Classifier_Animals.ipynb

{
  "cells": [
    {
      "attachments": {},
      "cell_type": "markdown",
      "metadata": {
        "id": "RvK82VEVkc6Y"
      },
      "source": [
        "**Project for Hackathon Lab (Prof. Guidotti)**\n",
        "</br>**Studnets: SHB ALF DMT**\n",
        "</br>**May 2023**"
      ]
    },
    {
      "attachments": {},
      "cell_type": "markdown",
      "metadata": {
        "id": "Vb-eUuE5kUsb"
      },
      "source": [
        "**Convolutional Neural Network to Classify Animals**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "m2u_KQZp3EIa"
      },
      "outputs": [],
      "source": [
        "!pip install -q kaggle"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "j3ZAivni4Jky"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "os.environ['KAGGLE_USERNAME'] = \"xxxxxxxxxx\"\n",
        "os.environ['KAGGLE_KEY'] = \"xxxxxxxxxx\""
      ]
    },
    {
      "attachments": {},
      "cell_type": "markdown",
      "metadata": {
        "id": "JUHH-LmpkSoh"
      },
      "source": [
        "Download dataset and unzip"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kQL726Eo4LVb",
        "outputId": "030e2ef9-d868-45f6-a76f-00d0de6c20d6"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Downloading animals10.zip to /content\n",
            "100% 584M/586M [00:26<00:00, 22.7MB/s]\n",
            "100% 586M/586M [00:27<00:00, 22.7MB/s]\n"
          ]
        }
      ],
      "source": [
        "!kaggle datasets download -d alessiocorrado99/animals10"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "WwVWKd0C9uGB"
      },
      "outputs": [],
      "source": [
        "#!rm -r animals10"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "V_8-CxI1-D04"
      },
      "outputs": [],
      "source": [
        "!unzip -q /content/animals10.zip"
      ]
    },
    {
      "attachments": {},
      "cell_type": "markdown",
      "metadata": {
        "id": "EDvRnFV9kLSW"
      },
      "source": [
        "Generate list of subdirectories of diseased leaves"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "5FF1jbxXiE5H"
      },
      "outputs": [],
      "source": [
        "import shutil"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "id": "hsoqNoLIjK1N"
      },
      "outputs": [],
      "source": [
        "path = \"/content/raw-img\""
      ]
    },
    {
      "attachments": {},
      "cell_type": "markdown",
      "metadata": {
        "id": "TYjqlq9ZkEEK"
      },
      "source": [
        "Delete diseased folders, keep only health folders"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 34,
      "metadata": {
        "id": "ziRUufxdzG1L"
      },
      "outputs": [],
      "source": [
        "IMG_SIZE = (100,100)\n",
        "RESCALE_FACTOR = 1.0/255\n",
        "BATCH_SIZE=32\n",
        "SEED=180298"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "metadata": {
        "id": "zhIgLGqrzBdA"
      },
      "outputs": [],
      "source": [
        "from keras.preprocessing.image import ImageDataGenerator"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 51,
      "metadata": {
        "id": "ccOjNu9N-emS"
      },
      "outputs": [],
      "source": [
        "gen_train = ImageDataGenerator(\n",
        "    validation_split=0.2,\n",
        "    rescale=RESCALE_FACTOR,\n",
        "    width_shift_range=0.1,\n",
        "    height_shift_range=0.1,\n",
        "    shear_range=0.1,\n",
        "    zoom_range=0.1,\n",
        "    horizontal_flip=True)\n",
        "#gen_val = ImageDataGenerator(rescale=RESCALE_FACTOR)\n",
        "gen_test = ImageDataGenerator(rescale = RESCALE_FACTOR)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 38,
      "metadata": {
        "id": "B6vtOYv20F1a"
      },
      "outputs": [],
      "source": [
        "import tensorflow as tf"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 53,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TMj9kv8GzNYY",
        "outputId": "5480f43d-cefe-4599-8a42-b9d44f032245"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Found 20947 images belonging to 10 classes.\n"
          ]
        }
      ],
      "source": [
        "dataset_train = gen_train.flow_from_directory(\n",
        "    directory=path,\n",
        "    target_size=IMG_SIZE,\n",
        "    color_mode=\"rgb\",\n",
        "    batch_size=BATCH_SIZE,\n",
        "    class_mode=\"categorical\",\n",
        "    subset='training',\n",
        "    shuffle=True,\n",
        "    seed=SEED\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 54,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "V7e2yPJrXVN7",
        "outputId": "baadbb0d-4778-492b-999b-b02666fba488"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Found 5232 images belonging to 10 classes.\n"
          ]
        }
      ],
      "source": [
        "dataset_val = gen_train.flow_from_directory(\n",
        "    directory=path,\n",
        "    target_size=IMG_SIZE,\n",
        "    color_mode=\"rgb\",\n",
        "    batch_size=BATCH_SIZE,\n",
        "    class_mode=\"categorical\",\n",
        "    subset='validation',\n",
        "    shuffle=True,\n",
        "    seed=SEED\n",
        ")"
      ]
    },
    {
      "attachments": {},
      "cell_type": "markdown",
      "metadata": {
        "id": "NEYLSmH6lvoL"
      },
      "source": [
        "Explore Dataset"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 41,
      "metadata": {
        "id": "YAZchg_Ml4iP"
      },
      "outputs": [],
      "source": [
        "import matplotlib.pyplot as plt"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 57,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 53
        },
        "id": "HmkR2FkAlwkW",
        "outputId": "79ca4ebd-2acd-4518-89ba-af07e4684bb8"
      },
      "outputs": [
        {
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'plt.figure(figsize=(20, 10))\\nfor images, labels in dataset_train[1]:\\n    for i in range(18):\\n        ax = plt.subplot(3, 6, i + 1)\\n        plt.imshow(images[i].numpy().astype(\"uint8\"))\\n        plt.title(int(labels[i]))\\n        plt.axis(\"off\")'"
            ]
          },
          "execution_count": 57,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "\"\"\"plt.figure(figsize=(20, 10))\n",
        "for images, labels in dataset_train[1]:\n",
        "    for i in range(18):\n",
        "        ax = plt.subplot(3, 6, i + 1)\n",
        "        plt.imshow(images[i].numpy().astype(\"uint8\"))\n",
        "        plt.title(int(labels[i]))\n",
        "        plt.axis(\"off\")\"\"\""
      ]
    },
    {
      "attachments": {},
      "cell_type": "markdown",
      "metadata": {
        "id": "dCv9EI1vmjqP"
      },
      "source": [
        "**CNN Model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 59,
      "metadata": {
        "id": "JKH_m3EboQBp"
      },
      "outputs": [],
      "source": [
        "# Default params for create_cnn function\n",
        "#NUM_CLASSES=len(dataset_train.class_names)\n",
        "NUM_CLASSES=10\n",
        "INPUT_SHAPE = IMG_SIZE + (3,)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 27,
      "metadata": {
        "id": "52CTm1jxp-3I"
      },
      "outputs": [],
      "source": [
        "from keras.models import Sequential\n",
        "from keras.layers import Conv2D, MaxPooling2D, Dropout, Flatten, Dense"
      ]
    },
    {
      "attachments": {},
      "cell_type": "markdown",
      "metadata": {
        "id": "nK_3Sv3IqWGo"
      },
      "source": [
        "General function to create a CNN multi-class classifier\n",
        "* Assumes more than 2 classes\n",
        "* Allows to adapt number of convolutional layers\n",
        "* Easy to experiment different architechtures rapidly and clearly\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 60,
      "metadata": {
        "id": "Fw9TnRmamlE7"
      },
      "outputs": [],
      "source": [
        "def create_cnn(conv_layers, dense_units, dropout=0, num_classes=NUM_CLASSES, input_shape=INPUT_SHAPE):\n",
        "  \"\"\"\n",
        "  Create CNN with general arch of:\n",
        "    A- Sequential Convolutional-ReLu-MaxPooling Layers >=1 of Doubling Num of Filters\n",
        "    B- Dense Layers at the end of the convolutional sequence\n",
        "    C- An Output Layer with Nodes = num_classes > 2\n",
        "  \"\"\"\n",
        "\n",
        "  classifier = Sequential()\n",
        "\n",
        "  # Initial Convolutional Block\n",
        "  classifier.add(Conv2D(\n",
        "    filters=32,\n",
        "    kernel_size=(3, 3),\n",
        "    padding=\"same\",\n",
        "    input_shape=input_shape,\n",
        "    activation=\"relu\")\n",
        "  )\n",
        "\n",
        "  classifier.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "\n",
        "  #classifier.add(Dropout(rate=0.2))\n",
        "\n",
        "  # Additional Convolutional Block with twice the num of filters as previous block\n",
        "  for cl in range(conv_layers-1):\n",
        "    # when cl = 0, power = 6, so 2^6 = 64 (since initial layer has 32 filters)\n",
        "    power = 6 + cl\n",
        "\n",
        "    classifier.add(Conv2D(\n",
        "      filters=2 ** power,\n",
        "      kernel_size=(3, 3),\n",
        "      padding=\"same\",\n",
        "      activation=\"relu\")\n",
        "    )\n",
        "\n",
        "    classifier.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "    \n",
        "    if dropout>0:\n",
        "      classifier.add(Dropout(rate=dropout))\n",
        "\n",
        "  classifier.add(Flatten())\n",
        "\n",
        "  # Dense Layers\n",
        "  classifier.add(Dense(units=dense_units, activation=\"relu\"))\n",
        "    \n",
        "  if dropout>0:\n",
        "    classifier.add(Dropout(rate=dropout))\n",
        "\n",
        "  # Output Layer\n",
        "  classifier.add(Dense(units=num_classes, activation=\"softmax\"))\n",
        "\n",
        "  return classifier"
      ]
    },
    {
      "attachments": {},
      "cell_type": "markdown",
      "metadata": {
        "id": "ntGCE9cMLhcf"
      },
      "source": [
        "Use of Colab's Free TPU"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 61,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Kdxlhxtcqx1m",
        "outputId": "4a62bcff-f7b2-4ff1-a535-0eeaab37607a"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d_9 (Conv2D)           (None, 100, 100, 32)      896       \n",
            "                                                                 \n",
            " max_pooling2d_9 (MaxPooling  (None, 50, 50, 32)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_10 (Conv2D)          (None, 50, 50, 64)        18496     \n",
            "                                                                 \n",
            " max_pooling2d_10 (MaxPoolin  (None, 25, 25, 64)       0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " dropout_9 (Dropout)         (None, 25, 25, 64)        0         \n",
            "                                                                 \n",
            " conv2d_11 (Conv2D)          (None, 25, 25, 128)       73856     \n",
            "                                                                 \n",
            " max_pooling2d_11 (MaxPoolin  (None, 12, 12, 128)      0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " dropout_10 (Dropout)        (None, 12, 12, 128)       0         \n",
            "                                                                 \n",
            " conv2d_12 (Conv2D)          (None, 12, 12, 256)       295168    \n",
            "                                                                 \n",
            " max_pooling2d_12 (MaxPoolin  (None, 6, 6, 256)        0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " dropout_11 (Dropout)        (None, 6, 6, 256)         0         \n",
            "                                                                 \n",
            " flatten_2 (Flatten)         (None, 9216)              0         \n",
            "                                                                 \n",
            " dense_4 (Dense)             (None, 48)                442416    \n",
            "                                                                 \n",
            " dropout_12 (Dropout)        (None, 48)                0         \n",
            "                                                                 \n",
            " dense_5 (Dense)             (None, 10)                490       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 831,322\n",
            "Trainable params: 831,322\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n"
          ]
        }
      ],
      "source": [
        "model_1 = create_cnn(conv_layers=4, dense_units=48, dropout=0.25)\n",
        "print(model_1.summary())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 62,
      "metadata": {
        "id": "dBmr42XssMR6"
      },
      "outputs": [],
      "source": [
        "EPOCHS = 50"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 45,
      "metadata": {
        "id": "ddAVm5N8twoI"
      },
      "outputs": [],
      "source": [
        "from tensorflow import keras"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 66,
      "metadata": {
        "id": "bJwrwgbnrCzr"
      },
      "outputs": [],
      "source": [
        "callbacks = [keras.callbacks.ModelCheckpoint(\"save_at_{epoch}.h5\")]\n",
        "model_1.compile(optimizer=\"adam\", loss=\"categorical_crossentropy\", metrics=[\"accuracy\"])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 67,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Iy87sxZ4sQK9",
        "outputId": "1290f4b8-f315-4b46-afdf-d37f484ba628"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "655/655 [==============================] - 130s 195ms/step - loss: 2.0715 - accuracy: 0.2599 - val_loss: 1.7982 - val_accuracy: 0.3633\n",
            "Epoch 2/50\n",
            "655/655 [==============================] - 129s 198ms/step - loss: 1.7642 - accuracy: 0.3821 - val_loss: 1.4730 - val_accuracy: 0.4795\n",
