowkXlKm5ISTSMlCKko/WkdPMVsUlQqJBanBDRGfEmk9MHY2mhRRZm0kZ+XyasumH6ij4rISklCCTqr9w+TgVT5Qx4oQRyoqzu89IobIuLniqI287Hp++e23+S/++T/nzcf3WTghAL/05a+ynLVcbqGqAuwUsALBZYITboaRtvL0uVAHR7Vcss6B1G9J5WOznwhQSqJaLDmeL1k0LaGqcFUEF1An5FwYD1DkMqsbS2hx2N838eVkUpSrGqEf1HbKRZCs+CKm4Nx5NpaMCxGcogw4KYB5Cqo4sgQIEQnRDL+LEELEi4eSLd4M2z2HWOFDZNPpZLx8yD54Ee8nGshkqk6ZDOny5EM3CVc0O0qydchdh3MDMRfOBBbHC04ePwAcJ0/fIiG4uubB608JTUs3bBm3G7abNTfPX7J++Yyj2czyt8eeqvKgkXG0CDTgkwbThzg9KHn6s27/Njvndf8aLZScKEVN9JIVvCKa8E4mFbgDHLEKZB8wd0FQZ1ZDwTuct3uMJ0zm+LU9UwiIQIw1OY14bz9zs14zq6t9t+8w67OvaRXjP+MDrpqjmEAFIDRzyDVl2BpdSq3BEutIKaBjh+Lw4umXns3jBUdt5MXlR8jcUa17o0GgiA+oCnXTgiiq5gnK3jLLYkTVHea9Z1cHD/CkZFPxAdFXxqNzMoE8d/uCce+05Ok1lDyy2fRsN4nT49pC1wtonuLH8mAGxtKjvkOrDicDw2pjL91IzoowUtfK8rjm6qpnfdMRK0cdAuiA7FIADrDaINQh0MRAzoXYNgxjz2w+o64dbzaB3/ndP+Th2RmNE5xA7eD145q1CG0lnC/mhLZlnEYq1dEZ81r5zg+e84++9ToX1z1HbUUEzp+8ybPv/AkyOh4+foovo92MCXQ5IeLI/Ujuhml0GHChwoX6i16qn3mFGKYxbGIsypgVzZaGgLNEBO/M2NsBYULQHsXrpECfeGMhBlz0oImiBS+OlEecC4QQcSGSRchpJFBw2fKevTrw4x5QiHP44C3OT3dqx8OsECLgyHlkb9Gh05hKnSns1fiRsQrs3HokFxiymb3mjrx2DK8uaOY2frq52RBEWb/4iDItXxoHXM4sI7gHZwxHLS4oMnURdwrau5yvQ6WFAOYxaoTqvchOd6B26mCWbObQFm7hcaFCxNJ2HGZSX3ZjXNUpctxsZlQnRWgu1plzUwycD0gwNwanDqeFkgCf0DHjRJg1LVIyomqK8kOsO6PPvQUeJn0RnPmRRm+8uJIQLeY96yv8okZTj+SEFHObAIixJWsmpZHrbsXJwzOiD5y+dp912lLHiA+eXAo+BLPBUbVuYckMw5aqijiZ7j8Um2AccB08wFvM2v0ebt5EqiCEqVXLdMHKztpBC0UzitlIePMh5upyw2pzw/lRzdwrcx8paWSMgWa7IW6uyDcv6V5+CE7JZUvZXJt0PmdSTraDCBBb6AZh241oVHL4bDn5odTD8zOKeFSE6AfWmxVNdDx98hopen7561+ndjPqWO+dGzzwlVPhe+tMCBb5c//ha7z3V8ry0VPOvvFbXG0yOSuPzxeUlBmSEpqGOja0i4d0Vxe8/NEPePD6I1AY+w1ZF8yWDwgKbhoz5mFg7Lfk8RK+8uiLXKqfefnKUlzGMdMNiTEbr3T34BMRwmTy7dRCu3ddbCavRxBCFcHZaMl7G2toFjLBQKIX8GHiq9o1lAsoGS9mCO4UfM5kjOccfGAY8jTOPcxyscbLOKE66954HCIeJ56k00hQFdWEYmjNo3ZD8IJmT+gVvdgwrIUyFrOEmC0m5b9QSmKMplyezVsWp0cMfcd2e812vSJl89ITkZ8A1IeK8cqYJgD7yS7lLag13nMpGZcTMnXsCpBzJkwbHPEGjnVS3DrxqMfWUw1Ap5JxLuBig3pv3TtxuJyhJJCMiHkP7gDe2G0RieTuMLNQP4tb+MknnaWp+KoilwxpnHiJFj8WqjmiiuYRN3kGaslIqFmPipsd4SqPpJ5ZVaM5sup7jqoZToE04mMkq/nVghKdn/h3GdU7tKwDroMHeKhdQIIj+tpUtDFMPkg7gcUUVTW19J1zlGw7tqr2NE3k+ibx6mJNHZWz1mwnqugtUFwuUQEp3kYgkvHTCEC8IL6mqj0uFEgDdRTK4Oj6kZSm3y2H2ck4PT4m1jMWmy0fvXhO9IXlbM6Dx4959OQRp2dHXP/wx+TrK3I5Mg8NEd46Cnz35ZrHD2run7a8/Yd/RHSOxdd/l/nZPf78/TXBBx4fN0jJDMlBPyAlU7en9OvMervBN3OcFwg1/dVLXr33fRuLieKDxwdPXVWE2eF18G42G9vBeod3gTFDUQN4+04StuROrCvtHMbRKsXymFXxMdyONESIsaJPAy4G84l0WMdPC+SRlBNFlOjUxAGAuGDWE94TxOMkkHNPQg8WZPgYJlqBgEsgHpFCstkQQsC5SB570mjnrhdHKRnN1kVyOLLCWAoMPeVqhcxH1iUT6goVs3IigIgnO0dW89Ucus4smnQyeQ8O792eg7SnpB3g+u8MhO92Ke+COwBNNopN3YAyQN3inGWJ+5xtk+PA+YgQoXhcFSil4L2Spu/3PuJjQENAfNwHN+h+kdmL90q2zZNm60SVA6SGwE+T7+jewEbFI97hK4v7ZNrsuZIoKeOrFhdbSh6RMELJpJIRJ5ydn9nIuyRKzjQIF1eXDCVS7Tu02Ah8r+USExmJcQSdCPlAE6R2dfAALzrj/4g4Zo3DBQ+hRhwUcfbgmcjeghEv8UIaE+KgqoXlkaMUx7PseLkZeHmTEArL+YziItuUWG02jMVUo+28YUbNrGl4/PgBPo3kF1dE7WiCGgG3EjLCdjAu06FyPU/OTnE3Heo8KQ+088j9x4/JY0F9xAXP8nTJ6m/+kuO3HzFbRhzCvcbhxfHuZeLp/YrL3/4thjjj2kUeP5xz5BNHD46ZVQ7FM24z2vfgAzFGggyEsaNbdSy+fsr6R+/SLI94+PQJoaqQKVHhNkbu8HzwsuZpIwNjymyHREEoKH6K8AMQtZtrLgWYjKBLxkdvY9QQwUU8gtdIShbUrkEZtaDiLYklF3Lf4UqmiR5RJcSItcqtUyjiqEIkxoiIZ8zpcDt43k/NUrGIQykgGSeeHBLJjSbC9wIhUMZkY0MptlFURYOQHEhJ5sFZevxYyKWDPpCdt5Ggt45hwTqt/XZLv+33gdo+TGkiwU8CUt2/HCLCK7sRrR2BPR9v55CwRwApWaJQGgni0GkjM6RkzwLASbDRue4cD8rEbb11YsC7qbOXLPVIoTg3/Tv0lnspwjCO1i0sxaw6DrA+d/x/NxZyEjROlhJQlPWm4/LlC3yIvLj4Ecf37pFSoq4qLl48I5fEt771DZwmu6Ziha8dQeB+s+Di4hWnR0t88NO9vSeljJOpg+1rVCw7p+hPA6KHUQcP8IIz6Oado63iXjEpAOIsxYBsoykRnDNuhfeQkwdRXBCqGpquh1lN1hmr1RYdI818BtGRS0VJhaZSTk5nFGoePH2Txekpqd8gV1es3n+f1dUVMQzUbY1zFeqSjXDzYZ5q7WJGzgXXVFSziqt+S11X9POK46MZSQNusSS8vOTlX/2A+a/9Ek0tVE742nHg//xghVvOeVU1yNExq1dXXI0BXj7j33/7AU6gB4oTRsywePX8A9p2QXf9jMsXL3F1S6gr8tDTX71imwdKTrb7m9IUVJV7X/7SF71cP9NyspOEK7ko2z5NUXq6/7x4j+idiCUtdl0EQb09vIp3OPEEZx3rIdsu2szGCs5bNFDqe3Lf4TVTuRqccV0kRKjnBG//njD2BvCCpxyqkz+Y16aFM+G0WGyimC2HJI/3wThD/Yj6jPgRxsHWbPKroyjivT2QigFvHRQNDi+KC3aMVBxjEVQTYzfQbzs0mxcfwe2zhw/VluPTldNtBJgge0uS3RqICEVsucdhQDQhQ08JHqPpF8Yx4XyNOAXJNmq1EQPee7IU8xacWna7LngeRsYxEZvWRrSa0TEhRW1M6z05G9Bxh7qzd58DLeQWd8v0fnEOiZGSEuIc7dEJFxcvyZoZ+47gA21dsVhYrm8IESfGGN4158CEXfOjM7qSefb+BR8/u0TTwPnZGV23om1nvHp1jWomhGCJVeL4td/+B7+QJfki6uABXlWZDYn3nuCUOKlldWqk706yHZSXiSsTJ1GAz2bGWDXCcvAch4qbLjOuYeYVGbeEUDE/WjAo1IuWeQ1hNud42aC552boqWJDWD6gS8JmfU3YDEjo8cEMlv2B3nhnp8dIrIhjppq1VNsbmqMls7Mzhu0N2QVKnlGdeMZnz3j5/j0evnmP4IRvP6z43390xfrFFU+bGd+7fsHmL/+C7yzgP/j6U772eEFG6AZlu9qyvbhAh0x3dcV6c8PNix/x+mmLqNDOTxm6DevNq6mj4hEXcX5GiDUhHt6IdsezQ8GHYDy8AkkhTjtp7zxkExeBjflcMNqC0Q3Mqw0R48RQqKtIyj2Zso8rS2Ni6IwY7RyIdyRV+lwILlItzwgxkLNSjQNh2BKiJxI4UIcgG/cJ+6BzAxqm2vbOOhbBRUpI5JyNE5qsi6c5o8NAyUb0l+BI3lPVFa5pbfzrzZtQxVHU7ls5ZYahR8mEKhBisJGjTve9fSfL/o2HyoBsKjPa3VFv9mPpiZenpdjapmR+a2Vkvd7QLJYUL6DGH6Vk0rhlUBPn4Rw+2GMz+LDjN6BT5GEIjqLQD+nW51GNQ+ZQshYU400e6C3f6m/74+582jmhxArvHDNV2iZydjI30/SJRiICZ+dHdl/aCyMmkonalCLlzMfPXyAEVjcD6jy+bhhywYWKm03HcnFM8I7Nes0Hz5/z+MmTn8Mf//+dOniAN2sqVJUQvBG+ZYpKgqldvId5AGZ4HBwk8AFCwSJTgnJ01KDiyVqYN45Z6/HSkLHYoePa/NYkd7j1SC4vGcee8foaRXBjYhYhzhtKymy6gWGzNZJ6M/+ilujnWvViRlGlrLe4MCPOZ8zv30NcQaoIMTCs1lytt9QxsH73Ay6Xc07vtSyD45/+yjn/zf/6b2ytr14yCzAbe379Kw9oa6HLhW7dM3zwHsN7P6CsB3xZk4aXtMcNy/unXL64ILRz2lixfPDQLDxCRJzgXLBOlB7e3VaLs7GrJpZtxTiO9KmQio0Oi+o0NZq87KYutjoDwMa7E/P9mgx6SZmSBVyZelP28TyOlJTwsE/M8N7hfMBVLfXyHFdF8vUNElaTym0CPV/0Qv2cyvvbcajIZPwqWIoN1qVTrwb4ipnf7gx6VRVRizE0Oxqm89XiFX2wY5gmoF7UulZpHBERqsqygn2clOI6dWz33Dudxsd3hAgHVCXdEVnsN++7z+5C22y9NWU2mzUkc0WIszkU27R4ZzwtH2uCD3v/1KK693nUCUSWrDidaAizxiYDOVlubR6mzrlx8HLOkwfioZ79P73uHgt7X4CIr9jPTmUC1GbKvoN6amCa2+tpdzzTmHn/vY/p+4G6bpgv5lxcvOB4UVOS0o0d3ThydnZOnlU8fet1Hj56/Iv+03+hdfAAb7ms7WQRwcUwcWHsZe+2PV2gttv2oA48xCioZnIqaDSqACjzWtBloIoQnWc7FOaLlvr4ntltDGskDzQRxjIwO/LEnWy+bhlGs5cYx8xms2EYB4I0X+xC/ZzKtzXtpE/rtiNhPkdKYSyF9aZn3ijvfPwRbcncXHWcq7L68fuE6g1my5ovHXn+2e+/zX/3r37Ik8eP+fVvfJ0/+NbrfPm04fLDNeuL55RXz5Dtijp1jNcvODluWMcF0s5p7j2llJF+veLk5D7z5RHjmGxEkyefQyCngfb03he9XD/T2q5v9qOQUnlSgm03wjKSVUk6WTU4N6krp7BYMQshZLIPcpiis6ipzMdkUWeKqUFLmUjtOnlLerPEEYECJRc219dmWloyTgIpCZvNYIkYHOaYyrmw5yKJWNqBc57i1B72qqg3EriUYkDEZ4qWafxkXoRgPMldpOKOU2y2E0rJeYpBM5+7UIeJU27dDhHZe4HeGVYcIq7bl+zBnU7n7RQJWYxjWoopxofBuLcxRFyEvttQVIlVNXWHTNts43YxIUtxyGRBtEMq4uzXlZTo+462irYhysVAHmpAL42UZJ1ZtPztna6/J+WwLqiGCpmOje7HkuaUAAAgAElEQVQEGbJTQO/oHHbmy10+H1DXFf/Ov/tbtz/TecbUo5OAyU33NNv52OZTDlTcuKuDB3g7W4DgPbG+DZtnsivY7+bEVLNOTEbtvOBxBIU0QikmsXaS0TpYRmcMSAbVEc8WxhtijDTzls16NY2eCkELXpRYGTG9jRHvPTFUVPWRBTTkAz3RosflQFwsKK4jiTD2A6N3LGcN33/3PWZ1i/OFk8WCII7oheu/+YDt2QnN2ZKvn1ac/f7X+PAyUYty4ga6H3+IXDynkR7yQPGR1C64mfe89c1v8sHqBsmZIELf9dx7/Jh6HsjdChkHI0OL+SOJOKzpf1gAT8cprknMNsIva7ZdzzA21N5RtJDV7B588Ga8C5M35OT4TpnixkCCwxVHKZBTwpeJ31QsVsgHR4gVzgfGJJQ+M+rAotkyrJ6TxKF9x3DT8fJyzc062ZjrQJ9xe1PpyZ7EunGeIiYCUNh7Dbop+9o54/6LCkXdNP5jP4UQL5TJxkZR88yTgjizgVAxACOYVYc42T0P73Lbd//AL2ZhfgG13dxMb+1Q7J1jITKd70qXM15s49J1Pf12i3Qds/nCNuQoTWjQnPdjdsSjzsMOZE+NAtUCvhBDNDP8ZGpoKYrHqAyUbBnY3k8qz8PloP7b1t6yTKZNz6679yn61J5W9RPfb/GAd83Tqypa3uyO8zcl7Jh/5OFOD3Z18ACvriqKKsF7Gxntfe+mtrDqJPCzSK2dsbXbRS0haOURLaRsmaneR5xTYrSYp0WoEclIuiKvNoze48uARwjqIUaUQhEBNd5TCJ7FPHByekTT1hzsoQjOBBYIpJHUZ7x41l1P0zScL5bMZjVJhNxvuHfvAdtc6F6+QseRvOoo3jEvytOcaKOnu7rAa0bLhuvrG5pqztGDByzOHlBOnkEVWbqK64vnKAXtBoa+Z3Z6go6WbKJT7mQWx1AyOWcObUgu08NDEHJOpKJs+oFcjL2iU0Re2anZxE30cnOpgmKB6mTyDvx5QVRIY0FcBXjQER0LrvH4GBAXGfpsI65NAh+JdWFQR9f1XF1e8fHFDSlHEyAc6N7mJ2rnQyfT00aVDHiM76glw8S506yIsyjFMqk/xTnjrqtOqk7z1XMu2PeWTHZq4Qhi9y4nppr+7DrcFl4VqylB5PZhD0wdvF1SUSIPHSSLDstZ2W57a8yp0tZzJBtgVtXJdLcYOHTWKFBVA+HTMXHiqKqKsd9OopjC2JsNTh4HA90+4JxHS2G8IwY59Ppp+4k9KNOpL737WtX9CNwES1P3jtvP3/0Ft892JtqNN7HX7kt2X6c2svv/Ad7f8Yp7sq1M8SfGf8g4U/VhbXIndvrc7ro93jk8iqjgvDIOjjQWnJjyNRdTqYX4yRZ+8JFSbCuuOZFLISsEVxFCYHnUcno842heUTVxUmkd5qHQEHAh4lQJo6dJI9k5jn2gy5lZ27DtN1xfb3jzq29zdbPmL77zfV47P+LeUaFbbalnc7Tv0WHENRG3uSYVYbNe4WNEokAMVDHAezeE5WMW0RM5ZfX8GaGdM9xs6Fc3bKeRiU5eY9aOyvaA/MrbX/Ry/WxLbl9c8HTDwJAiQ8qU4vG7kdX0oMo7R/5bU0hkGgmWNNpt1TvQYKBE7DzX7NApcF28MOSR9dAzpETtPW59w5EKaQJ4/TCSJ0AJB+nS8Rmln/mmCWGwB7+U/WiVaWzuQiBpmTpwgriCUwPFNmp05JzRMppfmOY7eGY35No5Hn7ycXbIy65jJpfBzu+S92DLwJpF8GnJpG67PwFDCNR1xWa94rJb408ds6YmjwP4yiY8Uwdgbxitt+NCEbHmUyk2gnfW1XPeEatIydnUvdn83Eoxgcch1r9tc3jCW7ffPD2rdd+z2/nn/bTd4L6/N723o0fsDvHuetDpd+nf8vP+7tdhoopP1f4wmlPrHunbaNbisXZv78b6u1gaF3ZteMG7yOAn76KhJ6dCylDSxGei2PPPl6nv4cmlRrTQzjzL5Zx7p8csZp75LFLFgPdCId3dZBxU+bqGnKGK+FlLzBlJhRgDNUKi5vLDKyRWaJ/44ff/hm989XVwkZurS2IZiE2N9zAOHVcXH5OHTHN8wvX1DWcnC8Z1QYaOiw8/ZnlyztWzZwzdliEPpHFgHEZ809DdbHHDBsbRjrebUgVCQGL1RS/Vz7xiXU2gQJDg8FWkG0duNlvO24h6tZFecRMdaMdZkh3hdCI0AxTrKIlYhy8EhID3AVcsbq9oQUSt000h5TypZC1YfXfN7eOyRD55Yz/gsgfM7R8q3Ppw7Yn2O0+6CTTsxgl3x1I2pjLeZM7WZd11mMrUndp1looqrpS9sEJ2Io/d6OuT/zuo2l5d3AFzZc+9M2B3+/Egpr7s00gIEKtAukr03ZZ1vaaezQh1JvhA9GHXGgItt/nhuxZ4yfsun6J2TxcBcWR1IB7IOB8IIZDyuI+QO7T62zZtd095e/9TPALd5Wbfiiwot8v/uT93+vRn/36548932+075Dp4gOemub0dXKGo2CjW2Uml4hDJu3S8CfA5U9xiOzYNFlEjmIWEKjTtjFKEnIRxzCa1d56cCjmPZgERa5pmzqz2LOewaB3HdUXdeHz0hKrBOaWQD/cp5wTUfI5QJcoS6XtKzlQFvMAb9+8xZMGVzNd/+ZcYuzWvPvqQ66sb3nzrqRmKivDq8objuiK2FW3dMIuB6ALPLi6ptleUdkFJG1YffYivG8Y0kEshek81m8N2TdUs8McWBo6bxATuMFv17aydiP3TZkaFbVe4XI88OoGgguiOdD6NCcWAQc62yy1jMUGFdwRvXJYik/rVQjz3x2fnyr+7jgB7oHomc1i/z0Z1ImS4s0M/zPr0ZT2ZdOxfs1fY6l7Zuld9Ip94lu3gnmVX38GCaoBjb3kitw+yycPa3p6Q3iesUQ60jbddXe7VyMC+RXTbJJoEJ0XJKaMZunGk2/b0fWa13gBC1dRUTUvJNrXxpVB0gAGz8XC3Yhe02MskNspjgmxq3n4Y8NPvzTlPFodKVR3exhL4fIS3H71+9jW//667UwRAJysUvftVn/EjlLsduzu/9pZ4+onPH+pjd1cHD/DsotZ9HFlOaRoz6Y5/bkBu+rybiMlOZLqRClEC0UeCOJIKWuzjikeLUAqMeAoOLYL30NSeyguLWaDxprydVY46OqRyk0FmoaoiWeVgHc2BvVO5RDvd4uQLlvseVZi3FXEwPtJmc4OkzPHJKVfrgZISq+cvWRwtYRxJeWT5+kM2r54TEDYXL1mIo8/C8PIF22FgGEcD494Rp/GI5EKsG3zd2ngrZ3SwmBudlHa8+doXu04/41o0Nbu7YCmFzbqjVJ6r9ZY+FdpSEGfjpFyy8UMnPlhRnfhLE/9LhJ0aXabO2w6aiXizjsh5ev8W4JUJtJRSMDs946F5Z6HgB2uCt69dR+LOh/TzIO2d7t30rbr7YpHbD+/HS7cvu0SK3XH6DGT4mZTyQ/XBk72Zt/vJxVam7TyW05uNddo0S7R4vN+SsnBxvcJXkePFETI3YVbAhHhhGvvobuQzLWMRu5c47J5uGyFniS6Ten93HEII9OOBjmj5nOeZyueiqp9Iv/j05mgHyvfgTz8F1D7v595+zU9suA7z9N/X4QM8YL/33SnRUHZzWXvg3I5rb2+ilgOJeBNc4HF4XAZ1DpgUuOIMRPoGfMT7QCQjuSe6xLzOVNFTB89s1uCCmPDAOXIaLI8QncYtB1j7NXVQPBIUdQ5XTNAiw0BOmcpZRueiqRFp6DYbvvKVt3BlYH11wfXNFTMvVFUkrV6xurhGt2tyd0NzckR/syG7SC6BnDKxccyOjyjjSIXlqxYEhkTSsrfzkCqC83tuzSHVMCbKNEoykrhQSubqRrnpeo6aYJnJwCe5K9N7u+3w/uM724Lpc9w+RJ3zpCkw3MmtLYeNqwzgeZgMlKd4p3SwDaRP1G0Xif262atbjtD+YbU/HLJXDd75SVMjaheZtevglU89HG9/0N92Vh9m7xpuXdNgZ4Qr0whwv4EXhw+RMV9SxchsvqSpGxO65IEPXnzIq6trVjcrqqNTGgfjOOBidUepfPtsARBxZE3INBovE2j3zpNJt/Qg5xA93LD7z/+zdH88Pu/av6Ut3H7FXQD3KeLCp77m9u1DB2//b+rgAV6epO3O3xJh7eSys8F9oltnN8q987yofX6a13s8HtA9n+J2SxC9yd/FFYImQlWIwdFUkRgDIUbjLXlnI+Ci6KgUNed/PdQGnnPGwYOpczrd1Jyi6vBS4aKShoTkTOk7dBxg21PGG3wdqJuWoJmmnvHqvXcp60u0zzD21PMZOWXERXyzwGnH8riiXs4RClo1NLMFkgfWY+bo5JxY13caInqwj7jV2qwidh24pmrIRRlzoevTJ0aF+0nfJ0DGNHxSncYkOuXJTpYQ9tORnT/bYG09cbJf312nzywm2AO/4Pxth+pgj8Cna+qIwtRFYrp/yDTeuzNSnLp207bkE/B7B/D2Hb+7B05ugY3c+a1/3yrE+hZM7e/7u+Si21ZoKZnQtNR1g69qKIXoHU2MCMJqteL65oZzprQXX6iq6rYZMI1+73a59yW7MbvsR0WfHDze+bf8vanbjrbeefcn4d7dtZQ7r2Q/fft0fdbHPo9n99M6fodUBw/wtt2ICMzaBgSKZtw0WlV/e38FLN5iumhLUcQr/w97bx4tWXLXd35+Effe3N5Se3VX761WS60FIWEhwGw+A8gYa8CDB8wqbDOyAMFhMAMMi1MpIxs87MYGjo1HYCEssQiQjBEIj9jBYIFaoLXVXV1VXcurt+dy894bEb/5I26+ynr1qvr1VtWVnd9z3nmZd8+I3/3FN35bGNW4ULEqIkpikxgPgNRLEcXZmDWKNR6I1qgkEZLEkjRaJElSFyiNQluWHqsB73y9qDg7S0XNHHYNWFgTa0QpYBMEj4jHZiniDY2lRdQ7siUIZUWoC5G6fISMtllcWqAaDMhshS4u0Ows4NXj1FKUJaKCTVt4pySJodVq02i1MY0maEXlKhqtxkRPAIILAe9nr/2nrTqT9XZjyRSlqFwM8p+2Lk3pu932oN2z6eltcQKVxPdGQ52gFIneJBNuUrx3MtjaOp5vdun1FPSSI/RSvHjtUp06ZucjlwLAdzy2dZtOPNo7vK52z17e18/w77kJkDQyqOOtLiXVSR1ve4lwqSYkIZA0myQ2pRwN40QdyGzK9jhnXIxxzpF4h80u52Q69d9Ms7jLPk8Qt+nEOityVQJy0+NaKwNddddeVucJJZZdR+0XV3+O5wLJm3mC530kUd45gpWpTpVLClLAiqkDziEEjYUsARWPCrSMA5PGtaNVQA3Wyk5ZCZsYsixFREjFY6wlaTQQ08Ta5FLcjBKr+2tAJAEJcZY+yxW1jQHv6mY3IGFne1wlTLAmVi8XsbjgKF1FJh6xik1TTMjwRinzIa3lJdKOQ7IGzscVAyTEMh9JFoPWG+02aQrtpQ7GKtYkdFqGynlWV1cpigLvXL06gCFNEo5z+w1tpmcaWlsbQtDafas7ViOdZLFNxdYptRFvJ2D/kjv3EsGL19whbCEgSYKI33HRhomrimjBQ2Lh8Z3J/IwTkj2dpTvtPNUI0xm0V3VETV91cnyYmkjJnu6A2XXF7o0kmzCxqanIRN9PKEI9wREbVzgwZrLqR70Wc62TXb028I6FdSL2XHqngClOV7sSRS6bwERSN/2UM0zwriVv+zHaX6YUnpoterqNdc/rzi7kisDGOeaYY4455phjjjluasyw2WiOOeaYY4455pjjuYk5wZtjjjnmmGOOOeaYMcwJ3hxzzDHHHHPMMceMYU7w5phjjjnmmGOOOWYMc4I3xxxzzDHHHHPMMWO4oQRPevLT0pPvuwH3/W/Sk9fu47jPkp589Br775SeDKQn9ul9wsvu8TfSk8/d57EnpSef90w9y74h8hZEvv9puM7nInLmaXiiyfXuRGSAXKO/4v57n7Z73iBIT+6Wnqj0JKm/v0968vX7PHfnWOnJV0lPfvsax17zHXmuQXryFuk9ddmXnnyu9J4+2d+Prqr33/SyP8eNxVz3PHuwrzp40pOvBL4NeCHQB/4KeLN29Q+fys21q69/KufvB9KTNwL3aVe/euq+X7ifc7WrfwC8YOpaJ4Gv166+t95/Clh4Gp7xfcCnAQ4YA78PfJN29Zx29cVP9fpXuacCIy6vG/km7eq/2d8F5CRwHPBABfwx8HpUTz+9T/o0Qnf1l8j7gLei+h+njnnK/fl04pl69/YL7eovAL8w9TwKPF+7+lC9/7J3ZL+QnrwF+EqgnNr8Ce3qy57SA18H1HrgCtnX7rNX9nfrqlrnvFW7l2Rfu88u2b+e2NWnE7xFu/qGG/NENx5z3XPz43EJnvTk24DvAl4PvIfYKH8X+GLgunT0cwRv0K7+R+nJIeCXgR8F/tEzfM+XTV6WJ4nXoPpeRJrAvwf+LfAlT8+jPc0QSVC9qVb2fg68e/9Gu/q9N/ohniReo119r/Se/bIvPUm0e3PJ/g3CayaT9+c65rpnNnBNgic9WQbeBPxj7eqvTu16V/2H9KQB/CDwZfW+dwDfqV0tatfiW4lk5TuJs6Pv1q7+v/W5bwHOaFe/V3rydUTr2GdO3X+HsdfHDoG7gc8GPgR8pXb1E/WxLwZ+DPgU4qz6x4H3A98NiPTkS6hZ+mT2Cvxn4ALwmdrVv66vcxQ4BdwFvIg4y71devKfgTuBd0lPfN0u7wAeAVLtqqt/w78AjgKrwPfWs5B9Q7u6Lj35FeAb6uc5WbfLe2tr5IuIVr5/UD/na7Wrf7H7OtKTB4DfrNv7F/d7f+nJLcDDwB3a1bV62yuIL/kJ7WolPfknpxc58fVfzDvf05M/5o28Tt/ILxPbf4+Lyt8Hvp/Ydx8iWvoerPedBH4G+BrgVuDXgG9AdTx1/j9nSn7QKD+IfFF93ecBW8DPovrGet/dxL75eqALnETka+ttKdADPgv4NER+DHgLqm9Aosyh+hAifw/4IeAOYBv4UVR/aL9t+VSwz3fPAN8B/B/AAeB3iZak9ce59huZsmpLT+5mSo53Hft11O+l9OT3680fqN/Nf0p8f96qXb29Pv47gW8BloCzwDdqV3/3Cf72/wr8lnb1305texDoalffKT15IZFQfQpwEfg+7eo7nsg9ni5oV8fSk6vKvvSulH3tRtmv3+0rZF+7l2RfepfL/pTuvEL2tRtlf6o/d2RfenvLvvSi7GtX37BL314h+9q9PrL/bIH0JAPOA5+jXf1gve0YcBK4S7t68Vr9e7NirntuDt2zHzxeDN6nA03gndc45nuI7sVPBl4GfCowzYxvAZaB24id8u+kJwef5PP+I6JyOgg8BLwZQHqyCLwX+C3gBHAf8Lva1d8C/hXwdu3qwm4TrHa1AH4V+IqpzV8G/J52dWXXsV9DJFSvqa91mStTetIBfgL4Qu3qIvAZRJP2E4L05AjwpcBfXuWQ/xX4L8SX6jeAn9zjGhNC9s1PhNwBaFfPA+/jEmGHOAD9l5rcfTHw3ceGXPyVt/OlwB8sFLwd+HLgT6/8QfJy4D8B/ww4TBzQfgORxtRRXwW8mjhY3c/jyA+yIz9D4GuJbfFFwDcgstuK8jnAA/X1p36ofg/wB8AbUF1A93TF/Czwz1BdBF4C/Pc9jnmmsJ9375uJVqPPIcr9BvDvnqkH0q5+dv3xZfU78Pbp/dKTFwBvAF5ZvwOvJg6GTxQ/B+yEVEhPXkbs//9av2e/A7wNOEbUCf9eevKiJ3GfpwzpSZuryL709pb9elI8wROS/SndeYXs15PYaewp+9q9JPt1P15V9ut+vN6y/6yAdrUk6tqvntr8FcSx5eI++/dmxFz3XLrus1b37AeP56I9DKw+jnn/q4hEYgVAetIjCvokeaIixnY54DelJwOi3/xKMvD4eKd29X/U9/kF4Efq7X8fOK9d/eH6+xj4s31e8231835P/f0r6+9PBgF4ifTklHb1HHDuCZz7E9KTHyIq7vcRYx/2wh9qV38ToLYqfuuu/Z9FHAy+Wrv6vse55/ulJ9MLV365dvU9RCH/FuCn6qDsryASS4gm+3+dBb4vC/xqeCPOGZadcE+i7JXg8TrgZ1Cd9MfPIfLdxEnB79XbfnIndk/kzcQZ0mSgq4A31e7V30Sm5Ecv+30PIvKLRIXza1Pb34jqsL724zTHFaiAFyHyAVQ3iErsemE/797riYP0GdiZHZ+SnnzNdXi+veCBBvAi6clF7erJxzn+26Un0+Ti17WrryVOXH5GevJ87erHiROMt2tXS+nJPwBOTixZwF/WFu//nTj5u174NemJAzrEmfyr9zjmdcDPaPeS7EvvStmfxO5Jb2/Z30t37nq3H5Te3rKv3Sj70ntysi89+YB2r7vs3yhM+nSC/4uoC39JevJd2lUlyuJkcr+f/r0ZMdc9z27ds288HsFbA448TgzHCeDRqe+P1tt2rrHr3BFPPjHh/FWucwfwiSd5zf8PaEtPXkU0+X4y15657Ant6lB68uXAtwM/Kz35I+Cfa1c/ss9LfMt0wPM1sLsNmrv65/VEC+T79nGtV1wlBu/XgZ+WntxDHFC2JsSa6Lr+8UeXab/uNYx++z6wgfHbf4kf+9IP83vIFbOZu4DXIvLNU9syLpeR6eD0K+RnV+zcpX4XeRXwA0QLQ0Z8wX9p1/2fSuD7lxIH2x9A5EHgu1D9k6dwvSeC/bx7dwHv3EXSPTFY/Lqjdu19K/BG4MXSk/cA36ZdPXuVU35orziY2u35duCr6wnjVwD/sN59F/Aq6cnm1CkJMdzieuJL6rAJS4xL+r09ZvJ3Aa+V3pOX/avpzlpfXRfZr11U36Xd6yb7NwpfslcMnvRkBHyu9OQc0Tv0G/Wu/fTvzYi57nl265594/EI3p8ABdEU+8tXOeYs8Yf/Tf39znrbE8UQaE++1LFg+8Vprp6QoFfZHnd21UtP3kHsyAvAu7Wr/Sd5rfcA75GetIhxGf+BaFG7nng98J3Skx/Vrv6fT+YCtZC/g2iqfiGXC/Bp4M13bfHm97yVr0drhdgDRL4N+MxdlzsNvBnVN1/jlndMfX4i8vM2oov6C1Ed17F0R3b/nGucf83+RPXPgS9GJCWa/9+x61mfSezn3TsN/BPt6h/t3lHHtlwNl71rRFfg0wLt6tuAt0lPloiW8B8kzoKfKH6OKHd/CIymyMVp4gTm85+O532q0K564FelJz/DVWRfu8+s7Nfv69Mq+9qNsi+9GyL7zzZM3HbngV+eipHcT//ejJjrnptA9+wH1yR42tUt6cm/IMZ+OOC3iab7zwP+jnb1O4BfBL5XevLnRKXxL4gJDE8UHyAy708GPkJk4vvFu4EfqRn8TxFnUS+qTecXgM+XnhjtarjK+W8jujbWuOSq3QsXgD3rRElPjhNN8+8FcmBAdNlOB5Lesw/T8VNFn5jt9LvSkx/Qrn7Xk7zOz9d/x4iJKhP8NPAvS0OahRiQawJf4N9ESYyN/DDRZT7BfwDeich7gf9BfLk/F/h9dIdIfxMi7yZaKL4HuCy+4hpYBNZrcvepRPf6Vesm7YGr9iciGdH0/m5UtxDZpu7P64F9vns/DbxZevJa7eqjEhOEPkO7+uuPc/m/Ik4C7iQG6P/fT+DRJm12heW3joO5DfgjYphEDjypGpHa1T+prQM/zOUTjHcTrUpfQ4yPgmh1H2hXP/xk7vVUID0RYvjCVWVfelfK/tQk8puk9+RlvyZ3T6vsS++S7NdyeF1l/1mItxLHpz6XE4b99O9Nh7nuuTl0z37wuIWO67i2byOa6y8SWewbuBTr8f3AXwAPAh8kZq4+4UKf2tWPETN33gt8nCeQil2/TJ8PvIY4y/o48Hfq3RO3xZr05P1XOf/PiDOLE8B/u8at/jWRzG5KT7591z5DbKezwDoxHuYb6n13EN0vj+33Nz0VaFc3ie3xhdKTf3mNQz8gsbjp5G8nE7CemQXg/drVR6e2vxP4wYsdjuYJvzN4MxuDf8XbiAkvr0X1by67g+pfEDOtfpIYx/MQ8HW7nuNtRCXyMNHVvl/5+UbgTYj0iROLJ5rN9OPAP0RkA5Gf2GP/1xCzb7eJltGveoLXf0rYx7v340R30W9LT/rEuNZX7eO6v0MkEg8C/5OouPaLNxJjjTalJ1+2a1+D6DZcJb6Hx7i2Av+OXfK3umv/zwMvZWrCWL/rX0C02J+t7/OD9b2vJ95Vx8RtU8u+di+Xfe0+87Jf9/uTln3pyYb0ri77Nbm77rJ/g/CuXfL4ToA6RvL9RAPGH0wO3mf/3pSY655nte7ZN0T12l6qZ/TmPfl54CHt6ptu2ENcB0hPvhe4qF19sskbNwTSk/8OvG2fsYFP8iaxDMyOq3eOOWpILO3xOp0qnTRLkF2F0+d49kJ68p+As3vFbc0xe5gV3bOvlSyeCUhcxuQFxLTjmYZ29akv23WdIT15JfAKYgD5HHNcV0gsP/KNxCLCc8xxw1CH2PxvwMtv7JPMcT0wS7rnRq5Fex7YBH7lBj7DHHtAevJzRFf5t97MsSRz3JyQnrya6Ba6QHRhzjHHDUEd4vLXwP+jXX3kRj/PHM8sZk333FAX7RxzzDHHHHPMMcccTz9upAVvjjnmmGOOOeaYY45nADcsBu964XWffoved8dtfOoD93H3PbfTXmzRDFD1R5TlmCovUFcS8oJKPYkYfCipSkeSGDRUhKpkNBpRjguqqqpXRBCstSQNS5pl2CRFJSMEpTSBKoB3ShlKHAnjMQQMpXeINVhr6bQ7iMDEiPpFP/yuJ1xu/tmO9/7ST6mK1DUWBBHFaADn0LKCskDzEcFVBFdgE894mOPKwPLSAVZWVkizjOUDy4yKgmShTdpeZOgDttMiSyDfXCcgLC52sEmCGsugP6TR7PhjvrgAACAASURBVGBtA6+Gs+dXOHz4KAudDs47rBUQUFHUAxL4u1/6jTPV/o+9663qnMO5kuA93ntUlRAC3js0BEQcQQN5nlNVJSEEXOXxIaB4yrLEOYcYIfiAqiK10DrvKYqCsiwpypJxnjMuHNvDMdujEYNxwdr2kH5eQJKggBiJ59eriiigKH/4kfMz1fY1ruIeUdAAOCg30JWPoeceoupfxDTamOQQ5sidyKEOZA0wKWDBphA8bG+iVQHbmwRfQVGQD7dYXbtAv3Bs9QuKQsEoqsrR2+/mk/72ZyPUTbxXSx95YKba/395+d0K0DCGZjPDGEOSGIwqiY3yZ41gDVHHu9hWxhhEBQ2eoIp3DjHUsm9wzqEKRgRFo/wGT1DBBwgKYLBGCSKIBAwBH8AHxVjBIBgxGBFMYnjHn56cqbYHKM5/SAH0ktRdgggaBFeVOHVgDNamWBK8CxhA8iF+1Mf2twlFTigrBANiENLYpsUACQ4JAaMBlYAPHg2BUMU7F6MB+XjEqCrZLEtWtrfYyHM+fuYU2+OCtNniR371j2eu/SeYeYJ3oNPi4GKTdqeB1YD1QqgqimGfIh8SigpXloSywqOMq4qqysHaaN8MFThlOBqCK3GVQ0QwxhCCJRhLkFoorWAAKyDGUIkSxBCCIBpQQiQVNaMLKKLEAW9GIVwa5RRlohYFUA2g8fuo36fRSNEqYIJQDEcUaYOFTofhMMeNYx/lro9ISlBwwZMttAmVkmQWV1S4ymHTlPFgixQlaSjGNpDgaWYpoh4r9QCril5jzLvZ4b3D+wr1Hu8qnPeXZhMh4KqCoAVBleAdwQdc5fAhEFRxvqKqKpxzWGtxLha1N8bsyKwx8b/sbI/ybK3FGkNqDWKU3BWYRgai8X4hDqiTv+cG6t+pFVTn0e3HYHMdzj2E7xcki8cxL/wUWL4j1tYWH88xzXhe6IN6aHtkANxyO9ZV6HafRIWO82xdOI+1Fu8LNCjNZgNjLKTZbAr540B32nx62+UfpJ54Xl0MZdfnPQ7Uqf9Sf6gnQirx+lc7ZRYR5FIJOr3sf/3NQJY1SDShLMeEvI8EgaBo8JCPkHwEwwHiHFYNGMEYS0Aoi4rgPIkREpPUEyaPBod6xYpBgDRJkWYTtYZKDAfaHh/g+XfcxcdOn6KfD69ru1xvzDzBS9OERpbinSN4jxuN8MWYYtjHFQW+KClGebRcFCPGeU5VBsqglN5RBYeKofSBBhUpSpZa0jTBokgVgICqYK1irEUEjCZYAYuBoIiNg6gqiMRmHw4GQCSDZoZJ3s6rXVt9jAas82jlCEWBFgVVMcYKOFeSGosJsLWxwcHDhynHBWXWQICqHFM1ckpXYVtNnCrjrSHWKq7VxDYbWO+o8hFkKWoM6gP99XXklhMECWBkR9mIgqphFtVtno+iVUJDlH/nCCFa4cbFmMGgj9cRIQSCV6rK451HxEaLXW3B8z5gbbQATs7P0mgVQSdtKagIKorgSQWaxpAlIKKUvkKLwGUma5gaEGcZE5IRQLdheA7dOAMrp3Dra4yGOQv3vBjz/E9B2ieIZbUU/AiGI2hZSBsgDdAyWvUWBSoH3oPNyDqHySrPwuKYstokSTwmEYy1NQnf1dA69VwzjJ2fOcXLZPe+q2Cvifdu8b20A+RyJnNpx141omdc5n2YrHA2TYil/qZIPZlUVYyrMGVJyMfYoFBWkOdIUWDKEu89kqaAAVGsESTNCFYQ9aj3aPBoUFADIeBdhYZAVZZAJHytNEMWDGVZUpYFtx05yiNnn8pqfs9+zDzBy7KUpYUOrVYTQmCwtoqOBriiIB+OGA9zytJTVSV53mc4KtkcwfbYgTrWRgVlPXgJjoUs5dhigyMLngPNhEWNs+tSS9IswZJGO5WAqkRCR8AYRUI9wEl0VTWy5sxbMYJA/MFACITgERS8xwRPVRaUoxwDbG9tYYOSpdFCt7K2ihpD0MD6+jrNdovROCfNGgy3tsmaDexCh/XHzrF4YInRYMjysWP0N/s0Eou6Cl8WDMYjcCU2BNQLYAhmosAFEYPM4GBXlkX8bcHhvaOoCoIPDAZ9+v0+ripJUhddtx6cV6rSRxeUNVREl4d3HiMSrXzOUVUljbRFo9XCWENiEyqJlm0VQYxg6r/EJlhrY/uqxueRGPo7GT9nd6ybyJQHLaBYQYePomdO4h47w/bqBknzAO0XfhrmvhcirWPERXgCuDEUAxjUFgZxEAKIAWNAE3ABSJCQoiHHO0dZRZdX1kxri6qJgdZBL2/o2RP3vaFTVjyZfJ9uit2k96o2uqnjw+XniexxwpXbdoilcgXpnDVIqK3PcDmTVqIXpyyoBjlpkmKCRwdDZDyGqsJWjmowxJcVRpWAIEExmQXvQAJYg9gUDQYRGy3bFRjjUQJBPRoi6TbWkKhiVUlUONBsx1AVEbaXl29QC10fzDzBW0wT2q0GraUFqv6AcmsLLQvyPKccDhkMRgyHJRvbQx7eyNncHHNwsc09d9/KHfcc5dCBI5jguXBxlUcubPGRTzzMydUBS60Gdx9Z4PmHDUc6CaaZ1DEvHtTiRfFqohvWQIrU7too6EmakiRTzT+jL/oOuZNoyREDRkFdCUWBliWhKlDnMCge2B6OGI0GjPKch08+wpFDRzl75izNVossy7AmJZQVj51b4cQtx1k5eZrWfffiROFAoOiPyDoZmqSQBNZXV2hmGeIdYhOCBlCJ8Rwi7PhpZwxFWQAKroIQyEcjNCjDfp9qPK7DDBQjhlE5RhVC8BhjcN6RByFJLGkjJTFCcI6xG1IWBVsb2yjC8sGDdBYXdmLzhGiRFhMnOUliadiExBiCGKR20tfBdzNM7mqoB+3D6DH01Idwpx9hcOEcpfN0brmX9vNegbn9+dA6TCR3CqGMfTaIVuhI6EIc3KyJ+4oYOkIV6n0eQ3R7NZzH+YCxCYlNSNK0nlhOk5Ib1B7XGdNqVS8z2V2KAb0UE6pXnLODevfOYbJ7t7J7xiKXnboXm5tZpQ/5iB02XUPq76qKFDlJ6ZDKo84hRQHjHFOM8XlOKMpo8DYWrb1g6qOBRCSgIngBa9M4brgKTMBIhUrU7cE7CErpKlwIBB8wIdCwCUvNFoNizC2Hdy/fPFuYeYK3tNAiccrGqcdolBU+HxGKkrIYUYz6bA5yPn5hyKmVggPtJq/+9E/ixS+4i+NHDuHFMQqG5QMdjnYsr7j7Vk7dc5A//eDH+OAjF/mfJwvWRhWfcuchjgoIJSQBQobWsUnWmDjgodFqZBN8MCRiMXq5+XoWobXim6jO1BjEx8SVMBox7m+TWUsRAsP+gHHl8c5hjWG4PWJza4uzpy7gykA+GpOPKxaXFjh85AhlUeFzz8pjKyQkLB0/wuaFi5TVCGcgWTyA+sBwe4uDB47i8hFWWpgsIWAuKfYZ1bPelVRVhS/GWDF47yjLkuAdSWrx3kOwVM4R1GCMJWu1SNMUMYbFJKPRbGBQXDGmGI9pJCnNZpOtzW02tjY5d+4sS/kyi0tLMdartuBRk7xG1qDZ9CT5iFIDxtjoqBG5xDdmtP0juVuDtUcJpz7K6OMfZnh+HVk6yOK999N63kuRo3dCY4nokh1GS1vwUBWQtSG1gIexi9tVwYfoni0K8DGGVT2EYGJQuxisNZBYkmYDm6VcYXWaUX1zJS73jkSOJ3vM6XYL4ZWE7GpW/ktH7u2OnTbWyRTJnFGnDQC2iK7RHVI9Ta5DIORDrI/t5auCpMzR8ZAwHlONcoILkdglBq2tdSQNVC2SxoQtYwwQCMFh04SqHCNOkWCwJNGFC+jEK+EDzgMOrFraWZOxd3s9/sxg5gmeFQijMUVZ4okxX6Eo8VXBymbJ+x8ds5GXvPpVn8RnftrLOdQ2+HKLfLxBv/DkARqNwOrWGicOHeS2I20+51PvY7mT8hcPnefR1W0WminNE4I1BqUJKnhj6uDamHhhgMRYjE0w0kQVNCi7VM9MYqJgRRWDos6hZbRS+LKgVECjJckVY/JRztrKGqPhmLPnLjAYVRw6eJBWc4HUerxLObeyjbWWCyuPUOUDgq5wiyQMBkOWDy8wVM/iwjLOwHgwYPHW26DMIQGxLYxJUA2IWJQwk60/Hg8iwStjPApKzAwkUJYlAK5SRCzN9hLNZoskSTHWYm2CoiSJjcQizVhcWKAoCoo8p9NeoLO4wMXVVdbW18nHYw4cPBhdubXAixEyk9BMU1KT4Fx1aVIz8VPNYsNPUJ1BVz8BD3+IwUMPsd0fkB29lYP3fhLJ7c9HDh+LBM6NofJQuOiCDS6auUnituBiXFJVxP3WgiugvwFJG6jwlacsHL7y0QqbWGyjQdJoYhuNKy1405jhPricRF2e07mnV1XrYy6z1F2dGF91in6ZlU+u3L/H91nCeH2Ny9/vSy7aEAI2OEQNiMVoiVTbuNEAzUsowahFTYJKgml2SFoL2OYiqpYYMu0RAhIqVAy+KmJOZAgYr9Hip4oGJdRWw4DifcD5gEokiFma3pD2uV6YeYLnypLhYEArBFQ8pnKM8xGr/QF/dWqAWss//bxX8pLP+GQ6zQbV9jr4isceO8vK9ohWYkm2N9i+uMGSTUispy3Ky19wF+20zZ9+9ByPnNvmYDOlfbhJRoyV8dai9Qwi1AojtRY1BiSQZJP1ieXSeDej0DoORhTEOygqQlmhZUVqLPlwSFVWDAcjRoOcclSwsbrNhYubBJNy7MRtLB84gnMGN66oRNAEGgtLLB5JsG5ENdzg5MNnOHpsGV/lHDi0RDWuCIngyoJmatEyJxgwiUGsxNgOuZQkMGsYj8exrEOI5R5QqJzD+xBDskRIkgbthWXa7TbWJjGEQCSGD3iHFa1dfErlKzKljqETFsIClXM479jq97FJQrPTxiY2ulgABBpZRiNNycsyOuF3RbfPrOivfpTw0Q9SPXaW0Vg5cM8DtE88Dzl6F7JwCLxA7qGsyV2ImYD4Ms5MXSwnhAaoquiaNU2wWTymdOg4R3FUPqeqxrGchAhZs0mSZIhtYNMmePZgGLMn87sR546XyIXWGa6yEyMXmZhMU7WJO7be8/gCusvaN0Xu9uB5zwmYqRi83T87VqkRQlmRJoKvSqphTpGPkZEjdQbJUmg2MIuLSLONbXYIJkUlxauCOghlTGpEyYclmUSjiQ8e9YoPAV9XaQgoThWvAa+BMngAmo0Gs4yZJ3iUgZJxDOQ0Qij7bJVjHrywjSYJX/vZL+H2F9+KXz/HY4MBm9t97Kjk9MUBZ9b7dKzjEVYwJqG/OebIgRYHjy3SaTV54W23kOeOP/nEeU5e2Ob2hYyD1qHYuiaSq23yKRghsYY0SZBGulOjCiIBkhlmeDH2IsTf6ANaOkxQ8nKMhhh4733AmoQkSSnw9EcV6dJhsuXDrOclH/7YafJxwGuCVyW1grFrHFhocdvBJscWF2i34eyZ82xvphhuo7y1YvX8Ok2TIi7WbGtYSyhSjMlQGwimDlyfQe3r69p3IoYkyXDOEQIYY2m2MqyxZK1FWu1lECFN0p1aeSAYDYjzqNE6O1awJoEENGtgrUGsIUkTjLUMRyNsmmDTFJtW2MqjwdDIMlrNJpuDQbQk1pjYr2c1wUhPP8x4ZY3CGZbvvp/m8duQ9hHAxBglSaO71RsoPJQjGG9DpWADOI8WY9SVUTqthWYDyEAsBIsWm5QSKNyAajwguApJDcZasjRFsgxrE8iy5wShuxZ06r/sUK+a2l2zaaaI34S41WRCa79rVG2Tap+XSq7EqncBRWOlBI2WbervZkanN+nEeidTmnWnA2KZJO886sb4cU4YjdHCQRkgGNQqlQ9kSYIkCcYmaNKIblofEK1Q52OJs6oiTRK0GBOci2WhfKzlGbwnKHhVXHB4DcS6F7G/jJ3ttR5mnuC5oqCoS3MkRiid57GNAjUN/t6n3s+9996KHzvOnHyEj55fZXXDIYXhwYsjPjEaE8RhjZIXJctJymfdcZD7NvrcetcxDmUZL7pzkbXxJg+f3eJ8f8BStoyxJg6KdXasuJjJI4nFJpZGux3jZkIdPzPDWbST8gETc7nWlrtQVozzHBHBVRXbW9uoJJSV54Mf+TiVpNiFZf7ikVNsjAqMtMmyRZqtxVhqxRicFx7bHnN6YwNLwV1HFrjvxO2E0SZ+rGytbTLsFxw6dIhQesqyoBE6Ma3eVYixMbPZmJnMs1CIGaxiCB6CCmnWIEkSsiwjTTNM0sSYJGZcYjBoXTbF48oK8VXkv7auYCiKV42lfdSS2IRG1qCzsEA+HjMcDMiWF3fcvD4ErAjNLMOIULrqsszxWZb9auUcNm2wdPQe7LF7wLSANuQlMIYKdFzCOIeVizAaI40WHL8NkhaIh8Qi0oBijA76lKdPYcbRyirNlLDYZJxAoXm08GnAmIRms4m1GSbNMNbCJNHiOQSBnVInIrE4sTGXYulENE46dyx2E29KAJSgIVr4VQFLIFD4ONExPmaYBx9qsiAEVayAqqvjwywqggkKEiYFFOKcXxWxJtZ8m0GkbjoW8XLr5kS/2KAY5zCVQ8uAdRCC4BFsmpG12tGFa1OcCiQp0mhhVPFFDmGMMTGW14oBEVyYuGE9znmcdziNXgbnfdyvoY4Nhz3KE84UZp/gVRWl97SylFJLVoZjTq/2+dsvfT4vv/c4UhVcOHWe01tDSnOQvzp3ho/1x6xXBaWHhWaDg0ttyu1NHhsG/ub8NpoXbIWcV95zB4tZxv0njjAaFKwMco4fWuBASLEiePV4LyQaZ2u1d4Asy0jSdKcm0OXZXTOGevA2k5jDAFaFqvJYY1lfXweErNGkP8g5+cgpSq90Dhzkr0+f5cKgINgORw/fwite/iruv//5nL9wgYX2Mptrm/zZn/0pG1sO7wIfPXsRQsEDtx5m5fwaG9t92gcP0elkjEZ5dAWMx5AkSJJikpTgBTFmJgc/oW5+AWMS0szUbtkUayfxoEIiBkOsxViGgLpY3Hiy2ogSkCQWGlUTcMET6gLKEDM32+0Oi0uOi+vrSOFJbEYwHsTjgiNNDM1mi3xzKw6c0885g20PICalccudsHwbki5ES93WCJUR5NuUp1YxJiPpLELnINx5FGwCrTZki1DlSLUNeQBbII0WjaRNWN+iuHiG4SfOUyw2MPfeSZ4o3muMvwOsybCtNknSjFZXktpS/dxAJ0ui41WENDFkScwIN3WR82iMCxMfLsDOpBtVjIkTmqCx7qN3gcorwzwmu9g6Xs97parPqQJkNhb/Tq2SWI3ZnKqRWMrENax1aMiOz3jmEMpix0IPXBaLJ0ic9PtAlY9jzdJxifH1aiCZxaYpYhKUWNzY1yWWTCNFsKh6vDdg4+S8KqsYshAUArGup4tWPKeK84qfxCCH6LI1QXDMJsGeYOYJHj7OJKqqYqAVj25sc+xAhxccX8A4x4UL5+mXyvbKmA+uDvjzc9to1mBUClt5yaAU1nPH4ZZlYamFTTMOdVpUm2MurPQ5cXyZ5x87hnWWj516DFc5QqIonioEvAVqy51FaZkooCEWykOvEoA7S4jFdomz49qsXhRjCleCCIOtAVuDAacfu8iFc5u0O8usnF9jmFcESWlkHW45foJ777mPe+9+HqnNWFo+wL1338X6+iof/bhnfW2NEs+5rTGHWyOOZpbRWs4jZz5Cp/My2q2sjinb5kCzhfgqur6MIIlFZ1DRJmmKhliEW0lIRLDGYuql8kxiMRikrk+oHggOQyARRY3g67g9CRIJXr0CSPCBysclmoJaVFOy5gLtBc+4rEhsDGBW76mCkiYJrWaLbTuoEy3i4CtGZrbId3L0TmTpVqAZkygGQ3Q4gK11hufOExZvZeGFnwQLLaTVhvZCLIlSldHipgIbW3FwdCWMCigdpnCkjZTQHrO2skbn8EFkMcOVBaZpSRoZNkvjEoo2Js0gszmJuRrazTi0CbEeY5oIRkIMzJepyU/tRYmBwhqXuQqRuHkNOBfwzuNcIK8Co0rwzmPrOmuVr0mhEcogNEzAGkgTQ5rEbGZjDMmUl3fyYbIKzCwiVMXUt93xiYL6gK8CrqpwY4+M4gpGwQjGKlZispAvS6QYgc1QX6CpQRcWSY4tIxcNZrBBIJA4h/qAEYdTot7xkdjlRYkLHjVQukAV4rWDBrzOs2hvamgIqCgB2BoXNLIm9996mFYSePjRRzi3VTDYFrY2Kj50cR3TbGKM0B94qmAwXsi9ZyFNOLac0G5nvPJvvYSHP36Sx86vsZDBbbcc4v7bj+OrkqIsUGOoglIGxTmHkYTg/U78UZo1EGuZTiOfVTeVaJ0VGDxSxbiJ4DzFuIirK4RACLCx1Wd1dYODBw6yvZUDDbBt2s0OR5ZvITMNTj18ilB4RvmQ82cvkGXR5XLw4DEITfr9Jmvjs4w0oyg9iWmBG7K5vk2WJRSDLR54wQsIVYUvbb2mZEzDJ509ZZtYGzPMgsalg4RYvsRa1ETXU9DoKnHe18W548AmtUw6CVSuAlcPTiEOgj4EShxKzBUw9fJ9WZrtJHIkSUKSJBjnoyUlTWi1W0wKHk+WPIvurNmDHLobSGBQwMYmenEN3JjRhVXSxkGyv/VZyPIypA5anWjhSxNod4AQY/QGW+A8yAY6LqhcAeM+qGMYzjOsFLYGLCYtkkQxnQ7N5QXSdotGqwmaYWwCu61FsyfulyGzk0lEDI8xMnHJTtyk9cKJ4dLSidF9GFd0CcGTjx2jKrphq0oZVYGRh8p5UvX1mrLxHl4MhQ8UBDR4EgNZmtDOkh2yl6ZSVwsglrHxYWYnN1Rl/WHvJeCC97hK2d7uY6qAFg4bwLYbaNAYW2dsjL8LNREL0VqaLLYpDTQOHsVXFZQlNiupigJPiHVOayLvXBXj7kRxIeBFCQJOA5V3OJ1b8G5qqAGPMizGhBC459YjHD/QZnVrwMrGmAsbOdWWstjskJPsFIofe0/QBBAaFg4ttLn9lmMclIo7XvoAhx54Pn/+G+/m7MoWS52U47fewe0njrOyvoFJUygV5yqC6E6KtgJibYyLuiy488qaS7MCIRICdR5xDq0q8sGABGFze8S4KNgcjjh9eoUs6zAajclIOXbkNk6tbNPuHGShfYjMNClHJWdOnSGEiio4Hnjxvdz3vLsQ28BVCVnWYWVlxNAZDhw7TDnKkbDOuTMrPPTQQ7zkgfvIh2Nai55gCgRIO50YbHujG+oZgMTAI0RDrJWrsT/Ex8xNqQlumUit8OoC0GXAOMWXJWNXELyvS3wp6kJUmt5TSk0KjYlB5ArJFMmDuCZtkljSoCTW0m61CSFMucVmmWu0ob8NozE62kY31qhWVgiNFtmdx5FOIxY1JoVhAe3lONGwBrzWyqsBG2vo5oiqLBlWBd7lMNygGheopIwHfVqpo3FkgbSd0ex0SNsdbF3pX02Ug+fA0iE7yJJLb7TE6IL4GaAmdKHeUJXRiuNDoKocg1GJd4HtsWdcOMaVw7lASDK8QBWUxVaLlpVI1AQGhYMEWklcOURDIC8d5agitUIjs7ScJTGR3IkoidmrHt9soKrr4O1g2kUrJi55WMUtm4M+EqCVNkiqODlEEkwa0KLAG4skEIoSO0mITgSSBhBrdhojMZlFlEDAhQrvSnyo8ERiV1QV46oiCPE4iYlos4yZJ3gmiYGsLihpamlR4otA36c8dKFgYz3nviN3c+DAIQ5sVXSans3tPiJCpTCuKu5ebnP70QUOHVzgdi1pLyyzcHiZW++8k4c++EFWL/Y5eqjgyKElnIEcwZcB6Y8gOGy7gUky0ARrUxIVYvjHTh7hzL7oRmMgswYHviKMx4w2+4w2N1EXGPVzTj8WyZ0rQT3ccvRWzIFbaG0ZjF1ioX2A5aUFFhcWGZdDtra3GOYDDh99BZ1mhyApKytbtLIOw8FR+qMtlg4cpkpHbPY3CVWFKwPr69sUd3gClosXVlg6dABNMtLFxZnMYhaJ68VGlRcQ9YiP2WuhrPCTpWElJmB477E2RRyoC4TSUfgCdTGpxSCE4HHO4ZyjDFUcOa3BabTqWWNpZBkBUI2WO2MN1pgYA2Umy5TJZX8zia0NcIBmUARc4aiO3s3CS18C5RasnoaRQFPg0CJsb0OV1+4oIBEox+jFFcK4YjgKjAZj/CjHrW3Q3zZoyLHpJpJBdrhFM0viQJlmsRQFgpu4Z2e1nfeANebSz61XP4CoX0Timqbiow52IeBdoKwqysKxMayoKs92SVzQPk1YWkwoMYyriiqAU6WZ2pigRVxURFRJE4shYElQY0gIlJVnmFd4H2hncTFLYwWrisyo4teyuqrdQonrxgYHjbSBbTS5sLFBy3sO2oyWTSmLCiEn9XHVCrUeTAPW1kjaTdKFJtXKKn79IuIGhNEILco6fjiuNeuq6CUqfEVBYJiP8EGR1NbeCj+7uqfGzBO8RhqzWFtLDY4ePkQTGBcDTm8MeOTcJstZkxMn7mH54GFu27xAaFb4Uc6hZpM1zTmcWu4/3uaeEwdYbiV88uIBklYbFg5w10tfyice/GtOXRiydOgCt959OwudJlYtGE/hYFSMaLXbJJKgTjHGksYkLSYp+vsqtXSTQtVDcEhwVPkIrSqCqxiNcjY2Njl15iyr22MOLR+BwnPk+FGOHz2Oax6ikwxZXjrK4uIyncVWjBE3gQurj3H+wjn+8i+Pc/9992MstDtN3BiaaYdRvokkDVot5fChY4xGG+RhwCceOk2j2WZ9OGLhYIdlEba3NjjYaiF+jwXBb3KI2DhTDkoIjuA9ZT4GHxAXkKAE7zEeTJKgNsGZuB4tNREsfIkriuhqFSE4T1VWhJrgibWQGKrg4/qQIS7snSYJ3iuKJQkWYwKJjeTPh6kki9k1XsfsWGdi8dbNVbwfkaQduHAe6a+g2dm4RNm4wGeAVITKEUzCuJ+DVnjAbQ8pywqvMOxvkG9tMlpfp7Ce1Xaq3gAAIABJREFUTrPJsjW0RclMSitpkWUNJEmjO9DHFUpmto2vAptMBGvqv4Y6scJEF6GLxK4oA+PxmPG4YjR2bOaesQskzSZHlzuk1pAmsDYsaSUtnFQ4I7isyUK7SVU53LCiCoGGV5qJwQOpMTStiaESqSEvKkaFp41FJCZqJHY2O8aEXXUup2sRoqhXcELwkDRaDHWV1Y010s5B0sygxsXJok2wPlrdbDnEnz9HuTVAFpvoYJOsGlKWI8p67VpXFJRFQVEUDPIRRVUy0JLKCNv5EGMTrM2iW149zRkND5lg5gne4kILrTzNNIXg2ewPKcWysjKCUHE4a9GoNkiqMfe3Uh41hhfcdRu3HHc8unqRe44c5u5bFxBruDNr8MLn3w+dNmotS0cO01pY4MGT50nallvvvI2FdhOrKYpjXHpUPEmS0M5auCpm9RRlebkFg9nNJPQhEjytCtRXWCM459jqD1hd28R54b577uHCY+ssLxzi1uMnOHz0CGsjoZVkdFqLtFstxmVBu5WytnGRU2cfJi+GXLh4nhc98ADttqXdzihCoJEtMhw1GRae4+0mRw4fZSO4GNOhygf/5qNUYrmjcYLi9BkOHz8el86ZQVN9AErvcVWFBoeoUhVFXL/UB0IV2yVNMlJj6Od9nMYY0dRmmMTS8IHUWoIP5KMhrqyo6ljK0kdXedLIMGmCYRLYDmJMnNR4ofIuhv9NrfkZ4LIMxpmEaUM5hLWLDM+sU4myfCwFP0SbDr+UcXHtQapCOfmxnH454i8//Aih0eLhk6fxxYjWwiKlNGi1WqRuxMKgz4E2LC+3+YKv/cccWjqGefDPSNNAsryM7RyAVgexMalIjSCY6O69ZNK6oc1yPaDeTXHamGghCipCqCKxC94zHpdsDUrGecWocIzKQGUMnYUmFUKj2UDFYJsNFqxjrT/EmJjtuToscWJBAv0yZooHawlWcM6TZSnGClpULHZatBopg7ykP3YsSorBE5LZ7ItRf3DVOYUQdQxBsMT4RIxhIx9xzhukEeiUTRreUwaHD2OSZoaUJeORQyTbWcKvCg7xDq0cRQiUuSMvPblTKg308y1yLQiNhG03wCQdGpriypyOtSQz/i7MPMFbXu5QDHKMGvJRzrhS1jb6HEozjt9ykHahJPk6jQMHecHtR9ha2eDInbezeKBN7odIVZAXFWZ9i1eeOEH7yGFYW0POnMI8fJJDwYGHM2tbFEXBwuHDpNKE1FOpoKOAMUKWZggO7x1FUdS6diqDdlYJHvXyZOoJ3pMP+oi1dBaXcOdXOXT4KNsbA1JJOH7kGEeOHOOOe+4mP7NGmlygkzWxgKSW0pf0R1sUVZ9AwV9/+AOsra1y5NDdLCwchUag2Vrg/2fvTWMt3dL7rt+zhnfaw5lPzbfu0LcH2223aTt2IHY6UhAggYOEFIIBxYggLIj4ECQMVqJKRfEHIiQH8iEIBdFSkg9ECKQgBiMsrABSCCHudmz3YLv79p3qVp1xz++wBj6sd59zqrr6dju+fW/3rvMvndr77P3u4bzDWs96nv///0ynJefzlk/fv0/Tduiuo2sa6q7mrF7yD3/9t/jGe2/z2R//QW7du8dqtaTKi496V33g6LwjBE/Xm4EaJYkb0wV826IiFFnOsu2YTKdUO1vs7u3T+UjTtMznS/Le0iHEgHMu2Q50HV3T4aJLNhKkUryOpO4tKpVlpe9Nq53GWoMxriezR9bW0mtF40YiJAGPq89YRMf4ldeRgRDciseLKfs/8TGm/89bPH7yFifD2/zKr/4jfuvr71DpnN987z10CMT4hKwcUJYFuybyKpGt3X22xlucfeOrdMXbbFWwe+sOev8WMtqBYpC6YsS+lycgugSlP+o98qEhtpcLtiiKoFPf0ugjvku2Sk3TpoBr6ZmtGlovjMeDvoey5e2zBbOQOKyz6ZJA4HQyxTlPBAqrcV2NC4HZqk4m6s5RZYamc4yrHJUbHBGbZ1hrUMYwn69YLB261OkYbSCiXM2gPvMcka7rsDpHx9R++db2DoNqwNGTE96oF2SzKQf1gJ3xmKK1mLqhM3OCi5ioUWiC1nS9M0PbtjSupe06Vr5h5pc0ylMbzaLVLOeOo0lLNS6ZhXPyGDDW4gYbO/oAL0CAl2lDB7SrGuWF1glbo4qP71eMYkWceGLXsmsrBlnO2+YEY1u2tm+Ty4iTx++Q1xN+dHeLWzcPEStEGwGF1pH727t89iXLyewxzapmfzjAVmPEp0RJlIZoBFUoMmOJLiSTRkkqL6TP3m2oR1UyC00ZHeeSofGTo2MevXvEctWRWZidzrhxcIt7d+5wcOsG5fY2240imt/FB4+0EVtYmnZF06zwoSPEltOzx8wmc3htSFHsJaNeaxE74GzZsrW9y2I+ZzgYUWYlmS6oyiFnixmnJ1MWixWz6QyTV8nqYMPg2hTYuqbFdy2tc4Q2Ob0H55lNptQ2cRB39vYgy+h8oHGOumlp2g5Ci6truqZhuVimNnzO41xH6zpESRJUdF1qbJ9bxCQRker5ScpYpPVoIL8gvG/e/v4m5AaCpmlX2I+9Sn5nC6bvcVJ7nnQjhrMJg9GYc1eyRPHobMWbT46oojBb1cSgMEoj9SnTqUYPMw5euo+OObMW6lXLa6+9xHBvF7V9EzEmfab0gVxvTSQ6cc6Qb0VD2MRjES9iC4mB6JLYy7UO5zxN66gbx3yVKi2LDqoyYzio6LzH64xRFXh8fIYSoemSQXfbdQiCSR0naVaONkSsgJGIc56pc/iQdv6qUeSZpXap24I2lsEAjGpYth1Ds5njvjHrxcQzAZQkP1Tvk/DB5BnGR/bKAdJ44v4edZHx1tvvcnR6zuHSM6hybGnQJnEch3lFLgqxChc8bXQ4CXgtLF1k6jxHq4bj5YJF27Bsap6cHDNrV+jjI6yJvLp/k0FmN/PUv4KND/AKnSG2YCkBglDQsbtTcHDzkP3xPpqc+ZMpo7Ii1xl/7PCQN06PaXwLNrA9X/KpouC1O/fRSqWG3wFYzonLGWVl+MRwn7cnimpnG11Y9GhAlIzqvMb5hmgUo909XNPhm4bgk3kmFy7+kbjBfjxKpFdwqtR0vu44enzMwd5Njt87obIFh3v77OyMGe2MkTzDFiW6yFFFRowd7arG5pbMWpQkf7aIJ0RH13W41tMu2yRARDhftHTRU+YZlTWMypLzVYl1ASse2prJyRStDQpF6DZv/0fn8U2Hr1vapsa1Td+dItCtGqzWDKsBISrCqiV0geCSys3Na5rJjPPFGavFFBUF6TNv+IgKoMTQ+EgXPDF4MuOwVqFR5ESUKLzSGGMx2qGVQkvyuhLpLVL6241EmUFnKLKc8tYB0S04O5oSd17l5t176Pr3iCFy54c/x5tvHrP/0idZfvlLRBfYyWEyb9G6ousc3nXcvHWIMpaXbhyAcWRtw+mTR+hqwGBXwFZ9WdiD90lcEOO6GRffgvGesGGHYG0/Evu/O3iPdwHfedqmYVo7OheZLztaD3luUXlGkOSLOZlOmS4bJou6t1OKGEkqca01MUZqH3EhmawYY7F5xqhKNltNU7OqG7rW03UeGwN5bhgYg7WGTGUsFpFVvXnUEKBXbz1994IgoBUiibcrEtCZoXIRhiO8a5kvprz68sssT+bMlitOdWQ6mXB0csbRk2Pu33+Jnd3tVBqfnENwlHmOFkO9qjk7O6OTyPHknPPpGUpBwKGyiNaBw719RuMB4/GYUm92VvsFCPAsKsuRLNCF5Mg/LEqqbEwxPiC7f5+BHiUrj6PH7LFkOINaHM5YhvdeZ7S7jaoKYtcRRSefqnKICRl79++xmyluWkuZReJsgtcrAp5xNSQEj6kK8t09msWSdlX35rPJUX3tnL6xPnikSTxEITMZGo1vHHdu3KStA7Hz7O0dMB4MKasCkxmi1fjgUVpwEtgdj1NZNnp2dvbor1ggXbitbxlUIwZW8+TJe0QN86Zl3tRsGYM2Bq0kDQLLBXjPnTs3aJc17771DqIyxGYf7Y76LkB8gNbhljVds0rUAKBb1UiIWJsxPT8neiGqKWQZpiiIUZhPZqymc+azM5bzSeqOICp54zlPVhQ0WUanFY1rMJmmDR1jkwyO19uLCEGDNYY8y3C5Twsl6I2+2VweXlUgiyWihLA4ZvrGI+pGsXN7G9ndJ+YTZHLGq5/8cfbutRyfd3zpa7/J17/8Fe5sD3BxTuc8pdW8fucOP/NP/yFmp8cc3LnLZz7zgxTGYHKPGo3AFsk3r22T7ySpPJsCPZ/ko+F9hEQbFuCldmNcmsnH1L7KhwBK6DxMVp6oNaOqYDiqOJ6tePv4nNJqzuYr5nVHjJAZIbMGrQyLxjFbtX2GLo1vEZA2CTNiVnH37m0OCk27mLOYTqjnc86WLbp21Lnm1laJ0ZrRIGO+aL71H/F9DK30ukD1HPS2PT4gISAqYAc5o0GJtA4VHavlgnxo8dRs3TjgdrWF/8rXyPb3mHcrFt2Epm555513uXHzFjtZhW4dq8WMs+O3uX/3BtEsKQcOQRF8ZP/mPi+/cp8sADqnLMrrAO/7HePhkE4029kYgjBZnqEiKBReEgFARgVyeJv4+ifJ772C7gKDsEScR413YXsXRttQFOlHe1jMkCfvoiePiNEzGOTw1lt4sySuHNpkeCNkVYHd28MUo9TuxrmnT/r4zO2GQYtCS/L/E20o8oL9nT0kTPnCV36TYT5EhUizqmnbhrZr0L5hWGSM0dSLJWq/Ym8w4HwxZ2//kGowZjqr+7KqUBQZNlPYPMec98GFsWzt7mPami5G2s4RfEeuhIPxFi/fusO0PuWNr73B3uGt1MtwwyA+EDuHqxt825EZQ7NKfpD1omYSloi2yYy6XRFjauckoqjbhqbrWDQ1jXM0zhEEup74bGKLWwJKYTNNqQrywhA7j3QenVl0b2YcJQV4pve/cp3rO2KsMywbCq2hKFHDEfXpCcePzrj/E59DebA+UI8/xvYn71LtvMTwQPPP/kzGpBX+2/O/znJxRtAZQ1PyqXt3+VP/3E/x6bs7rJqXqaoho1v30OUAmb8HWQaqABdgMUmTZ15clmqDB1+nhdG3ghp9OPvkQ0Iyle+zeH27MURwMdK0Dtc5Vj6QW4vJLTF6KgtPTmtOp54ugDGK7UHBYFhhioK33z1h2bQXrcYAjNKEtadjCJyenmLLIZ/41CfYPrhNu5gyPz1idnpCPZ/jved4suJglJHnmrFsHvcXwJgrocVTgV7KJscIRptE84gRMyrwKLaaiJbAdDFluWq4uz3kfDYhU5qX9oe0zjAc3+BLv/NlXKGoXrrN/s37QMbpG29g2yUjlriTr7OfCXXsaGpPXgy5O95lO2QYASdgosdsOC918wO8rQFqNEYrQ3ABc6oJjcPFjriaEx+/A65DBjYpz4Yl9uM/BEVFzAZgh0mFJrpfqzXQLRJhuZ3B9D1EaeLknNh1+GwMpsKW27jFlEFWkR/cQecDfLNCreRKg/V4oeLfUDukZCgqse9lqqm7wHKyIqxglG+RKc3u1piDnR2M1qxWS4rRAGuEcWFZLRuMtgy3ClSWMW0WvPbqJ/jCb5ylecwMGY3GoALz5RRP6Ds0JP+kLkLjHMZacB5L4M7d29zY3kEmLUvV4r1nMZ9/1LvqA4eKHnyL8jUFHu0UXdtydnzK2XyJV4bpakXjA2eLBabIyMuSzntc9AQi0nUMlaHcGXA2ndCGDl1YfGgxCG6+ILOWwJCqtLR1TWVzdN/bFlFElQI83bstp8m370m5aamjq1ACKoBrmT854tbHP0n7jbdRWjBdQ6Ffhls3kekpooS7O7v8yc99jtdkyd//9X/ImfP89Gd/gh966ZBROEeHlo//wCfxyhCtQCbJK89W4BTMlnByDqWCPQWm7DNYEYndxhL6n4cY41OLhxgF13WsljUxgFKaSMQbw9miJsZUSnUhEiJUuWFnWLJ1uM9g95B33z1iuarTGasU1loG4x2i6wgxKXJX9RIJgcnpMcent/jkpz7OeHuXIi8pRlvMn7zH8vSEVV0zXQR2Vf50ILRBiOopb+MrN2nOE9UTB2Li6cYgaGOR0jAuCkw2ZGiWEGB/v2BuBV+WuFXD7d19Birj//vC3yeLkVe2Mt5494SR7ShLg9/ZZjY7J7MDRoNdbn/sDrv7B/gY+yxuREjce7+p9JAem3l2XUFVjTHlkBACvmnI6hVRd5BliDH45QJRJ2jfJFPSrR3kG4AUMBolObYpYLCVSkldL5CYPoazRyCR2HUwmyDVCJMPUKNtMCXFYIgEQcodYvRE0YjJLpSET6ftNjOToSQJLZRA6DqOHj2hC4GDm4d4Lzx55xGDsqLIc4rM0mlFW68Qybh9uM27X37MoKiwmWW3LPHn8MM/9E+xWE559M57vHTrY9w6vIVIWqG3zoNLzcJDSOrlajhienzUK9ws+1sjtAgqwv7uHlVRUGygijZ5bQW0SqrYyWTJm++8y+lsQSeK6arFiWLuPSfiyDpHt5wCGqOF4DsKbejyAoygRkN016XsRdchRIpBjneOw8N9Cpsl3p9LbYW0CKHv96lI5rNWaZzSG0tJeAoxQp6DEXRuCNN3GQx300SWL6A+Qr7xNvglOIVSO9wuSm597sf5Yz/5w8jWDmp5zvL0LZrTjq39W5T7tyDPoG2Q4FOW0FbQBJjVcHaWzJPHI9AdEQXBE0Nvl/Lc7wnYD3XPfNcR/NPnl/epe0XjQAGDQcmODRzPl8zrjqb1LJ0nRmFcZeyMSqqtbcY37qGyAZPJ1zDakBcFg/EWq/mEg5v3KAdjXL3g0Vu/R14UqAj5YMjZ48fUr77K7vaI7M49svMRVhRGhNXZCcumwS4cW4OPZPd819E8x3bqKt1TaY3Wiqws6FyiKhltQGUoJQx1RswHuNYRqoqBzQgnU8Y399ga73Bw+DrBFbzx1d9k+u4JO9rSFJZP3f8hpsePCSGyt7WHUQpjU4cq5xvqpqZ2LrVTJNJtZA+jS2x8gJfv30QPR2itqY8eY6dTXASVW7wCbS22HBCnC6RpE0/FCtQtFDlkA8jydD8GaGpibuHsMeK6JKVazkCXsLWfSro7+6Az9HKepLRaEZYzlLZkxSC170r1KSASQrzkjGwYRHQKvpRiPpuxnE752Cc+jqgcvGJxNiPPShbzGcVsSDWsCMZQDMa89spdvvCld/CNwzsDsWNva0ScNPzhz/4zHL10QqaGGJMRI7StQ4nFassgT+R+CVBWA6qypLCGrLIMCkOWCbt729x9/TUoczYxwNY68Q9tXrBcrHj05AmiNXdfukc2GLFygcF4h8Y7Vq5DQmQ+m6FEsaqXnByfoK2hqwy+7Si0ZjwcsVzMiUoTfcdgNMBqjbWaLLdIlmOLAm36Lho+pmsqRJRI32FAngrwNjbYayNIBdWI3ZdvIKMdRG/D0sHuEDJgNITZDKknyKO3kNd+lM5GSjtAaUMwQvXax+GTGeIdktsU1JkSug4GN1OnjKjBN7B4BOcKbh4m42ljiVGBN7Dhk9lVPFv6T+a6Pi38TEaWK0Z4ZgIzH3rzbWFUZtzYGjI42KXaPaTcO2Q6XdKsVty4cZeDO6/hY+TrX/oHnB2/iwuO6fET2nrZBy2a5uwYtGZ2dsreuMJklmo4RMIBxigyBe3ZKa5tWNWbJ+6CFFC/PyJGK7Q2oAQnJPqSVqlNWWYQ0SijMUWB9pH9osJojel9CD/x8sfZFc3jd94FJby3qNkqKux4TNs1DCudvA+Dw7tI8BGJGmJH7IWOXm/mvLvGxgd42d17MBhBXZOv5ozGY+b+nFi3LJYdarSNGQhqaxuG+2DLVPbQLjX+xhPnRzALsFhAs4KuJuQFariFaIG8hGyMbO8S8xFUqawbkURyjgGplyAWUeFSOStJYCE6bmyhyvuAVkIQxWK5JM8seZEhJmdvf5d38gzRmiyz1MsVslhR7eygjGZ3f4tBYVhO5wy3SnzoEO8vSoD5jQExWJwPdE6orUargPaenarEdw63WpIXOUolTs3e3hZb2xW6zLh9+BLL6FIlbQMdza3NMFlG2VMCPv2ju2ib0UVQeUlUFmVyZrMZ5cozmUxYDSwYzSKvGIilI2CGGVYbFClEmERoVjUuVd156d4dBtWQoqgS6VwUojWpSVo6t5OCNtVtYoiXvCVg4xj+F4gQDbJ1F+k8vPIDsFAgJ6m8WhkoSqg+BcNjOP11WE4wpSFu76CKQ6huQpinjiFdRyKoK3A+ce+sgU4nvp0FmjkwSgfK6GQyFvzF13lR8NT5FWN/zglVaZk1jtNJQ9el3rNt5+g8jDLDbpFR7Wyzc/sl8u0DTLnF+fmcGCJFNWK5SsIj71rOT484OnpCcJ4yU9iYfO1CiCiVsZhPMOouRkXINTIcoYnormXZtTQTR9NsporWubXI5eK/S0hS4HsPOoAtMnwyuUi9fbUQtSZIKud2XYOSyHioaN2M1WSKuIBrF+xVOfv377KsV/jzI2yI7G6NaV2NxJAqdz4JLFsf8TEQLmrG8UKst6nY+ABPbt1Jg5+aE4ymHI2JCMujY7p6QReGFDrZPMjyCPIhcVBC6GDSwWpBbOaEtkUHT6jnqL3tFBBEoBimWW64lYK70RZSDaBLCiGUIroGokcZg+r5R9802G5oFkNi7yAfwQwKOgks5jPywqMkoLXifDJh1+wwsDlaNNJ5QtdiC8XNmzscnZ8TDvewmUJUiwqBjJyqzGi7yHyxBE8SC0QwOO7sbVPmGmtGnD2Zs6qXlFXOwc0DuljjYmRWT9FlSYiBvNg8Fa02FpvlKISB0gRMylgXFSYraT2IyRgMR5QotruG8/mMwdaIeb1ivlwynZ6jVwsUkdVySZZlVBKZ21my5MhSwrssMvIiB6XxUYjBQ3SIJMuKJLjQBBGc1oS4tjreYHgP8xMoWjh7kn4f7cGjRxBmKQhbFjB4BeYNbN1GiiGiA1QBhgZygXZNH4jQ+XRrNFgLyqbxLQaIOZQ3YbgN+Q7ofoHpQr/Q/Aj3xYeMEC6XDqlhSvLj9J1Hx4DygdmiYVZ3aOAw1xTmkltXjncohmPE9IubGDg/eYeuWwExLVq14Px6PA9psY6F6FPLua5NY5wolNFgNTE3xOGAMN7CLRYpUN9A+N4M+gLPnHsxdiAKk0NAMJnBk/wz0YqoBW0t0QR0F+maFatVkyo00aJaR+w80bV0zQwdWm5sDVEhoJTGSJHEXD7gfUfnPZ3vCCpemLNHNtd/do2ND/DIK0BBGZBqiFq1WO/IhhWha2iaBfr4EZkxZNWQOLbQLMC1xMWEWNeEZY3SFq81enuLaDIkK1Igl1VIWcH2LrEYw6BKg25oUlTT26GI0ZjcIvqyPHXphcfGBnjEJIdX2rB76wbHbz7izcdvs1OOiatI65Z0y8jW3nYqk0Qhtg6KiDbCay8fcvrFt/CrltzmNO2c5XzFsNojephPpyyWLRED3hN8jZGOl25sk9mIX604OXrEbDnjzp2bFKOSJ6fn7B2OGe/tMFksKIy+8M3aJCg0WV7gEJQ2dFHofECLgNZUZZmybUoTck1hhhTnBpNnFDop4QbGEpeatmkoMo01hq1RidvbwlhDsNC0HZlN3BlRoIl9M2+QEFLXit4HMarkdBw3c157Gj7A9gC+8f/CW18FvQ/3Pw1VmaoBXzuF8xVsLcG4voxaw84YOgVulvi/IYAxoHNYzBOvb1iCthBU4t8tzyHXcOtl2B6nKoTvF5JRoJMU8L0gUFf/1Li2a4poILpIcJHCGFSedktuDE4s5dYORTkiswWZNYgShoMcbS2rVbILoleDW6XQRhE0KKMIziHOpetAoMhs4voSCQR0fwiUApsVlMNdGn/20eyg7zLabn2Bx6fzGf0dHSLaWGIAoiDaJLGi1kRtiGIQyRCjCLrFGFJ/5S4gzYJIjZWGoCLKWFTw7AwKxHcEL4QQcc7hvKfzvWo/RCTGJLIRhTEqBZQbjM0P8HQGCGQFMthCFkts9HTLiqxp8fOG5ekEX+T4LsCqw1gD4pGmppvP8N4jOeRZjm8jejRMPLvBIA2w+RDyCsnztJruXPKeUqH3noqI0WhrU6nlwtyYy9tNDfCQVCbqXcs/+eM/yslbjzj66jeYnp7jVEfXON5+/DZNdBzklv3RAOkcWuDmbsXeQNHMp1TDQ2LMKAqDiKFrWxSpY4MxGsERQ0NVGHIVkdDStStWyxnVsGRnb5dAAKNYrBaoRYHRBmtt4oJsGEQ02mTJTkY5otL4zrNqG2arlrLyVNUQJZquaekWDrVs6M6XZAhFiKn0rTW6zMmDIcZACJDlGdpanLbkg76hPbE38Q74Xi0be6/H9f0XpYkFAPUMzB7oLbj7I4CGyTns3QWbJ2uSkYZVB24OxiYBRVn0nF8HbZM8HZyDVQNeoBilMUfrtE1sUql2ewzj6qJPJ971AV4gKci+qWzwYe+RDw3rQskaEpPbrgaGZUZZCHXraV0AVOpy4QUdQHzqkaqVQhvN1njE1vYW58dLRIEQk3lv8IiA1YoAKJMlD8Ig2Kpkd3cbpSVxkCMQuhQARoX4wGhnj3JDVbSrpn3f53WIKN1RlGmaFNGIzYnGIsqidY5giVGSlywGPDhXo3XoRUGSqjZBYTHomNPWzUXiJM2tvdF1TONPIHGB87yvNmzgwv4qNvPsuoqsTByUtkl8l8EALRpTNvi6IXQh8YmWK1zTovWMRjQxS7whLwo7qsiGI0Qy9PYB7O8huUkXrdKpXIJcKmNVMtZEkwZm7wguIDpHqSvq2YugLm5ugCcpUxlCQIiYXHH46h0KNLOdE8ZnM7785d/ldHpMh0Myg7aGsWuxeU4WltzeLfmNt96l2h5DXiLSUTepzyoElAhtU7NaroihYW93zLDKmZ2ecP74FHEtBwc7hDxwsjplsDMmE03WBMR6KmNQGxngKbT9Pp3wAAAgAElEQVQyRBMQpQi9H510ka7uqJsa5wNFUTHIS8RCI4bVagkxErwDY8ndIJU5uhYfHFrrlKnTGjEGUQpRaZL0MSSCdXDEEAjxiogoROTi1N/Q8/0qvIemBnUjKV0zkzL8RQblHTi8B8tV8rAbF/D1L0Jmk+gnL1Kgpmzy3iRC4WC+6B9PATWxBRNSQKgF6EBFqOve8Dh1cVDiN74ifhXf/KcKSjTFoEK0xkVN5mG2rFkskzNCVVoy5VCzM2SxjwzHiIaqqrj3ystMJyeokMyP8f1iRqWWk0kkp1AIUYSd3V12tsfE4PEIwafuIrF1xMahAhjlKba3P+Q9870B5zoWy0helpAZMlGoLCeagiA6VclUhu884nMkRLSOBK2QqFFRAwZFIAahbRxKQV7mBNfb3XQpcxqd4AIXButRJ/9UlEKrzb4oNm9WexbaQOwHtyyH4TaiVhS1oygG1IMpenpO8C3SQWw8OjOoMgdrU3usskTrHMbjZHisFaxWiQMzKNNnGJsGZwmJDN024LvEsWgdoWmTc7fAhfPj1dXDpp5nIsnjL9Bncjx10/Lo5D0U8PEf+zRdCHzlt7/Kql0wX0wYzUc43zIaDMF7DrZy4tee8N6jd9i9+zLKWLSJxMbjugBdZD7vOJ4tWR6/zY994pCh8iweT2gmZ+yOB1T7Q+aqphhXuBC5sbdPu1ymDF6WPX0sNgSJsAyiMqL3aBRBNF5DkUWa9UAoni4GjDEMd7YRm5qmB9fhXYtEjfeOtm3wweGDJxKTcEWvfazSY6IkKWd9SMrFGC+NZp/yJeNCSb6xZsdFBZUFvwcySAvBcS/iyg1IljJ5Z0topkALZgjjbRjtpnNSZRBWKaALIVUIfEfalx5UDtZBBYQsUUNCm7bxNikG8Sno22xP16egsE+JhqOKGKOJyiA6p7Q5ruvwwaMl0vqYqitWU0gNJ2/jtUbt7qEzw/3XXufo8WOO3n6T2LUE54kiSVTh0rgeoycA+XDAK6+9Qp6nYCMER/ABv2pwyxV+scAtl4hSNG39ke2jDwUXHnjPkvDAec9kNiEYhd3ZR5kCySsCCq8MQdlkz9QKEjxRut44XRCt0GJABQRL8JZ20SaHConE6BFJNJ+IQsfUHi1xwgM++r6ysXnj/lVsfoAHPRlIUoBXVdA6VGbxKmC3h6hCEYJHBYO4iLYGKTNi3xwaY/tMnekDt9TnkarsuTGSSEahvcwWtk2a5Fwqk8QuEW8vTqd1QHE14NtESFrRRoAQiN7TtTVYzWg05uz8mJ39Lf7QH/4xjs/Omc8WvPnm19keDrEHtxAPWYSP3RC+9OhNOoSD23exJgUWqwbmdUPTdKym54x15Pb2gGY1I9JhK40uNKq06OjQMWJzS902ZDZ5IYZNLZFLixKdxA2qH+RMRkAICHmI1E2LCwHnOpzrkL6MarTCh7TKjVi0MhTGEmLAdW0K8mJAxPcej8n8x4eABrREOp8yt8F7gk9yNdVbpLRdyya36ANScJdn4IHZmxAHYF6BapSCtujTmKQ8vPNVWE6g3AaTg+Sp7VihQQ1IGQefsnXzMygktSITAwTIe42zj9CGVL4lWXAoYq+mvZhtNx6plMrFuCo6B1LmU7QFpcgyy7YZs6qXGJOxbD1t1xGV4Oop6uhr6G4JOzcpiwE/8EOf4QvLmtPH7xIkdcqIPlyM5aIEU+S89vrrHN64QYiCbx2hbaCpCdMzurNjlOsojUMrCLKZZNTU6UMu181PzW9J+uIQmrZF5gvytqNAp+Oke79YYwGFhEBkidIGyQIutECaVwMRnWmGZkCjEl2nq9tLHnBIlmQmGkKvuJEoEFJXk03HixHghV7htO4aEZKKxvd8oSigbEZejMB7pMjTOWgLYkzN11Mmrk0DZwhQDlPJNytSJk98IlU7T2xX0CwRyaFepQ4X4gnqskS1Zp7K04mNzYMkjvd6f9N2SOvY2dlmmFesJhOGWyUr33Fr7x4SDafvPCbzgeMnbxO9ZXd3h4Ox4XyZcTQ5YkJguLtFFyNOeRa+xXVzRmHOp27vYrsl83nLYFRRDkp0IZwtz3AqcOPWIdPJjLppUFlJlefEvp3WpiH2q9nkvi8pa3ZhCSMYFRmaiuWqxjmPiLCYzzFaIcEljhGBGAVQiBI0yUTae4f3rufWQQge73wqRYWAiiBr3ktvV3DJyesJUrKeCDYUJksUEVPC1EJtIJrE3XUCtUtcovEQuoOkSt67CXadUe6pG2LS/RDSwtHo/vl19KL6TEQqA659B+mDcO89uu/A8KLAZIl7nYI8QZQhIonrpTUoQYkhxIjWFZ2HrWqIMZZHRyf4EDDthGy5ZDg/xW4dsJNXfObTn+GrRcnj996lqZMVRyr1GQbDEfdfeZlX7t3DekecrojNkrA4J6zmiK8ZWtC5QXSRApgNPSZZtnYlkOfdpDHDe9q2ZbFYUK1qsiiJPpUVOFFEpdMwoQ2hV+GvLZhiUKA1VhdIcHjXojKLhIDJLSpoxHnwgnhF1DppjnzixSvlUKHDsZkB9hqbH+BdKFnXRFcPzqVUbi9tjz3xMjVAXpdP+wvvwtYkBYWyfh9jUlZPJ/Um2BTghdATbSMxtNB1hLbDNS2hNzMWkSsZPNlwpXYfHEhq9xaioCMUNkOikFsLAstFQzmwvPfoiHJnyOF4xOxswvnJgsfn7+EE7u4dEJoFj977Gk27TTYc0i4C7XJGmJ9ye6wZ2Q5NktiXWyVOOd47eSeVxKJwdnLC/v4h8+kSjCarKjA6qTs3DMH7fr+vSxGCMgpjDKIMnQ8EHxhUBd6lbhdt7AjOI9GjiHjviah0PYRk0aHX3hNhbRDb/+ocoQ/66IUV3vv0Hj0PzzvfZ/USP2+jM3iDIgV5SsPO63DWQTFIpui2hGyVRFr7AxhnyRR9ME6+MxLAqnS7FkhEn8qOZW/NdOFM6NPxif1Y5VwahzyJKxY9WoXEZl9jg3c7gC0qnkobhQCiEOmDO2MQSQFgi0eriFI+eVDv7mG0ZTY5xTnPYj5BLWcobRiVQ37k/iGT3SHn8yVt16GUpsoztoclwyojm72LmgVcWxNDQLyjshZdbmHzDFEK71pC122sirPI8n7v9//36dTLxyK6aXAiNG3H4nzCzk2H9gEvQtQGjaarG4LWUJWERhFjMktHC9E5vHOpDWmIyRfS+bSQ7O2ZrFZE5zAqjUmBVG1QIdnXbGqDgTU2P8BL5K/Ln5A8oS46SfQTjIjqTVr7l11M+OmBGHuORujLvVr3dYB1sNYLK2IvxVYaXMpueOf6ElVvj/LM6LrJY21ccw37lVcQBdqgjUdi6rYQgyM3BsEzn5+z/+oeX/3673D/7l3u7u5ydnzG8fER4mYMc49dzKnnK9omp552hLOaHeO5UVr2twbs3txh63DMZDmBQgiVIRAYVEO088SmYVhV6LLqDTVThmvT0LkOtEKppFJDQLT0Vj0BkYgQe7sH8E6l7FB/nSgRvAv4rksWJzEAnkhIHJfgUcEnZbIPqF5Q4V0ghOQtmQK72BuOerTW/Xop9nqkDe5HaxMHjiyD8Q4MJN2qDFDJ9kTVqVVZlSdOnkgK5Ch60YQijRD9YnKw7m2VSoQpinPp+ehTLy4fk19e0GnB6jx0DZs90jwNvW4JCYkmEhwgiNIIOl0HaSVPng8QpejqJdo3lJXB5JphvsN00eBCcjpomobl5JwsyxjaiB1asmIMMVCVBV3TEpuafFDhvSOzOVoLmUAILmXDXd1nrzpAJyHTBmJQDq7s/4v/Ln51ztF1XU8hCSznC9p6hRlFuq4jLyukE7wI3nsyY4htUuaKSh0wQp8FTKXg/kdd3k/j1LpisP5wuaC/b+y4cwWbH+Cts3ehv/V9gHdB+k7hlvQu+2mi74O4mMpI9JNR/0taPSvdB4GyPlsuVbQxnXTrz36WYP5CQSVCrMJAluPbJpUmAsQupmxmTEFG1zSURcYbX/9dJpMzxntjPvaJl6j2tji8f5Ozowl5uWBpFzyeTAiNYxhhNMq4tT3gtft7HN7dx1YDZs2c83ZBmecMtsbpmPhImRfUyxXD7dRCyBEQnYi7m4bONQSJ2KIgz1MmKfaxAj5ijU58PB+AQJFZnCT1rO8CwXd0bZtsaKRXCkaPEEGSKlp6lWZSavaKQiJd8EgUMgy+X//4mLh/SdG4mWXxp1CVSSVLSNk8RxJioa6IelJmO6Xxk/o4xXOxfyyQOre3IH3DWImkobsXWlzkUHv03ptEwDmcc9iepvKiwBjzVFkw+H5/9mN8CgLSJK9sThCdvEq9R0JLbM9pukBoHYPcsGw9wbcUEsklsmocsetYdA1lnqNsJLYNmRbcoiY6R55pumWTeNzEC89J9JV5Y0MNIbV+f0WP9Au79WWwHkPkypypVCrLep9oJf7qePHM2PHsmS3P2UieeuR5W2weXpwAb81nCaknYezLGcmbS/oAL15m7kSlLMPlG6Uyl+sHCq0uy7hXsz996YqY/O5iCM+8z4uFuM4aoRCThCrJyNIDqVwnEYzSLOZzrAizpuGzP/ZZsiLj+OQ9jo8n3Lh1m717Bxzeu8Wt1U1m9QrxHhM0gkURGI4ypLC89fgMFzoa8QxyS9c15HnGYFAR50sGwwFZniWhjU0DrmygCWzdrFDeoYscR0CrdE4HekWhDyhJnLoQBaVSezEvkeigbhvq1eryeokRSK3nlKSsdvApUwfrTJ3vc3zpeUIiNYcQ+h+PD5tdFrnAVSqG6oVY8JxZ5YL9zQVR8iJmW48d8Zntn/uBz31vefblLwDUM/GFFpOy0PTVmousvUJplSx8RJCLcanDAvlAJz6eeIbGYbWwqhcYHxhmCucdJraY1mCdR8VUPlT9wsfQq8olxfHrao/0HEvZ0P7A78etjevEyVObxOfOk6lvNTz1gm8SbDzn7Z7/rb7tFpuGFyfAiz3pOPhE/l5n5fqTTZS6FFRcnFCRi0zcFSSSrroSDF5ZQa8nwv69Y7zsu/kiIq4DAUgKtixZdkTnEBcun/cBqwxWDC/feQmLpqtb3n7nXR49OaYNntxYhoMBb/7OV3nptVe4ceuAxWxONMLZdEExqvjqV7/C3Vc+RbNcMt6pWK5mKKUo8pzMGEKeYfIMnRmCIgV3SlDfZsX5/Yi1MpaloVRgpcAoTYg+ZVGjT85dIniSujVV0HuX9xiS5RpA5y9KHZIILr0FCr14IgWOIfRBX0wTpo/JLgURQkycvhDjhfDmGnAxzsD7z0EXmz1vzz3nhfK8gO/FgFFrIUrC5RC8DgiEGH0SHAXfPxpQKtlq4EGJJwaHipHKKoIyeB+wSrCZgRAwKlEgCIHcKIiCYh08ygXHTp45ZkqlqffFvQYuj0M6peNF0uVi/vxWr7ua+r+YptNiSp4b6T2T4X5qg82+QF6AAC9cBng+cYXiMyfTN2XwnlosPOdMWwd3Fxk8dRksxni58E7ywo0ncr4fIpEgKeODVojN0DEmIn8AWfMTAxhtMWLIlCU2iVDeLmpeunULpYTlcs5v/+MvslMWbI0GuLZm2S6p6wXjg0NOFwvuvPoyZ6dHDKsBw0FFVWrapibPcghCMRigjbkI7ESRODobOdRGXNcxm80wRUFeFj1V1IOk7iqhJ+GT9c3sdUAJiTuUGfA5rqlT5VCujKW9CnZ9rqeFTK+WJQ3cbt3o26WsXUysmJQ1V6lTgFJqg5W0H8Tf9Zxa1HPf9n3O3xeQIqKUuhyHL1g0ybIpTfd931i5ZGLFGJLfWkhzgkL1nmoRFQK+bZEI1vQGub3KMxIJLkAUTKZTaVEuuWDpX7wMQpRO801ftn0R0Ye/FzF4KnxdDezSHVlbbK2Jc8++ydN3nr6/5v6tk+dX1lCbOuI8i80P8EIvrljbDPR1/pQOjn3WoufAhHAxack6c3dF7ZoGSrnk363Nlp7N4MFT/LuNVgp+G4RedRn7qypahcKi8zwR8NuWaHTqQesC1lp855AsQ4tQ2YK90RaPT47wnefOjZuMByVWWVbzBeIhtyXLswXlaIvRVkWVV8xOz3G1Z2dvm7rJ8M5joqbMM0IISTmrLaIzUCaVVjYMEoXgPF03o6tKQmbQNp3riZzcO7tLsj1RfcInhJSJy6xFxUgdUkbOX+GupgH5MqBLCfL1uZ6uh9Bnr30I+OCTqazW+BgwNsdmtn/ti3J9vF8Q9pxfrgZz67LW7xtXZtAXCBfJS7l8IJX7UpB3ebWvB6bYZ+9M0ibbvFd7uxTAxYDJNDF6zDOBRgwQVTpYWmlEqz54iairhnwx9Kr2tLBhQ8Vd3xGeMshbz5VX5s6n8LxVzbcXZ8kzt89/drOx+QHe1QzeFQVtvJJtW6+21rjI5sFliuKqMfG6RLvm2Ij0dir0VhLqOsBbI66zYz3HRSX+ou4DLdV1RNeXQrRGZzY1BPeeEISdwZgcBXXN7cOb1HVN0OBCRCuLeMfWeMzx0SnDnYxuuWK4NcRFn46FyciioAwoHxMPTUUwJpUptUneWJuYRYoQfKCua+plzWgUwKbB0nuHUgaRxEEKPZ8O0qpZKZXakMVIyDJc7xcZ/NqSI154SF7lyKSOZBHfc8mipCDfh970uLcl6rwjdClIfGGujw/lz9zA8/ifANrYpzI2F0zGKJdUnPRIf/72Y5QSAipp6QTQlqgioiJKR7jSlSWux3m9DhLTtSNao/q00dqK5fKwxDTeqLV11Aty7j8HTzEIer76+kjFi23W2YE1b0/6rOlz3/GZW77t5bDpV8sLEOBd5eAlI+LLOv868EhloovS/rMrtEhaaaU7KYN3EdxxOYL0PL/1au06wONi8bW+bKOSREI2BmUzdNYRW4d3IXnRqdS30TmHaEOWZYTOUdqMwmiU0UStMUqotcLFiLQN41ITVudQWHSesXN4gG8btNIYbcB5pDe2VioZXwaV1JxyNQO7SUj1Z7zzNHVD13RY69DWXHDkIGXQBGBdWlIa8R4XHCKCsbY3LV5nvBXe+8uAfa3YRC7+rS+3dZbP943ZW9eCCM57nHux+qMmvN959u12xj/BOSps7vn9frjS8zvFbynQUkoTL4qmiYstKi1wFBkhOBSJI5qCM0UMjqjox3MuaTjrqk5M78+6HKuTwO5ieugpIFd53hc2HS9qBg9Yl08hhcsxhIs4jv4YrJ+72P7KS58+oy929rOPPn3/6nz9AkBe6ODjGte4xjWucY1rXGMDsXnEo2tc4xrXuMY1rnGNFxzXAd41rnGNa1zjGte4xobhOsC7xjWucY1rXOMa19gwXAd417jGNa5xjWtc4xobhusA7xrXuMY1rnGNa1xjw3Ad4F3jGte4xjWucY1rbBg2OsCTh/KyPJQoD8X0v/+aPJQ/8x2+9mJbeSj/ujyU/+19tv0peShf+WC+9Wbget9/NJCH8nl5KH/5e+V9rnGN50Eeyn8pD+UvfASf+7/IQ/nT38F27zuuyEN5SR7KXB7Kd62JtTyU35KH8rnvcNs35KH88e9w2+t9/+2/w3dl33/Y+L4wOpaH8rPAnwM+CcyALwC/FB/E/+vD+Pz4IP5t4G9f+T4ReD0+iL/bP/9/Ap/4/b6vPJTPAz8LtFce/r34IP7IH+gLf4C43vcfHOShvAH8mfgg/u9XHvu5/rE/8t363O83yEP5ReDfAQ6Ac+D/jg/iv9o/92vA34oP4t/46L7hB4/+3LgB+CsPfz4+iH/2o/lGCd+t6z8+iD//AXy994U8lL8IfCw+iP/Glc/9F76T1z47rjx77cYH8U1g+AF8x18DfhJwQA38PeDfjw/iI+CXgP9MHsoHPfb+u8BSHsovXHnsL8UH8a/8Ad/3At/v+z4+iD/4B33/b/GZEVjytNXyB7rvr+J7PsCTh/LngP8Y+HngV0gT8j8P/AngQwkyvsv4K/FB/PMf9Zd4Hq73/fc25KHo+CD6b7/l9w/6Ff6/Cfzx+CD+njyUm8DPfMRf68PCv3Q1+P+o8QJc/98r+LPxQfwb8lB2gf8O+GV5KP+A7+6+/5H1IvkFxzfte+BPfZc/80Pb99/TAZ48lC3gLwH/VnwQ//srT/2P/Q/yUBTwH5FW/NvArwI/Hx/E02/z3n+RKysMeSgvA18HbHwQ3TPb/hx9lkUeyt/rH/5iH43/28BjUlbhbr/9LwD/ATAG3gX+vfgg/urv82//n4D/NT6If+3KY78BPIgP4v/Qr+r+GvBZ4Aj4C/FB/Du/n8/4Np9/ve8/gn0vD+VTwF8HPgO8A/wn8UH8u/1znwdWwH3gjwJ/Qh7KCfBfA68D/zPPNOGRh/IvAn8ZeBn4bdLx+Y1v91nPvMfP8UyW8Womtf9eS+AV4KeALwL/CmmC+tOkY/SvxQfx17+DXfDjwK/EB/H3AOKD+B7wX/Wf+Uv9+/+kPJS/Cnye1Hiojg/if3jlu/1d4P+ID+Ivy0O5TTpWPw3MgV+OD+J/8R18j48c8lAy4D3gj8YH8R/3jx0CbwD344N49H7H9w/42d/J9Z8D/ynwJ/vn/g7wC/FBbPry1t8iTZi/QMpM/mJ8EP+b/rWfB96OD+Kf/w7Pr0X/N/50/3f+7PockYfyg8BfJV2PHfCfA/8I+EVA5KH8y/TZ+XUGGPibpPPyj8QH8Tf79zkA3vz/2XvzKNuzq77vs885v/EOVa/e0O+9njSgqQFJYObExiQRCivGJsTBAVvAIpjgAQc7gOOg5aLwwBSIYRkCy46zhEFMJgJBMBLCYDAIhBAiEohBUqv7db+h31hVd/gN55ydP87vVtV76m5et7rVrav3XeuuW3Xv7/5+9+7fGb5n7+/eh9S/HmAYV2RH/i1wH/BzsiNhsMtPcmTMGn7DPyZ5na8Arx+iD7cN3dZrsiM/Dfwd4C+R+tSebut8GDMfABrZkX3gHMmj9xeHj4+BbPgdXw78G+DHgc+/1fYD/j7wd24ZX08DHwSqI7Z/M/Aa4FdIfe8yEIHjwDuA7yLd33Wx/d8avs+HBru87ajtgf92+J5fodv6zlvPM4ypvzDY+8du9/pHbH+vbuvV4bVPJZH7s7qtvezIVwHfCJwm2f5rdFsferLzPt81eJ8NlMCbnuSYrwO+iDThnQWuA9//bH0h3da/MPz5Kt3WsW7rTxx9X3bkZcDfBT5dt3UCvJY0ID9VvAE4cG/LjrwKuBv4f2VHRsAvAW8ETpFWHD8gO/LA07jOE+GO7Q/P+1GxvexIRpo83zqc++uAHx1+1wpfRgrdTEid/GdIA+YW8FMkYrU636eQBvr/iTQg/xDwZtmR4jav9VTwJcDrgRNAC7ydNNCfIK2Mv+c2z/NbwJfLjnyj7MinHdXZ6LZ+M/DrpFX3eAhfvgH40mGxgezICeC/At44vPZzJMJ5N/BfAl8vO/Lap/kbP6rQbe1Ik/TfOPLylwK/PJC7J7y/z8Dlb6f/fzMpxPVq4FXAZ5DawAqngQ2S7f9H4PtlR449ze/zPwA7wDHg/aQ+gOzIBHgb8IukMegTSPb5ReCfAz8xtJWbpBe6rS3w/5DsucKXAP9Rt/WxW459HWlS/8LhXDeF04Yx4fuALxjGnc8hka+nhKHt/nfABZLtF7cc8pdJ7WETuAZ8IYe2z4G/PpCC7yItfD7IU7D9sJj61VtefvHw/C3Al5NI1J8Mz79NIjPrZPsnWoQetf2bgX/5OOdYEbKveyrkDm6y/Zccefl1wI8P5O6vkEjzF5Ns/+vAn3mN5zvBOw5cudWrcwu+Fvhm3dZHhobzLcBflUHc/xwgAAXwgOxIptv6odVK8wnwDbIjN4483jC8/mbgpbIjLxn+fx2pw3Sk1d2HdFv/b91WP3hGfhr475/B33HH9s+O7X/m6DWBHzjy3meRVuLfrtva6bb+B+DnuXkg/Fnd1t/QbY2kwT0D/oVua6/b+u+A3zly7NcAP6Tb+tu6rUG39Q0k8vVZt3mtp4I36bb+rm5rQyIFjW7rD2sKIf8E8Cm3cxLd1h8hkc3XAv8ReExu1grdevw7gF0SeYNEBH5Vt/USyRt4Urf1W4ff+EHgX/Hsh2CeLn7mlvb4NzkksKt9019HIvTw5Pf3I8Xt9P+/TtIPPabbeplEwF535P1+eL/Xbf0Fkgf16S4g3qTb+o7h+/woqe1D6o8XdVu/W7e10W3d12397ds85xu5uS182fDa00EEPkl2pNKk4fqDp/DZ7xvGgt8nkbufInmibsV/0m39haFP3QfkR2x/g2SLNwPfRgrpPpntv/rI+PNZRxY9qzGQYXH1IuCtQz/7GlK04EXDffgjUhTw362R7f/BExx31Pb/lkSqj+LPk2z/5bqtP/9nXPNdt/Tzo7ZfRbUsaSxe9fWvBb5Nt/V9g+3/OfBq2ZH7n+xCz+sQLXAVOCE74p5koLkfeJPsSDzyWiAJlj/qGNzaX08iO58oO/IW4B/otp5/go/87/o4OjDd1kZ25CeAvyE7skO62X91ePt+4DOHhrmC47AxPBO4Y/tnx/ZfpI+TZDH8exY4N5C3FR4ircJXOHfk77PAo7qtesvxK9wPfIXsyNcdeS0fPhdv41pPBZeO/L18nP9vWxStQ2LN4GX8ouHvd+u2vuUJPrIaGH9peP7e4fX7gbO33CtLWv0+H3FT21hBdmQB/EXZkQskL8kqjP5k9/cjxe30/7Pc3N4euuXaV2/57IKnL46/+ATnuRd4skXck+FXgFp25DNJ7fXVPLnH8nGhKYT614BvAP4v2ZHfAP4X3dY/us1T/D09kjQkO/Jfkzzfj9xy3FEbnALyI/cnkLylbyV5sP4s2/9r3dajIdpV3/rZ4fkeUlvrgPcMr91PCgkXQ58qSJKQu7m5HdwOnpe2fxLc2v7KW/rG15I8kL96G+f6VH18Dd7PAj8oO/JCEhnfHYg1JNt/r+zIdx85XvgzbP98J3hvJ61Iv4gU5nk8nAO+Srf1N259Q5K264kwB+oj/59+el/xw6Hb+kZSiDoH7rUAACAASURBVGhKCpt8BzevbG8XbyARh/8ELHRb3z68fo7UmF7zTHzfJ8Ad23/0bX8euFd2xBwhXveRQiIrHCVzF4C7ZUfkCMm7j8MJ7xwp6+6f3Xoh2ZE/fxvXWuGm+zXoRZ516Lb2wE8NHrxPIoU/9HEO/RHgvUMo/RWksDWk3/+gbutLHuczH0tYEdiLDN6S4fUnvL/PAG6n/58nTTwrj8l9w2tPFR9J+zrHE3tkH6+tHL65rUF25CdJC7hLwM/rtu4/zXO9BXiL7EhF0kT+K5JX5+lgZfv6SY65RLL3Cpakj7uPpOG7Xdw6Fm8Oz3+FRCKPkudzpNDqV2rSx30p8I26rb/5OOf9WLX908XXAv9QduT/0G39+0/nBINj4SdJff3l3Ow0WPX1p6QtfF4TPN3WXdmRf0zSD3jS6qQnaWw+T7f1m4AfBP6Z7MhX6LY+JEms+Tm6rT/7xGcG0irnH8qO3EcK8fyjp/DVLpFc1x/GwgcN093Ab5BEmUtS53vK0G19++Ad+25uvtk/D3y77MjrSLoASCugmW7r+57OtR7n2nds/9G3/W+TVoffNKzU/jOSzubTn+D4t5NS/P+e7MgPDMd+Bml1DGmge5PsyNtIer2atAL/tad4rd8neURfTQrLfMtH8iMliebRbf3Kx3nvK0kT1a+RJp/XAp84fF84vP8H0G19RHbkd0j36ad1W5fDW+8A9geC+H0kb8QrgEq39Wgo+/mOHyHdg31uXqw84f19ksnytnCb/f/HgNcPtleS0P1HnsblPpL29fPA9wye+/+T5MF8YAgVXgJec8si5la8kbQguErSFD4RPqzdrSA7chcpLP420pgzI3nIjyaQvVC39UO384OO2P47gc+RHflNkpzqrOzIdw62fzNJb3xiuD+bJCL+d0l9ZXo71+KJbf/FJF3vLx459gdJWtoVb/g1kv54bWz/EWCflOX8y7Ij367b+r8+zfP88PA4RdLcrfCDwD8ZIhl/ICkJ6vN1W3/qyU72fNfgodv63aS4+OtJA/85UiNerdK/l9TY3yopu+i3gM+8jfP+Ekkb9P8Bv0saKG4X3wK8YYiff8kt7xXAt5M0FBdJN+rJCMw3SSrauHrcqr34YeCTOTJwDoP355NWrueH63zHcO1nDHds/9G1vSaN3xcCXzD8hh8gaToeN9wwHP/FwFeSRNd/jbTCXr3/TlKG878kJcC8fzj2KV1Lt/VPSNlrbwP+lI+8TMO9JBL+eNgjDWwPk3RF3wn8LT2s/fW9JJ3nddmRo9mwbyDdqwMyPuhl/hKJgD9I+p3/mhTKej7i525pj28C0G09R0pYUY6El5/s/j4TuI3+/0+Bd5L68XuG7/iUi2N/JO1r6I+vIbXli8PnP294ezX5XZUdedcTfP63SQuJs8C/f5JLfRuJzN6QHfmGW94zJDudJ/XDz2XIxiS19YdIWeq3jcH210g6x8vD+V/Goe1XIv/fI9m+A35Ut/UGKVxZyo78kye5xN+UHZmR7hmkxeJR2+vw3vzId3oTiZQdlx3ZI433v8Ka2f7pYrD9a4Av+DNs//u39PN/ceQcv0EiqO/SIxmyg+2/A/jxwfbvJY3dTwpRfVLv5x08x5CU8v41eqcQ7kcdd2z/zENS6Y/fB145hGCfqfP+BRIRv/8WTeJaQHbk3wDnH08z+rEI2ZEfBt6v2/qtz/V3eTYhO/J64LJu6w89199lhduxvezIfwDeeJv6tOclno+2vx08k7Z/XodoP94hO1IDf5ubMy3v4KOAO7Z/djB4Dl/xTJ5TUjLG/0wSjq8juXsByVN7W5nIz3dIyrJ/GSkpZq2h2/q82u7vdmwvO/LpwKeSdHgfs3i+2f528Ezb/nkfov14haTU6csk/cHTTR+/g6eBO7b/2IGkwqI3gDOkYrdrhSHU817gu3RbH3yuv88zhIuke/bTz/UX+TjEk9peUqmotwFf/5HqOO/gqeHZsP2dEO0d3MEd3MEd3MEd3MGa4Y4H7w7u4A7u4A7u4A7uYM2w9hq8X/qOr1YjYAgU1rFR1/gu0PY9x46dICw9YdEyHo2ZnthivLGFqWp29/fZvPcMBA9NhxiDjDfAOAgRqgrIIPYgHlShFyhHUBpSRYEIxoCpoFcIK2+pDpV9jnhPBbj/VcKa4TNeekotLfduZVQibBUVL7v7DPef2GI6mTCdjplMxkwmE6qypKwrTD2BvICygqKEPE82zCrIMshcehYLzoHLIatBQ7Ijmu5L16XjfEi27z3pgNVDjzwDmy9aK/v/oy/5NI1RAYO1FrGGIjNYExjXY4q8wodIVY0IMeBDACPECFkEoyAi1KOK0gntcs4iGCR6vO+xNsOoEFEwluWypfEdWVmixrG7N6OJnlmzIC8K+hhomjlFZhG52dTf92O/s1a2B3jLzldpWZacefkruHbhItcfepje99TjGo9y6sw99EZZND2v+OzP4dTZU+wverIYISz53V/6FbZOnOL8ww+BRpq2o6wqfNfSdi3nz13krlOnuHH1KllesLd7A9XIdDxis644eWKL3dkS7VuiKFubx/Ah0vU9dV3e9F0/53/7obWy//d89V9OnVr1YJQNITCb7XPt6jWMZIhkXLx0EWstzjpijEwmE0A5tlHjDOTOojFSlTU+BELwtE2DiKEsKmazBVVVIAam0ym97ymdocgsVZFzfOsYqspoNGbV5L33ZFmG71OO0ee+/gfXyvYA3/q3v+y2QoOqCqIYI6hGFCWKHtgn+P5gHDKSI2IRI+jwWessag0ighiHiiGEiBiH8QaMgCjReyQERCOeJn2WNPq//vt/fO3sv8LaEzxYTd9CVCXGiDWWGFtUFbGCoqgqvvdoiGnyUYjeY6xBBUQVYjysqqZxIBMDWTCkxhQDeE2zY2YSOQmA6PCQREyMGR5H/l5DiEBmLWARFLGWEOPQIVPHNCYREOccYgYDq6ZH20IIkFvIyvRaCCAm2VOPPITDexCGh0aIAtjhsx6sSV/sJqL9RKWaPnbR9CENdhiK0hKDgjEUJmPZdLSdxzmHdRlZlpMZgw+KsxY72NFYSxciXdeTu4wyz1HfY6zFuox+2Q5t2iS6bAwRxfseMYIo5EWBtZY+hgNi9/EgDcmyDCOGGw8+SN8uyasCGmi7lqKuKeuS6WjMmbpidv4ccf8GfduQVzXtYs7Ju8/SNi3VxoTY9ShC6Dr6rqNrO0ajEd57xFqWiwVZ5tBhjFNgf75AYwQjGEnjnzEWa8Pa29/3PYikcWZoc77v6bse5xzGOOazJUWepzFeIbMO62waRoykZi2wYmZyZC0Iqa2HoU2LpM8YMemggw8DyPBnIjG3Lm7WEhoYJsjDZzny903HKqvmqKqH9lE9sPdNnxLh4AOrv8XcdL2jV765pa93u78VHxcEbwXVSNBIYYaBMASMsSCCxojve6JGRBVrDKHrseMqNSJjBwJhILOJuOX10AjjYbA7euiaRExMORA3B6M8PRtzMCEOowKwvgQvxEieCyYYjEu/WUzqjMmcFmMM5uD3a5qUQkCWi+QxzQqQDvoW/OCViz1EBZdBlkOzTDY1wwBr033F+3ScGAh9InhqDu/b0eenuw368xRBHB5P3/fcmM9BPVYMpStx1uEyKApLXTbkeU6RF1ibo+rpAWsNRZaBsdi8AGMIfU8Ui+QOH5UglqZbkuXCsu0IEtEYMTZDAZdlGJOzbBuyImfRtOzOu2FFfmTiXEOUdYUVCxoJfU/XLIhRcEXJZLqBNUpVZxRVgWrEdz11NaLcOs74rrM88od/iKsKyr5m3u9Tjyc0iwU+RKJ48rLm8pXLaAi0bUeeOZw1LJcN0XvKZUNVFuSZQ0Ngb3eXohoRo0cE+r4fJtTn2lLPPJIjP40lUdPY37YtaPIOdW0P6ilyR5ZlZNZRliVlWVFWJXkmiAaMpHkizwrm8wV5nhFiGIgcN5E1kbRgjQOJs0fGdOHQ65SI3kfVHB91RN8iwxyXbCQkTwkcRlBS348xoiE5TGII4AwxJGp2SKrltoixroIyyEGARg8W8Xp4zMcJPg4I3uChGzx4IUYkS23Ah9TBMQYfPL7v6boWFotEOhTUB2Q8ApslspHnMBpDPYKiAimTh8734OfgbwAC4y0osvR3NQHJgeEcxgLmyKrjsMGvGwShzA0ExWYZXe+JkO7F4C211iEyOMyVRJD7Hs0LsALSw+w6LPfAd6hapB4hozG4Aoo8kbYQh9CtSc928AYGn0jcUa+eX3n3BnK3fg48ZvMl1hpijHSdcuXigquXb2BFyApDURgm44ytzTGbG1OKwuGcIbM5ZZmTuYyuscSoFNWYajTBOkfTRVyeoSbSiWfe9bT7C5zNUDGJJCJ43xOM4AX29ue0vqfzkcWyHVbtutaepNF4gm8aFrMFlx+7TNd3iMmZZjmTzU0mG1OaRUOZ10xPnMKMJzTzBfmx0/iuY/P0Wchzlnv7LP/4j7lx9QrNcslivqBpGlBBjaVZtnRtx7JpKfIMi9K2HUEVHyKigbIohgUuFEWBYFCV4R4815Z65jHfnx0sHMUYVCPNYonLHFYVo5HcGfI8J89zKlcwHo+p6xqxgs0MGnzyRKlixNA2CwDGowoRQ/BKXVWoRjKXHXhPrdE0v/hEoI1Jc9AKMshIEu9YQ+MDzd4uiEle/OEeGGOHqE16feVYUY0IkbS4D4QO1DqMmOQDGY5DLN53GHUr12oK64o5Mnveak8BYnKq6srqh25BWU/zH+DjgOCRVg4GQIkxPYrVgBcDMoTsoioqgilK6tEIU5eoFaQqAYO6DPIKKUooCnAlyDiREBcgZnB9Hzod9GBZInlik0aMPP394V+QQz3YesGSOmY0CtYhIUJIq1xjV8Q2EV1jDH3bDmRcsF2PzXJi0xJ9i41dCvFqhrgcd/IUkuVoXYFkyTNoLBT1oVtIdSDf/RCJjcmr54/sw70K8a4Z9q7dIKjSdp5rV1rmy46yKpLXqLA4SfKDy1dn7O03lIUjzy1GhKrOGVUVo7pOLbNbkrcz6mpMjMIomyLG0sRIVtVE2xECLBdzSjFEVXrv6TSw7D1d73F5jnE9zmbDyHpLCGfNYESYLxZcuXyFR85fxEfhBZ/wYo6dPIHLLL7raOcL5sZQ5yXFVkW1eQY7vQsTAhMpid2ceQOtj6jLyMcZnRrmu3vU1YjZ4gYqSnQGIuwtFlhAg6dpG5x1FJljY6LIYHPvPb335EWBkfWcAvZ2dwd9V1pkiklkrsgy9pcNRWaZ1CUiQlmUbG0eY2O6gbUWHwcBqkb6dknwHmtgczpmPp/jnMUYS28izlREDVhncUZQZyicQ0Toug7vPUWRoYM+2Nq0YFJVrLVrOe4A+Nl8IHYpWiWraNXKo2dkIHok8qZpwR2CR6PgjcVYAwrGGEKMuDxJeNQlp4xxjqAOsRnRWFQ8KhbEEAGLRT8sSHuEaN+qg19DrGfvPgJBhkYF1qQb3nQdWZFjjaUoCmw5oswK8rJmfPouQqeIcYjNkY1JIgdZDj4Sr9/A3r+ZPHEUHMRmRcCWsHkfVF0KJTpJ4VwZwrMoyd13ZBUxfMuE9QvT5sYQvdKJ0vQ9VZYPWkhFWK3m0ko7hECzXKAixCjEGDExoH0PAnlZkxuHFZM8cft76GQCe93gaR2nBAwboZOUmBEH7eSK3HVtSr7wXfo/xsMQ7ZrBi6FpOq5eXWBNzqm7NigrS16mRYZRsGIxZoQrkwfVq+D7juuP7WN1xnRcYx3kRU5eFVT1PtZaZrN98qImGiD0iHUoEZtn+BgI3qNEBE1CdTLEGuaNAC7JHw9svn62hxQCvXz+In/wB+9jb95y1z33kFdV0iSKYba3x+bWFvlkE2tyrC0wWYWGQeNYlvjFLuI7Tt19DzeuXOLihYs8cv48VhzXdndZtC0xQvA9sfdkeU7XdWTW0aoQfMRmsLu/oK4rEI9iME7xi+ZISGu9sLExPWhWq3bmnKNtW6bTKaMiI8ssZVGSFzlFWVJXJapKjhBCQKOncDVds6BZLMitw44qxAghRDLjaPFpnhCDswbBYe0QnpXUJ9IjDI4GQ55neO9xLkOfcJvWj23ML1/DOoux6SE2jfMHzyupEtD3XdLA+56ua/A+kV+RITnMGIL3ZLmjKAtcnoExw/kdxuYYmyEuJVmo2JRwYTMElyLDokRNxD15BHVNR52bsf4ET83Qr+xBPD+iSe+yOWIynaIqjMYTbJZjsjytLooSqvGQEBGH8KxBbJ5Cf6bgwHzCEG61iQhmpAvpQCLsQASFRCYej8gd1fGtEfLCplWZy2haj+bF4KFb6WOTZiuEQLts2d/fRURwtiTGCKFlMq7puh4/aIYym5ICNHikWaB9j9jU6SnyIXGlhMYcinAPvHcdtPtJI6kpLJBmufXr7iYr2b86p6gqqspQlzkKHD9+CkTomobF3gwl4mxGxOBcTl4WZNWI0HUsQyS0Lc21q0iMnDxxjKrOqOqWvGxwRZaSxbMCsY4+hCSJCGEYtD2BRNZDCGRu8LTq0QF2/WwPcPnCBd797vfyoUcucfree5ke2yCzhlyU/evXyQrHbDan35tTbXaYjU3yvCLO95hdfASalkvnHqYPnmXT0DQNqsnjvWx7rl+/MSTRBESgWS6RZTMkyQAh4iR57OqqIM5TxmdUUCNUZZmSktYQTbO8OclCIYSeqsipyoxxkVMUBXlV4oqcrCjIyyJFcXwkV4dGh4Ye9Y6YGaw1kBm6rjsIPGR1QSBJfNCIasDaHOscVlIG7sprl2UpgSPG5PGzVpA19aAu58thDEjauiTFsRhnsUMyi4rQ9j0+JodKjBEferrQk+cFRV7QaiBGpWs7YlCqukQ1YjNLXuZYI5RFTpbnZHmBsZaIYMsc8hz1BnEZqGAwB147YSURWU+CvcJ6tq4j8F5RSSGLIJBXFc4IQsAag+97irJGMosbpzIcZjRKpTecgz4kvV2eI9UYceVA5iCVQgEOEq65OTtT8kQOV++ppiQMjYfZm0oiJsrgFVwvGAv9sqfOaoL3GDFpRWfM0M2SDqiZLbh88VGads7dZ+9hNJ1iixIUsrKkxhDbJcvlgqYPtP2CzAeyvicb1QBo2yJFPmjrBr3dSodnBFqfEjX6JnllgUP94/qx6wf/+FGqScnxUyN8l8hxJHL56hWWy57lfJlIV55Rtj1ZVVKPHN6n7NvRuOb45hTfdNT5iPn+nKuXLnL56mVGxR7TY2PqSYXLc2pr6bqe3vf0fY+Q9EhBI20fEGdRAVQJR8Ljq+zCdcTv/e67+MOHLuHKmnrrGMeOH+PY5gZic0wpTDc2ESOUG5sEDJ0PxMUck43IR1MuXvkApsiYX97l+vXrxBh47MIFnLVcm+0xW3RUVcl8vhjsDRo9y66jyDKMKgEhz5TZfEld5IMemcELFdbV9FhrDibwGAGULM+oq4LpZMQoL5In1TlM5sjzDCUl16lTQh8IGhBjyEY1Yg2x9xADzlrECd7Hg7CrSEro0Bjouo4yzzGkrGUQ+q4ncxk+eqy1FM5ijKzr2obNMy9AY0hJLsETgx+8zB1972mXLQB5UVLmFWoyxGZ4DSz8gjwvqOsRMSpN0xK0wUZBbEbft3RNpGl7VDuIPXmepdC5CPUoZahnZYHLc3BZijAYS5QhaiRpUb/OGmD4eCB4fQSb4v1WBI3gigLVjs57xHtyY8nrCSYrIRshrkjZmfUozfuqqUTHTbXThgQJTdH+m1PAj6TIH2jBVq+7gVPcklixpgOtYOijoQ0BEztE2zQZ6WGgOsTIYxcvcuXSJe65716qyTHqU3dhNo6BB2yNqKKhJet7+t3raLdM55dhBWaHzNi+h9FoyFAmcW9j0usaUzKGKRLhuylKvn43IK9yqsqgoaXvI60aZoslMQpt29P7iPeKsQGNC/LcUNdVSkkyhum0ol/2KcQ6EcbHNzl591kuPvoIj37gQ9yYXWE6LRhPSpbNgiwfo6r0fUeqvwchBlQNzbIDI9js8cUJ64h3vOdBelfy8gdewpn7zuDKEsky6skYawSxOUVZ0e/PyfKcxYc+QDSG3VlLNq5YLOdcvXCJdrZP13cs5w1FntM0+yxmM/rQM7+2TBn/MZBZQ98GrDV0XjESwDhuzBeUzrAq7WSNJYZIPao/rPTHuqAs8mH9nAYaMcJ0PGJjOqGuKoo8x4nBicE6l7w+1qZcK7EIgnUpQSmqxWGI9PimwTkZNHRDuRmJZFZRUUyW5gSjfiibksqz5HmRqnlIkkCkMVAPMsnXDUVVsCp/skqmOnjEgXgHT55blEjfB3xQfBdo+kBe53QYuj6gtqCnJ2ogquLKGqNK5hxNOyczoBrY3d9HNLJseuz1PYo8Y1RVlHVFlufYIk/1VY1Bh+n7sHrDemLtCV7fe5ymn2msQ2KOhIKsLiirmrysUVsQg0mFEW0EiUCfhPijKpG9Ax1dNjwfzYCFx52qjtbqOTjMPP6AqrqWJKPrU826ZdtS2UgYOnkfUmHdGJUbN25w/vyjvPD+F3LPC19EffoMMj2ewuEmJm9qjIirkRjJRxPwS1guoZkBqcglokPIyQ4hb00JMBpJccRVeZrh3h019/qZnnJk6bue/d7jo+PGfM580RPU4IEghqyoWC4WhBDIW8Hs92liAvb3Fuxe22c0Ktjcqrhy6RLGZpw8e4atF56l2V1w/txF3MUZZ85sUtYNkjvUCEryoISQkmx63xNFqJ3DrWodrnLa1nQVfXkZeMGLT/GqT/s0JuOce8+eocxdKpFSjunbjqws0DAlGks9MbjRmHDxEhfe/z5cnlHkjnxjA3b3aE1H07X44ImS6q7N53Mwq8LRAUcqSdRoizPCUnuMBHyesjxdlmH6nsJA03Xr2OwB2BinxUYcvHjOWsbjEVVRUGQOZyWVNA0B3ynkaeFt3KEwX1WH2pwGjxA0lXbqfYtoqum5ars6TAVOLBoCmaT6k8bEwYuYQuViDNZm+D59Lq5pGmfKJRSO5rceLTuaJBtJ/qQxoiYSQk/rW/ZnM6TIoFtisRSS08xn9L5jsrGBj56gkZyCRZcyxLu2Jy8nlEVG1za0IdI1gcV8l9zuUVYF9bjG1SOKqhpKdkFYb363/gTPZhm5deQ2w1mLsxnO5WQCRk3ahIJIv+gQ77CxQ6tUykS8Jg+SsYe7Jay8dwft9jaGyJsmML35z6dyno9B9F2PRiUYAZcyn1JUOglnQ/A8dukSVVVz1+nTVJtbSDlC1EAXoY+gPhE1UwERySUlsBibQrAhDN66eFj/LoldBnI3/G/M8P7j1B1cQ3KdZxn7TaBZdDTtEh+ETi0zVXpRisqRlRldAyEKXgSi4ozjrhNbzHdvsLy+YH9/wXy2YFxnjKqCRxYt0+MbFKOCzbPHufLwZR56+CqnTo0YbdSITfclDNmC3vdEwGSW0HtUDnU5civRXiPcdc8ZPu81/wU2y9jcOsFoskk3n1FsnER9IJsUZEWJ07QAcvWEdjnHh57T97+AG9eu4Gf7aFB29/bJi5JF06Vw66hifmWBzXPazpMICbTRp4K9CL3vUR+pioymV7Ic5ouGEMLgueieaxM9ayjLfJDW6sF4kzmLs0JmDGbY0cNZixqLjRYdwgrOuaFYdMSYROKyPE/Fug30y1Q6ZRXH0RiJQ2KAQYm+xxibSIwP5GXOijDaIUSYyoOsJ7mDgQSvEhwPnjnURdrBbjEQY0o8FITlcknftlw9fwkVS1XXiBqW7RJVZW93D5c5FvMFZijS3i7mNM2CjY0pTddhRJhsHk86u2aJxo79xZL9xZy6njOajKnqOmXl2jUdfAasPcELUfESERMwGPqYvEi5FoTOol4w2VCEcpyR5xXWTpFiDLYCqVKSRJbzpC4fffyXPxy3hmVXHwys4+1QjfTeY7IsDbJ5QZbleBWCSWUkZssFn/CCFzGabmDyHFEHiyaRO29SOZqqBGdSKDaYdD908Nb1fUqzN90hsVuFz6OmTOZowGoid3YgeHEIEq92KVkz3Lg6Y7kM+KhEK5TTMT46bNNRjSvK3HFsMmV/f550cwJFmVGMarwzMBqxe8VjMsNy6ZnPGqajjmpUEIPHDluObZ4+zuXzl7n42D5nRciqkiieKEoklSFMmksI/fJAp7ruFf1Ho4p3/967KcuS41vHOLExZlI67r12nbvvvZ+srqmtw7c9thxz49IF6q3jjDa3uHrhUaYbx1gsljz4J+/H5hmPXX6Mvb1d9mcNi+WSEA89VPNFQ4iKtYbcpSQXZ4TMWJadJyjoXKHMCFGHUjjlOq5rgLSLyIr0QtLkGUkhV2JKPonWYIss1eCUlMkfQyTEFhAylw3kg5TVbDN8jGSqB3VTRROBMdYlRbFGjJiDcKT3PS66VC5kgIik73d0R4Y1Q/QdK6nSqtDxgd52IH3GpG3bYvBohOViCao4DEShD5G5NpA55r0nxEBYLFMBZVVCWDAZT9LYj7I3nyEqoEIhOZ16bOkoq4Jcc9R7+i5w+eJlqrpkPB2TF/lzaqdnG+vHKG6BDxYrGSZmiB0htqL3GXk2oSgn2Lwiq0bUG1tk1ZhoMoLk2FjSN4ItckxW3HROAQ7SQA9eeBzcNHqmukqEPiUAaEiFeUMS7hIjHDvzLFjguUUISt8nPpVjqEwSwnZRCdYwWy7Z3DzGiZMnk0YpKjRN8px2EXoLVZEi410PBHCDLeOw169zg7YuDuRNhu3fVjuHrMKyg6hcNe1LG1Yla9YTzawHMbjccvzMFuXxk/zpI1fJjaWoS+46eYJPfvknce6RiyyWPfhIMS74tE//dF78kpfy1l98C32IPPDJD9DuXeXCBz/E/txzfdmyMWrY3KjJs2T7Yyc2uXrxCo9d3mfreCQrC6IIgUhUg3pP7Hqsy9KAf3QHgOfQRs8m3veBh2mbB8mzLDVLa9mcjKlHFffdfy+f/KpP5BNe+hKmkw2ycqpKSQAAIABJREFUbJcTJ++iOnYXeTVBbc7lRx7EVSPGJ0+ze+0K+7M5y6bHWMeyaej7wKLp0jpl0OGFEGi9p7AWxRCjx5J0T6LKXCPWODof8LP9tfRcQyrrc6DBAzLnMJBq1VmlI2KMQ7KciMOIxThHDD3GpTIaoe8IviPPU7084xxZqYns+R4VIXQ9K6mOjwGnMW0LN2SNewK6hLqucW5VKguMNcmzta6tP4SDkTU9rzz1h6TPGIEYWc7nxEDK6Ne0ZVwXoA2eZWjp2iXBJNLcRyVzgstyrDGMT5zg3LmHh+QXg4ZUlHrRXiRzjrLKGY8qqiKjMBmmKMmzivlizv7sKqOqeq4s9FHB2hO8rjOgJTFmlPkYYyYYlxOlopMx3SJSisXbiG1bstqhIuzeuEYXI+X+HHMx1eOp6horhqIuk0zPpU4v5SDoR24Ju96KmEp0LFKCwOHetOsbpvKq9EGxQQmqGOfSfpjGIZLTtC3jYxOKskBCQBfLpN+IFm0imBJxHlgkUhZtIsWhgb473FdWLOgqEWbIrnB58t6ZQWDu/WFyyyrTdk1X0JA0WgHPyVPHuOeFZ3HTE5y7PCdoj7WWcw89zHve9Qd03qMoJiax8h/+8fs4d/ExFk3LaDrltf/NF7Lcu8S/f/ObuPjBc/gWLrctMcDmpEjeCeOop2Nm1/bI3JIJFpM7FCFoxBg58J6saVP/MFy8uqBUcLlnsjFmXBbYXLjRLPnAO97NL7/zvZw8tskrP/GlvPJVD/DAJ72S8e4+IUS6tmF3f8GVC5fYvb7LfNbQeQWbsVwsMS6nb5ukY0XoVfExHmSCNn2gcGDzbNiSLCABRAxZ1yAmle9YawgIMmwbZhFVnDvcJnE1FqQajgavyb7Gpl0onLFYEzEx4jBDbTywLiPESFlWtBhC29J3AVHFqGJtmgcyk4FLCR4pmzQO146E2K91iLabz1OG8lAHb3UvdFhsgBK8Mtvfo102qIduvsSrssSw38MiJo+zKzOKLJWasV1PUVbkeUGe56jLcKMxvu+ZN226P1ZY9j2VD3RNSztfMhrXjMY11ljKLKOcTJAh63mdsfYEb29vQV1VOBfpw4L9RhExjCab9JfnlNUEmSnFKNBGT68eD1y4cpm92T4AWZ5z8sRJqiIj9h3HNzbYrEeULmNra4v82JT8rk1MnqVaefAEK2MHZU3yJMmhp+lgT9r1Q9NH+ihkylDvyB7URGpbT9dHNibHMMbSNQ05NhFfLxAEbED3PFJmqFlC26VdQ9pF2rbMCFQVUk5SOjw2EUHbD3vWDh4968D4g/uiq23L4KAO3voRj8jJM8e5694T1JOKaAUjSp5bxCiOCEHTFj4ChhTSsyh716+ynM2xYnjn7/wmk5EyHhVsnNhgvteyf2PB3m5L3/SMpyV5pmSlw1UFs1mHyxpyLYkm3UZjs1QeRw6m1wOs6zQ3LQtO5kKZO6KxFGXF9MwZbOux5Ywbe3v8ySMX+MOHHuU973s/9977Th542SewMd3g7D33cO3yY7TLllnTMO88ktfsN9fZXyzoup7GexZdIJDqDfYhYEQQoxgkLZg6xSkDwR70xhrTTn23ryv5mMNBduSgwQsx4FaEjkH/OSRbGWvThjfe45xDu4AzWSpPrKlosQ89Ym3agWSo9SjWkBc5y9ATWsWqENQQ+5iSKYzB+IA4iJqSLdQLalKxaWD4DuuHdm924O8ADrYsOyR8wmw+Q0Mkz3KiQJUVdBpYKGhZUtY1xgqTuqIqTNp+b9hX3GUF1lm6rmOytYGzOTeu7xK8Z7aYESOU1mEidI1n2V5n3jZsbW2lnauMkGWOTLLn1E7PNtaf4M09j914mGXbUVVTRuMpo9GUcPkGbRfYPHYcV1Z4gfNXHqMLPV3b0XYty2ZBVZVkLuOxC+epski72GeUF9xz6i4q57jv7nt48ctfTje7jh3XZCc3YVwdivmPQjQRubpevTA8rd8Au0IfFB+FEMH7cLBqNTZt91NVNSKOpvFIVLwXhCR+dsYRwxwXKzRWyVs376CbE5p9jDUEK/i2o5gY1OWY4MAUSFamHSucAC4RPDckYMRUDy/6fiB3aSeLdfNnbG1NOX5izHRrymhcMm8ioypneWOPejxCQkk5WxLnEROFqrbko4qsLFnuztCouEJol3vkGskzw/ETxxjVPaHt6JuO+SLgw5KtYwVGMooqp+k6lk2fat9ZweQFgkUkpOKnrJr8+rZ7gE955UvR/T02nKULEabHuNEFZn3HsdOnmHU942xBb5U/fvAcV2dLLl7dRWLk1Z/0cuazfU6c2GI5n2MRLl29yt5iwXze4MoCLz19iPgYiZpItKrS9Z7MGNRHTJHjnB22BkyFd32IaevmQdi+jhR7uVwORC5tlxXFoDZ58vPcIGaoQ3cwBA/1OJuGwlrysgBRjM1RnzR3RgPO5kRC0m0PY4e1lizPCL1PmbKqmJjObV3atcH7gHNp9wsTI3mWkjd0HTfBBrQdFs/pJgxlSSJBAkGSNjuPEQPEpseIYzwesxCDdg2jzWMU002sAaeB2OzjjGIF5ssFXbukKAuigSx3CIbxdIIYwe479navE42hrCvCYkEMhma25HL/GGVZUJUFdVWSmXUb9W/G2hO8D1y4QOc9zuWMcey3PYtzF/ARNqabXG+WLLynjZFruzfIcseJzU0mdcakLMgcqZhid52NjS1Gx06kCv+zK9THNlkur3Lj/AfZ2Nwk0zFkHeoLZHOati6jYBCUQViAtuDGpJYfgEFLRgAzfg4t9ewgqCGQkh2cpGoExshBplpR5GnfxqYnNktMq1hjqUcjetPTdR2m76Bp8EYwbUehHhc9+4slvQHbjeh7wRUleZdjc3dz+HW1R604yARsRPH4fplCJ0NtptFza6pnHJujmlFuycUg6qmrMce3Nrmxu0TIqEYTNjY7RmVDDIHJxib15gbdsKputOHkiU026oKCJdNpjTWCM3Bsa8SVSx29F7o20nWBKsuw1mKdw/eBrgtIZnDG4zWV2DU2LXrSrYlH/l4/fO5//uf4oz99iP3LV6Fp2ASKjU3GBk6cPMWVa/tcunyF2LUU1tDP93n4XMfp03fx3j/5IBcvX8YhhOjp+47pqKZ2Dikqlr0nqBCto2mXxBApyoLee0Bo+p7CWnrvsUMpDu89xhpcjLS+HzRhrCO/I7P2sJa8rrapSnQ2+EAkUhRZUsnENCZpiDjn0hZxIe3dKwA2YoYi9bFXJAYM4ENA1OCshaIgxEgMBh3eV01reqeKHXbTEFUg4L0e7rKxhigkHxKAhrp3YSgsvCq1PdjDayQrSsrRBJOXLPseJaIiaNfhyjJtbUlG00WuXb/ObDbHOks9Ktjc2iBzGcEnu1ZlRWanWCK+abBlxiTfpN+bMVv0+NiixtIOGdb1gbNlPbH2BO+PHn4Yay1333UGf/0qxyZT7tra4vjJ0xhXUpQjrjUNe8sF040NXnD/fYyLDPp9Shcw2hFDqs5fZEKRRTbPbLIxnRL6nlygLjx1rkhuk5fJCsSONIEtgQb8HK5fh7KCsh12WoiHCRZR4cTp59hazzx6H5LmQgTEEDUOW5GlFWye53RthwmeKgi66MhsztwHeicUVUk+2UBGGxR1hfQLFo8+yu5yxu61a5wYj+g66Jqesqhw0yl2MoGxHu5Ba1bhcDc8kicvhDBodAzGrd9Aa/uW9oZSl6kobqRlZFI1+fk+bG5sMJ1sINNRKvtgc4xL5YHapqEqc15w7xlGDvJVTUINaOg5dnxCaAMXLlzHiGG22+GsI8sybObQLu03bBU0REJUiiptB3WYPKgHz+uIT3z5fdiq4uKFy/z6r76dbtbwgnvu4TNe/UqqjU3OP/Iou48qZZHx6I2WfYQiwkMPfpAgllnX03UtzlhOnzrJuWv7pHRYz7jMMTpMhMahEli03SArS2H3iKLGpG0aIe3TaYUQNWXcrrbwW0NoSIuHwYGUdHAYomRJVycGTIZq8mp2Pu3jW1QV3hpMH6HvkKh06sFHCJGw6AfZdCJ2PipGLDYT6mxCbDu6tqPvOqxzqE3hWTP0g0iXsmyDOSj0vo4ojiQm6lAiZpXxraqEGPCqlGXFaLpJVtX0CKJKNR7hsxzf93gxQCQG5fqNfS5e3sVkBdPRFnlZ0bcBKUAJKSt8GOvryZjGCNEIdVEiXcDGyKxb0s+XxDKn67u1lUatsPYEL8sLTh4/joSe6ajiZfee5p6z97B18ixluclovIUdj6mOb1KNazIHzkRiv6S5fhFnI8v9q6nooo34sMTlOdaVKSXbB8oCKAJkPVJGyCJcT6t2nEAzQxcLKEdIvQl7s+Tej8Oj99AH5MRzba1nHmliSZOLy3I4stF0NohwfQgsm4boI35vSe5yXJVjNktMXqUtbm7MmJgKa0fY8QnyCOViydW9GziT4aqCYGa4EMg2R7AxgUwTx15l0epKFCKIs6lUwZAUs370Dk5ManyIzB+5knJRcEivnA4tV/caJhvH2I3K7t4e1jhUG2yecfrUXZwYjzl2bMq9d42w/R4mLynVEAuHVBnWe/xGyXJR0i17MmOJrU97bBaWTuNQ74tUWyxPYSwzlIv4eAjRbtUFL773Hvq25wu+8LW8+efeysV3vZv65CledfYepqOKSSFMTWQ/Nzwyb7i+9BSZ4fQkxwTPY0EJKJeuXqftOrIso+8956/ukVnDZlViBYoiAwn4vscKyLBFVtt5CpsSuULUQZOatGdh2MJrHRFVD4LPguBchnUZWIeYpAX2Lhv2FhfEGoIovVGQDJUmbWflA3E+Y7Y3gyh0XYMSqcqSvCxQV2GLChVofQcxpKQiZ4d9mYeMWYXQ9yg2FWCXeFgbbg2xXC5TpuygdUy7epi09+6Q85ZnhvFkmop9k4ofew1ghTzPUr36EPE+7X7UdA2IYLOS0/fcyye/4iW853d/i+DTtnIuL+h9AOuw1lG5nNxYJuMpxaTB3riC2RVm8zlts6Q3wrJf31qQ8HFA8P7cqz6F3MBmIZyc1jzwohdS5wXTsWU6KXCFEuyczObE2S7qG0zmsMbgnEVRqnpKGyCrK6xxmLIE3yTRvrb4rmHZ3aBaBtj3ECpk2aGzJfQLdLlEqiksWvT8B4jLWRL/dh20HvFJm+Re+tnPtbmecThr6UMkxmHCTz4zBMhdRq6GGAKma9mfLQizhkwyxJfY0DGfK66O2GqDvGhxoaU7f5Hm4nna2XXUdmiZYy0UGcj+PvTLlDGbRg3wAQo3aCKHcK0xgxB7PSe4/5+9N/uxLLvy8749n+EOERmZWTnUQHaRJQ7qlmCLcguGDf+LfvKzX/0kA37oB0G2BNmGYdmQuiGxmyySxZqycojpDmfYox72icisFvUmdoFXXMBFBJORWXH3PWeftdf6re8HIH1EhYwJhSIzOmdsDHx81vFx1+HLkYcdjHJNDIkwTOAz6puv0VbxoJWsg2Qcj8zTxHE/Ms0TefboEHm06VgpwXFKtKsLhsM183REK4E1qg4vl4KSAmMUWkuUPm3Ny7uxWTUcjresG80Pf/IJ3XrD//g//c/8L//r/8bDJ0/IMaPJ9KrQykgMiZnMHKBxDkfhybplLLCfqodnSImUM0ZrxhCZwhEBrDtHbxWi1AOUEJmUM85UZ4Z67xWyqJW9uEDGTzbJluLtnS1k5dQpXauYErLW4AxFaYRQSBKqBMq8R5cd41Ct4a4uL7m5vGYeRnLM5BgpJWEUrDrH+kGVNeiuaoRlqVPis4/EGHBWYYWlSLEMutSpZ3nXwz3RKKK2xWP+dpX4ThepTNU5Si0pJVdMDSCX7k51/FD4eUYurXWl9FIcgJvbPb/69FdM3qONJOVMCAGlDEUItJKkQrWj04rVdo1rYNUZri8lN7s9+5iYfPxPvIPTiJNP8NZq5nzd8ac//Igyz9g84VIh7ROHcSALg207xtcvqn6okrvIqepVvPekFKvzgrVkHMYpbKtRG4t0DqUyfvBMh0vU7TVq3aG2a1CJeHNLUQ2mf4D4+paiemT0lJureoJTEmEMpT3NB18jBBFFg6S5Yz+JyqUyWiMyhBCYxpnpMFDmiBFgpYZDYTpeQfo1281D0vYr4jxx3L1hDkdK8oThFmMkTz98Tv/wAjlPlNkjUnqn/Q0UDa2u1maqttJP2ewbQCOZQ6gTk6WKk8+fPEIQUSKSKDhlSK1jLiOzTJSYSNOMihL78iuurl+inKUISQye1lrW3ZaSI9Jpjn3DzW7gONyy6msFPAUPIi+8u4IxCrcYuhvr/iM8xKniItqup1ev+ekPPuTzV2+42G74J//kH/FP/+L/4J/97/+S8+2KUKAo2HQSt4f9lNFSEGNg9p5N3/MnD8/YjzMvb45cDwM5RnyIlFJnMWPO3BxGdlLQaEWnBb1TOFMHmVLK9wcaKSSFfC+VONUkI8TyFlMqCyFlnFUIqUkChKrwYamAlIkhkMPAtL/h5Vdfsx8KQq6QpkGun2Gdh5QpIRD9zOH2hq9fvqG9HHj+fmZzHmj7plZO0zKdngV+mDAFMAqBrPv90rk50dQaANe4+wnmkr/doqVUSzjbGKQSlAUZI4uq3sBSkkpGSI3SGkQmel9dKMkICn6e+OKLG7TMGNdByZXQoFR9jsua4DmlkNSWvFaSzXqFoSBSRg4TJp3mkMtdnHyC92Td83C9gtsbWqUYD3uk61Ha0q02tG2PmAaadoXImehnQvT46EEIfAikVE+7SklImZwDbedoVh2ma+kuLuj0lnS7YzrsScOBNnms65DSkIcA49eUrBAEyrZBrB5Vka1kAT+e5kZbyozCIJTANk3lQy8PG6U1ear6Rh8zxzkgI3hZeP3VN+TDTDru0GVmfPYENWzZF8VfffEZn129oCvwnm74aLXmxZtLPvmTp/z4T39cfWmRi0PFgkPJpU7RLn604r5te5rrDvW0vFr1DHNgjAHVN5y//5wUI5vNmuvbHS9/8xk+JmyRiATGWUoKKKDrJcEn9tcj3brjQdehpUCSmGPAJ48PhRICRmSUSGwePabkwO7Na8h1qEIZXW2ipICc3j7YljbtierMaV3L0+cf8Orla7aN4Deff8V/9cmH/H9/ecG//rd/xZ/96IcI4yhmZtVrHq5gFzw+w5vdgBEFzYHXqgJhnz4447jfkxDExbD9juIlEISUCCkxBMkYE+etoukqpLdqyt9CdgulDgecaEh9N8xT9b9FCGLKGC3JSqGlRMSMijP4wP52z8sXL9kfjtUia/WIpntAFgIhMqM4ooSghJGkBJ06x23W7A9Hfv43v2XdKT768Bmbs3O0MWipiN7jva+DFa1DoatNmtbIZcignOgJUy981wJ1K17+/C02pSJ7UqpWlndn7Tvrt1JKbXU7jUyQc+LsbMs4e+YwMeze0DrL+myDUhpjLQhVq4NCkpd/00iJKJkUPUlVqzjXOvq+rRrs+Y8VvD/oOFMS52fCOKG0pbENymmUsmSh8LmaP49ZIpUlakkqplqVyUJnNPPhlulwTcmRvFSOvGhJMSEOM4ebkbOHD+vNiyWOkemrW4K4RRVBoy3FH2uyYU213WoclFzFwCV9+8F3QvGh0fwmRFAOIdWS4NUpWqUUQlU463GcuLo9IELmsD+SdwPMHt05UqP54ouv+HL8NW98ovjEIyP50Uff4/31GY8vtvQus91o7OqstmDvdpVUxdK0sMC/arKX8mJ1drpCcyUFPnrOz87YKsU4z7RdRyyFFy9fY4UmDp6iNM26RamGZt3i/VA3PtXRnjkub17gZ8/USvq2QWuLT5GQE1kINBJBxoqE9AekUvStrVIjJVDWYJxDNpoYI+VvP9hOdP2Jga5fs+kPDFPkvYsNv/38BX/+0x/wf/7lr/jFr3/D958/JqYrWnnJ8wbS1vDikAgpc+4EZxoO44BQhuHmmr6zbG3L9f7IcZ7vk7a8rKgQglQKtz4yhIyPgucbR8PChls4hFKwWPWdZpi7A/PCNRclEbxHSI02uh4+omfY3XL5zStevLwiZImyfX0uXO+4vj5wmI5sNj2QMUpWrePibiWkoO87FA+4efMNn/7iM56+H9lu13V8dhGkhBAJSmG1IqZMUpki3vUzP70Q4a3f9P37vGOQLv87p4QQ1cNXiAIhYaXGZIlQhkShhIjMqZIV2oZnTx5xPI6AwHUVkI/IaKOZJo8PCafrZ5xiAlW5nkJWs6MsCsJKmr5OnKc0fCfr83cVJ5/gTbfXIAuqaVC6w9o1wmyQbYdYdWBX2G6L7VYo6xDWgihoJZiHPXkeMAt0Mc0TyY8QIrkUUkzMxwk1Rz57c8mDhxcIrZAahFbM0ZNDIBrQUi6IAosYJsSs3laP7ka9TjD++x99n+uf/xasou0aoL5tKWu7ImlFACZf2E+C68tr4ngk5sQuRZRpOIbIqxvPujX87INn/MMPP+aTiw3f//gHNNstRQmsFchwWxNJ1y+2cKW+8uJiIe8SuiXBi0sL90STPKN1bc3ub5HaknJm9/oNzllW1Amzn/75P+YQAyrMWFVIMnF9mXj+8Se02ydcf/M5zl4yHSaGONMoQ05z1XktCbSI1dtTlozJAS1BWltZYjndT7ZpoxHOfGup360snVooXasY5w8fQyporZElcZgmsjT8xf/1//PZN9d8+KhHlh0bNfLxSvLYCeaiaEpGpkhJhT3V67TpOp4/fUb89WcoLdkPE+kd4/p31zKUzMvjzMF7Pli3fPjQLWiOO/jvCbenRPWird9DQSL00pZNgeA9b16/5sUXL9nvZoS1FATfvHnF7eFYkwpTmENEG4NzrnppK4fTisZq1l2LkRnZOtL2Acf9nhdffYMAzs83ZDIxFnJJBK3r5wQVwUJepnFPc9/Xd3hR8tvq3X3LXIAWKEQFrC8+vw6JzaByYfaBohRaSciRIsA0mqwcytbhrYgACdY6tDEgNPNUhyZySuQYsX2D1RK1eKIjMrlEjFV0nUX8sUX7hx3ztKfpG4x1KNuCskjbYLo1bnNG01/g7BZ7dl6J+52l+BGRZso4EpMgTAmFqkJd7arLVQzVJihX3s44Drz4cmKz3WCspnO2aj/GidHvWW1WleTvJ1abbU0kgTtYUzlRTcZ772148kXDrTZoVcvmSlSdhNYakaBCHCKHuOfrwyXCtMS2AWtYmZZNiDx+T/O0b/jkgw9479F7fPTJ36d5+h7u7KzW9Q+vYOcRpgPd1OMaSyUPBWGuiXQMFXR8l/TldP8ZnFpIrREx8uBsw+3xWK3LhoHDfsdwsydMkd32kg9//EM2T95jurnk+sWXbJVk03c02w05PiUlTQqByy8+r/BjUoVMl1LZYUohpaSIjEBUVp42dOsOSSGnyl2TMb1jun6io8vvhDSWYhp0Tpw/fY/DZ19QROHJ+RqRE59//wM+/folv/jylg8fPWTbvGYzjqwkFCOYgdupkHxmnGd0Y8kikSm4xkKQyCI4zDM+/m5f5VwKx1D41dXAEOHjhz0XK4vS78oTTu/aj/KtJZmUEpTCWFtF/eORy5dX/PrT3xITJKm4vb7m9njk9WFgGCeUFBRdMAv1CqHJGDSSjdWsjOL8bMtm29BmTedajHZM45GbmwOIgm0MRtVDVgwR70NtGeaTv/QpKd+71si7ifmFWFA10VX/qaSmLIM/WhZUCDip8FmRtCYXMK6FkpmTJySPUNWJx0h9NxBOLhltFUUoip8Jc6TThirWqQxDkwopeUQKiBRROdG5006BTvvdAWSBUC1eGEKWyBhQcYLJEg8OZQJajRTRIrQkkxiPr5Ep4ifPeByQKaGEwE8zcRoJ83hPPR+GgePxWK2CvCeHxLpv0G2HsZI4VS7SpGeUtQglWWVBGmbS4o1aRN2E7He9Vr+HeLja8PHFAz69HVBC3lFKlgRPYaPgOCfkNPL99885TAdmJIKCk/DJ9z6gl4oyez54/if84Kf/ALd2rH/4Z4hxh9CKvN/BcKyVO2vfcaxIkCRMY03sZMXaVL50guRrkgcnqcVTqtCvWkKcOdusKVJxHEe26xXbvufq1SXjcc+Ln/8N5f3n3Lx+xbS/pT/ryMeRUd7ijzMPnzzj1Re/RseF7VUK0zDhS8bYpqINqKm0MVVvqU0l+2cFbd/S6xXee1iYWG85eKe37ndRLZIlYRpJIXF+tubHUjJ++Jx//i/+H37y7BxJ5i9/+5Kff3FF7xoeUnhkJlQpHCO8HuHNkNHWYIQg5szucAApCDEgRaFZqkPpd7Rc62BFZT6+3I8cZ8+Pn2z5+MkGc8IIsJDftkitkXWgqyT8/sDrz3/L1y+umb3hzeGW3XBgCplDzlweR7TW9OsND58+Y90ZDtevCccdMQauj4EwZmRpSa+vyWmFaDqsVlgjkbIlxMDtYeJBY6oLgw91ajcJ0jRTbKp2WbWH+V0v1e8lYoxvbQl/h9ZWaY0o1aOXLEgloXLBkHEZhmkiFijGMlMIOTP7hA+F1i0DHDFjrKqOLUKSKVilSEpAivTW0SuFLIlSIpTa7o0+kBdtZE5/1OD9QUeOghAFISSCDDWR8gFpAnKeyfNEkoa0n6BpKGPGsMYo0G6P2xTSdMQfbis8NGVSjIQQuLm55rDfI6Uil8LhcOB1Aff4MYNPNH1D8IGQBJPP6BJRumJBdN/WiuKJx+cvblg7RecUMdaJzpQrNoUisErTO8fFdks2E3/6Jx/w5Vdf0UnBR48fcbFWrDcXdGfPeP7JT7H9GsqEGK7qTT56eHOFcAo6WydkrQGjoMQloRNgXU38QqyvGChhWhK82rY9ta3WGEOMkZwi43BgvTlHAfM0kkpme77B6SPT8cjVl18yHUbIGRkkL3/9FU2/5/b2hvZ8g9Cax++/z3zcMUwjIhZkKuQlQc5ElKnYGVEEyQeEqUBdEy1t2yKVWAzX34lTLuR16zox3veY1Rbdtlzd/Ibd9TU/+7MfMY13HwmsAAAgAElEQVT/lnH0aG349OtLDsPIz28zvyiSZmGAzUlQlETLOkQxp8RvPv+ymqY3DTHNQKFzhpgL0+wRQpCXxDnlylszUtBqyVmrUaLcQ49PNVSj3x4eZCH7EX848Pqbl3z99St8Mby63nFzvAUE7eaC3TxTkuDpB+/Rr7b82T/673j+ox/x8vPP+e2/+3959eKX+EMm7QP99gEuzFy/viJ3E2ebjq53tI3GZstxrF2dpu/JJTNNE72z1SIuBnJ6lwd5ejHOY3XuoLp1SCnvvyKgRIHKEqc0WYgK45YSWzJNSLhY8IvJk5eiqm1yoRRJzmpx8o0IWYH5KWWkEMQwU7ynF5oHpqFTGgiEkpCyDmsICiklULI6jJxwnHyC51MipUwKGS/C4qogsVlgEzjZI8MGBeiiiaLgrEHkSIkFP83Mx5HpcCQc9wyHW25u3nB7c03xkU3Xc7buQEBjDXEMzPsR0QHOkdEIWUW1edF+zfOEshqxeKDev04w/uLf/Du+994ZqmkRxiCVrUOtPqJ8RBvH+tE57qxH68T68g06Hultw4fvPWXVrrD9lq7XmOmaEo8ILaCxYFdwe40wFrYO4RaundWgZWUV+lhHtuTCo0gsujwBMROGRWRbyslVUMdhxDrHdnvGcX9kd31Jt9pgmoab3S3Hw4H5MEDKHP2B5ANKSG7f3NCuOoQyPPzgffqLc2zX8+rTX3Lc1QQipUjOCSMdWkuKrJZaKWXUkmAUVRBS4P3MNBtKTt/6/e70R6d55YMIvh4sup4yeQKVBfjsYsMUPM/ef0IshXn8nGnbcKULu8MeISw3syfmqqkrKXI7h3ecDwRSJAYfgYzThtZoYkpI4Zh9XNpiBSsFZ43holE83LRsO82mbVGykE/YSsEuEphCweRMPuy5/OpLvvztC7LpuJ1GXu2uWK06npy9h+7PuXr9itYaXKeZh0v+73/xz1CffsZZd8Z8zAjbcH5huY4HmtWax+aMg1VcvXmBLBFFz+Zsg7GGxjUEGZinib7tKEoScqYRqiJA3tVfn2BsPnpadXChauGyjxVIHONSuYuEKZFERluHazRRS3RWdLEwE4kpMgcPxhCFwIeIVQaBIJdMyBElDFLJetDMULynyYIHxrEqteWbdCZR6lAjpXJXgRgCJf5Rg/cHHXMCP0emPCNiwet6g8UkiLFA1BxcxA3XNNOKLDJSBIyIHA87drfX7HdX3F5fcvXNS+I0IdKEy4mtcpx3G9QU8FpirUGXgiSjREGVhCDfM4Akd4bWM8YabNGne4RbYvCFX78euXi2BQylSFJ66+IhqJWmtuuQSNI88uOf/IDz/hE6FRpjEbajadpKONlsYb2CzQUMI3QO0XbQyKrsjfNyVcea4CUPxi4Ef/1Wb6ccKIuSilN9ypWUyCEwp4TRgpRgOB7oZXURMZ3EKMl0HPBzQKpCSYE4F8aS8X7icP0aZe29xZDSClVyBRc7XV0ABGRy1dclyCKTqZgDqxtAMI0zUv7H63zSV3/XLXa7BdE3GK149qO/h4qeX/71r9h0Lf/13/+E169eMs2VF7jfHzlfNVzdHng1esiZkgqDTxRRmXfAvcenVhqnNY1rmL2n7xv2xwElCk4IbI5sO8O6sXS24juUEhh1t/WXt+r3E4o47JceuUDFyPDqFV/85kv2Efx85M3NkbP1Az788Am9aphlSxaSrmmRolBK4Pb1V/zVv/8UguQf/PB9PngsUUZwIwXCCFYrS2vOgUi43XH7ZkQUx+pM4TpD0/XEYSSlhGscviSS0Cip3k6Rn+bWg+26eo0u3ZGyvO46JYLaHp2HAX8YIBVKazFInFZshaQUuPIzIYUKq06JXCpsuuLFEimGOoBRCmmaaVLhTBvOraYhk3yosGUSKQRESpQYETlDiJyul0iNk0/wjrNHl4GoArGxGKMJMRBixoVImCPaXGEPmrRrESJzHHak5NnNRy73O16/fIk/DJT9yHnT8vzhI86aDl0kUmqEtmAEVguIEZ0yloIuiUip/C9ZIb85J7yvujyl5J3nMpRykh/G95485tX1NXEeiN4wowkxkin3FRzXNHVUnsKTJ8+QOaOTocyxgiy7Dnd2Ae0a0W9BKso0gZ8QjYFVU6t2eQYiqAwpQpqr9s68Y6peMlDBrxhTyegnWj3VzhJipJECiaBzDp8KfhrR1iCspnGW1aojxUCYPHGaq+uBuJt0FiijcP2atl8xjwOlZIZ5qsXQmEAKYq7if0FGFIlSkhQSSdcpxAq5ejep4J2K1Gmuf73uFlyPtmjtYLVm99XXDLtbPnl6zm5M/A//7Z/zr/7NX7P7m1/y0abBGcX6Ys30ckdrwM+JtZZMGfY+knIiCSrcFUHyM3rVQWMrxiPZKmDPgbPOYQVYXde+axxtY9FiQdWcYHIH0OSpYjBLwR9Gvnn5mt0UydqSQ6QzjnXb09uGTjkQBqMMSWZ00xHGidY5OjugTKJpI323ZgqKlAuNhF5DVo6w3XK9m5mOgVs1I5Rkc9ajm4Zx8oQQiOOANIZUMllUGzk43fN9KeK+PSsWreeddRyUCjhOCWccYRgY9gfyJMiuRbsGpTSrXChachMCx5JRtedFFhktZF3BeUKkhEVgcmEtDWtjUNmTRSamAEkiZUGkRPKeEgI5JEQulJT+E+/gNOIUc4pvxeWbK2g7knXk2GKcwsW5XiopkUNEGsNxgsOomcPMcTgwzyPzNHL15msuVhsuHlzw4LHFpIxKEhsNSRSMbWnPz8FPBD+hUm39Sl8oAYSzddJHSkpJxFi1AN57tFFIKWp+caIJ3qNGYx9fcDMN5NiRjKruCjEtwvCEqYaxWCtpzRZC9bVPOqGcwnRb6NZLMlZ/VoyHujuuNtC2YFVtvY6htmNjBj/Un48KQlOTvpKgBKAykuo0891412lFt2kpIbHdnpNiZB6OmFQIITDPE0JoVptNraYagWgspawpQmKdxXuP0grbNti+R0iJMZrjYU/T1IqRlJBSwmqJNAZKIaZEyQWlZN1ApULIZVr63l6A5fsTPkMr+9bM3GhETOSUEcOOP/3zf4wQijcvvkE3hn84fp9No/j0V5/jo+DCSNquYRgOHL0kpsjN0WNc4jBMpJIpBbQoaKkI84w2hjwHOqNwJDbdil5JemexSmENbHpHYw3WLFWkE82t9TI1Xwocjgdudkfsao0UCqsC685grEbnKuuQCh6sNlzj0U1Pv/b0yvDDUPmp7627ykmNCmdbXALlA0pLeuMI7QqpJe9/731++9Vfsxp6Prx4gOhmhmFClGrJlRddGfD2cH+KIfW3qnZVD1mHXir83FJ0IquI0xppLcN+z3wcySFXNJlUWCFpheImJUYKQSQCGV3q81vkOljRao2Tkl4pVI5kZjwRUQoiVe18SR4RA2n25JhByDrkccJxijnFt+LqcCSHzGA9K+9prcZYzTB7rB1wtkXahiEnxmlgnke6xrFpHY/Pz+lL4uF6w1nTouKSFBYoqsG2DTjDHBL5ONMUSRkL2SfatqEIS9A1Eczk6mCTCyVBmhORGS0rJORU7/QzIzB9TyQhSyaXQkyFMNcp4qwVyR+QRmPMOSLV6VqZwViB7BtwPeLuPvSeEkV1p1ivEU0DxlTzbq2rNg8Jc4ZpqKSUYsB7ULImd2mGEhBEak8xn2QVr2ka5jRy2O0RQJj9cpjItNYSfOL26gpypmssUi1YGSGIvqCUZr3dIo1GSr14PWqabkXOEdc4pIDby2ukVhRRp+OiFERKrQIKUYddlglCqJWjpX5xinn129BmKVsUUBqkRmTB+u/9FGwD45H1NFL8zM8ay/NHW372k4/51//+V3z5xdfoVceL6PExcNY4HnQNrw4TVyRCzsSYKbk6w8QQkAiUAmMUXWPZNpZ163jvbIOzCkpGa4EW4KxGLFWQ07vyYZ4DQkBKmeubA6bt2RpbQbjW0jUdZEkeR5K2CKs4c5bLw5F5DKyaDiULT1NDv77AdS1RWQ7jSCcNnbKE40hMCSU0rTQIldlfveb5s6fMfiTHgNaCrrUUUZ8DJUtSUSeLZrqLOeXlDFe7B0KKt1O1UuBzQRlTuwAiIduWM+uYx4lhOOKPR4SQGKlYCYkqgiglqUjmlAlSgDIoJascwUiELrVjEzMlTJQSKKSaXC5osxxClT0sjL5T3PffjZNP8L7eHTjYwEob+sagHRQhaLqWnDOqFC62D9jtjpw/2GKMpNGa7z+5oBWKYhs0ColkTJ7SG/rzM9arFTElUk4cr66RQrJpO3wueEYwAqMkkcrvKVKRQiLHRABmRrQomKa9K2ScZPRO4nNg068YRdUghpiZ5pkYItFoTIlIXQ2/NQlZCtI2oE0doLjbCGcPjUMoQzEWoU39c0FN7lKFXFIiZT7CPCCatg5XiJnaLgsgE6iyIPJ0rfidYK+kcQ5VBONhxE8zjdVoa/EhEEJESzCqrmEsCVUEUqn60grnGnIpWK0XbGAkpMgcPdvtCh8mSs5Ipyo3bGl5GKUhJ+bg0UqhlFwKWSfM5fidsSR2iNqeFiCMgZAhJ0rTILsOu+rYnq2xTjPeXnH+8GcMux1/+YvPKX/9GX78hjKPlJJ52DWY0oBUxJyZ58rjjAi0rIn7g94hpOTpecej8y2rtsVqRc4RJQTOaELwi67yu16j309c3RzqNHHKxCw5f3BBCANvYkBZi240nXHIpBnHESkiD3vDi11ivDngti0pZUzforqGWQqGcWa82fFhe4alVGeecSYHj1GwWjVsHp2TG8X1kJnDEddImtYx+zrwB5UDevK3gtL3EPN8p8O7w/ikAlliBEip0NqQfKCoAlLQNpYcfE32kq9JmJAY+bZKV0yLdi0pJ4r3FZAcE3PxpOgpPhDDiFBQSBAqs7YkKKkghEQqXQ9hJxwnn+Dd+Kk+u0vCx5kwRtqu46OnT5Eh8mjV8f75OelJpl+3nF+c1RZHSOgYiNJg+y2y6RBo+ucfsTl/hCyR+bhjunlNGGZWK0UjFahEe9YirSYSEFLTNA6fE34OxDAjUUgqDV03ppokn+gNb60lj76iOnzEmJ4QErP3+OBx0SAMiFRI80SQEdc0b/PdlCghVXFuTKA9nG0RIoNTNVkToT5IhaiAYx/heEvxM+gGkcTiQ6vrz6mFj2cimFxdLU4wwbNaI0xGrFZ4Vdt4cfIYrZFKMMeE1Xox9q6VZKUVYTGyN9bhlMRaS8gZJ1u2zlW/5jCRY2YYBzbrFSl6wuzxPlQLoOodULErRaGVXjxR34rLT9WH8z6EWBxUdJUKpFhbtkoBCaENzeOHNGdn4Cfc+Zrx5hz36gp93vL8+RMevveIX/zyN8y3N1wfRoaQ6LsOKwrHOaCoTglITYoeay1rZ1Da8OxBz4OzHmtsTdJLJsaAKOBMRTSVEy3hFVHxGUUoNps1OQty0mz7M5q2p7G2estOgVQ0kUSL4oOzNV/eDuxvJ6wqKDRzSYTi2e8OPJCG97ctMnuCzAQRETITc2bVNiAj4zQACWTBtRaKIovaOBBCEN9lr53gvgOQRGWZIt5KMO6+SiVBVoSPEIJx8gi52Bqatg7eSUjHAe1n0uLpG1JE5FQ7DUIh+g6zOUfmzPTmEjHOyFwhxqJkRBHEKQAFHWsVLy0UjcrrtOh+890s0N9RnHyCt9409G3Le+sNF33Peb9lu96yXq9pteKib+kk6LZekEIIVFFka8kP1pj+AfbsMdvnH4HrIWW0Nhy+/rq6Y2jH6uyMtdEYQLcKYwST98zBU3S1WPFTwIeR4CeMblDaIGSduBVl0YydYPRdR54SksoPTDnjQyDmTIwRH0LVaMVIKgNIjbYaGSZE9hA0BAnxWE9yOcGqhfNNFbFbUzdJuUzJBg8Z4n4gDzPWUidozdLKNYvBejnWSkqkPoRPcKPVUlJUtYSz1pBCIPrINIxVeyolQgqsdhQpiHctjBgJfubysMM6i9IaFnaUEJKcc9WO5kLMGWk0qQRiqOsoVTX9ltagjEZqiXO1BfyupVbV55xwohcjNG3F9ChVIdwp1O+FqjZv2latnjXIkuldQ7vZkI4Hwjjw5Lzl+//NT9jvbjkGuNzNvNmPHK5fcr2fkWlCSwmiauqcAqsE21WPsZrNdoVdhoxizMSoFybZaXenXNtxN8kjhCL4qgVt1Yq27VFKUUom9RETCrfXe9ZNg1t3PD7zTDGSUqzAXqkpovDBasNGO5rgK02hMTTbWh3FVFlJVpl173i0OqPrDUoXchbM3tehAylPcq/521GMetuFvmOMLhecUppcEjnDOE8k72lW7TIUI3BNSyoF0RtskyjTROMjSsI4jjVBnkfCq5nh9WtySUgfMDlTZJ06zynXzg0LsUEItFVopRDGYJ1DuhZc852t0d9FnHyCd7HuuOg7np+t2bYdj7ZndF1H6yxd12K0QjtH6yqsWJgOmjXu4jHt46f0T57h1lvSnDBaMr95xXzzCqcStihK29CsnqFkQpeMSZYiAsOxUHzG9S1CK1ScySWTciBERcHhWgeijpILcZq7bWM0xjQ0KmLmGYRkTDBMsVaKgiegySJQYsC054tH7Fy1drEggqIswn2RCjx8uCR1d/Y3y9OqpHozh0SYZvLksUXUh6lUC5OsW7RRDSQoPkG6g3KeVshFikARhNkTsiRpgWkMWQv8NGGMwjWOAngBNIoQI1IIpsOBm+trvA9oYypSRim6vqe1jrZzFdyaMzlFQvDVsgxASUzXIKymiOpsYYw5We/N3xna1IqyEFVCwHItllTbVMskLFLB5BHrNaSEsg5pDdo5fvrD94m7PdPFGh8TX7265Xv5nKtLyxwS4zRWqUNMSFUgB0QutI3DWIe1Ete4RWsnmX18611bTja1RiyHEkqF6GphsK0Gu6LtOnLKKKXIJPwYeNh2CKmZQkTLkS0LFFcKnHO8ev2KlZPI4nGdozvfMs0HyBljdL2HFl2qsQrbWLTTCJnIc1qmSise5GQX/Z0oWkG+G64QS5JXQcdZKebBI0vGzxGrDX6cK8dRa6xpCLmgbEvpHXGaUbkwDQO56xDB43KdIC+FejCVM5JCUfL++ha5VPzpXWJpLcq56pAhBdiGdOKfxckneA+05oGElSi0SqCMxDQW0zpM1+CaFtv1mK4jK4XuzjHbh2w/+hjz9BGqscy3A3F3BYcZmWbicIvKMI8z1jma8zMQCcKISooU9xDqQ021mhgjw3xk8kPt0GjoOkPbaWCxbzpRvflKQatktVraC1LMjBlGH/EhklIiKcE8ZWxra542eTIJKUXtIcUqjK1DFBLEu3Do5WtK9UWB5Ek+knx9qAmV698RAlwH6w1M9URPof69E7zR26Zq6ErOSCORWUKSaKXIgNusyCkRwoQQgsYaMgolJG3T0C+HICFk1ZsuXrIFKissR4wxNG1DjAETJFLU6lCiiv+FEKAkuWRmP397ne8SjFO88AH0gs6WsupHC+AWzc89rkfVqrQx1SNZANYg6AE4f/6ctN2TYiTOnocXZ1y/fsODXjJMAe/ra5pnUgrVZcd71MLldM7hnKn6JykxWjD7+Lcu99P7AKS17wwySJRSKG2RrkU3Dilkhd7mhLYWUiH5wjTOdMbWa7ek+zPk9589Zxp3OFN9mbUSbHRPzgkpq85UWEnXt7UtKRVFGQKKKR1JVM9m8ZYNVOP0lh5YtKb3zDveVu+kIsdIERB8rC3rGEFCTgklJMPNFaFkdNuSYosyhqZf4bVi+94jConx8ho1J3QW2Fw47m5IKaBkhakoJRE5I3OphRUKaE1SmlQgLM+L4P9oVfYHHXqB2Wql0EKipERphXYGt+rp11va7gzbb7DrFbpZobsVyo+Ezz9jGgfGqxuMrCLztK+WZU2/wmw3qM1maQMk6C0MATkPCFWwWtG0hnGC2c+M80hjNcZpComby9dszs8w6/V3vUy/t+jaHi0nlNQ8WK85FoHPmZgy3ntScstGKwlBIZVHFlHbVkZXeG5Otatxx7wzsrZi1bJ9KHMPNSUnmEZKjGQkwpqKqBkOEEOdus0FQqCESB4H/FAF7N13vVj/mSOnWHV1Qlbdis7IIphSokiDci0xBvAT8zwTUmGaxqrHywUJ2EXbKKRAK4EskGNEiowIgZwThIC0GmcMKUWMNhQlyIJqSSdANw3auvvf7ZQ9aO+jxCoNuJvgVhp8qNeoUovpeq6VaG0h6+qbLEWtgDgHEpSUyGnCNA16f0A/fki72+NDYr/fA+BDIMVETLHaKaZ0X9hunMWnTCkZgUJLwTTNlHKashAA2zT3LWiBREmDVI6iLNqYesjJhVIyUhRyDCQym3VtFVpridGTc5WWGGOQpUOJgtQKJagctbJMI0tBkYUQC8iCVhaKYZoO5JixygBVR/xu3fRUbwPXtPcyjLJU1ErJGOuYx6EOchlDTAmWyVptHEopjsORoiRTqt7B6+0ZZMnNbodqetYXF4guk3Tk5uaGR48fwbbHX11hfSKVxLA8a1MKlJKQSpJKIc+enGHy9TB053hyqnHyCV7txWu0UFhjcFLTSEunHeu2w9oWazuc6QmHkenyhnDcI0tAyUiZB5gDxjT4mEFo7GqFbAxqs0X0NcEjjiDmejGHUPdsrZECdrtbpnEi57Kc9hTDOPL4wVl96EmJONEpi4gmBU9rG54+fMxNjHz5+pbRB8bZk1LCR4HRDh8F0iesztXWKi/idGfeIlCcpSZ1asGjwDLzXifTSqDMMyXEysZzLQgD1y9rQpgydGd1EGO/Z7y55Xg8kNIpJngJyMQM/WpNmG8oOS55cN1YJYKUC841hBiW6p6oCJsM1qqa3ElLiDPh6Ekp0/cNhUwKCUr1OpVGkUUBJWi6NSEmppxBJErwzGF+57e7Mx8/0RIG1KSt5HrNLjog5AJ9FlVoTgiAql/LMtldRP35mKBIijYIC8wjqm1JIdB1HWoeae0D5hiY5okQqp1Z9NWScfITSiqkEhgkIPElIoSk65uFT/Ydrs/vMZS6m7AXKKnRuuqlC6oeOmStVOeckSUjKUhdEJ0ip0zOM0UUbGNQqkEg6JwghVCTwRAgFYSUSCEoCEqRIOqBqAhbLa9jwaAoOVTdK2KpHi5xoh9ATqlCjhFIpe/vdalV1Z+7QkiJsoDms4DGWvw8LxO4ddI1TxMqJeQ84W9vuS6F6c0bnFFIIWly4Ktf/g2ma6sVmlBMw0jxnnmYkaWQkscYjdYGKSQ5FsocKDkyev9dL9XvNU4+waOI5aRWTxCkBFNkynsOUuKmxPF6wOlLrAyE4ZY0HVFkRp/o2zVh2DHOntX6HLs5o+16lHVLkqGrmFpR/21ZfTj7fkWWmePxyDAcGYahIiespQDWNTjXIIypVaUT1SbdjiyWSgpjJDZU9ty1D4whE0MhKohWowTEVIgyoRCIJJHOIYys3rJLNW4pmS4PRF2/xgTC11G1cYSYUJ0D7aoAPRbi1RtMBNYRYmC6fMPN5S3DOJ5sRamUxTFkQdTkBVVwhzB4K3yWxCTuNUKCWjlCQIypTr2FgDb10GKMwVrDMAwUyr3hdykZkRXTOGKblvVqzfF4g7KuOmTw9r/57vcnGYV6vYZY94pYvWNh+aJl1YWG+a3EQGtIAZEyLIMAoqSqZQyyTtsribEaZItRGhdCdegJtV2bjSGFmshLfTdUVOUQjav7zzzPlBN1sYCK3wCQQiKNRSgDSiCpWkWtM8Z2lGJJcWaaQvUjF5CVQCmLdi2lVN3YPI3E/VTtFYtA1PI0LKiZIiS27xBKVZu+DHOcMVphUyZ6EEVglLr/HcvdPnaCIcsC8K+QlLurHlEK1tiqdbSOlDLWGHyOIBWp+CpPlRJyobe22r0JTxkGfIgUozHrnpCqX+142LG/zkjpaGyL0Zqm79nd3BJzwipNCpmSKx2gaVtkUUQfmMbxu1ukv4M4+QRPLrTqlHItj88jRTcI5fC7a/KwR6Gw1pFVIk4HUgqEXLCm5bibGPY7GtsilEA7h+k3sD5HnD2A8y0cbuEQIU2gRd1QZCHnQIz1IoyxbrAIibG2nia0ru0Zpd4S708sfnHtWduOaZ6wSiBjwBpFSJkxRqYUcRGaFMmyIjuSTCQkilJ1eE1TW1zO1q9QH5ZJQdH1zyqoDUKhzAlZFNZ0ZD8hlhH7w87j/CXNYSbHyOU3r3jx5hqh1MlutKUU5HJKvqPJA98adnj3neeFOVVKxRbUb7j/e/emH+Ldf0Ms/9d9T2whxJfaunpHYP1fVMRIrTYLGA/Q9rC08EDCNNfJWqnACDBlqeTVtUOK6rhjKqhbSImUkqZvCcOA05oUY3UbcRU6Ta7csTll2qZlmAasdSh9N70M5Ex7VwnnNC/9+wRvWTMpqzxHLkoOpRV1+FijrMB1Bp0KfppQqjILU64VzhIzqoCwHTlEckkYW6dmpartxUzd93Mp1e0lVPSH0aLimmLGaY3Wekl8yskeKmFRzxTukeb37zRlhFIorSlK4HR9PqssUVLRIGAcKQLKMuyVC/h5REuJAsLkod+gpGQYZ6RU7G93rAykrGg3dsE8Gfa7kWbVI5CM04TUihADxhomP5I4XZkC/BeQ4Flbufk+RnJKRD+R7EBSsT6EtMAYzWGMJD/xH9h782BLtqy877f2kJlnuvfW9Kbu183cghZGOCRBhC0gJGxCFm0RkmWMbYawkZGsluwIQpZACl8uEmHCkgySsWwskBsLgS1hwEDIAUIC2cKYQRgEppuh+/UbeK/musMZMnMPy3/sPPfeqldVXd39hn6nzldx65yTmSdzn5Ur9157rW+tXXk72AtKjD0pKJIzzjvGe3vUu3swGiO7e8NsOxZtHrkSkuk8RqaINfRHN+n7nq5d0XVtGTwpRFNjDDhXsr2GtWo3Ef/i5g2++PKI+dFRKROQE9YIfUy0MbDoOxoLMSVCAucK8TlbSxJwzpZxyNcl/KGmcJiG5d2MjOcAACAASURBVN8QCjk9xVL2JChqKnLtkUlNOLrF6mhBzCVz93BxyPSkFFm+fvMONxcLvK8GLsiGYTDMRMy58iRlG8jd7xneK0MixXqt4HUW3JmzTc7VttK7rnX3h/V1zzwVG2hJPAz53O8OfaELhL64iYwddNaU/sNW5VUzSF32ubpk3PbtUMpHEePAC84HZDQh3b6Js56TwyP6WDiRta+ZXLrEYj6n71pmsxk3b9ygqkpWb9su6dqS3LWpWE8+RMrEwtpSjNsOyVY5BLJCJTXel7VhNXdYb8g5EUPEupKcUqg1BvEVIgZrFS8WEoi1iLdkMiFFEEsQQSUwcRYbIt0yDOszl35/aNmbKJ3XH9bqXZPIdd8RumIgGxTjK6q6IuVM13ZoKvdmPJ0SYiBjCH0PrnCpHZRkmKS0OWONoUuQxROx5FiKslvrsM4wGo/L9yma7quKjBJzovYj6smI4+X8DZfNGwnZ5FnEFltsscUWW2yxxeOIzXQbbbHFFltsscUWWzzG2Bp4W2yxxRZbbLHFFhuGrYG3xRZbbLHFFltssWHYGnhbbLHFFltsscUWG4atgbfFFltsscUWW2yxYdgaeFtssUGQA3mfHMhf/UQ5z+OG11D+XyQH8tJr0abhfO+QA5nLgdiHHDOXA/mU1+qabwa2+v8mQuR9yGsgM5EvQl473UfkHYjMkQfr/rD/La3798PG18H7aCAH8mHgScpa6Wu8T/f1vW9Oix4zyP3lj26G/Af9+lrd1588t+1rhm3/+pvVrk80yIF8I/AngSvAIfAzuq9fPuz7aeB7dV+/601s34c509MA/N/An9J9ffHNatNHgu7rC8B0/fl+ctR9nd7nq68Ztvr/aPiE1n+5v+6jn7i6j96t+0iRIXpOhvr66v6bha0H79V4j+7r9NzfRhgXbyG8B9Xpub+t/B+Ah3lj3qqQA/lq4CuBLx4Mjt8L/JM3t1X3xXuG9j0NXAP+2ze5PQ+EHMhGTuS3+v+m4T2DQfQJr/vIZur+o+Kx/vGPAjmQCrgKfKHu668O254APgy8U/f1hhzIlwJ/Ffgk4Ncps/l/+ea0eMMgZ/JHi/yRM/mjegN5tfzRt6b85UA+E/jvgd8D/A7wDbqvPzLsex+wAt4JfCHwR+VAbgHfDXw68I+4Z3mCh+nmw651zzm+hnu8LHIgCny67utvD+1aAp8M/AHgV4A/DvxF4Kspg8BX6L7+v48ggt8H/Lju6wcBdF+vAv/jcM1vGc7/+XIg3w68j7IkQKv7+vXn2vYjwE/pvn6bHMgzlAHoC4A58G26r3/rEdrxSNB9beVAfgD49vvt/wjy/zDwnZQB/Wngh4E/rfvanvv+1wN/geIx+Ubd1/9p2P5HhvN+KnAEfLfu6zcN+z4JeA74WmAf+LAcyFcN2zxwwD1y1H197z339N8C/jrwLHBMkdtf/zjF9RGx1f+3kP6rtsiDdf+h/bLcX/fRM91H7tZ9tOg+8mrdR4vuI/JJ3KP7yP11HxlkqPpepNxPVH8bebXuo6+/7r8e2HrwPgJ0X3vgfwH+w3ObvwL4J4Nx97nA3wW+DrhEUdofkQOp3/DGbiL0wfIfjLv7yh9568lfDsQDPwr8BPAE8GeBvy8H8q5zh/37wLcAM+DnKR3j3wMuAv+QMrCsz/dA3XzEa300+HeBvwxcBjrgZ4FfGj7/APDfPOJ5/h/gq+RA/rwcyO8976XRff1LwP8FvPecd/17gK+QAzHDb74MfDHwfcO2H6UMuG8D/hDwn8uBfMnH+BtfBTmQMfDlQ7vv3fcofcN/AHwJZbD6DIoM13gK2B3a/h8D/50cyIVh3wL4KmAP+CPAn5YD+bJ7mvCFwGcO5z/FA+R4L74b+Drd1xnwu4F/+hAxvCbY6j/wVtJ/ebDuP2K//FHpPvJg3UceTffRMxk+JEL03cDXoW+c7r9e2Bp4r8YPy4Ecnvv7k5w9ROvF9b6S0qkA/CfAd+q+/pzua9J9/R7KA/75b3zTNwI/jMjhub9T+Z9b3PBV8kf151BN6Ce8/O/SL+Bvn9v3+RSuyLfqvva6r/8U+DGKQbvG/677+jO6r5niefDAt+u+Bt3XHwB+4dyxD9PNR7nWR4Mf0n39F4P36YcoXoX/Wfc1Af8r8LmPchLd1++lDLZfAvwz4LocyF94yPE/T5nF/6Fh078H/LTu6zWKN+SK7us3D7/xQ8DfGY75ePHDw/07Av4N4K/d55hH6Ru+Q/f1Rd3X2xTD5bz8A/DNw739RxQPzLuG3/3Tuq+/qvuaB4/U91MGtfP4Jt3Xhe7r6mP4fQH4LDmQHd3XO7qvv/QxnON+2Or/Q/AW0f/SRz+C7n+Efvk7UH0RfbDuoxrQu3Uf1Z9G9VdRzYNH8L66j+oC/dh1H5EdVO+gr5nuv+HYhmhfjS87TwJeQw5kCXyRHMgrwKcBa1f+O4GvlgP5s+cOr4BnXveWbia+DH21/JEif+T+8kfeMvK/S7/W4Z/h4zPAi8PgtcbzlFnsGufJzM8Av6P7dy0o/fy59w/TzfwI1/pocO3c+9V9Pj8yiVn39e9TvCke+LLh/S/rvv74A77yPRQP7z8eXv/msP2dwDODIbGGpczgP158me7rTw4elj8K/DM5kM+655hH6RvO38/n79l3S/c1nvu8ZJCjHMjnAd9K8TBUQE3xYJ3Hx0N8/+MUj8q3yoH8S+Av6r7+7MdxvjW2+v8R8BbQ/9JHy5nuI/fX/Y/QLz9U99H76z7yxuk+UnQffU10/w3H1sB7dKwfoqvAD5zjybwIfIvu67e8aS17PHCX/M9xNV4EvgXdCPm/DDwrB2LODTzvAH7z3DHnB7NXgLfJgci5Qe4dwAeH9w/UTTmQP/AI11pjAYzPffepj/J3fUzQfQ3APxw8GL8b+HHu4VgN+F7g1+RAPocSlvnhYfuLwHO6r5/+OrYxAT8oB/KdwL2ZoI/SNzx77v07KDrwKPg+4DuAPzzwAL+dEg68q3kP+f7D9qH7+gsUjpsH3gv8g3va+npgq//n8Amv/1p0H3mw7n+Efvnj1v2BB/ia6j5adB95Q3X/dcHWwHt0fC+Fy3BCCRGu8XeAH5ID+UkKJ2QMfBHwf+q+nrzRjdxgPFT+yKvlj77l5P9zlJnqfyEH8jeAfw14DyXUcj/8LBCBPycH8reHY38/8FPD/gfq5kd5rV8B3i0H8nuADwDf9PH8yIGUju7r19xn39cAN4Y2LiihqncP7YXiGbmrXpXu60tyIL9ACdv/b+dCkj8PnAwD5N8CesoAOBoMmI8bA23j3wYuAO8HvvTc7kfpG/6MHMiPUe7FX6KE8x4FM+D2YNz9fgo37Sc+iqa/So7nflMF/Angx3Rfj+RAjiker9cbW/1/K+m/fGTd/wj98p9BPnbdH4y711T3h6S+PwH8GKpHyBum+68Lthy8V+NHpRT8XP/9EICWGle/RLH+T13cuq+/SKlZ9B3AHeC3ga95w1u9OfjRoejk+u+HAIY6S6+SP7o58h8Set4D/GHgJoWf9FW6rx94yPF/jPJ7b1MIzz94bv8DZfPRXEv39TeBbwZ+Evgt4J9/nD/1WeBnHrDvGPhG4AVKDbD/mpJZur7m3wT+HTmQO3Ig57MBvwf4bM64mWvv2pdSuFrPUX7nd1HI2x8vflQOZD6091uAr9Z9/f/OH/CIfcP3UQanD1E8T49aKPY/Bb5ZDuQE+C8pXoaPBg+S4xpfScm+PQb+FIUQ/7piq//AW0P/Sx99TvfRu3X/Efvlj0v3kY9P9xG5gzxY9wfj7g3R/dcLovpwb+UWZ5AD+bvAy7qvf/kjHrzFaw8p8ke38n+rYvAO/QrwrwwhqNfqvF9A8fK+8x5O1ics5D6Ff7fYbGz1f0Apk/K19+Vbb/GaYRuifURIqS31x3jEbKgtXmOU+kZb+b/FMXhOPvO1POfAE/vPgO96SwxuWzy22Or/Fm8ktiHaR4AcyF8Bfg34a7qvz73Z7XnsIGfyR7fy3+IMUorVHlKKpd6/4OoWW2wotvq/xcOwDdFuscUWW2yxxRZbbBi2Hrwttthiiy222GKLDcPGc/Au79T3dVFaa5k0NaOqpq4ctbXUzuKswRmDNVL+BIwxeOswxpJzxlqLaqZuaiRD17XszGaMmgZB0KyklNCUmO7M8EZwziEoQrGqBbDWoKqogLOGr//uH5T7tfWtjD/3pX9QBXjq8hVCSOzOdgndgtnlJ5hUEw5vXefk+JDLT72d8c4FJrt7GO/QoFTWEmOPNCP65TFSj6hGM3Ls0WmD9hlJAfEjUruk7Ra4ukJiIHYRZwxtF3GNp+1arDP4ukbcCDEVqT/h8M4rpBhRVd77zX9to+T/we9/3wPd89Y5rLUYaxExRKP4usJYR1XXgMHXY1KMkJWYAtYEdNHiraIEkrVY4/HWEa0naSK0C46P57z0yk3+j3/8U/zzX/xl7syXiIWnnnoK4zwvXnuFazfvYKxnNpvyBX/wC3nf9/+DjZI9wN/7hv9IQ8w453DekVIipYQVgzWWFAPxZMHN517kpRde4u2f/Tlc/ox300wv4eqa2C9ZHd8mL5ekvqWLgVXsuX3nkBtXrzFxDgvcPjmmjYnsPJ/5rk/lyd0JN59/jsbB5U99J74aY63FOosg5xaEOcOXH/wPGyX/7/q2b1SshyQQE3gHef04CIhFETSDWAHNKIqYjBAguaFamuG0bFqMYF35LAARNeve3INYCAHIiFWQWL6eM4iBFEEEsgVTlXOnnq/9839lo2QP8BPf9V/dp+9RUEWyAEIUIUvCoCigAhmDqGBUsApGhKxF3kXSZbzMg8QEkPWVtJwHkeH+PBr+za/9ho2T/xobb+A9CKpa/li/nlU/1OE/VYqirBXn7NuICKqKEQGFnDMKGCmKenpkVjDl2NKxDt/NuRiD3O/8m4O3X3k7WTO70ynRwKSZsDwJNM4T+p7ZZIfJeMZoMsXXNUYEDREjjhgCxkAMPSkpjVhyvyQq2GRRyWAtoV/gxxN2xw0pBVqEqpqifc9s7AmxY1LV9Ks5/WKOSgsJskS8b6jqjX2+H4z7UDMKXUOH97l0noPOopSO0wh67klZvxdVEIOIYKyhrjyjUYP3Z11MDIGmqqi8R6RcL6XMyfHx6/pT3yz0oceII+dMDLH0EzEjoiiRHHoOr1/nhQ89x4Unn6aa7ZCTErolXgRWS/JqCSmQuxVd23K0OGZ++xBdrTjOESPCcrWki0r2juc+/Dz+k59ldHGX5bVrpNUK6zxiwKoFWd/nxwFDBy5nb9G1cbbesEYx8s62aenI19bDXV2E3P19oRiPcu7AuwaT9dfk3DOmr97/WGCQqcrd0i+WGWa4XXJqRD8c69u5viNb3I3H2sDLWqy48qKn49j5h/f0UdSzB31trKkqDDM4VUVzLjM8PVNO1Yxih22yHg5f3cluaKf77Gf9LlDoFwsE6Nue0e5lprPLxNyT+oj1Dd5b8iAdo2C9QxVyajHG4FyFBfpVS7WziwBtH/BVxaRxJDHkPgIGV42QrNjRmK5dUfuKoIqxFodn2S7pFidEERJSZtePWfdwqu+qnLfh1tt0/Xk9hq11d5jQIIIME5P1kyIYEIMxgveOyXhM5V2ZBAExRawxOOuo64o+lMnVtavX2EQYMXhfkbIWr35MRaeJpBgI8wU3XrnGZGeHC089ReMrbB+IsmTVd/TzY1K/RPsVx8eHtGLo5wv6xYI+ZW6vOpaLBTlEnPc0OXPt2g1mjefZt18m5Ex3MqcaN6hd6/j5m/044d7+ljOr4HTXw2SzNiPO71bIihqLqJJTIqcIKZA0gMmgJZKTcyKHgKZE3/bEqISYqEiv8e98K0BO/7dqSmTrvEE8TC6Lp9ScOkJ0a8p91HhsDTwYjLxBdc7+P3vm9ZzSnTn3ZfDUlVAsdvC651wMxnLQ8CqkrKfXKjM9wJw7n3A2mG4ibA2qVBOLhkAzmhJiT0gZ19S4eoRzFc5ZciozaANoSkXSqmhONPWIvu+oXUNSJa2OqdSiQVE1SFwSYwZjcM7QtXOCH+GcJXY9tqmxxoA0jGaeajYjdS2xWyJIuRGPEU49CbqecnCXR3s4ilMPCINr2ggaB8+e5tPvoMVTbUSwxlJXFVeuXGZ3NuXm4TG2qnDWUfkKawx1VZFyxDvH/OSttuDIo0FTJLkim9QHjBVSisQckRA5uX6Hto08+fZncaMxOfZ08yMat0fbnkDfk7ue9uSEbtXSWcd82bIMiUXKnPSJo1Wgb1suzISR96S+5+Ubt7h0cUblK1bLFZOUkJxPLfjHyoMn9+lb9Z4PCpDRnFEymiKkiCYlK6TQl30hDDSdRM6Rvl+SY9FhQej7MIQMEzF0VM4imDLBSYHQLoghQI6kFOn6hLjHiwZ/ftxbR7/InBtwdfCG5nPx16HfkTKBPB/vknvOe/bp8erPH4TH1sA7DdHq2sA6e3+vftxt9gGqg1e+rGAiIkOIVk/thLUDOud8er1hxz0Nued1wzDemaEKqe/xWUkpYk2DE4O1lhwjmjJZDWIMOSlqLUImhh7rPeIcOSWM80g1QkMg93OodqCbExGkqbBVRW5bui7gmgoZOUywYD1du8Aah/iKNnSoQkhLUhdQzRvrQX0Q7nJcDBvWOlqoB6bQCETIrD14gjEGRRAxIJBR+hARjdiqdMDeWibjEc++7W08+7ZnuH7zFl1WmqYh58R4PKZerQgR2q5jnMb3aeEGQHXw3uTBSZxBMzkmuvmSO0eHjGdTqroqIe6cCd2c2LaklDAKXoR+vqILgXk/59ai4/rhCX40Qn3F9OJljg4POWmXTKyjFsutw2NeeeU2T48quq4nxYitSoRB14PqpnY4A2LfF49PluI5Y5jE5AQ5nxpvUDxIpB5VIcdA7hZIiuTcg0BYteQUQJW6HmOsBSNICIyqmtRGUugRBbEOsYZpU2PEIEDKiZQCIhHrlNAlEENVe1La7PtwNwa/3ODUMJqLMZcGT2jOp2Myw1gqomQDxjnUyF30p1Ns7bkH4rE18ADyKQ+PYSL3AOtOzoy/4q3jzPXG2sArYZi7vP0D1+7srHeHZk9DYKfX3jzYpgYg5kjbdjSj4oFDBE3F+LXWEkLAubU6KkkzxjmyZqwYqCsESz94+UQciMEa6PuWqt5BQ6Tte+qmJmMgFcJuThFyxjpP2/WIK4kFOxefpMUP5OvNlP8DsdbDezx4DM9Eya8vWmuMlMQiY5C6hkYRA8ZkTA5oSOR2hXQrxIATQ+09Tzxxhc/4tE/j5Vde4eqtQ7yzjJqaRdsyqhqSCqvlihD6N0kIry9EBB0SeMRZNCZi39OvWo5uH6ICk50pYgyiSmo7NCfCoi/9iRQierfsOA4rosK8jywUpO0LbxWlGTmiqTheLnlmtkO/WPHi1ZtcePuT5NQR+0DVlBDi/RIsNhHx+g00ZwwJ1XRq1J4FWkt/LK4iKWjoAEFXK4g9KUVS7lERjPEY4wotQSxiPFEztpqQjSEmQWUIIlqPrw3GlISWFIpXz9sKcUrf9zS1o29XgJLkMet3gDUPL6cEKZGTQk5ozmQFVYtoxqAYUbJRhIy4avDgwek4/TiK76PAY23grbl3p0kW66jV6T+5h40H5zkAa2NNjCkevNPBUk45SymfhbEUPQ0FGlMyaI0xCPmUp7RpyLl4x0QM1ahBrKHBQcqk3GKNYBCqqirHSgl1K2W2plmJMYGLiCrGWJJaMB7rHbFbYr2HPqAxMNnZAUCsJSclrxaE1GOco+sC3lfkyqG+InYBsRMwm2pe340yxq09yWeJP2se3jrJUICUEp5MCIG68iW8bQ3JlMxbg5C1A3wx8GwmdQGTDc54kvSM64p3vP1pnrp8EWMcyxjZm005OjlBlMJZyonlcvHmCOR1hliHxozGSCRDH8ghELtA3/VMpjOqpkZzJvWBFAM5RnIbi7PPCMtVy+3Fgufv3GbnyhXmXUc9ntCHgPM1MScgUdc1fYjknLgyHnH9ZM6t+ZTLtRJDIKdEIhbDY0P7mvMwEsEoogK44vnRtedaiXHt4evJKRLbxXAfVqSUqHyFYIpRbC3OuYEepqhmvLWItSiK9YaYwRpDzh20GbUeNQ4Rj8aekCPWOpyLhL7DWAMihPA49DyceuWG0ZGsQyg8BDQXnmKZ8CtWDZDJmkmiiHGDwQeighhB19kYWzwUj7mBp+e8FmvDbr2Tux15w34ZDLyzEK9ijTl1HcdU3M4hJubLFYKSYqSqKliuGNUVo7rGUgZXMYK3HvJmkm1DKtwfca6kw2tGSIVIa+wg1yEzU8CYqoRU1lx+ldLJqkNTRA1YWyFUpJwQX0GGcHKMa2piKvdEkhKPj8lhRbV7maQRaxxJBBNB0hK1FXEYYDcRd3mLT2kI679iePvKo4C1DmPWk5KEd56UIsZCiB19yvjKF6+qcYgYJEeMKGogi0W8RbNAiAgJI4m9nR2euHyZLkE98MCqqmTRxhBK+DJtZk9dJo1a9DQVmkLsetrlEjEGV9dDIlEkayT3kZQSZBAxpBg4nh8zT4nsPV1Vk0SovcdZR8zgZlcIt67iBExdEXLi4qjiWht4+dZtphfGxD4U74icUUpOG3j3m83BaUmUdYRmPWWn9EGpRzQT+56wPCHniFgPMeGMw5livLnKI9ZixJAz2NqjYjDOl4S6wfPkfIVoLLzUnImpx3oBZ4eJK6zaFaNRg6sqQrci9P3m2yj3qtYwmZSUyTGgKUDfY/LQL+WES4kkSpaMeotNQs6Fk60OjK8wYs+Nx/e/7n3DuY8ZtgbeaWD2PB/vjHenazfwMAOUc5xYESHFRFNXGCmzuMV8xclyxXQyKcaEsRytWmZSMguPFytCTIyb4hVJKL6uIW8m2bYiFa5WzsUoMILkwt0qIWxQMjnrwFcBEmC0GBO5zPlStogZslP6Hk0JO7Koc9B1SOVJeCSB5EyMEU9FrgySIilE1JoySxRBcibnnkQYEm02rzeIMZaZ8mDQnU5mCuGOnANiGqyrEWMwlAzwGAZPz2AEpqxYXyHTCcaOUCDnHpMzpGI8WOdJKRSvqiomOTyZ0WjEpUuXeOnaTaqqYjqdMG1bRs0JHJZJzmZObYqNoSgYIfc95ExKEVQZjSeDPiualNAHUtcPfY3F2lJm5Xi5YA40OzvQNOzMpiVMZRz17BIyfZLbL7wfITNuGmJKiFh2vOGwCxyvEsv5isluwIsbsvo3T9fvhTlf8mSYwBdvdSr1B1Mi9x1hMSd0LaaqBuJ/6Ye7vsdVxZAQMagYfF0V7pj3YFzJwPfF2yRRETVYHCl1gKIpkFLEicU4T1WV5ywrmBRwKWNNfLNE9PriHsPOwNn4mhPEiIZIjAHTB8wwwKacQCMqhoxBg6DaYwTQhGQdxhFzVsIJYJ0ot+YRb76KPxIecwNv3QlzzmNXTH8xxbBbq5DBnNbuKoWOy4yj8h4xhlUfOF4FUi7h1+P5nHFT03iPqTyrrmM8HmGqmr7rsL3BUMpG2D5Q19WbKInXD+vCMFbAGkgpD0YbQ5ZxGd5VZYhel0xDZytiTOUhV4U+lIcaiKlwwOw6tLKc45o9pCqcDbGCYLCayFnpug5fTzHOoDnR9T3KULA6puLN2kD0fcupElP6v8KdE4yYIUykZ3zRVOoK5hTJKZKMIKlku1XNCCuuJMTkjGQhY0EyxhhSCIMxmEtmYDSYZHDOsrMzY9p4DpdLrNlhb3eP67fv4K2hj3d7GjcJfd/j16HBwUOBGIwtRY6NOFIW+uUKYsnSFDIZoVNh1UVutj3qLTFmvDPMpmMWi5bxdMaFZ57i+kmHNRlrDI3zmL4HESZ1xfFixe12RT948FCGTP/NlPddWHupB97zOtktp0AOPblbsjw+IoWIWIdFICeyKq5yiPXFuMMgxmGsRUUQY8la+pdSGEgw1pFTxGAKv7jMWompB2PIRrEUD3lSRYYwuYhS+820ROSe92c+kpJ4pH0g9z1JAypKzAnrPGZck5xBcFTZQpdJcUVOLSlmjFiMi6VY9LqExXrcvodMtcVjYOCZIdtvXXlnbUSUDB3BGlMGOzO8N2eJ18YY0LKiBZyFNkRKSrxxjqO2Z75sUYSYSkFfI4AzVJUvg6gqs+mUrMJkMiKmjPGepIq3lj7Ecq1NhLFAmbWlGMiFT4sZXlNMw2yrpMJnzYgx9KE/5buUMF4io9SuAmtQ7+hXPZIS+JpIxoaeHMA1HiXRhTnGGnzVgBSOWNu3JTSGgaDYbLAbOt1zXk5nujLUkwIzzHSHba5CvSPbcp9IZ1xSTYkcE2IMbbvCdwGxvsguZmxtycZismK9w9mBsB57cIqkTD1quHTpIk9evszhc89xcnyEbSY4a6mqikVbyO2bCCeZnIaMzZzIMRNCJGkuej3w8dKyo1/15BBw3tIBx11PFyK9KOPxGD+ZFFk7S1VZUrfE+czFCxUvkko/5Qwml0xmi9DUFSvNHM2XPBkj1Oe4w5sp8jMM3ueiy0NJnxRJfUe/WtAu50OI1IKWOp117fG+Og37iZSxQXRIrsBgjEONRcVgRSBHJAE6eKKtRWxZiUGjkjKoKxOplFM5cy6ca7UWp5vpv36YeuWc0BjQEMmSEO/xVYOtKox3dNbTh4yNGbUZZxt819IuD3EhEE0HokjVlPsha27fOlnsIzTgMVD/NTbewPPODd5bGZIf7n6/Nu6MMadG1pkRaBDNWOcKhyUXY3E+X5GMo+962r7HGoszBmcNdV3hrMXbct66qrHGEnNiNB4TYqKu6sIDcZ6qGWNQ0oY+6CaWEgWIYJwntz2EWDrDXLgYKiDeEXNGU1uyDVcnaH+M9qGUNYglrKvVGPGuEKibMV49KhW2KbNvrCEGECM4bwkxYCUR0goDVNNxmUUn0JSpcHdzLzcIzo+G6exZd7YucFwc1blwjmqPqUpCgAy0gnVR7uxKYWIRCKsVIS8wVhA1pVf4fQAAIABJREFU1PUeqCOTh2QaQbzBrVchsZbKe5pRw3Q2papK9uFyscC7CmtLqZyYNzNMFUIsnINcdC3FSGg7Yizb+xjRLtDPVxwdLwgxYrylVSVag6kqprsX8PWIau8CaTrl+OQQzZHlyRHzw5s0e5dp/ITaGrwzaBxCVznjpHBgTxYrVsslzWhyN8dkg6E5DXadntZszLGnW56wWi4IfQ9qiu6FQFV7rLElWxZTvHcKmobSVxmySKnZJrZQ/EQxSqF8GIemBGpBynJmxq3DkqV/MYNDIQ8exZzzxk7szwKnp2vdAFJoHzEQQ08GqvEOrqphNCInoW9XJJNo2xZCwI4dWSyjuqKSHdK8JM1JlOLNqwxZSu1UTq95rgH3weNi3MFjYODV3t91R++J2g8e3lPfxlDfq2TB2uFII0IbIykpXVYOF20x1lKmaUbUzjKq6/LgrsNUwKRusN4xGY1oQ0JMWf8TW7KFnLOIkRLmtZv5oONckXlWUlrXODLDjLpU9lcxpFiyZ6VbwfyQxfWrrE6O6HPmwuWn0SAY77Bi0AS2W+Iay2pxWDxw3RiqGTKbQqKQdl0zPOwR60aodaXDb8PQtoo8qu4qnbBJ0Mw5zh2vei1Kb4hdjyTKWqXGINaQU5nYWGMK4b8PZE2otZh6jBEhdC0aOyRErEghRONPw5CEgBWD9x5rDTuTKUftmpcEFos99SxuHqwZwrFtIIWSJZhCyQbv+57QB2LbcXQy58bJgk7A9ZYswnjaMJ2MmT3xBBHDeG+PO30k9gGTApI6Dl9+kXi4ZDqeceXCLpbA4fVrkIUUIyJQuQqSDFnPCUHOJrKn/20gNKGp9OplMhnJsSOGFavVCvSsZJVzBmdKIkXx7qwdAGuCiZZnyZyOEiXSIwK6picYjPihvp4ja8RYwWQIKZG7jrppThM+nPdDgsYmTi3PB0wLklhMTiUSEwIpJVzdUI2nZCO0XV9CsFno2w6XheOjFRM7YdWtwDcENYzHY3Q1p1fwDOWbnL97VbnzA/wDsKlqfy823sC7X0UAOd0np1k95pQtVrx61loELWFWY+i7CA6O+wS2eH3WXrvQ94SuZbazQ06ZcV1zYXeHNvTs7O3SLpZc2JkS+0jlPVXt6duWcd1gpRRM3lAaGGqLiim5hP+yrOu9lgGPjCYQDZiw4uTay/SH18ldj+QKO57SSYVxFuMdWZWmGaEG1Db4iUfnS8LRAutXmNSTp3tk64d1Ud1gcJewSRqYvilnMn0pkbKhZSNSHPiNa4OOM2/C+lWsgzaSTCZ7Bw2YyoGxxJwxWPoQMM6CLZyxCgfWEFYnpPkd8qolx4StKqrZlFy5IXQCkJg0NU1dMx2PWcYFhMjJ8fFpmZZNLREk69VucilV0i9bYteTUvFQtKsVtw+PubnoaI0nG8vObILGxGg0wVpLXdc4V5FEmM12OUyK5ISVzFgysXL0tUWJXLh4geXhHUzI5FSoD3U9BhViH0vhalOSnTZ9iDurb1oyMzUEutWKtu0IIaBZ8L4a/Etm+CteO1FK4d2Bn1o8bRQOY1asNYhY0tqMMUMN1JTICZytURyqEWshiyGjZWm/WLzVIgZf1Wi2b5KE3ljIOoM/RWJKqBGqaUkYSgi354fkZcTbMS/cOeSkT9w5WjF55YTdKrOzO6aZTXhip8FhyDGhEhAXkcqdRhxOx3YezsXTe143FZtv4J0R8O7aeu+tXSdVQJl5Q1k0vaoqjucLFikTUiLESFN5uq4vAxSKM8J0PMYJ7O3uEkNgfnLMhYt7xL5nOqqxqvjKEVMg9cpoPCKmhDGu8HA21MITVzowg5BjqR5vxJBSgJwKH6lvyatDTl7+EP1iXrh4YmlDTzfPTKLSuJr+JOGnDWBpag/JgjrMZAyNEm69jFm9QkVGJhdQ48B7VEotMs2lY1EodZSIsOqGhm7egBdTGYKy6inJv0wmzlWNj4UPFESwlUdVqc0IcTVqDBEGXpFjdbKgP57T14dQlQEs9y05dMRli0WZeIdzY3Iq3DoxwmQyZjqZ0Pcdzlqaqi7cSOcGHuxmDnIKhNAjMdG3HX3XkUIgxshiPmd+MmfVRzrniPUYJw7TTPC29FujesRoNCIaR64reutKeY5uxc7eRd792Z+FufwMH/71X+fa8x/mk975Nrx1uOIWJ+dEIpGBk5MVs72MrYaA2eap+13IOSOYIoO+p1vO6duWENKQ0FX21U2DMw5rXXk+UsJILMkUKRdPtAqqgrWurCsc48BrBRCMq9AshW4cExoiWFtmsSYPSTWJFBPODqvAABhTygptNNZZRkPmfozknKnGE9Q52j7TdYFwu+XGrSNevnOVl6qGO3XN0s4Y9ZELq47x8SHvvpDwoWV3p6GOmRwise0w3iHV3UmK61WyH8XI22RsvIGn5/7Te18BZ0u5knWYbh0uMmIwzrEKkauHC3SoE5ZiYpUSk1HNqKmZjEeM6wqysre3Awhj31A3nspaNCWcs1ijWGdBTFmAHcqsI8eyZM2G1mJLMZaQt1LKQlhDzgFjISpoHwjzO5y8+AFEIUTDSuc0dsTvXL3Byaqn8hV7vqHXTHPlEnu7S57cmaDjC4x29sCXmTdPPEmYH7O4c52xJmTvEiGVjDdyRnKpgl6W21LQAGkI127g475qu2LIDckqp8sBDa9kRW1XiroawauW1USG+mwYj1CMlNXJMcfXr9LevkG7WuDrMa0GvDNcuHyFqqpLCRAMGA85YgbZjkYjdvd2GTU1nQYuT3e4dv06d47nCMUjsolIMRBjQGIihUjoIt2qJ4Sevg30fcaPdrB9JlQeKw7nBOssvrJUe1PcuGZvNOO4DSxChGoEcYEaj9QjdmZjZrMZL3d9KYGSFSOCwUCKhD4gkwkpJUIIeG8ppOTNlPkaMkRGMpBjRwodKURSTKccbOdqnPWFaxdL5pd1vqyCY3zJ8h9sE68lOcAo5H6FdksAFIOpa0QNqEGyFFsmrSt2lILt1vnTeps5F6NP1hyhTcc6qUGLlxPAVBXJWBZdS+wShzeOeP72nN8Y73E0u8KRWG71icnFhl3NTI9vMnvpNlfqXdqdMbPxmDA/QWNA+gjWI8UWL1j7ax4D8T4MG2/glUFnjWHWNfDuzBAeWtfnkYF7Vwi0hpiU28cL2piZjSo0RZwxzEY1F3cm7E5HzKYTxqMJInDx4kWqpqZbtViBNCy/1bYto9oXXpMVjChOwE9GiLF0XVm6aRNhjAdKeQJSgqxlhYkwrD+bM93t6+Qu0Eli1Qof+ODLuAi3F0tOQkcUw27lGO/OSPM7zHZGxE97lrepAAnrJzhbIVWFv/g05ugV2uObVN6hZgfnxuBKiQ+J8bQOmKgQBg/jJvYDYb68y6DTlE+NgLVHL1iDna25Yj3R1+Q+40aFNxlCT3t0zMmt28wPb7C4c4M4XxCyIOOGqq4Az6Wnn8I0DSqGEAt9QcrwhzGOixcusjOeMl/exhthb2eHV27cwrvi1d5IpFTqLcaEaKEmlBJgSlAhGV+KdkuHhIRtXEkK846dK5eop2NyVg5v36Ge7OCAZjIj9Mcslgue++AHeTImjm7fRHKmX7bknLCuLn3bcO0QI1VVE0IJj1kZIhYbjJxK0lVOmRTK2r4hxmGZRIOIIyel7yPeFWPOuYrkHMbVaBJ0mJAKUla+SGVZvn6xpNKSjNRUI6xtyFWNxrZk1Bo5LUujKGItIrZkUg9ltMxQ5H7jHXgADFQFhBAjWMMyRFJqOTlZcOfmIXdOFnR+yvM7u6yqGfH4mHg8Z5Ez7d4u7c5THHfKyy9e4+lZw6Und6Cq6LsVpu8xVQWmJFSeN/IeoWkbjY038Jx/8E88TakeyLbGGtYf+xDoU+Jw2eKcRVPkyt4O77hyiVFdPBvNqGY6nVB5z3g6BYW6qYlNQ+47VMs6rLPpmBA6lIxZzxjJxL7DGkNde9yGdrhl0lqK50pMZZDLlI63X9DeuUbwDrlwkfTyNa6+/wXef+s2bYo8aRuyGLSuaLEYdcjhkspXvPjhlxldNlSjY2bjPap6grlwEVNVmJ09jGS6o5v4qib58emDb4QhtC7DereWTfTeAXQni9PixnfVAx0I5MYWw06tIK7wWnKKhaO0XCJ1IqxWdMeHEDp8PWb25NtZjo4Y+RGj6QwBZrs7THf3iBTDsT06wjc11pXQVs7KZDLh0sWLvPDyVdp0zGwywXtPVVV0q83MojUq0Ce6VUvSUnfRWE+ix812efrtF+j6yPGN69AGXEw0TU1uaux0yuVn38btW0d4a/Bi2JtOOKwbctXgVbl94yYnixUnt25iU6Q7OUGGOp2GQvAnZWIsyRVVVdGHQFVms2+2eF5XaCxJXKRI7AKh7wsdZEiniykXWTlPRhDriWJO+Xd9inQJViESQo9JmfF4zMRWxEWHzUI1nmCnDYwmyGiMyQrtkpQYPKRDOSJK5EallEsRFM2xRIs2s+s5xWm3MyyJmFUxvuJk1TFfHHF485g7129zZXyJdudtRFfTJ0e4dQN54SWe/ax30XKBpbVcG435vNknc+O5FxmJxVXKTuUIoacKERkqZpxd956VLjZb5e+LjTfw7ndPC3VCXnXAut/LWpYauz1fklJibzJmd9LwSU9eYm86xhtDMxpmyQLjUcO4qoihp9KMNUp2BmNLyn0IPc77snh4Dhjri9va2nOz6c180k9nsjFBV4qwZk2E1JNPDtHVkrqacHK04oWXb3KjXWJtRTZw7OBw0XJrNUcz2MM7PDHZYTzdY/niVcLL78ddusjbn7jM07MZMwExCZoKd/EK6fY10vwE9Q2uHjFUOygZV/msCOrGQilZqgPvR8z518JNNcOamuIduW9Lwd1igZNWK8JqSR8D1lfs7OySJOEmM1wzohk1jJtRKf7tKnzlkdgTUy5FxHMmxozzHuccly9fpqlr7qxaMI5RU3M4n2+srZFiQmMidB2aS41HrMWOJlx58kkuPv00Lzz/Evb2LfbGDixcmu1g9nbZu/IE9XSXnWgI8wUKzKYTJrNd4vEtJt6yOrxNdzInrlZUYlgcHeOdxTuHM8UojDGTYiKul0sc1l3e+PVorUMUYt+eZm2GVCoZ5CzDZK8kV6QMfVLaruPq0ZKbi8iNw2OWfU/tHaRIbQ0XZzM+ZW+PC94y8TXZGgLAagmrDgkdloxkRyaQyRjHkOiRi8cuZ0SG0iuYzb0PMoyn64/GnC7xpgixT8xvH3P4ynUmfsRTV67wm9GSpSKfzOlefh77oQ/wrs/9NOLFPX7xzhFXveB3L9McXuflq1fZvTxjNtkj9its7DGpQnBko2gx208zatdrET9uRt7mG3jy8CIM55cry7lopEhm3gX6pDSVp6kcl3dnkCJGldlkhAg0dY31jtpbTOzwmsn9sI4nuVSr9yOa0ahkC62W+FgSDIrxl3FVTUiZDaUhDaVQCnlZQ8A0HggQWvJqga8mrFZzrl+9xu/cPmY5mTKbOmY7CV2seGkx5yQNjMk+srIdR7/9PO7WHZ6oHJfmS67fuM6nv/NtfCrC1GYqf4FcTah2LhPmx0i3RMWC8yW5YqhyL7LZizY1o2aoA7Q26OR0/WNjhuV+jICz+KYmI4Qc8ZqIJDSXQsmjelzI56YsAF75Umy3qT3e2OIxcrZkBXpfspRjj6scfiisbK1ltrPDqG64evsINxJmo4br55YX2jRY6+i1hOQygjFQT0ZcuHSFencH09TsPfUEt+7cog4R5wxjsfhmgvM1xlZMZrschcBkNCZWlotXrjC//jLTUQPtkuPjE4xmvHN0iwWT6ayUpTEGZy0xxdPwJJQIA9bCufprm1imRgWypJJYJXkoPuzQPKxhKop4w4nCrZtHvHLnkDttz/WlsgyFvuC9Y0ccuU+caM/t+YqToxPe9cQFxpcuUFcJJgYjCm3AdB0YQ8gZtSUrN+UAsWSh27WBB8O63PDYWByD98RWNUEzh3cOOb5xmyYb3nbhClXliAptVlguybdvYdtj8uI20xxJKbOox1xdBT7jiWd44YXfonnqMsukjL0n9B3O14i4dR33LXgMDDxY51iszfd12UU5t19Pax4VNzIs2zJ72x03PH1pl2cu7TGuPesCA7V1eGvxRrChHzI0SpgrpkwXOowR+r5HcDjvqeoaqTx9uxxWzDBoKsuVbaobw3tPzkoyLVIPNdJSRk+OICdWbeL68x/ixatXWTUzPvndn81vfOiXmT09ZnU9s4qphFUG4yTGyOHhEU+MG05q5aQLPOPGHN0+4mV1vL0e42djpB4TxeOaKeIsWQNxyITTlMCUGlRlFYvN9KC6pj415Iw1w/szL976d4dUarP5cVPqNKKkPuCGbFddL3IRQwnfIkhypL6jQ/HOlUw2a7DOgbWkqMRUsubWxcVnOzMuXLjAbzz/EkEz3lmcG2pDbiAyMoT/XKncr4lmZ8LO5Su0MbA4OSGKMtvZYXX9OuPRmLxcon0khogmMFjEWWxdBsbpdIxvpohZMZpOOTlZYp0ryTFGqJsKZwy1Nagp9yDFTI6ZqJnKGHxVIcZttGlRSpaEon+qWMCJIZrS9yc8R63ygZdf4cbRCTFEcDW9WjAl6lDVDb/v8z6P337/r3P71k1CVo4yvHKy4ELlaUYVfjKB0DO/dkRde+onLoMuCzUBQwgRo+BNhbEOQYb1gDldPm3jMdBiClPElChKVkwbuVRPmY5GuLqm6QK1ySxdQo7vEHPLb/zar3CJHboLT0PTcJQDvnFcme5x9bde4lM+59OpRyPiyQna9+TKY9WechvP18d7HLHxBp5Z23WyzpE9C82elVCRUy+eiGHZ9aScmDnPE7tTdkcNdeWpbPHWTeoaZwxGtZD2pXg6VEutqa4vta4ChVyN9GiqsM6RYkDEltITg3dLY8RuaFZbigml8IKygThvyW2L5AjGoe1tDo8OCcbzu/7Vz8PPZtjnEzXCMmTy/8/em0dLkuX1fZ/fXWLJzLdUvdp779kYxmKGHYRYJGSPLYPFETIymGVsIxnrgKSDbWRJ2I8nMT5IMpsQ2DIIM9LAEdYIEIuOWAYh0DBi8EjDDIhh6J7eq6uqa3lbbhH33p//uJHvZb2u6qme6e7qyZffc/K9yIjIjMgbN25872/5/qxDNB7ERoIQUkS9cOq155g2A17z4Jupn/so13ZvUF6+SDHI1qhkKsqqIjkhxSbH/xmXr1UyqCiRWYbX4j3u6pX+geXuEHmgTShoyqXjUiKEgEuKRXBVSWwCyVhSSjRtQwoNIUbaJmKtYRQC3o0ofU0yBlcXyHSI1DVFXaHJZt2vlLrARyjLgpPra1hr2BvuI9ZSVyWTBU2yiCki1uZMTs1iuUJiMt6lXl1HeiXbl5/jxOnTFEaR/V0q72j3R8SyRk6DGKVvC1QThbOQEvXqGs3uPr2VVaqdISblOsBiLf1eD5k0lEWB05pRGud415T173Ksu2T5j1spSC0KQgttIKWAkq2pRiE0AVCqtVWiFrTPbiOuJDURSQGjJid/iWE0nvC7H3g/Elp6zhBUOH/PBfaHO3zkxg2KXoFbWaGgoL/exzpPjIFEIqZITJEUY366OIf6TncyRSSn7i7wBZiDgqhiDOyPG5564mmuPXuVk9HQD1lpwWikIOFC6FKzHAl49KOP8uTQEv/En6Ja6bE92sP0HZVz9GLB7uVtVh88jRSOMJ2Qao+REplzx+rNw9+xwuITPDtTbT8czZ4XeNnFoSdVooFpGxnUFfed2eD8xhobgz6VNVTOUhbFQb3OSMqB+saBycuIyTFmzYT5AliaIs0IjC0Q50hBMV1xd4ktMTavYKu8cki7Y/JdFYkSsSnlhItqgDa7DHf3KPsrPHjfvaxduJ/n9q+hhWUYI7vTKUVdMx6OcmKKgnEOUcEYxVvBrfc495mfQ3/vYS795rt55sozyCCxPtxmsH6e5E+B9LvAZiV1blljXBZf1m4AWMSBVmbEIuUJjN7qlV2Is6lPbANqDTYq7XhCkyLT8ZgUQhb0blqMMVRFSRtbUtVQVhVYgIAmpTAGQ66mkEtFdTNpY9jY2GBtZYUbe3s5azFG2uliEjyNLXHSAAZX9FAn3YSiE/i2JaJKv6oZFyVrG2cgQWUF3d8j7uxgjMcYm7MQNWGMsLZxgqe3n6VvPYPBCrvjISKK9Z5eVTEdT/HeYanwTUu0Ngv0drqDWeNzMa2mMxjVXM1DE0VZ06ilbSeMpw29ylH3S8p6jVNrq2gTmKaQLdkG2mmEFFnp1RBaJCWsdYCwcWINd2qdvndcfO4K42ee5dzaKmuDdYK0NNMRIUGIDaFtEVok5ZhsG8B0VYsk1z5b3Bi8m9AFtysUdcVoMgE91OkMIVe4GJQWmghljSn6gEFTpB2Ns+RPSoxTxNmKyhlWTcnulRsMz6zQX6tpb+whoQXr6QakY4/FJ3hHav3JLZagi7+zuX5kjJFTa336VcmgrBDNlkBnDJK6mqcqSFdT0BWGlMAWJUlyJQDnHO10TGob0ETTRJImyjKhwWOtzXFKUt6c2r1gEGdy3UURjJqsF99ZMmPbEEKiXj2NP3MfRVUhE0O0nigWJWKcxbmCtmmo6h5f/y3fwr9817sIl59gfH2P9fvOUa/2OP2Gz8WnwJPv+WU++KHHuWdjhzc+DHhwxuLKCqN54E7JIMaB5Fn2gU7TgmHatIdErpvFzGtBdkEJiGYXoC9LUopAZ3ULLc1oxHB3j7ZpiSlmcWILrSikQD8qbYzodErTRsxojG9bql6Z44xQooLpEjtW11bYOLnOY89comnjQie55JEnUVUl0ltjfeM0/bV12qBE67h67RpnT57k+rMXWV87QYlBYsN0b8jaYIV0/TqmWkV6FakNVFWFccJgrYerV9ib7OG9pyxrRBK9XkW/qkmyi/eWNkUK54m+7IR6DdbnIPdckhEWs+eTybAIRVkhtmQyjoSg0Da42lOQJzXnz53l+vVdVuqSoJ5JC1amkCJOJ5hgccaRUqLXX+HCmTMU3jLc36O+934e/d0PcuXKJfr9HnW/T78qUFWqqqBtYhatFou0LdaCFUFspviCHh+LkoCaPPb7umafPYzLNdoLFCOw2ithZ0TqDXAnz9M+9xSaxogtKWyJimGiiojFicHHiB2O2HnyWQZvfj2pLDDTBvU5znTeijd3GscKC0/wZpa7Tvns+dtn2YSdFaONidIXrNYVtbOoKoVzeJsrR6ckWaF8Fr+UIjFMEefY291hNBrPZaklqrLAiiXFlma4T9u0lFVN1RsgziPeEyYtaUELrkc0kwVncWqJTLHWUSCMxyP6RUH//MM0g3WERF1UlL4mjPfRcUMYTSiLOluPQsMHP/Dv2Nu+xpnCUlWOqRqKfo0tLP2HX8/5G1e49osXuXjpKr264kJdMbAV3hkK5wnSxVwSQSIicWHv+rZtmWm5S1d1RbIuEOihAHIwhlBVVJVHg6AxElJkMhwx2dkhdhUAIPfrFCLDi9dYcRY5W1D0+0hVo9Iy3rlBHO8j6ysUdU005DrLmusu13XN+XNn8b//KOO2yfVaF9SKEULE+JzQUPYGaFGyMxwSm4Dr1UiMxGaKiwFJuW191WO4s41NkTK2hGF+EKoktGkx4uhVJSdOn+HyE3tMRuNM2BT6vQGltbTGEnoVe1HxvqCoeogYVKGsKmLipiSLRbwBYgSxBfiCGIXJeI/YKEYN4n0OkxkH7jlzmkuXr3Hj6iWqssf6iVX2R/vs7O1mzUCTiTFiOXXuJKdOn6CQyPbFjzLcucFAphR1SbQN+22iGpzIpeQKT1X3aKcNru7hvEUldlZFUKOYxVerAboxoxt1jTU47zh5egOu7hIVUhswIXGy9qwlZacs4Oy92GeeIAyvUZ44gzVZXiwpJAwxRExqqCYt7TPXmZzbpTo5IO7vo00L3jObYh3D5NkDLDzBO2rB46as2o72dZpFSbNAa1k4CmvpVQWDuqRwOTMnxZgrAnRJ2DEldocjbmzv5Hq1ITINIVtJUsJZQ+HyazDo413OpNUUEYWy6tE0DcY5UlhMF60zNo9iRsEKWnrSdEIaT9CdHXwb0KKCIpcUq4uaftkjlJbyQuLJy48xJeF8zj6++vRTxOE+1amaot+jOnmWft3Dh0iLstLv8dCFs0yHYy5fvU5vZYXKV4TCUfbXwHts9kUS4/SwRN0CjgAuNlmao9OfmsWZHmTWWgGxGOcySS4LsIYwmTIZjRnu7zPa3aFtWtSV+VpGw97l59Bnn+P+T/1UemsbaPKo7WFPecr1FUaXLjO+/hzm1FlwFsTnup6qOGc5vbGB945JaNGkC6sBaZwH40hqUFXqqia00xyOIZGVQY/da6OsyxgDo71t7nnoQab9Aeod1uTyV2kyzqEffoorPAVw6tQJ9rfXufjss2g7Ya2uWFtZzW1scmzjtAn4JuDrOgtZx5iTYBLdOHg0PnNxIBiM98RyQDtpibZl3OwymSbKaWQynOKcofKWB86fYm9nJ2t2aosXKEx2o4p11Kvr1HWfN953nnK8S7tzg41myNi1nLl/g1GbWKlqbOFJwMpglRSnGBOoXUE0Di0HaGxIaQppiuRCMnnycwyQtTezBXN9bY29sE1TOWLMJfV02rJWKhdS4jlnGZ07i334dbjtdeSe+7BVTQ3UzqFiSN6TBIokhGtDnnrf7/HaL/gMvPWEEPA5sLqLDVnYIf5jYuEJnhwQvLnLe5NZvKvR2dUutMZSOMNKXVH5nJFmuviBhOapock6Pte3d7i6vceN/RFVXeOKgkkrTKZTelVBmkZWjKFtIpNrO5TesrraRydjpNMEKuuKwpe0zWISPF8XOfbNJJLNljzaKdPxLkwbrHGodUQDiKEsawYr6+ywQ3SWtdpzadJQVxX7wxGPf/gPOFsXrJ0/A0WfC+ceolBh+PQVwkfej370D1hrRmyXhvtP3wco0+mEUg1vwtnWAAAgAElEQVQhzK6zJYlDTeoScBbTT9JomCNypstE5iBbPJJr0xqN2NLjqgpppsTxiBRy7FZwPZjs4xC8GCbjKbsXr7DRRMLOEOM8dmWNUdPQqmJqx8qF0+w9pzTjKa4uMc4SAZvAimVjfZX+yoDLuzu0TcOi5nMWgz5tFAb1AFtksXM5yKCM+LqmV5WYkxv0T57k0u//HpO9XRJAVZKMwZeWdjQlNIlqUKEJnMJKr+LMmTNMr1+jHe2zUVWs1DVpPMSK4IoSXweKtqEqFe+EFBWdSeS4HFOWvbQL2P4mYoseUykZttm69vSlS8QY0DQmTBp83cPsGApXsjoYcOnaDuM2y84oYI1n7cQZXvuG1xOuPkO9d5l2L+Aksr5WUoaCam2dE8aACv1+SZhOcnUkY0EjRizJCTGFnHTjsj5faiZoCmg6HgQviw4bLNCvK+yJda6ORzTjlpBylRATGh5eqfhQu0dzegWvr6ccn6ftr2G9UDRT+saQSEyN0DiHGkcSx/TKNk+95wM8+PmfRrAJFzM5P6qPsJgj/e2x8ATv8Iq+8KUVMVhrKaxlpa7p1xXeSK49GGM3CGZNpTBNPHPpCqOmJRpL8BXPbI/orQhNG9nZG/HwYMCN/R0mSVmrPP2yotHI9vY+/X6BMYL1jqLfJzmLOP9yt8RdgXYBjBYhTVvi/hgdjbFNQ+0LYlEQp3uYwqPO4sqC9dP3c/nx92N7A85cOMHk0g7jJlJ7x4oo9953gvL0KWIsOTk4zeSjjxJ+/0PsPfUI+9evUgxKVs+eY/XUBQqrGJNATa7fiSEZA8Zi1d2sxrlg0MIdqMcnjaSYSElzybJ0WLLMOkcKCesKjGqOrxPB9XpUDTSTKUwaUs/Qjoa0w32G08C1J54hiSU5S1rvU917nlrXaIpEsbHG9NoeJiU0RSyd4dAKvbrizKkNPvz447Qh5JjWBUS5MqB0PYx4stnYYKSgKiumKVL2euzFqwzWTxDbhmJtjSjCidUVAgm3so4pLLp/jdKAaQIaFZWWQjwn+z3MfffR7O7QTxETAsSALwtiMhQx0IsB4wzWGFLMoSVy4BY3C9v3RQyxGHD98g2e/MiH8VY5tV5ReEtdeKoSnMkxwEyGnHYtWkeG04aEstLzOGs4USuro8tUbkihhrLuYas1XNWjVAFjcNaQ2inStpRWCG3oPEeKaotJU9RamhDQBIX3JGlJtLNSPwuLwzs7x6wbhJVBj7IoePbZp6AS2pCYTMeUK1MeXBtQP3WF4vR5zNkN4rTCmhrrDbK9w4nT64goWnnMqZMkU9BcuYRPgYv/4RHcyT7nPvMNncUu/13QLn5HWHyCd4cQEawxeJtdtKbLQEwpdgHJikoktIFLz11nezRiHJRpUhrjGCVlsr1PioHtvSHvv7GNiHD6zHk2zl3gynPP4rXh3Gqf0ajBisH7KaGZUJw4iS3Lu90ELwtUc+xWO55iMZhpoNkdYkdtrklrLKEJpCAEFWKvZOX069AnHyX6IeXZk5yZNuxc20OdpagrNh58mKtXb7BOn97jH6Hdu8HV555kb3+f9RMn2Dh7jv76BuXJM2hscSkSmkR0U4rCHYQfadSF1qIaDkddVQo9LMl3Ky4VEynGrKUmCRWDrWsKtUz3myz1MW2hrgh7Q4iJy8Ndru0Nqfe3eard53qacuHcaT7/j30Jpx98iEZbTFEwaRucFaSYVXYRisJz34VzFD4nGy3qNajXT+EGK7RNIo6GYB31YJ00HuMEUCEkQauS3WevsnH+HqqTa9y4+DRxZ48UAmVd4uuS6WgC+0OqqkesukQJazG9HlEDfjpmunsDQ6QsS2gj1lvqusI6ixHTVZXRrhbq4mpvAqgKqg5pJpxeqSlrl8fxLkyATo+urgs0Kisezq2dJPkK0ZDrhTtLv/RYXxI4j5qEcQWmKJBen8pYNARMmhJTm8uPGZvJiDEYU9K2LSklNExBE6FVQojZusdxE2nLIuu+cMQUGKz2qIuS0bM3kHZCHO1yb3+Fh0lMmoamqEiF4lNBRaSOU+47fYJBlRicXcUUBWl7n/GlMzz62+/n8miP+rlr3GtK1MySLHK5Mj08hWOFY0Dw5NbvblJKyTp4WaW/ynpT5GD8lBLG5mQBTZHd/RE7k4YboynTBDeGE3AVk/GYGzu7TJoWFUOMASOGiVS85Qv+OGcefB2XH/sIl688wZoXrOSA8+lwj3o6xRaLSfBmLkFXlUx29kjdgGck66+1u/s04SLTyYhGQXfW8H7Aaiq4ePEi0zimjbmwtxXP2ftew5kT97L3oY/whtMWfeL3GJGoas/JMw9RDvqsrW0grgINFHUPG4UokZSUMJxgqxKxilUQYxe2kkKa3py4M0soOpBm6OJRxRhSDNlSrYB1VCt9pE1U9RgtShpXMb4xZDocYkrDE5Mdnm1bdkdXmPgBJx58mEeffJztX/kXfN5nfw4Pvu5TEEeuHNK5BWe1gI0I9549y4mVNYb74y6udfEQBIwmjBVMWWaxoBhIojlOyHj6q31WNtZhOiaOxjDoU4nDiKEuLJPxiNFkjO9VyDQSdnZII0sxWOkmKok2BdJ0zHRvD1d5in6dk1qsw5VKXdcH8igaIuJBjLDIdVyMtahCYRS3toK1votDDFirGJHOopmQQrDeY4t+zva3dDWqhajge6vUpEykiyInvVif7UPGIJiDrGQjFlsW2c1OwFISxlPaZp+YyC7FBKaTbMIt5tgDcNQ/mrokL2MFX3pOndoghsi+vcGoGcNQqFzBnzh/mmsffZZ24yzJCSt14qRPvPlND/CaUzVhsst4MiFMBNNMWT+zzhs+99O58Jr7qU4McKUnzlfIOa4BeBwDgidHpABuR/AAvHN4Z3NlCc1ad0EjxIRIom0De5MJTUzgLE8/cw1T9Hjo/gf59x/8QBZIVrLGlQoe2N/b5qOPPMIX/tHPwD9wP1cEwo3LbO/u08bAPfc69q5eougNXqEWeWVhy5qYYpbZaCPt3jXG158m7QyxRUlURUY7sHcdFxLJWNQ6LuyNmTxxhWujISEEetaw6hxntocM/uARvvDkPdSDHjYGTK/i1PoGMuhj6pKyKhADcbqNxgnRVSSxGPG5CPm0BZtvfmPdwhoysptIDu+Bo9nkBzxPCDESU8L7AlcPMOUqDHeZiuJFiDESd3fpeWHHtOjJE7gAMSW2G+GLv+jL+MBv/DzS7/Ebv/ormMmEC69/AFeXoJGYpiRXYmyJDS0bqytc2Njg+s4eZbGYw1AmWgmNXfZqCCCKKT1Itmq7whOalrrfYz9G9i5foxZoYiJMpx1RUYq6wvpEagNhf0RoGnCG2IyhmWQXb1lgywJbeOqiwJclTTPNVkDnsgX1wJKrC63BFmMiNSOcN6ir0diSUpuz5nPAQK68YnMllZgS1uasclv2Ce0UV5bUvRWCKTA6AVNj+msgBh8bSCEnMeFwRZ8gDqzP1iMxpBhJzT6pHaMhSwLFtgVxYDwxKRLjXW6plwm34K2zsqHZVQtrq6vsbG8zODFg79o2IY6YjK5zvqp5631nmPTXqaqSXiEU0x3WteHG04+xs30Z5y2rqycY9PuYwjO47wwr957JxQasOfRWLG4XvyMs5sg6h9nM9ZZjmRw+7py1B5INmfxrLrEVUqeZqExCYNIG9kZjnt0eolg+5fVvpF7fIMWsuxRTQn1NWXjSaJcUI1evXEJToC7g/nvOM6pKdi5dZG845Or1HepezXT3xivXKK8gQppkEenpCJohpDHViXVk7QzJCqZtMFcvwyQSNDAa7eMRVoPyWRfOMx5OGEtiXyMnfdZpq3XKmXMbVN5T2oQWnrKMmF5FchbnbHazlJ6gCZNajChJTa5HGVvQlhSyRtWizqGNsTfPaA4NdzctiEinf6eIsRSFR7wlFJZ+v6JYH9DeuIE1maDsTlo2Hno9pmkZ7+7y5Ef+kHf/zE/Qqy2nPuNN7Ic9nnnkETbOrqNaI3UfY0oSCRHBGKEuHOdPneAPL16EZjFH4aIsacYTvLdEjTlJK7VYX3ROo0SvX7OzvUtphBP3nAMMk51t4mifEBJx2tKMJrT9KckIxiTaZoy3gnEOJVdOUIGi7uP6JSHlS1vVNa70GC+URYmVTvxVdeGffcZasAZXVEgag3EENZjOWtyGgDHS1apVnHUYm2smizFY57Mlzlpcbz3Le3RZ5xqzzFAM2SUrImAc1vWIKWX9xxBI0wlxOsaKYpwQEiSbRZDFSOdCXMyrcMvHbfffIDl5sfDUdY2JidQ0eOsY7u3hrz7Nfb0NhttDnBNG+zvUJYzHe0ivh7eRfl1xarWGskAFIoJoQoybJc4ufB+/Eyw8wTNH6rzeaml+zfysNqVEQEkpC6NMpg0xRHb3Jzy3vc/nv+WzeOD++7k6HGGdwba5jq13Qq907I4UZ8FKwgr4sqadBGrviXWPqirY3t7j7LmW2hUvazvcLbRtRCDH/KSQi26vnSC1jiIpTRpBVSPGknoFLpQ4DHspYpKhd6JH3euRrKUdjSg1Id6w0i+ofI01ivZqjF/BlisEyZIUqGR9tiR55qwGCoeKoFlqHmeKTOjvdiO9TBCbJzfzlrosVyBdTdrDV066yNmDxghilKpXIas9vIfJcJ3Y7HL1mW3qskexNqBWZdLr0VtdRZrEVCPnT59i/dPeAvt77F2/xkZ9DlPmuC9rc6F3RLCinOhXFAZGC8qwY4hoSoQUwPgc/yaCAQrrcJUwmTaUhcc5S5hMKFZWSVZY2ThFXVaM9/bo+xJpI1hhZ2+byvscK6kR7yxNA5NpQ9tG+oXFe0eMkeRsriUsWY7DYG4Wlj7i3VgsaE6qciXSjLFlgXMVGgOkhFcIMRDbCcZairI8iBcWKzjblZaELkGlIcWQXb0JUmiz5IyarC2poClhNRAnY8J0iAkjnAgUBSCYEAixRVMiaa7yMtNqWzTculfpgWSJEUHUUBQeyoITJ9cZj0YEAjbs4UeBIiYoHXH7GoPzpxkzosJhnWfQq3FCp4SQx7k0E1TXLu7uWMU33hoLT/BuckPclDA5/yaTABWYUwvLcSqSzf1taGgmU0ajMaNpw5seeoiTKxXEMWt1yaC/wvXxLpCI0yH77RhrhJXBgNMbp7Oi/LilmTRMhkNubN9g3I4JsWW4t0axuvaKtckrCUdB0pRdsa5AXB/RiBibH4DBUqys40KknQ4RjfjC46wljKcY6zCqOOOIK33KtiE5oSpqnOshtsRWHlOt5hvaVgRpUQ0k43OCQFY9Ra3B2JiDo4AonQt+QeG60lTS1Uo+cJHMdPBm5LYjfAd+lSQkFUzVo+j1kMqhKPvbV1g5dYowbqh6FWpLbgwv0/iKpoTXntjgjfc9SPSWcOY0091nIEb6/Zre+lq2cLfd3ZWE9X6d3ZHVgsafpoTGmDNWDYDFWaENLWnSUPqCEFtWV1fAWYY3bqCTFo9ByjJnMtucpRxjyu7alLClJSHoTObrgKRnySdEDscwhNTJfqh2SrEHYtcsML/L1RGKsiRSE9sWb6WrMwsqOQ5ajAFrGTeBwiSsM7lGrLHoZIS1U2KcYMi1fLFFdrFiUDMT0rW5Qk4zJjUj0mSEtYJ6jxEQk69L4T02lIhMmLYB28UhHxfo3F/EIIB3Hi2LzrKa2B96rCi+iPSi0sSGUys9SjG43oDV1T4UNUVRZ9d6d11yMEru4waylfoY69/NsPgE7+Zgu6PhR4ebZGbo01wIupMOU4SokbbNKe6KZb1fU8QJ451rlFVFvbHKax56mJ3tq6hOs3ApeUB2tuC+C/egbWLn6jUuP/E4YbhD24xwDs6dWsV7T7OgOngmaFcBJGFtj9QLMBpnywINxnsSNeoSzju8M4gVrIF65SQiNTrZRaTFr61jd3ZIYUIqaqRax3ghxH1SGGFdgXqDczVCAuky54JiTCdWHSaQckB0lqxeXBRleUjgZh18lmRhTLbidcTPeN9phEUkKcQmW3x8SZwGqkHF2dc+xM5gh5VpQnoOU5Ss1I4brafcOMn5k6cgKdOotETW7n0IrxNwEDViuFmXaqXfp1eVDJvFfMhlCw3MfvFssnnQBjlKP9eJlXwdVBUxJluTTL5mRgxBO8mH2T5yWOla6JJY0MNjKt04JIdxd3Qkb9HJHTl7PzVjXFmTYo0zjjAd43yRk7xcARG0i7WeCeKHqFgTsxxNihhpERuyO9ya3JQm11em8w6kOEU05DHfO8T2ELQL2QmIJEgRIbtzFTC2ZTqZgF3MR/DRcVWPrNVOgN1YS1mWBHIy0ql0kmkzBgslQtEKtIp3nqKo8L0+1pY4W4C1GLqKIAeSKHrTQRfUA37HWMzeNYejF3jeKzErY5b3yws6k5ToPmyki52IefLbq0uaEEgou3vboJHUNDx07gzbD72exx//Q9q2oSxLjDjuO38PRZjwxIfeT7O3jU8B66H0Jf265MTJkxhbEtKCFlxXJcVEDDmL1fmC5Fq0CYBgezU6EWxKME3gHULElnmgNG5AMgknDVINkOkIO2kx/QGmWgGJeOOI6hBbQVkgbcr6bAgp5bI1QS2G7BZRMVjxWLGdG5OFZHq28DAjccZgTCYLZmbNmxE/srU6xUgILYQGKxHTKhQF0+mE4bNXaFPE1iXTZkQcjYmTCdVgjfMrfVxVMNm7gQk94nCCX+tRDQa4soe0AZcgaUDUHByzqipWej2eHS5m/Oms/q5qQpglNejBRFNRjMnxj1bm9u3I2sG1yWV0ukmoOchCz5CDLGW6usMHm45qgMpBfsXND8IFRNOM8LZHDK6b3JiuGlELIsS2RazF+QIRyTWpSThXZE+DL7oJkQVXgLEHcZOS0k0xXjPZFfCZMEqCGNFmgpEuocPkWF8lYp3P31V4wmLObW7fu2ads9MfNUbAOYxUiBH6OsBOHarZNd6WObO/LCvEGaSscbbAGEvq7qdb1ZxF8nrprNzHFQtP8OzBSDhH5m6z70ytIbuxuv26ZAtm5vy2pS4LIoo1lr3hPsPRRymKggd6htUH72FnOKSuemys9KhsQodXCOMxlTNotKSkB0XCQwuS2izGu4CYPdS0G97gcJ4lIuih6bRDl91qDCqZmNBV/cDkwGlbFqTCo9ZgvSCi2OizNcSaTjs0B0zHtiWmBMbRTKd0+qMkTd25LWYMDEA16Gf34JE4w2w96q5ISp3rDtq2wfgCJ4nURJAWU61kWZm1NSaPPYY1nrLnGO3uMRicxK2uEdWg4yGr/R5aW9y5dep+D19XYDKpwxgkpM5tGUASZeFZ6fUxeu3uNNDLjAOraZojWMyNP53lKIV04DrXNLPgzT6Tx53UtjCz1M3JP+SA8pmLfd5qmO86mblyuwfqTTF4CwxvLYQRzhmidmO3yUkmYl3OgEW7zOKE6zLOzUzQPrZYX3YucHIG7mxs0tiNTza3r8kuWyQTiuw9aHPGfmpBY84cNRZxJVgllzRPpGZy19roZcXHIFUiczeDCGIdvsjjvvUlJuVYumgiiuKKAjCIK3ICjUkYBNXD50ea/0q9zSksuOX6KBae4L2QFMChi4PugTeLX5mtg5SUhGZdIyNUdYlpWgShqkqszRpIIQSwno2NdfbHE85euIcwHjLo99G2Zb9tMCnl2oPW50wt60kh5cyrxSxkAcy5o26DeQp4QMdlfuZ1+OEcNzvna3/+F93uDA7ikhBZVOm7m2C6vkxK2R3dmW9UZ8LHqVvOLrtgxznjUyymDVmzrgg4UeregEm9wniYtdbKVDOa7lPsG/or65jSEHWKDQWFs9R1gTdZd81IZzGyXXyTtkBETH6g+qP1ohcFXVJL0uxGOvQW0PXBjuCl9iAOMmmaq58951JPM4uHyRa/ubslZyabg0D/GebJ3KExTw9fC4zpaIItckKD8TXRuFmQFsZYjBViIpPppBjnO8KQUE055i5OUeOQGLMbsNNHNbbISTNkN6OaImfTknL5sckEQ0ScgDoQj7ZTYozZdes81jiS8dl4cFxwMHYfgcCMvFljsI4uOS7iulAEtZlMYxxqQE3sSGA2ABxGzs8Okwni/CGOUUsfYOEJ3k0k7iiJOLpzSjkgWRNJZ4rj+QFonEFSjh2oywIDlKWlLIoufb7KZWt8yfr6KkXtKfobTPZ2afZ2KUTxdZl1kDA5a0uy3p7MpuLHAQd34eEUS7vZ3K1bQI8s3nxFO1Pg4T7zXzJj6c+z0h2PaVwI4ZDAzQLwZ++ZTZ4zOTBWcHZWwirr25k2EZsG4x3eWnpnzpGuKGG8T+/MBfq27CRnFDWKLxy9lRpfOJwTspREi7Eml2QSi5BdiTPLrjOysIHmpnMDaDd57PjFITQTtpA6+RgxxKiIN3PPQuHwg/lahdAeTIq6A3VqAbNjdV8/f+vMLLazPsAiyxzT1SDPJCHH5+bxOSZFDGAqRCJCRI3F+IIUY7aykcf5mCKmEzRGDNolaGAtiqDiu3hIQWNAY0OcjvASDuKwMSbv58Ca0JGOREKwRZVrmx8bPP/JOzOqzNaKtTlZbjYJSt117MIT5onaLfvwgQt47v8id/SPgYUneDOz+tGuNR/CMiMJsxqNSieRkiIx5PqdposXEHIZM2vAG4u3YK3gnMVaj/MFWMd0PCLGFmkaenWFc47paIpKoihLZlEcVgRrdHEdhTp/p83/f96OHD7MZuvkIOFvlhk4I+q3CDTi5rtbjmyefZHhuNzxYdp0HCFbykzXv40xh6+5BAxBiCFQFBW2KomTEUybPABbQ7naZzxeo21bXG+AH2zgqwKjgaQNvihALKmZdtI0iaIoEY3Z9WjnrErddhWlmSyomwo4sLLRPcfmXK+qmovRa8ouJcmExMxKI87GpTmLszHmsHavdj16LiP6wEUrcnCsmYci/+smucfBnCEGYxJiFFIDYjFicwKLttnVh8OkSAptl3iUZbUkKdgKYx2x87yIdcySYpAs/6Ri0XaCTkcYEi4FMKCdRQpyWztfEFtyHJDGAyuqXdAa5LfHvPflZgvbzcu5z4sc9tWbuZoc+fzzbXQ699njSvQWnuCZOQI3j8MECzm88LNEC1ViAkIu0D6zOGTzscEKOGOyWCOCIVv3IJLSBFXFYuitrhKbSAyB1AScsbiqE2Ls0vi9Fbwvc1zSAkKftzC37abp2FET3s2DwOFD7iDK4nnHueP79zg83IBeV6JqPqniJhy4azMB0wjqTFaC8J7Qenwy+CRZtsO09NdXSaMhw6uXGajFVycxdUXhV3He5eBya9DYYIyD2GJMFvpV8kMtxkCMk3zNnEEX1Ioxe0DdNLVRmElz5mFJDi2qJrsIZ7peoAdxqnLTfXDUctHF4M0seDMyx+GT8XB55p6dvWARn3xt21J5h5K6smQB6QSHDQmIhKQYW3Zxcop11UFcqgBJ52WFTE6UUAESGluMxPw+NgfkXYxFnEdiyB5cATSgSTDWodhcs1Zz5l5iMfv+x4+OlQnIwQPiCCE82m27G+12bth508Bxw8ITPBHb8bZbXeDDjoF0LhWyBlibUq4dGLMyubXZdG8FnHV4k612zubKF0ZMt49QljVNG5iOJjjj0DYRU04Jdz7P2JIGBMV1BbCPCee4yW8kB7feLXL6jl6z5yVizBPzWxznwICX9z0yt7vNhxYL3vsDdxyaDhMqtBOIORKHNXOVp85iZ7yDpMQ2kFow3lIN+sT1NcZPPsYwXMR4MKvraC2olRyH5DzGV9jUksQRQ5ut4jHkh1pSkmZ9OO/sXEboYuH59SLm2lukC8+Ye151cXTzXX8mq5EtGpqzaLtssJml42bBaj2YkGYe1/X/NOsHHEyUFvouECGEiEUgSc6ONWR5mZTjS733Obba2TwZ0YiSl2NqERKxTRhfoJrFwFHNgscoGMVoFrdPxmDzAVAMiYSxPl/nOEbIZRFTApUcv2eMPSiZeHxwtNd1hO5jNMO8U3fmhcnPjjtov4Xu6C+MY0DwDq11MkfmZkbd+eWDmJmU3Ucp+2rntKeytc5KztD01uJcJmjGumzLM5nkeee656jgixxwnlKuZyso3jlICSsG1YBZ4CyLwwfN3PsjxOvAKjHnnp3b+WD5qBNW0Vu7m7pjPJ8D3o40Lh5SO+EgNfwoRA6GR51ZeFSybqDtNL9msh4KhAjG4L1j9cxpmsmE/d3r7D/3LGm8TzFYw59cxxiD90V+uKU2ZxsaS0zhIP5rdh9aaxj0e3OZ7guGmbd0NoXRmaFaOEjm6kIP8oYuxXu2/cBHJQf3hmRxzu77D92xBwRP04zTHUwcD0hkZzWcJdUsMiygIRAjGO9RGsQ6lJitcaqQEsYmkjGdqzuXOFMUa0yeehpIMZczSzF0oQ1CCi0Gl63jJhsSQDtLocGWPZLm2rZ0FXNUcv1bYyXXxg2RuOgX4paY98Ic/f238cwcIWnzfXqJ20OOS9r8EkssscQSSyyxxHHBwsb2L7HEEkssscQSSxxXLAneEkssscQSSyyxxIJhSfCWWGKJJZZYYoklFgxLgrfEEkssscQSSyyxYFgSvCWWWGKJJZZYYokFw5LgLfGSQ7bkx2RLvvPV8j1LLLHEEku8MpAteVC2RGVLXPf+12RLvvEOP3uwr2zJfy1b8ksvsO8Xypb8wUtz1ouJhdfBezGQLXkcOAs3yYv/mG7qN9+dM3rp0f3Gb9RN/ZW5dW/r1v2xu3VerzbIlvx14M8Dp4Ft4D26qX+u2/ZrwDt1U3/k7p3hS4/j0P9frVi2/d3Fsv2fD9mSrwG+FfgUYA/4APB23dR/80ocXzf1x4EfnzsfBV6nm/pIt/03gDe82O+VLfkx4GuAZm71o7qpb/6ETvhViCXBez6+fJ78LPHCkC2xuqkLVW9HtuQbgK8D/qRu6qOyJeeA/+Iun9YrhWX/v3tYtv3dxbL9O8iWfCvwvwDfBPwimQz9p8CfBl4Rgvcy4+/opn773T6JlxtLgvHirWEAACAASURBVPcxIFtSAJeAL9ZN/VC37gzwOPCAbupzsiVfBnwn8CDwH4Bv0k394N05408csiVvBP5P4C3AM8Bf00392W7bjwFj4AHgi4E/LVtyDfiHwOuAf8ERefIXap8XOtaR73gbR6yM8zO67rxGwEPAFwK/A3wleZD6BuAy8NW6qf/+Dprgs4Ff1E19FEA39RLwf3fHfHv3/Z8nW/J9wI+R9dQnuqn/49y5/Szwr3RTv1e25ALwA8AXAfvA9+qm/r07OI+7juPY/18tWLb93cVxbX/ZkjXgbwL/jW7qT81t+rnuhWyJAb6N7OVYB95N/u3XP8Z3fwfwWt3Ur+3ePwg8Bnjd1HBk37fRjfmyJb/erf6dbtz/78hj+jt1U+/t9v+rwF8CVoGLwF/UTX33i/ztvwD8S93UH5hb90FgUzf1p2VLPoU8ln8m8Bzwv+qm/r8v5hivJJYxeB8DuqkN8E+Ar51b/dXAu7sb/NOBHwX+e2AD+AfAz8qWlK/4yb4EkC3x5Jv4l4AzwLcAPy5bMm8K/xrg7cAK8D7gZ4B/DJwE/imZWM2+77btc4fHejH4KuDbgVPAFHgv8O+69+8CvucOv+ffAl8vW/I/y5Z8lmyJnW3QTf0bwG8A36ybOuhcOO8Avrob9JAtOQX8SeAnunU/Ryac9wBfCvwV2ZK3fpy/8RXFcev/ryYs2/7u4hi3/+cDFfDTL7DPtwBfQZ7kXwBuAD/4cp2QbuoXdYtv7sbdn5zf3j0zvhn4bN3UFeCtZCL+YvEO5q63bMmbyeP2L8iW9IFfBn6C/Lz6r4Afki351I/jOK8IlgTv+fgZ2ZLtudef5/ABPqt893VkQgPwF4B/oJv6W7qpUTf1HWRy8Xmv/KnfMW76jcAPzW37PGAAfJduaqOb+qvAz5MHthn+uW7qe3RTE9ny5oHv001tdVPfBfz23L4v1D53cqwXg5/WTX2/buqEPDhNdFP/UedC/kng0+/kS3RT30kewN4K/GvgSjc7vN3+7wN2yOQN8o3/a7qpl8nWwNO6qX+z+40fBX642+fViOPQ/1+tWLb93cWy/TM2gKtHLWpH8E3A39BNfVo3dQp8B/BnZ4kVdwERKIFPlS3xuqmPzzwwt8H/dORav6Nb/7PA62VLXte9/zrgJzuy/2XA47qp/49uaui8Qf8M+C9fpt/0CWPpon0+vuJWcRiyJSPgS2RLngVeS+4IkF2V3yBb8i1zuxfkWc2rFTf9xpkpvHt7AXiqI28zPEGexczw1NzyBeAZ3bypqPETc8sv1D7pDo71YnB5bnl8i/eDO/2iWYBvZ2X8im75A7qpv3ibj8xmfr/c/f/+bv0DwIWOSM9gyVbAVyOOQ/9/tWLZ9ncXy/bPuAacki1xL0DyHgB+WrZkfuyO5ESVVxxdmM5fIRPNN8mW/CLwrbqpF2/zkf/jVjF4uqkT2ZKfBL5WtmSLbGz4s93mB4DPPTKWOw4J/6sOS4J355g9wC8B7+qsRJDJztt1U99+187spcVF4D7ZEjNHvO4HPjK3zzyZexa4R7ZE5kje/cBs9nTb9pEt+cI7ONYMQ6A399lzL/J3fVzQTW2Bf9pZ8P4jcsCx3mLXdwK/25n030h2W0P+/Y/ppr7uFp/5ZMJx6f+vRizb/u7iuLX/e8mWyK8gh7bcCk8B/61u6nuObuji6m6Hm8Zx4CUbx3VTf4IcFrNKdpf/bbIF7sXiHWTS9m+AkW7qe7v1TwH/Wjf1P34pzveVwJLg3TneSY6j2uPmTvPD5JnMr5Dj0XrAlwC/rpu690qf5EuA3yInK3ybbMl3A18AfDnZ1XgrvBcIwF+SLfmhbt/PAf5Vt/227fMij/U75JnZW4APk2dqHze6pAx0U992i21vIwfQ/jp5QHor8KbufCFbBh+e/4xu6tOyJb9NHhj+mW7quNv0PmCvI4h/j5yN9kag1k2dd2W/2nFc+v+rEcu2v7s4Vu2vm7ojW/K/AT8oWxLIMdItOa74j+umfhvwfwFvly35Bt3UJ2RLTgN/VDf1n3+Mr/8A8FdlS+4nh7X8tRdxarNx95GjG7oYvHuA9wATssfGHt3vTqCb+t7OMvnd3Gyd+3ngu2RLvo4cmwk5RGlfN/X3P55jvdxYxuA9Hz8nW7I/9/ppAN3Up8gB+8qce0039f8jZxL9fXKg6SPA217xs36J0MUafDnwnwFXyfF5X6+b+uEX2P/PkH/zdeDPAT81t/227fNijqWb+hFyZtevAH/IJ56qfx95MLgVdoG/DjxJ1sD7O8D/oIf6T99Pjje5IVsynw37DuCPMDcodPF/X0YeCB4j/84fAdY+wfN/uXCs+/9dxrLt7y6W7d9BN/W7yRp4306e7D5FTmKYeSa+n+yq/iXZkj1yYtrn3sH3/jI5HvqDwPvJpOlO8R3AO7qYua86sq0Evos8vl4iJ0G8EHn8tiPX+uqR7f+IPJa/c+7c94D/hBw/fbE7zt/ujv2qhKjeytu0xK0gW/KjwMVb+e6X+OSBZPmD3wE+rXPBvlTf+0XkAeGBIzGJC4Fl/797WLb93cWy/Y8XZEu+HvgL+kku/r900d4huriCP8MdZmIu8epFZzl840v5nV0yxl8GfmRByd2DLPv/XcGy7e8ulu1/vCBb0gP+IjerS3xSYumivQPIlvwt4HeBv6ub+tjdPp8lXl2QLNa8DZwHvu8un85LjmX/v3tYtv3dxbL9jxck65M+R473+4m7fDqfMJYu2iWWWGKJJZZYYokFw9KCt8QSSyyxxBJLLLFgWPgYvJ/8oe9RYwxRQJKi2r2sgAiCIDGhqjhjiSmiCiKCqpI0oqlBYgJJxBBQDDFEVBMhNrRtS0qJlBKhbZlOpuzv7nL16lUe+ehlLu5s89ykoW0TMcbbnutoNJLbbvwkxQ9851/WN7zuNQzKnGjUhkDbtnjv8MZgRFDrSSq0bXMgMJdSIIT2pvaKQUk6dw1ViZrXkRLaXYMYIm1oaZopTdPQTBumzZTpZMp4OiXepKt8iO/50Z9frPZ/xzfm5lSFFCAESC2IQtvkCrplAc0E2hZiALVQlvl9XYNz4NdQC+Mnn+J9NwbI+U/nj/ypL+UPf/NdfPBX38+FEysMH3mMMB6xtz9l2F8htg3t2gZv/Iqv4vR9n8Jr3/TpjEbP8b7ffD9rrmIyfIx7X/MW2tDyqz/zU3z7D//jxWp7gKsfVNBbqybq3IIqGAMym28LpIC2LSSFWREFX+R9QoTpEKzN17IoEFfl6ynd543pPqf5vSqIPTyGRtCUNyeBk29YqPb/3r/7v6uIQbt7XVOA0HTvBQE0zdogQkqk2IImNEZiMyWFhhgiKeaxKKmSYrevKqlr2hQTIUTaNjCdTtgfjtjdG3NjOGVvNGVvPGHaNPnEROYuU15+dne8UG0P8JVf9CY1RjBiMNYgIhhjMMZgrc3/Z++NRbr/xhhE8v6z/8ZI/j9bbwTIfVnm+nnqngXWWay1eb0qKd16vJ/hb/39BRx7Oiw8wRMxeYhL3c0pAibfYtbkwU7UdANgJnaH0Pw6evn1yPajn1BFk5JiIqWu491ylF98yG1vnYW9p1492L8GaO7vxmRC4B2kCI3CdArjEZiU94sOrEKvB+0YSg/WQ7uNtj0e/91H2b/3i1g7e5Lfefe7eOr3Poy/sc3lD3+E5/YadpJgvaFud5g0idF+w3v/4Q+yt77OG7/4S3n9W97MRz/0W4ynI976n38lV4dD/u0v/BNG8ZNd+P926O75W3X12TrtyFeMYBOIy9cJBQK0LXE0ZrS3Q4oBTcp0PKIZDXHe4wtHvbZGffI0pt8H7xFjD288kTnS1x0rD3SHJ3H7m/STGEJKkdg2GGMY7e/hDEzGY6wIRelx1qExYIygKeVnhWYCYVye9EtSUJPbNLaIEVQNhkRCs4HASCYUKCL1oeFAMlmJqplQoh1xyecnHcFbWHRd7YU25/+H7TP7yMf46E2ffmEsdAt/TPz/5L3Xs1xZdub32+aY9JnXAbhw5VCFMl1tqtmGrumH5MyQQ00opJhQhELPE/pr9KagXjQPVIxiZDgxGgZHwwhSJJtkd7UrX9VAFVAArjfp85yzjR72PnkvUNVNDsmq6k6siIvrcTN37rPOt9f6vm+tPMDz3iEIFzCAVBKvAqATsYJEzKe2MkilQpXIhdOtECIM1Ar/2dnuowZz9T/Lvwh4vHdY57AunPQeW6rjMqGd+9Kn9bcf1zWvoz8IFZv6BfA+VCyqBWgFPgmVICdBJ6B0ABxFBZ0emAKUxo+OcI1t7owytrausH94hzu3HmDu71DuHvHmqeWDUjBzjjSRPLPZ5LntJvOTOXcOpywOPuRw9p/YPTnkC89/Ba8TXnvtVQ4PT2l0+9x5/cMf/zx+WkOKRw6DfHRPnsNZ4fsuVFKFQKQZPm2gOj1aWcriZJ/RyQkHB4ecHA0Znp6is4REKy5tXeDqleu0L24hB31EmgWgGA+xZ39DnAN04uxxrlh47yjnc2aTMaYs2bl/HyUFplgwXxRkjQatPCNLNd1OG7wlzTKSNA0VI6mRysVOgcPFipL3LsKQevV8uIdICFUlyMnA+fDzQmCdDZ2KCFsC5o4AbyXBdR0fn4DP8/4/7qOHf1+c+8zHz8RHfsQ/+mt1fOzyih/zvdWKlQd4EPeAEMuPAZRSsQ0bvuohnM6EwDmPwIOwYbqeO9sxIiZIIcQjQO+sG+ZjqThU8EKSwD+8sR+XUFIRbirE9TqfHH/UBf4PXaeAqB+/1X4kTBmqdfWbd+Bc2NNlEb6vNVgHVkImQGShNYuANAVXwqJiUZygtp4kb8H4nQ/o7O4wH074i0rx2tQysRbnwS0K7i8qhsCXr/YZpZqdgymznT2+/2d/xmR4QIM+o8kBm60elc4R1Tuf8UJ9QqHOGemfv/YfzQOC2D6v27FnlT8hBEiNbHdJqpKWSBgXFR0PVkGxmLN3fMLuwSkPdg954omrXLh2hc6lK4hWO/6+DO1ZBODC5z6eaoUPoH7FYnZ6zMHePuWiYDyeMDo9xVmDloKymDM7FZzgmc8XeCHJs4ROq0F/0GVjY4O82QxnI6mR2iNtAHrehyqcRyJwy/uHiK+VlIo0lQgBSiukUhjrABmpP1CDPFjR4ulDEasbkfK0/OrfId0H7ObrG/jZj4aqCkI8fE0JIdAqtHuX4Hl5n65/9tyCr/zaPyYAr95NQoYT1fmTU6jwgV+CuMjvci7w8YxDOBt/p26dgJQK523kDJz9f1opjNLoyDNACtxjsJF+VPT6PfI8R1KfwM4B5Hgqdg9lu/qGU7c4AlALnEeW3Dvn3Bkfw9ag8fxrW7++deH/MYzRyRmLoG7NwVnLVqcgXPieqSAhtnJF4OpVBb6YYA/3uG87DJ6/yWLvNsWdd2gh+Y5t8M6koMobaGdxVYn3ntI7vnPvhE4nYXujzUQknBxNWOwM+evdv2Bw7QovPneT/f0xa31H8dBoylWKcwDvI3K2c3tSuMCJNDYArrryVt/YlIL2Gkm3TwI0nnyGkw/vsD4eMzzcp9Ub8fpb77HzwfvcPdzjmd09nnn6lAvPPIPeWEckOrR+hQgPpP5/hQygX/54jtJPY7z2ne8yHY9QUmHKElcVlMaSNxporQOWlhKdCYyzJHmLRVlycjxieDpC6QSpFVtbG3Q6LWSS4rxDOBfuI86BPwNqLnL9QuoRaKURafjmhu8i5ITSGGTsaNTt2dWt4IVsf/4AL4R86PsxoeNFzal2eC/x4gx7ee9xCGRdJBFnbe4zNqNEqvqewhklIfwH1AWc+uOHHsIn9fR/QmLlAV59AVnrwiFWyiVAOAP1IogpAG8d1jqcMVTGhqTrPUmiUNLjI3gQ5/h3SqklkVMqRZqlNBoNOt0O3e6Iw9mMTHnmtjoHLFd9a4VoNpsB6C5BtohJLlQmfDwNewE+AmjvHQ4BUiFiEnVe4L1ZvmYytp6kEDjhl1wjgViCbmoxhvPLi/l8tXXlw8b2rDiXMQEwUCwCME4kFCWoNJDtmwkgwUZ+3mTC7GTB7v1bNNa3+d5ff5OOgdmFJ9iZeH7+t36Ov/z+q0z3HqCsIJGK0nnGZcGr7x9z6ctNmh1FQZOqrCiPj3j7rfc5PBzy5KVrTA/vYsziRzyBn/Kw7pF1f7Q1FHlxQoTm03yBFwqhksili6+fMQilQSSgJEm/xVZvI/xOVTDZ3WWweYk33nqDw90HPNjbJ5GKJEnYTBPo9RBJBPhKsmzTLgHf6gG8Ox/cIYtrm6Yp7WaDTt5EeEMjS1FJQqPRIElTZtMJSkmsA2MqJEHkJbxn/8Eu40bC1uYGQimETBDeoJTA1vkFEEi8cEhqIUDIUUIGIO29Z1EUjxQXVhlguOUBHeE/IpyAcyy6mKNroHdW3qvb2g7notBi2Rj3odMmCTxJAFkD7hoshrxXV1kfjo80e1cyVh7ghU1Qn5rk8mUOrdNQuQstVQfeUS1KShsu9MV8DtahNeR5Tp4nZ8RYHypNUkgsoZJnrV2qhdI8p9vpsNnrcDSaM68qyggugccG6KVpGp+rJ+AwiawVTpEMbp3A2sBXlFJRuXhxIkP1Uwi8CNlQENdMPHqzXJ7T4sHQna2tYAnqV329Hwq5hLQsySpxPZFAaUOrFkLlqCIsnvShqmQdTOacjg2d7Yuc7O0xswkXt7fYaV7mxZsvcvPrX+bW3j3GkzHT2YKWlkityFLJ6XTB7umE9bUBU+tJOg2SUQM1nzM6PmXPW9rK89SNZz+jBfqE4yFuHSzJvP7cDwi5rNpVzjCdTVAiJcsykjxHJBnIClH4kK11rMYlWahg6Iz2kx1evLjNtZde5u2/+Sbvvv597u3uotMUmSSs37gB3Ryh4+svBfio2g3ti09vTT6l0CIc6pM0Ic80WaJoZZLeYItms0mSZVTWkSSaVitHSElVluAd3gWul7UGawymmLOYzUgbzdB10Cl4i/QBZOBj8cDFmpKUyAjqBFBJw3hRcm9/SFkZKmsx1lEZG4oIKxjiXKqpQdfye+cEJnU6PqviBXpTOL77KIT11PU6v0z7/mzbxjOKwCN8yG3nz7VCynN/+6FH+Y//xH/CYuUBXr2LRGyXCiGw1kZbE493gYtXVY6iKJhM54xGpyyKAPB6rRZaC5LpjMFgQKORI3DLFqNzDillaBkCzofTSpokNFtNur02g+GE0aKgsOFxnLf5qN9WNZSULGueIrQyZKzieSEw1jGdV1SViUo0g7GWNNVYE8nmdTU/tm0DqTZyYep27TmeYyBCx4ggPgD4R0r0qx5SnmXE+mRsHdgKSkPgYwkwbgm2qSqoFBhgPsctDMdDS+f6Oq5YkKc5g+vXOR4O+IWf/VkObUlrWjCeTMnxOGPYbGfo0jNaCO4dFTxxWTOpUqZ5iUgVjYXHWE+5KGG9x3A0/SxX6ZMLeX6vndEGztpHBE6kAJTCS8l4NGZhPFoltDttWq0OaZah0gzh1ZJrBARhTKxai0aL7lPP8IV+nyRNuHvrXT7c2WM0W/C88Wy/9DlEv3tWtTt/0xUf6R//1Md0NieVglx5ummTdjunt7ZGo9slb7ZRSUIW7ZWSvIE1Bp1mCClwNoA2ZypcWWLTJB6RYv0orqGQGimDO4P3HqkC0hAIhJJYU3FwOuY7737I+zuHFJV5bLjBUqhYWJHn3mKFrW5RL1utPEQl8T50ysQZqQdg2bp9iIfn6+tKgA99NfHQAq+6kOXHx+oDPOJNX7BUxjrnMMZgKosxnrIMoOLw6Iij4YSqqphPJ0ilODkeMp4MWR8M2JrMuXzpAnkjXfKn681T8ynCiQHQCY2WpDsYsD5ZMJwuqKpwcnMEZe3yfe3ltoIhosRpqUDzRJDtKUzBbF4ynpYsioIk0Wgdjl+lKVFSBJHGsqUe1sjWgJ36SxLvTfC48vVfFTjvsM5Qm9Q4Wbd962PjZ7Agn2bMZgFALNeuzq6RdK9lAHyJD0AvzUFnkOQwG+FNSVlWnE48/UaD4ze/y8V2j7Ubn2P+7X0uXbrA6d27VNM53locgtwZzHhEB09mDUfjBVmWc/WJpxlN30SmJ2ilSJXEWI+SOYVTf9sz+SmNGjid32jnDhl19cyGA4rOWyyqiju7Bwih0UoyGKyxPlgjSzOarTaNZhNsI/hyZj6UqqQGFEIo0vVtXvrlXydRkr/6q28ynM+pnKfRH7DWfhaRppyx/c/xMlctnCHRmkG3wealC7T6azS7faTWSJWw9AFccqeDu4JzDqRFuEAnQQqckkgfnBi8lKEd6zxeBGaxUOCjX6cQEuE9p6MJ33v3Lm/f3WU8K5aHzuXdYkWXvY7zwA4hzwG6c+3pCPTCDVMsOXvnGQzh9iGWS1aDOx9/w3mP9O4hfp9fVsvPgb/l737iT/0nKlYe4IULkWg35aI/HRSFYTadMRyNub9/RGEFo9PTQBIvS7wp6XbaeGPJdMpsNufo6BhTFvR6HdbWuySJWqZwIVQomLhwoXsZfPYG3S7TtTnD0ZR5UVKUVfTGC1WopXDgs1uiTzxqUnH9JCsnOD4dM54tGI9nWA9lVQUxhhSYqiLPNGmakGc5iQ5V1qBmDm0RVfPu4v9/1nq3OO+Wr3UQjob3QsoAqB38KE7GSkUJoX9RA6jI+ZIKMoINSiSLh5/T0MgjAPCwsIxOK0SSozD4To/1/oDBM19Av/Gn9NZa7P/J21DMOK0cB3PD59qCvoYMy10Bw3mBxbOYHJMo0FLirWPhDCme6ckhfnr6mS3RJxofVxn7yF0m7mtjAt1DCHaPTrBGo5RgODWcjgrarSab/YJua06j1SBttYEOghykCUAPECIh6a9z7cWXuH33Lu+99x7z4h7+L77Jz3dadK9dRahz+2EF27MAqYR+K2NtfZ1Gs0XebEW+lsVWAXDVXETvwdmgYvaR/O9sFQoCMc/4eE24ugoexXgigsRw4BSYsuKNW3f57jt3OB5Nl9xsIQR5EkC7lpI0UWSJJtGrebgR0aWCmg8dBY6RoAQQAV3dro3dGcRZZwu5rNCdfe/c36gL4fF1kx9biA6w8OPrJ6t81w2x8gCPiPK9FVjnozO5pywM+0ennEwWPDgcMpnOUc4gRODbbVy8QKfZpNVqsbu7QydLABjN5hhjGY5GXL1ykUZaCwNCJa5WqMmo8EwbGWtrfS7NSwrrWCwWGBsmWtjIxXAueOatYtScxCUFTEgOjoa8f2+f8WSGc6CTkOSKyjEejykWC9qtBu1Wk36/R6sZgJ8kqJu9t6Soj5TeQ2IIFVprXXS3CcakzgsECiE14jG4sIFQiYuJMdijuGDHsSigdKBdABfzKoguksjJkz4IMKxgNC7JGx2O792hXJRcu7hB8+JVMgH/+U/+PQfvvcFaIhBK4ZDkWFrCkaqg3zDGMz0dUrkJ1XhMWhmaCkZJis8ywLMoq896pT6hUPyIO0sMf8aHlA4qQ7fdYTZfcHBcILQiO5ki5S6tPOOFG0/RaeZ0mw3W+2s05iVJrw+NRrgRJhq8QUjoXH+K5z//Je7f36Waz9l9cJ/b332VF5Qi29wMQB7ivXb1ck8zVaxtrDHY3CRp5HhnsIXBhESNcxa8iJNxaiuruhEbDoo+5hvnPN4GEZ47r9KvRQPxt4bjCd/83ju8c3eHylhEbKc385Tnrl/ixesXyTMdpzeI5SSHVQyp9XJtHqrcLQk7IQs7D9hQ4JDSRY72snQX27JnlbhHj+FLepNzOAKfdVkVFITryi9rf+d+8ZN65j9ZsfIAz7lgTGnwOCcpiorJdMa9vSPu7R5wMp1STCbYsqDRaNDvdnji2hWuXrtEK8tJEsmT1y4zGZ+yWMwp5gWT4ZjR6YQP7T2euHaZNFWI2m4i7kAZzUOVVrRaTTY3BkxmcxaLDOtsFBW45fvglbR6IQheeMGGJggp9o9OGI7nzBcF1njAkqQJ49EBo/GE2WxGnirW19c4Ojllrd+h1WrSyDOUUmjhqXDIONoGiAKMcPqrwXPdvhUELpr3Mip2H5Or+3gUQJ07f+yVAVC4ChYGZAY2gGa0QtjwOlFZvITJxOHEnMmHd+knLVoXL5D11+gmitf/8N8zPRmR+IonOimZq3iyZWlgSQUkQqGlIvUCjCVblGx6g13vcKQbFL6isJasMp/VCn2ycU4N+PEhImEoUAvMbI6WEik1e0d7GAFKaUxVoYRnupizsdZna23A9rxg0B7THI1otDs0NtahkSOECn67acrTX/0a4/GEv/nzP2OymPL9H3yXPM945gtfQqdbseokV/Jmt77ep9PrhvW1BusdpioxVRUmGsVCnFBndA7hg8jY1eb4nHGkq7JcVmSlCACGugWJ4MHBEX/+3bfZO5lgnUECeaZ59upFXnnhKdZ7nZiTVjPPPxpSqKWTQQ3WamAXQF8sjlIXAc7GlElZV+viz8fPzyquYcN676OBdKBfaSlx1uFFLRaPFT8hHoKG/mM+WtVYeYBXVRXGWAye6bzgw/s7fHBvh939YyazBWkCjTRjvd+h127y9JPXuXh5m2a7FU5YHvJM0GylFMWC2XhCp91kb8czGY14sHPA9uUttJQIGRVV5xSeEkmSJnQ6LTbW+hRFibEWYyPIi2DErjDAC62ncOGNpgvG4zlVaSiLgqIoKIwN62Iso9GI4XBEWc6R798Jran1PmtrAwbdUNFrNxtkaUKWZaRJupTfe+GJTJhlRdUB1lusqzDOY4wJLdoV7Mh+JFzMoIrQivNxJBlEiorCl4swjswYWMwg7weOi3EYoyhKQZ479oYl69cv0Ny8SNpqc/nKVU7vvIPzpyTSsNGFV1oJubfYUuCtoRKCVjOjqwWLWcXlsmTtYo564gLfuT3izuGYmfUPqdxWKpbt2B91I4nEovj8nXMc7e3T0IrRdMRwvsB7gbMOgeNkNKTXabEx6HHj+lWe3L5EM81ot9psm4r2xga+2UB4iVASlSS8HhuR8wAAIABJREFU/LNfY3x8xGs/+AG3799j9md/hvNw8+tfQ7U7oV0vV69N2Ot1yZoNjKk4OjpiOl+wtb5GUVSUiwKlNccnp2RZSqIkMtHUVtBKq3BAj7w5F3M0CJwPL5cU4cBYecnd/RPe/GCHg9MxeE+vmfPkpQ1evnGV7QsbSxut2uftsYia5hyvgeUwsljVtFH8JaRDK41EI5FxfGio/MnIFw7ai3qmbJwQEkFeEi24vHeUVYlcTu5x53h//iPdnsclVh7glWXJojIMR2Nef/uH3H2wz2w2Yz6f00gz8rRJt5mzvbXGjaeuMVjbRGc5WoYTiIobREqNkk0SnWB7lqyRc/Bgl+HJKccnQ7Y21oPnj49KKhXPDFKgvCfLEga9NlVlli1ZFwGesQ6zqi1aHm7TjifxhCsdrVZOv98J49ycYzKeYMsFWnQwtolAUFYlJ8enHOwf0mmm9Hs9Bms9NtYHrK2t0Wm1ydI0upcDwrLkeHiPw2GcZVEVFKVlVphgbUDN4VjhC18WNQEyUlFEAHbWwdzgEwXO4uYV5E2kjTcfCVQFi9M5utGiNHOKSpB2uySdHlKnPP38C6TDB7xfjZEZnBzOaSeSnlIUruR+KZmjubrZZztV2MWcbsOz+YsvM7v+HHvmVfaHYwpboHTyGS7SJxgfAa7+kQ/92ZekJMkbCASLsqQwFccnx5jKUVUV1jmUlLRaTXrdNofHIx7sHNFuJlzYXKOwFVeqgt7GBrLRQKTBM08nCV/80ivcuX2H49Mhh0dT3nj9dS5eucDajRsIna8kDy/LU1xlODydMJ2cMh7PON4/YWE9jSShsjAvS5SQOGtIMk1DK4SEXqeN1sECK3CCyzCvlkC9KYyhWBSUleXB1PPh8QTn4WI749rFNZ66eoH1QY8k0bEVvCxVxeqhP8P+KxqVc6HquYwzg2dEmDgRwHMAYl7Ml1U2GZ0XagaDVjoK8BK01g+5TlRVidJ6WS2UUgXWg5Qopc9A3sfpnVb5BYix8gDv+HTI/uExu/uHvHfrfabzAukdrTyj1WzQbqY8dfUS1y5vszZYQ2qFxAczywhK6iqzEIIkTVBW0el2yYQiVYrDo2Oa7SbN/CxZCsHyNKGVxmpLs5WHMr3zUWgRvdmcX5JxVy2kVEv5k3UWayx5npI1c9IkRWtNFSt5670WV7e3mE6mDIdjJpMJRVkyGg/xPgNnGU+mWGvQKvIcvUd2e2RpCqjQjuWMmxEU05airJjMSk4ns2W19CFot4onvMI+oqI94xl5mYQKhQyCHzsvyaoKZANEAbrJojxlvH9AaQWtrE/S7dekGG688Dzp4W1+5mtfZvf7r/IX/+7fkJYFm7mFNOH2ApzwvHx5k2fdhK6fsf5Pv0z7d/8VR2/e4o3pCffbbYw1EZSvYCw5uXCOUHQuosjBA1Ij0oyLFy6y8cEOF/rrHOwfsTAl1loWZRFyT1WhFxW37u+ye3BMJ1NsX9hgNq84Ph1x9fKI7QsXyTttSDVCabqb63zxpecYjU8ZTibc3zvke69+h5/t98kHmwipoP1ZLNAnF94ZysWc46NDhPPMjeLq9hX2R2N0o8V8Pse5OdY7vEzwQjGZzsFbjJe0mjlaWSSgokebkIJFYZjMC+7tHdNvpUxHc+xoQkd7eq2MBo7ZaIy0FXmekaY6th/VsmV43qdzVS2yTk+GGGuQj/AMA0faLu9/PuYj5wOVR0iBEgqlZKDjKEWeppRlKLgkSYKUEh2Fdy6aiWutkUKitEIIubx3139bqY9WqVdz5R+OlQd4B0fH7Ozt8WDvCLwjT8NpoJkl9DpNti9u8cTVbZrNJjiHFB5shRMS78JGEVJEDtmZj49KNEm7yWBrk7kpmIwmNNI05O/IKajLwlImJFqQpf5M4VmTe12c3LCiF/r50rj3njzP2dpKcHgSqdBJgq2KID6pQnWt32mwsd5mPB5xcjqk0SB6FwpMZWg2MtIkoYzt96qqSHQSLVnOQHnNx6uqikVhGM/mjMZTzMeYi64gvINywTl1CyBjO66eXmDxlcc4iZPgywJsgdAanMTagmamGB+WtCiZ7B/hq7BP+2t9ymdfRn/wfW4+dRW+8kX233+dvq64PZHsKMMLVxJ+r1fx0nxM8ltfRvw3/yNsf5XWvT8gPx6ifYuW98z9igK8muj9seCuRn41TymMjuv01riytc1Wd4e1TpfChEH1QkmMsxhrqKqSJEnorK2x1u2iGxnfe/sW775/lxdvPssLz0x5Ynub7togzBPG88RT17m/d58ffrDLdDziw/u7/PB7P+DmFz6P7vRWbv8XRcFiPmE4HCOTFiMjee39Dzk+PcU5H2gbxqCUpjRVVHJ6lFTcOwiK124zZb3TpNttkacJrqg4Hk54/8EB88ozGk3wztFPQ0t3UTmOT4YURcG83aKRp+RZgtaSJEkeKtiu+nzyoipRUlEZizXFUj3rvUcrRZbnCMEyF9d2nVpp0iRB1UMBfKjy1UUQU1VhkECWRc4eccCAxMkgDPM4nJIIpYK6VsgVlBH93WLlAd7pcMhivqCqKrI0IU1T8iyn3czYHHS5sn2JNM3QWgMCZ2wUBgiQDmc9woURKU6EaQz1XDypJI1WiwsXthgdDynmCxqNRuA2PeSPF04SSZI+RBA9e1vhCQvLWnu4eHu9DnlNdMaCcFgvsdqjhKBaWErnkMKS54qtjR7tTFMsZjihmM0LUp2Qak2apFhrKMuCJEnQSuOFxyGxXuAi566ogp/bfDanLKuP5Tt+VJ+1AqHycxSweHc5pxbEOJyDsrCQK4xxJN5C2oJGG6kUxdxSTCxUBenpGFfGGZxSsPXc85iOY/+P/x1ba32uX/kGh/fusffOkGb3lP/qhQGvPJGiX/4VxDf+O2g9Db4gTWdsdHus7R4wCdq3FY9HwV38GpwJMaLVhsobbG6u0Ws32NxcZ7yYMZ85irIA7xlNhgxPTpBKcHoyotftcOOp6wz6A/Z29yh+8Caz2YTFdMaz16/T6bbRmabd6/HE5W3eu/0hhXc82D8leeMtGo2cJ198kVVj4ZWVZe/ghMnCobTk7r0PsZVBiTDb1Ivgp2k9JFoFpawHF8a5oKTkZKrZPRmTakUrTzFVxWhWsDno8tLNyxzcv8eiqIJiVIWxXEoEgV1ZVeESsw4lIVHl2QhFWL7uK5r1McYyrwrqaR4s+XASg0OUBiEEi6JASYlKNEoFL8fAC5XI+poREiEliQw5Pcjy6nuKQkmJ9Z5U6jilJfCvlyIZsbrr/LfFygM8rYMLeaoVSgj63TaNLKXVarDea9PMNFK4qKISwf+SmvssY+JVoATCnc0epG6pKkHeaOO6HlMuorJHRYHc8rz2EaNHePj0tqoAT8qz5ytESKYi8lpMZRCAsQt8saCcTVksFsyKirIoMFWFdRakI9MOnTRo52nwwUsUSgk8Fu8NzlsQcRqDCH531ljKoggcpmibYiqzstXSj4Q5PxqrLmuea9dqTVUsqEqJNR7vJ3SNQLQGYFOceZXRSUVRQios6aRCnoxi9U8htEZfeZ7u57/M7T/831Cyz8QoLm02eeXmBb7+jc+hv/BLiK0vxdYv4Aw62WCQp7TwdKRnuqqq5qXxNhHPfcwh4vwNXwhQgn6vzVqvjVnMSaVmhg8iLi/QnTZlVbBYlMymc4bDUz68+wH9wYDLFy7x3DPPsn844g17C1cZnrh6ibVBFyWg2+0y6LbYOTzkdFaSJ5Ifvvc+SiQ8+dQvfkqL8unE8cmE49EclbV4sLuHtIZUeoROUEnK3FQI59lqtxF4ZmVJZS2zssQ5T1lVlMYwLwVaSobTBZmSPLG9yQvPPUW/36XXzJhNZ6g0Qahaze9jASBM7PHeRY89u/TZW94Hop3WKoaznjzLEEKideiaKZ0gVRpar1ov+XTBisYtbWmEECgtl8WRINQTSy73GVAWaB2kMeE+I6IKN3Talu1wfhQF6kf5461OrDzAK8sCJQWXL25SLeZ0Ww2ajYxms0mv00ZJEVVRYOJoLBXbq96Hkr3HYuLGW54K4lxTSdh8Wd4IZWU8QomlakiIMwWRlHEWrq9J/mdigFVV+dQXo6QeWxNOuSgBFhaTCZPxkOP9Y46OhgwnCywSmWQYa5mMhzgzpdtqsNYXtFrR+Fh6pPIkiYqDvWujSxFHqnoqU1EUJWVZUlUlZRHmrn5krWMSWLmoinOfxNmjMgnkfwGUM6RQWFMyL0q8F7QXJaLTgX6PKsk4GVVIo8hkRW++QO8fQVlCIyEk1JTWS7/MMyqheP8NLl+6zlOXLrP28teQ69cRIokgxoMfQ/k2/jv/gW6e0heeYeRRrnz8KFb90koCiHkg0YpeK2ej1+PB7mFs70le+fwXee7LX+DBgz3+5I//iNPdHWxR4DzsLHY4PR2C93zjK19BJhnfefstZsWcJ65eZtBpovMm2xc2eXB4zKGfcTJacPvOLo1Ggyc/zbX4FKI0FpGkjIqS2WyOxmOl5PozN7hx83nefP0H3NvZp3fxCoN2k6KYM5uMODk6ZFGWVNbhhUQKQbeZ0283uXZpi4tba3R7XZRWqItb2CpU/MIYTIX3YSqJs+HrzlqsqaiKEmdM9HuLrXn4WMy/CtHr9cjSjDRNkSrM5gXwXiKVREm1BHr1+FAhBFLwyFjJM3Amz407W6pzfX0JBY5kEN7WwPmMd/zwof4xyDcxVh7gNbRCdVrkaYZJFZ1Og2ajQZ7nJGmCFxKPpqosQiqUTtFKIQjDpoULO8grGUEfwVTR13y6sLmkViQyD457IrZwRVD0hPm0wbetbpXVgCJs0NXdcDV4DYlNoqVgXiwoy4JyPuF074CdnT0eHI+5/tznuNRdZ1YYFkVJWVW0JiOOdu4xLWZM7+1zaXuNTqdBoiSJysjShEYjJ8tStFZUpVm2vo0xsYVbMl8sKMoy+g2uaFZ9NNJuVHLWzzfapNSqyTRBzBZhv5sSWxpcpVBb12CQ41ONKmHNO/oioe8U2WIO8yHkzaVCVzY6tL74mzQ//6sMvEAozZlsrQ4BpDA6Rux+APt3GCSa4cIiV5WD97Gt2XOxbJfXPx6qz1ma8dL1a4yGY0pT8cade4xGh6xtDvidf/F7lKUgLSf85//z33JqHNPSUThBYafcun2byWjEP/+tX6PTHfCd195hOJnz5PXLXFrvcfXaNb7/7rtILZjOSu6dnNL8cJdvfOJr8enGYlEhdMp8PEI4SyEVWkh++P4dZsZzenBAkiUkrTYXrl7BO8N0eEo7bzAbHlOZMiiW2w021gf0B31asdoXWomQZRmikcduT+j0eBe6QdiQe7xzGFNR6gW2qqIllgmUkBVOQ3mak6YpSoXWazB1VkugB+GgLWyYZ5EISW1/7IXCe7sshtSmyILQFVvyFwG8jfw9gRBhlFz9JqVaVtHPWuOPPtLVvffCYwDwOu0mxlqkF+hGQt7IyPOw+YQIhHOhw3tJ4GVgXCgwWQMxMeBsmEtIKMHXrVpPBHCAVApPEFEEg8bIw4scvGDZc8bBW75f1QoSBB+veiwQLloFVMynYyYnx9x7/y6VbPL8l79O5TR/+a1vs7OzG1RuCIQ3rPd6XFrvMTcFewcTpJK0UDQanjxLg1otCQ7xVoKQHucNxhkqYyiKktm8oCxLnKsFB48+0E97ZT6FEDXPkXO9DiLuEHi6qEVJpiSuKYOSU4DYeALfz0hkxmY+YMtN6LRzsu0B+jd/HdFuPPqHApVBZX/LA6rAb+AXDXIvaJmCdSyZXNE73cfus0cA3cdUL6WS9Ltdrl7cYm845u7+IQ+c4d6Deygn0M4xyOHZp66yezrnnd1TxidjUl8xPdzj/fGQ//v/sbzy0ktc277EW+/eonIWby+x2Wlx9epldo7eQuqEk9GUN2/d/QQX4bMJKxTtToe9g6NggJsk6EQxni3Y39ujkSV0en2anQ69wQAloZVnpFVBqTwygV6/Q5ZqGs02rW43KDgFSKVDrrc2CuoINk3yrCPjjAlvzuCcJUkSqsUiUE9MbCfCyoK8Zt4MgE7JpUVWGClmzw42oh5VVnveRZNpUc+uqHmK7kycwllDoK7RudhVCxo7tQR4KBdNC89m4T4ObdnzsfIAL08UpbekafDQaTRC5U4qjbcCpMI4G8x4qxJjSqypSFSCkgqrFFYJJGe9f0sYY4MIE53w8hyJVMW9K8JgaudDFa++kH24tQbF59njXFUOnncB7Lq4XiKqBk1ZMj45JW+1uPrU87x1f49vvfo6Im9R+ha3br9DkucU8yn30pT59ctsb3YxpmR4MkFLjS1LhINEp+gkCdwNYxEKHBbjDKUxLErDPBpMO1e3zx+DyHmY+1VLwMMnCCdQOiHv9klzSXU4QQ0uQec6Ik1pdfu81Nqh8eJN2v/6v0d99TcQa08FruN/aXgDx9+BP/k3qPmEfrPD4KTCC0979VUWZ/Eo6JPynI0N4eXRiixPGXRabA26bA167B61eO/d9xgdnFBai6tKXnj5i7SGhlHrQ8T+Ibv3PmQjq5CU7O3c509PR/yz3/g1Lmxf4tbd+4DFX7nGtSef5dbdHabTOVIIptPJp70Kn3hcvLDBh/snKDztRsYsa9Hrt5FHxxTlnMvPPIuVKb2odtU4ivmY7fU2crON1GGUGEohVULebKFkOLSL2JkRSkVxRQ3uzgRMXkqskjgTKBAi8bH65JbG3qvsw+mcj44UDlsFazC8RwmHFPH+WPuV4oOQgmClVY84AwJoO3//5GzFlsyGujziPU7YIFZSEq8UUoVWsPzIIfJcFXCFY+UBntaKNGmhdE3sTEBIvFQQL9RiPGMyPqYqKkxVhrmFZYWzhm63R299gyTJgtFimHqFl7ESJBRenJFAhdDBEUHEU108Y+DPnTLqB3dOabvUia9YOFeTjcN6SCEQDpSHrfUNyk3NW3dv8//95ff52q/8Nl/71d/if/+D/8Dbb7/OaDIHa5nOp/zgnTsgrvHUxTbj0xGNPKHdbkSrg2BDIACrA5fDO4+1jqKsWMxLiqLCVJbSBA/CMwrMaq47AFLXDOXAeUwU5Drw5/IM9EX47rdRaYbqNEgGF+D6haBiEz3aP3eF7rUb8C//NWLjRgR2f5/1cmBvwQd/AW+8jTgY0jg+5QKGVi4ps8cJ4T0SSyPqs8qySBKyXofLFzc5mUzZOR5yOpnx5g9/yLdefZWXXnge6S1eaHyW8uIXvsjTiwV/9B//GFU8oC0dSa/DpPL8+be/xTd+7uu02x3eeOcWSiZc3Brw9DPP8Obrr1FiaTXzz3gR/vGj2czZ3lxncrBPq9Oh1ezT2ljn2rUrKGFRzXWGwwn9Zk5DeJQt6DQ0DZ2FUW9K4Z1FqOhhh0OoNLQalVyqQ6VSkY4QeHg4T1DqeaSXIdERWoXO6WjKe26/ryj32lUVZVkFn1IRTaOFCC4HkSd3vnPlcEtbq9BwOFsXD4SRfuJhUOzP9b3O3VeFFBFgC6RSaF2DvPA6ndFhVx3ePQYAL8+yIJyIpokIhVJZMPSvSg527rMoFuTdAd0LWzQ7HRaFYTocYYoRH96+xe7ePleuXg/VvyQJ1SIdDXyVqHuw0TFbIElq1AYitmvtI5tJsKwC+vj5KoY9p16SQpAoRaLU0sdueHzKnbff49d/6df52m/+cwqR8NJLNzm++wqno1NuvfN6sJHRKb2rz1FWh3Q7ErMocZXBWIONAhgZX2cVzZWdI04CMBjjMNYzmS0w9mHO14ouPTy/FcBcmgRwV/dnTXybWkSvC1sXod+GrevwuS+FcWUkqK//E/iN34akxd9vleKedxO49edw+z4UCdgxvY4jbSQU2lE1V3SSxaNV+R93M6+/J1UABXlOd9DjycsX2T095f7RCbfvtvjTb36T5597HuGDJ5ixlrt7JzgTuF4N7Wkmmus3bnKwMHxw+11+8Mab/OzXvo4h5c3b75Mkgn63w3qvQyvRVGb1ZgFvX9hg98EBmRJsbGzQ277GwbxANVvgPZPJjAvdNoMEZDUlkZ5+r4NzNrRYvUfoMA1BKoVONCpyyqSUUZBuAkdMBnFFSOdn+U6cQY6lK4OUElfPyV7Rrg0EgHe+jYpUCCmx9TDJKGL0tZ2Jq9W0D9uG1UbF4mOcKPzZ8i7f19VUISVCgdQS7zQuVvJCLAl8q/wSAI8BwFNxY0klUTrBC43zlqIsOLq/Q9Zqs3b5GirJsEiEyskzS6km9Ne3wQp279/h7bfe4tr1a3S6XfIsQyUaqTUSEap5y6jVovFU58qovDVx08Y2mahrRzXZdjVhRlUnM4gy9+BGnqQZUirWvOJLL7zIc698hfX1Dd69fZu//NM/YjQ6YXv7Cu+88X1SFdraF6/foNiB1B2TiiBiCSaYrv4DsYUSUHNVVRHchVm/3kG70yHL83PCxYc5kSsVrQEYD1MfKwv1PFob3hsJ6xtweRO2r8D1G7B1E8jCduz9DMhln/e/IGp26gj8KUxfhYNvwXgX1kvE5y4iC0HLzGmlGt9evQpSiPp6rz/9OMAXKz/GhfeBSISQCeQNNjc22N5c4+rmOnc213j39e/ywc4uo5ljerBDaRW3vvceh6dzGmLOZl/QaDS4ceMGv/K5V/j9/+X3+eDBDms/vM3axhbv3b3F5lqPp7e3uPH0Uxzs7TI8Pf0U1+TTiUarjbEPyBLNYG2dK9ev0htPeLB/SFUsuDbocuXCOtIuyJWL1JtwSHTWI6VACoVKdDjQS4VO0tD9QUCt1PSRh4cMIO8h9lgAfGLpyRamynhjl8BmVYtIpTFAUMUqqXDOUjmHsxXOhXuhq+1RvF8CXhH9IJdg7ry9iTtzrQCWYkaIgg1JzP/hvVQSvI40Kn9mbcbKLvtHYuUBXo2npAqt1DAmxXC4v0d3a4tWs41zjul4xMJA/1LOyXjI8fERG82ColjQW1tjNi+4/f77XL/2JP3+gNR5dNBko3XtlyfjhR8+985AfWEv83x98zuX71e4hDcvK6zz4eITAqkFeSsH0ac00BhIklaHPE/IWhk6U5wc7HCwe48HD+5jpQJhEVLQa7dwl5/i9M6UtX5Go90iUQJhKxRhaH1oj3twFl8VmKqgrCp8FHisDTpB8QbU4Hpl27STSWhtiEBoRgtQCrQClYLvwEYXtvpw5Vm48DlQW2cHDvXEP+CPCyAF1qH1dfjKJbjwn2BcgV0gFg4W46AB6a/YnKxlnF3rD8dZi2l5wBOwnHahgtpeZDlJV/DK5z9HYQXvvP8Bh8fH/F9/+H/wzGaf/Qd/w2Q646qas96xrGWeXuLJNy/x+a/+PJdv3OR3fud3+bf/6+9z+/Yt2t0eZek4OBpyqd9jq9elmSWMmo+KZn76QyUpWZaSSE233SbXEjJNa7NLokK+SaVBJBm+VmkKcD609VT0bfNIpI5jxmJboG6xCinD4TL0FJEorLHhfxPgCca8zlmctThr8D6oaM8O9qsZpbVRvQqCOH/dOnx0phBKhbGgOkUlgR8v4kSnAIQDTz3kahGsUoxFWIe3gQPvYhdMRM57SHVhXaUPvO9aYbvKa/3j4jEAeGJpqmi8w9uK0fCUdruL9zCfjJlN5+wPRyTNDjMJO7u7iNJSTQoayrJ7eMig2+fgoOTw+BSlElqtBg2dIrWKJehgwoiSMac7PBbhXVTW+odP8DEpnMVqninKsozeUOFEppIUJRVaZRgDJ5MZpXWM9n7I2pPP0Gl2yPImQilm41M0oWzfanZod5qkTY05XceoBXm/SZo2kCqJVVOWXIvaBynYEoS2gNYJp6eTMJroI490BTPAoBHAnFYBNNQu8aL2xOvA+jNw4QIMXo7g7ryA4u+7JvXvNUA4EDkkbXh6k8BHXYCbheqHOwJm/6Cn+RMb4lxLaBn+kQ9jVUES8oPlLC8kaQAW0wlr7SbPPnWdg5Mx7732ba7/2j8j6Q1IZiO6SUlXeaQ1NPobXLv5eZ557nm66+v8zM/8DO/+4Fu8/dprlIs5Es/O7i4vXLuA8Cn9bptOumpzLCDJc1pZwmZ/QOo95ckezUaGbKmo7gxzer2QseXqwEu0iq3BJEFFvnZdUarnpga/tQjCpQ77GJbqUGdD69bHdm8oHoU8Zh1nIIYV7RwQPWU9WHfWuQoH/DS8RR88pfWSv66FDe1Z53BYnHUURYV3jkRIFOCtXZoiO3E2dGBZwVNR8CIk0jm8q19bv9J+sz8qVh/gxeOxM4HoX5oK7wRaJSwWBXfv7aKV4v7xCVYdc/z9EcPxmDRNSKRk0GiQJYpUzkkaLYQUTKdTtE5IcotODUJkqJgohFJx4kWcsce5i/ghU9PlPysdZVkGU+japFJqpNAIbzGmQquEdmuNajbi6N3XaG09QavdZTabo7BBVi81l689Sa/fg/mYvJEzPj5Eqm2yvEma5WHUnJQoadFanePJhESqVBhH5Aw00ibwGKz+YCMAOS0DyEs1ZBnkOeQtSLegdQmyayAvRLL4P2YCFATkIkF0whs2fK4gcGHG8WurHOJHfHz+07raF70K628qRdJscWFjjS/duMHB3hHSw7f++pt84fMvs7AacfABHS3IWh2uf/6r/Mbv/rdsX7qIwdHrdrn57LPcv32L05Nj0jTldDTEVgVadkiVpLe58Qk//08/tFYoU3L96eusbW+Q5GkQTUgdxkiKmKsJYymtNSil8ELinEOnaeCNAZxvF9YcsNpyAxAEG61gjwUeifemTv44a6LhcQA71jlsVXd3Prs1+iSjcm4J7KTWYaZ75KobPAqLMxZjSnAOYwzGW4y1VFVYL+f8clJFrgWpkmhEmBTiHF6ctXaJFTwlQAm1vITqb/+YOvpKx8oDPK2y4NbvPSaeCLKsQbkoOT484v7xCbsHR4ynE5SQWCFwFoZHx1hgX6dsDXoMGzOeuLSBTnKsc3hvcS54jEkpo5GjRHpAitCWDEVigDNi6I98pKsO5qauAAAXYklEQVR5pZs4JozIv5NKoRDgBEKGuaapyinsmJ3XXiW9OubmzZt892/+cslmyZotXvnq10izhHLu8LaglTXJVJtmu0PWaAR1tBIYHwylxVLkEgU2UlAWhrV2C63CtvdLBvCKuhDevAlZB7Iu6C7oNog2iAaQEappRyCSTzDbPfofP5JyRO+T+sM/+XFOObs0n679CkWYNiI8yHabze1t5gvDtUsXeXBwzMIVfPfNd/jlb/wTTFmSMuXZZ5/hiz/3y2xdvEiaZiwWMxpac3XrAt1mTlEUqKyJM475bI6WEq0kSbJ6twGZZAy2r5C1B6TtDt47nAnEfxdhWa3OdB5QGpFmgTNGTfVQsXMe1ZtSh9cpgjtRi/GFiHy68BoKCGDOOZx1WGPDx8YEYUxVURbl0gN1FcPXYz4JFcsghAsVOBO7KlVlsNZSlCWz+ZxZaZBKUZlgBO09KBVGneXakWtFv9kik4qlZCya6LsozMjThLzVDDPlhV8KGYFznZ3HJ1bvyn4khJDLDealxFYOpGU8nfH+g0N+eO8B82LBYjLFCviFX/0l3nj9LWYHu0GgkXvu7x/S67VoZ5rN/oBeq4lSAikcQsSSvwwzVoOv1Rlx/8wBQSxPGh8XK3uSMwbrPMpH0nJs1YZM6lFKkKYJmcw5PSnZ3/krejrl5ZvP89Z7t9DA137+57nxzFPMT04Y733IfDhkY9Ch1WqQZBlpmgWbFCmx1oVClKsQPvBh0lQjLJSFYTqfkeio2jz3UqwkD+/iv4R6jLwwQAUsgClwHKpnzoJaI/YG+fTPtSu47n/nOFepq/3wJGetXemD4MI5ZJJy6fJFfuGVzzGeDLl/OubtH97mz/70P/Jf/6v/gVe+/FW2LmzS7TTJsrC/rXOYakGeCNqJYmwD4R8XwE6qNXmShur3ioVsdOg98VyorFkbPNaVxFu7vMm7iK+kTqKnXRLNvgPdJsxFjVwwarc1FcQAhPzunUN4B9YuVbXeW4QInm7OVlhrqMqKytTVKY8xblWPlQAUplpOefJ4jAnCxqKscD7ckxdlgVKKJE1J2h26XoShBEVBohWdVhspBXmW4WxJoiRaSmbzOSp2ZwTB7F5IiZfgtUIoFb0lWVp/fhywW+X1r2P1ruxHwhM2glOSqijAwXS+4PB0yu39PTqDAXlV8WA8pfCO+aLCRaUnHi5evowpCsajEfcPT2molHaWI4RGqpQkSeMg5WB6iZARWAik0DhXPlxG/qwX5FOOk9E4eBulKfWA6GU1T4bTmasEzVafdtphvnNI4sa8uNnhYvfzJN0eTz91HbPzPqPDD9BSsL01YPPiOnkzQ0V+ZZKmCAGVUSghQlK3Fi3DbE9EmGqSpsk5Pk1dvl/Rk7QI5rYBNSSEql0D6BJckCWouyBahFTwGPvRfRLxt53azlM2lpU7YOmdBHgZqks5ZKrNtUub/OKXv8D/+81vY69c5+133+EP/uf/iXu33+O3/8XvkT75BIIc5yyL0QSxKGgLTy9LKReGyhuEMyQS8lTTyFLyZPVsaupxed74s9wrFEK55RrLWP1RaSPy8aL0SojQzgW8M1HcLOqbSajYWRu4X3hcBHZEwGerKnDFKhMqeTbwyYLiPxYboqhmFdMOwHQyoygrjHOkeY6xluFoRGUMqVRkWUo7zWg2GmRpSpZmVJErnSYJeEei9f/f3rn1yHFdV/g7l7r0fS6cm0SLsmhBii0gTgwY+RH+gXnMQ35BXoI8J0gAI44fAkUIDEN27JgWKXLImenp6a6qs/Owz6mqHg3zJEt2uRbBaXZXT8/wVJ1T6+y99toYhMw56spQZBnWGW4JYKQNnLhoPGucpShzGmlwTjdGxkZyZ9ACl6EO+FsweIKnOyndte7u7qiqilevLvn8V79leXjIj//mx/zXf37K5ssvWF+t+Zd//jdCaLiYT3FZwZOnT1nMZ3z205+yrXZc393x2Dms92STCXlZ4n2mlVbOktrKtnYdKQMDb13wdf0YJvV78eoNdWwkrZC4s4pFF84BFeVsxsk77yLXa+zVC26vrsHvyDNo3nyOn8xY5LnqIMNGU+JZFgl2Fj0OjTaxNtp2zhnIvSPPdFftvOfk7NEgU1IP40OU2EUdXIuo9TLbeHzBn3ck7Q+E1r6n/dKhP9/fNveTTtcY8A5TGxanj3jaCFUt/Pyzz3j27LfsMPzTP/4D//Hzf+cnP/kJH3/8McvVCnO7oXr+jEm142Ba8mL9Jev1LfPcs1xMmU9K5lPt7DM4RF81rMfYTK1KWpsMaQuN1PnAdTq71oQ45seNjb5sDRYIPR82/aiuQ4WEoIUVTU1TVzRNBbEndl1rsUBd1eyqmrqRmNodJjJncGVOHi2p1re3nB4dklyByrJgUk5w1jKZTHDWsFlvmOQFTV2pjhoIjWZhsmkJaEVyPptgvaUJtaa/xZIXBVkZN/m7nRbcGYvLtOhCCAQxJFVOq84ZOAZ/pzNWb/YhNEgT2G42fPHi91w3O370w7/m3cfv8ux3v+fm+Jgbm3Gz3jGbTFgtPK6YcvbOd5jPJtxcnHNz9YbTI208vThYMZlNcXmGzdRrD2OwojJpE3+2Mf0+ePtoo0j65BsclW8Ol6+v2VVfNVJVcmdxLmBsRRMs5dmKVXif7GbC3S9/yfbZGkxDMZviKBG5prm+pGkq5PiYvFiSZRrBM8mlXARvLZnzOGPw3pJ7iwQV/m5ut9RZZ1vTnpVB8psMqIEdsEXTs3doqrZCSd/Ft/bbDR7xmty/k0in/I5rBiIPX4itFk9UMJprZOPo7IQfTUuaes3Lly95c7tlvdnw7PkL/v7v/paD2YK//PgH/MX5Kd85OGA6ybg4O+FXL5/TbO+YzZYspwWzXCPfDHHD45S0GgmdyD55k2IxMS0tGtpRYmdsd6pS+7EYaaNJTgggkbhJ03SfLxLTwTWhqbuCirozTq6qiu1ObZuqKlbRfrOj8o3h6HAFmOhiEFgcH2GNdhxyuWcynWCNwTuHNYZqu8PFatlgPSFoFM7H6HJoKt2kx+IsqWusiOrfsxKf59R1jUjAScCihTbG2VhUY+Kj/n5DHff7GODM3ofNPKGu1WMnBF5eXvLq9TXf/+Ff8eS9JywWcx4/eZ/N5Wte3f6C7a7h4vQRJ5MdO19wcLzC3m5YTHOoSo4PVywOl0yXcyaTCTbLtF1N9FZLpfSQcvwWxHXbhngE2I9qDRShFqoasB5nDanDmwg4Y4CAsY76dsPu5g5fwNHFEy63G3bXFbXL8QHurl8TNtfUTY1dzJnMV5R5QZFlao2CpWlqpA6YIGTGkllPZiyF92AE2dXc3tz0HM15e/RkEPg0PnqU7JV06dmcjgDuXZwjvi60koQHjnVO2/fCCW4/bZsejR4z1uN8Qzkt+eh7H3B3t+WLL1+x2dVcPXnM716+5PPPP+dnP/tXXl+ccfX0e3z/6Yf4zHN0dMTV5gWPDhYcLw8oshyyDDPECJ6JZtGhVpuqoPpDkwoiQmjHN5nj6pO4OLWttGJ/cdPoZ0lQ+w00WocE1bHWFaGuIxFBg4SoDCX1nhW0fWLbyYKHN/5DwGw6addZI6juWQwSAnkR+7zXNSaaShMEfGwDF8CIw1lLiERZQsAaiZp6g3FaOKkBFIcEwUoMumIwRj9XelXPQ86UvQ2DJ3jeewRo6oaq2vK7Z89597tP+eDph8xmc4qi4Oz8jN8fnfHB+pLjxYbVas7EbFgsz5g5w83NGzZ3NfPFgtlsynQxZzqftzsEbL9CttN20XRt7ff6D8b3pQxMm9cdIGZlob5zpptcJlalafWxBjJEaqrthvpuQ8gajHeY0kFTsVu/pmpuaaoaKTPOT05YrJZ478mzHGNUUJsaWqfuFqn/oXdW87XWcnO9pqq68R72fP8YLbJ4Ww/ZgJK9EX8Q3Pe93IvOvaXoqn1fL/IXpHvuNeVu6sBqtuD7Tz/gaDnnzfUtd9uGxwcnvDM/YLPbUW3v+OwXv+Dm1UvMNGcymfL4nXM+fP8xR4slPs/089zwbgOyu4WoqzPGR1VCrJ+VaLgbbVIQH9PpTczEONXSNXUvA+MQq+sMVRVL8NN6lrS83bk01hCaroViK0n5StP7YaL0qpcjkq5oTac+dU3QKlchkjvtVJRZF6te9foPsYWeiJpLG2tiRM/EzkRamZt5S5AQyaEgEg2Sk99nuvcMe7F/EMOb2ffgnEdEqK1hfX1NYQq++/QjZrMZLobxl4spIdwxLT1FVpBPMnZ3wvnpO9jbO6hrVvMZs1nJ6uCAcjbD5VkM4fd648XKnYSky30YprfmD/fCO3204m67JSnIO5IHzhucM5hQEahoqNmub7i8/ILNly9p1ldgLbf1mloasumMw4sLTs7PmU1nOOe69LdI3BnrzbBrLm2wFkrnsVYw85kSwR6GW02Vo1dg03vUqGnbl7bV6I342hH6T1JkP2q7UmTvfrTuPrl74Ns1/+QoipKj1QEeyzJ/zfXNNTNrsBywA/Jyypcvv+TXv/4VeeX56JNPuH51yXsXJ8xnU4z3WnE4wCpaJEBTY32OoJmDYB3GOGi2SvZiNSyGuCGkjd6lcjgRiZYp8XlvvcGYqMcxrVtCGwuPT2zbdivKdazt3jvUZQdw1lJXWnzivY9kLKZKg5DaVmqGXHWMID0HipgNswaf560ZtWodBecsMTCKlQanTB2cQYiSqfSGXgQvRJLfYsgngT8DglcUKry8W19Rb7Z89INPyI1FqgqpG6qd2gW8/8FTPv3iN4TdFuenLFcXHEwX3G3W1NstRwcritIzX8xVzGm1QCBNaCPQxCiSiHr4hOjBI9Z0eun4+OAiPkCcnhypNuKB/6wxgrOQeUMjFVvZsWt2rN+8Yf3mkm2zpaqgMpZiPmNxesLp+QXL5Yosj+1tkuVBCNGfUNqF1cUG014E62KroTzrjXs/5jpEbOmInEWne38zMdyNxR8H+qnvnvgntSRLLwf0i7FqjXIf7XvSP2Krp8mUqfMUeclyNmOzuWG93nB8uMBmHslLrr7zHm6y4vrqkvNHRxyXhvfeecR0XqrptffDJHig2jpE572JaUDb38xIS8Q6/tyP7ptobZfIOOrZbRwiVY/QpehQV1hnSO3PuuidSW0suZeVHyBCE9qxTH6BffmSjT53IhrtlKDkzVqrrgtxHJ1z5GVBVVVYl5GMo501bYrcWe1iAaihfmxZGdr7rPSmX+++a2TwS+BAZ3YHk3tsXbPd3OF9htjAr//7U2bP56yOT5gdHJKVJYtpyZOPPkEa4fD8HG8tr59/QV5VLOczrBXyIqecTbExcgQ6gTX3L9FXSRC052C6ntt+eMQL7H7UbsDX2enxMbd3W51TvbVVSbHubH1m8T5Qh1vumhuqesMmVNw0W8Q4JosVB6fHnDw+Y3m4ZDIt8JnDWfBGkCBUjeosTTS8NNaRZZ7MuXbHGJwgscXQoFfXFin9OtSr648cNoXweuSuPRW919pjD2gh2whDfD2YuKBYjPOYwmC8x5clk+qAZbXh6HajRWWZYVcJszzj9dU1FyczitMlh/M5WV5ooYGzg0zRthG2eHM3GqbbH952uDs9XrtoE3ug0juHfU2kGuvFpx2Zk/h9iROauNYnVwXbuisMfw3yrmtDRpTPKN+Ko2pM20bSGu3lnhLfzmratWm044XFEOqdEmXUIsU6tTs2VrM0EgfdWi2W6VdNp0KXEH+Pbo8/7JMwwJl9H+p5Zq3DFzl1dctyOYNQ8+J/f8OrF88RkzGZzXh0eMj84IA3r99we3VFs14zWc6ZlDkiDYvVkqzIVdvVc+o2MVoXWm+17qLZ6zX4tnTs/5vK/dPG+ekJz1++1BuOxKbdEnfVGJy1ZHnGYjlnt7tjs76ivgImOfNZQTGZsjo44tHFGatHR5TzCT53ajSNYBrVyhAr2pI3lbWeLMvIMkeza9o0jfPDtSb4KkZi960i1Pu6nz0N0D3CZqJIqc/x0trRavASaTEaiUrFXWLA600yt0Xc1DSqSyot86KgOjvG2prCWbzPML5Q/Z0daIq2RT8Fnma+1ciexJBcq4VM6fFeCn0PMS5nbORnMSvQ6sbonWPTrveqw0utGlUXbGLMcKhrkTWazQqRjOl/N5IrET3eqO2Jd17buVVCExp85gmx1WcqtLDWxCIL1e21EpxE3ntdFvVHCSEVyoh2FmmaEO/RHbke6vgnDHlmA9EeRYSynHL+7hl1MFSbCmlqXr96zW67Y13dInWFbSqe/c/nLGcrZkXG/GgV6yeESSyucEWOdXZ/7ovs6bqkXTxGHC4XbKuqHQ8Nx/fGJ9qlFEXBweEhPss4PDlhu61oEDKfUxYl+XRCOZmSJd9B69SGIDSxkXfQCRxUkOu8U4LnM7aVvsc4Lct/COPZGvG1IxGzFr1onE35PdfZpSSicM9nrbsb9clKJHjpzhZTgFG5jqlrbAjgLNMsBwlIqLSwIPUmtg5sBvLwnPhTRqyV1XXGdATMiEQvO2glGpKa0O+FU/lKIrWnw6PH51IKOBm5p1O4V1Rme6laY7UwDAa78NheNNoI6nAQMymIUIfYq12EWqpI2GIWvTd+3jucdTEIYqI0KrYF7RVQpH2PQHs/sBgCUDeBula7lva+I0Om1x0GT/CaRrVZxlmcszS7CpqaUDfMyhJfV0xnU5oAmQhTn3FwsKTIPYSaELSSqpzNcEUWhfsRvYhdEoqG1nSptzyEt+VgH9olDgtlkTOdljoO7aJo2rFLi573njIvcEvLdDKlroUQJ691Vo2lnXaucM5GSU1oxdIhmY3Gz7XGat/b2Kw6hEb1GUbahbePgZ+GEd8G9vRe98hCihoZibo7AbHdsT2RXi/flyw8ml7UyQo4iSnDVDhgOqNl52JK0Om/XYp42O7vUCFtjC2SsV60Tbr3pMxCy/sSIYtaMdOmeaPOy+gmXyIxTGlaTCrQMFHDJ3s6PGs1a1F9w8PwTaOpa40sQ6uLbgMhopG7pJW21iCN6u/KXPtie+ejzUoXtZNYsOK8Ru9Mq3mM8wihS8qqQkJEN/5p87+3kWL4naUGT/ASeai9wzmPtbuoCwhk0wm2Tn0IPc465n5BmWc4I4g1eJdTTCcUZRFz/r2+dlHWAbSh5yQcld5rpl+1mdKx6UJ7QHYzJFgr5HlOVQckdDtZonbOWodEj7wsNzifkeUBEaMBEAMSd4PeZVjno69K6BaKlijHpIloD8kUyjfW0MQq2yzu+oa/dxvxreM+kdg/+NXXDQ9ognq5pL2gkoEQ04vO9A66jgiGJpIXNUtW2wij73cGskw1eA9seP7k0foL9qKmJi7YaRweyrJI+6X9thTrMW1VRfosfYPaeHSEQXob2cBXrVJsskq5x/mHBEE0wtYGO6TVxYkIzlgaiTYn0enCWot10cPO2XacUm9g49Km3etpaSuUDWK6IAsmIDQE0R64dVXRxAKO/Ujt8DF4gicIxjt8VlDXDS4Xcqmo7lQHNplMwJooa1GjRUujYWGXkWUZ+WRCnudY71vtXf8nEEvpu7mvsWYR1SA0vfkMMWNwf3EZ4BoLqKiVtJPT19LOOJiYTrUZYg3G6dh5GyuR2z8Ki9X+6xI6TUdy+TdxRycNRmrUEgSIO/A6qLA3y94WTR0x4mtG6Btq30vVGnmYYPQhck9/1z9o9gmM9D4vdb/oLzhYJXNWWk9IXA7Wa6p2YOiSrYJFCDHaE0Ra42GI6dk2VcvemLbWKNIjx2jUzsYWZhj2rFJaXbZ+gJ7qNj0bbVKSVQrDpRrWupgZjzfG9Ne6OMQBj95vnTMYq8bGiX8lQgeoJMparHW4TCN7kgolYmVyMN08sNEKRZpGO2lEUplIYwjp/MLQbwZm1IqNGDFixIgRI0YMCwMWX4wYMWLEiBEjRvx5YiR4I0aMGDFixIgRA8NI8EaMGDFixIgRIwaGkeCNGDFixIgRI0YMDCPBGzFixIgRI0aMGBhGgjdixIgRI0aMGDEw/B+/tdZ9uZl0kwAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 720x648 with 30 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UPhHsD_mQpws"
      },
      "source": [
        "Testing Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0eFMfCgpQq6M"
      },
      "source": [
        "from keras.preprocessing.image import load_img\n",
        " \n",
        "image = load_img('/content/gdrive/MyDrive/Infections/Test Images/Cell.jpeg', target_size=(150,150))\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z258xJnLQtzy"
      },
      "source": [
        "from keras.preprocessing.image import img_to_array\n",
        "# convert the image pixels to a numpy array\n",
        "image = img_to_array(image)\n",
        "#image = np.repeat(image, 3, 2)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "D-SNX5daQu4g",
        "outputId": "9030c11f-ae73-4128-ef57-f8196332545f"
      },
      "source": [
        "image = np.expand_dims(image,axis=0)\n",
        "image = np.array(image)/255.0\n",
        "print(image.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(1, 150, 150, 3)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GKE_54E8QwDm",
        "outputId": "52478eb9-54f5-43f8-8848-5767ae0d1b9a"
      },
      "source": [
        "pred = model.predict(image)\n",
        "print(pred.shape)\n",
        "print (pred)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(1, 5)\n",
            "[[4.8816949e-03 9.6760726e-01 6.9175308e-05 6.9940317e-04 2.6742440e-02]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WKCI9HMkQxPj",
        "outputId": "5167f1f8-b497-4d61-eba1-df88dce90d3a"
      },
      "source": [
        "predicted_id = np.argmax(pred, axis=-1)\n",
        "predicted_labels = dataset_labels[predicted_id]\n",
        "print(predicted_labels)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "['Cellulitis Eye']\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}