Projeto_14_Geração_de_imagens_com_GANs.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Projeto #14: Geração de imagens com GANs.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/DanielMSousa/curso-deeplearning-na-pratica/blob/main/Projeto_14_Gera%C3%A7%C3%A3o_de_imagens_com_GANs.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "M27qF7CTrBqc"
      },
      "source": [
        "# Projeto 14: Geração de imagens com GANs"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1CEdoaEDUy7V"
      },
      "source": [
        "- Código adaptado da documentação: https://www.tensorflow.org/beta/tutorials/generative/dcgan\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zKmFmyaGunc7"
      },
      "source": [
        "# Etapa 1: Importação das bibliotecas"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nkilXv26_KTH",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        },
        "outputId": "34b58f08-4353-4500-e0f0-047ed6c272a3"
      },
      "source": [
        "%tensorflow_version 2.x\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import os\n",
        "import PIL\n",
        "from tensorflow.keras import layers\n",
        "import time\n",
        "from IPython import display\n",
        "import tensorflow as tf\n",
        "tf.__version__"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'2.8.0'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 1
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uAE6Icc0uylP"
      },
      "source": [
        "# Etapa 2: Importação e tratamento da base de dados"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5P_ciLXawMZg"
      },
      "source": [
        "(train_images, train_labels), (_, _) = tf.keras.datasets.fashion_mnist.load_data()"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Lj_7M8Ta2fjK",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "1f37774c-79ea-4be2-f9f4-fa044fb3f293"
      },
      "source": [
        "train_images.shape"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(60000, 28, 28)"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DC8ZLSBK2iM4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4699bd47-6c57-4520-e7f1-898fcf3d517a"
      },
      "source": [
        "train_labels.shape"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(60000,)"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tUbmhQed2V9f"
      },
      "source": [
        "train_images = train_images.reshape(train_images.shape[0], 28, 28, 1).astype('float32')"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Tv261Ul43Q_G",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "7f9a16d9-5e26-4042-eb94-89380a031e8e"
      },
      "source": [
        "train_images.shape"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(60000, 28, 28, 1)"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wf9jgEoE2tUc"
      },
      "source": [
        "# Normalização entre -1 e 1\n",
        "train_images = (train_images - 127.5) / 127.5"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EbRYvHSa2Nkd"
      },
      "source": [
        "buffer_size = 60000\n",
        "batch_size = 256"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xUcE_UfXP7ex"
      },
      "source": [
        "train_dataset = tf.data.Dataset.from_tensor_slices(train_images).shuffle(buffer_size).batch(batch_size)"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HWxs6fjiQE5c",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "3a73561e-b660-4178-b571-9ca856af159c"
      },
      "source": [
        "train_dataset"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<BatchDataset element_spec=TensorSpec(shape=(None, 28, 28, 1), dtype=tf.float32, name=None)>"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LlszUhNNyrl_"
      },
      "source": [
        "# Etapa 3: Visualização dos dados"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ronVvCdJsYc5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        },
        "outputId": "71f6a664-11a0-471a-913c-0560c8f8a717"
      },
      "source": [
        "plt.imshow(train_images[112].reshape((28,28)) , cmap = 'gray')"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7fd64025bf10>"
            ]
          },
          "metadata": {},
          "execution_count": 11
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Y0GmpAjG3GiH"
      },
      "source": [
        "# Etapa 4: Construção do gerador"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QkPICwnUR7nX"
      },
      "source": [
        "def make_generator_model():\n",
        "  model = tf.keras.Sequential()\n",
        "  model.add(layers.Dense(7*7*256, use_bias=False, input_shape = (100, ))) #12544\n",
        "  model.add(layers.BatchNormalization())\n",
        "  model.add(layers.LeakyReLU())\n",
        "\n",
        "  model.add(layers.Reshape((7, 7, 256)))\n",
        "\n",
        "  #7x7x128 - filtros\n",
        "  model.add(layers.Conv2DTranspose(128, (5, 5), strides=(1, 1), padding='same'))\n",
        "  model.add(layers.BatchNormalization())\n",
        "  model.add(layers.LeakyReLU())\n",
        "\n",
        "  #7x7x64 - filtros\n",
        "  model.add(layers.Conv2DTranspose(64, (5, 5), strides=(2, 2), padding='same'))\n",
        "  model.add(layers.BatchNormalization())\n",
        "  model.add(layers.LeakyReLU())\n",
        "  #saida: 14x14x64\n",
        "\n",
        "  #7x7x64 - filtros\n",
        "  model.add(layers.Conv2DTranspose(1, (5, 5), strides=(2, 2), padding='same', use_bias=False, activation='tanh'))\n",
        "\n",
        "  model.summary()\n",
        "\n",
        "  return model\n",
        "\n",
        "\n"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LwNlRTqyepj_"
      },
      "source": [
        "noise = tf.random.normal([1, 100])"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "noise"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GSUuccN3vW7J",
        "outputId": "301631ed-037e-497d-a3a1-2cc28075259e"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(1, 100), dtype=float32, numpy=\n",
              "array([[ 0.6834917 ,  1.8964229 , -0.3376083 ,  0.7360506 ,  1.3924062 ,\n",
              "        -0.16605078, -0.39168137, -1.223969  ,  0.05295035,  1.409398  ,\n",
              "        -0.3873138 , -1.1444573 ,  0.5533998 , -0.64711833, -1.5368384 ,\n",
              "         0.43597084,  0.9307647 ,  1.1285906 , -0.73260784, -0.04978077,\n",
              "         0.40679362,  1.0454975 , -0.1467054 ,  0.30255607,  1.7163224 ,\n",
              "         2.2085059 ,  0.33938628, -0.40227935, -0.04670194,  0.15987553,\n",
              "         0.5590347 , -0.303033  ,  0.72294694,  0.21242024,  0.06426553,\n",
              "        -0.14722596,  0.6874513 , -0.3020488 ,  1.3166778 , -0.7251194 ,\n",
              "         1.453139  ,  0.17024966,  0.12348117, -1.2499163 , -0.7477073 ,\n",
              "        -0.32049003,  0.01993686, -0.22732347,  0.61268425,  1.2689421 ,\n",
              "        -0.12206834, -0.0338158 ,  0.34359753, -0.1433937 , -0.26652125,\n",
              "        -1.4686936 ,  0.04503652,  0.4183516 , -2.2542586 ,  1.0401782 ,\n",
              "        -2.401728  , -1.2504282 , -0.9241306 ,  0.8954959 , -0.47416887,\n",
              "         0.17908856, -0.344409  ,  0.81370497, -1.8147153 ,  1.2451346 ,\n",
              "        -0.8043174 , -0.14924034, -1.5194061 ,  0.82998097, -1.1285126 ,\n",
              "         1.3058673 ,  0.11672904, -1.8338733 , -1.2378244 ,  0.3169897 ,\n",
              "         2.331531  , -0.7809674 , -0.2512529 , -1.3030323 ,  1.1531914 ,\n",
              "        -0.19438033,  1.3271313 , -0.22013013,  0.6675695 ,  0.93735105,\n",
              "         0.3537958 ,  1.264588  ,  0.44684765, -2.9369154 , -0.91614074,\n",
              "         1.9821719 ,  0.32093212, -0.86555964, -0.30546024,  0.74781466]],\n",
              "      dtype=float32)>"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wMqKlmhBevD8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b9da0887-a0ad-492c-8095-f3faecce922e"
      },
      "source": [
        "generator = make_generator_model()"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense (Dense)               (None, 12544)             1254400   \n",
            "                                                                 \n",
            " batch_normalization (BatchN  (None, 12544)            50176     \n",
            " ormalization)                                                   \n",
            "                                                                 \n",
            " leaky_re_lu (LeakyReLU)     (None, 12544)             0         \n",
            "                                                                 \n",
            " reshape (Reshape)           (None, 7, 7, 256)         0         \n",
            "                                                                 \n",
            " conv2d_transpose (Conv2DTra  (None, 7, 7, 128)        819328    \n",