            "Epoch 3/50\n",
            "655/655 [==============================] - 126s 193ms/step - loss: 1.5746 - accuracy: 0.4566 - val_loss: 1.4457 - val_accuracy: 0.4922\n",
            "Epoch 4/50\n",
            "655/655 [==============================] - 126s 192ms/step - loss: 1.4597 - accuracy: 0.4981 - val_loss: 1.3177 - val_accuracy: 0.5573\n",
            "Epoch 5/50\n",
            "655/655 [==============================] - 127s 195ms/step - loss: 1.3679 - accuracy: 0.5286 - val_loss: 1.1967 - val_accuracy: 0.5894\n",
            "Epoch 6/50\n",
            "655/655 [==============================] - 128s 196ms/step - loss: 1.3116 - accuracy: 0.5528 - val_loss: 1.1346 - val_accuracy: 0.6044\n",
            "Epoch 7/50\n",
            "655/655 [==============================] - 128s 195ms/step - loss: 1.2395 - accuracy: 0.5753 - val_loss: 1.1470 - val_accuracy: 0.5944\n",
            "Epoch 8/50\n",
            "655/655 [==============================] - 131s 200ms/step - loss: 1.1863 - accuracy: 0.5940 - val_loss: 1.0283 - val_accuracy: 0.6487\n",
            "Epoch 9/50\n",
            "655/655 [==============================] - 130s 198ms/step - loss: 1.1483 - accuracy: 0.6088 - val_loss: 0.9909 - val_accuracy: 0.6565\n",
            "Epoch 10/50\n",
            "655/655 [==============================] - 128s 195ms/step - loss: 1.1209 - accuracy: 0.6249 - val_loss: 0.9356 - val_accuracy: 0.6821\n",
            "Epoch 11/50\n",
            "655/655 [==============================] - 128s 195ms/step - loss: 1.0711 - accuracy: 0.6341 - val_loss: 0.9169 - val_accuracy: 0.6888\n",
            "Epoch 12/50\n",
            "655/655 [==============================] - 127s 194ms/step - loss: 1.0476 - accuracy: 0.6385 - val_loss: 0.9199 - val_accuracy: 0.6906\n",
            "Epoch 13/50\n",
            "655/655 [==============================] - 132s 201ms/step - loss: 1.0218 - accuracy: 0.6515 - val_loss: 0.8787 - val_accuracy: 0.7055\n",
            "Epoch 14/50\n",
            "655/655 [==============================] - 129s 197ms/step - loss: 1.0056 - accuracy: 0.6599 - val_loss: 0.8942 - val_accuracy: 0.6969\n",
            "Epoch 15/50\n",
            "655/655 [==============================] - 130s 198ms/step - loss: 0.9941 - accuracy: 0.6622 - val_loss: 0.8373 - val_accuracy: 0.7104\n",
            "Epoch 16/50\n",
            "655/655 [==============================] - 129s 197ms/step - loss: 0.9460 - accuracy: 0.6795 - val_loss: 0.8515 - val_accuracy: 0.7162\n",
            "Epoch 17/50\n",
            "655/655 [==============================] - 128s 196ms/step - loss: 0.9462 - accuracy: 0.6809 - val_loss: 0.8124 - val_accuracy: 0.7317\n",
            "Epoch 18/50\n",
            "655/655 [==============================] - 130s 198ms/step - loss: 0.9228 - accuracy: 0.6859 - val_loss: 0.7751 - val_accuracy: 0.7431\n",
            "Epoch 19/50\n",
            "655/655 [==============================] - 149s 227ms/step - loss: 0.9107 - accuracy: 0.6956 - val_loss: 0.8499 - val_accuracy: 0.7114\n",
            "Epoch 20/50\n",
            "655/655 [==============================] - 127s 194ms/step - loss: 0.8996 - accuracy: 0.6980 - val_loss: 0.8145 - val_accuracy: 0.7328\n",
            "Epoch 21/50\n",
            "655/655 [==============================] - 127s 194ms/step - loss: 0.8798 - accuracy: 0.7029 - val_loss: 0.7559 - val_accuracy: 0.7489\n",
            "Epoch 22/50\n",
            "655/655 [==============================] - 127s 195ms/step - loss: 0.8720 - accuracy: 0.7041 - val_loss: 0.7773 - val_accuracy: 0.7414\n",
            "Epoch 23/50\n",
            "655/655 [==============================] - 127s 194ms/step - loss: 0.8563 - accuracy: 0.7105 - val_loss: 0.7535 - val_accuracy: 0.7532\n",
            "Epoch 24/50\n",
            "655/655 [==============================] - 127s 195ms/step - loss: 0.8439 - accuracy: 0.7152 - val_loss: 0.7862 - val_accuracy: 0.7362\n",
            "Epoch 25/50\n",
            "655/655 [==============================] - 126s 193ms/step - loss: 0.8339 - accuracy: 0.7209 - val_loss: 0.7785 - val_accuracy: 0.7441\n",
            "Epoch 26/50\n",
            "655/655 [==============================] - 127s 194ms/step - loss: 0.8122 - accuracy: 0.7240 - val_loss: 0.7281 - val_accuracy: 0.7596\n",
            "Epoch 27/50\n",
            "655/655 [==============================] - 127s 193ms/step - loss: 0.8185 - accuracy: 0.7263 - val_loss: 0.7094 - val_accuracy: 0.7674\n",
            "Epoch 28/50\n",
            "655/655 [==============================] - 128s 196ms/step - loss: 0.7985 - accuracy: 0.7303 - val_loss: 0.7813 - val_accuracy: 0.7429\n",
            "Epoch 29/50\n",
            "655/655 [==============================] - 128s 195ms/step - loss: 0.8006 - accuracy: 0.7317 - val_loss: 0.6792 - val_accuracy: 0.7691\n",
            "Epoch 30/50\n",
            "655/655 [==============================] - 128s 195ms/step - loss: 0.7823 - accuracy: 0.7396 - val_loss: 0.7197 - val_accuracy: 0.7615\n",
            "Epoch 31/50\n",
            "655/655 [==============================] - 127s 194ms/step - loss: 0.7688 - accuracy: 0.7402 - val_loss: 0.7334 - val_accuracy: 0.7546\n",
            "Epoch 32/50\n",
            "655/655 [==============================] - 126s 193ms/step - loss: 0.7638 - accuracy: 0.7417 - val_loss: 0.6848 - val_accuracy: 0.7741\n",
            "Epoch 33/50\n",
            "655/655 [==============================] - 126s 193ms/step - loss: 0.7627 - accuracy: 0.7468 - val_loss: 0.6907 - val_accuracy: 0.7729\n",
            "Epoch 34/50\n",
            "655/655 [==============================] - 127s 194ms/step - loss: 0.7635 - accuracy: 0.7451 - val_loss: 0.7079 - val_accuracy: 0.7676\n",
            "Epoch 35/50\n",
            "655/655 [==============================] - 127s 194ms/step - loss: 0.7490 - accuracy: 0.7515 - val_loss: 0.6902 - val_accuracy: 0.7703\n",
            "Epoch 36/50\n",
            "655/655 [==============================] - 126s 192ms/step - loss: 0.7372 - accuracy: 0.7559 - val_loss: 0.6726 - val_accuracy: 0.7779\n",
            "Epoch 37/50\n",
            "655/655 [==============================] - 127s 194ms/step - loss: 0.7279 - accuracy: 0.7541 - val_loss: 0.6847 - val_accuracy: 0.7727\n",
            "Epoch 38/50\n",
            "655/655 [==============================] - 127s 194ms/step - loss: 0.7399 - accuracy: 0.7525 - val_loss: 0.7233 - val_accuracy: 0.7605\n",
            "Epoch 39/50\n",
            "655/655 [==============================] - 127s 195ms/step - loss: 0.7155 - accuracy: 0.7601 - val_loss: 0.6660 - val_accuracy: 0.7823\n",
            "Epoch 40/50\n",
            "655/655 [==============================] - 127s 194ms/step - loss: 0.7137 - accuracy: 0.7584 - val_loss: 0.6496 - val_accuracy: 0.7886\n",
            "Epoch 41/50\n",
            "655/655 [==============================] - 127s 194ms/step - loss: 0.7059 - accuracy: 0.7644 - val_loss: 0.6424 - val_accuracy: 0.7919\n",
            "Epoch 42/50\n",
            "655/655 [==============================] - 126s 192ms/step - loss: 0.6947 - accuracy: 0.7629 - val_loss: 0.6522 - val_accuracy: 0.7892\n",
            "Epoch 43/50\n",
            "655/655 [==============================] - 127s 193ms/step - loss: 0.7015 - accuracy: 0.7674 - val_loss: 0.6335 - val_accuracy: 0.7924\n",
            "Epoch 44/50\n",
            "655/655 [==============================] - 127s 194ms/step - loss: 0.6883 - accuracy: 0.7690 - val_loss: 0.6735 - val_accuracy: 0.7775\n",
            "Epoch 45/50\n",
            "655/655 [==============================] - 148s 227ms/step - loss: 0.6890 - accuracy: 0.7696 - val_loss: 0.6246 - val_accuracy: 0.7938\n",
            "Epoch 46/50\n",
            "655/655 [==============================] - 127s 194ms/step - loss: 0.6814 - accuracy: 0.7709 - val_loss: 0.6397 - val_accuracy: 0.7892\n",
            "Epoch 47/50\n",
            "655/655 [==============================] - 148s 226ms/step - loss: 0.6790 - accuracy: 0.7732 - val_loss: 0.6485 - val_accuracy: 0.7901\n",
            "Epoch 48/50\n",
            "655/655 [==============================] - 127s 193ms/step - loss: 0.6747 - accuracy: 0.7735 - val_loss: 0.6267 - val_accuracy: 0.7961\n",
            "Epoch 49/50\n",
            "655/655 [==============================] - 131s 200ms/step - loss: 0.6787 - accuracy: 0.7742 - val_loss: 0.7969 - val_accuracy: 0.7360\n",
            "Epoch 50/50\n",
            "655/655 [==============================] - 127s 193ms/step - loss: 0.6604 - accuracy: 0.7835 - val_loss: 0.6461 - val_accuracy: 0.7917\n"
          ]
        }
      ],
      "source": [
        "history = model_1.fit(dataset_train, epochs=EPOCHS, callbacks=callbacks, validation_data=dataset_val)"
      ]
    },
    {
      "attachments": {},
      "cell_type": "markdown",
      "metadata": {
        "id": "3KtRjDDPq1a2"
      },
      "source": [
        "**Save Model Weights & Training History**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 72,
      "metadata": {
        "id": "WL0CEZ0h31Bb"
      },
      "outputs": [],
      "source": [
        "!mkdir -p saved_model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 75,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RrBTKMY13HvX",
        "outputId": "d43047c7-ed7a-4e52-848b-2a591b0fedd3"
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 4 of 4). These functions will not be directly callable after loading.\n"
          ]
        }
      ],
      "source": [
        "model_1.save(\"saved_model/trained_model\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 70,
      "metadata": {
        "id": "GOqTsFwzh8bo"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "np.save(\"trainingHist.npy\",history)"
      ]
    },
    {
      "attachments": {},
      "cell_type": "markdown",
      "metadata": {
        "id": "c_KNWy8fqYmx"
      },
      "source": [
        "**Plot Training and Validation Results**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 71,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 410
        },
        "id": "tFXu37Kkliob",
        "outputId": "fe71f4c5-dbb8-4e1e-c444-f39e5e1208f9"
      },
      "outputs": [
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA04AAAGJCAYAAAC90mOkAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAADo6klEQVR4nOydd3gU1fqA39nd9EogFZJQpUgJ0gSUJkVAFBEF8Yooyk8FUbn2il67XkWvvWIBRRCwEJVQBBGUZigCoZMAIRAghdQt8/tj2CXLpm3YZLPZ732efbI5OzN7zjuz5dtzzncUVVVVBEEQBEEQBEEQhArRubsCgiAIgiAIgiAI9R0JnARBEARBEARBEKpAAidBEARBEARBEIQqkMBJEARBEARBEAShCiRwEgRBEARBEARBqAIJnARBEARBEARBEKpAAidBEARBEARBEIQqkMBJEARBEARBEAShCiRwEgRBEARBEARBqAIJnATBCRRFYebMmU7vd/DgQRRFYfbs2S6vkyAIgiC4CvmcE4SKkcBJ8Dhmz56NoigoisKaNWscHldVlfj4eBRF4aqrrnJDDV1DcnIyiqIQFxeHxWJxd3UEQRCEOqIhf8799ttvKIrCggUL3F0VQXAaCZwEj8Xf35+5c+c6lK9atYrDhw/j5+fnhlq5jjlz5tC8eXMyMzNZsWKFu6sjCIIg1DEN/XNOEDwNCZwEj2XEiBHMnz8fk8lkVz537ly6detGTEyMm2p24RQUFPD9998zY8YMunbtypw5c9xdpQopKChwdxUEQRAaJA35c04QPBEJnASP5cYbb+TkyZOkpKTYykpLS1mwYAETJkwod5+CggL+/e9/Ex8fj5+fH23btuW1115DVVW77UpKSrj//vuJjIwkJCSEq6++msOHD5d7zCNHjnDbbbcRHR2Nn58fF198MZ9++ukFtW3RokUUFRVx/fXXM378eBYuXEhxcbHDdsXFxcycOZOLLroIf39/YmNjGTNmDPv27bNtY7FYePPNN+nUqRP+/v5ERkZy5ZVXsnHjRqDycennj3WfOXMmiqKwY8cOJkyYQKNGjbjssssA2Lp1K5MmTaJly5b4+/sTExPDbbfdxsmTJ8t1NnnyZOLi4vDz86NFixbcddddlJaWsn//fhRF4Y033nDYb+3atSiKwtdff+2sUkEQBI+jIX/OVcX+/fu5/vrriYiIIDAwkEsvvZQlS5Y4bPe///2Piy++mMDAQBo1akT37t3teuny8/O57777aN68OX5+fkRFRTFkyBA2b95cq/UXGiYGd1dAEGpK8+bN6d27N19//TXDhw8H4OeffyY3N5fx48fz1ltv2W2vqipXX301K1euZPLkySQlJfHrr7/y4IMPcuTIEbsv6rfffjtfffUVEyZMoE+fPqxYsYKRI0c61CErK4tLL70URVGYNm0akZGR/Pzzz0yePJm8vDzuu+++GrVtzpw5DBw4kJiYGMaPH88jjzzCjz/+yPXXX2/bxmw2c9VVV7F8+XLGjx/PvffeS35+PikpKWzfvp1WrVoBMHnyZGbPns3w4cO5/fbbMZlM/P777/z555907969RvW7/vrradOmDS+88ILtwzglJYX9+/dz6623EhMTwz///MOHH37IP//8w59//omiKAAcPXqUnj17kpOTw5QpU2jXrh1HjhxhwYIFFBYW0rJlS/r27cucOXO4//77HbyEhIRwzTXX1KjegiAInkRD/pyrjKysLPr06UNhYSHTp0+ncePGfP7551x99dUsWLCAa6+9FoCPPvqI6dOnM3bsWO69916Ki4vZunUrf/31ly2wvPPOO1mwYAHTpk2jQ4cOnDx5kjVr1rBz504uueQSl9ddaOCoguBhfPbZZyqgbtiwQX377bfVkJAQtbCwUFVVVb3++uvVgQMHqqqqqomJierIkSNt+y1evFgF1Oeee87ueGPHjlUVRVH37t2rqqqqpqamqoB699132203YcIEFVCffvppW9nkyZPV2NhYNTs7227b8ePHq2FhYbZ6HThwQAXUzz77rMr2ZWVlqQaDQf3oo49sZX369FGvueYau+0+/fRTFVBff/11h2NYLBZVVVV1xYoVKqBOnz69wm0qq9v57X366adVQL3xxhsdtrW2tSxff/21CqirV6+2lU2cOFHV6XTqhg0bKqzTBx98oALqzp07bY+VlpaqTZo0UW+55RaH/QRBEBoSDflzbuXKlSqgzp8/v8Jt7rvvPhVQf//9d1tZfn6+2qJFC7V58+aq2WxWVVVVr7nmGvXiiy+u9PnCwsLUqVOnVrqNIFQXGaoneDQ33HADRUVF/PTTT+Tn5/PTTz9VOHwhOTkZvV7P9OnT7cr//e9/o6oqP//8s207wGG7839VU1WV7777jlGjRqGqKtnZ2bbbsGHDyM3NrdFQgG+++QadTsd1111nK7vxxhv5+eefOX36tK3su+++o0mTJtxzzz0Ox7D27nz33XcoisLTTz9d4TY14c4773QoCwgIsN0vLi4mOzubSy+9FMDmwWKxsHjxYkaNGlVub5e1TjfccAP+/v52c7t+/fVXsrOz+de//lXjeguCIHgaDfFzriqSk5Pp2bOnbSg4QHBwMFOmTOHgwYPs2LEDgPDwcA4fPsyGDRsqPFZ4eDh//fUXR48edXk9Be9DAifBo4mMjGTw4MHMnTuXhQsXYjabGTt2bLnbHjp0iLi4OEJCQuzK27dvb3vc+len09mGullp27at3f8nTpwgJyeHDz/8kMjISLvbrbfeCsDx48edbtNXX31Fz549OXnyJHv37mXv3r107dqV0tJS5s+fb9tu3759tG3bFoOh4hG3+/btIy4ujoiICKfrURktWrRwKDt16hT33nsv0dHRBAQEEBkZadsuNzcX0Jzl5eXRsWPHSo8fHh7OqFGj7Mapz5kzh6ZNmzJo0CAXtkQQBKF+0xA/56ri0KFDDnUprx0PP/wwwcHB9OzZkzZt2jB16lT++OMPu31eeeUVtm/fTnx8PD179mTmzJns37/f5XUWvAOZ4yR4PBMmTOCOO+7g2LFjDB8+nPDw8Dp5XuvaSv/617+45ZZbyt2mc+fOTh1zz549tl/O2rRp4/D4nDlzmDJlipM1rZyKep7MZnOF+5TtXbJyww03sHbtWh588EGSkpIIDg7GYrFw5ZVX1mgdqokTJzJ//nzWrl1Lp06d+OGHH7j77rvR6eT3HkEQvIuG9DnnStq3b09aWho//fQTv/zyC9999x3vvvsuTz31FM888wygfTZdfvnlLFq0iKVLl/Lqq6/y8ssvs3DhQtu8MUGoLhI4CR7Ptddey//93//x559/Mm/evAq3S0xMZNmyZeTn59v9Grdr1y7b49a/FovF1qNjJS0tze541kxEZrOZwYMHu6Qtc+bMwcfHhy+//BK9Xm/32Jo1a3jrrbdIT08nISGBVq1a8ddff2E0GvHx8Sn3eK1ateLXX3/l1KlTFfY6NWrUCICcnBy7cusvetXh9OnTLF++nGeeeYannnrKVr5nzx677SIjIwkNDWX79u1VHvPKK68kMjKSOXPm0KtXLwoLC7n55purXSdBEISGQkP6nKsOiYmJDnUBx3YABAUFMW7cOMaNG0dpaSljxozh+eef59FHH8Xf3x+A2NhY7r77bu6++26OHz/OJZdcwvPPPy+Bk+A08tOt4PEEBwfz3nvvMXPmTEaNGlXhdiNGjMBsNvP222/blb/xxhsoimJ7A7X+PT9b0axZs+z+1+v1XHfddXz33XflBgInTpxwui1z5szh8ssvZ9y4cYwdO9bu9uCDDwLYUnFfd911ZGdnO7QHsGW6u+6661BV1fbLW3nbhIaG0qRJE1avXm33+LvvvlvteluDPPW8dLfnO9PpdIwePZoff/zRlg69vDoBGAwGbrzxRr799ltmz55Np06d3PrLpiAIgrtoSJ9z1WHEiBGsX7+edevW2coKCgr48MMPad68OR06dABwWO7C19eXDh06oKoqRqMRs9lsGypuJSoqiri4OEpKSmql7kLDRnqchAZBRUMIyjJq1CgGDhzI448/zsGDB+nSpQtLly7l+++/57777rON9U5KSuLGG2/k3XffJTc3lz59+rB8+XL27t3rcMyXXnqJlStX0qtXL+644w46dOjAqVOn2Lx5M8uWLePUqVPVbsNff/3F3r17mTZtWrmPN23alEsuuYQ5c+bw8MMPM3HiRL744gtmzJjB+vXrufzyyykoKGDZsmXcfffdXHPNNQwcOJCbb76Zt956iz179tiGzf3+++8MHDjQ9ly33347L730Erfffjvdu3dn9erV7N69u9p1Dw0NpV+/frzyyisYjUaaNm3K0qVLOXDggMO2L7zwAkuXLqV///5MmTKF9u3bk5mZyfz581mzZo3dEJSJEyfy1ltvsXLlSl5++eVq10cQBKGh0RA+58ry3Xff2XqQzm/nI488YkvBPn36dCIiIvj88885cOAA3333nW3I9tChQ4mJiaFv375ER0ezc+dO3n77bUaOHElISAg5OTk0a9aMsWPH0qVLF4KDg1m2bBkbNmzgv//9b43qLXg57knmJwg1p2ya1so4P02rqmrpTO+//341Li5O9fHxUdu0aaO++uqrtjTYVoqKitTp06erjRs3VoOCgtRRo0apGRkZDmlaVVVLHz516lQ1Pj5e9fHxUWNiYtQrrrhC/fDDD23bVCdN6z333KMC6r59+yrcZubMmSqgbtmyRVVVLQX4448/rrZo0cL23GPHjrU7hslkUl999VW1Xbt2qq+vrxoZGakOHz5c3bRpk22bwsJCdfLkyWpYWJgaEhKi3nDDDerx48crTEd+4sQJh7odPnxYvfbaa9Xw8HA1LCxMvf7669WjR4+W6+zQoUPqxIkT1cjISNXPz09t2bKlOnXqVLWkpMThuBdffLGq0+nUw4cPV+hFEAShIdFQP+dU9Vw68opu1hTk+/btU8eOHauGh4er/v7+as+ePdWffvrJ7lgffPCB2q9fP7Vx48aqn5+f2qpVK/XBBx9Uc3NzVVVV1ZKSEvXBBx9Uu3TpooaEhKhBQUFqly5d1HfffbfSOgpCRSiqet7YGkEQhHpE165diYiIYPny5e6uiiAIgiAIXozMcRIEod6yceNGUlNTmThxorurIgiCIAiClyM9ToIg1Du2b9/Opk2b+O9//0t2djb79++3ZUcSBEEQBEFwB9LjJAhCvWPBggXceuutGI1Gvv76awmaBEEQBEFwO24NnFavXs2oUaOIi4tDURQWL15c5T6//fYbl1xyCX5+frRu3ZrZs2fXej0FQahbZs6cicViYefOnfTv39/d1REEQRAEQXBv4FRQUECXLl145513qrX9gQMHGDlyJAMHDiQ1NZX77ruP22+/nV9//bWWayoIgiAIgiAIgjdTb+Y4KYrCokWLGD16dIXbPPzwwyxZssRuEbbx48eTk5PDL7/8Uge1FARBEARBEATBG/GoBXDXrVvH4MGD7cqGDRvGfffdV+E+JSUldqtDWywWTp06RePGjVEUpbaqKgiCIJSDqqrk5+cTFxdnW8RS0D6bjh49SkhIiHw2CYIg1CHOfC55VOB07NgxoqOj7cqio6PJy8ujqKiIgIAAh31efPFFnnnmmbqqoiAIglANMjIyaNasmburUW84evQo8fHx7q6GIAiC11KdzyWPCpxqwqOPPsqMGTNs/+fm5pKQkMCBAwcICQkBQKfTodfrMZvNWCwW27bWcpPJhHVEo9FoZOXKlQwdOhRFUSg70lGv16PT6TAajXZ1MBg0zSaTqVrlPj4+WCwWzGazrUxRFAwGQ4XlFdW9Om0qW/eKys9vk6qqpKSkMHDgQHx8fBpEm5w9TwBLly61c+DpbXL2PKmq6uDA09tUk/NUVFTEihUrbB4aQpucPU+FhYWsXLnS5qCiuhcWFpKQkGB7/xU0rD4yMjIIDQ11en+j0cjSpUsZOnSo3XuyNyEONMSDOABxYKU6HvLy8oiPj6/W55JHBU4xMTFkZWXZlWVlZREaGlpubxOAn58ffn5+DuURERE1+nCyWCxcccUVNGrUyKuHmVg9NGnSxGs9iANxYCU4ONjrPVTXQdkfGYRzWH2EhobKZ1MNEQca4kEcgDiw4oyH6nwueVTg1Lt3b5KTk+3KUlJS6N27d53VQafTERUVVWfPV18RD+IAxIEV8SAO3I34FwdWxIM4AHFgxdUe3BqCnjlzhtTUVFJTUwEt3Xhqairp6emANsxu4sSJtu3vvPNO9u/fz0MPPcSuXbt49913+fbbb7n//vvrrM5Go5ElS5Y4DLXxNsSDOABxYEU8iAN3I/7FgRXxIA5AHFhxtQe3Bk4bN26ka9eudO3aFYAZM2bQtWtXnnrqKQAyMzNtQRRAixYtWLJkCSkpKXTp0oX//ve/fPzxxwwbNqxO613eXBdvRDyIAxAHVsSDOHA34l8cWBEP4gDEgRVXenDrUL0BAwZQ2TJSs2fPLnefv//+uxZrJQiCIAiCIHgrqqpiMpnsEu14GkajEYPBQHFxsUe340KxejCbzS5JkuFRc5wEQRAEQRAEobYoLS0lMzOTwsJCd1flglBVlZiYGDIyMrw6GY/Vw4EDB4iPjyc4OPiCjqeolXX5NEDy8vIICwsjNze3RpmLrItkefsiheJBHIA4sCIequ/gQt+DGyry2XThiAMN8VBzBxaLhT179qDX64mMjMTX19djHaqqisViQafTeWwbXIGqqpjNZk6dOkVhYSFt2rRBr9fbbePM+6/0ONWAilKfexviQRyAOLAiHsSBuxH/4sCKeKiZg9LSUiwWC/Hx8QQGBtZCreoOVVVRVRVFUbw+cFJVFZ1Ox6FDhzAajQ6BkzN4b2L3GmIymUhOTvb6CXfiQRyAOLAiHsSBuxH/4sCKeLhwBw1h3SNVVcnLy6s0l4A3YPXgKjz/yhAEQRAEQRAEQahlJHASBEEQBA+mxGRmzd6TbDjhvcNxBEEQ6gIJnARBEATBgykxWbj18018tVdPUan3ph0WBMG1NG/enFmzZrm7GvUKyarnJNbc/gaDwesn23m7B3EgDqyIh+o7kKx65XMhXlRVpePTv1JQamb5jH60igqppVrWb+R1qCEeau6guLiYAwcO0KJFC/z9/Wuxhq6lqjY+/fTTzJw50+njnjhxgqCgII9OlGFNDlFSUsLBgwfLPbfOvP9Kj1MNKCoqcncV6gXiQRyAOLAiHsSBu1AUhZgw7YvAsbxiN9fGvcg1qCEevMtBZmam7TZr1ixCQ0M5evQohw8f5ujRozzwwAO2ba1BZXWIjIz06KDJiiv7iCRwchKTycTKlSu9OlsNiAcQByAOrIgHceBuYkK1wOnIKc9etPNCkGtQQzy41oGqqhSWmtxyq+4X/piYGNstLCwMRVGIjo4mKCiInTt3EhISws8//0y3bt3w8/NjzZo17Nu3j2uuuYbo6GiCg4Pp0aMHy5Ytszvu+UP1FEXh448/5tprryUwMJA2bdrwww8/XLDj2sS6ppergidZx0kQBEEQPJyYMD9AepwEwdUUGc10eOpXtzz3jmeHEejrmq/qjzzyCK+99hotW7akUaNGZGRkMGLECJ5//nn8/Pz44osvGDVqFGlpaSQkJFR4nGeeeYZXXnmFV199lf/973/cdNNNHDp0iIiICJfUs74jPU6CIAiC4OFYe5wkcBIEoTyeffZZhgwZQqtWrYiIiKBLly783//9Hx07dqRNmzb85z//oVWrVlX2IE2aNIkbb7yR1q1b88ILL3DmzBnWr19fR61wP9LjVAMMBtEG4gHEAYgDK+JBHLgTW+CUW+LmmrgXuQY1xIPrHAT46Nnx7DCXHKsmz+0qunfvbvf/mTNnmDlzJkuWLCEzMxOTyURRURHp6emVHqdz5862+0FBQYSGhnL8+HGX1bO+I68sJ/Hx8WHkyJHurobbEQ/iAMSBFfEgDtxNs4ggAI7leW/gJNeghnhwrQNFUVw2XK4u0el0hIeHo9Npg8uCgoLsHn/ggQdISUnhtddeo3Xr1gQEBDB27FhKS0srPa6Pj4/d/4qiYLFYXFt5F2L1UFzsmt54GarnJBaLhePHj9fri6QuEA/iAMSBFfEgDtxNVIgvAMdyvSeT2PnINaghHsQBaEkRjEZjhUkR/vjjDyZNmsS1115Lp06diImJ4eDBg3VbyTqgKg/OIoGTk5jNZtatW4fZ7N2LDIoHcQDiwIp4EAfuJipYC5xOFRopNnrnOZBrUEM8iAPQAoaCgoIKA4Y2bdqwcOFCUlNT2bJlCxMmTGiQgWZVHpxFAidBEARB8HDCAgz46rQvBlmSIEIQhCp4/fXXadSoEX369GHUqFEMGzaMSy65xN3Vqvd43qBNQRAEQaglXnzxRRYuXMiuXbsICAigT58+vPzyy7Rt27bS/ebPn8+TTz7JwYMHadOmDS+//DIjRoyoo1pr8wzCfeF4MWTmFpPYOKjqnQRBaHBMmjSJSZMm2XqPBgwYUG5vS/PmzVmxYoVd2dSpU+3+P3/oXnnHycnJubAKexjS4+QkiqIQEhKCoijuropbEQ/iAMSBFfHQcBysWrWKqVOn8ueff5KSkoLRaGTo0KEUFBRUuM/atWu58cYbmTx5Mn///TejR49m9OjRbN++vc7qrSgKjQO1DFzHcr2zx6mhXIMXingQB6A50Ol0Xu0AXO9BUV016M9DyMvLIywsjNzcXEJDQ91dHUEQBK/C096DT5w4QVRUFKtWraJfv37lbjNu3DgKCgr46aefbGWXXnopSUlJvP/++9V6Hld4mfFtKgs3H+GhK9ty94DWNTqGIHgzxcXFHDhwgBYtWuDv7+/u6ggupLJz68z7rwzVcxKLxUJGRgbx8fG2FI/eiHgQByAOrIiHhusgNzcXgIiIiAq3WbduHTNmzLArGzZsGIsXL65wn5KSEkpKzqUOz8vLA8BoNGI0GgEtja5er8dsNttN2raWm0wm29AZi8VCgKod7+jpQtsxAPR6PTqdzq4Mzq1zYzKZqlXu4+ODxWKxm3CvKAoGg6HC8orqXp02la17ReVl22SxWMjMzCQ+Pt5hkruntqmyuldUrtfrSU9PJzY21vZa9PQ2OXuedDodhw4dIi4uzuagOm2yZl+zWCyoqoqiKKiqatcmRVHKTcFt7dE4vz+ionKdTlfhsV1RDlBaWmqXPryq49T3NtWk3GKxYDQabefUeit7jZ1/fVaGBE5OYjabSU1NtXsxeiPiQRyAOLAiHhqmA4vFwn333Uffvn3p2LFjhdsdO3aM6Ohou7Lo6GiOHTtW4T4vvvgizzzzjEP50qVLCQwMBCAhIYGuXbuydetWu0Up27ZtS7t27Vi/fj0nTpywlZfmKICef/YfITn5gK28d+/eREVFsXTpUrsvCwMHDiQgIIDk5GS7OowYMYKioiJWrlxpKzMYDIwcOZLs7GzWrVtnKw8JCWHQoEFkZGSQmppqK4+MjKRPnz7s2bOHtLQ0W7mzbUpKSiIxMZHVq1eTn59fZZsAoqKiSElJaTBtcvY8DR06lC1btrBly5YG0yZnz1OPHj3YunUrW7dudapNp0+fJiYmhjNnzqDX6/Hz8yM/P98uoAgKCsLHx8f2Y0fZ+iiK4lAeGhqKqqp2XgDCw8MxmUx2w4B1Oh2hoaGUlpZSVHRuaQGDwUBwcDDFxcV2P7j4+voSGBhIUVGR3fpLfn5++Pn5UVRUZHecgIAAj25TQEAABQUFdtdSddtkff6CggIMBoPdtVdYWEh1kaF6TmI0GklOTmbEiBEOi4B5E+JBHIA4sCIequ/Ak4bq3XXXXfz888+sWbOGZs2aVbidr68vn3/+OTfeeKOt7N133+WZZ54hKyur3H3K63GKj48nOzvb5sWZX/2NRiNvzFvGR2l6OsaFsuiuS23be0uPk9FoJCUlheHDhzvMZ/DUNlVW94rKAZKTkxkyZIjttejpbXL2PKmq6uCgOm0qKioiIyOD5s2bExAQ4NG9M6qqkpeXZzfXyxt7nMxmM/n5+fj4+HDo0CHbUL2y11heXh5NmjSRoXqCIAiCUBOmTZvGTz/9xOrVqysNmgBiYmIcAqSsrCxiYmIq3Mf6i/D5+Pj4OASeer0evV7vsK31S6aVcD/tS8OxvJJyg9eKAlpnynU6Xbk9ihWVV1T36rapqvLy6qgoSoNrkzPl1iClvGvJU9sEzp2nyhxU1iYfHx+HZALWL+DlPW95VJSEoLzyio7tinJrEGFtT3WOU9/bVNPyso9bb2WvC2d+8GwY4ynqEEVRiIyMlCwl4kEcIA6siIeG40BVVaZNm8aiRYtYsWIFLVq0qHKf3r17s3z5cruylJQUevfuXVvVdEBRFFrFNgYg+0wJpaaGt5BlVTSUa/BCEQ/iAM71yHmzA3C9BxmqJwiCINQZ9f09+O6772bu3Ll8//33dms3hYWFERAQAMDEiRNp2rQpL774IqClI+/fvz8vvfQSI0eO5JtvvuGFF15g8+bNlc6NKosrvKiqStsnf6HUZOH3hwYSHxFYo+MIgrciWfUaLq7Kqic9Tk5iNpvZtWuX3Xhab0Q8iAMQB1bEQ8Nx8N5775Gbm8uAAQOIjY213ebNm2fbJj09nczMTNv/ffr0Ye7cuXz44Yd06dKFBQsWsHjx4moHTa7AbDaTlpZGTKj2hSDTC9dyaijX4IUiHsQBaD+kFBUVlbtorTfhag8SODmJxWIhLS3NYQKdtyEexAGIAyvioeE4KJuutuxt0qRJtm1+++03Zs+ebbff9ddfT1paGiUlJWzfvp0RI0bUab2t/mNCtXlTmblFVezR8Ggo1+CFIh7EAWjvZSUlJdUOGAYMGMB9991n+7958+bMmjWr0n0URal02YXq4qrjlIezHqpCAidBEARBaCBYe5yOeWGPkyB4K6NGjeLKK68s97Hff/8dRVHsUrNXhw0bNjBlyhRXVM/GzJkzSUpKcijPzMxk+PDhLn2u2kICJ0EQBEFoIMSEWXucJHASBG9h8uTJpKSkcPjwYYfHZs+eTffu3encubNTx4yMjLStKVfbxMTElJtltD4igZOT6HQ6EhISGswCjzVFPIgDEAdWxIM4cDdW/3HhWgILb+xxkmtQQzy42IGqQmmBe27VHF521VVXERkZaTeEWFEUSktLWbBgAaNHj+bGG2+kadOmBAYG0qlTJ77++utKj3n+UL09e/bQr18//P396dChg8Mi0wAPP/wwF110EYGBgbRs2ZInn3zSlhp+9uzZPPPMM2zZssWWEtxa3/OH6m3bto1BgwYREBBA48aNmTJlCmfOnLE9PmnSJEaPHs1rr71GbGwsjRs3ZurUqQ7rhVmP7evr67KserKOk5Po9Xq6du3q7mq4HfEgDkAcWBEP4sDdWP2f+OcYAJl53hc4yTWoIR5c7MBYCC/EueZYzvLYUfANqnIzg8HAxIkTmT17No8//rgtMFmyZAlms5l//etfzJ8/n4cffpjQ0FCWLFnCzTffTKtWrejZs2eVx7dYLIwZM4bo6Gj++usvcnNz7eZDWQkJCWH27NnExcWxbds27rjjDkJCQnjooYcYN24c27dv55dffmHZsmWAlq30fAoKChg2bBi9e/dmw4YNHD9+nNtvv51p06bZBYYrV64kNjaWlStXsnfvXsaNG0dSUhJ33HGH3fEURSEwMJDiYte8J3rvzxE1xGw28/fff3t1phYQDyAOQBxYEQ/iwN1Y/UeF+AJwzAuTQ8g1qCEevNPBbbfdxr59+1i1ahWgJUX45JNPuO6660hMTOSBBx4gKSmJli1bcs8993DllVfy7bffVuvYy5YtY9euXXzxxRd06dKFfv368cILLzhs98QTT9CnTx+aN2/OqFGjeOCBB2zPERAQQHBwMAaDgZiYGGJiYmxLPJRl7ty5FBcX88UXX9CxY0cGDRrE22+/zZdffmm30HijRo14++23adeuHVdddRUjR450WE/P6qGwsNBlySGkx8lJLBYL6enpdOzYsdyVp70F8SAOQBxYEQ/iwN1Y/fds3gaA4/klGM0WfPTe8/uoXIMa4sHFDnwCtZ4fd+BT/TlG7dq1o0+fPnz66acMGDCAPXv28Mcff/Dss89iNpt54YUX+Pbbbzly5AilpaWUlJRUew7Tzp07iY+PJy7uXM9beQt8z5s3j7feeot9+/Zx5swZTCaT0+vS7dy5ky5duhAUdK6nrW/fvrZMidHR0QBcfPHFduc2NjaWbdu2ORxPVVVKS0vx9fV1qh4V4T3vqIIgCILQwIkI9MVHr6CqWvAkCMIFoijacDl33JyclzN58mS+++478vPzmT17Ni1atKB///68+uqrvPnmmzz88MOsXLmS1NRUhg0bRmlpqcs0rVu3jptuuokRI0bw008/8ffff/P444+79DnK4uPjY/e/oih1kn5eAidBEARBaCDodArRtpTk3jdcTxC8mRtuuAGdTsfcuXP58ssvuemmm1AUhT/++INrrrmGf/3rX3Tp0oWWLVuye/fuah+3ffv2ZGRk2C38/eeff9pts3btWhITE3n88cfp3r07bdq04dChQ3bb+Pr6Vjl8sn379mzZsoWCggJb2R9//IFOp6Nt27bVrnNtIYGTk1hPnDdnqwHxAOIAxIEV8SAO3E1Z/7FhWuDkbSnJ5RrUEA/e6yA4OJhx48bx6KOPkpmZyaRJk1AUhTZt2pCSksLatWvZuXMn//d//2c3X6gqBg8ezEUXXcQtt9zCli1b+P3333n88cfttmnTpg3p6el888037Nu3j7feeotFixbZbdO8eXMOHDhAamoq2dnZlJQ49orfdNNN+Pv7c8stt7B9+3ZWrlzJPffcw80332wbpucMiqLg5+fnsqx63nVFuQC9Xk+7du28dtywFfEgDkAcWBEP4sDdlPUfG+adKcnlGtQQD97tYPLkyZw+fZphw4bRqlUrFEXhiSee4JJLLmHYsGEMGDCAmJgYRo8eXe1j6nQ6Fi1aRFFRET179uT222/n+eeft9vm6quv5v7772fatGkkJSWxdu1annzySbttrrvuOq688koGDhxIZGRkuSnRAwMD+fXXXzl16hQ9evRg7NixXHHFFbz99ts18qEoCgEBAS4LnBTVVWkmPIS8vDzCwsLIzc11esIagMlkYv369fTs2RODwXtza4gHcQDiwIp4qL6DC30Pbqi48rPp1aV7+GD1fiZf1oInr+pQC7Wtn8jrUEM81NxBcXExBw4coEWLFvj7+9diDWsfVVUpKCggKCjIZUGDJ2L1oNfrOXjwYLnn1pn3X+lxchJVVTlx4oTL0hp6KuJBHIA4sCIexIG7Kes/Jsw6x8m7epzkGtQQD+IANAcmk8mrHYDrPUjgJAiCIAgNCOscp6OSHEIQBMGlSOAkCIIg8M4779C8eXP8/f3p1asX69evr3DbAQMG2Famt958fX35z3/+Y9smKyuLSZMmERcXR2BgIFdeeSV79uypi6Z4PTFeOsdJEAShtpHAyUn0ej1JSUleOeGwLOJBHIA4sOLpHubNm8eMGTN4+umn2bx5M126dGHYsGEcP3683O0XLlxIZmam7bZ9+3b0ej3jx49Hr9ejqiqjR49m//79fP/99/z9998kJiYyePBguxSzgusoew1ae5yO55dgMtf+uib1BU9/HboK8SAOwPVJETwVV3uQwMlJdDodiYmJXpfi8nzEgzgAcWDF0z28/vrr3HHHHdx666106NCB999/n8DAQD799NNyt4+IiCAmJsZ2S0lJITAwkLvuugudTseePXv4888/ee+99+jRowdt27blvffeo6ioiAULFtRx67yDstdgk2A/DDoFs0Ul+0ztLD5ZH/H016GrEA8X7qAhzAtydRpuT8XqwVV476uqhphMJlasWIHJZHJ3VdyKeBAHIA6seLKH0tJSNm3axODBg21lOp2OwYMHs27dumod45NPPuGGG27gr7/+wmQy2dbmKJu5SKfT4efn57BoouAayl6D+jKL4GZ60TwnT34duhLxUHMHPj4+ABQWFtZGteoUVVXJy8trEEHghWD1UFqq/Yh0ob2Q3pmn8gJQVZX8/Hy5EMWDOEAcWPFkD9nZ2ZjNZoeFBaOjo9m1a1eV+69fv57t27fzwQcf2LJYtWvXjoSEBB599FE++OADgoKCeOONNzh8+DDHjh2rraZ4NedfgzFh/hzJKfKqeU6e/Dp0JeKh5g70ej3h4eG2YcqBgYEe22NjsVgoLi7GYDB4de+jxWKhqKiIvLw8AgMDLzhFvwROgiAIQo355JNP6NSpEz169CA5ORnQfrVduHAhkydPJiIiAr1ez+DBgxk+fDhGo9HNNfYOrCnJM70ocBIEVxATEwNQ4RxPT0FVVYqKirx+npPVQ1BQELGxsRfsQgInQRAEL6ZJkybo9XqysrLsyrOysmxfICqioKCAb775hmeffdbhsW7dupGamkpubi6lpaVERkbSq1cvOnfu7NL6C+UT64VD9QTBFSiKQmxsLFFRUR79Q4/RaGT16tX069fPNgTRGzEajaxatYohQ4bg6+t7wceTwMlJ9Ho9vXv39upMLSAeQByAOLDiyR58fX3p1q0by5cvZ/To0YA2tGH58uVMmzat0n3nz59PSUkJ//rXvyp0EBYWBsCePXvYuHEjjzzyCB9//HGttMWbOd+/N/Y4efLr0JWIB9c40Ov1Hu3Q19eXHj16EBQU5NVD9Xx9fenZs6fLgkcJnJxEp9MRFRXl7mq4HfEgDkAcWPF0DzNmzOCWW26he/fu9OzZk1mzZlFQUMCtt94KwMSJE2natCkvvvii3X6ffPIJo0ePpnHjxgB2DubPn09kZCQJCQls27aNe++9l9GjR3PFFVfUXcO8iPOvwVgvXMvJ01+HrkI8iAMQB1Zc7cHtIagziy4CzJo1i7Zt2xIQEEB8fDz3338/xcV198FgNBpZsmSJR3ffugLxIA5AHFjxdA/jxo3jtdde46mnniIpKYnU1FR++eUXW8KI9PR0MjMz7fZJS0tjzZo1TJ48GXB0kJmZyc0330y7du2YPn06N998M19//XXdNsyLON9/bLj39Th5+uvQVYgHcQDiwIqrPbi1x8m66OL7779Pr169mDVrFsOGDSMtLa3c6HDu3Lk88sgjfPrpp/Tp04fdu3czadIkFEXh9ddfr7N6e3OKz7KIB3EA4sCKp3uYNm1ahUPzfvvtN4eytm3bOmSsKutg+vTpTJ8+3WG/uvyhy9so69+6CG5WXjEWi4pO5x2Twz39degqxIM4AHFgxZUe3Nrj5Oyii2vXrqVv375MmDCB5s2bM3ToUG688cYqe6kEQRAEwZuIDPZDp4DJopJdUOLu6giCIDQI3NbjZF108dFHH7WVVbXoYp8+ffjqq69Yv349PXv2ZP/+/SQnJ3PzzTdX+DwlJSW2xRgB8vLyAK3rztptp9Pp0Ov1mM1mLBaLXX30ej0mk8n2y2rZrr6y5aBNJNTpdA7dgdac8edHvBWV+/j4YLFYMJvNtjJFUTAYDBWWV1T36rSpbN2r26byfHh6m5w9T1bKHsfT2+TseSrPgae3qabnqayHhtKmmpwn69+K6u7Na8vUJQa9jqgQf47lFXMst5ioEP+qdxIEQRAqxW2BU00WXZwwYQLZ2dlcdtllqKqKyWTizjvv5LHHHqvweV588UWeeeYZh/KlS5cSGBgIQEJCAl27dmXr1q2kp6fbtmnbti3t2rVj/fr1nDhxwq7cYDCwcuVK8vPzbeW9e/cmKiqKpUuX2n1ZGDhwIAEBAbY1TqyMGDGCoqIiVq5caSszGAyMHDmS7OxsuwAyJCSEQYMGkZGRQWpqqq08MjKSPn36sGfPHtLS0mzlzrYpKSmJxMREVq9eXa02DRgwgMsvv5yUlJQG0yZnz9OIESPo3LmznQNPb5Oz56l37940b97czoGnt6km5+nvv/8GsHmo723SWUppH3CS1uGwXd+Jg0ezHdrk7Hmy1tHqoKI2XXbZZQiux2AwMHDgQLvFHWPCtMDpaE4xnZu5sXJ1RHkOvBHxIA5AHFhxtQdFddPPf0ePHqVp06asXbuW3r1728ofeughVq1axV9//eWwz2+//cb48eN57rnn6NWrF3v37uXee+/ljjvu4Mknnyz3ecrrcYqPjyc7O5vQ0FDAuV9eVVVFVVV8fX0xm81e92ty2XJrvcv2PHhym5w9TwaDgdLSUhRFsTnw9DY5e570ej0lJSXodDqbA09vk9PnyVSE8cReTPoA9I3iUfQ+lbep9AymrJ2Qm4Ea2xXCE6rXprwj6Fc+B6Vn0MX3wBLXDXN0Z/ALqbpNej3mrB2wdznK/hUo6etQTNpcI8uld2O+4tw6TJWep+IcTJnbUeMvhbPn29rW0tJSTCYTBoPBVpfy2lRYWEh4eDi5ubm29+D6xurVq3n11VfZtGkTmZmZLFq0yJaqvSLmzJnDK6+8wp49ewgLC2P48OG8+uqrtoyDVZGXl0dYWFiNvVh/TLT6B7jrq038vP0YM0d1YFLfFk4f09Moz4E3Ih7EAYgDK9Xx4Mz7r9vC0Josuvjkk09y8803c/vttwPQqVMnCgoKmDJlCo8//ni5eer9/Pzw8/NzKPfx8XHI6V5Rzv6yUarRaOTnn39mxIgRFeaEd0W5Tqcrtz0VlVdU9+q0qTrl59exKg+e2CZny41GI7/88ku5Djy1TeDceTIajfz666/lOvDUNkEFdS88hf7oZvTZe+HkHsjeAyf3Qt4RfAAfQFX0KGFNITxRC4jCE8E3EE7u07bN3g1nss698Rr8Yehz0ON2UJSK23RoFXx3BxSd0gp2/4wO0KFAVHto1h2iO4LFjK60AF3pGTAWQmkBlOTDkc3o8w7bHzQgAopOodv+Hbqhz4HOvr3lno+t32L45WFodxWMn2P3kKIopKSkOFwL57fJEz7ACwoK6NKlC7fddhtjxoypcvs//viDiRMn8sYbbzBq1CiOHDnCnXfeyR133MHChQvroMZagJqcnGzn37aWU553JOQoz4E3Ih7EAYgDK6724LbAqSaLLhYWFjp8ybF+uZFx84IguBRVhRNpsPsX2P0rZPwJqqX8Tf3DsZScQa+aICddu1VGcDT4hWoBWPID2vGveQdC7IcuYzHDqpdh1SuACrFJ0GksHNkEhzdBbjoc36HdqkLvB837QqsroPVgiGgJ/70IzmTBgdXQamDVPv7+UrvfckDVz+fBDB8+nOHDh1d7+3Xr1tG8eXNbFsEWLVrwf//3f7z88su1VcVqEeeFazkJgiDUJm4d+OjsooujRo3i9ddfp2vXrrahek8++SSjRo3y6NWdBUGoJ6iqFkTsWqIFTDmH7B+PaKX18DRuDU3aQJOLoHFrTD4hJC/5iRGXd8On4OjZ4OkQnD6k9fpEtNS2b9wGmrQG/zCwWGDDR5DyFOxNgfd6w6i3oP1V2nMVZMN3k2H/b9r/3W+DYS+CT5lJ/vlZcGQjHN6o9WYZ/ME36NzNJ1D7G9EKEvtovV9l6TAaNn0G2+ZXGThZjqaiy9pOKT785d+fyy9IdMOid+/ePPbYYyQnJzN8+HCOHz/OggULGDFiRIX7uDRxkbEQy9YFtDm2DhhhK48M1n5dzcwpsh27LPVu2OsFDuW13ldVVRIXIYmLznfg6W1y9jyV58DT21RbiYucWePJrYHTuHHjOHHiBE899RTHjh0jKSnJYdHFsif/iSeeQFEUnnjiCY4cOUJkZCSjRo3i+eefd1cTBEFoKKgqLH0C1r19rkzvCy36wUVXQpuh0Cix/H2NRlB0EBoLjRMg4dKqn0+ng17/px1/4R1wbBvMuwm63gwdx8DiqZB/VAt+rpoFXcY5HiMkGtqN1G41ofMNWuC04wcY+V/wCXDYpLDUxHebDhO84lWuBX41d2P+xhwu71yzp2yI9O3blzlz5jBu3DiKi4sxmUyMGjWKd955p8J9XJm4SG8u5qqtM+gAGEteYfX6VPLz8zmQB2DgyOlC27Hre6KVC00eA9oXIm9OXDR06FAAr05c1KNHDwcHnt4mZ89Tq1atHBx4eptqK3FRYWEh1cVtySHcRW1MwPVGxIM4ADc7OHMC/lkI2xZovToJvSCxr9azEhrn3LFUFVb8B37/r/Z/lwlaz0+L/uAXXI3dL9CDqQRWPg9/vAWUeUtuchHc8IXWy1UbWCwwqxPkHYbrP4eLR9seyswt4vO1h/h6fTrFRQX85TeVcKWAr9u+Sb/h42gabh9kVdfBhb4H1zWKolSZHGLHjh0MHjyY+++/n2HDhpGZmcmDDz5Ijx49+OSTT8rdx9WJiwyvtUApyUeduh5zo1aoqsqRnCIG/Pd3fPUKac8NdyrxTXnl9fnXZDg3ZN/6vA2hTZXVXRIXSeKiytqk0+koLi5Gr9fbHHh6m2pynqqTuCgvL48mTZrU7+QQnkxRUREhISHurobbEQ/iACpwUHIGVjwHp/ZpWd98g7W/fqHa34BG0KwHNG5ly85WLUoLIS0Zts6DvctBLfPl6Pg/sPHs4tmNmp8NovpCh6ttmecqZNUr54Km4a9CrynVr9NZLuhaMPjBkGe1Xq1Fd0JuBnS8Thu6V43ArSwms4Xj+SVk5haTlVdMZm4xBSUmzBYtI6hZVTFbwKKqmMwq/f360585pP78Me9sakax0UxRqZnUjBxMFu1DaFLYdsJLCrCENuXGcTc7JJJwiQMP5sUXX6Rv3748+OCDAHTu3JmgoCAuv/xynnvuOWJjYx32cXXiIjUkDkrSID8TQ2RbAOIa6VEUKDWrnCwopUmw4/NZn7O65fU5yY+qquTn5xMSEtJg2lSTclVVKSkpISQkxOFHDE9tEzh3nlRVpbS0tFwHntomcO48WYes+vv7Ozjw1DZBzc5TcXExAQEBdh4qS2JUGRI4OYnJZGLlypWSpUQ8NDwHplLIO6IFHdUMZsp1UHAS5l6vJTCoipBYaH4ZNL9c+xvR8txzF+Vo84SsyRYyt8Kun6D0zLn94y6BzuMgrCkcWgeH/oBjW+H0Qe2WOgeWzYTBT2u9SOW8afP76/DbC9r9oc/XKGhyxbWgqiqFsZdy+saVlBzfTW5Ye0qPlFBqKqLUZKHUbKHUZOFMiYm8YiN5Rda/RvKKTeQUlnIst5jsMyVYnBhHsFbpRH8/aJ+/jr+y95HHuUDt0pYRTL6sJYM3fQL7QJc0ocKgqcG9HpygsLDQ4UO7rhMXqSExKNlpmHOO2D7YfQ06mgT7cSK/hGO5xRUGTg0Fb74GyyIexAGIAyuu9iCBkyA0NE4fgm3fQuFpGPhY9XosTKXw5bVwaA00aavNfel0fcVzeioi97B2nOzdWrrrgY+BxaT1QJXkaamxS/Ih76iW1CA/U0tMsG2+tn9oU22/nHQoyS3/OcITtWCp8w1awgUr7Udpf4tzIWO9FkT9s0gLoL6fCus/guEv288/WvcOLD87z+SKp6FP+Rk9XUWx0cw/R/PYkpHDtiO5HMst5nRhqXYrMFJqLpu1b12Fx6kKH71CVIg/sWH+RIf5E+rvg04BvU5BpyjodedufoY2nPy7NY0L9vJRj0yOtBiLv4+eVpHBtI0J0c7ptyu0AydNuDABHsKZM2fYu3ev7f8DBw6QmppKREQECQkJPProoxw5coQvvvgC0BIX3XHHHbz33nu2oXr33XcfPXv2JC7OyWGjNSVYW8ZDyc+0K44N8+fE2R7Ijk3D6qYugiAIDRQJnAShIVCSDzu+hy3fwMHfz5Wf3Avj54K+ipf6zw9pQRNAdpo232fFfyChj5aUoMM12vC6yjixWwua8g5rAdDNi+DskKFyMRbB4Q1wcA0c+F27n3dEu1kJirStiUSj5lqShvielfeI+YdBmyHabcCj8NcH2lC8zFT4dBh0HAtDnoG0n+HXx7R9BjwKl8+ovH3noaoqaVn5rEo7wdp92Zw4rmNl4TYaBfvRKNCX8EAfwgN9MZosbDmcQ2pGDjsz8zCaK++B8DPoCAvwwc9Hh69eh69Bj69ewdegw9egI8jXQGiAD6H+PoQGGAj19yEsQLtFh/oTE+ZP4yBfdDonhkD6TYDlz9Irfxlccp/9Y6lfA6rWKxjR0ilHnsrGjRsZOPBclsEZM7Rr45ZbbmH27NlkZmbaTVqeNGkS+fn5vP322/z73/8mPDycQYMG1Wk6cjXk7HDA/GN25bFh/mw9nMux3KI6q4sgCEJDRQKnGlDROEpvQzy42YE1dfbfX8HOH8FU5otR4mVaj86eX+GXh2HEaxUHGxs/1TKrocDYT7SAZus8LZhJX6vdkh/U1v7pcI0WvASE23Y3GAwoRzfDN+O1BVobt9GCpvD4yuvvE6BllGvRDwaizV86vEFLlNAoEcLiHdNnO4vBD/pOhy7jtUBw85ewfYE25M90dm2by2ZA/4cBOJ5fzOrd2eh10CjQl0aBvkQE+dIoyJcgXz15RSbW7M1m1e7jrNp9gqy8kjJPpmNHTqZjHc6jSbAvSfHhdGkWTkLjQO34gdpzRAT6EuDrhqUVOo6F5c9qQWze0XPJNSwWSP1Ku590U5WHaSjvCQMGDKh0iN3s2bMdyu655x7uueeeWqxVFVTY46Ql8cj0krWcGso1eKGIB3EA4sCKKz1IVj1B8ESKc+Gn+2H7d+fKIlpB0o3aMLbwBK0H6ttbAFWbu1PeMLRDa+HzUdpwuiuetu91yT2iBRlb5mmJF6zofKBlf2h/tZYG+9g2+OYmMBZoc45uWgBBjWut6edTUGKisNRMo0AfDPpy5jCV5Wgq/PKoFgwC9J7Gmf4z+fWfLBanHuGPvdkVzg/y0SuYLard4/4+Onq3bMzlbSIJ9jNwurCUnCIjOYVGcgpLySk0YlFVOjUNo0t8OEnx4TRrFFBp1jm38ckwbZHfoc9Bn7MBwME1MHsk+IbAA2namlAXiLwHl88Fe9n5I8z7l5Z05fZltuKPVu/n+eSdjOgUw7s3dXNhjQVBEBoGzrz/SijqJBaLhezsbJo0aVJudhBvQTzU0MGp/WAshugONX/iwxthwW1a4gRFD5dM1HoDmnW371XqcI32JXjp49r6ROHxWpmV3MPw7UQtaLp4DFx2v/3zhDWFvvdqt6x/tEBsxw9wYifsXQZ7l6H+dB8oOhSLSUvdPX5O1RnsLoBio5kdmXlszchh65Fcth3OZe+JM1h//gkP9CEiyJfGQb40DvIjItiXRoE+Z4fO+dIoMJbwQXNoevw3srNP8H52d5Y9v4xi47m5RZ2bhRHib+B0gZHThaWcKiilxGSxDbFrExVM/4si6d82kh7NI/D30Ze5Fpp67uuh8/Va4LRt/rnA6e+zvU0dr60yaJL3BPdiCY5BB6h5RykblreN0V6Pu47ll7tfQ0KuQQ3xIA5AHFhxtQcJnJzEbDazbt06RowY4dUXonhw0sGZ49o6PZu/0IbYXfGUFqg40/NgscDaN7U03xYThCVoQ+vie1a8T++pWnKEDR/BwikQEgfxPbRhcd9MgIITENMJrnm78rpEX6zdBj4G2Xu0IGrnDyiZW0C1YGl3NbqxH2tD41yI2aKSmnGalB3HWb37BGlZ+ZjL6RJSFE2r1tNjZP+JgiqO7Ac0A7T5IC2bBHFNUlOuSYqjeRPHAKGo1MzpwlIMOoWoUH/HejaE10OHa+HnhyFzizZfLSRGO8+gLcpbBQ3CgQdjDoxCB3AmS3uvOHsO2sVqgdPB7AKKjWb8fdwwFLSOkGtQQzyIAxAHVlztQQInQagJ2XvRL5zCoJNH0QVthaTx9hnerBiLYN3bsGaWfRrt5c9oX1Cvead6We/ys2DRFNj/m/b/xdfCVbPs5hqVi6LAlS9p6wLt/gW+Hg+3p8DKF7TnD2ysJY9wZghWkzbQ7wHo9wDGE3vZ8MvX9Lj2QXTVCJpUVeXQyUJ2Z+UT4q/1DmlzfM4NsysoMfH7nmyW7cxixa7jnCootX/6YF86Nws/O/wtjI5Nw2gc5GfrHTp55uzfghJOniklp7CU04Va71FOmb/BfgZGdIpldNc4OjUNq3T4XICvngDfgAofbxAENYZWV2jz4rbNh7BmYCzUFuFt1sPdtROqIjgKFUXr/S3MhuAoACKD/YgI8uVUQSl7ss7QqZlk1hMEQagpEjgJ3oPZBAXHz018rynbv4MfpqMrPUMIwJrXtFtsFy2F98VjtPWJtn2rTbi3ZomLuwSGPQ8n0rRkCzsWa2m7x8+pOFuZqUSbu/DLI1rvkCEARryi9QBUt7dKb4DrPoHZI7Rg6cMB2hwpnQFu+EKbD1VTwhM5Edq5wrV9LBaV3cfzWX/gFH8dOMWGA6c4nl/isJ2iQFiADxGBvhzO0dYtshLqb2BguygGtYuiR/MIYsMcF/MDaBLsp61TE13z5ng9na4/Gzh9q2U0BG0YaH2ckyXYozNQYgjF35SrJfg4GzgpikK7mBDW7jvJzmN5EjgJgiBcABI4OYmiKOWuRO1teJQHVdWCj+XPaOm5r3oDut/m/HFMJVr66g0fa4dN6MtO/0toZ0lDt3+FFpRkboGlT2rpuPMOa/uFxWuJFzpepw2fSewDUR3g25vh+A74cKA25K714HP1PbwRtsyF7QuhOEcrj+4IYz+tPMV3RfgFw4Rv4aMrztXrype0RWerQW6RkVMFpRjPLsJaarZgNFkoKjWytyiQhX8fJb/EbOvNySnUtt+RmUdukdHuWL56HW2igyk2mjlVoCVTKDvMDiAhIpDB7aMZ3EELlnyqSvrgZjzq9VAZ7UaAT+C5BYQVvZaRsBo0GAceiqIolPo3wf9MrkNK8rZnA6e0Bj7PSa5BDfEgDkAcWHG1B8mqJzRsDq2FlKe0NNdWfALhrj+cW5Pm1AGYP0lbCwjg8ge0tX+s6yMVnNR6kLYtOJexzTdEy1J36V1a6u3zycvUgqfDG0DRwYDHQEFbi+nkucU3CW2q9TBddj/4OM6vcYqsHbD4Li1IG/REpT0JJSYzK3YeZ/6mw6zafaLcuUXVIdBXT7fERvRsHkGPFhEkxYfbzbMwmS3kFBk5XVDKyYJSmgT70ioy2Ovf7N3Gd7efW5D4ouEw4RuXHl7eg8vHJV7mjofdPzv8OPTthgwe+m4rfVs3Zs7tl1ZyAEEQBO9DsurVIhaLhYyMDOLj4716sl2995C1Q+th2v2L9r9PIPSeBunrtAViv58Gt/xkm0BdKTt/hMVToSQXAiJgzEfQZrDm4NAhzUFQY+gxWbvlZMDRvyGhNwRHVnzc0FiYtASSH9CSRqx87txjPoFauu+kG7WFRysYCuc00R3g/1ZV+LCqqmw/kseCTRl8v+WorQcIIMTPgK9Bh49eh49BwUevLdCKxURUWBCNgnwJD9AWfdUWf/WhRZNgLo4LrbTHyKDX2YbZlTNLzCOo968HZ+h0w7nAqWvVazdZaVAOPBCLxUKBEqwNHz6vx8maIGJXZsPucZJrUEM8iAMQB1Zc7UECJycxm82kpqYSFxfn1RdivfVQnAu/Pg6pc0C1aEONut0C/R+BkGht+NG7feDQH1qmuV7/V/nxVr+qZbEDiO+lDZULawZU4iA8vurFX60Y/GDUWxCbpM2HiukESROg/ahaSetdVGpmd1Y+2WdKyC0ykltkJK/IpP0tNrL9SK5d2uKYUH/GXNKU67o1o1WkYxILo9FIcnIyI0Zcjo+Pj8vr6ynU29dDTWg1EOK6akNG2wyr9m4NyoEHYjabOZxnoT1oc5zK0CYqBEWBkwWlnMgvITLEtdkv6wtyDWqIB3EA4sCKqz1I4CQ0HDI2wHe3QU669n/7q7W5RU1an9umUXMY+iws+TekPK0NWWvcqvzjlQ2aek+DwTNBXwvBgaKc661yISfPlLAjM49/jmq3HUdzOZBdUOECr1Z8DTqGXRzD2G7NuKx1E/Q6GTLnVeh9YMpv7q6FUAOKfcK1O/mZduUBvnpaNA5if3YBu47lERlSSU+4IAiCUCESOAmej8UCf8zSghzVrGWJu/ZDSOxd/vbdbtPWpzmwWhuyN2mJ45C93/97LmgaPNNxcVg3U2qykH6qkIPZBWTmFnEsr5hjuSVk5RVzLK+YrNxi8ktM5e7bJNiPuHB/wgJ8CA3w0f76a3+jQ/24on00YQHe23skCJ5KkU8j7c55Q/VASxCxP7uAtGP5XN5GAidBEISaIIGTkyiKQmRkpNdPXK83HvKPwaL/K7O+0RgYNQv8K0m5q9PB1W/De320RA7rP9ASOFhZM0sbNgcw6MkKg6a6cGAyW9h2JJftR3LZn13AgbO3jFOFVfYcAbRoEkSHuFA6xIZycVwoHeJCiQq5wAQTZag314GbEQ/iwN0oikJgVEvYh8NQPYB2MaH8vP0YOxvwPCe5BjXEgzgAcWDF1R4kq57guexJgUV3aos9+gTC8Feg67+qv+bMhk9gyQxtbaS7/tCG7P3xFqQ8qT0+8Ano/2Dt1b8cVFVlf3YBf+zNZs2ebNbtP0l+cfk9R0G+epo3CaJpeAAxYf5Eh/oTE+p/9r4fsWEBBPnJbyNC/ULeg8vHJV4KT8ErLbT7TxzX5lCe5Zftx7jzq01cHBfKkumXu6DGgiAIDQPJqleLmM1m9uzZQ5s2bdDrXZTpzANxiwdV1dY92rcS9i2HfSu08pqub9TdOmRvFSy+G9pfdS5oGvBYlUGTKxxYLCp7jp9hc/ppNh48zdp92WTmFtttE+JvoHtiI1pHBdOiSTAtmgTRMjKIqBA/t/+SJK8HDfEgDtyN2Wxmz6Es2hr8UUzF2jynRs1tj7c/m1lvz/EzmMwWDPV8bbSaINeghngQByAOrLjagwROTmKxWEhLS6NVq1ZefSHWmYczx2Hvcti/UhuOdybL/vGeU2DIf2q2vpGiwDVvw7u9IeNP7QZaBr4BD1e5e00c5Bcb2Zyew+ZDp9mcfprU9ByHuUi+eh3dmzeib+smXNa6CR2bhtXbBA3yetAQD+LA3VgsFtJ276ZtcAzkHNTWiSsTOMU3CiTQV09hqZmDJwtoHeX6rJ3uRq5BDfEgDkAcWHG1BwmchPqJqQRWvwZr3gDLubWEMARA877QciC0GeJ8L9P5hCfA0Ofgp/u0//s9CAMeubBjnkfGqUKW7cxi+c7j/HXgJEaz/ejYAB89XeLDuCShEb1bNaZ7YgQBvt77JicIQs1RQ2JQcg46ZNbT6RQuig4hNSOHXcfyG2TgJAiCUNtI4CTUDQUn4c93tXlEF18LPgEVb3toHfw4HbJ3a//HdNbShrcaqK2lZHDxGiTdJoGxSEsokTSh+nOkzsNktnCmxER+sYnM3GJ+SzvO8p3HScuyn4ydEBFIt8RGXJIQTteERrSLCWmQw2YEQXADITHa3/MCJ9CG66Vm5LArM5+rOtdxvQRBEBoAEjg5iU6nIyEhwasXEwMnPexJge+nnhtm98ujWoDS7VaIvOjcdsW5sGwmbPxU+z84Gka8qq3HVJtzeRQFet9drU0LSkxsychh46HTbDx4iv1Z/rywfRX5JSYKS83l7qPXKXRPbMTg9tFc0T6KluUsJOupyOtBQzyIA3dj9U9WnFZQTuDUNlrrZdp1LK8uq1ZnyDWoIR7EAYgDK672IFn1hNqjtBBSnoINH2n/N26jDcHLTT+3TfPLtR4fvS/8/NC5D/tLJsKQZyGgUZ1XG7SkDScLSsnMLeLgyUI2HzrNxkOn2JmZj7mKPOD+PjrCA3zp3rwRQzpE0/+iSMIDfeuo5oJQv5H34PJxmZe1/4OlT0DHsTD2E7uH/tx/kvEf/kmzRgGseXjQBdZYEAShYSBZ9WoRs9nM1q1b6dy5s1dPtqvSw9FUWDgFstO0/3vdBYOf1gKkfSu0XqXdv8DB37WblYiWMOpNaNGvTtoBkFto5NuNGWw/mktmbjGZuUVk5ZZQaraUu31cmD/dmkdwSXwYujPHSepwEeFBfoT4+xDib8DHi4bdyetBQzyIA3dj9d8lKBodlLsIbrsYrcfp8Oki8oqNhPo3rIWu5RrUEA/iAMSBFVd7kMDJSSwWC+np6XTs2NGrL8QKPVjM8MebsPIFLalDcAyMfhdaX3FumzZDtFvuYdj8JWz+XMue13c69H+48vlPLiTjVCGfrDnAtxszyh1mpygQFeJHXHgAXZqF0y2xEd0SGxEXrtXPaDSSnLyNDrHd8PFpWF9Aqou8HjTEgzhwN1b/nTrGnA2cHBfBDQ/0JSbUn2N5xew+lk/35hF1Xs/aRK5BDfEgDkAcWHG1BwmcBOc4u5aSbncKFx9ei/77H6HolLYIbcFJ7a/p7DpE7UfBqLcgsIIP57BmMPBRLZOdsUBLzlAHbD+Sywer95O8LdM27K59bChXd4mjaaMA4sL8iQ0PICrEz6t6jwRB8HxUa3KIvEzt/fq8+aHtYkM4llfMrgYYOAmCINQ2EjgJVWOxwOENsOtH2PkTnD6AHmgNcKKc7f1C4cqXqp+hTm8Afe0FTaqqsvf4GdYfPMWSrZms3XfS9tjlbZowpV9LLmvdxO2LyQqCIFwwwWcDJ1ORlnAnINzu4bYxIfyWdqLBJogQBEGoTSRwchKdTkfbtm29I0tJ5lbY+AnsSoaC4+fK9X6oLftzSh9No6at0AVHQVATCGyi9S6FxNZsQVoXYTRb2HE0j/UHTrH+4Ck2HjzF6cJza0HpdQqjOsdyR7+WXBxX84DNq66FChAHGuJBHLgbm3+/IPAPh+IcLdnOeYFT+xht4nPasXyHY3g6cg1qiAdxAOLAiqs9SODkJHq9nnbt2rm7GrVP5hb49EowFmr/+4XCRcOg3VXQ+goUvxAau7F6mw6dYv7Gw5wuLCW3yEhukYm8IiN5RUbyS0wO2/v76Oga34heLSO4vns8TcMvfB6V11wLlSAONMSDOHA3dv5D484FTlHt7bZrF3s2JXlmPqqqNqiedrkGNcSDOABxYMXVHrw7DK0BJpOJtWvXYjI5fjlvMORnwdcTtKApoTf86zt4cB9c9zFcPBr8QtzmQVVVvlh3kBs++JNvNmTw6z9Z/Ln/FDsz8ziSU2QLmsICfBjcPopHh7dj4d192Pr0ML6ecin3Db7IJUETeMm1UAXiQEM8iAN3Y+e/7Dyn82jZJBiDTiG/xMSRnKI6rmXtIteghngQByAOrLjag/Q4OYmqqpw4cYIGu/yVsRjm3QR5h6Fxa7jxG4ehHuAeDyUmM08t/od5GzMAGN4xhj6tmxAW4ENYgA+h/gbb/UaBvuh0tftLaoO/FqqBONAQD+LA3dj5D7EuguuYWc/XoKN1VDC7juWTdiyfZo0C67imtYdcgxriQRyAOLDiag8SOAnnUFX46T4tEYR/GNw4r9ygyR0czy/mzi83sTk9B50Cjwxvxx2Xt2xQw0wEQRBcQmis9rectZxASxCx61g+u47lc0X76DqsmCAIgmcjQ/WEc6x9C7Z8DYoerp8NTVq7u0YAbMnI4er//cHm9BxC/Q18dmtPpvRrJUGTIAi1wurVqxk1ahRxcXEoisLixYur3KekpITHH3+cxMRE/Pz8aN68OZ9++mntV7Y8KhmqB9DubIKIXQ0wQYQgCEJtIj1OTqLX60lKSmp4i4ml/QIpT2v3r3wJWg2qdPPa9lBYauJoTjF/7j/Jsz/toNRkoXVUMB9N7E6LJkG18pzO0mCvBScQBxrioWE5KCgooEuXLtx2222MGTOmWvvccMMNZGVl8cknn9C6dWsyMzOxWCy1XNNz2Pm3DdWrIHCyJYhoWCnJG9I1eCGIB3EA4sCKqz1I4OQkOp2OxMREd1fDtWTtgO8mAyp0uxV63lHlLq7yUGIys/SfLP7cf5LM3GKO5hSRmVtMbpHRbrvB7aN4Y1wSIf4+F/ycrqJBXgtOIg40xEPDcjB8+HCGDx9e7e1/+eUXVq1axf79+4mI0BaVbd68eS3Vrnzs/Ft7nCoKnGK0wGl/dgElJjN+hobxxaohXYMXgngQByAOrLjagwROTmIymVi9ejX9+vXDYPBwfaoKRzfD/Fuh9Aw0vxxGvFqtRWsv1MPe4/l8vT6DhZsP262xVJZgPwOxYf6M7tqUu/q3qvVkD87SoK6FGiIONMSDdzv44Ycf6N69O6+88gpffvklQUFBXH311fznP/8hIKD8LJ4lJSWUlJTY/s/L03p/jEYjRqP2nqjT6dDr9ZjNZrveK2u5yWSyTXi2Zo7q378/BEZhANQzWZhKitD7+KHT6WzHbRygJyzAQG6Rib1ZZ7goyj5BhPX8nZ+FysfHB4vFgtlstpUpioLBYKiwvKK6V6dNoP1arNPpKiy3tqmsg379+jn49tQ2VVb3isoVRWHVqlX06dPHto2nt8nZ8wQ4OPD0Njl7nlRVdXDg6W2qyXkqKipi7dq1Ng/l1f38dleGd326uQBVVcnPz/fsLCW5R2DrPNjyDWSnaWWNmsMNX4C+ej06NfFQVGpmybZMvlmfzsZDp23l0aF+XN0ljpaRwcSG+RMXHkBMmD+h9ah3qTwaxLVwgYgDDfHg3Q7279/PmjVr8Pf3Z9GiRWRnZ3P33Xdz8uRJPvvss3L3efHFF3nmmWccypcuXUpgoBbIJCQk0LVrV7Zu3Up6erptm7Zt29KuXTvWr1/PiRMn7PZXVZXfN+9kADp0qoUVP86ja78RREVFsXTpUtuXhSYGPbko/HMkh70bf7M7xogRIygqKmLlypW2MoPBwMiRI8nOzmbdunW28pCQEAYNGkRGRgapqam28sjISPr06cOePXtIS0uzlTvbpqSkJBITE1m9ejX5+efmZPXu3duhTVaMRiMpKSkNpk0DBw4kICCA5OTkarVp6NChnDlzhqVLlzaYNjl7nnr06OHgwNPb5Ox5atWqlYMDT29TTc7TihUrMJlMNg/ltamwsJDqoqhe9imXl5dHWFgYubm5hIaGOr2/0WgkOTmZESNG4ONTv7/Y21FaCDt/hC1zYf8q4OxpNwRA+6tg0BNa8FRNnPFgtqh89ech/rs0jbxi7QWp1ykMbBvFjT3j6X9RJAa95+Up8dhrwYWIAw3xUH0HF/oeXNcoisKiRYsYPXp0hdsMHTqU33//nWPHjhEWFgbAwoULGTt2LAUFBeX2OpXX4xQfH092drbNizO/vFqDhREjRqAoCvo3O6LkZ2K6dSm6+B4OvyY/+9NOvvwrgzsub8FDQ9vY1c2Tfk0u2yarg+HDhzskD/LUNlVW94rKAZKTkxkyZIjttejpbXL2PKmq6uDA09vk7HmyWCwODjy9TTU5T4WFhaSkpNg8lFf3vLw8mjRpUq3PJelx8gbMJvh0KBzbdq4ssS90uRE6XAP+tfflZevhHB5ftJ1tR3IBiI8IYHyPBMZ2a0Z0qH+tPa8gCEJdERsbS9OmTW1BE0D79u1RVZXDhw/Tpk0bh338/Pzw8/NzKPfx8XEIPPV6fbkTmysaEmkwGCA0DvIzMRRlg05nO7atfnHhQAa7juVXGOiWV67T6dDpHH/oqqi8orrXqE3VrKOiKA2uTc6UW7/8lncteWqbwLnzVJkDT20TOHeerEFHeQ48tU1Q8/N0voeK7leFBE5Ootfr6d27t2dlKdmzVAua/EKh9zTofANEtLigQ1blIa/YyGu/pvHln4dQVQjxN/DQsLZM6JWIvp7NVaopHnktuBhxoCEevNtB3759mT9/PmfOnCE4OBiA3bt3o9PpaNasWZ3UwcF/yNm1nPIcF8GFMpn1GlBKcm++BssiHsQBiAMrrvbgeeOj3IxOpyMqKqrcqLnesvkL7W+3STDg4QsOmqBiD6qq8n3qEa747yq+WKcFTaOT4lj+7/7c3Lt5gwmawEOvBRcjDjTEQ8NycObMGVJTU23j8Q8cOEBqaqptvP2jjz7KxIkTbdtPmDCBxo0bc+utt7Jjxw5Wr17Ngw8+yG233VZhcghX4+A/pIpFcKO1wOlEfgkn8kvK3cbTaEjX4IUgHsQBiAMrrvbg3TZrgNFoZMmSJU5l4HAreUdhz6/a/UsmVr6tE1g95BUUs/HgKT5Zc4B7v/mbAa/9xr3fpHIiv4SWTYKYc3svZo3vSlRIwxuW53HXQi0gDjTEQ8NysHHjRrp27UrXrl0BmDFjBl27duWpp54CIDMz027ScnBwMCkpKeTk5NC9e3duuukmRo0axVtvvVVndXbwX0VK8iA/gy0t+Zq9J8rdxtNoSNfghSAexAGIAyuu9iBD9WpAeRMx6y2pc0C1QEIfaOI4zr4m/J1+mnkb0ln9j0rWHyswW+zzi/gZdEwb2Jop/Vs2mPVBKsKjroVaQhxoiIeG42DAgAGVZgecPXu2Q1m7du0cMrnVNXb+Q88uglvBUD2AQe2i2HUsnxW7TnBt17oZUljbNJRr8EIRD+IAxIEVV3qQwKkhY7HA5i+1+xfY22Q0W0jelslnfxwkNSPnbKkCqESG+NGlWThdmoXROT6cpGbhhAV6Z2YxQRCEekEVQ/VAC5ze/W0fq9KOYzJbPDK7qSAIQl0igVND5uBqyDmkJYXocE2NDnGqoJS5fx3iyz8PkZWnjYP31esY2SmaRoWHmXT1QOIbBzukfhUEQRDciC1wqrjHqWtCI8IDfcgpNLI5PYeeLSLqqHKCIAieiQROTmIwGBg4cGCFaQ/rFdakEJ2uB9/Ayrc9jxKTmVd/0bLilZi0lJZNgv3416UJ3NQrkSbBvuTntyYkxHuDJo+6FmoJcaAhHsSBu3HwH3o2cCrO1dbxK+czQK9TGHBRJItTj7Ji13GPD5zkGtQQD+IAxIEVV3vwbps1pK6yJF0Qhae0BW/B6WF66ScLuXvuJrYfyQOgU9Mwbu3bnJGdY21zllRV9QwPtYw4EAdWxIM4cDd2/v1CwScQjIVagojGrcrdZ2C7qLOBUxaPDG9XRzWtPeQa1BAP4gDEgRVXepABzU5iMplITk6u/xPuts4DcynEdoG4pGrv9sv2Y4z83+9sP5JHo0AfPp7YnR+m9WXMJc3sEj14jIdaRByIAyviQRy4Gwf/ilJmuF75mfUA+l8UiU6B3VlnOHy6sA5qWnvINaghHsQBiAMrrvYggVNDRFXPDdOrZm9TqcnCMz/+w51fbSK/2ES3xEYsmX45gztEe+1QPEEQBI/GmlmvkgQR4YG+dE/Uhuit3HW8LmolCILgsUjg1BA5sgmO7wBDAHQcW+Xmh08Xcv0H6/jsj4MATOnXkm+mXEpcuHTxCoIgeCzWtZwqSUkO2nA9gBUSOAmCIFSK2wOnd955h+bNm+Pv70+vXr1Yv359pdvn5OQwdepUYmNj8fPz46KLLiI5ObmOaushbJqt/b14NASEV7rpL9uPMfKtNWzJyCEsQBua99iI9vhIWlpBEATPphopyUFLSw6wdt9JikrNtV0rQRAEj0VRK1vlr5aZN28eEydO5P3336dXr17MmjWL+fPnk5aWRlRUlMP2paWl9O3bl6ioKB577DGaNm3KoUOHCA8Pp0uXLtV6zry8PMLCwsjNzSU0NNTpOquqislkwmAwuGcIm6rC7l8BFdoMBd15C8yW5MNrbcFYALf+DIl9yj3MmRITz/zwD/M3HQagS3w4b9/YlfiI6mXfc7uHeoA4EAdWxEP1HVzoe3BDpVY+m9a9C78+ChdfC9fPrnTfy15eyZGcIj65pTtXtI+uYSvci7wONcSDOABxYKU6Hpx5/3Vrt8Lrr7/OHXfcwa233kqHDh14//33CQwM5NNPPy13+08//ZRTp06xePFi+vbtS/Pmzenfv3+1gyZXUVRUVKfPZ6PwFMyfBF+Pg6/Hwzs9tblMppJz22xfqAVNjVtDQu9yD7Px4CmGv7ma+ZsOoyhw14BWzP+/3tUOmqy4zUM9QhyIAyviQRy4Gwf/odXrcVIUxdbr5OnD9eQa1BAP4gDEgRVXenBb4FRaWsqmTZsYPHjwucrodAwePJh169aVu88PP/xA7969mTp1KtHR0XTs2JEXXngBs7nioQUlJSXk5eXZ3QCMRqPtZt3fbDaXW24ymWxlRUVFrFy5EpPJZFduNBqxWCwOxzYajaiqiqqq1S4HsFgs9nXZvQze6wM7FqPqDKh+oXByL/xwD7zZBcuaNzGeOYVl0+fa/kn/AkWxa1NBUQkv/7yTGz5YR8apIpqG+zPnth7MuKIVhrNXQnXbZDQaWblyJUVFRTVukzXDSUXlFZ2P6pynsnWvrfNkMpkcHHh6m5w9T+U58PQ21eQ8FRcX23loCG1y9jxZ3xutDiprk+B6rK9Fu8xR1qF6VcxxgnPD9VbuOu6x56hcB16IeBAHIA6suNqD29Zxys7Oxmw2Ex1tPyQgOjqaXbt2lbvP/v37WbFiBTfddBPJycns3buXu+++G6PRyNNPP13uPi+++CLPPPOMQ/nSpUsJDNR6WBISEujatStbt24lPT3dtk3btm1p164d69ev58SJEw7HWL16Nfn5+bb/e/fuTVRUFEuXLrU7QQMHDiQgIMBhLtaIESNsXzasGAwGRo4cSXZ2NuvWrUNnKaX90fm0PvErAMbQRNbG3soZv1gSs3+jzcml+OVnolv2FOaVL6EzF2JBzz+GznQCW5uyiuCLPXoOF2jdlP3ifRkZc4YTO9aRvAOSkpJITEysdpsuv/xyAFJSUpxuk5WQkBAGDRpERkYGqamptvLIyEj69OnDnj17SEtLs5U7e56cbZOz52no0KEODjy9Tc6epx49ejg48PQ21eQ8bd682c5DQ2iTs+fJWkerg4radNlllyHUEWXnOKmqlqK8Anq3aoy/j46jucWkZeXTLkaGUQqCIJyP2+Y4HT16lKZNm7J27Vp69z43pOyhhx5i1apV/PXXXw77XHTRRRQXF3PgwAH0em1uz+uvv86rr75KZmb561SUlJRQUnJuKFteXh7x8fFkZ2fbxjHqdDr0ej1ms9n262nZcpPJZPsFzmg0kpKSwogRI1AUxe6XOb1ej06ns/0ibMW6WvH50W5F5T4+PlgsFsxHt2L4/k6UEzu1B7pPxjL4Gcx6f9u2irkUw46FqGtmoZzaC4Cl7VWoN3xha9OqtOPc880WCkrNhAf48OKYTgxpH1lu3cu2tbI2qarKzz//zJAhQ/Dx8al+m8r0DiqKgsFgqLC8ovNRnfNUkzY5e54AkpOT7Rx4epucPU+qqjo48PQ21eQ8FRUVsXTpUpuHhtAmZ89TYWEhKSkpNgcV1b2wsJDw8HCZ43QeFzrHyWg0kpyczIgRI869J5tK4Lmz84UfOgCBEZUe47bZG1ix6zgPXdmWuwe0droO7qZcB16IeBAHIA6sVMeDM++/Tvc4NW/enNtuu41JkyaRkJDg7O42mjRpgl6vJysry648KyuLmJiYcveJjY21fSmx0r59e44dO0ZpaSm+vr4O+/j5+eHn5+dQ7uPj4yBQr9fbHduK9QvA+f+fX1722Bdartv1I7rvbtcWsQ2KhGvegYuGoeO88ZU+PnDJzShJE2DXEti3At1l98PZdnz391EeXbgNs0Xl0pYRvDm+K9Gh/g7PV1FbK6qj0WjEYDCU67HCNul06HSOo0MrKq/ofFT3PFVVfqHnqTIHntomcO48VebAU9sENTtP5Xnw9DaVR2VtKs/B+XX35knKtY3DuTH4QWBjKDypDderInAa1C6KFbuOs2LncY8MnKDi69PbEA/iAMSBFVd6cLrHadasWcyePZvt27czcOBAJk+ezLXXXltucFIVvXr1omfPnvzvf/8DtLH5CQkJTJs2jUceecRh+8cee4y5c+eyf/9+2xeDN998k5dffpmjR6seww0ektEp7yi8cymU5MJFw+Hq/0FwpFOHUFWVt5bv5Y1luwEYnRTHK2O74GuQNOOCILgPj3gPdgO15uW9yyBrG9z0HbQZXOmmR3KK6PvSCnQKbHpiCI2CHH+MFARBaGjUala9++67j9TUVNavX0/79u255557iI2NZdq0abZx/tVlxowZfPTRR3z++efs3LmTu+66i4KCAm699VYAJk6cyKOPPmrb/q677uLUqVPce++97N69myVLlvDCCy8wdepUZ5tRYywWC8ePH7cb2uJSVBV+ul8LmuIugXFfOR00Gc0WHvlumy1ountAK94Yl+TSoKnWPXgA4kAcWBEP4sDdVOjfughuftU/LjYND6BdTAgWFVbvcZzXW9+Ra1BDPIgDEAdWXO2hxt+kL7nkEt566y2OHj3K008/zccff0yPHj1ISkri008/rVZWnnHjxvHaa6/x1FNPkZSURGpqKr/88ostYUR6errd3KX4+Hh+/fVXNmzYQOfOnZk+fTr33ntvub1TtYXZbGbdunWVZvK7ILZ+C7t/Ab0vjH4X9M51LxaUmLjji43M25iBToH/jO7IQ1e2c/nwmFr34AGIA3FgRTyIA3dToX9rSvK88ucBn89AD05LLteghngQByAOrLjaQ40H/RmNRhYtWsRnn31GSkoKl156KZMnT+bw4cM89thjLFu2jLlz51Z5nGnTpjFt2rRyH/vtt98cynr37s2ff/5Z02rXb/Kz4OeHtPv9H4Ko9k7tfiK/hFtnr2f7kTz8fXT878ZLGNLBMxcyFARBEFxAo+ba3/UfaAuit7i80s0HtYvivd/2sWr3CUxmCwa9DO8WBEGw4nTgtHnzZj777DO+/vprdDodEydO5I033qBdu3a2ba699lpbmmKhmqgqLJkBxTkQ0xn63ufU7qcLSvnXx3+RlpVP4yBfPr6lO10TGtVKVQVBEAQPodutsON7yNwCX1wDw1+GHrdXmJq8a3w44YE+5BQa+Tsjhx7NK08oIQiC4E04/VNSjx492LNnD++99x5HjhzhtddeswuaAFq0aMH48eNdVsn6hKIohISEuD4z1D8LYddPoDOcHaJX/dSR+cVGJn22nrSsfKJC/FhwV59aD5pqzYMHIQ7EgRXxIA7cTYX+AyPgtl+h0w2gmiH5AW3hdFNJuccx6HX0v0ibV7t8p2cN15NrUEM8iAMQB1Zc7cHprHqHDh0iMTHRJU/uDuplRqczJ+DdXlrK2P6PwMBHq97nLEWlZm75bD3rD5yiUaAP8/6vNxdFh9RiZQVBEGpOvXwPrgfUuhdVhbX/g2VPg2qBZj1h3JfnkkeU4cctR7nn679p1iiA1Q8ORKfz7i9egiA0bGo1q97x48fLXZz2r7/+YuPGjc4ezuOwWCwcOnTItVlKkh/QgqbojnD5v6u9W6nJwl1zNrH+wCmC/Qx8cVuvOguaasWDhyEOxIEV8SAO3E2V/hUF+k6Hm+aDfxgcXg8fDoDDmxw2HdIhmhA/A4dPF7Hh4KnarbgLkWtQQzyIAxAHVlztwenAaerUqWRkZDiUHzlypE7TgrsLs9lMamqq67KU7PgediwGRa8tcmuo3roZJrOF++b9zW9pJ/D30fHppB50ahbmmjpVA5d78EDEgTiwIh7Egbuptv/Wg+GOldCkLeRnavOeCk7abeLvo2dEJy0b38LNR2qryi5HrkEN8SAOQBxYcbUHpwOnHTt2cMkllziUd+3alR07drikUl5DyRlYcraH6bL7IS6pWrtZLCqPLNxG8rZj+Op1fHhzd3q2kAm8giAIQjVo3ApuXwYRraA0H/avdNhkzCVNAViyLZNio3d/8RIEQbDidODk5+dHVlaWQ3lmZiYGQ42zm3snu3+BghMQnqClH68Gqqry7E87WLDpMHqdwls3dqXfRc4tkCsIgiB4Of6h0Ha4dv/AKoeHezSPoFmjAM6UmFi6w/EzXxAEwRtxOnAaOnQojz76KLm5ubaynJwcHnvsMYYMGeLSytVHFEUhMjLSNdk5dnyv/e04Fgx+1drlg9X7mb32IACvju3MlR0dJ/bWBS714KGIA3FgRTyIA3dTI/8t+mt/9//m8JBOpzDmkmYAfLfpsAtqWPvINaghHsQBiAMrrvbgdFa9I0eO0K9fP06ePEnXrl0BSE1NJTo6mpSUFOLj411Ssdqi3mR0Ki2EV1uBsRCm/AZxXavc5fvUI9z7TSoAT13Vgdsua1G7dRQEQXAx9eY9uJ7hFi8lZ+DlRLCYYHoqRNh/phzMLmDAa7+hU+DPR68gKtS/buolCIJQh9RqVr2mTZuydetWXnnlFTp06EC3bt1488032bZtW70PmlyB2Wxm165dFz7JbO8yLWgKT4DYpCo3/3P/SR6cvxWAyZe1cHvQ5DIPHow4EAdWxIM4cDc18u8XrKUlh3J7nZo3CaJbYiMsKnyfetQ1Fa1F5BrUEA/iAMSBFVd7cDpwAggKCmLKlCm88847vPbaa0ycOBEfn+ov2OrJWCwW0tLSLjytoXWYXvurK1zB3cqerHymfLGRUrOF4R1jeHxE+wt7bhfgMg8ejDgQB1bEQ8NysHr1akaNGkVcXByKorB48eJq7/vHH39gMBhISkqqtfqVR439tzw7XK+ceU5wLknEd5vr/3C9hnQNXgjiQRyAOLDiag81zuawY8cO0tPTKS0ttSu/+uqrL7hSDR5TCez+Vbvf4ZpKNz2eV8ykzzaQV2yiW2Ij3hiXJIsRCoIg1CIFBQV06dKF2267jTFjxlR7v5ycHCZOnMgVV1xRbhKleknLAfDbi7B/FVgsoLP/PfWqTnE888MOdh3L55+juVwcV3fLXgiCINQ3nA6c9u/fz7XXXsu2bdtQFAXrFCnrpCtv7xKsFvtWailgQ+KgafcKNysoMXHb5xs4klNEiyZBfDSxO/4++jqsqCAIgueQkZGBoig0a6YlNVi/fj1z586lQ4cOTJkypdrHGT58OMOHD3f6+e+8804mTJiAXq93qpfKrTTtBr7BUHQKsrZDbGe7h8MCfRjcIYrkbcdYuPmIBE6CIHg1TgdO9957Ly1atGD58uW0aNGC9evXc/LkSf7973/z2muv1UYd6xU6nY6EhAR0uhqNctSwDdMb5fDrnhWT2cLUuZvZfiSPxkG+zL61BxFB1Vscty5wiQcPRxyIAyvioX44mDBhAlOmTOHmm2/m2LFjDBkyhIsvvpg5c+Zw7NgxnnrqqVp77s8++4z9+/fz1Vdf8dxzz1W5fUlJCSUlJbb/8/LyADAajRiNRkBzqtfrMZvNdsNMrOUmk8n246XZbCY+Ph6dTmdXDqDX69HpdLbjWjEYDKAzoCb0Rrc3BfPe5ViatLctLWIymQC4pkssyduO8X3qER4edhEK546tKAoGgwGLxWL3w6m1vKK6V6dNZetenTZZHSiKUn5by7TJio+PT4V1rw9tqqzuFZXrdDri4+MbVJucPU/lOfD0Njl7nnQ6Hc2aNbM7jqe3qSbnyWw223kor+7nt7synA6c1q1bx4oVK2jSpIntxFx22WW8+OKLTJ8+nb///tvZQ3oUer3elk2wRpiNkJas3e9Q8bDGl37exW9pJ/D30fHxLd1JbBxU8+esBS7YQwNAHIgDK+KhfjjYvn07PXtqyQ6+/fZbOnbsyB9//MHSpUu58847ay1w2rNnD4888gi///57tdczfPHFF3nmmWccypcuXUpgYCAACQkJdO3ala1bt5Kenm7bpm3btrRr147169dz4sQJW3lSUhJ6vZ4VK1aQn59vK+/duzdRUVEsXbrU7svCwIEDCQgIYFdRJJ2A7PUL+fNUS0aMGEFRURErV2oL45otEOxjIPtMKUs27UeXtdN2jJCQEAYNGkRGRgapqam28sjISPr06cOePXtIS0uzldekTYmJiaxevbrabVJVleTkZDuv57cJtC9QI0eOJDs7m3Xr1tXrNgUEBDjVpmbNmvHrr782qDY5e54CAwPtHDSENjl7nkpKSuwcNIQ2OXueli9fjslk4vDhwxW2qbCwkOridDryRo0asXnzZlq0aEGrVq34+OOPGThwIPv27aNTp05OPbk7uNCUr2azma1bt9K5c2f0+hoMm9u7HL4aA0GR8O800Dke44+92dz08V8AvP+vS7iyY6zzz1PLXLCHBoA4EAdWxEP1HdRm2u3g4GC2b99O8+bNufrqq+nbty8PP/ww6enptG3blqKiIqePqSgKixYtYvTo0eU+bjabufTSS5k8eTJ33nknADNnzmTx4sV2XxLOp7wep/j4eLKzs21enO1x2rFjB126dEFVVed+TT66FZ+P+qH6BGKasQeDv/ZDXdkvUM8l7+LzdemM7BTDrBvODedz96/J5/c47dixg86dO3P+VxtP+oXcFT1OW7ZsoUOHDrbXoqe3ydnzpCiKgwNPb5Oz5wm05YIuvvhimwNPb1NNzlNxcTH//POPzUN5dc/Ly6NJkybV+lxyusepY8eObNmyhRYtWtCrVy9eeeUVfH19+fDDD2nZsqWzh/M4LBYL6enpdOzYsWZfkKzD9NpdVW7QlFtk5IH5WwCY0CuhXgZN4AIPDQBxIA6siIf64eDiiy/m/fffZ+TIkaSkpPCf//wHgKNHj9K4ceNaec78/Hw2btzI33//zbRp0wDNhaqqGAwGli5dyqBBgxz28/Pzw8/PceFzHx8fhyy1er2+XKfn925lZGTQqVOnCrPcVlge1xmCIlEKTuCTlQrNL3PY/vruCXy+Lp2UnccpNEFYgP2xrCNQzqeiule3TVWVn9+mmjioqO71pU3OlhuNxgo9eGqbwLnzVJkDT20TOHeejEYjhw8fpnPnzg7P7altAufPk16vL9dDRferwunB6E888YQtAnz22Wc5cOAAl19+OcnJybz11lvOHs67sJhh1xLtfgXZ9Gb+8A+ZucU0bxxYL9KOC4IgeAovv/wyH3zwAQMGDODGG2+kS5cuAPzwww+2IXyuJjQ0lG3btpGammq73XnnnbRt25bU1FR69epVK8/rUhQFWpxNS76//LTkF8eFclF0MKUmC8nbMuuwcoIgCPUHp3uchg0bZrvfunVrdu3axalTp2jUqJEts55QAYfWQmE2BDSy/aJXlp+2HmXR30fQKfD6uCSC/GqcLV4QBMHrGDBgANnZ2eTl5dGoUSNb+ZQpU2zzhqrDmTNn2Lt3r+3/AwcOkJqaSkREBAkJCTz66KMcOXKEL774Ap1OR8eOHe32j4qKwt/f36G8XtNyAGxfoC2EO+hxh4cVReG6S5rx4s+7WLj5MDf2TKjzKgqCILgbp3qcjEYjBoOB7du325VHRER4TdCk0+lo27ZtzTJHWYfptR0JevtuwWO5xTy+SPM6dWBrLklodP7e9YoL8tBAEAfiwIp4qB8OioqKKCkpsQVNhw4dYtasWaSlpREVFVXt42zcuJGuXbvakl3MmDGDrl272pJLZGZm2k1arg9csH/rQrhHNkFxXrmbjO7aFJ0CGw6eZtex8rdxJ/XhGqwPiAdxAOLAiqs9OJ0comXLlixatMg2BMLTqM2JyZViscDr7eHMMZgwHy4aantIVVVu+WwDq3efoFPTMBbe3QcfvXdf6IIgNExq8z146NChjBkzhjvvvJOcnBzatWuHj48P2dnZvP7669x1110ufT5X4rbPprK81RVO7Ycbv4G25a9jdddXm/h5+zF6tYjgmymXes2PpoIgNFycef91+tv5448/zmOPPcapU6dqXEFPxmQysXbtWodMIlVyeIMWNPmFnvtl7yxf/XmI1btP4GfQ8ca4Lh4RNNXYQwNCHIgDK+KhfjjYvHkzl19+OQALFiwgOjqaQ4cO8cUXXzT4Obgu8V/FPCeAx0a0x8+g468Dp/hhy9GaP1ctUB+uwfqAeBAHIA6suNqD05No3n77bfbu3UtcXByJiYkEBdmvL7R582aXVKy+oqoqJ06ccEh1WiXWYXoXXQmGc5mU9p04w/PJ2poYjw5vR+uoEFdVtVapsYcGhDgQB1bEQ/1wUFhYSEiI9h66dOlSxowZg06n49JLL+XQoUNuq1dd4BL/LQfAps+0eU4VEB8RyLSBrflvym6eX7KTQe2iCPGvfkaq2qQ+XIP1AfEgDkAcWHG1B6cDp4rWshAqQVVh54/a/TLZ9ExmCzPmpVJstHBZ6yZM7N3cPfUTBEFoALRu3ZrFixdz7bXX8uuvv3L//fcDcPz4cfcNf/MkWvQDFDixE/KzICS63M3u6NeSBZsPc+hkIW8t38PjIzvUbT0FQRDchNOB09NPP10b9WjYHP0bctPBJwhaX2ErXpx6lC2Hcwn1N/Dq9Z3R6WSsuCAIQk156qmnmDBhAvfffz+DBg2id+/egNb7ZE30IFRCYATEdobMLXBgFXS+odzN/H30zLz6Ym79bAOf/nGQsd3iaRvjGaMlBEEQLoT6P5mmnqHX60lKSqr+Ao9FOfD7f7X7bYaATwAARrOFt5bvAeDuga2JDQuohdrWHk57aICIA3FgRTzUDwdjx44lPT2djRs38uuvv9rKr7jiCt544w231asucJn/lgO0v5XMcwIY2DaKIR2iMVtUnvp+e70YDlQfrsH6gHgQByAOrLjag9NZ9XQ6XaVZdMxm8wVXqjaps8xFFgtsmQspT2trNwH8a6Gtx+nr9ek8unAbTYJ9Wf3QQAJ9Zc0mQRAaPnX1Hnz48GEAmjVrVmvP4UrqRVY9gL3L4asxENoM7t+uLY5bARmnChn8+ipKTBbeHJ/ENUlN67CigiAIrqFWs+otWrSIhQsX2m7z5s3jkUceITY2lg8//LDGlfYUTCYTK1asqDw7x5HN8MkQ+H6qFjQ1uQhuXmwLmkpMZv53trfprgGtPTJoqpaHBo44EAdWxEP9cGCxWHj22WcJCwsjMTGRxMREwsPD+c9//oPFYnFbveoCl/lP6A16X8g7rKUmrwRrogiA55fsJL/YeGHPfYHUh2uwPiAexAGIAyuu9uD0N/ZrrrnGoWzs2LFcfPHFzJs3j8mTJ7ukYvUVVVXJz88vf1hCwUlY8Sxs+hxQwTcY+j8Mve4Eg69ts3kbMjiaW0x0qB839fLM1dcr9eAliANxYEU81A8Hjz/+OJ988gkvvfQSffv2BWDNmjXMnDmT4uJinn/+ebfVrbZxmX/fQIjvBQd/h/0roXGrSjevT4ki6sM1WB8QD+IAxIEVV3tw2RynSy+9lOXLl7vqcJ5H0Wl4/zLYNBtQodMNMG0j9J1uFzQVG828vWIvANMGtsbfx7vHngqCILiKzz//nI8//pi77rqLzp0707lzZ+6++24++ugjZs+e7e7qeQ7WtQb3raxyU2uiCIBP/zhI2rH82qyZIAiCW3FJ4FRUVMRbb71F06ZePL75wGrIPwrBMTApGa77CEJjHTb76s9DHM8voWl4ADf0iHdDRQVBEBomp06dol27dg7l7dq189pF22tEq0Ha3wOrwVz18LuyiSJeOLsuoSAIQkPE6cCpUaNGRERE2G6NGjUiJCSETz/9lFdffbU26liv0Ov19O7d2zE7x7Ht2t82g6F533L3LSgx8d5v+wC4Z1Br/Aye29tUoQcvQhyIAyvioX446NKlC2+//bZD+dtvv03nzp3dUKO6w6X+Y5MgIAJK8uDwxmrt8sTI9ugUWLX7BNuP5F54HWpAfbgG6wPiQRyAOLDiag9Oz3F644037LLq6XQ6IiMj6dWrF40aNXJJpeozOp2OqKgoxwey/tH+RnescN/P1x3kZEEpCRGBXNfNMzI9VUSFHrwIcSAOrIiH+uHglVdeYeTIkSxbtsy2htO6devIyMggOTnZrXWrbVzqX6fXep22L4C9yyCxd5W7JDYO4qrOcfyw5Sjvr9rH2xMucU1dnKA+XIP1AfEgDkAcWHG1B6d7nCZNmsQtt9xiu918881ceeWVXhE0ARiNRpYsWYLReN7whaxt2t8KAqf8YiMfrtYyFN17RRt89J69hFaFHrwIcSAOrIiH+uGgf//+7N69m2uvvZacnBxycnIYM2YM//zzD19++aXb6lUXuNy/dbH2fdWfu3xnfy2RRPK2TA5mF7imHk5QH67B+oB4EAcgDqy42oPTPU6fffYZwcHBXH/99Xbl8+fPp7CwkFtuucUlFavPOKQ0LM6FnHTtfkz5gdOnaw6SU2ikZWQQo7s2jLlg3p7iEsQBiAMr4qF+OIiLi3PInrdlyxY++eSTBr9khkv9W+c5HU2FgmwIalLlLh3iQhnQNpLf0k7w4e/7eeHaTq6rTzWpD9dgfUA8iAMQB1Zc6cHpbo8XX3yRJk0c30CjoqJ44YUXXFIpj8M6TC+0GQQ49rzlFhr5eI3W23Tf4IvQ6ypeUFAQBEEQ3E5IDER3AtRqZdezctfZXqcFGw9zPK+4lionCILgHpwOnNLT02nRooVDeWJiIunp6S6plMdhm990cbkPf/LHAfKLTbSNDuGqTo6Z9gRBEASh3mEdrrd3WbV36dkigksSwik1W/j0j4O1Uy9BEAQ34XTgFBUVxdatWx3Kt2zZQuPGjV1SqfqMwWBg4MCBGAxlRjkeOzu/qYJhej9tPQrA3QNboWsgvU3levAyxIE4sCIexIG7qRX/tnlOK8BiqdYuiqJw14DWAMz58xB5xXU3v0KuQQ3xIA5AHFhxtQenj3LjjTcyffp0QkJC6NevHwCrVq3i3nvvZfz48S6pVH0nICDAviDrbCrychJDZJwqZP+JAvQ6hQFtG1Z2EwcPXog4EAdWxIP7HIwZM6bSx3NycuqmIm7G5f7jLwWfICg4riVAiu1Srd2uaBdFm6hg9hw/w5w/07lrQCvX1qsS5HWoIR7EAYgDK6704HSP03/+8x969erFFVdcQUBAAAEBAQwdOpRBgwZ5xRwnk8lEcnLyuYlmFjNk7dDuxzhOhP0t7TgA3RIaERbgU1fVrHUcPHgh4kAcWBEP7nUQFhZW6S0xMZGJEyfWeb3qklrxb/CFFtoPpOytfnY9nU6xZdj7ZM0Bio1m19WpEuR1qCEexAGIAyuu9uB0j5Ovry/z5s3jueeeIzU1lYCAADp16kRiYqJLKuRxnDoApiIwBEBES4eHf0s7AUD/tpF1XTNBEASv4LPPPnN3FRoura+A3T9rgdPlM6q929VJcfx3aRpHc4v5bvNhburlpd8RBEFoUNR4wF+bNm1o06aNK+vimVjXb4pqry0aWIZio5m1+04CMEACJ0EQBMHTsM5zyvgTSvLBL6Rau/noddzRryXP/LiDD1btZ1z3eAwevn6hIAiC0+9i1113HS+//LJD+SuvvOKwtpNXcOzs/KZyEkNsPHiaIqOZqBA/OsSG1nHFBEEQBOECiWip3SwmOLDaqV3H9YinUaAP6acK+Xn7sVqqoCAIQt3hdOC0evVqRowY4VA+fPhwVq927k3VEzEYDIwYMeJcdg5bYoiK5zf1vygSRWkY2fSsOHjwQsSBOLAiHsSBu6lV/62sacmrP88JINDXwKQ+2vIl7/22D1VVXV0zO+Qa1BAP4gDEgRVXe3A6cDpz5gy+vr4O5T4+PuTl5bmkUvWdoqKic/9UsobTb7u1+U0NLZueFTsPXoo4EAdWxIM4cDe15r/1YO3v3mXgZPAzsXcigb56dmTm1Umvk1yDGuJBHIA4sOJKD04HTp06dWLevHkO5d988w0dOnRwSaXqMyaTiZUrV2rZOYpOQ26G9sB5gdPh04XsPX4GnQKXtW7ihprWLnYevBRxIA6siAdx4G5q1X/zy0DnAzmH4NR+p3ZtFOTL7ZdpvU4zf/iH3KLaW9dJrkEN8SAOQBxYcbUHp/utnnzyScaMGcO+ffsYNGgQAMuXL2fu3LksWLDAJZXyGKy9TWEJEBBu95A1m94lCY0IC2w4acgFQRAEL8MvGBJ7a3Oc9i6Dxs6ty3T3wNb8tDWT/dkFvPzLLl641nFouyAIgifgdI/TqFGjWLx4MXv37uXuu+/m3//+N0eOHGHFihW0bt26NupYf6kkMYQ1cJJseoIgCILHY5vntMzpXf199LwwRguW5v6VzoaDp1xZM0EQhDqjRrlBR44cyR9//EFBQQH79+/nhhtu4IEHHqBLl+qtKu7pOCaGsB+mV2Iys3ZfNtBw5zcBXj/hEMQBiAMr4qHhOFi9ejWjRo0iLi4ORVFYvHhxpdsvXLiQIUOGEBkZSWhoKL179+bXX3+tm8qWoVb9W9OSH1wDxmKnd7+0ZWPGdY8H4JHvtlJiqp1FcRvKNXihiAdxAOLAiis91HhRhdWrV3PLLbcQFxfHf//7XwYNGsSff/7psorVV3x8fBg5ciQ+Pj5lAif7HqdNB09TWGqmSXDDTUNu58FLEQfiwIp4aFgOCgoK6NKlC++88061tl+9ejVDhgwhOTmZTZs2MXDgQEaNGsXff/9dyzU9R637j+4IwdFgLIT0dTU6xGMj2tMk2I99Jwp477d9Lq5gw7oGLwTxIA5AHFhxtQenAqdjx47x0ksv0aZNG66//npCQ0MpKSlh8eLFvPTSS/To0cMllarPWCwWjh8/jsVUCsd3aoUx9uO1rdn0+l8UiU7XsNKQW7F5sFjcXRW3IQ7EgRXx0LAcDB8+nOeee45rr722WtvPmjWLhx56iB49etCmTRteeOEF2rRpw48//ljLNT1HrftXlHPD9fY5l5bcSligD0+P0pJIvbtyH3uP57uqdkDDugYvBPEgDkAcWHG1h2oHTqNGjaJt27Zs3bqVWbNmcfToUf73v/+5pBLvvPMOzZs3x9/fn169erF+/fpq7ffNN9+gKAqjR492ST2qg9lsZt26dZhP7AFTMfgEQqMWdtvY1m9qwPObbB7MtTPcwhMQB+LAingQB2WxWCzk5+cTERFR4TYlJSXk5eXZ3QCMRqPtZnVpNpvLLTeZTLay4uJim/+y5Uaj0faFoWyZ0WhEVVVUVa12uXW4nrpnma3MmqnKYrHYbWstP7/uwy+OYlC7KErNFm3IXklphW0qW/fqtMnq4PxtK21TJXWvbpucOU/Otqkm58n6WiwuLm4wbXL2PJXnwNPb5Ox5Ks+Bp7epJufJ+r5g9VBZ3atDtQf9/fzzz0yfPp277rqLNm3aVPsJqmLevHnMmDGD999/n169ejFr1iyGDRtGWloaUVEVzw86ePAgDzzwAJdffrnL6uIMyvGzGfWiOoDuXPx5NKeI3VlaGvJ+bRpeGnJBEAShcl577TXOnDnDDTfcUOE2L774Is8884xD+dKlSwkMDAQgISGBrl27snXrVtLT023btG3blnbt2rF+/XpOnDjhcIzVq1eTn3+uN6d3795ERUWxdOlSu5S8AwcOJCAggOTkZLv9R4wYQVFREStXrrSVGQwGRg4chIqCcmInvy36nEK/SEJCQhg0aBAZGRmkpqbato+MjKRPnz7s2bOHtLQ0W3lCQgL/Gd2Rta+tZOOhHJ78/Bf6RqsVtikpKYnExMRqtwm0L1ApKSnVa9PIkWRnZ7Nu3bnhhzVpkzPnydk2OXuehg4dCmDnwNPb5Ox5so6AKuvA09vk7Hlq1aqVgwNPb1NNzpO1jlYP5bWpsLCQ6qKo1VzK+88//+STTz5h3rx5tG/fnptvvpnx48cTGxvLli1baryGU69evejRowdvv/02oEWk8fHx3HPPPTzyyCPl7mM2m+nXrx+33XYbv//+Ozk5OVVO3rWSl5dHWFgYubm5hIY6P//IaDSSnJzMVQGb0a+dBd0mwag3bY/P/SudxxZt45KEcBbe3dfp43sKVg8jRozw2vGz4kAcWBEP1Xdwoe/BdY2iKCxatKjaIxvmzp3LHXfcwffff8/gwYMr3K6kpISSkhLb/3l5ecTHx5OdnW3zotPp0Ov1mM1mu2Em1nKTyYT1I9xoNJKSksKIESNQFIWyH+16vR6dTufwq6p1wvT5gUdF5T4+PqifX41yYBWWzhMwj3oLRVEwGAxYLBa73kZreUV1//j3fTy3ZBfBfgZ+md6H2PBAhzaVrXtF5WXbZHUwfPhwFMV+mHxlbaqo7s62qTrnydk2VVb3isoBkpOTGTJkiO216OltcvY8qarq4MDT2+TsebJYLA4OPL1NNTlPhYWFpKSk2DyUV/e8vDyaNGlSrc+lavc4XXrppVx66aXMmjWLefPm8emnnzJjxgwsFgspKSnEx8cTEhJS3cMBUFpayqZNm3j00UdtZTqdjsGDB9tFq+fz7LPPEhUVxeTJk/n9998rfY7yPpwAu645Z06SyWQiODgYJVPrcTI36YDl7HH0er1tmN7lrRvbjt8QLrzyujGDg4MbVJucPU+Kojg48PQ2OXueynPg6W2qyXkym812HhpCm5w9T9b3RmtdK6p7NX+r80i++eYbbr/9dubPn19p0ATg5+eHn5+fQ7mPj49D4KnX69Hr9Q7bls0UpSgKISEhtvNdHhUFtM6UK1c8BR9fgW7r1+j6TrNlldXpdOh0jqP/K6r7rX1b8uOWTLYczuXJH3by8S09HNpUluq0yepAp9M55aCiujvbpuqcp+qUX+h5MplMhISE4Ovr6/AcntomcO48VebAU9sEzp0nVVUrdOCpbQLnz5Ovr2+5HsrW3ZkfPKvd41QeaWlpfPLJJ3z55Zfk5OQwZMgQfvjhh2rvf/ToUZo2bcratWvp3bu3rfyhhx5i1apV/PXXXw77rFmzhvHjx5OamkqTJk2YNGlSpT1OM2fOLHc4xNy5cx2GQ/z999/ldguuXbvWsVtwwZWQf5Tf2zzOqeC2AHTv2Yth72/hTImJf3cykRCsbV+Trs7jx4+X29V56NChcrs6d+3aVW5Xp1NtSkxkxYoV5XZ1Llmy5IK7b6VN0iZpk7Tpsssuq/Yve/WB6vY4ff3119x222188803XHPNNU4/j0f1xH07EXZ8D22Gwk3za3yYHUfzGP3uH5SaLPzr0gT+c01Hh54iQRCE2saZ998LCpysmM1mfvzxRz799NNaDZzy8/Pp3Lkz7777LsOHDweoMnBy9XAIi8XCsf07SJw3AADjAwfAT+tp23AohwkfryciyId1Dw2wZdRrCL8mn/8LuU6nIyMjg9jYWLtfEDy5Tc6eJ71eT3p6up0DT2+Ts+dJp9Nx6NAh4uLibA48vU01OU+lpaUcPnyYpk2b2so8vU3OnqeSkhKOHDlic1BR3QsLCwkPD6/XAcKZM2fYu3cvAF27duX1119n4MCBREREkJCQwKOPPsqRI0f44osvAO2HuFtuuYU333yTMWPG2I4TEBBAWFhYtZ7zQgMni8VCRkYG8fHx5f6q61JO7oN3eoLFBLf8BC1qPtf4522Z3D13M6oKDw5ry9SBrWt8rDp1UI8RD+IAxIGV6nhw5v3XJStC6fV6Ro8e7XR2uyZNmqDX68nKyrIrz8rKIiYmxmH7ffv2cfDgQUaNGmUrs37gGwwG0tLSbJPhrLh6OITRaOTwpl9IBAhPxCf4XNakVXu0RW/7XxSFn59vuc9ZHp7Q1Xl+HY1GI1u2bKFp06bl1t8T2+RseWUOPLVN4Nx5MhqNbN26lWbNmtXo9VSdck8YOqAoCtu2bSM+Pt7u+T25Tc6eJ51OV66D8+vuCT0KGzduZODAgbb/Z8yYAcAtt9zC7NmzyczMtOul+/DDDzGZTEydOpWpU6fayq3b1wVms5nU1FS7HzFqjcattPm9Gz6GlKfgjhVauvIaMLxTLE9f1YGZP+7g1V/TiAn157puzWp0rDp1UI8RD+IAxIEVV3tw65LCvr6+dOvWjeXLl9uCLovFwvLly5k2bZrD9u3atWPbtm12ZU888QT5+fm8+eabxMfH10W1CS06+4F53vpNq9K0YS0DGnAackEQhIbOgAEDKp2LdX4w9Ntvv9Vuheoj/R+GLd/A0c3wzyLoOKbqfSpgUt8WZOYW88Hq/Tz83VYiQ/zod5F8jgqCUP9wewg6Y8YMPvroIz7//HN27tzJXXfdRUFBAbfeeisAEydOtCWP8Pf3p2PHjna38PBwQkJC6NixI76+jr08tUFYUYZ25+ykWIBjucXsOpaPosDlbeQNXxAEQWjABEdBn+na/eXPgqn0gg738JXtuCYpDpNF5a6vNrH9SK4LKikIguBa3B44jRs3jtdee42nnnqKpKQkUlNT+eWXX4iOjgYgPT2dzMxMN9fyHIqiEGE8W5/ojrbyLYdzAGgXE0pEUN0EcO5EURQiIyM9YthNbSEOxIEV8SAO3I1b/PeeCkFRcPoAbJp9QYfS6RReHduFPq0aU1Bq5tbZG8g4Vf21VUCuQSviQRyAOLDiag8uSQ7hSVxw5iKzCV6IA3MJTP8bIloC8L/le/hvym7GdG3K6+OSXFtpQRCEBoJHZY+rQzzWy4ZPYMkMCGwM01PB/7y6qyoc3gil+dBqUJWHyys2csP769h1LJ+WkUF8c8elRIX6107dBUEQcO791+09Tp6G+fguMJeg+gZDeHNb+a4sLT1v2xjn1rLyVMxmM7t27bLL3uVtiANxYEU8iAN34zb/l0yExq2h8CSsfetc+emDsOoV+N8l8Mlg+PJaOLC6ysOF+vsw+9aexIX5s/9EAVf9bw1/p5+uVlXkGtQQD+IAxIEVV3uQwMlJ1GNacgo1sj2Uyc6x+5gWOF3kJYGTxWIhLS3NLo2xtyEOxIEV8SAO3I3b/Ot9YPBM7f66d+CvD+CzEfBmF1j5PJzaf27bXcnlHuJ8YsL8+XrKpbSJCuZ4fgnjPviTbzdmVLmfXIMa4kEcgDiw4moPEjg5iXJ8BwBqmcQQJSYz+7MLAGjnJYGTIAiCIADQ7iqI7wXGQvj5ITj0B6BAywFw7Qcw+n1tu70p1T5kYuMgFk3ty9AO0ZSaLTy0YCszf/gHo9m7vwQKguBeJHByEiXrH+1O1LnAaf+JAswWlRB/AzEyFlsQBEHwJhQFrnwRfEMgsp3WA3X/PzDxe+gyHtqNAJ0BTu6FUweqfdhgPwPv/6sb9w1uA8DstQeZ+Ml6ThVcWAY/QRCEmiKBk7MERmD0awSx59ZwSjs7TK9dTIjXZC/R6XQkJCR49aJq4kAcWBEP4sDduN1/027w2GGY+hdcdj+ENT33mH+Y1iMFsHeZU4fV6RTuG3wRH9zcjSBfPev2n+Tqt9ewMzOvnG3lGgTxAOIAxIEVV3uQrHo1RVVtK6W//Msu3vttHzf1SuD5aztVsaMgCIL34rHZ42qZBu/l99dh+TPQZhjc9G2NDrE7K587vtjIoZOFNAn2Y+n9/bxi+Q9BEGoXyapXi5jNZv7++2/MZSaZ7S7T4+Qt2Dx4cbYWcSAOrIgHceBu6r3/NkO0vwdWg7G4Roe4KDqE76f2pU1UMNlnSnjy++12j9d7B3WEeBAHIA6suNqDBE5OYrFYSE9Pt8vOscuaUS/aewKn8jx4G+JAHFgRD+LA3dR7/9EdISQWTEVnk0fUjPBAX16/IQm9TmHJ1kx+3HLU9li9d1BHiAdxAOLAiqs9SOB0geQXGzmSUwR4zxpOgiD8f3v3Hd9Uvf9x/HWSdNOWAp1QKLNskFXLENlQ5IrjiooK7gEIcr2KiuL84RVEVLw47lWuAwEHLoaUjZQNZQiUMlug0BbonknO749jAqEtbUratM3n+Xjk0fTk5OT7fZ+TnHxzzvl+hRB2URRoNUi7f3TNdS2qUxN/JgxoBcDLPx8gNbtyR7CEEMJe0nC6TkfO5wAQ7OdBfW8511oIIYQoVavB2l87uiUvy8QBregQ5kdGXjEv/rgfF7tcWwjhJNJwspNOpyMyMtLaO8eR89ppepEhdfBi3mu4OgdXJBlIBhaSg2TgbLUi/xYDQNFD+hG4dOq6FuVu0PHuXV1w0yusPpTKD7vP1I4MqoHkIBmAZGDh6BxcO81K0Ov1tG3bFr1eD1zuijwyuJ4zi1Xtrs7BFUkGkoGF5CAZOFutyN+rPoT30u474KhT2xA/nhnSBoDXfvmT89lFNT+DalArtoUqJhlIBhaOzkEaTnYyGo3ExcVhNBqBKxpOLnbE6eocXJFkIBlYSA6SgbPVmvwtp+sl2jeeU1ke69eCruH1yS408s/v9rJ58+aan0EVqzXbQhWSDCQDC0fnIA0nO6mqSlpaGqqqoqoqCZZT9VyoRz2wzcFVSQaSgYXkIBk4W63J/8puyY2F1704g147Zc/DoGPzsQv8dOBCzc+gitWabaEKSQaSgYWjc5CG03VIzyniYm4RigKtXexUPSGEEMJuIZ2hXjAU58KpOIcssmVgPZ4b3haAn0/p2Hc60yHLFUKIq0nD6TpYTtOLaOiDp5trn0MqhBBClEtRruhdzzGn6wE82DuCXhEBFJkVxny2nQ/XJGI0ufb4NUIIx5OGk530ej1du3ZFr9e77Gl6YJuDq5IMJAMLyUEycLZalX8VNJx0OoVP7u9O/xZ+GM0q78Ye4e+fbOFkeq7DXqO2qFXbQhWRDCQDC0fnIA0nO+l0Opo1a4ZOpyPhXBYAbVxw4Nsrc3BVkoFkYCE5SAbOVqvyb3EzKDpIOwwZyQ5bbICPBwse7ct7Y7rg62FgT1IGI97fxMJtSS51nUet2haqiGQgGVg4OgfXTrMSjEYja9euxWg0kvDX4LdtXbDhdGUOrkoykAwsJAfJwNlqVf7eDaBJT+2+A7oltzAajaxbt45RnUJY+cxNRLdoSH6xiReX7ufh/+0kNbvAYa9Vk9WqbaGKSAaSgYWjc5CGk51UVSU7OxuTyUziX6fqtXHBU/UsObjSr3hXkwwkAwvJQTJwtlqXf6u/etdzULfkYJtB4/pefPNIFNNHtsPdoGPt4VRi3t9E3NF0h71eTVXrtoUqIBlIBhaOzkEaTpV0OiOfvCIT7gYdEQ29nV0cIYQQovZo/dd1Tic2gLGoSl5Cp1N4pF8Lfp3Yl7YhvqTnFHHff7cxb20iZrNrf5kUQlSONJwq6chfp+m1CqyHQS8xCiGEEBUW0gV8AqEoB5K2VOlLRYb48tOEPtzVowlmFWavOsJD/9vBpdyqabAJIeou+cZvJ71eT3R0NEfTtJ56XPH6Jricgyv31iIZSAYWkoNk4Gy1Ln+dzuG9610rA083Pe/c2YV37uiMh0HH+oQ0bvnwD+KTMxzy2jVJrdsWqoBkIBlYODoHaTjZSafTERQUZO0YwhV71IPLObhyby2SgWRgITlIBs5WK/Nv/dd1Tru/hNTD1724imRwV89wlj7Vh4iG3pzJyOfvH8fxv7iTdeo6kFq5LTiYZCAZWDg6B9dOsxKKi4tZtmyZtSvySBdtOFlyKC4udnZRnEYykAwsJAfJwNlqZf5tb4HGPaAgA76+HTLPXNfiKppB+zA/fpnUl+EdQig2qcz45U9GzfuDFftT6sS1T7VyW3AwyUAysHB0DtJwqoSCIiMn0vMA1xz81sLVu7gEyQAkAwvJoe5ksHHjRkaNGkVYWBiKovDTTz+V+5z169fTrVs3PDw8aNWqFQsWLKjycl6t1uVv8IB7l0DD1pB1Br6+A/IvXdciK5qBn6cb8+/rxvSR7fBy03PgTBZPfrOboXM38uPu0xSbzNdVDmerddtCFZAMJAMLR+YgDadKSM0Ho1nF19NAqL+ns4sjhBDCgXJzc+nSpQsfffRRheY/ceIEI0eOZMCAAcTHxzNlyhQeeeQRfv/99youaR3g0xDu/xF8QyHtEHx7DxTnV8tLK4rW697maQOZNLAVvp4GjqbmMHXJXgbMXs/XW09RUGyqlrIIIWoHg7MLUBul5CuAdrRJURQnl0YIIYQjjRgxghEjRlR4/o8//pjmzZvz7rvvAtCuXTv++OMP3nvvPYYNG1ZVxaw76jeF+36Az0doPex9/zDc9SXoq+crSgMfd/4xNJLHbmrBV1tP8d9NJzh9KZ/pPx1g9qoEhncIYVSXMKKaN5BedIVwcYpal66IrICsrCz8/f3JzMzEz8/P7uerqsqbv+7nv3HJjI1qylu3daqCUtZ8lgHFfH1dt/EoGUgGFpJDxTO43s/g6qYoCkuXLmX06NFlznPTTTfRrVs35s6da532xRdfMGXKFDIzM0t9TmFhIYWFhdb/s7KyCA8PJz093ZqLTqdDr9djMpkwmy+fOmaZbjQarZ0aqKpKXl4e/v7+mEwmm84O9Ho9Op2uxDn+BoPWMLn6NJaypru5uWE2mzGZLh+FURQFg8FQ5vSyyl7q9OStqF/dhmIqxNz1fkwxc9AbDOh0Opu6llUnVVXJz8/H19fXpiz21qmg2MwPe87yycbjpGQWWKc39HEnplMoMR2D6Rbuj06nlF+nq9bTlWWvSJ2uVfZrTc/MzMTb29v6XnToenJCnezd9vR6fYkManud7F1POp2OjIwMfHx8rBnU9jpVZj0VFRWRk5NDvXr1rK95ddmzsrJo1KhRhfZLcsSpEk5c0D5IXbVjCAsvLy9nF8HpJAPJwEJycN0Mzp07R3BwsM204OBgsrKyyM/PLzWXmTNn8tprr5WYvmrVKry9tUHVmzZtyg033MC+fftISkqyzhMZGUnbtm3Zvn07aWlp1umdOnXC39+fjRs3kp2dbZ0eHR1NUFAQq1atsvmyMGDAALy8vFi+fLlNGWJiYsjPz2fdunXWaQaDgZEjR5Kens6WLZfHXfL19WXgwIEkJycTHx9vnR4YGEjv3r1JTEwkISHBOv3adepDQsd/Ern3LXTxX3HkXDZeI16nWbNmFa5Tv379MBqNrFix4rrqNH7gQPqFKfywaR970hX2XlS4kFvEV1tP8dXWU9R3V7mvlZnW/qrd66lr16521cne9RQTE0NeXh4bNmyoovVU/XWyd9uLjo7mzJkzHD16tM7Uyd71FBkZycGDB0lPT68zdarMeoqNjS23Tnl5eVSUHHGyU3FxMVFv/s7FQoVFj93IjS0aVkEpa77i4mKWL19OTEwMbm5uzi6OU0gGkoGF5FDxDOriEac2bdrw4IMP8sILL1inLV++nJEjR5KXl1dqw8nRR5yKi4uJjY0lJiYGRVFq9a/Jys7P0a/8JwDmu75C1/5vFfqF3JLBiBEjShz1vN46FZvMbDl+keUHUll18BzZBUYa+rizbGI0gX5eNepIBmjb35AhQ6zvRWf/6l/d256qqiUyqO11snc9mc3mEhnU9jpVZj3l5eURGxtrzUGOOFWznEIjFwsvX+MkhBDCtYWEhHD+/HmbaefPn8fPz6/Mo3AeHh54eHiUmO7m5lai4anX60sdvNHyBaCi08tq0Noz3XIKUEWnl1X2a9bpxscg8xRsmYdu5TRoORCDR70Kl1FRFIfXyc0NBrUPZVD7UPKLOjL6o80knM9m+i+H+OyBHuXXqRRVtZ4sX35L25Ycup5KUVO2vWtlUFvrBPatJ0ujo7QMamudoPLr6eocyrpfHrnK0U6JqdrAt0G+HgT4uDu5NEIIIZwtOjqaNWvW2EyLjY0lOjraSSWqAwZOh/rNtG7KN/zL2aWx4eWuZ+7dXXHX61h9KJVvtyc7u0hCiGoiDSc7JZ7XGk5tgkv/9UsIIUTtlpOTQ3x8vPV8/BMnThAfH2893/6FF17ggQcesM7/xBNPcPz4cZ577jkOHz7Mv//9b5YsWcIzzzzjjOLXDW5eEDNLu7/133D+oHPLc5V2oX78c1gkAG/8dpAT6blOLpEQojrINU52evWXP1kQd5KH+zbn5VvaV0EJawdVVTEajRgMBpfuRUwykAxAcoCKZ1AbrnFav349AwYMKDF93LhxLFiwgPHjx3Py5EnWr19v85xnnnmGgwcP0qRJE15++WXGjx9f4dd0RI+vdXIbXDQWDv8GTaPhwRVwjbpVdwZms8rY/2xjy/ELdAmvz/dPRONWA7orr7Pbgh0kA8nAoiI52PP56/x3eC1z5LzWY0dbF+9RDyA/v3oGKazJJAPJwEJyqDsZ3HzzzaiqWuK2YMECABYsWGDTaLI8Z8+ePRQWFnLs2DG7Gk2OUlfytzH8bXDz1sZ3il9Y7uzVmYFOp/DuXV3w9TSwNzmDeWuPlv+kalIntwU7SQaSgYUjc5CGk506hfnR3FelTZC3s4viVEajkXXr1pXam4+rkAwkAwvJQTJwtjqbf/1w6P+8dj/2Zci7WOaszsggrL4Xb47uCMC8dUfZnXSp2l67LHV2W7CDZCAZWDg6B2k42enZoa2Z0tFE+9CaeYqJEEIIUadET4DAtpB3Ada87uzSlHBr18b8rUsYJrPKM4vjyS107S+qQtRl0h25EEIIIWouvRuMnAMLYmDXArjhPmjSw9mlsvHGrR3ZefIipy7k8cTXu+jU2B+zql1foaJdD2XQ67jthsZEyqn+QtRa0nCqhLL6inc1koNkAJKBheQgGThbnc4/og90uRf2LoTfnoFH14G+ZH2dlYG/txuz7+rC2P9sY1NiOpsS00ud7/M/TjB1aBse7dcCva7qLtiv09tCBUkGkoGFI3OQXvWEEEJUG/kMLp3kUgE5aTCvOxRkwqAZ0PeZa/ay5wzL9qWw4+RFdIqCooBO0QbkVRQ4lJLNxiNpAPSKaMC7d3UhvIFrXy8tRE1gz+evNJzsZDabSU9Pp1GjRqWOgOwqJAfJACQDC8mh4hlIA6F0sm+qoB3/hWVTtfsR/WDomxDWFaj5GaiqypKdybz+60Fyi0z4uOuZMaoDf+/RxKHdRdf0HKqDZCAZWFQkB+mOvAqZTCa2bNmCyWRydlGcSnKQDEAysJAcJANnc5n8uz8IN/0T9B5wchN82h9+fAwykmt8BoqiMKZnU1ZMvomeEQHkFpl47od9PPrlLtJzCh32OjU9h+ogGUgGFo7OQRpOQgghhKgddDoYOB0m7YTOY7Rp+xbDh93RrX0dgykPzCbISYXzf8Lx9bD/e9j6MSRtc2rRLZo29GbRY9FMG9EWd72O1YfOM2DWeqb9sI+txy9gNrvUiUBC1Cpy1ZgQQgghapf6TeH2T+HGJ2HVy3ByE/otHzBcMaDb9yRQSuND7w7jl0F4r2ovbomi6BSe6N+S/m0CmbpkL4dSsli0I5lFO5IJ8/fk1hsac9sNjWkTLD3wCVGTyBEnOymKgq+vr0PPR66NJAfJACQDC8lBMnA2l80/7AYY9yvcsxi1YWv0qhEFFVDAuxEEttOuhQrpBKYiWDQWMs84u9RW7UL9WDapL98+eiNjeoTj62HgbGYB89cfY+h7G4l5fxPz1x/jZHpuhZfpstvCFSQDycDC0TlI5xBCCCGqjXwGl05ycQCzCS6dBA9f8G4IOv3lxwpz4PNhcP4AhHaFB1eAe83r0a6g2MTaw6ks3XOG9QmpFJsuf0VrF+pHTMcQRnQKoVWQHIkSwlGkc4gqZDabOXXqFGaz2dlFcSrJQTIAycBCcpAMnE3yBzMKp3IMmL0b2TaaADzqwd0LtQZVSjz8MhFq4O/Gnm56YjqF8tkDPdj+4mDeuq0j/Vo3Qq9TOJSSxbuxRxg8ZyND5mzggzWJ5BQaSyxDtgXJACQDC0fnIA0nO5lMJuLj46WXEslBMkAysJAcJANnk/wrkEFAM7jrK9AZ4MAP8MecshdmLNTmObGxagpbAQE+7oyNasZXD0ex86XBvHNHZwZEBuKmV0hMzWFO7BEGzF7Pj7tP23QoIduCZACSgYWjc5CGkxBCCCFcQ0QfiJml3V/zBhxebvt4fgb88R7M7QzfPwRf3grn9ld7Ma8W4OPOXT3D+eLBXuycPoTZf+9CRENv0rILmbpkL3d+HMeBM5nOLqYQdV6NaDh99NFHRERE4OnpSVRUFNu3by9z3s8++4x+/foREBBAQEAAgwcPvub8QgghhBBWPR6Cno8AKvz4KKQe0jqMWDUd3usIq1+FnHOg6EE1w/LnatRpff5ebtzZvQm/P3MTzw2PxNtdz+6kDEbN+4MXftzPhdwiZxdRiDrL6Q2nxYsXM3XqVGbMmMHu3bvp0qULw4YNIzU1tdT5169fzz333MO6devYsmUL4eHhDB06lDNnqqeXHEVRCAwMlF5KJAfJAMnAQnKQDJxN8rczg+Fva73tFeXAFzHwfmeI+xCKsrWe+EbPh4k7wOAFSXHaWFA1jIdBz1M3t2LtP27m1q5hqCp8uz2JoXP/YMV5H1b8eZ7zWQXOLqZTyPtBMrBwdA5O71UvKiqKnj17Mm/ePEC7iCs8PJxJkyYxbdq0cp9vMpkICAhg3rx5PPDAA+XOLz0XCSGE88hncOkkFyfIvQCfDYCMU9r/zfpCn6eh1RBtoF2AjbNg7ZtQL0QbdNej5vZmt+PkRWb8/CcHU7JspjcJ8KJHswC6NwvgxhYNaS1jQwlhw57PX6cOgFtUVMSuXbt44YUXrNN0Oh2DBw9my5YtFVpGXl4excXFNGjQoNTHCwsLKSwstP6flaV9oBQXF1NcXGx9Tb1ej8lksul1wzLdaDRiaV+aTCZOnDhBmzZtUFWVK9uder0enU5nXa6FwaDFbDQaKzTdzc0Ns9lscyGboigYDIYyp5dV9orU6cqylzX96jopikJiYiLNmzdHr7/ce1FtrpO960mn03HkyBGbDGp7nexdT4qikJCQQIsWLawZ1PY6VWY9FRYWcvToUVq2bIler68TdbJ3PRUUFHDs2DFrBmWV3cVGwKg2JpOJxMREWrdubfOZ7ErszsCnITzwM+z5GiJjoEn3kvNET4I938ClE1ojasjrji+4g/SMaMCvk/qycv9ZVu45zrFMlcPnsjl9KZ/Tl/L5Kf4sANEtGvLEzS25qXWjOns0Qt4PkoGFo3NwasMpPT0dk8lEcHCwzfTg4GAOHz5coWU8//zzhIWFMXjw4FIfnzlzJq+99lqJ6atWrcLbWxvDoWnTptxwww3s27ePpKQk6zyRkZG0bduW7du3k5aWZvP8Vq1asWnTJrKzs63ToqOjCQoKYtWqVTZfFgYMGICXlxfLl9tehBoTE0N+fj7r1q2zTjMYDIwcOZL09HSbxqOvry8DBw4kOTmZ+Ph46/TAwEB69+5NYmIiCQkJ1un21qlr1640a9aMjRs3VqhO/fr148iRIxw5cqTO1Mne9TR06NASGdT2Otm7nnr27EliYiKJiYl1pk6VWU87d+4kPT2do0eP1pk62bue1qxZg9FotGZQVp369u2LcDyz2UxCQoK14eqKKpVBg+Yw6OWyH3fzhBH/goV3wZZ/Q9f7ILCNYwpcBfQ6haHtgzCe3MmcsTEUmCA+OYOdJy+x69Qlth6/wJa/bu1D/Xi8fwtGdgrFoHf6lRsOJe8HycDC0Tk49VS9s2fP0rhxY+Li4oiOjrZOf+6559iwYQPbtm275vPffvtt3nnnHdavX0/nzp1Lnae0I07h4eGkp6dbD8fZ88trcXExsbGxxMTEoCiKy/yafHWdVFVlxYoVDBkyBDc3tzpRJ3vXE8Dy5cttMqjtdbJ3PamqWiKD2l6nyqyn/Px8Vq1aZc2hLtTJ3vWUl5dHbGysNYOyyp6Xl0f9+vXllLSrXO+pesXFxSxfvpyYmBibz2RXUqUZfHMXJP4OLQbA/UuhBh+puVYOZzLy+c+m4yzankx+sfY5Ed7Ai0f7teDv3cPxcq8bX7Dl/SAZWFQkh1pzql6jRo3Q6/WcP3/eZvr58+cJCQm55nNnz57N22+/zerVq8tsNAF4eHjg4eFRYrqbm1uJAC2n2FzN8gWgotPLWjH2TNfpdOh0JX8BKmt6WWWvqjpZvviVlmNp81+r7DWlTvZOv1YGtbVOYN96ulYGtbVOUPn1dHUOdaFOFZ1uKePVGVxd9rp6apCo44bPhOPrtNvh36DdqNLny0nVeuDzDS79cSdrXN+LGaM68PTA1ny19RQL4k6SfDGfV37+k3dWJhDTKYTbuzWhV0QDdDp5rwpxNacem3V3d6d79+6sWbPGOs1sNrNmzRqbI1BXe+edd3jjjTdYuXIlPXr0qI6iWul0Opo2bVrqFxNXIjlIBiAZWEgOkoGzSf5VnEHDltD7ae3+yhehKO/yY6oKJzbB4vvh3bbwwQ1wZtf1v2ZaAszrCZvetetpFckhwMedpwe1ZvPzA3n91g40beBNTqGRJTtPc/enW7lp1jrmrErgRHru9dbCKeT9IBlYODoHp/eqt3jxYsaNG8cnn3xCr169mDt3LkuWLOHw4cMEBwfzwAMP0LhxY2bOnAnAv/71L1555RUWLlxInz59rMupV68e9erVK/f1pOciIYRwHvkMLp3kUgsU5cK8XpB1Gvo/rzWk9i2G7Z9B2iHbeX2C4JHVENCs8q/33Xj4c6l2f9AM6De18ssqh9mssvPUJX7cfZpl+1LILrx8iu0NTeszpH0wg9oG0ya4nhw1FnWOPZ+/Tm+GjhkzhtmzZ/PKK6/QtWtX4uPjWblypbXDiKSkJFJSUqzzz58/n6KiIu68805CQ0Ott9mzZ1dLeU0mE3v27LG5hsAVSQ6SAUgGFpKDZOBskn81ZODuA8Pe0u7/8R7MaQfLpmqNJjdvbWDdh1dDcEfITdU6lMjPqNxrXToJB3++/P+a12DbpxV6amVy0OkUejVvwNt3dGbH9MF8cM8N3BwZiE6BPUkZvLMygWFzN9L3X+uY/tN+1h1OpaC45m5r8n6QDCwcnYPTG04AEydO5NSpUxQWFrJt2zaioqKsj61fv54FCxZY/z958qS1G/Arb6+++mq1lNVsNpOUlGRzMbUrkhwkA5AMLCQHycDZJP9qyqD9rdC8P5iKoDALGrTUBtOdeghueQ/Ce8K9S8A3FNIOw5L7wVhk/+ts/RhUM7QcqB3dAljxT63r9HJcbw6ebnr+1iWMBQ/2YusLg3jj1g4MiAzEw6DjTEY+X29N4sEFO+j6+iqe+34vF3MrUb8qJu8HycDC0Tk4tXMIIYQQQohaQ1Hg9k9h2ycQ0QdaDLw8WK6Ff2Ot8fTFCDixEX6bArd+VPGe+PIvwe4vtfvRE7XGU1EubJkHv0wCNy/oeIdDq1WWID9P7o+O4P7oCPKLTMQdS2ft4VTWHU7lbGYBS3aeZs2hVF4Z1Z6/dQmT0/hEnScNJyGE05lMphJddNcmxcXFGAwGCgoKXPa0CEsGJpPJpbu+FS7ANwQGz7j2PKGd4e8LYOEYiP8GAppD/39WbPm7FkBxLgR10BpNigJD39QaT7u+gB8f004NjBxxvTWxi5e7nkHtghnULhhV1a6Jmr70AAnns5m8KJ6f48/y5uiOhNX3qtZyCVGdpOFkJ51OR2RkpPRSIjlIBlx/Bqqqcu7cOTIyMhxbsGqmqirh4eEkJye77C+ulgyOHz9OQEAAISEhLpuFM8jnUQ3MoPUQiJmlXQe17k2to4jOd137OcYi7WgWQO+Jl49SKQqMnAPFeVqHFEvGwb2LoeWAEouojhwURaFnRAN+ndSXjzccY97ao6w9nMqQORt4fkRb7otqhk6nkFNoJOFcFgdTsjl4Nosj57Np2sCbqUPaEN7Au8rKV+O2BSeQDDSOzsHpvepVN+m5SIiaIyUlhYyMDIKCgvD29pYv2rWcqqrk5eWRmppK/fr1CQ0NLTFPbfkM/uijj5g1axbnzp2jS5cufPjhh/Tq1avM+efOncv8+fNJSkqiUaNG3HnnncycORNPT88KvV5tyUVUwqqXIe4D0Ltrg+dG9C173vhv4acntGukJu8Dg7vt4yYjfDdOG0vKzRse3wSNWlVt+SvgaGo2037Yz85TlwCIDPal0Gji5IW8Uuf3MOh48uaWPNG/JZ5udWPQXVF71ZoBcGsjo9HI9u3b6dWrV5kDQboCyUEygOvLwGQyWRtNDRs2rKISVg9VVcnNzcXHx8dlG3+WDBo0aABAamoqQUFBpQ6uW9MtXryYqVOn8vHHHxMVFcXcuXMZNmwYCQkJBAUFlZh/4cKFTJs2jc8//5zevXtz5MgRxo8fj6IozJkzp1rKLJ9HNTiDwa9Bximtl7xv74WHVkBwh5LzqSrEfajd7/VYyUYTgN4Ad34OX98BJzfB7y/C2CU2szgjh1ZBvix5PJpvtp3i7RWHSTifbX0s2M+D9qF+tAv1o1VQPb7fdZq4YxeYuzqR73ed5uVb2jO0fbBDPztr7LZQjSQDjaNzcN0kK0lVVdLS0nCxA3UlSA6SAVxfBpZrmry9q+50jeqiqipGoxFVVV264WTJwLJOi4uLa2XDac6cOTz66KM8+OCDAHz88ccsW7aMzz//nGnTppWYPy4ujj59+nDvvfcCEBERwT333MO2bduqrczyeVSDM9Dp4LZPICcNkuK0Rs/Dq6B+U9v5jq+D1D/BzQd6PFj28gwecMtc+HcUJP4Oiauh9WDrw87KQadTuD86gsHtg9mUmE7j+l60C/WjgY9tA/C2Gxqz4sA53vztIKcv5fP4V7u4qU0gM0a1J9Tfk6x8I5n5xdZbVn4x4Q286RkRUOHP1xq7LVQjyUDj6Byk4SSEcCpXbWjUZbV5nRYVFbFr1y5eeOEF6zSdTsfgwYPZsmVLqc/p3bs3X3/9tfVXzePHj7N8+XLuv//+Ml+nsLCQwsJC6/9ZWVmA1ti0/Kig0+nQ6/WYTCabrnQt0y0NVcvzLK6cDqDX69HpdCU6YLH8+mo0Gis03c3NDbPZbNMBiqIoGAyGMqeXVfaK1OnKslekTpb7qqpWuK7VVicMcOeXGL4ahZJ2CL66HeO4ZaheDS6XcfOHKICp61jMhnrwVx1KLbt/Mwy9HkfZ+hHqymkYwzeC3rZTliszqM711MjbwG1dQqzTry4LwIiOIfRv04h5axP5zx8n2XgkjUHvbuBaerdsyHNDW9MhzK/cOpWWQVVue1Dz3k+lZVCr66Sa0G98G2OLQajhN9pdp+JrvJ/s6ZxKGk5CCCHEX9LT0zGZTNZB2C2Cg4M5fPhwqc+59957SU9Pp2/fvtYjb0888QQvvvhima8zc+ZMXnvttRLTV61aZT1i17RpU2644Qb27dtHUlKSdZ7IyEjatm3L9u3bSUtLK7GMjRs3kp19+VSp6OhogoKCWLVqlc2