ImageNoiseReduction.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "ImageNoiseReduction.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wxgAJtDfUInK",
        "colab_type": "text"
      },
      "source": [
        "<center>\n",
        "<h1>Image Noise Reduction with Auto-encoders</h1>\n",
        "</center>\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hp-DzmFX9bEu",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "\n",
        "from matplotlib import pyplot as plt\n",
        "\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras.datasets import mnist\n",
        "from tensorflow.keras.models import Sequential, Model\n",
        "from tensorflow.keras.layers import Input, Dense, Conv2D, MaxPooling2D, UpSampling2D, Flatten \n",
        "from tensorflow.keras.callbacks import EarlyStopping\n",
        "from tensorflow.keras.utils import to_categorical\n",
        "\n",
        "%matplotlib inline"
      ],
      "execution_count": 73,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "J3lF3X9I9k-Z",
        "colab_type": "text"
      },
      "source": [
        "# Loading and Visulizing Dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8KnSu8aj9oE3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# loading dataset \n",
        "(X_train, Y_train), (X_test, Y_test) = mnist.load_data()"
      ],
      "execution_count": 44,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rjhboMUv_xyf",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 85
        },
        "outputId": "33556405-ae9a-48b5-a43c-efede45efc88"
      },
      "source": [
        "print(f\"Shape of X_train: {X_train.shape}\")\n",
        "print(f\"Shape of Y_train: {Y_train.shape}\")\n",
        "print(f\"Shape of X_test: {X_test.shape}\")\n",
        "print(f\"Shape of Y_test: {Y_test.shape}\")"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Shape of X_train: (60000, 28, 28)\n",
            "Shape of Y_train: (60000,)\n",
            "Shape of X_test: (10000, 28, 28)\n",
            "Shape of Y_test: (10000,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SXePWHE1__Ex",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# defining visualization method\n",
        "import random\n",
        "\n",
        "def plot(X, predictions=None):\n",
        "    nrows = 4\n",
        "    ncols = 4\n",
        "\n",
        "    # Set up matplotlib fig, and size it to fit 4x4 pics\n",
        "    fig = plt.gcf() \n",
        "    fig.set_size_inches(10, 10)\n",
        "\n",
        "    for i in range(nrows*ncols):\n",
        "        # setting up subplot\n",
        "        sp = plt.subplot(nrows, ncols, i + 1)\n",
        "\n",
        "        # displaying random image and label name \n",
        "        rint = random.randint(0, X.shape[0]) # generating a random number \n",
        "\n",
        "        # displaying image \n",
        "        img = X[rint]\n",
        "        plt.imshow(img, cmap='binary')\n",
        "\n",
        "        # removing axis ticks \n",
        "        plt.xticks([])\n",
        "        plt.yticks([])\n",
        "\n",
        "        # displaying label if predictions\n",
        "        if predictions is not None: \n",
        "            plt.title(f\"{np.argmax(predictions[rint])}\")\n",
        "\n",
        "    plt.show()"
      ],
      "execution_count": 108,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qf3h763Q_2QR",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 578
        },
        "outputId": "e49c227b-acfe-47d9-c7cf-b40bd0498b32"
      },
      "source": [
        "# visulizing training dataset \n",
        "plot(X_train)"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 16 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UBUkmcvwCBF1",
        "colab_type": "text"
      },
      "source": [
        "# Data Preprocessing"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9UakJKUpGTWI",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "5d9fe1f9-85e8-4841-cd98-5345f71abdff"
      },
      "source": [
        "# adding channel dimension \n",
        "X_train = np.expand_dims(X_train, axis=-1)\n",
        "X_test = np.expand_dims(X_test, axis=-1)\n",
        "\n",
        "print(f\"Shape of X_train: {X_train.shape}\")\n",
        "print(f\"Shape of X_test: {X_test.shape}\")"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Shape of X_train: (60000, 28, 28, 1)\n",
            "Shape of X_test: (10000, 28, 28, 1)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hZK93lkwCDT1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# normalizing images \n",
        "X_train = X_train.astype('float32') / 255. \n",
        "X_test = X_test.astype('float32') / 255."