            " nspose)                                                         \n",
            "                                                                 \n",
            " batch_normalization_1 (Batc  (None, 7, 7, 128)        512       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " leaky_re_lu_1 (LeakyReLU)   (None, 7, 7, 128)         0         \n",
            "                                                                 \n",
            " conv2d_transpose_1 (Conv2DT  (None, 14, 14, 64)       204864    \n",
            " ranspose)                                                       \n",
            "                                                                 \n",
            " batch_normalization_2 (Batc  (None, 14, 14, 64)       256       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " leaky_re_lu_2 (LeakyReLU)   (None, 14, 14, 64)        0         \n",
            "                                                                 \n",
            " conv2d_transpose_2 (Conv2DT  (None, 28, 28, 1)        1600      \n",
            " ranspose)                                                       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 2,331,136\n",
            "Trainable params: 2,305,664\n",
            "Non-trainable params: 25,472\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Bqx0BOjGfYn-"
      },
      "source": [
        "generated_image = generator(noise, training = False)"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "generated_image.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ssreefnSz7om",
        "outputId": "016f2341-85cb-4df3-f5ff-9cc995e534b2"
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "TensorShape([1, 28, 28, 1])"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jvfESeBafsrL",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        },
        "outputId": "bb61428c-4355-45dd-a1f4-f6419c1f4d7f"
      },
      "source": [
        "plt.imshow(np.squeeze(generated_image), cmap='gray')"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7fd5c8f35310>"
            ]
          },
          "metadata": {},
          "execution_count": 18
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "53qDZFRn3-S1"
      },
      "source": [
        "# Etapa 5: Construção do discriminador"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pzXywbIxg4Sk"
      },
      "source": [
        "def make_discrimator_model():\n",
        "  model = tf.keras.Sequential()\n",
        "  # 14x14x64\n",
        "  model.add(layers.Conv2D(64, (5,5), strides = (2,2), padding='same', input_shape=[28,28,1]))\n",
        "  model.add(layers.LeakyReLU())\n",
        "  model.add(layers.Dropout(0.3))\n",
        "\n",
        "  # 7x7x128\n",
        "  model.add(layers.Conv2D(128, (5,5), strides = (2,2), padding='same'))\n",
        "  model.add(layers.LeakyReLU())\n",
        "  model.add(layers.Dropout(0.3))\n",
        "\n",
        "  # 6272\n",
        "  model.add(layers.Flatten())\n",
        "  model.add(layers.Dense(1))\n",
        "\n",
        "  model.summary()\n",
        "\n",
        "  return model"
      ],
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FECxiL_7iHA1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f72e80f8-4dbb-49dc-d672-1e21db6b5ec8"
      },
      "source": [
        "discriminator = make_discrimator_model()"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d (Conv2D)             (None, 14, 14, 64)        1664      \n",
            "                                                                 \n",
            " leaky_re_lu_3 (LeakyReLU)   (None, 14, 14, 64)        0         \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 14, 14, 64)        0         \n",
            "                                                                 \n",
            " conv2d_1 (Conv2D)           (None, 7, 7, 128)         204928    \n",
            "                                                                 \n",
            " leaky_re_lu_4 (LeakyReLU)   (None, 7, 7, 128)         0         \n",
            "                                                                 \n",
            " dropout_1 (Dropout)         (None, 7, 7, 128)         0         \n",
            "                                                                 \n",
            " flatten (Flatten)           (None, 6272)              0         \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 1)                 6273      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 212,865\n",
            "Trainable params: 212,865\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3kCEOHsLiaYq",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "57ccf3aa-f59b-4216-a721-030d3c62f23c"
      },
      "source": [
        "decision = discriminator(generated_image, training = False)\n",
        "decision"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(1, 1), dtype=float32, numpy=array([[0.00162938]], dtype=float32)>"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JbFoRCw471Yt"
      },
      "source": [
        "# Etapa 6: Loss function para as duas redes"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OMWmRYdTkUuB"
      },
      "source": [
        "#From logits indica que o valor talvez ainda não esteja normalizado\n",
        "cross_entropy = tf.keras.losses.BinaryCrossentropy(from_logits=True)"
      ],
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LbJ4u3_xkoqC"
      },
      "source": [
        "def discriminator_loss(real_output, fake_output):\n",
        "  #Compara o resultado das imagens reais com valores 1\n",
        "  real_loss = cross_entropy(tf.ones_like(real_output), real_output)\n",
        "  fake_loss = cross_entropy(tf.zeros_like(fake_output), fake_output)\n",
        "  total_loss = real_loss + fake_loss\n",
        "  return total_loss"
      ],
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cr6coEqAlcEJ"
      },
      "source": [
        "def generator_loss(fake_output):\n",
        "  return cross_entropy(tf.ones_like(fake_output), fake_output)"
      ],
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7FwviROulzjc",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "cd3776d7-c91f-40bf-b4e7-de24649dc716"
      },
      "source": [
        "generator_optimizer = tf.keras.optimizers.Adam(lr=1e-4) #0.00001\n",
        "discriminator_optimizer = tf.keras.optimizers.Adam(lr=1e-4) #0.00001"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/keras/optimizer_v2/adam.py:105: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n",
            "  super(Adam, self).__init__(name, **kwargs)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "B4hthLwsmD5T"
      },
      "source": [
        "checkpoint_dir = './training_checkpoints'\n",
        "checkpoint_prefix = os.path.join(checkpoint_dir, 'ckpt')\n",
        "#essa função faz com o que o modelo seja treinado salvando os modelos no checkpoint\n",
        "checkpoint = tf.train.Checkpoint(\n",
        "    generator_optimizer=generator_optimizer,\n",
        "    discriminator_optimizer=discriminator_optimizer,\n",
        "    generator=generator,\n",
        "    discriminator=discriminator)"
      ],
      "execution_count": 26,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-wRQqOeB5Zh5"
      },
      "source": [
        "# Etapa 7: Treinamento do modelo "
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "epochs = 100\n",
        "noise_dim = 100\n",
        "num_examples_to_generate = 16"
      ],
      "metadata": {
        "id": "3Cp4BrkpXVHl"
      },
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "seed = tf.random.normal([num_examples_to_generate, noise_dim])\n",
        "seed.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "a2da3mrnXecL",
        "outputId": "ba66c22a-4148-4a2f-e603-79c734a1f2e0"
      },
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "TensorShape([16, 100])"
            ]
          },
          "metadata": {},
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#Isso permite que o tensorflow saiba que ele processará isso\n",
        "@tf.function\n",
        "def train_steps(images):\n",
        "  #gera 256 seeds aleatorias\n",
        "  noise = tf.random.normal([batch_size, noise_dim])\n",
        "  \n",
        "  #permite que você controle o gradiente do modelo\n",
        "  with tf.GradientTape() as gen_tape, tf.GradientTape() as disc_tape:\n",
        "    #chamar o gerador e passar os numeros aleatórios\n",
        "    generated_images = generator(noise, training=True)\n",
        "\n",
        "    #passar para o discriminador as imagns reais e falsas para classificar \n",
        "    real_output = discriminator(images, training=True)\n",
        "    fake_output = discriminator(generated_images, training=True)\n",
        "\n",
        "    #calcular o loss do generador e discriminador\n",
        "    gen_loss = generator_loss(fake_output)\n",
        "    disc_loss = discriminator_loss(real_output, fake_output)\n",
        "\n",
        "  #calcular o gradiente das losses para atu alizar\n",