XhQEDBuDl5cXy5cttnh8TE0N+fj7r1q2zTjMYDIwcOZL09HSbBqSvry8DBw4kOTmZ+Ph46/TAwEB69+5NYmIiCQkJ1un21qlr1640a9aswnUC7QtRbGxsjayTZ9DjDMr7F4YLieR+NpJNzZ/DpPfALz+JAcfXgqJjQ2F7sq9YJ2Wup4FT0e9dhO5CIoe/epbjQcMwGAwMHToUwCaDmraeLHVqW5TI853gx5M6DmZoX/b1CnjqVbwN4KUHT4PCyRwdcccuMHr+Bbo3MjMy3ExEYNl16tmzZ4kMqqtONeX91LJlyxIZ1OY69fQ4QdjW9yiMX8rq1q9XuE6WMlpyKK1OeXmlX4tXGukcwk5ms5nk5GTCw8NduqcSyUEygOvLoKCggBMnTtC8efMKX0BfU6mqSlFREe7u7pU62hIREcGUKVOYMmWK4wtXTa7MoLCwsMx1W9M7QTh79iyNGzcmLi6O6Oho6/TnnnuODRs2lHr63fr167n77rt58803iYqK4ujRo0yePJlHH32Ul19+udTXKe2IU3h4OOnp6dZc7PmF3Gw2c/bsWZo1a4bZbK59vyY74Fd/s9lMSkoK4eHhJQa7rFF1unQM3RfDIf8S5paDMf39K/TLn0G3bxG0H03xbf+pUNkNBgPs+h/Kb5NRPf0xPrkdvBui1+tJSkoiNDTU+rlc6Tol74ADP6IGd0Rt0gtdg+boDYYqOTqTkVeMl4cb3u76EsM6nMsuZtbvh/llb4q23vQK90U15elBbfD3MpSok06n49SpU4SFhVkzcLUjToqicPLkSRo3bmzNoDbXSR/7Errtn6CiYPzHMfD0q1CdCgsLOXPmjDWH0sqelZVFo0aNKrRfkoaTEMIpamPDqbxG0YwZM3j11VftXm5aWho+Pj514novuPa6remfwUVFRXh7e/P9998zevRo6/Rx48aRkZHBzz//XOI5/fr148Ybb2TWrFnWaV9//TWPPfYYOTk5FfpRoabnIhwseQf8bxQY86HdKEhYCeZieGQNNOlR8eWYTfBJfzi/H3o8DLc4sDOS4+vh23u0LtAtfIIgvBeER2m3xt1KnCJYKRlJsPVj6PO0Nk5WGQ6cyWTmikNsPnoBAF9PA/f2asrdvZrSvJHP9ZdD1FyfD4ekv45k3fcjtBrksEXb8/nrmj+TXwej0cjatWtLtLJdjeQgGYDrZZCSkmK9zZ07Fz8/P1JSUjh79ixHjhzhH//4h3VeyylbFREYGFjrG02qqpKVlVXrL0R2d3ene/furFmzxjrNbDazZs0amyNQV8rLyyvROLJc21Fdebjae7E0tSqD8J5w1/9A0cOhX7VGU9No+xpNADo9jHhbu7/rCzh3oPQcCjJh+XPw2SA4sbH85R5ZBd/cpTWaGneHJr207tRzU7Wu0GNfhs+Hwhcx2thT1+vXKbD1I1j75jVn69jYn68fjuJ/D/WiXagf2QVGPtl4nAGz13PPp1v5de9ZCo2m2rUtVJE6lYHZDCn7Lv+fvL3CT3V0DtJwspOqqmRnZ9f6LwfXS3KQDMCxGaiqSl6R0Sm3ipY/JCTEevP390dRFEJCQggODiYhIQE/Pz9WrFhB9+7d8fDw4I8//uDYsWPceuutBAcHU69ePXr27Mnq1attlhsREcHcuXOt/yuKwn/+8x9uu+02vL29ad26Nb/88st1Z1yVVFUtcYpYbTV16lQ+++wz/ve//3Ho0CGefPJJcnNzrb3sPfDAAzadR4waNYr58+ezaNEiTpw4QWxsLC+//DKjRo2qtl4F5fOoFmbQZhj87cPL/0dPrNxyIvpC+9GgmmHlNFSz+XIOqgoHf4F5vWD7J3BmJ/zvb9rYUsbC0pd36DdYdC+YCiFyJDy4Ah6JhWnJ8NAqGPIGtL1F6/3v9HbY+E7lym1xNh6O/fVDxeHfwHTtC/UVRaF/m0CWTerLZw/0YGDbIBQFthy/wKRv9xA9cy0zVxzmYEo2eYV1oNFQSbXu/XAtF45Cce7l/09XvOHk6BykcwghRI2QX2yi/Su/O+W1D74+DG93x3wcTps2jdmzZ9OiRQsCAgJITk4mJiaGt956Cw8PD7788ktGjRpFQkICTZs2LXM5r732Gu+88w6zZs3iww8/ZOzYsZw6dco6TpKoOmPGjCEtLY1XXnmFc+fO0bVrV1auXGntMCIpKcnmCNP06dNRFIXp06dz5swZAgMDGTVqFG+99ZazqiBqixvGakdyMpMhMqbyyxnyOhxZCSc3oSQsBxTIOgOrXoCEvy6Cb9BSO7Vu/3fagLzH1sEdn0FQu8vLOfAD/PAoqCbocBvc/tnlU/HcPKFplHYD+PMnbTDeTXO0sjfuVrmyb557+X7+Je2IWAVOw9LpFIa0D2ZI+2DOZOSzeEcyS3Ykcy6rgP9uPgUYmL1/DaH+njRv5GO9RTT0wc/LDQ+DDg83HR4GvXbfoEOvUyg2qRSbzNZbkVHFrKq0Cqong/U6S0q89tcrQNtGTu/SjkI54fpyaTgJIYQDvf766wwZMsT6f4MGDejSpYv1/zfeeIOlS5fyyy+/MHFi2b8wjx8/nnvuuQeA//u//+ODDz5g+/btDB8+vOoKL6wmTpxY5vpZv369zf8Gg4EZM2YwY8aMaiiZqHM6//36lxHQDHpPgo2z0K95hRY+fTB88iQU5YLOAH2fgX7Pao2f9qPhl0nadVGf9NcaXb0eg32L4OcJ2pGrznfDrR9pA+6WpcNoOHg7/Pkj/PQUPL5BG2PKHheOaQMDAzTrA6c2w8Gf7L5+pXF9L6YOacPTA1uxPiGNhdtOsfVYKnlGhZTMAlIyC4g7dsG+sl0lwNuNe3o15f7oZoT6e13XsoSdzsZrfzvcBnsXQ2EmpCfYNvqriTSc7KTX64mOjq6Vgzo6kuQgGYBjM/By03Pw9WEOKFXlXvt6KIpi7QShRw/baxRycnJ49dVXWbZsGSkpKRiNRvLz8226WS1N586drfd9fHzw8/MjNTX1uspZlRRFwcfHp1aP4VSbyeeRZEDfZ2DPNygZp+iUcUqb1qQXjHofgttfnq/dLdq1VD9PgKOrYeXzsPdbSNkLqNDtAbjl/Yr9mh8zG05ugrRDsP5tGGznjwdxH2gNtdbDIPop+PJW7VTBke9du9FWBoNex+D2wQxsG0h6ejp6bz9OXcznRFouJy/kcjw9l6QLeeQWGSksNlNoNFNoNFFoNFNkvNyDm7tBh5tOwc2gw02vo8ho5lJeMf9ef4xPNh5neMcQHuwdQfdmFR+Ut7rVqfdDyl7tb+PukJ6obXPJ2yrUcHJ0DtJwspNOpyMoKMjZxXA6yUEyAMdmoCiKw06Xq26WblVBa+Rc6dlnnyU2NpbZs2fTqlUrvLy8uPPOOykquvYF1W5utj1VKYpSoovlmkRRlBJlFtVHPo8kA9x9YOgb8MPD4OGnNWK6P1R6A8g3BMZ+Dzv+A6umXz4VqtdjMPxfFT8Fyqch3PIeLL5PO+Wu7S3QpHvFnpuVAvELtfv9pkLjHuDdEPIuaF+MWw6o2HJKceW20LCeJ92aBpT7HLNZOyVPr1NKNIaMJjOrD6WyIO4EW49fZNm+FJbtS6FjYz8eiI4gplMo9Txq1v6rzrwfzObLDafQrnDx+F8Npx3QfXy5T3d0DtI5hJ2Ki4tZtmyZXaMM10WSg2QAkoGF2WwmJyen1Mc2b97M+PHjue222+jUqRMhISGcPHmyegtYDcxmMxkZGTW6cVeXyXtRMgCg050YH4wltt07FHcdd+0GkKJAr0fh8Y1aJxCDZsCId+y/bqTdKOh4p3bk6KcnobigYs/b+m8wFWm9CTa9UTvC1PYW7bGDP9lXhqtUZlvQ6RQMel2pR5AMeh3DO4aw6LFoVkzux5ge4XgYdBw4k8Vz3++j+xuxPPXNLlYeSKGg2FTK0qtfnXk/XDwORdlg8ITAtlo3+FDhDiIcnYM0nCqhTnTt6ACSg2QAkkF5WrduzY8//kh8fDx79+7l3nvvlcaFqBLyXpQMANSwG8hT7BjeIDAS7lmoHfWp7GlnMbO0MZ7SE2D9/5U/f/4l2Pm5dr/vM5endxit/T30G5iub11W1bbQLtSPf93ZmS0vDOK54ZG0aORDodHM8v3neOLr3fR4czVTl8SzPiGVYlPFPutVVWXXqUv83/JD/LDrtMN6gKsT7wfL0dDgjlrjuklP7f/0I5B3sUKLcGQONeu4ohBC1DFz5szhoYceonfv3jRq1Ijnn3+erKwsZxdLCCEcx7sBjJqrdWMe9yG0HaWNVVWWHf+FohwI6gCth16eHtEPvBpAXrrWUUSL/vaXRVVRzuymfu5x+59rhwY+7jx1cyue7N+SP89m8eves/y69yxnMwv4cfcZftx9hnoeBqJbNuSmNoH0bx1I04a2Ddqjqdn8tOcsP+89Q/LFfOv02IPnefuOTtT3dq/SOtQKZ/dof8O6an+9G0DDVloX5ad3QpuhZT61KkjDSQghKmH8+PGMHz/e+n/fvn0xmUwlBkKNiIhg7dq1NtMmTJhg8//Vp+6V9mtjRkbGdZVXCCGqVNuR0HkM7FusnbL3+AbtuqurFeXB1vna/b7P2B7l0rtpy9nzlXa6nj0Np4Is7bV3foEh9U9uQsF0sgu0Hnhd1SqPoih0bOxPx8b+PD+8LbuTLvHL3rMs359Cek4RsQfPE3vwPAARDb25qU0gwX6eLNuXwsGUyz+i+bjriW7ZiA1HUln55znikzN4b0xXols2rNLy13hXXt9kER71V8Npe7U3nBS1ToyMVXFZWVn4+/uTmZmJn5+f3c+3DKTl6+tbY3tSqQ6Sg2QA15dBQUEBJ06coHnz5tYe6Wory+CvOl3p58e7giszKCwsLHPdXu9ncF0l+6brJxlonJpD3kX4942Qcx78w2HQK9r1T1f+oLT9M1j+LNRvCpP2lOw97+hq+PoO8AmEfySArpze0FL2akew9n9vO0gqoPo1QXkqDjz9HVTBijObVf48m8XGxDQ2HElj96lLGM22X7kNOm0w31tvaMyQtkF4Ja3nSFEDnliRxfH0XBQFnrq5JVMGt8FNb9/VNXXi/aCq8HYzrfvxxzdB6F+9ze78An6bAs37w7hrDw5fkRzs+fyVI06V4OUl/feD5ACSAUgGFrV2x+RAkoFzyXtRMrBwWg7eDeDv/9N69stMhh8f1TqBGPoWRPQBUzFs/kCbt/fTpXc53ry/NtBpbhqcioPm/Up/rbQEbfyoMzsvT2sUCT0eQm07EhbcgpJxElY8D7d97PCqlkenU+jUxJ9OTfyZMKAV2QXFbDl2gY2JaaRkFHBz2yBGdgqlgc9fp+PFzYNVL9HGuxG/Pr6d12PPsHhnMh+tO8YfRy/wwd1dadawlCN411Dr3w8Xj2uNJr2Hbdfj4b20v2d2gdlUbuPakTlI5xB2MhqNLF++vG5ccHcdJAfJACQDC1VVycrKctgFvbWRZOBc8l6UDCycnkOzaJi0Szva5O6rXaOyIAYWjYU/5kJmknY06Yb7Sn++5XQ9KLt3vZw0+PpOrdGkc9OOao1fDhO2wY1PYPQJYVPg/aiKThuj6s8yllONfD3dGNohhDdHd+K/43ty/43NLjea/lwKq17S7uel4xP3Dv+6szMf3dsNP08De5MzGDJnI33/tZbhczfy94/jGP/FdiYu3M0LP+7j3+uPsu34BZse/Zy+HTiCtWOIDtp2YRHYVtu2inIg9eA1F+HoHOSIkxBCCCGEcBw3L+j3D7jhAVg/E3YtgMO/aTeAG5/U5ilL+9Gw52s4+MtfXaRfcUTBWKiNG5WZBA1aaA0mv9ASi7hUrzXm6Mno497TTutqeqM2fpW9CnPg0K8QdgMEtbX/+eVJ2go/Pq7db3EzHF+vnc7YbRwjO7ena9P6PLMonu0nL3L6Uv61loSbXqFDmD89mgXQtYkfWdceLrDms1zfZOkYwkKn1wZxPr4OkrdDSKdqK5I0nIQQQgghhOPVC4Rb5kDU4xA7A46s0E7D6/HwtZ/XvL92XVJuKiRtgYi+2nRVhV8nQ/JW8PCHe5eU2miyMN/0T/TH18C5ffDzBG3Q34qeUlyUqw0QvPl9bVBeFK3zi5unQYPmFVtGedIT4du7wVSojaU15itY8oDWwFzxHIz7lcb1vVj8+I0cT88lM7+Y3EIjOQVGsguN5BYayco3cvhcFjtPXSItu5D45AzikzP+egEDP6Ru45bOYYzoFErj+tc+ZS2vyMjZjHwiGvpgsPOaqipxNl77e2XHEBbhvS43nHqWsz05kDSchBBCCCFE1QmMhHsXwbn94OELXvWvPb/BXRsMN/4bOPjz5YbT5ve1U+8UPdy1ABq1vvZy9O5w+2fwyU1apxM7/ws9H7n2c4rztTGm/nhPu84KoF6w1uHFvkVw4Hvo9gDc9E/wC6tI7UuXk6p1gpF/CRp3hzv+ox1JGfZ/WllPbtJOVexwG4qi0DKw3jUXp6oqyRfz2XnqIjtPXWLniYskpmYTn5xJfHImby47RNfw+ozsFMqITiE08HHn4Nks9p/J1G6nMzmWloNZhRA/T/7eowl39QgnvIEdY4I5kqpe0aNel5KPN/nrOqcKDoTrKNKrnp1UVcVoNGIwGFz6QmjJQTKA68ugrvWqp6oqiqK49LZgyUB61bOf7Juun2SgqTM5HPkdFt4F9UJg6iHtaNWisYAKMbOh16NlPrVEBlvnw8pp4Oat9c7WqFXJJxXlad2gb5oDOee0aQER0P956HQXnN8Pa9/UGjUABk+tDH2eAR87uwwvyoUFt8DZ3dprPLxaOzpnsW4mbHgb/JrAxO2ld+teDlVVOXspl9hDaSw/cI4dJy9y5Td+nQLmUloA7nodRVcM3Nu3VSPu7hXOkPbBeBjK6eHQkS6egA+6ao3fF85ojekr5WfAv5pp9/95DHwalbqYirwfpFe9Kpafn4+vr6+zi+F0koNkAJKBhaXR4MokA+eS96JkYFEncmgxQDsdL+ecdqQodgagakeMrtFosrDJoNfjkLACTmyApY/BbZ9A6iGtY4Hzf2p/Lx4H9a8Gg3+4dkSp672XOyUIuwHu+wFOboa1b2inEMZ9CDsXQPQE7eZZgR89zCb44RGt0eQVAGN/sG00AfSdAvELteu4/ngPBk6vaGo2fA1mxvWOYHyf5qRmFbDyz3Ms25fC9pMXMasQ5OtBp7/GoOrcxJ9Ojf3x93Yj9uB5Fu9IZlNiOn8c1W4NfNyJat4Ak1nFaFYpNpn/uqmYVZXWQfXo1jSAbs0CaBVYD53uOvcFlqNNQe1LNppAO2oZ2BbSDsPpHRA5osxFOfL9IA0nOxmNRtatW0dMTAxubm7lP6GOkhwkA5AMLCzjRPj5+blsw+HKDET1k/eiZGBRZ3IwuEPbGO3UvOXPatOa94fhb5f71BIZ6HQwej7Mj9a6sJ7Xo/Qn+jeFvpO1Ti1K+7IOWrfqD66Ao2tg7evaF/wNb8P2T7QBfXs+Cu6lnN5WkKl1drHnK0jepnWxfc+i0o9+uXnBsLdgyf1a9+1dx9p9XdXVGQT5efJAdAQPREdwMbcIo8lMkF/pZ3vc0jmMWzqHkXwxjyU7k1myM5nzWYWsOHCuzNfbk5TBkp2nAfDzNNC1aQDdmwbQLtSXBj7u1Pd2o763O/5ebhUbk8rSo97VHUNcqUlPreGUvK3MhpOj3w/ScBJCiGp0880307VrV+bOnQtAREQEU6ZMYcqUKWU+R1EUli5dyujRo6/rtR21HCGEqBbtb9UaTgANW8Fd/7Ptltoe/o1h1Pvw/UPa6V+BkRDUAYLba0c1gjto1zJV5McvRYHWg6HlQDj0C6x7C9KPQOwrsOXf0P+fWuMLtFP79i3WjniZCrVpOje44zOtp7+ytBt1uZe931+Ee76tXL1LYe0GvRzhDbz5x9BIJg9qzaaj6ZxKz8XNoMNNp8PNoGDQ6XDT6zCrKgfOZLLr1CX2nc4kq8DIxiNpbDySVupyfT0M1PdxI9TPiyYNvAgP8Ca8gTdNArwIb+BNPXcD+hM7qQfE5YezfvkhzlzKJy27kJ7NA3jsppb4e7lpHUTs+QqSdzgsm/JIw0kIISpo1KhRFBcXs3LlyhKPxcXFMXLkSPbu3Uvnzp0rvMwdO3bg42P/+evX8uqrr/LTTz8RHx9vMz0lJYWAgACHvpYQQlSZlgO10+aK8+Cexdqpbdejw23aMt3rlTtoaoXodNBhtNaRxf4l2rVJmUmw7B/amFVFOVrnDxaBbbWe+Tr9HeqHX3vZiqJ1xT6/NyQsh8TVWmPNCQwKDDj9CRRma0fCSmm8xnTSejcsNplJOJfNrlOX2J10iRPpuWTkFZORV0RWgTaWUnah1itg8sV8tp8s7RVV9njsAQVm7vFgv3rc+sj2kxf5ZlsSEwe04oHWPXAH7bRHU3HlG9V2kIZTJRgMEhtIDiAZgGtl8PDDD3PHHXdw+vRpmjRpYvPYwoUL6dGjh12NJoDAwMDyZ3KQkJBKjGEiag1Xei+WRTLQ1JkcDB7w1Bbt2iNPf/ueWlYGdi6nQvQG7XqojnfA7i9hwzuQmaw9Vi8EOt2pNZhCOlW8O3TQjopFPQFb5sHK5yFiM7hVvDMlh20Ha16DzXO1+94N4ebny5zVTa+j41/XTY3rHWHzmMmskpmvNaIu5hZxJiOf05fySb6YR/KlPJIv5nM2I59gcxoBSg7FGGjarjs9A+rTOMALb3c9//3jBEdTc3hz2SEW+Huw1s0P9+IsOH9Auw6tKnMAUF1MZmamCqiZmZnOLooQLi0/P189ePCgmp+fr00wm1W1MMc5N7O5QmUuLi5Wg4OD1TfeeMNmenZ2tlqvXj31zTffVO+++241LCxM9fLyUjt27KguXLjQZt7+/furkydPtv7frFkz9b333rP+f+TIEbVfv36qh4eH2q5dO3XVqlUqoC5dutQ6z3PPPae2bt1a9fLyUps3b65Onz5dLSoqUlVVVb/44gsVsLl98cUXqqqqJZazb98+dcCAAaqnp6faoEED9dFHH1Wzs7Otj48bN0699dZb1VmzZqkhISFqgwYN1Keeesr6WmUpsW6vIJ/BpZNchKhDCnNUdd93qnp0raqajNe3rPxMVX2nlarO8FPVd9ur6tZPVLUozzHlrIjdX2mvbbm9GqCqp3dV2csZTWa1YO9S7bXm9y3xeLHRpC7afkrt9Vas2uz539R10/up6gw/9cgvs1VzBfflV7Pn87eO/CRRfcxmM+np6TRq1AidrgYMDuYkkoNkAA7OoDgP/u86xsS4Hi+erVB3rwaDgQceeIAFCxbw0ksvWTuCWLJkCSaTibFjx/L999/z/PPP4+fnx7Jly7j//vtp2bIlvXr1Knf5ZrOZ22+/neDgYLZt20ZmZmap1z75+vqyYMECwsLC2L9/P48++ii+vr4899xzjBkzhgMHDrBy5UpWr9a6zfX3L/kLa25uLsOGDSM6OpodO3aQmprKI488wsSJE1mwYIF1vnXr1hEaGsq6des4evQoY8aMoWvXrjz6qG2vVuoVXb6K6iefR5KBheRQAzJw99GOMjmCpx/c/iksfQKyTsOKf8Km2dB7EnR/EDxKH9/JIRmc/AN+naLdv+mfcOEo/LkUlj4Oj2/UOrFwML1OQZ+2T/unlI4hDHodY3o25W9dGvNF3AkOrIvkZvby5/bVbKh/G4/0a2Ezv6O3Bdd8R10Hk8nEli1bMJlMzi6KU0kOkgG4ZgYPPfQQx44dY8OGDdZpCxYsYNSoUTRt2pRnn32Wrl270qJFCyZNmsTw4cNZsmRJhZa9evVqDh8+zJdffkmXLl246aab+L//+78S802fPp3evXsTERHBqFGjePbZZ62v4eXlRb169TAYDISEhBASEoKXV8md28KFCykoKODLL7+kY8eODBw4kHnz5vHVV19x/vx563wBAQHMmzePtm3bcssttzBy5EjWrFlTYnmqqpKbm4vqWkMD1hiu+F68mmSgkRzqYAYtB8DkvTDyXe2ar5zzsGo6zO0EG2dB3sUSTyk1g+J8OLIKVkyDDbO0sZDKcuEYLL4PzMXatWE3vwgj52inHqYfgdWvOb6eFmfjtb+hXcucxctdz1M3t2LcXXcB0EN/lFFdSv746uhtQX4aFELUDG7e2pEfZ712BbVt25bevXvz+eefc/PNN3P06FE2bdrEr7/+islk4q233mLJkiWcOXOGoqIiCgsL8fau2PIPHTpEeHg4YWGXP/yjo6NLzLd48WI++OADjh07Rk5ODkaj0e5uwA8dOkSXLl1sOqbo06cPZrOZhIQEgoODAejQoQN6/eWLqENDQ9m/f79dryWEEOI6uXlqY1h1Gwd7F8Gmd+HSCW1Q3rVvQoOW0KQHNO4BjbtDw0jteRlJcGItJK6CExvBWHB5mVv/rQ3w2+Mh2+7X8y/BwjHa37BuWlfuOh14N4Bb58E3d8K2+VoX4C36V6z8qqqNnXViAxzfoF2T1HIgDHnNttMPVb08htM1Gk4Wvi2jAIUmpIKSCVT8GrDKkIaTEKJmUJRKjY7uDA8//DCTJk3io48+4osvvqBly5b06dOH2bNn8/777zN37lw6deqEj48PU6ZMoaioyGGvvWXLFsaOHctrr73GsGHD8Pf3Z9GiRbz77rsOe40rXT3uhaIomM3mMuYWQghRpfRu0O1+6HKPdtrc5vfh/H64eEy77VsMgEHvzhC9H2570m2f79cEWg3Sxj5KO6x1OrH9Exj8mtYFutkI342HC4ng11jrBv3KU/JaD9EaWjs/h5+egic3a4PRlib7PBxZqTWWTmyE3Ku6J9/9P+3xmNnQ/m/atKwzkJcOOoPWRXx5PP207uRT/4TT27U6VCFpONlJURR8fX1ddpBLC8lBMgDXzeCuu+5i8uTJLFy4kC+//JInnngCvV5PXFwct956K/fddx+gnVt95MgR2rdvX6HltmvXjuTkZFJSUggN1bp23bp1q808cXFxNGvWjJdeesk67dSpUzbzuLu7l3taQrt27ViwYAG5ubnWo06bN29Gp9MRGRlZofJeSVEUdDqdy20LNYWrvhevJBloJAcXyUBvgM5/1255F7WBfc/sgtM74cwulPyLeJvSURU9SngUtBkKrYdqjQxFAZNRGwNp3Vtw8bg22G74jeAXpo0d5eajDdDrW0pvrEPegGPrtCNeK6fBbR/bPn7ugNYT4P7vtVP9LAxe0CxaG8i4QXPtSFn6Ee21243SGlCW0/QC21W8B8HOf4fM3tppjFdx9LYgDSc7GQwGBg4c6OxiOJ3kIBmA62ZQr149xowZwwsvvEBWVhYPPvggfn5+tG7dmu+//564uDgCAgKYM2cO58+fr3DDafDgwbRp04Zx48Yxa9YssrKybBpIAK1btyYpKYlFixbRs2dPli1bxtKlS23miYiI4MSJE8THx9OkSRN8fX3x8PCwmWfs2LHMmDGDcePG8eqrr5KWlsakSZO4//77rafp2UNRFLtPFxSO46rvxStJBhrJwQUz8G6gHQlqPUT7X1W1Rs2lkyhhN5Q+/pXeAD0e1Dqx2PwBxH0IyZYf6hS44z8QWsbwGh714LZP4Ivh2gDFkTFaw+foGtjyodbwsgjrBq0Ga6f0NempdTFv0XqYdo3W5rlw6FftqFRwx7+e16Xi9e/7TJkPOXpbkM4h7GQ2mzl16pTLn6oiOUgG4NoZPPzww1y6dIlhw4YRGhpKYWEhL730Et26dWPYsGHcfPPNhISEMHr06AovU6fTsXTpUvLz8+nVqxePPPIIb731ls08f/vb33jmmWeYOHEiXbt2JS4ujpdfftlmnjvuuIPhw4czYMAAAgMD+fbbkiPOe3t78/vvv3Px4kV69uzJnXfeyaBBg5g3b16l8lBVlcLCQukcwklc+b1oIRloJAfJAEXBXD+CU4aWmD3KGbfKwxcGvgRP74au92mDA494B9rGXPt5TaOgzxTt/q+T4d/R8M0dWqNJ0UOH2+GRtfDYOm35EX1tG02gHVEa9DI8tl67nqkgE05t1h6rwPVNFeHobUFRXWwvl5WVhb+/P5mZmZX6dbS4uJjly5cTExNT4tx/VyI5SAZwfRkUFBRw4sQJmjdvjqdn1V7MWdXMZjNZWVn4+fm5dPe/lgyKiorKXLfX+xlcV8m+6fpJBhrJQTKA68hAVSs+SK+xCD4bqF1jBeDuC90egKjHIaCZfQU2GbXOKta9pXVg8WRcxa5xKkdFcrDn81dO1RNCCCGEEEJUvNEEWk98d/0PVs+AJr2g+zjwLOcIV1n0BujzNHS8HbLOOqTRVBVc86dRIYQQ4ho++ugjIiIi8PT0JCoqiu3bt19z/oyMDCZMmEBoaCgeHh60adOG5cuXV1NphRDCSRq2hDFfa42eyjaaruTfBMLLHzDeWeSIk50URSEwMLBu99RSAZKDZACSgYWiKBgMBpfOoS5lsHjxYqZOncrHH39MVFQUc+fOZdiwYSQkJBAUFFRi/qKiIoYMGUJQUBDff/89jRs35tSpU9SvX7/ayizvRcnAQnKQDEAysHB0DnKNkxDCKerSNU7C1rXWbW34DI6KiqJnz57WjjLMZjPh4eFMmjSJadOmlZj/448/ZtasWRw+fLjC1xIUFhZSWFho/T8rK4vw8HDS09Otueh0OvR6PSaTyebCZst0o9Fo0xmHXq9Hp9OVOb24+IpugdF6mwIwGo0Vmu7m5obZbLbp6t7SYC5rellllzpJnaROUqeaUqesrCwaNWok1zhVBZPJRGJiIq1bt0av1zu7OE4jOUgG4JgM6sJvN6qqUlBQgKenp8v+undlBrV5nRYVFbFr1y5eeOEF6zSdTsfgwYPZsmVLqc/55ZdfiI6OZsKECfz8888EBgZy77338vzzz5f5vpg5cyavvfZaiemrVq3C29sbgKZNm3LDDTewb98+kpKSrPNERkbStm1btm/fTlra5QElg4OD6dmzJxs3biQ7O9s6PTo6mqCgIFatWmXzZWHAgAF4eXmVOKUwJiaG/Px81q1bZ51mMBgYOXIk6enpNjn4+voycOBAkpOTiY+Pt04PDAykd+/eJCYmkpCQYJ1ub526du1Ks2bNKlyniIgIIiMj+f333+tMnexdT8OHDyc+Pp7Tp0/XmTrZu56ioqLYsmULFy5cqDN1snc9tW7dmnXr1pGXl1dn6lRV6+nKjMojR5zsJD21aCQHyQCuLwOTycSRI0cICgqiYcOGVVTC6iG96tlmcOnSJVJTU2nTpk2JhkNNP+J09uxZGjduTFxcHNHR0dbpzz33HBs2bGDbtm0lntO2bVtOnjzJ2LFjeeqppzh69ChPPfUUTz/9NDNmzCj1dRx9xKm4uJjY2FhiYmJQFMUlf022ZDBixIgSP2DU1jpdq+xlTQdYvnw5Q4YMsX4u1/Y62bueVFUtkUFtr5O968lsNpfIoLbXqTLrKS8vj9jYWGsOcsRJCFEr6fV66tevT2pqKqCNK1Rbj9aYzWaKioooKChw6YZTYWEhFy5cID09nfr167vMkViz2UxQUBCffvoper2e7t27c+bMGWbNmlVmw8nDw6PEoMSgffG4+kcIvV5fapaWLwAVnV7Wjxv2TNfpdKVu42VNL6vsVVknRVHqXJ3smW758lvatlRb6wT2radrZVBb6wT2rSdLo6O0DGprnaDy6+nqHMq6Xx5pOAkhnCYkJATA2niqrVRVJT8/Hy8vr1rb+LteV2YQEBBgXbe1TaNGjdDr9Zw/f95m+vnz58usU2hoKG5ubjY7+Xbt2nHu3DmKiopwd3ev0jILIYSoHtJwspNOp6Np06Yu+6uyheQgGcD1Z6AoCqGhoQQFBZU4RaA2MZlMJCQk0KxZM5c5ynI1SwatWrWq1Q0Fd3d3unfvzpo1axg9ejSg/XK7Zs0aJk6cWOpz+vTpw8KFCzGbzdb3wpEjRwgNDa22LOTzSDKwkBwkA5AMLBydg1zjJIQQotrUhs/gxYsXM27cOD755BN69erF3LlzWbJkCYcPHyY4OJgHHniAxo0bM3PmTACSk5Pp0KED48aNY9KkSSQmJvLQQw/x9NNP89JLL1XoNWtDLkIIURfZ8/lbI5qh9g40+N1339G2bVs8PT3p1KlTtQ4yaDKZ2LNnj80Fbq5IcpAMQDKwkBzqVgZjxoxh9uzZvPLKK3Tt2pX4+HhWrlxJcHAwAElJSaSkpFjnDw8P5/fff2fHjh