      ],
      "execution_count": 47,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YAK-SsTu9pmK",
        "colab_type": "text"
      },
      "source": [
        "# Building a Classifier"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mczmrUPF9xn_",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 425
        },
        "outputId": "fb2d27c0-ff68-41cc-b9c3-f1faf283369e"
      },
      "source": [
        "# defining the model\n",
        "classifier = Sequential([\n",
        "    Conv2D(filters=6, kernel_size=(3, 3), activation='relu', input_shape=(28, 28, 1)),\n",
        "    MaxPooling2D(),\n",
        "    Conv2D(filters=16, kernel_size=(3, 3), activation='relu'), \n",
        "    MaxPooling2D(),\n",
        "    Flatten(),\n",
        "    Dense(units=120, activation='relu'),\n",
        "    Dense(units=84, activation='relu'),\n",
        "    Dense(units=10, activation = 'softmax')\n",
        "])\n",
        "\n",
        "# defining loss function and optimizer\n",
        "classifier.compile(\n",
        "        optimizer='sgd',\n",
        "        loss='sparse_categorical_crossentropy',\n",
        "        metrics = [\"accuracy\"]\n",
        "    )\n",
        "\n",
        "# displaying model summary\n",
        "classifier.summary()"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_4\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_8 (Conv2D)            (None, 26, 26, 6)         60        \n",
            "_________________________________________________________________\n",
            "max_pooling2d_8 (MaxPooling2 (None, 13, 13, 6)         0         \n",
            "_________________________________________________________________\n",
            "conv2d_9 (Conv2D)            (None, 11, 11, 16)        880       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_9 (MaxPooling2 (None, 5, 5, 16)          0         \n",
            "_________________________________________________________________\n",
            "flatten_4 (Flatten)          (None, 400)               0         \n",
            "_________________________________________________________________\n",
            "dense_12 (Dense)             (None, 120)               48120     \n",
            "_________________________________________________________________\n",
            "dense_13 (Dense)             (None, 84)                10164     \n",
            "_________________________________________________________________\n",
            "dense_14 (Dense)             (None, 10)                850       \n",
            "=================================================================\n",
            "Total params: 60,074\n",
            "Trainable params: 60,074\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5CUh0hFQERvX",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 615
        },
        "outputId": "cf5ec529-3bf5-4b2d-df7e-4d270358387a"
      },
      "source": [
        "# training the model \n",
        "hist_classifier = classifier.fit(\n",
        "        x=X_train, y=Y_train,\n",
        "        validation_split=0.2,\n",
        "        epochs=150,\n",
        "        batch_size=32, \n",
        "        callbacks=[EarlyStopping(monitor='val_loss', patience=5)]\n",
        "    )"
      ],
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.5986 - accuracy: 0.8183 - val_loss: 0.1931 - val_accuracy: 0.9389\n",
            "Epoch 2/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.1733 - accuracy: 0.9462 - val_loss: 0.1201 - val_accuracy: 0.9647\n",
            "Epoch 3/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.1186 - accuracy: 0.9634 - val_loss: 0.1060 - val_accuracy: 0.9678\n",
            "Epoch 4/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.0945 - accuracy: 0.9706 - val_loss: 0.0846 - val_accuracy: 0.9737\n",
            "Epoch 5/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.0797 - accuracy: 0.9754 - val_loss: 0.0772 - val_accuracy: 0.9774\n",
            "Epoch 6/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.0696 - accuracy: 0.9779 - val_loss: 0.0772 - val_accuracy: 0.9760\n",
            "Epoch 7/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.0619 - accuracy: 0.9800 - val_loss: 0.0696 - val_accuracy: 0.9783\n",