        "  gradients_of_generator = gen_tape.gradient(gen_loss, generator.trainable_variables)\n",
        "  gradients_of_discriminator = disc_tape.gradient(disc_loss, discriminator.trainable_variables)\n",
        "\n",
        "  #aplicar os otimizadores e atualizar os pesos\n",
        "  generator_optimizer.apply_gradients(zip(gradients_of_generator, generator.trainable_variables))\n",
        "  discriminator_optimizer.apply_gradients(zip(gradients_of_discriminator, discriminator.trainable_variables))"
      ],
      "metadata": {
        "id": "FsgCUTNQXjVe"
      },
      "execution_count": 29,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def generate_and_save_images(model, epoch, test_input):\n",
        "  predictions = model(test_input, training=False)\n",
        "  fig = plt.figure(figsize=(4, 4))\n",
        "\n",
        "  for i in range(predictions.shape[0]):\n",
        "    plt.subplot(4, 4, i+1)\n",
        "    plt.imshow(predictions[i, :, :, 0] * 127.5 + 127.5, cmap='gray')\n",
        "\n",
        "    plt.axis('off')\n",
        "\n",
        "  plt.savefig('image_at_epoch{:04d}'.format(epoch))\n",
        "  plt.show()"
      ],
      "metadata": {
        "id": "UAPzBo4Mf0ap"
      },
      "execution_count": 30,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def train(dataset, epochs):\n",
        "  for epoch in range(epochs):\n",
        "    start = time.time()\n",
        "\n",
        "    #parte de treinamento\n",
        "    for image_batch in dataset:\n",
        "      train_steps(image_batch)\n",
        "\n",
        "    #parte de visualizar resultados\n",
        "    #Vai gerar uma imagem a cada batch (sem objetivo de treinar) e salvar\n",
        "    display.clear_output(wait=True)\n",
        "    generate_and_save_images(generator, epoch+1, seed)\n",
        "\n",
        "    if(epoch + 1) % 15 == 0:\n",
        "      checkpoint.save(file_prefix=checkpoint_prefix)\n",
        "    \n",
        "    print('Time for epoch {} is {} sec'.format(epoch + 1, time.time() - start))\n",
        "  \n",
        "  display.clear_output(wait=True)\n",
        "  generate_and_save_images(generator, epoch+1, seed)"
      ],
      "metadata": {
        "id": "prbdP9n7fQR-"
      },
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train(train_dataset, epochs)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 247
        },
        "id": "BdRwwPJhhw4e",
        "outputId": "1af3f64a-3044-4480-8e35-050c7fb5665d"
      },
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#carrega o ultimo modelo\n",
        "checkpoint.restore(tf.train.latest_checkpoint(checkpoint_dir))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "q5GoQLMxjKAI",
        "outputId": "6f5515ee-e2f4-4291-e77d-ae40b365b040"
      },
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.training.tracking.util.CheckpointLoadStatus at 0x7fd6401cfa10>"
            ]
          },
          "metadata": {},
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def display_image(epoch_no):\n",
        "  return PIL.Image.open('image_at_epoch{:04d}.png'.format(epoch_no))"
      ],
      "metadata": {
        "id": "NqzOimSBjY8P"
      },
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#mostra a imagem da ultima epoca\n",
        "display_image(epochs)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 304
        },
        "id": "u3QuWvgTjpja",
        "outputId": "e3749086-eaf8-43e7-e6ce-c91de7d148ba"
      },
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<PIL.PngImagePlugin.PngImageFile image mode=RGBA size=288x288 at 0x7FD5526FFC90>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAASAAAAEgCAYAAAAUg66AAACT9UlEQVR4nO29aYyd93Ue/tx9f+++3zv7kMN9lahd8hIncWqnSYMiRhG0TVOgKNJPRQIELYoi/Vb0U4EA/dAiSJAgRfB34GZzHDmyLVmyZMo0JVIkZ1/vzN33ff1/mDxH79CSTVGcuUPqfYAByeEs7/L7nd85z3nOObrRaDSCBg0aNIwB+nFfgAYNGj670AyQBg0axgbNAGnQoGFs0AyQBg0axgbNAGnQoGFs0AyQBg0axgbNAGnQoGFs0AyQBg0axgbNAGnQoGFs0AyQBg0axgbNAGnQoGFs0AyQBg0axgbNAGnQoGFs0AyQBg0axgbNAGnQoGFs0AyQBg0axgbNAGnQoGFs0AyQBg0axgbNAGnQoGFs0AyQBg0axgbNAGnQoGFs0AyQBg0axgbNAGnQoGFs0AyQBg0axgbNAGnQoGFs0AyQBg0axgbNAGnQoGFs0AyQBg0axgbNAGnQoGFs0AyQBg0axgbNAGnQoGFs0AyQBg0axgbNAGnQoGFs0AyQBg0axgbjuC/go6DT6cZ9CZ8Ko9Hogb7uk9ynTqeDwWDAaDTCcDiEXr9/dgwGg4e6Rp1OB71eD71ej8FggOFw+Il/xmHc53GEdp+HB80DegxAY2E2m2EymTAajTAYDB7a+BBGoxF2ux0Gg+Gx3zwaHk8cSw9Iw09iNBqh2+1+6lPKYDCI4RkMBuh2uxgOh2M5/TRo0DygxwSj0Qj9fv9Tez1qA2Q0GtHr9R4q/NKg4VFANzqGR9/jHg4cFgf0SX/+x/0cvV4Pk8kkYdzDGiCNGzmIz8p9PkpoIdhjBL1ej9Fo9KkWCknsbrcr/9agYVzQDNBjAJ1OJ1mw4XD4qUOmjzJi/B0AtJBMw5FBM0DHFDQGo9FIjMNhkcX8+VarVX6v2WyGwWBAuVxGv99/5L9TgwZAM0DHFupwS6/XQ6fTHaoh0Ov1kpLX6/VQFAVWqxXtdhvNZlPzijQcCjQD9Ahhs9lgMpk+9c8hUUzvx2QyQa/XixHg/wEQr+hhPCP+nrNnzyISiSAWiyEYDGJ2dhYejwc2mw23b9/G6uoq/vRP/1RS9hqeDPD9WywWGAyGsVyDZoBU0Ol0MBqNIvT7pJvaZDLBbDY/0ush98O/M4OlNkDAvlHq9/sHvKT7M2f84M+x2WxwOp1IJpOIxWLy56VLl+D1emGz2eDxeOByufD//X//H/r9/mNhgNRGG9h/Rsz2PSmkOz1VvtOHuTeLxQK73T7W7J1mgIADG3JiYgKNRgOlUgnNZvOBwh6DwQCDwQCz2fxIDJDa2KgXmc1mg81mg9frlQVIhXQkEsHe3h5WV1dhNBqh1+vFmI5GI7RaLXQ6HQyHQxgMBrjdbrz88sv45V/+ZWxtbaFSqaDRaKBcLiOXy6HT6cDpdOL8+fOw2WwIBoPQ6XQol8uf+v4OEyaTCRaLBWfPnoXL5YLZbEahUEA6nUY2m0Wz2fyJ71E/a5amHFfeiwdQKBSC2+1GtVpFu91GoVD4REZIp9Ph7Nmz+OIXv4hXX30VW1tbh3zlHw3NAP0jaDz0ej1cLhdsNhuy2SwajYaI/+4nhHnSWiwWWK1WGI3GT32aqMlndWil0+lgsVhgsVgOhHl2ux1WqxUGgwE2mw0+nw+KosBoNKLf78NoNMLhcKDX66Hb7aLX60Gn08HhcMDpdKJcLmN3dxflchkmkwn9fh8OhwNerxeKosBsNqPRaCAWi6Hf76NarR57L4jG2uPxIBAIIJFIYGFhAUtLSygUCiiXyxJ60GtsNBoYjUYwmUzodDrHzgDx/ZtMJthsNuj1erTbbfR6PfT7/Z9peNTr0mAwwG63IxqNYnZ2Fu+8884j9dw/CT7TBogvRa/Xw+FwwGq1ot/vIxqNIplM4ubNm8hms2i1WnK60PPodDrQ6/WyUb1eL0qlEtrt9iO5tvtDQKPRCKfTCavVCqvVinq9jm63i1gsBovFgmazCavVipmZGczNzcHpdCKXy8HlcmFmZkaMY71el1AqnU7jm9/8JpaXl1Gr1ZBIJOByuZBKpRAMBuHz+bCzs4PBYIArV67AYDBgd3f3WIdi9F7sdjsCgQCuXr2KqakpzM/P4x/+4R+wsrKCmzdvwmQyIRKJoFwuo1KpYGlpCe12GzabDZVK5SM9pXFBr9fDYDDA7/fLIbO3t4fd3d2fkGWoucL7fwa9aavVimQyiRMnTuDSpUt47bXXsLu7e5S3JPhMGSC+yFAoJPEveY+TJ0/CaDQim83C4XDA5/MhGAyiUqmIFxEOh9HpdKR8YTAYiIfU7/fx7rvvolKpfKprvD/9Tv6Hho//Fw6HYbfb4fF4MBgMkMvlYDQaYbVaMTU1hXg8LiFHKpVCr9fDYDBAKBSC0+lEPB7HYDDA2tqaZL/oAYTDYfT7feRyOTidTpjNZsRiMayvr3/qd3DYeOaZZ3D+/Hk8//zzCAQCCAQC8Hq9cLlcuHTpEqampnDp0iUYDAa4XC60Wi10u13k83l0Oh00Gg28+eab+Ou//utx34qsV5/PB5fLhTNnzshh12w20el0UKlU0G63UalUhAao1+sYjUaIx+Nwu92IxWLI5/NibOkxeTweyXY+iuTJw+AzY4AYO1ssFoTDYTidTng8Hjz11FM4deoUrl69Cp1Oh/fffx8GgwEWiwVerxftdhsulwuBQAALCwuoVqtoNpswGAzodDrY3NxEqVRCNpvFe++9h16v90iul7wEDQ//NBqNMBgMCAQCwss0m030ej0YDAaYTCbE43HMzc3B4XAglUphZWUFjUYD3W4XHo9HjMz29jYGgwHsdjvMZjOGwyFMJhO8Xi/S6TTK5TL8fj8URYHP54PT6YTBYPjU9WiHiTNnzuCXfumXcPHiRbhcLgwGAxiNRpjNZszOzkorE3U5ynA4RL/fR7vdRiaTQalUwt/8zd+MnbDmgUJDevLkSbhcLvh8PuH50uk0ms0mcrmcrO9KpYLRaIRz584hGo3izJkzWF5exvb2NtLpNKrVKlKplHjUJpNJy4IdJiwWC2w2Gy5duoRYLCYkLgBMTk4ikUhgeXkZo9EIPp8PvV4PnU4HXq8XwD4/5HA4ZJHW63UkEgkoioLRaIRGo4G9vT0xTp8GNDQmk0k+eErp9XrEYjFcu3btQDjWarXgdDoxNTWFM2fOIB6PQ6/XY3V1Ffl8HgCQTCahKArW1tawurqKO3fuYDQaYXp6GhaLBaPRCGtrazAYDCiVStjY2EAqlUK320UwGMT58+fhcrkwPz+Pzc1NVKvVT/dSDgE6nQ6JRALnz5+H2+0Wr67ZbKJQKEhmjB6l2WwWDqXf70On0yEYDMLr9cLhcKDVao3N2Op0OszPz+Ppp5/GzMwMnE4n1tfXUS6Xsby8jPPnz2NmZgaBQABGoxF+vx+VSgW5XA5zc3Pw+Xxy2NhsNjkYA4EASqUSisUiarUaFhcXUSqV0Ol0xnKfT7wBIqkciUQwMTGBcDgMh8Mhp6DNZsNwOEStVgOwr+Uh+cuKcRqEfr+PTqcj2SSerP1+H+VyGZ1O51M1CFN7PSaTSbwdGhpuEgCiySG343A4YLFYoNfrJZxIpVJoNpvi1QQCAWSzWdTrdZTLZVitVthsNvFq+v0+ut0u6vU62u02+v0+arUaHA4HTCYTfD4fpqenkc/nUa/XH9k7ehTgczMajRJOkB+hh8NDhzwJ32e32xWv1uPxyDPvdrtjMUBMbgQCAczPzyOZTMJisUgoDexzXb1eTzKjfr8fer0evV4PkUgEfr9fkiWtVgv1eh21Wg12ux0OhwMejwdmsxndbhdOpxOBQODI7xN4wg0QMx1nz57Fz//8z8Nut8vGdjgcCAQCaDQaWF9fx2AwECPARUy+hFmRWq0mp4fL5YLFYkG320WhUBAi92EXLMMrs9ksrrf6epjVymaz+Na3vgVg31uampqC3W6HxWLBrVu38M4776BQKIiRCQaDuHr1KiYmJjA5OYnJyUmUy2Vcv34dzWYT9Xod+XwejUYDqVQKNptNsmfRaFSyZ6FQCKFQCBcvXkQmk0GhUHiUr+pTgc/I4XCg2Wxid3cX4XBYDofhcCgiUXJqbG/SaDRQr9exubkJAJiYmECv10MgEEC73Zai3aOE0WiEz+fD3NwcXnjhBTn8Jicn0e/3JQmyuroKr9cLp9MpRldRFFmTPIDK5TLeffddrKys4JVXXoGiKLhy5Yp873PPPYdTp04d+X0CT7ABYtg0Pz+P6elpeL1e4UF8Ph8GgwFKpZJkPKLRKCwWC4zG/UcyHA6Rz+clrDKbzfB4PHC73fKyO52OhGSf1gOimIzfPxgMxEtrt9tSwU6dCk9nnU4nJ1ij0RAj2W63hVTd29uDwWBAoVCQlDxDEJKRFosF1WpViG2GJZVKBbVaDe+//z78fj98Ph8ACCE+bjATSS/XarWiWq3C7XYDOFjt3+v1JIFA/ZTZbJbMEkl4p9OJUCiEXC535J6exWKBw+HAxMQEjEYjNjc3UavV0Gq1kM/nYTab4fV6Za1YrVYhnkkd0FPa3d0Vr87r9WJhYQEWiwXA/uHVbreRTqexu7uLYrF4pPdJPHEGSF1Y6ff78eyzzyIej8Pj8aBYLKLZbCIYDKLf76NYLKJYLKLVamF6ehqKokgo0u/3kU6n0e12UavVEAgEEAqFxFUtFotSJ1WpVOTfn8YA0aCMRiMxEkajEY1GA61WS+rCGDK2Wi3UajU4nU7odDoxPlyMRqMR1WoV6+vrSKVSsFqtmJ+fl7CLBKTdbpfQy263IxwOo9Vqod1uo1QqoVwu4+2338b8/DzOnz8P4PgYIIPBgEgkglAohNOnT8Plcgl5zv9X19INBgO0223JgtpsNgl1KdL0eDyIx+NYXV090nvR6XSw2+3w+Xw4ceIE9Ho97ty5g5WVFVSrVSiKIkkBkueUj9RqNXn/KysryOVy2NvbQyQSwTPPPINoNIpTp06hVquJF8XvWV5eFq7wqHGsDRA3o/rf6r/TQ6ALSiVzOBzGzMwMQqEQLly4IBva5XJJWEaPg5kQhj5ut1u8mlAoBIPBgJ2dHdnkwL53REFfIpFANBqFz+f7VC67+j6ZImZ8z1OL6VVyBOrm9ABgtVrh9XqFJyAv0Ol0JEyhF0jNi9vtRjKZhMlkQr1eh9PpxPz8PBYXF5FOp1EsFmE0GnHixAnEYjHEYjG43W7Y7faHus9HCfJS/+bf/BsEg0H4/X5RpTscDgm7yG0xrB0Oh9JfmwdOrVbDaDRCIBDAqVOn4HK5sLy8jFQqdSjXzvfBNRWJRBAIBHDhwgV4PB54PB60Wi2Uy2U4nU4JsykV4UGXyWTgcrnEEHm9Xjz//PPo9Xp48803oSgKQqGQvK9yuQyj0Yhz584hn8/LzxmX8PLYGiD1FAie/HSb6S2QePX5fLBarZKlSSQSOHHiBLxer5zmlUoFDocDAISMpLKY4MIcDAZiBCwWC2q1mqSp1R4SjZnNZoPD4XhkqUwa1ft/Ln83yWcKztQNxshxDQYD2Gw2AJAMx3A4FHK50WgIea0oivw+PkfWm1F86Xa7ZZE/qpKThwX5Mo/Hg0gkgjNnzsgmI4FOr4aGejQaSUKBrUaADxXn1Mbo9Xr4fD44HA64XK5Hes00NpQ9WK1W+f9QKIRwOIzJyUm4XC45NEejkcgfCL5rg8EgWVen0yncZSQSgV6vx8rKCmw2m9R7MeNnNBqhKIqElzTQ48CxNUA0LBT+8VQ4d+4cyuUyMpkMPv/5z2N6ehpzc3Ow2+1QFEW8EIZOf/u3fwu/3494PI7p6WnodDqk02kYDAYEg0EpY1haWhJFNF98r9eDyWTCxMSEaERqtZpwQlQWV6tVeL3eTyXmYuhIpeqZM2fQ6/UOpPWDwaAYG7PZLPVd/X4f7733HpxOp2hEaISMRiNcLpdsvnw+L5uuXC5je3tb0ryNRgO9Xg97e3twOBw4efIkotEojEajpPa3t7cxGo3EmI8DwWAQyWQSX/nKVzA/Py9kM3k6HlCdTgcmkwm9Xg/tdhtOpxMWi0XS75VKBd1uF/1+X7wjvvNH1dkAgBhLFit/9atfxYkTJxCNRjEcDtFsNrG6uord3V2k02ns7e2JDGRmZgZerxdGo1HUymazGW63GzabDffu3UOj0UAikZCDlvd58uRJAPvcF0PzZDIJl8uFYrGIbDaL3d1dkVyMA8fWAJEkpRd08uRJifPJu5w9exaxWExcb/UpYzabYbfb4XK55MThKUL3mx6UwWCA0+kEgAM1XtyoDPHUXgZLLsxmM0KhEObn57G6uvqps0M0RHTRrVbrRzaOZ8rdZrOh2+2Kl6M+ybi5aEz5XPV6vfA/FosFg8FAtER2ux1utxsWiwXtdltCGkoS6KofJgf0Ua1nGYr6fD7Mz8/jwoULuHDhgmxien78foan9GrU1fH8OhoYg8Eg90Uvguvik94n3xv/7vf74XA4kEwmYbVaYbfbMT09jVAoJOuVXQkURZHrd7lc8Hq9YnxIkJtMJng8Hsno8iBgKGUymYQvpCfE59dqteTd0gNjaPaoBLSfFMfSADH8cjgcsgm/+tWvYmJiQmqfLBYLFEWBwWBAJpNBvV5HNpuVzRuPx8X4UCPCE57hRigUEq9jenpawhv+brrmzJA4nU7RX9y+fRuDwQCRSAQOhwOxWAzvv/8+dnZ2PvW9A0Cr1YLVaoWiKKjVamJMqGuhQXC5XJId40nPDdhqtaDT6SSsHI1Gwi0xrR6LxTAajVAoFJBIJIQAbTQaaDabcj+Kooj3xfdzGFAbf3qiOp1OkgDPPvssXn75ZXzlK18BsL9xUqmUkOoMvcj1MBxjgS4A8RAZnnHzDodDWK1WKS5WG60HBYuCGSY+9dRTiMfjeOqpp+B2u+H1ejEajdDpdHD79m3odDq43W5YrVbEYjEYjUZYLBaEQiE5KHd2dlCv18X4zM3NSUYvHo9LCMVQmZndp556SsJwvq+1tTUMBgP4fD4R1lILNw4cSwPE9g/nz58X7iefz6Pdbks2h/VLwH4Mqw5X+PIZkjSbTckasSaKaWp+b7lclg0cCoWQSCSEqN7d3UW9XsfW1paQ1dVqFTqdDtVqFXq9HqFQSLiUBwU3ssViOdD7h2pknsBsC0Jj2el0ZBM1Gg0A+5uq3W6L0aHxVJ/uAETdm81m5d5ZVd3pdGCz2bCysgKTyQSdTodsNovRaIStrS34/X6EQiE5FA4T5NeYJv+lX/olTE9P49KlSwgEAsjlchKCptNpCUnV3QK4XsgF0htkPRSFfCaTSbJAgUBAukP6fD5Eo9EHvmZFURCNRvHMM8/AZrNJqMdsFst/OI2kVqthOByiXC5LOQ3T8NQ0cR4cAPFs1tbWUCwWUa1WUSwWReeUTCYxNzcHABLasdsBPchms3lA60aF/GG/z4/DsTRAZPMXFhbECFSrVZTLZVEoq7kaPjx1nQ9d03g8jlarhUajIQQsNyTT2+12G9VqFZ1OB/V6HTqdDuFwWE4Fpuzv3r0Ll8slqmSj0YharQZFUeDxeISkfVAwNHA4HAeawhsMBsl4qTU/9AwYGlEjxNOfxun+VD7wYWhHFXepVJLTP51Oo9Vqye+m5+VwOGSBm0wmzM7Owuv1ynV/EqizPvTQ1NIDdcZTXf/G4stnnnkGCwsLOHPmDKrVKnK5nPTCKRaL4s2QJNfpdCIy7ff78Hq94ins7u4KT0gxXj6fl3VD0pZ1Vw8Ku92OWCyGF198UbyaGzduoFgsYnt7GzabDYqiSLjD4ZAkzrvdroTUPCiYbVWTyN1uF6lUCrlcTtqIUN3scDjE06c+7d69e5JEYCaN7zkQCIjnOA4cSwOUTCZx6tQpvPLKK+IG37lzB8Vi8UCIw0wQFci9Xk/aY9TrdSFhq9WqiLiomGUmqFwuo16vIx6PCyE9HA6xsbGB27dvo1KpYGpqCjqdTsofms0mqtUqDAaDqIRtNhva7fYnasfhdrsRDAZx9uxZNJtNUbDydOamJQ8G7Ge0Wq3WgT5AwL4xZWaOdWPqEIa6EaZzA4GAiNXoXZInY6sPJgCAfY6hWq2iUCigUCh8It2IxWJBMpnEK6+8grm5OUSjUayvryOXy+HHP/4xLBYLnE4nisUiOp0OPB4Put0uyuUyXnrpJVy8eBFOp1Nq1HZ3d7G2tiZdG5PJpPQqqtfrMBgMiMVicLlcoh6+desWFhcXUS6XYbfb4XQ6RenN50ROjaUpbM72oKBn7nA4MDk5iUgkIte9trYm/ZharZZokphltNvtkkgxmUwiDBwMBmIkgsEgms0mstmsUAVXrlwRIz0YDPDBBx/gzTffRKFQwEsvvYR+v498Pn+AjmDozmtNpVJIp9MPfJ+PEsfSAJEEJEHG058V6gRPOnpJTInTmvN06ff7B5p5kWBk3xhuWvUGrFQqomxmWp5uq7ou69PEzgwvWH/UbDbFs1P/XDUhS3dbrdrm1zDsVHdEVH+/Wv9CUr3T6QjfcX+bV3o+AGST0tB/EtLS4XAgFArh3LlzmJmZQSQSgaIoyOfzB8j2SqUiei1mbk6ePImZmRnx0Pg++R4oI1B7tuqWrF6vV4SZfPdMsXMTWiwWuFyuA4kM3ucnSU/3+33U63Xs7OxIYzO2yyAP1+12DwgfARzQVHk8HiGdueZYeMzEA4WxDodDSG6j0Yh6vS7V7vSo+N65HtQHWq/Xk8hAK0ZVgeRYoVDA5uYmMpkMbDYbXC4XXnzxRaTTaWxsbCAQCMBut2M0GsFisQgB12g0ZPPVajUEg0GcPn0awP5mZOEptSMmkwmLi4uo1WqoVCqoVqsolUo4d+6c/I52u41cLodIJCK6I6asmZWgjuZBwcXP8opOpyO6H5/PJx6HOsOhKIqElfS4uJD9fj/8fr94EhRPqjVTjUbjQJHmaDTC5cuX4fF4sLq6KjySWhPEkO1hsybT09O4cOECvvrVrwoH99xzz8FsNuPXf/3XUavVUCgUpCSEYRkPg1arJX17arUaIpEILl26JJuJGq1kMimfY+fGYDAoJTMvvviiFGACHwpd1V4in8/DTI3d29tDsVjE4uKiHGihUAherxdzc3Oy6bmm2G87EomIdxQMBjEajZDNZuH1ehEKhYSjazQakjGLRqPweDy4c+cOcrkcbt26hVQqhc3NTZw6dQoXL17EuXPn0Gg0YDAYMDExgUgkAo/HA2BfkMjSnXH2yj6WBojxOQV33Ah2u10WB0lT9UlPr4DhErAfpnW7XRSLReEeqIUxGo3iMZFoZtYkGAwiHo8jGAxKapNZBqvVCqfTKSQmF5t6gz4IqNvY2dmRBlOsNVP3KOYC4emsrphnVorfww1OEhqAkNEkPvmsGLJRlKjujaOWHHDjqLOEn4QDorfhdDolNFb3Xe50Omg2m2J86OFQv0SPgClr9mKi8SUPAkDC7GKxKAp3s9l84H5ZMkOPjh7CYDAQwSrlGJ9EcMksIbVivM58Pi/Pvd/vyxoiz8ZwmF48APGG2fKF98l3XavVEAqFpB8VuchLly7h7NmzUg3Q7XYRCATkd7FGTt03iM93HDiWBojuIxcQ05c8vdhCgacVSxbU1c7MBrndbjSbTTEgw+EQ29vbAD7UA+n1eqytraHZbEKv1yMej0v1OGu/2u22ZN6Ypu/1esLFsKbqk2TB1O1RgX1SmuEYDQ1LCeiSk/NRu+kAhMgm0U4jTUPEzcc+Pk6nU1TBPA3J+XAz0C1nOQeNGzN3Dwqe9Ha7Xdrb8ueNRiPpMMAQkZ4gvUlyfaPRSDYQkwadTkcaqtPQGI1GFAoF8dIoXCTPVi6XxdCpDRCTDRQzMjR/UKifNVEqlaDT6bC0tCRkMjVLFAw6nU7xSMkhkt9jWp8yEhLH2WxW6tVIWM/NzeHixYtYWFiA3+8XuoKlNwxv+f4tFovUDI4Lx9IAZbNZ/OhHPxLtg9VqxSuvvAKPxyObcm5uTkol+HkuYPIg3Ni07lzwsVhMfi43O/unMAvidDrR7/dRKBTkFA4EAqjX62g2m8hkMtLulItnaWnpE5Gz6vQ3DY7a46O3xzCNRq9UKomh4IJldkv9QX0Q24aoNwa9IoYrAA6I2lhfRDEfq8Qp+vwk3Mj29jZCoRDW1tbkcwyVGo0GXC4XFEURD4/vrlwuy0FEmUQ+n5d3R4TDYdnQJHinpqbkGTNzSYPFrA8zh8weUh9DTzYajWJiYuKB7/PjoA7z+G8ac71eL9ldejrAwSJaHqxcAyaTCRsbG7IOzGYz4vG4qODZ+ZDCUp/Ph1qthkajgevXr6NSqWB9fR3r6+u4d+8e7t27p3lAalATUa/XZXGePHlSwg3GwQRPL2Zv1OlndSkDuQWGcjwB+ZKZEucGK5VK4oVR+EiOKJ1Oo1arSdP6RqMhadEHhdpToQFiozC1gWBWi/dAPZM6VKInpU5rAx+K7miAaHBpjIH9TToajQ4I0sitqGukaNx5HQ8Kls6srq6Ktgf4UFHOCR3qMIsFwWoSFcCB0FCt0mZYSG2Q+nPU4zDsYuLh/mkhlHSoxYifxNP7WeDz5jv6acSvWhFP48NroeGs1WpiLOk9FYtFIZ/Jm0ajUckQ3r17F/l8HsvLy1hdXcXdu3fHOmDgWBoguujdbhfValUmADCLoe4CRy+GGQaKuM6ePYvRaIQ7d+7IAqPgjzFzOp2WjcvFwROfm1l9etHL4v+piVyeop+kqvj+cgOdTodMJiMbR32/6j5F6u/h5xnacLPSqDI9TYPFzCHvjV9PQ8dnQKPEscw0aG63G4PB4BO1nmWt0fe//33hZE6fPo1IJILLly8jkUhgamoKoVBIQkmr1Yp4PC4/g5qwZrMpHJI6TOXP39vbQ6lUkpE8zz//PDKZDHZ2dmTMkNlslkbvP/7xj7G0tISNjQ2YTCZcvHgRkUhEwps7d+488H1+EpB/ehCo1fGVSgWZTOaAgaLXSM5Q/X00YFw3PPT457j7ex9LA0TwJZEcVJ9oFosF9Xr9QCq+3++LcaJVpxBNvfkqlQparZZkmO7f1Pe7w/f/32Hfr9rIcbEwdUpjoV5saoOoNpgMz9T/T3Dh8aQkL6EWXwKQhU2Dpr7OBwXvg6JJlhuUSiWMRiNsbm5Kdz8WhRLcROyCoJZGqDdYLpfDe++9J4cLCVcWnWYyGQlfqINZWVnB0tKShCxGoxHNZlMmomxtbR2L1rNcc2oP6knBsTZAanDRkXz+WXj77beP4KoePWgAeK/cjOqTTk24q8WG6p8xGo3kOfH/6TGoyzTIhVAIx5CEJDEFmiT7acwfdhPwPS4tLQH46e9JfU9qHNYhwOfD30svUcPh4bExQJ9V0HioT0FmotQSBP4f8GGlOr049fdyg9GjYWaRkgcWbvLryKUpigK32y1tJY7q3o8Saq93HL//swjNAD0GuD+Dog63SCrz/9SELb0d9c+gYaH+g0QzxZ/MAKq9JQAHSFn15580PKn3dVyhGaDHBOoyCfIp3W5XqtbV/6dOxfN7+TXMuAGQokd1ES91RMB+SMKSEw7s297exvb29tik+xqeLGgG6DGCmhxXE8EfZYA+KsOmhrqGjqEZU9z1ev0nyGiK3drttsj3NWj4tNCNjqHP+XHk4+OCB32kD3uf92eJPsnX03Dc/zNYCc4QTa2N4dfwg5zRYd/ncYF2n4cHzQN6DKE2Cg+yaD7qaz5KdkBFNkOvjzMyx/DM0vCYQjNAjzEepSGgAdKg4ShxPKbLadCg4TMJzQBp0PAZxrh5q2NJQmvQoOGzAc0D0qBBw9igGSANGjSMDZoB0qBBw9igGSANGjSMDZoB0qBBw9igGSANGjSMDZoB0qBBw9igGSANGjSMDceyFuyTqDONRqM0NX/55ZelWdbt27ext7eHd9999ydGHRPqcbVutxsulwszMzMyqZPDCj9p64nDrJ5m5brVaoXNZsMLL7yAiYkJXL16Fbdu3cIbb7yBdDqNZrMJq9WKZDKJF198EQsLC4hGo/j617+O5eVlvPHGG5+ogf5H4Siq/tkH/MqVK5idncVXvvIVGU1dLBZRKpXwd3/3d3A4HDh9+rSMp1lZWcHu7i6+853vIJfLIZ1Of+w6+FnQquEPD8fSAP0ssGJbURQ4nU5cuHABiUQCiUQClUoF9Xod8XgcDocDjUYDhUIBmUxGvl/d0JwTO+fn5+Hz+TAzM4NisYhUKoVqtYpms/kTvXXGDXVrDA7fi8fjMmF0Y2ND5qxzVhS7JHKeGL//ON0XYTab4XA4EA6HEQgE4PV6cerUKUSjUYRCIZl+yv7Y586dg9VqxeTkJILBIBRFwXA4lAGDHHy4vb2NUqmEra0tGTekYbx4LA0QPZepqSlMT0/jt37rt+D3++F2u/Huu++iXq/j+eefh16vx7lz53D9+nX8wz/8g2w4jpQxGAxIJpOYnJzEr/7qr2JiYgLT09PY3NzEnTt3pAvgcZpCwKkY7AnNcb0vvPACLl26hBdeeAHvv/8+9vb2sLa2JlNMee+cJ89RLcfp3oD9+3O73ZiZmcGXvvQlPPPMMzhz5owYUs68UvfKPnfunMwt49yxS5cuSW8jDn/8xje+gZs3b+KP/uiPUK1Wj929fxbxWBqgs2fP4sUXX8TExAR8Pp/MkuKgvna7LXPWr1y5IuN4VlZWUK1WkUgkZPDhwsICzp07B4/HI2NpOHnjuJyS6okYHBhosVhgMplQr9exvr6OP/7jP0YgEEAgEEA0GkUwGMTk5CSy2Sxu376NXC6H4XCI5eVllMtlKIqCZrN5LMbOEDabDT6fD//u3/07JJNJzM7OwuPxYDgcolKpyFw2Dg+8f2wR8GHPbH693+8XD/DatWsIhUL49re/Db1ej0KhMOY71vDYGSCDwYCZmRl8+ctfRiAQgNVqFY+GTbT4wRHOnBrJMTWRSASdTgeZTAbJZBInTpyAzWYDAFnQPGmB8TfgYr9mo9EoY4k5177b7SKbzeLb3/42Tp8+jaeeegrxeFwmjd69exdvvvkmMpmMhKL9fl9GPzcajbHfH8Gw65/9s3+GSCQCm82GSqWCRqNxwGvlvZPPUQ9OZJvZWq0GvV4Pr9cr4RpnpicSCbTbbc0AHQM8VgbIYrEgmUzC4/Gg2+2iWCzKyccJmSdPnsTJkydl/G4mk0GxWESr1YKiKAAAr9cLAAgEAjCbzUin0ygUCrDb7QAAp9OJ559/Hq+++up4bvQ+cJqnoihQFAWJREI8IIPBAIfDgVOnTiEYDEKv1+P69etoNpvwer1ot9uYn58X0jocDmMwGKDT6SCVSqFYLB4bA/TlL38ZV69ehdvtxnA4RKlUQrvdlrHKJpMJwWDwwKx0k8kEr9crI5ppiGKxGID9Xtb0GPV6PRRFwX/6T/8J3/nOd/D7v//72uyvMeOxMkAmkwmRSARutxvA/kxznU4Hj8cDYJ9gZYak3++j1WqhXq+j0WjILHCS1zqdTjylWq0Gp9MpM7I43pncybjIWvI9FosFFosFbrcbTqdTxjXTCPFaeS/ZbBa1Wk2azfPeOGGUz9JsNkvWcNybUKfTIRgMIhaLCW/DsdDAh89Cr9cfCL3o6XI8EdvI8nmox2tzjNHU1BRisRiMRuMnnvKq4dHisTJALpcLzz//PCYmJmA2m8WFvnTpElqtFtbW1lCv19Hr9bC3t4dms4lyuQyXy4VAIABgfxSN3+9Hv99HvV5HuVxGsVjE008/DY/HIyetz+eTCaQccXzUsFqtcDqdcDqdsNvtmJ+fx2AwQKFQgNfrhV6vh8fjgcViQT6fRzqdxnA4RKFQQL/fh91uh81mg9PplJCDaW2PxyNzvtrt9lg3IY0LPZVGo4HBYIBisQiXywW73S6Zr06nI++NnmE0GgWwfyA1m010Oh34/X4JV9WTTmmwbDYbFEVBrVZDq9Ua272PEzxcxznh5LExQDqdDhaLBTMzM3C5XOh2u1AUBUajEZVKBbVaDblcDh6PB16vF+VyGZ1OB+12WxZ4rVZDr9eD0+nEcDiU7Aj5nuFwiGazKRkiYD8cq9frR+4hUO9jMBjgdDrhcrlgMBjQ7/fR6/VEC8TMVrFYlAXFyanA/qZrNBrCo1it1gNhK79nnLDZbPB4PIhEIggGg+LN0LMzGo0HRnMPh0MJt0hMEzRklB0A+4dOv9+HzWaDwWCAzWZDLBbD5z73OfzoRz/CysrKOG77SGG322G1WlGv10WGEgqFkEwmsbS0hFKpNJbremwMEDfiyZMn0e12USqVEI1GYbPZkM1mRbvj9/sRjUZRKBTQarVkwfZ6PZRKJSFgAcjoGX7NYDBAq9VCq9VCoVAQziCbzR75KUHi2WAwQFEU8XhIkJPzIMG+ubkJk8kkk03JeTAUbTQaqNfrspmJcRsfYN/IJxIJJJNJ4W5GoxHcbreEXc1mUya42u12eDweaaLPwYoMMU0