107tyZp59+msmTJ5fadXlVqUv5V5ZkoJEcJAOQDCwcnYPTG06WgQZnzJjB7t276dKlC8OGDSvzmoe4uDjuueceHn74Yfbs2cPo0aMZPXo0Bw4cqJbyms1mkpKSbHr6cEWSg2QAkoGF5FD3Mpg4cSKnTp2isLCQbdu2ERUVZX1s/fr1LFiwwGb+6Ohotm7dSkFBAceOHePFF1+s1tM261r+lSEZaCQHyQAkAwtH5+D0htOcOXN49NFHefDBB2nfvj0ff/wx3t7efP7556XO//777zN8+HD++c9/0q5dO9544w26detmHahQCCGEEEIIIRzNqZ1DVGagwS1btjB16lSbacOGDeOnn34qdf6rx8rIzMwE4OLFi9aL0e0dKyMvL4+srCyXHSsDtB608vLyuHDhgk03jrW5TpUZK+PqDGp7nSozVsbVGdT2OlVmPeXn59vkUBfqVJmxMq7MoKyyWwYadLHLa8tlySMrK6tSz79y3+TK48q5egYgOYBkAJKBRUVysHzuVmS/5NSGU3p6OiaTyXreuEVwcDCHDx8u9Tnnzp0rdf5z586VOn9Zo7M3b968kqUWQghxvbKzs/H393d2MWoMy4j34eHhTi6JEEK4porsl+p8d+QvvPCCzREqs9nMxYsXadiwYaXGW7GM7p6cnOzSPR9JDpIBSAYWkkPFM1BVlezsbMLCwqqxdDVfWFgYycnJ+Pr6yr6pkiQDjeQgGYBkYFGRHOzZLzm14VSZgQZDQkLsmr+00dnr169f+UL/xc/Pz6U3RAvJQTIAycBCcqhYBnKkqSSdTkeTJk2uezmyDUoGFpKDZACSgUV5OVR0v+TUziGuHGjQwjLQYHR0dKnPiY6OtpkfIDY2tsz5hRBCCCGEEOJ6Of1UvalTpzJu3Dh69OhhHWgwNzeXBx98EKDEQIOTJ0+mf//+vPvuu4wcOZJFixaxc+dOPv30U2dWQwghhBBCCFGHOb3hNGbMGNLS0njllVc4d+4cXbt2LTHQoE53+cBY7969WbhwIdOnT+fFF1+kdevW/PTTT3Ts2LFayuvh4cGMGTNKnP7naiQHyQAkAwvJQTJwNslfMrCQHCQDkAwsHJ2DokqfsEIIIYQQQghxTU4fAFcIIYQQQgghajppOAkhhBBCCCFEOaThJIQQQgghhBDlkIaTEEIIIYQQQpRDGk52+uijj4iIiMDT05OoqCi2b9/u7CJVmY0bNzJq1CjCwsJQFIWffvrJ5nFVVXnllVcIDQ3Fy8uLwYMHk5iY6JzCVpGZM2fSs2dPfH19CQoKYvTo0SQkJNjMU1BQwIQJE2jYsCH16tXjjjvuKDFIc203f/58OnfubB1ALjo6mhUrVlgfd4UMrvb222+jKApTpkyxTqvrObz66qsoimJza9u2rfXxul7/mkz2TZfJvklT19+Psl8qyRX3S1C9+yZpONlh8eLFTJ06lRkzZrB79266dOnCsGHDSE1NdXbRqkRubi5dunTho48+KvXxd955hw8++ICPP/6Ybdu24ePjw7BhwygoKKjmkladDRs2MGHCBLZu3UpsbCzFxcUMHTqU3Nxc6zzPPPMMv/76K9999x0bNmzg7Nmz3H777U4steM1adKEt99+m127drFz504GDhzIrbfeyp9//gm4RgZX2rFjB5988gmdO3e2me4KOXTo0IGUlBTr7Y8//rA+5gr1r4lk32RL9k2auv5+lP2SLVfeL0E17ptUUWG9evVSJ0yYYP3fZDKpYWFh6syZM51YquoBqEuXLrX+bzab1ZCQEHXWrFnWaRkZGaqHh4f67bffOqGE1SM1NVUF1A0bNqiqqtXZzc1N/e6776zzHDp0SAXULVu2OKuY1SIgIED9z3/+43IZZGdnq61bt1ZjY2PV/v37q5MnT1ZV1TW2hRkzZqhdunQp9TFXqH9NJfumpdb/Zd/k2vsm2S+53n5JVat33yRHnCqoqKiIXbt2MXjwYOs0nU7H4MGD2bJlixNL5hwnTpzg3LlzNnn4+/sTFRVVp/PIzMwEoEGDBgDs2rWL4uJimxzatm1L06ZN62wOJpOJRYsWkZubS3R0tMtlMGHCBEaOHGlTX3CdbSExMZGwsDBatGjB2LFjSUpKAlyn/jWN7Jtsyb7JNfdNsl9y7f0SVN++yeCwEtdx6enpmEwmgoODbaYHBwdz+PBhJ5XKec6dOwdQah6Wx+oas9nMlClT6NOnDx07dgS0HNzd3alfv77NvHUxh/379xMdHU1BQQH16tVj6dKltG/fnvj4eJfJYNGiRezevZsdO3aUeMwVtoWoqCgWLFhAZGQkKSkpvPbaa/Tr148DBw64RP1rItk32ZJ9k2vtm2S/JPslqN59kzSchKigCRMmcODAAZvzZl1JZGQk8fHxZGZm8v333zNu3Dg2bNjg7GJVm+TkZCZPnkxsbCyenp7OLo5TjBgxwnq/c+fOREVF0axZM5YsWYKXl5cTSyaE63LlfZPsl2S/BNW7b5JT9SqoUaNG6PX6Er1wnD9/npCQECeVynksdXaVPCZOnMhvv/3GunXraNKkiXV6SEgIRUVFZGRk2MxfF3Nwd3enVatWdO/enZkzZ9KlSxfef/99l8lg165dpKam0q1bNwwGAwaDgQ0bNvDBBx9gMBgIDg52iRyuVL9+fdq0acPRo0ddZjuoaWTfZEv2TRpXeT/Kfkn2S6Wpyn2TNJwqyN3dne7du7NmzRrrNLPZzJo1a4iOjnZiyZyjefPmhISE2OSRlZXFtm3b6lQeqqoyceJEli5dytq1a2nevLnN4927d8fNzc0mh4SEBJKSkupUDqUxm80UFha6TAaDBg1i//79xMfHW289evRg7Nix1vuukMOVcnJyOHbsGKGhoS6zHdQ0sm+yJfsmjau+H2W/JPslqOJ9U+X6r3BNixYtUj08PNQFCxaoBw8eVB977DG1fv366rlz55xdtCqRnZ2t7tmzR92zZ48KqHPmzFH37Nmjnjp1SlVVVX377bfV+vXrqz///LO6b98+9dZbb1WbN2+u5ufnO7nkjvPkk0+q/v7+6vr169WUlBTrLS8vzzrPE088oTZt2lRdu3atunPnTjU6OlqNjo52Yqkdb9q0aeqGDRvUEydOqPv27VOnTZumKoqirlq1SlVV18igNFf2XqSqdT+Hf/zjH+r69evVEydOqJs3b1YHDx6sNmrUSE1NTVVVte7Xv6aSfZPsm1xx3yT7pdK52n5JVat33yQNJzt9+OGHatOmTVV3d3e1V69e6tatW51dpCqzbt06FShxGzdunKqqWrevL7/8shocHKx6eHiogwYNUhMSEpxbaAcrrf6A+sUXX1jnyc/PV5966ik1ICBA9fb2Vm+77TY1JSXFeYWuAg899JDarFkz1d3dXQ0MDFQHDRpk3TmpqmtkUJqrd1B1PYcxY8aooaGhqru7u9q4cWN1zJgx6tGjR62P1/X612Syb5J9k6vtm2S/VDpX2y+pavXumxRVVdVKHAUTQgghhBBCCJch1zgJIYQQQgghRDmk4SSEEEIIIYQQ5ZCGkxBCCCGEEEKUQxpOQgghhBBCCFEOaTgJIYQQQgghRDmk4SSEEEIIIYQQ5ZCGkxBCCCGEEEKUQxpOQgghhBBCCFEOaTgJ4SIUReGnn35ydjGEEEIIQPZLovaRhpMQ1WD8+PEoilLiNnz4cGcXTQghhAuS/ZIQ9jM4uwBCuIrhw4fzxRdf2Ezz8PBwUmmEEEK4OtkvCWEfOeIkRDXx8PAgJCTE5hYQEABopyvMnz+fESNG4OXlRYsWLfj+++9tnr9//34GDhyIl5cXDRs25LHHHiMnJ8dmns8//5wOHTrg4eFBaGgoEydOtHk8PT2d2267DW9vb1q3bs0vv/xStZUWQghRY8l+SQj7SMNJiBri5Zdf5o477mDv3r2MHTuWu+++m0OHDgGQm5vLsGHDCAgIYMeOHXz33XesXr3aZgc0f/58JkyYwGOPPcb+/fv55ZdfaNWqlc1rvPbaa9x1113s27ePmJgYxo4dy8WLF6u1nkIIIWoH2S8JcRVVCFHlxo0bp+r1etXHx8fm9tZbb6mqqqqA+sQTT9g8JyoqSn3yySdVVVXVTz/9VA0ICFBzcnKsjy9btkzV6XTquXPnVFVV1bCwMPWll14qswyAOn36dOv/OTk5KqCuWLHCYfUUQghRO8h+SQj7yTVOQlSTAQMGMH/+fJtpDRo0sN6Pjo62eSw6Opr4+HgADh06RJcuXfDx8bE+3qdPH8xmMwkJCSiKwtmzZxk0aNA1y9C5c2frfR8fH/z8/EhNTa1slYQQQtRisl8Swj7ScBKimvj4+JQ4RcFRvLy8KjSfm5ubzf+KomA2m6uiSEIIIWo42S8JYR+5xkmIGmLr1q0l/m/Xrh0A7dq1Y+/eveTm5lof37x5MzqdjsjISHx9fYmIiGDNmjXVWmYhhBB1l+yXhLAlR5yEqCaFhYWcO3fOZprBYKBRo0YAfPfdd/To0YO+ffvyzTffsH37dv773/8CMHbsWGbMmMG4ceN49dVXSUtLY9KkSdx///0EBwcD8Oqrr/LEE08QFBTEiBEjyM7OZvPmzUyaNKl6KyqEEKJWkP2SEPaRhpMQ1WTlypWEhobaTIuMjOTw4cOA1rPQokWLeOqppwgNDeXbb7+lffv2AHh7e/P7778zefJkevbsibe3N3fccQdz5syxLmvcuHEUFBTw3nvv8eyzz9KoUSPuvPPO6qugEEKIWkX2S0LYR1FVVXV2IYRwdYqisHTpUkaPHu3sogghhBCyXxKiFHKNkxBCCCGEEEKUQxpOQgghhBBCCFEOOVVPCCGEEEIIIcohR5yEEEIIIYQQohzScBJCCCGEEEKIckjDSQghhBBCCCHKIQ0nIYQQQgghhCiHNJyEEEIIIYQQohzScBJCCCGEEEKIckjDSQghhBBCCCHKIQ0nIYQQQgghhCjH/wMYBc1VE/UoxwAAAABJRU5ErkJggg==",
            "text/plain": [
              "<Figure size 1000x400 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plt.figure(figsize=(10,4))\n",
        "\n",
        "plt.subplot(1,2,1)\n",
        "plt.ylim(0,1)\n",
        "plt.grid(True, \"both\", linestyle=\"--\")\n",
        "plt.plot(history.history['accuracy'])\n",
        "plt.plot(history.history['val_accuracy'])\n",
        "plt.title('Model Accuracy')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend(['Train', 'Validation'], loc='lower right')\n",
        "plt.text(len(history.history['val_accuracy'])-3, history.history['val_accuracy'][-1]+0.02, round(history.history['val_accuracy'][-1],2))\n",
        "\n",
        "plt.subplot(1,2,2)\n",
        "plt.grid(True, \"both\", linestyle=\"--\")\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.title('Model Loss')\n",
        "plt.ylabel('Loss')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend(['Train', 'Validation'], loc='upper right')\n",
        "\n",
        "plt.show()"
      ]
    },
    {
      "attachments": {},
      "cell_type": "markdown",
      "metadata": {
        "id": "mLsZo3MN3bgS"
      },
      "source": [
        "**Load Model & Make Predictions**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 129,
      "metadata": {
        "id": "tW8Eea-X_TH3"
      },
      "outputs": [],
      "source": [
        "animal_classes = dataset_train.class_indices"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 76,
      "metadata": {
        "id": "eiGPyTTR3dW5"
      },
      "outputs": [],
      "source": [
        "classifier_1 = tf.keras.models.load_model(\"saved_model/trained_model\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 77,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JiAnwLj_4YjS",
        "outputId": "838d1037-88ae-4d03-dbe1-f3209e1e735c"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d_9 (Conv2D)           (None, 100, 100, 32)      896       \n",
            "                                                                 \n",
            " max_pooling2d_9 (MaxPooling  (None, 50, 50, 32)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_10 (Conv2D)          (None, 50, 50, 64)        18496     \n",
            "                                                                 \n",
            " max_pooling2d_10 (MaxPoolin  (None, 25, 25, 64)       0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " dropout_9 (Dropout)         (None, 25, 25, 64)        0         \n",
            "                                                                 \n",
            " conv2d_11 (Conv2D)          (None, 25, 25, 128)       73856     \n",
            "                                                                 \n",
            " max_pooling2d_11 (MaxPoolin  (None, 12, 12, 128)      0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " dropout_10 (Dropout)        (None, 12, 12, 128)       0         \n",
            "                                                                 \n",
            " conv2d_12 (Conv2D)          (None, 12, 12, 256)       295168    \n",
            "                                                                 \n",
            " max_pooling2d_12 (MaxPoolin  (None, 6, 6, 256)        0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " dropout_11 (Dropout)        (None, 6, 6, 256)         0         \n",
            "                                                                 \n",
            " flatten_2 (Flatten)         (None, 9216)              0         \n",
            "                                                                 \n",
            " dense_4 (Dense)             (None, 48)                442416    \n",
            "                                                                 \n",
            " dropout_12 (Dropout)        (None, 48)                0         \n",
            "                                                                 \n",
            " dense_5 (Dense)             (None, 10)                490       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 831,322\n",
            "Trainable params: 831,322\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "classifier_1.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 78,
      "metadata": {
        "id": "c5rKf0hy546X"
      },
      "outputs": [],
      "source": [
        "from io import BytesIO\n",
        "from urllib import request\n",
        "from tensorflow.keras.preprocessing import image\n",
        "import matplotlib.pyplot as plt \n",
        "import numpy as np"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 81,
      "metadata": {
        "id": "hhUPddt-57Hw"
      },
      "outputs": [],
      "source": [
        "def load_image_from_url(URL):\n",
        "    with request.urlopen(URL) as url:\n",
        "        img = image.load_img(BytesIO(url.read()), target_size=IMG_SIZE)\n",
        "    return img"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 140,
      "metadata": {
        "id": "rByM30usAUng"
      },
      "outputs": [],
      "source": [
        "animal_classes = {v: k for k, v in animal_classes.items()}"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 148,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bV3To13aCYIG",
        "outputId": "1b3822fc-ff69-4add-e637-ad63e7dba1eb"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "{0: 'cane',\n",
              " 1: 'cavallo',\n",
              " 2: 'elefante',\n",
              " 3: 'farfalla',\n",
              " 4: 'gallina',\n",
              " 5: 'gatto',\n",
              " 6: 'mucca',\n",
              " 7: 'pecora',\n",
              " 8: 'ragno',\n",
              " 9: 'scoiattolo'}"
            ]
          },
          "execution_count": 148,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "animal_classes"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 141,
      "metadata": {
        "id": "nRaKYcUr6BRD"
      },
      "outputs": [],
      "source": [
        "def predict_animal(model,animal_classes, img):\n",
        "  img_array = image.img_to_array(img)/255.0\n",
        "  img_batch = np.expand_dims(img_array, axis=0)\n",
        "  prediction = model.predict(img_batch)\n",
        "  return prediction, animal_classes[int(prediction.argmax(axis=-1))]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 146,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fJ1Tcmpy7j7f",
        "outputId": "5b43d93b-7b26-404a-e7ae-14ae07f9ffec"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Image URL: https://metronews.it/wp-content/uploads/2023/03/animali-ragno-botola-1568x1307.jpg\n",
            "1/1 [==============================] - 0s 31ms/step\n"
          ]
        }
      ],
      "source": [
        "probs, animal_pred = predict_animal(classifier_1, animal_classes, load_image_from_url(input(\"Image URL: \")))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 147,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        },
        "id": "10A3NMXW767w",
        "outputId": "d9e217f4-6e44-4101-95a4-452191ab0fcc"
      },
      "outputs": [
        {
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'ragno'"
            ]
          },
          "execution_count": 147,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "animal_pred"
      ]
    },
    {
      "attachments": {},
      "cell_type": "markdown",
      "metadata": {
        "id": "AWswZipyCbKH"
      },
      "source": [
        "**Save Model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 149,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YEOEVcolCcVv",
        "outputId": "bec75b06-8cfa-420c-e5fc-874a8260a479"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "  adding: content/saved_model/ (stored 0%)\n",
            "  adding: content/saved_model/trained_model/ (stored 0%)\n",
            "  adding: content/saved_model/trained_model/assets/ (stored 0%)\n",
            "  adding: content/saved_model/trained_model/saved_model.pb (deflated 89%)\n",
            "  adding: content/saved_model/trained_model/keras_metadata.pb (deflated 93%)\n",
            "  adding: content/saved_model/trained_model/variables/ (stored 0%)\n",
            "  adding: content/saved_model/trained_model/variables/variables.data-00000-of-00001 (deflated 13%)\n",
            "  adding: content/saved_model/trained_model/variables/variables.index (deflated 64%)\n",
            "  adding: content/saved_model/trained_model/fingerprint.pb (stored 0%)\n"
          ]
        }
      ],
      "source": [
        "!zip -r /content/saved_model.zip /content/saved_model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 150,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FWkEvhdMCwDK",
        "outputId": "1e5c4a79-4aac-47ff-88a1-c450e90446f7"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 152,
      "metadata": {
        "id": "HI_xdI7IC1gB"
      },
      "outputs": [],
      "source": [
        "!cp \"/content/saved_model.zip\" \"/content/drive/MyDrive\""
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "provenance": []
    },
    "gpuClass": "standard",
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}