mk/ybWc1arSYek91ux8zMDH7lV34F5XL5iTVA6nfr8Xjg8XiQSqVk+sn09DReeukllEollMvlsVzjY2GA9Ho9Ll++jIWFBZjNZrTbbRmUNxgMkM/nYTQa8cwzz2A4HGJnZwflchl6vR5XrlwRviAYDGIwGAiJXSqVMBqNJKXLTWq1WkVZ63K5RLh2lKAK2Gg0ihGh8dHr9ajX68hmswiFQkIuM+QA9jfw+vq6/LypqSmcOHECd+/eRa1Ww+bmJgqFwrEgYV0uFyYnJw/IKQwGA6xWK7rdrhwQwD53pZYjqDkcGmwA8j2UZfD5kSMym80IhUKSeHiSsLCwgHA4jOnpafj9fkxOTiIcDsPtdiOXy0nW1GQyYX19fayZ0MfCAOl0OoTDYYRCITE69Fj4d3IBu7u7kj3R6XTC91B81+/3ZeGScwD2Nyy/jrqRUCgEALKoj/J+1dofXtdgMJByglarBYPBgFgsJvdAg9XpdNDr9VCtVgHsb9pkMolIJCIq6XK5fIDkHSccDgdCoRCsVuuBufPqe2Gqne+G4RfwYepdHXbx0GEWTU1EAzjAg/EZP05Qe3gAhNuzWq0Ih8OIxWKiHj916hR8Ph9cLheazSa2t7dx+/ZtVKtVFItFmYg7Djw2Boihw8bGhqSkm80mBoMBnnrqKTkZ6VLTKBWLRVmYzIjV63WYzWZ4vV6Zfc7F3+l0EAqF5PQgz3TUoNrZYDCIe2y329Hv99FoNETR/HM/93MIBoPodruyoNTGmXzH5OSkGCumuFut1kPXRz1KTExM4Itf/CJ8Pp8YGoYPJJ2pcHa5XOIlMdSiV9PpdGRjqj1Gi8UCj8cDvV6PXq8Hq9UKr9cLp9OJubk5JBIJ7O7ujnUjPihoeBVFgcPhQLVahV6vh8/nw/nz5/Hss8/i9u3bkgUtl8u4ceOGvOsTJ06gXq+jXq9L4mKcJPxjY4CCwSACgQC63a7E8TzlKpUK+v0+yuUytra2kM/nhTMhYavT6YSM5ULjRu71emi1WpLiJcFbrVbHcjJykfFEq9Vq6Pf7MBqNkp5mStloNEKn06FcLqNWq6FWq4nXwPDKYDCg0Wggn8+L10di9jik4O12O8LhMMxmM4bDoXicNLLtdht2u13S5gSv/eMmuDKpwASG+l5ZujExMYHz58+jVCodWwPEe+92u/B6vTh58iQ8Ho9kN3U6HZxOJyKRCOx2O2q1GtLpNBYXF+Vg5fuu1WryjG02G9xuN7LZ7Nju7bEwQHq9HmfOnEEymUQul4PRaITH40G1WkW73catW7eQTqdx8+ZNOdF9Ph/MZvOBU7FcLqNSqcBut0vKll9PTQ0LNZkxIE901CDJ7PV6JVWs1+vRbrclw0fuo9fr4f3335eMHj2+drst4dnGxgZWVlZgt9sRDAbla8ft/QCAoiiYmpqSFLqiKOj1eiiXyyiXy2g0GojH47BarZLVNJlM4uX1er0D2S/yZ6PRCMViUQ4hKqIJnU6Ha9euwe124/bt26hUKmN8Ch+PUCgERVFQKBRw9epV/O7v/q4YIHJ+g8EA9+7dw49+9COsra3h3r17WFxcFG/SarXKM3O5XLh27Zpwh+l0WiOhfxr0ej1mZmZw4sQJpNNpIRvdbjfMZjNu3ryJdruNqakp4YSA/ZCr3W7DZDLBYrHA6XTC4XAgEomgVqthfX1dyhqGw6Fsap6U+XweqVTqyL0gqpt5D/z9zWZTjAa9pF6vJ8a03++LB8ef0+/3kc/nZVNOTExI+v5hDeujTN0zvOZhQVK51Wohn8+jUCig0Whgfn4eAITDo+dH7m40GomHxM/3ej1kMhkJwxRF+Qk+j+UtxyEb+HH4tV/7NZw5cwavv/46TCYT/uEf/gHdbheDwQAXLlxAOBzGqVOn5F5+7ud+DpcvX8adO3dgtVoRj8exvb2NXC4nNMTdu3flWbBEaRx4LAyQTqdDIBCQMoJWqyWaFpPJhJ2dHSnDoEter9fRbDZRq9VgtVpht9tFmJdIJJDL5bCysiK8AvkktUtfqVRQKBSO3EvgiQXgwHVRNsBNZjAYREbQ6XSEiFWrhweDAUqlEiwWC6xWK/x+/6eucXvUBojeCeUSBoMBzWYThUIBlUoF7XZb3gm5H3o4fF8kZNU1YSzXYYU9w1H+vzrTeFwNkE6nw0svvYRXXnkFzWYTm5ubuH79OnK5nBw68/PzmJubA7D/zq9evYrRaIR6vS7taijsZAJja2tLPMJxCW2Bx8QAdTod/M7v/A6cTqfwIb1eTwoNd3Z2EAqF4PF4kEgk4PV68cMf/hDlchmtVgter1c0M71eD8ViEeVyGf1+H06nE6FQCOl0WniSbrcLp9OJVquFSqUyljCFhaLFYvFAuMWN6HQ6EQgEYDAYhAujmK9er6Pb7UoBJ/UxZrMZi4uLyOfz8vmPM0LqjfxR1/YoQK6NRvLP/uzPsL6+jkuXLqHdbmN7exunT5/G/Py8cDj0cIfDoXg+DodDwjJec71elzAzn8+j1WohGo3C7/cDgIRudrsdiURiLImGB4Fer5eSonq9jr29PbzzzjsIBoPwer34wQ9+gLt374oQdTgc4tKlS/B4PFhZWRHpQS6Xw8bGhiRdzGYzgsEgotEoarXa2Piv4/nU78NwOMTi4iIMBoOchoPBQEIwpqSz2Sx8Ph98Pp+c0mazWU49eg+dTgf9fl9OXtZH0UANh0Mh/Uj4jgODwQDtdls2mjrtymunF8D7sFqtcp/0EihqVKepmXr+OGOilukf1v2zPs9gMMipfO/ePfh8PrTbbayvryMejx+4RvI+NMhqyQKNptqr4vNgQoHPq9frCeejVocfN4xGI6yurooeLZ1Oo1qt4sSJE5ifn5cMIbN7drtdDiAKb5vNpvClXCtWqxU+nw8TExPI5XKaAfppIB/DvxM84YF9roS6F6vVKp30eMoPh0OUy2UMh0M4HA45OXlqMoxhh0Gz2YxyuTwWFTQA2fxU+/JzvH96AN1uV2rG6CVw8zHEoNhuMBhgfn4egUBAOgN8XD8g8myHuTAtFgump6dht9uRz+exvr6OpaUl2Gw2lEolvP/+++K5zc3NiTAR+JCkNxgM6HQ6B5q0kR80GAyYn59HKpXC5uamHDz9fh+FQgH37t1DNBoVAeRxxHA4xH/7b/9N3gUPoi9/+cv4V//qX4l0ZGtrSw6gP/zDP8TNmzdx48YNAJDIgfdos9lw8uRJXLt2DV/4whcAAHfv3h3L/T0WBgj46FBA/TmWarDkgtoGEq70agBILxkWYpZKJeFPmEVRt209ajBOJymrhjqNTJKZhCRTzmrFL9XAalecOqifFXYctudHT2VjYwN/9Vd/ha2tLTQaDWxubkqG0m63w+fzCVGs9gD5d36+2WweKMVgpwCn03lAz8VnlUqlMBqNYLfb5eA5jri/5MRgMGB3dxe3bt0SISUTKdSC5fN5SV7QI6YHTJ1UsVjE3t6eiFLHgcfGAP0sMKbt9XrIZrMSgqhJR246VkgzHc8sAF8kNzE9iaMGSXR1iKQ2hmovR13r1u12Ua/XhaTlz2q1WqJ32tnZQafTgd1u/6mL7qj0QQaDAXfu3MF3vvMd4SKWl5fld7vdbiQSiQMqaW5Cptf5d+qFAIghcjgccLvdaDQa4u0OBgPUajVsbGyg1WrB4XAc8DSPG9QeON/L4uKi1MT5fD5cvXpVOh6QvP+osJSFvpVKBel0GktLS8hms1ox6qeF0WiEw+GQJvNcnBaLRariaXh2d3dhNBpFY8PFxwphk8mEYrEopPYHH3xwpCekmmxVf46bUp26puAQgDQd42ZlDVmhUIBOpxPPwO1243Of+xzee++9jxWhHYUBMhqN8Pv9aDQa0qng/t/JREC73ZYWGmwjwq/nc2A3BLZU6Xa72N3dlaQDC43X19fx3nvv4Vvf+pYcOAzxjxtoQCwWC65cuYJGo4H19XWsr6+jVqvhzJkzKBaL0gFSp9PhS1/6El566SV885vfhM1mw+nTp/Hmm2/KOrZYLNIze319XcqUxoEnxgAxo8LmXeo0K110tSKai1nd6FytCWk2m3A4HPD7/QfI26PCzyJ/6cmp6+GMRiNcLpcI8trttoRf5FKA/RA0kUhgZ2dnbM3WCHo16ha4atwfBtOzYfipJtvV3ROZGaQWjMXGpVIJm5ubWF9fx8bGhmTUxpmK/jio791gMCAUCqFarWJ7exu1Wg3tdhuBQADtdhvValWkFs8//zycTidu3rwpE1/W1tZk0ovNZkM0GkWpVBKPWCtG/ZRgER7rt8iL9Ho9UU7fL0unolitraFiemdnB/F4HBaLBV//+teP9F7oAd2/KNR1UqFQCDMzMwAg1d7BYBATExMyeoeegMViOVCMaDabkUgkhFv5qGzYTytxeFRotVq4desWqtXqx/6eyclJXLt2TXRCap5OPdWD1fP8KJVKUnagKAp8Ph9++MMfIpvN4m/+5m+QzWaRz+dFAT/ucpSPAr1QHor0VmhsW60W3nzzTZhMJiiKIuvm2rVrmJychM/nQ6/Xw/b2Ni5duoTLly/jnXfegV6vRzweRyaTwa1bt4RvGweeGAPElC7VwfdzJ+rZWMyW6HQ64Q7YrIoLvVaryUSGcXhAP21D6HT788yCwaAUHVosFmmkxpIGXrvaa1A37eemVVfF30+6H6aHRK7rp3kfbD9Lb4elJiwrUJPt6rYcbMDG7FGtVsPy8jK2t7exu7uLWq0mXhIw/sEDBI0N9Tr04EejkaTVub5poJxOJ65evQoAItRsNBpIJpPo9XpCTVgsFlkvgUAANpsNvV5PFPTjwBNjgFjd3ul0UC6XpSRBLcJT9xBmCFMsFlGpVOByuaTP8mg0QqFQwPT0NDwez1ji459lgEKhEGZnZ3H79m3odDrxZmq1GnZ3d9FsNkUP5XQ6DxRysqJep9PB7XaLR6gmLNV//qzreVgMBoOfWfDLNhNMn/PrSTxTHc2CYhrdUqkkxHM2m8Xy8jK+853vYGNjQ6bdUhN2nAhol8sFvV6PWq0mfbvpze/s7Bzw1PV6Pex2O2ZnZ/H7v//7ou3a3t7G9vY2nnvuOcnoMgEzMTEhh8+9e/cOJGnGgSfGADHbU6vV0Gw24XK5DkxPYOqRdUHU2PA0ofFhJkVRFCE6j6NGhF5MtVoVjYe6X5A6VOGJWCqVpNiz0+nAYrEgHA4Ln6BO0wKQhvY8fR81aDx+ms6Kng29AXJ8TDurG5X5/X5YLBb5eovFgmg0itXVVVy/fh2ZTEaMlHrDjZsHU4PdLnl/bCtCL4+g99NoNLC7u4u/+qu/gs1mg9lsxrvvvotcLodAICDcF5XTfD61Wg1ra2sIhULiBY0DT5QB4kDBWq0Gt9t9oMZHp9PJaccKcXWoxjCF2QSKFI/jxAR1LRe5AJvNdiAMIRh+VioVEWqazWYMBgNYLBYEg0FsbW3J16qJTz6H0Wi/r8yj3qQfx3Xd/zU89RlO8l3RKwL2DRDfOZ8Bx/j0ej3cu3cP5XL5J4SVx+Vw4YFBY8PiXIbSHwUeooVCAW+99ZYMMPjOd76DVCqFUCgkiujp6WkkEgn4/f4D3pTP50MulxtbQeoTY4CA/cVks9mg1+uluXkulxMxIglHbtZOpyPzxUlW+3y+n2jTeVwWKcGQa2JiQhr0syaMhCJ7v/DaSVSymZXJZEI0GoXdbsfq6ipyudyB6RTqAtjDJKN/VrqfBlHdmvWjrofKaLZgBSDvlg3YPuqUH6fnw7B4OBzi/PnzmJ6exvXr11GtViWJcL/xof6L74cDC0wmE3Z3d5FKpaQDAAW2zIJynXc6HTmk/X4/1tfXx9aW5YkxQExBAx+qhOlaq8VrzJ7wgbMimDwCpy2oFcfA8XLTyWtxbBDDTMoIeM3kwdS9ktXcB/kw9T3SAJF3UP/Ow7j/n/UzaUBpgPg5dZ0avSNKDyi8pAdLT+ko3h+fobr+jno0rjGql3ntRqMRU1NTmJ+fx97eHvL5PPL5vHQvYOaLkhCq+I1Go6i8Jycn0ev1sLq6Kq2EWTfpdDplppw6pGPJEuUZ48ATY4AURcG5c+ewu7uLbDYrsT7rY9hBsdPpyNxwEnPMDnQ6HSGr2YEQwAFjNm6oW0iw0JZkpcfjkR7YLLSkseJJyM5/Gxsbkq4n/0MjxrCLYdo4m5bxXtm72O12iwEih+d2u+UZZDIZ7O3tIZVKSeP5j6t3OwzwPVCTxpluU1NTKBQK2Nvbw8LCgnQidDgciMfjWFhYQDwex+TkJDY3N/H1r38dc3NzuHz5slS1/4t/8S+g1+uRy+Vw4sQJJBIJGcGkKAq++c1volgs4itf+QoikQi++c1vYjgcwuPxwOFwwGw24+7du1I/GIvFcPHiRbzzzjtH9nzuxxNjgAimaClwo1egbuNA40OPSf19FPGR0KXbflzCMJ72NKRqT46eDUlYZjtYdqH2FoD9US2KouDtt98WD4kbiKEBuZZxgV4DvVi+24+SWlBWwVOenu1RHRx6vR6RSAQvv/yyiEA5ONLlcsmAzKmpKTkQGV7S22Tv5mAwKNqteDwOYF+S0Gq1pHTi3XffRbvdhtlsRiwWw8rKCiwWC3K5nGQ1OUGkUqlIfyS2IrZarUgkErDZbGNb30+UAeKCZBih9lrUGRem4LnR+P/0lvi16vT0cTJARqMRjUZDJqDyc9QAqTkSFumSoGfIqdfrEQ6HceLECXzzm98UA8bMWL/fF46MhmscHiAnW/BemEC4n5xVc0XqZ0BV9FHVtSWTSXzta1/D1tYWstksgsGg6JLoZQYCAej1eikCZctdFlI3Gg3pXWSz2TAxMSHeeLVaxfLyMpaWlkQRbbVace7cOSHoNzY2kEqlRIg6HA6RzWaRy+UOFKnabDbpRjAOrRvwBBkgFmQWCgXs7u4ikUhI3E1Qqu/1eqW3ULVaRblclobeat6oUCiMfWzx/aAXoC4y7XQ6QjCORiPp/UIup9lsSkMrm80mBoyeAvkk1gRxQbJnzlF6EfeD70KtV2FYxhHOfGe9Xk/GNofDYXi9XkQiEXi93gM6sMOC0WhEtVrF22+/jfX1dWQyGczMzMBgMEjJA4cq8B6azSYymQxu374Ns9ksDfGGwyHu3r2Lt956S8SJ5LjoxdDQDodDLC0tIRgMIhaLYXV1FZ1OBz6fT7RW7B/t8Xik0JriRJayjANPjAECIO1a1VMk1CEUT0N6AOp+yuoeOjw5m83mga6CxwUMEdX1U9xc5K+40ShPYEqX7jY3A1ugmkwm6aujrjxXd2McB0gi8z4ZQtNLVR8wrOlqNBqi5bLb7QfGOx82ms0mtra2kEqlZDoLC6A5lYWeUDgcRqvVQjqdlnehDqkpq1CPjFJnDVmmcX/dHw0UVeCc8qtuvMbpKOPuBPnEGCCm1Um6shyh1WrJpFOm4TmYsNVqiQfEscVOp1O0FdyAbOtxHHrGcOM1Gg1Uq1V4vV5YrVZptsYMibrZfq1WE4/H6XQCgHh/pVJJ7pstHD6qbm5cqNVqyGQysnFpSNhsfzAYCEcCAJlMBisrK7h48SIcDoe8M4pKD9MIMfX9wQcfSG2ioiiSjWR7EHZl2NvbO+AVcY3SayHJ7nA44HA4cPnyZXS7XWQyGXkWkUgENpsNdrsdxWIRu7u7qFQqIsewWCxwu91ymJbLZZjNZpw8eVIEnTTo48ATYYCYGeJcdPVsdCqfqQehZ0RvyGw2y6Yk76Fu80Al8bhPCoILmYpgemyskeJiUitndbr9Jvfqpmxqz4idAXZ3dw+ob9VQ9yI66vslZ8ffz2xmsVgU74gcBu9F3Sv7qDggtTDQ4XDAbrfLAXjmzBkRyaplIEyI0NuhZ9tqtRAKhRCNRqEoCmw2G6ampqTtCL+eKXoq4pvNJubn52Gz2QB82DWBHlY4HIbNZsPc3JyMKv9ZWqzDxPHYVY8AVqsV0WhUvCC73S58gcvlgs/nk1G/nC/f7/elhzQA6Y7IzcfhhoqiCIcybtDQMuW+s7Mj4jt6egxB6aLzZPV6vWKsyuUy2u026vU6fD6fNAJrt9tIp9M/0axsXCS01WqV9rHc4DRAW1tbaDab+Pmf/3kxrLFYDAsLCxJ+q8ttDvv6R6ORjD6mF+p0OjE7O4uvfe1rKBQK2NnZEY+u3W7D6/Xi1KlT2NnZQTabRTqdRqVSQb1exxe+8AX8y3/5L4UysNlsQg1UKhVUKhV84xvfwNbWFpaXl8WY/Pt//+9x+fJl8aLq9bp4RWxBGwgEDuiCxoUnxgB1Oh3kcjlpxUrRGvkSdhikR0OxmLo6nEQ2Tw1OVhg3D/JRaLfbqNVqBxYl09BqEaXNZkMwGJQBf/QoSDw7nU74/X6peaOCmqlhtXxhHAaInhwr+UulkuiXmO0keL8ulws7OzsyiklRFJw+fRqLi4uH3vmPXjZD5cXFRaTTaWSzWelYyezU5cuXEQqFEA6HJVTsdrvw+/34/Oc/j9nZWTidTuRyOZTLZdy8eRPlchn5fF4mxq6vr8vavnjxIl566SUsLCzAYDDg5s2bKJVKoqHq9/tYXl4WMprPJpfLHeoz+Wl4IgwQNx5H2ACQF8r6Gs5QUrem5P+pG1jRQ6J2hrL4cYrx7gdJR1ZMswvi/bolhmsulwuNRgP9fl/CNnJhDFk4eZVdFElwq8OFcYAhCbNY7B2kVm0DBw2V0WjE7u4urFYrpqenZQb8zs7OkbQeVR9mrEy/c+cOgA9bqXg8HjzzzDMyq469q61WK1wuF774xS8KV1kqlbC9vY1XX30V6XQaqVTqgKaNXvHk5CR+8Rd/UTiku3fvIpvNCmekbk3DzqFOp3Ns7ViBJ8QAjUb7c6CWlpZkWCFnZrMfcK1WE+FeMpmU7MHe3h6KxSJisZhkBYxGI+x2OzY2NrC7u3tsRhgDEG9uZWVFRi/T9aeHwlQ7CeiNjQ2k02mUSiW4XK4D43wGgwHS6TS2trak4RnDWD7HcXk/AHDr1i385V/+pWST2OnA5/NhampKxjI1Gg2Uy2X87d/+Ld555x3Mzs7C5/Phvffeg9Vqxa/8yq/g3r172NnZGct9qNcPuZo//MM/lEOAGT0aiL/+678Wb4pENb2/jyqo7Xa7+Iu/+Au8/vrr8vvy+fxHVrqry0UMBgNqtdrhP4CPwRNhgIB9j4fFl2zXCUAEhcx49ft9BAIBqS3K5XIoFosIh8MwGAzSmExNYB4nISJPenUzfS5Aej38GmZ9KHRjWMUwjIRzs9kUMpN6muNicCm45IFRqVQOKN2HwyGKxaKkv3d2dlAsFnHixAlpQu90OhEOh4XIHjfocafTaQAH+y8xQ6auYftZ4HNgDdnjhCfCADHepuDKbDaLSK1UKkm8zBMEAMrlMj744AO0Wi0p9ItEInA6nSiXy9jb25PG9TabbWxjSz4K7HXD9OxgMEC5XBbvjdeqLt4cDAbSL4alAf1+H+vr6/KM+HU01MeB96KxpKhudXVV9D1LS0uoVCrY2NjAzs4OvvOd70BRFESjUczOziISiSCRSAhhz2zncYNa13Oc1tlR4IkwQGzDEY/HJcNFVTPFVwaDQcjWU6dOYTAYYHJyEo1GA51OB6FQCAaDAXt7e8jlckilUpibm0MgEDhQMT5uMGxSFAWJREKyWo1GQ4wSNSDqFrTqGegsrlUX3TJ0peFRV5eP0/tLJpO4cuUKYrGYaJyo+OXs80QiIT2umf6enZ2FoihQFAXtdhtra2ufuc39OOCJMUBMdzJjEI1GJSOiKAoCgYDoRubn52G32+FwOFCtVmXmdi6Xw+bmJra3t7G6uoqFhQWEw+FjZ4D6/T78fj9mZmYkvdpsNuF0OuH1epHL5aT5PON8l8sFq9UqOhTgw0JPekRszs4iXvIEwHjS8DqdDvPz83j55ZelQp+hFctsms0mLly4INdOY8lM1GAwwNLSEt577z1pyKbh+EA3Og49Ju7Dw5y40WgUV69elYyQevQwiTi2omBDLhKYJPnYT5qan0QiAbPZjKWlJfm6B8GDPtKHuU+SlNPT04jFYtIF0GAwwOPxIBgMSoaEHg4blun1+gPNrjqdDprNpmhm7t27d2D8r06nk3ldvHf1vR3mfRILCwuYnp7G/Pw8HA6HCPJOnTolRHkymZT2tIuLi9jc3MSdO3dEK1OpVFAqlbC2tvZQRugo7vM4YBym4InwgIB9Q7O6uopwOAyfzydK0XQ6jVqthnK5LC0JGJKw/omaEnoXJAk5L/vjZlaNA7zGTCYjRpLTP0keq1tW0NiQXFZ/MHuk9h6ot6EnpG51MY5ncO/ePSwvLyOTycDr9cLr9WI4HEq7CnpC7PK3tbWF27dv4/vf/76Qsg9K5mo4ehxLD0iDBg2fDRwPYkODBg2fSWgGSIMGDWODZoA0aNAwNmgGSIMGDWODZoA0aNAwNmgGSIMGDWODZoA0aNAwNmgGSIMGDWPDsVRCf1pJu8ViwX/5L/8F8XgcpVIJ2WwWu7u7UmBptVrh8/kwPz+PYDAokyOXlpbwb//tv5XyjYeFJt0/iE9znywo/r3f+z3pY8wWtGy3UalUYLPZpOKd/a+3trbwB3/wB1hZWcHKyspDX8NRvk/WuQUCASwsLODixYt47733ZNggi2/Z8+nu3bvI5/PY3NyU7pUP2wNbK8V4hPD5fDLcLRaLYXp6Wpp/b29vy3gap9OJYDAoY3QPe3aUhk8GzlV/5ZVXcPr0aRm7VKlUpGMg6+NY1c956Yqi4Nd+7dfwrW9961MZoKMExwgB+0MIt7e3pY3M9PQ0kskknnnmGRnYuL29jXK5LG2GAUgL4cdhHT+RBojtOfR6PdbX11Eulw+0nazX67BYLKhWq7BYLNIWs9VqSQPw49ALRwOkxQjn3S8vL0vNGodGplIpAPsN7P1+v8yK7/f7iMVi8Pl80v/7uIL3+Zu/+ZsAgD/7sz9DPp+XomLeTzqdxne/+120Wi20Wi2Uy2VYrVacP39eipL39vZQq9Wwt7d3ZBNBHhZPrAHiXKtUKiWNuXkqsFdyo9FAIpFAJBKRvtJmsxlGo/FYzADTgAMtS+v1OjY3N2UqCDtf8t2yPS0b85tMJkQiEbhcrgPjio4jzGYzXC4Xfv7nfx6DwQB/8id/IlNYPB6PeOy1Wg27u7syZMDj8cDpdGJmZgZms1lGSOVyORQKBemCeVzv+4k1QKyKr1arMBgMSCQSMoqm0+nA6/Xi8uXLSCQSMiUgHA7j7NmzWF9fx9ra2rhvQ4MKDK1CoRCcTqeMUuKsM5PJJEaJs7Y4fBI4/nwbZ3y99tprMJvNmJqawvz8PK5evXqgKyQb0L311lu4c+cO4vE49Ho9CoWCdHLweDzw+Xzwer2oVCoyrJAtfI8TnkgDBEAWJadA8E8AMk3BbrdLk/dwOAyHw4FEIjHWKQEaDkI9Mpv/5hgiNkvjIEqn04lms4lWqwWXywUA0gWRUyGOqyfgdrsRDAZRKBSkeyWnlNCrY0uR+/tF03tn62E+j0gkArvdLo3sNQN0BGDvGvZ6np2dRb1eR6PRkJfk9XphsVhQr9eRzWYxGAyQTCbh8/nwwgsvoNls4saNG+O+FQ340PPpdrsyJqnZbMoobYvFgjNnzkhT9uXlZaTTaXz5y1+G0+mUZvsceXwceSCdTofnn38ezz33HG7evCldPb///e/j1VdflVl1J0+eBLA/VSOXy6FarWI4HMLhcMjseM4Ac7vdeOmllzAcDhGPx/HWW2+hVCodO27ziTNAnKtEq08Ck4bH5XJhenoaVqsVRqNRpoOyjzQnRj4JYDiinn/2cV9nMBgQCAQQCARw4sQJyahkMhkUi0XcuXNnLFwChyjW63XpYDkcDmXChXo2GL0Eeg2tVksM0nEaraQGrz+dTuP999/HxsYGqtWqjJACAIfDISFlp9NBpVJBq9WSA5UdQNm1st1uw2AwoFQqSedPl8sFr9eLcrl8rJqzPXEGyOFwwOfzodPpoN1uizGq1+sIhUKwWCy4evUqer0e0uk0Wq2WjHWxWq0Ih8Pivj/O0Ol0MJvNssCbzeZHbkC2NbXZbJiZmcG5c+fw67/+6zCZTNDpdLh+/ToWFxexsrIylgmx9HLYSpXTYBVFEc+AXRFJuDLUajabuHnzJjY3N9HpdI5l+MVRUGtra9je3pZRyd1uV3qa+3w+2O12yeRxkMJgMBCj6/F4JNzkLLe9vT0oigKPxwNFURAKhWSK6nHBE2eAuJmYIclms3A4HLhw4YLwCBw7o9frceLECZw8eRKdTgd7e3uSNXvc4XA48K//9b+G2+2GxWLBzs4O8vk87Ha76GUURRG9lMPhECJ3enpaZk194QtfwMTEBN58803s7e0hk8kc6X0UCgUA+xyO3W6XSa+RSEQG+bFlrsPhwNmzZ8X72djYwLe//W1ks9lja3yAfc/NZrPJ2u31etjd3YXdbkc0GkWtVpOMGO9TURSZcApAPCdODjEYDNjY2IDNZhPCPhwOY3t7G+12ezw3/BF44gwQyWemXev1OhRFQTAYlH7Km5ubACBENNO8dG+P0wt6WJjNZly+fBmBQABWqxXr6+tIp9MyGZUz4cPhMJLJJBwOBxqNBgwGg4x5HgwG8Pl86Pf7iEajaLVaR76ZW60WKpWKZLaYWCC5OhwO0el0xNPj9I+NjQ0UCgVsbm5KJuy4gaEvkySKokCn08koJB4UpVJJ7oH36XA4ZJIvJ+AOBgOhFjhSaTAYwGAwwO12y7s/TjheV/MIwBc5Go3QbreRy+Xg8/lgs9mwvb2NYrGIu3fvIhQK4dlnn5VpFwaDAd1uV/igxx1msxkXL15ELBaD2WzG3Nwc6vW6GGeLxSIzyenGc1ZYuVyWUMtiscDtduMXfuEX8MYbb2Bra0tCnaMAeQ69Xg+bzSYTTgeDAfL5PJrNpozgTiaT8j2FQgGFQuFYTXklOMGW2SqHw4Fr167hxIkT+O53v4terwev14vRaIRMJoNut3vA26Ew0WKxoFarQa/Xw+fzwefzIRwOY2trSzRC3W4XxWIRp06dwvz8PH74wx+iXC4fG4/wiTNAJC05qhjYf2FUjjKDoigK9Hq9zBQPhUIye/xJESFy/LTRaJQMEFXE5IcAYG9vD/V6/cC8cP7ZaDREUTw9PY3Tp09jZWUFlUrlSO6BoSBLLQBIUkFdNjMajYSY5rDF43baEw6HQ+oR6alzJFSr1RLPhcaXym6OW6L8gIJLjuFm4qXZbKLRaMBms0ndXKPREE2c2Wx+4BFTh43j+YY+BZi2rdfr6Ha7Mnonk8mgXC6jVquhWq3C7XZjMBhgb28PW1tbCIfDMBqNKJVKEm8/zuj3+1Ki4PV65f4DgYBwKcPhEN1uFz/4wQ+wvb0to6n9fj8URZHyFACYnZ2FyWRCMBjEH/3RHx35kD9Obu10OvIxHA6F6zMajfB4PKjVami320K+HkdEIhEEAgEEg0E0m02kUilsbGwgnU5jd3dXjJDRaES328WFCxcQDAZx+/Zt9Pt9mf6rXqec99ZsNrG5uSlENeelLS8vo1arwWAwwOl0agbosBAKhTA7O4vt7W2ZmR4Oh2UsMwCpftfr9YhGo3A6nQiHw2i1WrLQH3cMBgNsbW2Jt9Dr9WQD88Q0m82iug0EAuj1enA4HIhEIiJ2o9dEz9LhcBy5Z0Eur9FoYDQaSXaIk13VNVMMLwOBAHw+37FUQFut1gO8I7NWzWYToVDowOExGo2QSqVQLBaRTqcxHA5FdgDsizDJFQGQekZmzfR6PRRFkcOX339cRJlPlAHS6XTw+XxIJpNYXFwUL0iv18Ptdsu/WXE8Go3g8Xjg8Xjg9XqFZ3gSDNBwOEShUBAjY7FYYDKZDgwlJA/EWri9vT3YbDaR8DcajQMGiAubxOlRgipfAEKi05AWCoUDgxS5IY+rnMJisQjRT6NJry4Wi2E0GslARQosDQYDyuUyAAgZz7ov/rxOp4NGoyH1jDRAZrNZJv5aLBaptj8OeOIMUDQaxcLCAnK5nGS5OE2TStpcLodOp4M7d+5IdoBkJSuKH3dYLBa8/PLLUBQFtVpNFOIUJLZaLVSrVYxGIySTSbhcLsRiMREtVioVFAoFuN1uOTFtNhsCgQBcLhfsdvuR3YtOp5OpqKzwZlnBcDgUo0jxHsdO5/P5I7vGBwULSoH9PkZzc3P4zd/8TfzoRz/C4uKihLxTU1PSeoStNlhw6nA4UKvV0Gg0xDCTcG+325iZmYHH40EsFkOlUsG9e/fESF26dAn1eh07OzvHQg/0xBggbjBWT3P0ss/ng8vlgk6nk7CCJB5dYJK06szE4w663jabDeVyWTwg1lJxo9Ig9ft9qbgejUZCQvO5Go1G2Gw2ABhLmEp5Ba+J3g4AIdR5iNBrO47hl3odMuQ9d+4ctra2sLq6KloeRVEAQHgurmvWvFFASsJaTcbz7+p6MqfTKcW86rrIceOJMUCM/YEPsyQAsLCwAEVRDsTNsVgMLpcLoVAIlUoF6XQaJ06cgN1ux9mzZ2Um/OMMZlYGgwEymQympqYQCoUOpKRphKrVKkql0oHK8mAwiEAggHK5DL1eD4fDAY/HA7PZPLZyFYbIDCvJYTBrx3eu0+kQj8dRrVaPzUYDPjwkmYl0uVyIx+N46qmn8PrrryOfzyOVSmEwGMDr9QrRPD8/D7fbLRkwvV6PRCIBu90uaXqSzaVSCYVCQcK1ZrOJdruNU6dOYWFhAV6v91jVgz0xBshgMMDhcKBSqWBtbU08oGazKZoJk8kkFfB6vR5er1dib9Y5MbZ+nEF3u16vw+l0isEwGo0SurBGbjgcimfBJm6j0UhO1W63CwDCpfGEPcpFTN2MTqeT3813eH+nA6amgQ8r6Y9TMzL1PQCQrBX1PAx3/X4/yuWyVPcDH76DYrEomcpqtSrPwmg0wu12S2sOj8cDq9Uq5H02m8XKygqy2eyxeR5PjAEymUzy8O/cuSOK2EqlIqEZRV+dTgc6nQ6hUEhiabrulLEfp0X7SUF1d7Vahd1uRzweh8PhgMFgkPSroiiykbnonU6n9KUZDAbodrvy9RRrsjPfUfEHao+B4SPbT5CEtVqt0rKCHwzDaFyPw7ukIVUboGq1ig8++AClUkk8PIvFglgsBgAoFouo1+vCBzF8C4VC8Pl8EqIxoeDz+VAqldDtdhEOh+X3dLtdbGxs4Ec/+hEqlcqx8YKeGANksVgk3c5MGLAvsut2u+j3+1JlvLS0JCeny+USEVi324XP50MgEIDX60WtVhMP4HHC5OQkpqenxfPR6XRoNBqo1+sHsnw8NVmoypOUHgU3MLD/fLvdLmq1mmRUjgKKosDr9YpKnR6OutIf2Pd8isWiGJ9WqwWTyYSXX34ZS0tLuHPnzpFc74OAzfK2t7eFdwyFQkgmkyIe3dnZgdFoxMLCgrTRePbZZ1EsFnHz5k0pKnW73eh2u9jd3cXMzAzm5+eFgqAHfPHiRdEZJRIJOJ1OrK6uHos0/BMzlodFiuQIHA4H7Ha7kH0UaVWrVeTzeRHSMTzji2d5gtvtPrZK2p8GejJ+v1/SsQaDQTJDav5ETVzy3/QcSEKznQMb+tfrdXQ6nSMzzAwh6ZENh8Of2DgMa3jQ8O86nU4M8XECi4CNRqNwdSy1MJvN0Ov1kinz+/2yDmm49Hq9ZLyofqZX43a74fP54PF4xNOPxWIHild5KB0HPH477GNgtVoRj8dFD5HL5WAymZBMJuHxeBAIBJDJZJDP55HP51EoFIR4pVdAdz0cDuPSpUt46623pMXB4wB20gP2PQJmv3w+H8xms2S6BoMBisWibGaXyyUZFRoiZpui0aiojLnQj7K3zvnz5/Hss89K+tpoNEJRFAQCAWmxQREeOSy9Xi/K4MuXLyOdTh/JtT4IdDodJiYmMDU1hUqlgmq1iv/9v/+3lGcEg0GUSiWsrKzAZDIhHA6Lscnn8+h2uwgGg1IsXCgUJJSmiNRgMKBWqyGVSsl4n0wmg52dHaEgjgueGA+IPA9JR8rZyR+wCX25XEan05GTUk1U0p1XFAUTExOSdn5c4Pf7cebMGZw+fRpzc3NyupK8VGcCmWrnvQMQg8RnwWptGudWq4VUKiXdJo8CPp8PExMT0mKDIRa9AHUCgffIbKher0cwGJQq83GDnqXL5YLf7xeh5ObmJrLZLKrVKmKxGJLJpLQZKZfLMJvNomZmobDZbBYx5mAwEAEpw7hSqSSevKIo0ry/Wq0eq9asT4wBolaCC5Czo3ha63Q6ZDIZbGxsHGhcRSUptSSj0QjhcBjXrl2D1+sd9219IiwsLOA3fuM38LWvfQ1f/epXpbi2XC6Lp8fTUt3WgQJFPier1SpqW8JgMCCTyeCdd95BqVQ6MgMUi8Vw5swZ8ejUNU61Wg2lUkmMkHouGIADZSbMmI0TNPihUAiJRAKhUAgAsLi4iKWlJayvr+Py5ct49tlnYTabUavVsLa2Br/fj8nJSWxvb2NnZwfNZlMGFPKe/H4/Zmdn8fTTTyOTyeD9998/wAWS+1xbW8Pu7u6x4H+AJygEI0fALAFLB6iAZtfDTqcjnBCna1LEBgCNRkMWyeMiSDQajdIJMhAISEsR9v/xeDzCFdDQqFt8ckNTqEjDTBHbcDgUvuGo27LSE6M+iQcHs3C8XnpqvFa1eM/j8SASiaBQKIy11QqzValUChaLRZ47BbH02vv9vqzJbrcLr9cLv9+PGzduCE9ptVpl9JTNZsPp06dhtVqxtLSEdDothw6zbKPRCBMTE7BarWK4NSX0IwZPd3IYrJRut9uoVqvywtUqYH4dv6/VasFgMMDr9R57PZC6SJQ1bYqiiKEtl8tC0KrJeABSF8b2DYPBAE6nU7gfbnz2V1aHrEd9j/eHhDRAan0QQ3B6vOp2tIqiIBqNjr3XE6+HNXpq5TaNEK+f6Xq2Ffb5fHKA0MtngsFqtUqZxsbGhqTuWbKysbEhnr3f70exWMQ777xzLApSnxgDxPiflcbAfmXwysoKJicnsbCwALvdLmldu90uBonpZbPZDJ/PJwV7PGGOg4bkfthsNtjtdrz44osIhUKIRqPQ6/V47733JIRi9oPEPI0r+ZNarYZisSibtVKpSPaQi7PX68kmj8fjePHFF3Hz5k1sb28f2X0qiiKbsdlsSijJDUhPjc3JAMh1D4dDPPvss5iYmMB//a//Fe+8886RXPdHgUJYvjs211crutU6K667YrF4YH17PB40Gg3p+ROPx/HCCy9gfX0dr732GgqFgpD2rHX0+Xyo1+tYWVlBLpcbu+EhnigDxD435HK4APV6vbisDKtYSa0WuPGE4enCj3FogbiRdDrdgfsij+B0OmG325FIJBAIBOD3+0X7wYVrt9tlogKbnaufC1tBsLbr/kWp/vdoNILdbkcoFDpAZh82OMWWHiufCZ8FJQS8Vv6p9ixcLhfm5uaOtID240AvyGKxCC9HL5aJEjVvSTV/tVoVSYTdbke320Wv14Pf74fH40G73Ua5XEYqlZKwjJ4WU/IsV2Fy5jjgiTBAOp0ODocDyWRStBUszZibm8PU1BQSiQSCwSD29vZQqVSk9zNPpHK5jF6vh0ajIYve4XDA6/Uil8sdqRek5mAMBgPOnz+PcDgMRVFkdBA9G/aSaTQasNvt8Pv9EoIyJKMHQSMLQFTR7XZbdCOctcUwSy1EZL/haDQqae+jALkf8iNOp1M4EHoK1PzY7XZpsWs0GmWcMw+fceu6aPSZsQoEAmi1WvD5fPD7/fB6vbh58yZarRa63a4Ym1wuh1KpJK2F2dtqNBphYWEBOp0Of/7nf45sNoudnR1MTk7CZrPh7bffhslkQigUQjgcllo+dTHxuLt/PhEGCPiQiGXowJPQ6XRKWl5dQU0FKqXvNFwkrrkxo9GoaGYOA/RoCE5oDYfDsNvtcLvdmJubk7HDXID0ACqVinA7drtdhviRF1F7OhzOyPscjUaSMeLPIK8CfDiBk5wLy12OkhtTFxazIp/clPpDPUO+2+1KGMrDhC07xgl6bsCHxoglLgwhyWNxeonFYkEwGITNZkM+n5fODfSAdnd3AeyXdNCzcbvdcDqd8myKxSKi0agYoKPm8X4anggDRI/B4XCIkE59KhoMBina42YnOckQhSEIG5U1Gg2EQiFMT0/j3r17h3JScEHabDYJjebm5hCJRHDhwgVEIhEkEgl4vV6YzWZpNuX1eqXKmVwWQymOJ1Zni5rNprSipQdEyQKrrNkzW12vRA5I3Uea43uOAtTCkNuhITEajULYsncym+eT61IfLPQkxm2A1IcNuxDw3tRdKwFIz3Kj0YhEIiHJBRofvn8qpgGI5+fz+eB2u+Vzo9FIuhio5SbHgQd6IgwQORz2UiFfYTAYZGNSI6IeW8JJGOqTn/VQ3ISHGXpFIhFRXZdKJdRqNUSjUZhMJqyurqLX68HpdMpsLHIv6+vrB6a5qofzVatVWWQ7OzsHihHJAZlMJni9XjHazJAxi6gmnrlhmIFSz2k/TNDbYWMxtcJb3VFwOBzK7LPRaIRyuYy9vT0xlOoMmvq9jgP05igSJEelFooygULjRG5Hr9djd3dXvD2+DxLZc3NzIi0ZjUYoFovS1aDRaKBSqaBUKh0Y1cPnMI6Jt8QTYYAYdvF0v1/HAnzIJagzJzRW9JrUi5MnLivFDwNMD1+8eFG0G1arFcPhUAR2rVZLCg/J5XQ6HWm+Hg6Hpf6NLjdPzna7LW4+QzL+H5+DwWAQhTEXJsMuVp2T7FWTv4dthFgUy+6NTL/znalDRd4PdU18Xuqv5ZpgX+xxgIkP9aGnroxXq9VpnJrN5oHJFnzH6lBUr9fLNGCG2eQGGQ20Wq0D38/nx7WvGaCHBBcW3Wv2SaF3Q7GhoijSFY4EJSeFMrwBIO04DAYDTp06BafTib//+78/lOF2k5OTOHv2LL785S/LZtre3ka1WsXu7q7ok4rFooRIDLNIBFO7UygU0Gg0UKvVpG3q5OQkWq0Wdnd3hbglkelyuUQftby8LOpmkt3ValWMn8Viwfz8vGzgo9CPMDSlwWXZQa1Wk8brDIvJa9FY8ZQfDAZoNpviHUUiEUxOTmJtbe3IRXjshNjpdIRbTKVSEkbS+LAp/d27d8VQNhoNmZrKTpcEDRQAZLNZLC4uwuv1SgcEZtRKpRJ2dnZQKBSkk4Ga6xsXHnsDRDD1yI3GBQl8qK8AICcjP8htAB+mbnkyeDwetFqtQyPtisUidnd3cevWLSFNW60WBoOBkM5sIsZwghyQ+iTnAqaGhx9sL8taIWZAgP1+xPV6HcViUUIytrBQl2q4XC7RRanJ8sP2gFi5zQOCBwwN30f9firg6UWynxFDuUQigfn5eZkWMg5wzambyXNMNpMLag+J98IeQFarFX6//0CjMovFgna7LV6OoigHWqrc72kBHxqfcafjH3sDRBfSbDbD7/eLd9DtdmXRMhPCGJyEH18wR9be31M4GAwe6oC727dvY2dnB6urq/B4PHC5XDh//jyCwSDm5ubgcDjgcDjgcrlQq9WQyWTg8Xhkymmz2ZTm4kzPR6NR6e/Dlg8TExMSkuZyOakLyufzyGazeOWVVxCNRrG7uyvPiS1YSdLT0zoqV91qtSIWiyEQCAi5rC6gZciprtxnRXggEEC/3xdvicmIq1evQlEUvPnmm0euiOZGV4sOvV4vqtUqfD4fJicnEY/HxWsnd8nGcb1eD2fOnEEgEMDMzAxWV1eRSqUQDAYxHA6lUJVSChpudXfLXq8nhk+tIGct3Tjw2BsgSvD5EnjSMwukbs9ZLpdlhAv7B5GU9vv9QvCpiUs17/Goycter4dqtYqNjQ3JSm1tbcFutyMYDCIYDCISiSAej0NRFMzMzMjAPV4TW8x6vV7YbDbhGNiqgZmsRqOBZrOJ3d1dNJtNNJtNOBwOnDhxQnof0c33+XzicVmt1gOGnOHpYWeU3G43rly5glgsJnVT7NZoMpmEVAdw4JBQt5elp8HNyKr6o9YDqWvVWILBcLnX60FRFHS7XWQyGSkgppHodDqw2+3weDxwu92i/SoUCtjd3ZX1Hw6H4fV6xWsnWc/3yIQMQ2tynPSexqUHeiIMEMOrarWKUCh0gGhWu+z1el2adVNNTPeeJ406PakmbQ+jrSc5CnXPoZWVFdn4sVgMs7OzeOGFF2AymTA9PS1hCWdKMdx0u91CuvPEZKq62Wwim82iWCxie3tbSEyXy4VkMinTMEj8ulyuA1mjj/s4TDidTszOzsq8Np7g1CbxPtV1a+rrYgsRPks22qf6/ahBY0idD6URnIxBKQVDa/UYanrpLA1qNBoolUrS84r/x7CbHBjvnz+Lo8kZutMAjVMX9NgbIKJcLuPGjRtwu92IxWLiBbHL3Gg0Ekk7RXV+v1+yJhxb6/V6USqVkM1mMTs7C7fbjbNnz8Jut2NpaelI7oVCyO3tbSEWnU4nZmZmEI/H8fTTT0tP4FgsdqABP2eA0eOj+03Phrqf0WiEYDAIi8WCYrEoxpAKaYYM6sGOwIek/2F3CqjVarhz5w7y+by0RTEYDCIs5SYG9vkshphsWWEymSRsJEdGr2Ec4QavgZMsqOGpVqvY2dmRLg3UsnGiBxXtPp8PP/7xj0XpXSgUUKlUhPfK5/NSikNSnuR7t9uV51Ov18UL7Pf7wjmOC4+9ASIHRLeWL4gbiNoXANJ2lfEvNRJ8KcCHvY/T6TSSySSsVisikQhKpdKR3pdaKdtut2UeOBtUxeNxKckgN8PG5UxJqxXc9Jh4+jNLxFQ/q7GdTueBjB9riDjayGKxSKuHwwTfK4V6ahkFOT2GZTREnU4H2WwWuVwO9XpdDBYJaXIrR51y5r3crzJnWMywmQpudh9QS0m63S4KhYJ4zAyz+F7L5bJ8nVpsyt7nvH+GWjTM4xYkPhEGiE2qqHje3d1FMBhEp9PB9va2zIFnQ3X+O5VKSTe6U6dOSZe6fD6P733ve5iensbU1BQuXLiAfr8/tkpqLqrV1VWsr6/jBz/4gcT3zOCpyy/oVpMjUWthuCjpRdyvF7m/qPP+8hWqdg/71LTb7ZicnBSuh6d1LpdDIBBALBaTFqUmkwnVahX37t3D3t4eyuWyzEDjzPtMJiP8xzg2nLoHNwcnMBymkLLb7SIQCGA4HKJSqQgJvbe3h0KhgGKxKNkuerEMs4APe2LxPbIOMJlMYnl5GalUSpIv9IDURb7jwGNvgIB9b8HhcODkyZOIx+NSIex0OqW0wWazYWJiAjMzM7h7967wJslkUsg7nj5erxenTp1CMBiE0+mEw+E4Fs3JmNFgWHU/V0Ajw5ogtZdCI6JWMpMr4ilIb/GnbdCj6iHjcDgwOzsr18M2uvV6HX6/H9FoVIwSPUGLxYKZmRm0221MTk6K10Z+KBgMyqY96l44/L0sfmaB7WAwgN1ux5kzZxCJRKQNa7ValWJotVqaMgmHw3GghMflcokHzH7Y9IQ53UUtvKWROmy1/8/CE2OAXC4XLly4IKEJ++1yCoDdbseJEydQq9Xw/vvvYzgcwu/3IxwOQ6fTyfyw4XCIUCiEa9euIZFIiLboKFtQPAjUZRP34zCVvke1aRVFwenTp8WAsJd3uVyGz+dDJBKRVDs5vzNnzoiXxjBRXY4RjUalRGMczbgoiKUBajab0Ov1CIfD+PznP48XXngBd+7cwebmJnZ2dkRawdIbJkN0Op1IFBiqBQIB6e+UyWRE9dzr9ZDL5dBsNg94teMIRT8KutFxuIr78DDiqEAggNOnT+OZZ57B2bNnMTc3h36/jx//+MfiMbz66qtYXV3Fzs4Opqam8JWvfAVzc3NC5JLvyWQy2NrawpkzZwAAv/M7v4PNzU2sra090LU86CMdtwjs0+Iw7zMYDOL8+fMHpkWwYJhTTra3t9FutzE9PS0hKXmTbDYrnA9Db5fLhW63iz/7sz+TaaFHdZ86nU5mf1EcSk+cGc9gMCj8Ti6Xg9vtRiAQkMJa9vUplUqid0ulUsJ5stUq+SG1N0uS+6fdyzhMwRPhAQH7WbAf//jH0lclEAig2+1ifX1dyL7FxUWkUin0+32USiXcvHlT0pQsyuSiVadJ7969i2w2O+5b/Ewhl8vhtddek/KZ+fl5aczmdrul6JTcHw0QdUB7e3syBYVeATc+RY1HCYY7JIyZEmcBcaFQkLVIjoZ8HjVAwWDwQIpdp9NJuQy7enKIoTqsur9p23HCsfSANGjQ8NnA8elMpEGDhs8cNAOkQYOGsUEzQBo0aBgbNAOkQYOGsUEzQBo0aBgbNAOkQYOGsUEzQBo0aBgbNAOkQYOGsUEzQBo0aBgbjmUphlYj9fGIxWK4du0aFhcXsbm5KeOKn3rqKezs7ODevXsf+716vV7qprrdLiqVCvL5/ENL9B/0+9gu4mF/j06nQzweRyQSwec+9zkEAgEEg0GkUilUq1WsrKwA+HDYotFolLlphUIB1Wr1wDjuB6n6f5j7/Kys20eJY2mANHw88vk83njjDemBpNPpMD09jT/4gz/An/7pn+L3f//3P/Z7jUYj/sk/+ScIBALIZDK4efMmcrncoV/zp13YRqMRzzzzDM6cOYOvfe1r0vPp5s2b2N7eBgBp1u/1euFyueDxeDAYDLC2toatrS2srKzg3r17KBQKqNVqY52FpeFDaAboMQM7APIU5/gcr9eL+fl5vPLKK7h165ZMUyUcDgfcbre0dU0mkz/xNQ+Co25j4XA4oCgKLl++jEgkgldffRWKosDn82FmZkYmS3BeWrlcRjqdhk6nkx5QJpMJVqsVXq8X+Xwe169fl6mjx7VI87MCzQA9ZmAjMs4I4zhfh8OBZDKJixcvYmtrS4wLm5RxsgLnzQPAzZs3P9HvZl+doxzla7PZxHA6HA688cYbmJubQygUQiKRgN/vRzqdhsFggMPhQC6Xw97eHprNJjweD86dOyftXN1uN4rFIpaXlwFAugGOsyfyZx2aAXrMwGGBL774Is6dOwe9Xg+/34+dnR3UajV4vV5p4sV2DoqiIBgMwu/34+7du6jX67h69aoMMnzQuVAul0s2vHqSx2GA/Zn8fj9CoRDeeecdeL1enD59GnNzc5ifn0e9Xken05FxO5xx1Wq1ZHro66+/Lj9zenoasVgMVqsV2WwWS0tLWFtbQzqdPtCO9rjh/pliTxI0A3QfHuZlc+IoQ6PDgMFgQCgUgsVigcPhwNTUFObn51EsFqHX67G0tIRGo4FwOIxQKIRisSgDBsPhsAw+ZHtSo9GIYDCI06dPY3l5GZVK5Wdeg9frxcmTJ2Uo4mGCYSUJZ/Xo7Gq1Ks3IOMuN7WrV/XKGwyGq1aqMw+n1erDZbDK6qdFoyPjifD4/ttlYn2VoBkgFzsWia/6gvXLD4TDm5ubwwQcfIJ/PH8q1ORwO/MZv/IYYugsXLmBiYgLf+973kEql0Gw28aUvfQlf+9rXUCqVMDU1hXQ6Db/fj4sXL8q4lmeeeUa67H3pS1/C888/j//4H//jAU/ho6DT6XDt2jX89m//Nn7v937vUBu0scXo008/jdnZWUQiEayurqJUKmFvb08ajHGs8cLCAvR6vTST6/f7CAQCsFgsMpTQarViZ2dHSPtwOAyn04lgMIh0Oo1vfetbKBaLh3ZPGj4amgH6R7ALncfjgdlshslkQq1Wk2mSP80jmpiYwJe+9CWZ1XQY1zYajWSoYjqdxrlz5xCPx/HUU08hHA7jxo0bMk6IHfI4L8xgMCCfzyOXy0nTdnbTq1QqMsblp2E0GmFjYwN///d/f6jGx2Qy4eTJkwgEAnA4HCiXy9JStdvtol6vYzAYwOPxyERUemMcU8M56JyHNhgMUCqVZCoIOyi63W4oioJYLIY33njj2BqgB/HGH9cwTTNA/whOWA2FQjIIbnt7G+VyWWY3fRxOnDiBX/u1X8Obb76J9fX1Q7m24XAons69e/fw1a9+FVNTU/jSl76E9fV1bG9vy+iefD6PRqMhTcxHoxFSqRTu3LmDaDSKWq12IGP0oOHU7du3sb29fWgblZzV1atXYbPZAAA7OzvI5XJwuVxihL1eL5LJpJDhHC2kJsd7vZ5M/iyXy8hkMvD5fLBYLEJYRyIRmTH/h3/4h2NpVP8owAGGj6O0QDNA/wi9Xo9kMon//J//s6S6vV4vdDod3n33XayuruL1119Ht9uVIYicMPmDH/wAv/u7v4ubN28eGOr3KNHv9/HBBx/g0qVL+F//63/hzJkzMpDR6/Xit37rt1AqlZDJZGA0GmWiK0MTbraJiQmEQiGZ0JlKpcQj+lmLl/zKg3hMDwPyW5xC0mg0ZM47s2/9fl8mjLJnN42N2+0Wj48Tb9kPvFarIRwOw+FwiBaoVCphfn4efr9fJqeQa3ocoNPp4PF44Pf78eyzz+Lu3bt49913x31ZnwiaAfpHcNM+/fTTaDabyGQymJ+fh8VikfEud+/eRbvdlozJaDSCy+VCuVzG66+/jlqtdmgkNGc5eTwefPGLX0S320WpVJJxQidPnsTq6ipSqRQAyEjf0WiEXq8Hv98PRVFk/hn5Eg65exADxFFAhwWbzSbTHoxGI1qtlvA49HAoBWDIQZ6OWTM2p+dBwcGGDNf0er2EaeT5OFfdarU+NgaIDfip74rH40in0+O+rE8MzQBh/ySJRCIIh8MYDAYwGAxQFEU8oVqtBo/Hg1/+5V+WhX379m2YzWZ84QtfwBtvvIG/+Iu/OLQBbxzd8qu/+qt46qmnYLVa8Xd/93f4/ve/L2NqXn75ZZnxXi6Xkc1m8e1vfxvJZBIvvfSSbOx8Po9isYjXXnsNHo8HiUQCTqdTRh2PE6dPn8b09DQSiQQMBgN8Ph9arRba7bZod6anpzEcDpFOpyW75fP5oNfrZUAlpQVms1nmbiUSCWxvb6NareLMmTMwGo0YDAZwu91wOByYmJhAuVxGqVQa66C+B8Xs7CwmJiZk2uuf/MmfoFarjfuyPjE0AwTIoDe3240f/vCHMj+di7jX68FqtWJqagr1eh31eh3AfkiSTqdRrVYPbdGqT/psNou7d+/i//2//4fvf//7+OCDD+B2u9FoNHDnzh35WnJGJIvv3r0Lm80Go9GIcrmMcrmMtbU1BINBuY9xbzq9Xi8jl71eL9rtNjKZjAxf5MhhejFGoxF2u12mf5L/6PV6omtiap7fX61WxYPls+p0OhJuM+Q+buCIHk6/5TTcwWAg01ErlcrYD5CHgWaAsO++X7lyBVarFf/jf/wPuN1uRCIRXLlyBcFgEN1uV8Y1p1Ip7O7uolaryQDDUql0KNfFcAPY3yjf/va30e/38d//+3+X6z537hxcLhc2NjYwOzuLM2fOyDTQra0tbGxs4N1335XwifqYRqOBQCCAVCqFQqEw1vngNChTU1M4c+YMYrEYNjc3ce/ePQnHyuWyDBnkbPVAIACj0YhUKiVhcq/Xg16vl1KVTqeD4XCIdruN7e1t9Ho9JJNJuFwuKIqCYrGIXq+HeDyOer2ON954Y2zP4eNgs9lgMBhQq9VgNBrhdDrFQM/Pz8NqtaLVao31HT4sNAP0j6jX6zI/nNqRzc1NZDIZeDwe6PV6pNNpdDod2O12/MIv/AIqlQpSqRTu3r2LTCbzSK+HnAw9E51OJ8T3aDSC3+8XcaFer0cul0O/30elUkE6nUav18Mrr7yCQqGAd999FyaTCTabTWaN1+t1GXf8SaaEHgaYFq/X68hkMigUCshkMqjVamKcotEoOp0OqtUqAMj0UI5hZkKAz4tpeqvVCpPJBAAyWbXT6cDn82Fubg7b29tYWVmBwWCA0+k8Vh4Qs1unT59GMBhEvV4X3orvf3p6GsViEd/97nfHPuf9YfCZN0D0MprNpkyhtFgssFgsyGQyGI1GmJubg9FoRC6XEzLz6aefRqvVwo0bNw5NfAjgAEFMF1un08HpdEo2q9friWaJHIbdbselS5ewsbGBt956CxaLBTabDT6fD91uFxsbG2KExq0ApgFqNpvI5/Oo1WqoVqtCFhsMBvj9fuHf9Ho9Op2OeDeUERgMBtmgiqJIuELS2mw2A4BopMLhMO7cuYPl5WXMzMzAbrcfKwOk1+thMpkwMzOD2dlZ1Ot1tFotlMtlRCIR+P1+TE5OCu9FScKDQB2uj/Pw+cwboEQigVAohM3NTWSzWczOzsJut8Nut6NUKqFer2NtbQ0ejwe9Xk/KLa5evQqTyQSv1wu73f7Ir0u9KFjnpP43F6Lf74fBYIDBYEA0GsX8/LxwVgsLC1KqcOrUKZw4cQKKoiCfz+O9996TTNC4tSPM5hQKBfR6PeRyOUxPT+M//If/gFwud0BecOrUKTnpOYZ4bm5O9E6scqdRLhQK8Pl8iEQiuHr1KhRFkWLd1157DR988AF2d3cllGHx7rifCb3xRCKBa9euYX5+HtlsVpTgHo8Hk5OTuHz5Mvr9PgwGA773ve/hr/7qr37qz7VYLDCbzfD7/eJhl0qlsfFHn1kDxDYWiqIgEAgIscz58PQ8zGYzLBaLnKScLd5qtQ4Qm4eN+zcEyVav1wuDwYBsNguXywWbzSaue7vdPjAHndyIuiHXuDcar8toNKLb7UqfI5PJhKmpKTEkvG6r1SrXzYJb8kQMwxwOB0qlEnq9HgaDgYhMmaJnKMNsYa1WQ6PROJL3+CDg9brdbiSTSSiKAovFIoaR748tR3Q6HaamprCxsYFEIoFisShrmGQ9dU4MTYF92oEh7bjWwWfSAOn1ejidTsTjcemPYzQa0el0sL6+jlqthnq9jsuXL2NychLxeBx2ux1utxv5fB7VahXpdBrlchlvvPEGtra2jvwems0mDAYDXnjhBXg8HjgcDrRaLSkHGQ6HeO2110RouL6+jmq1in6/L5v8uPAFDMHo2bjdbvh8Pvj9fiwuLqJQKCCXy4lhYpmFx+OB3W5HpVKRnj/BYBAulwvb29uSlg6FQgiHw9jd3UW9Xsfbb78tYkqGbO+//75oosYJho1utxsnT57E5z//eej1euEaaVwKhQJsNhtyuRxsNhtsNhuuXLmCcDiMr3/961hZWUGz2YTP58O5c+fw3HPP4fz58zh37hzq9Tr+/M//HNevX8fi4uKRtle5H4+1AdLpdNDr9XA4HKLZ+agHSZ5Hr9fDYDDA6/XCZDKh3+8jk8mg3W6LtL/VasFms0lNGCulXS4XLBYLAEgPHnIt/PxRIhAIIBKJIJPJyEneaDSk6FSn0yGTyaDRaAhXROUsSVlg/BwAAAkF+NHr9VAoFHD79m1ks1nhbNhapNvtStaH/A9T7uwDpPZM7XY7IpEINjc30ev1xItgar/f72N9fR3dbnfs5Qxc0yaTCZVKBYuLi/D5fDAYDMhkMuj1ekIHZLNZ3Lt3DwaDAZubm+IVeTwekTPEYjE8/fTT0obk+vXryGazeO+997C7uyvGRzNADwG+KJ/PJ2ERcNCdpNGhRN9isWBiYgL9fh/5fB5bW1siYLNYLGi32/B6vZiZmZH2DuVyGR6PB4qiYDQaifiNHBA78B3VS9Tr9YjFYjh58iS2trbQbrexubmJWq2GSqUimR9yQ7xfAIhGo6jX6wdIyHEbIRoEclLNZhOpVApvvfWWGAi1h9NsNqHT6aRGj9xcpVIRfoQZr8FgAKfTiYmJCaTTaQnJaOhI7JJPOg4eENcruzfSeORyObmXWq2Gvb090Tndu3cPsVhMukRaLBY89dRTmJmZwYsvvohisYhCoYC/+Zu/wcrKCm7duvWJOj4cFh4bA8RYt9PpSAdAl8sFo9Eobik3ETeU1WpFIpEQ/qDVasFqteIXf/EXUa/Xcf36dbhcLjidTjz33HNwOBy4evUq+v2+1Hux+yD7DNPjqNfrGA6HmJqaQrlcxsrKClqt1qF31yNfoj79dTodJiYmxFtgyNhoNOB0OjEzMyM9fFgflUwmJdvEFhbjAj0gn88Hl8uFu3fvYjgcwul0yrtgRwKDwQC73S6Zy9FohKtXryKTyeD111+XqvlgMIjhcIidnR3MzMzA6XTi8uXLSCQSePfdd6V+DIC8s+FwKIrqozbIPAjofV+8eBH9fh+dTkfq4yiQZVcEs9mMdrsNk8mEa9euwWq1SpsVm82GWCwGp9OJ0WgkMgzW0xkMhmPRCfKxMUAkHOnRsEZIr9eLK62GWi3LF8kWDNyQmUwGwWAQHo8H0WgUNpsNfr9fdCisnGa/Yf6bwjZ2I1QUBVarFd1u90gMEEsMuFFYOtJqteBwOODz+SQstVqtcLlc8r3FYhEGgwGxWAzFYlE2XbfbPUBYHyX4O1kLBnxIxLKdKhMAVDGbzWYJ2UKhkHye98BSmmq1KodIOBwWroleFX8/PUJu8nFsTno/TI50u10pvHU4HLBareLtMZs1HA6lyJgJlMnJSQSDQTidThGycj+otULHQXLw2BggFgmORiMhiff29qDT6WSx8GGzBzIALC4uCoH827/927hw4QJarRbi8Ti+8IUvSOr1j//4j1Gv13H69Gnkcjmsra3h1KlTCAaDmJycRK/XQz6fRzwex9TUlHhVzWYTgUAAHo9HMmOHBd5XLBZDJBKRplvAvrYlGo0iGo0iEonAbrdjcnISlUoFt27dgqIo0jdaURRcuHABy8vLuHHjhnhGb7/99liMEPk5apVYn6XT6eBwOGA0GrG9vY1CoYAf/vCHcLlc8Pl8KJfLGA6H6Ha7sFqtmJmZQS6XkxYqo9EIjUYDtVoNxWIRZ8+ehcfjkY1K/oNdIumBtFotyYoeBniY3u9pseC4VCrhxo0biEajmJycxMLCAvx+v3gvnU4HwWAQiqJgZ2dH5AmLi4tYXl7G1atXEQ6HYTQaJczk7yX3dRzCTeAxMkDqF8UHx3geOGjNdTodEokEzGYzarWa1Bixu97Ozo5oT+iK8mQYDAaw2+2Ix+MIBoPSS7nVaqFWq0FRFNhsNtEDkQfy+/2HKkjkM7BYLIhGoxJ+8vRWd/5zOp2SamW2iCppp9MJh8MhXAoLMIvFoqTCx2GAeO02mw1Wq1VCLOBDnVC328Xu7q602gAgzdaazeaBDU2ymT/T6XQKp0fukB4FW3zQwz7MQwT4cC3fTxmoP/r9PiwWixwyFKKyy4HBYJDrpJfu9XoRj8fl/3jP9XodlUoF5XJZMn3UTY1bjvHYGKCPw/0vE9h/Ic8++ywCgQCq1SpOnDiBixcvYmJiQirCgX2vim49i/o6nQ5isRiuXLkiLRr8fr9wJz6fD06nE/l8Xmqr2J85lUpJfH5YoPfCRVur1WAymaTvcyAQQK1WQ6fTkRA0EAjgxo0bWF9fRyKREI9wdnYWHo8Hq6ur2NnZEU7hqE9Gk8kk1+TxeOB2u6HX65HNZmXyRzgclvQ6vVGbzQaTyYQ7d+6IXohGmmUcoVAIsVgMk5OTqFar2Nvbk3oqr9cLYH/t0HsGcOiiPPXzpdFTG0+GmIFAACdPnpQSlEwmI83YstmsEOdM3U9NTSEajUoWMZlMot1uY29vD0tLS9ja2pJe4WazWQ7tcXpDj70BIvgi6cYTDocDxWIR77zzDt544w3odDrMzs4K13Dr1i2sra3JS3E4HDJZotFoSAP0YrGIpaUlJJNJ+P1+tNttGAwGqR2bmZnB9evXDz2u7vV6wt3wRLdarQgEApLVqNVqEhq2Wi1ks1nk83mUSiWUy2VYrVa8++67cDgc4gXSmxonB9RsNtFoNGA2mzEYDFCr1eBwOADgQBU7OTiKEkulEoxGI3w+n3hDACQsG41GeP/995HNZlEsFg94iB6PR7iSbrcrFfg0CoeFjyqFUHc+YFhcqVRQq9XQ7XalPfDW1hbu3r2L7e1tzM7OCi/JWr+VlRWRZqyvr+O1115Dq9VCr9fD888/j06ng3v37klnh729vbH1QXosDNDP6ner1k64XC4JT+jaV6tVbGxsYH19HaPRCP/8n/9zUZO+9957uH79Os6dOwev1yutHPR6vehq0uk0MpkM7ty5g2azKZohm80mBDRTpYf9HPr9vrT/UDcI43wramEoVBwOhyiXy7KQ6/U6SqUSut0uEomEFLOqw4BxgHVMzOqQvzGZTGKQ+F6Yeua10lApioJKpSKfN5vNSCaTAIClpSWk02lUKhXpsMgwhWRuo9GQynrg8DRSH0cb0JNheMRMLD1alppkMhmsrq5ieXkZoVBI1jLJ63a7jWKxiGKxiJ2dHdy4cUP6X586dUrIeobfpVLp0KecfByOvQHiSyLZylOe2RuDwQC32w2n04lEIoFf/uVfxnPPPYdUKiWbLZPJ4O7du+j1ejAajbh16xYikQhOnDiBiYkJ2O12nDp1Sk7GwWCA27dvS9uNRqOBVqsFnU6HaDSKhYUFKUpl6FatViWDcpgwGo1ScmG1WhGJRA7wJp1OR7QznU5H0uvkgVwuF6xW6wG+iAtVvamPEnyOwMGUOJW8TqcT4XAYzWZTujkyY8ReOeRBaKDsdjvC4TCeeeYZrKys4M0335S1xPfc6XREXkEdFSUJh6mNUns8/CA39/zzz8Pn88Hj8Ui7VXI2NMzNZhOhUAhOpxP/9J/+U0SjURiNRmSzWSwvL+Ptt98WpTRFtIFAAF6vFx988AFqtRref/99VKtV+RiXDOPYGiB6NewPzH8z3d3pdKS9qM/ng6IoiEQimJ6eRjKZFJ1OtVqV9DrbWHIDM4RSFAVer1fGvHCCBJW0bOal1+vhdrslra0ufmQcfdghGL06m80Gu90uBkX93ACI0G44HMJiscDr9Up7Cp6yw+EQrVZr7GOK+W553Uy9M9VM1To3Ce8f2DdY7IxIb8lqtUJRFLjdbvGk2JSeWU8AEnazIyQ9D17TYUNdo0cDQ30b+S01+G+DwYBIJAIAMt9sb28P5XIZhUIBq6urqNVq2NjYQKFQkHdsMplQKBRQrVZRLBYPKOc1Evo+sBn5yZMnkUgksLOzA7vdjmeeeUbmQ33uc5/DxMQE3G63cAPAfgfATqcDnU6HXq+Hs2fP4nOf+xz8fj8cDoe0H+WEBYfDgXv37mEwGCCRSCCfz2NpaQkXLlxAOByG3+9Hr9dDpVJBNBqF0+lELpdDpVLB8vLygcyFmsx81GDK2Ol0QlEUacXB8MvpdMLtdoveo1gswmQyIZFIYGJiQnoK8UQtlUpoNBoi3gOOvjRD7QUwzcwezcFgEA6HA2azGWtra3LtLpcLwWBQWnfQW+CcL5PJhNnZWVitVrz55ptYXl7G+vo6kskkwuEwtre3JatJWUOpVBKBnvqaDhN8zjy4DAaDGBIejrFYTEYRkaxntsvtduMv//IvsbS0hFdffRWdTufANJB0Oi0kNA1asViUNDyft9aO4yMQjUbx8ssvY25uDtFoFNlsFgaDAZOTk8jn89JC02QyodFoSJElK9fb7bbI86n3YJsKj8eDRqOBnZ0dqaHZ29sTEprTNNmuk83dAaBYLKJWqyGdTouSmO0e3G73odWF6XT7I2s4Hlk9/YGivV6vd8CjYZEqewCxRowhrMVigdVqlbauh2k8Pw706Ox2u+hxeB2szWq1WjLBlBXtTKMzfLFarQfCcnqlXAfkeOr1Olwul2xoda8gckFHtSHVRo7apHw+L831KR6k5IDcZqfTQa1Ww3A4xNbWFra3tw80K2PXA36OjfipAaJnPO4aQOAYG6DJyUn81m/9FiYmJiS1zLqfdDqNYDAoqlbOwSoUCvD7/XC5XJI2Z/V6uVwWNWkymUSlUsG9e/dEuNbv94VLqlar0uaSdWDAvrufyWRQr9eRSqVEqZpMJrGwsCAu/mGAojy2i2VDLpKHJGrZYqFer4vyOxqNiuHkhIlWqwW73Q5FUYTkPGoDpFb+stiXHgg3EIADzdjYB5oeMolXhmlErVYTsrXZbGIwGKBSqRwgprnJqRfiejqKjan2suipN5tN5HI5mEymA2UoPBzC4bAYJb1ej2q1iuXlZWxubv6Efov0AIADU0LUHCqvQ/OAPgLtdhupVAqlUglms1kmPhgMBqRSKSwvL0sf4IWFBZRKJbz//vtCTFLrQ6PA7EG/38drr70mvXJY0GgwGFAsFtFqtWSKAsldRVFEEs9N63A4MBqNZMCdzWYTzuIwoO570+l0pFEXyVd6OWzJQaPCRcdmVvPz8wCA3d1d8TwWFxeljetRh1+cVErjyZKZdrstLVd57+Syut0ums2mGF2Oo2GNHLNkFosF8XhcasnoNbJOrNvtStMyeli8rsMc9KfOeBmNRvHq+DuB/aSBTqdDtVoVb02n08HlcmF6elq8O34f8GFJCTOkbMuiNjr8+3HwfoBjbIAYUjHjsbe3J/Htzs4OUqmULFYWVrKFKslnuulsIsbM0OrqqpyWXLAOh0MmSTgcjgPpXqb4ucCpUtXpdPB6vVLkd9gvlYuNqXYS7Fx0dOOZCWOcz81dKpVEdEcPqtFoIJfLIZfLHXn9Ezch+R+1AWy329Knm5uU74FZPnXDfv6bIRVngVHrpC48NZvNontqt9tifNia9rA5IIaaXFMk0On9MSvIAmp6fGqCnSO1aUzUglzuAYpK1S13j0sTOuLYGiC+BHaDu3PnzoFT3+v1iqx8aWlJxs3QAzp37hzMZjPq9Tqy2SwymQwmJiagKAquXLmCYrGIe/fuie5ncnIS4XAYFy9elNM3l8sBgJCb8Xgci4uLWFxcxMbGhrQ8oPdDj+0wwCZqBoMBlUpFFhdTx+rSCxpHhiCcBcaKcofDgTNnzmB1dVWEmCRhjxIUjgIQj4anfrPZlOJaGhKTyYRcLie9rOkRhkIhAJCwrd1uS5MuRVEAQAY4spyDYXa9XsfOzo54FOrSnEe9UWkgyRUy/Ov3+5L9oqfH2kJ2CfB4PJibm5OEy9raGu7evYv19XXk8/kD10peEDj+s+KPpQHiqcdsj81mE+NBLoACO2p7WAsD7J9goVAIOp0OuVwOXq8X09PTklUh7xEKhcTljUajUBQFk5OTKJfLyOVy0qCe38c2qO12W2qvut2uGD9K4w/rmQQCAYTDYQQCAVEGq098enwkdBkSGgwGKeBkhoybOJ/PS0+ccbSgoGfCU5qblClpcnokiHnq0xtl5TslFJwWy41MDRc5LvIn9CjUVfD0NB41L6Lu8UMOiroshktsRaLWktGIcoRQJBIRg8zCUobY9+O4Gx7iWBognniRSESEaKwPYtFnKpWSSZedTgeRSATXrl2TSZrsEJhKpTA3N4eLFy9KG1LK+8nrcKFyqkQmk8HOzg5efPFFxGIxIaM3Nzcl43b69Gmpnbp9+za+973vHaoXYTKZcOrUKczPz+P8+fOSwWFGJ5fLwWKxwOfzSRjmdrtFK8QWDTabDe12GysrK1hZWcH6+rqEIEfN/5C/YtW2+nM8UHjdm5ubqFarMi/e4/Egl8uh2+0iFoshGo3KmJ1cLick7NraGnK5HFqtlhDWTqdT1oG6V/RoNBID9ShDFerZePi99NJLQhvQ02Jbl0ajgUqlgvfeew+hUAjT09Po9Xqw2WxYWFiQEiHKMBh6P644lgao3+9jeXkZ//N//k/Mzs4ikUhgbm4OPp8PoVAILpcLsVgMPp9PivEcDgcCgQAajYa0YvD5fJiamhKhIgvxOOCNosZ+v48bN26gUCgglUrB7/cjFotBr9ejUCjg//7f/4t0Oo3t7W1ks1mUy2XcunULBoMBvV4PmUxGYvLD2sRq150EM/AhkRsKhaR+ijofpu69Xq+kuwOBACqVipRmsIXIUfI/9HxYw0bhIb0BdQjJzBDLBdSeEDVC9JA6nQ6KxaKo4Pkc6ElRUEo9TTAYFEPDsIv8zMOCng5JX/I2oVBIDCs9VnqvvB/WY7F5Gq+DnS3Z9TOTyWB7exs7Oztj6V7wKHEsDRA72X3jG9/AmTNnMD8/L/U7anUy26Iy9c0sGd1rDrRT60ro6ajrZkajEW7fvo2dnR1sbW3h4sWLmJubQ6fTQalUwl//9V9je3tbeugyBODvVKc1DwvsfW00GlGtVg8QkBQnUuHKUgIA8n/8HMNYFqzSAB9FNbSa1CWx7HK5DmiTqGxmGNRoNMQroHEisco2G9zUbK3K4YYcVUStE98Zjbbb7RbCnkb40xDP6oJorgtmTkOhkIR5AKSwlhwQdUs0UrFYDNVqFYVC4cDsuk6ng729Pezt7SGTyWgG6LBAFv/OnTtYW1vDD37wA0kbAx+6tWazGR6PB91uV6aDMrMF4MCi45gWpivV8TY9mG63ix/84Af4kz/5E6kL40lDnoJZBv79YfGgRKE6bUvvi7F/qVSC1WpFOBxGqVSSynd6gOzxs7q6iq2tLXzlK18BAKysrMjzOmrhHZ89K9nJVfFwoOZLPY6HXhHwIU8UjUbh9Xqxvr5+wPhw3DLvvdfrwWw2I5fLoVAoyIFFfqjdbqPVaglX9LAZTYvFAofDgXA4LImJ+5MDDodDZnvF43HRo/G+KP9gz6l6vY6TJ0/CbrfjzTffRC6Xw+3bt7G4uIjt7e1jM0roYXFsDRDwoZiq2WyiWCwCwIHGUYzdeZLl8/kD2YuHjeMfpKfPUZ86w+FQMjk8LflBla9aakBvgupoNqyn4aTm5iiND//8qHQ58KFhp1fGD3qcvHeqlkk0M8OXzWZF/asW+AGQcJltWtWeCDNufB4P26iLIW8oFDqQAAAgRpCyEPVhpg6B6RmpW3AA++ueXs/Ozo6EpI+z9wMccwP0UVB3Q+RLpMF43F7GJ7neTqeD7373u4jH4zhz5gzK5bLwPexjMxqNRFvC/jAMZ1nvRQN9v6d4FFCPRqIRIemrLqNQN1lj5b7dbke328XW1hY8Hg9CoZD0bVJ3AOBYbb/f/xPdASjmq1ar8Hg8iMfj0oOI01Q/Tfp6MBjA4/Hg2WeflZ/JzFcwGBTNESUUJJHZ3gUAgsEger0e1tbWhCtj6La8vIx0Oo1sNju29hmPGo+dAfooPG6G55OCJGmxWJR5YBRXqhcosyxOpxO9Xk8KFllwW61WZeT0cegJTF6LtV0cMsAyEUojbDab1HHRiPT7fdHvsAd0pVKRn82yDjXHReUxm85NT09L+w3+DA4zfBjveTAYoFQq4ebNmwfS40wSkODmPZHDisfjACAcHWsMaYCazaaovQFIEuI4KZofFk+EAfosYDAYoFAoyNgd1nxxBpbJZEIsFsPU1BTcbjeazSZmZmbg8/kQjUZx4sQJqSsalxz//lolANIQTF1cWq/XoSgKksmktEtZXFxEt9sVBXOn08HW1pZkNNm2FPgwxU+tEH8fvRFFURAOh3Hy5Ekpv6CXmE6nH/q5sFbwtddeO1AsbDQaEQ6H4fV6ZYABJRKhUAjnzp070CKlUCigVCoJWZ5KpVCr1SScIyE/zommjwqaAXpMQNXz3bt38X/+z/8R/oMTMFhusbGxgRs3biCXy+HevXtwuVwIhUJCuBYKBeFK1BL9o7oH/smm/h988AGsVquMEWIZjdFoxAcffCCpdmqVWq2WGBIaUQpE1RXhLEZVb1AKETnAUKfTYWtrC5lMRoqMP+2MNIa36gGInMBLA7ezsyNdGQKBAG7duoVoNIpAICB1XsFgULozUNlcLpclTH1Y46PO8h0H46UZoMcI3W4XxWIRt2/flt42DCu48KvVKvL5PHZ3d1EqlWC325HL5aSzYy6XQ7ValZYMRwVmHhna8PcXi0Xpa8RwiX1/2GaUBDRwUFVMXQ9/nvpnfFR2iN9L74jZKAocOVjy01SKM1y+//cyScBQkfPfWZ/H5xGJRKTSnypo8mAk1ikr+aTXpq5vU7+PcUI3GvcVaNCg4TOLo+9ApUGDBg3/CM0AadCgYWzQDJAGDRrGBs0AadCgYWzQDJAGDRrGBs0AadCgYWzQDJAGDRrGBs0AadCgYWzQDJAGDRrGBs0AadCgYWzQDJAGDRrGBs0AadCgYWzQDJAGDRrGBs0AadCgYWzQDJAGDRrGBs0AadCgYWzQDJAGDRrGBs0AadCgYWzQDJAGDRrGBs0AadCgYWzQDJAGDRrGBs0AadCgYWzQDJAGDRrGBs0AadCgYWzQDJAGDRrGBs0AadCgYWzQDJAGDRrGBs0AadCgYWzQDJAGDRrGBs0AadCgYWzQDJAGDRrGBs0AadCgYWzQDJAGDRrGBs0AadCgYWzQDJAGDRrGhv8fuz+d5HDSiV4AAAAASUVORK5CYII=\n"
          },
          "metadata": {},
          "execution_count": 35
        }
      ]
    }
  ]
}