            "Epoch 8/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.0549 - accuracy: 0.9825 - val_loss: 0.0672 - val_accuracy: 0.9801\n",
            "Epoch 9/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.0509 - accuracy: 0.9844 - val_loss: 0.0587 - val_accuracy: 0.9833\n",
            "Epoch 10/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.0470 - accuracy: 0.9852 - val_loss: 0.0616 - val_accuracy: 0.9798\n",
            "Epoch 11/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.0435 - accuracy: 0.9863 - val_loss: 0.0577 - val_accuracy: 0.9828\n",
            "Epoch 12/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.0398 - accuracy: 0.9873 - val_loss: 0.0512 - val_accuracy: 0.9839\n",
            "Epoch 13/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.0373 - accuracy: 0.9881 - val_loss: 0.0564 - val_accuracy: 0.9829\n",
            "Epoch 14/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.0338 - accuracy: 0.9895 - val_loss: 0.0614 - val_accuracy: 0.9822\n",
            "Epoch 15/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.0324 - accuracy: 0.9901 - val_loss: 0.0554 - val_accuracy: 0.9837\n",
            "Epoch 16/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.0300 - accuracy: 0.9909 - val_loss: 0.0549 - val_accuracy: 0.9847\n",
            "Epoch 17/150\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.0287 - accuracy: 0.9911 - val_loss: 0.0534 - val_accuracy: 0.9836\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IdNbCdyuGQlc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def plot_acc(history):\n",
        "    # Plot training & validation accuracy values\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', 'Valid'], loc='upper left')\n",
        "    plt.show()\n",
        "\n",
        "def plot_loss(history):\n",
        "    # Plot training & validation loss values\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', 'Valid'], loc='upper left')\n",
        "    plt.show()"
      ],
      "execution_count": 56,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l5gBK4JQHf_2",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        },
        "outputId": "edf0a193-e4fc-4d8a-c41f-d3ebc0085a65"
      },
      "source": [
        "# plotting training curves \n",
        "plot_acc(hist_classifier)\n",
        "plot_loss(hist_classifier)"
      ],
      "execution_count": 79,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXxc5X3v8c9vFkkzsrXaBmyNLQMmYBLARmEPhWyXkBQnbRYckkBIwyvc0mxtU0hvE5re3OamaZqmpQslCWQpvpSG1EkgZGkWSALYEAdiMOCAjeV90Wbty+/+cc5II3lkj2yNRp7zfb9e85pzzpwz85Mw89XzPOc8x9wdERGJrlipCxARkdJSEIiIRJyCQEQk4hQEIiIRpyAQEYk4BYGISMQpCEQKYGbNZuZmlihg3+vM7OFjfR+RmaIgkLJjZlvMbMDM5k3Y/qvwS7i5NJWJzE4KAilXLwKrsytm9gogXbpyRGYvBYGUq68B78lZvxb4au4OZlZrZl81s71mttXM/peZxcLX4mb2OTPbZ2YvAG/Mc+yXzGynmW03s/9tZvGpFmlmC81srZkdMLPNZvb+nNfOM7P1ZtZpZrvN7PPh9ioz+7qZ7TezdjNbZ2YnTPWzRbIUBFKuHgFqzOyM8Av6auDrE/b5B6AWOBn4HYLgeG/42vuBNwErgBbgrROOvRMYAk4N93k98AdHUecaoBVYGH7G/zGzV4ev/T3w9+5eA5wC3BNuvzasOwM0Ah8Aeo/is0UABYGUt2yr4HXAM8D27As54XCLu3e5+xbgb4F3h7u8HfiCu29z9wPAX+ccewJwJfBhd+929z3A34XvVzAzywAXA3/m7n3uvgG4g7GWzCBwqpnNc/eD7v5IzvZG4FR3H3b3x929cyqfLZJLQSDl7GvAO4HrmNAtBMwDksDWnG1bgUXh8kJg24TXspaEx+4Mu2bagX8FFkyxvoXAAXfvmqSG9wGnAZvC7p835fxcDwJrzGyHmX3WzJJT/GyRUQoCKVvuvpVg0PhK4JsTXt5H8Jf1kpxtixlrNewk6HrJfS1rG9APzHP3uvBR4+5nTrHEHUCDmc3NV4O7P+/uqwkC5v8C95pZtbsPuvtfuvty4CKCLqz3IHKUFARS7t4HvNrdu3M3uvswQZ/7p81srpktAT7K2DjCPcAHzazJzOqBm3OO3Ql8H/hbM6sxs5iZnWJmvzOVwtx9G/AL4K/DAeCzwnq/DmBm7zKz+e4+ArSHh42Y2eVm9oqwe6uTINBGpvLZIrkUBFLW3P237r5+kpf/COgGXgAeBv4d+HL42r8RdL/8GniCQ1sU7wEqgKeBNuBe4KSjKHE10EzQOrgP+KS7/zB87Qpgo5kdJBg4vtrde4ETw8/rJBj7+ClBd5HIUTHdmEZEJNrUIhARiTgFgYhIxCkIREQiTkEgIhJxx91UuPPmzfPm5uZSlyEiclx5/PHH97n7/HyvHXdB0NzczPr1k50NKCIi+ZjZ1sleU9eQiEjEKQhERCJOQSAiEnHH3RhBPoODg7S2ttLX11fqUoquqqqKpqYmkklNNiki06MsgqC1tZW5c+fS3NyMmZW6nKJxd/bv309raytLly4tdTkiUiaK2jVkZleY2bPhLfhunmSft5vZ02a20cz+/Wg+p6+vj8bGxrIOAQAzo7GxMRItHxGZOUVrEYRT5N5GcHeoVmCdma1196dz9lkG3AJc7O5tZjbVG3vkft6xlnxciMrPKSIzp5gtgvOAze7+grsPENybddWEfd4P3ObubQDhLf+Kort/iJ0dvWi2VRGR8YoZBIsYf6u/VsZuwZd1GnCamf3czB4xsyvyvZGZ3WBm681s/d69e4+qmN6BYfZ29TM8Mv1BsH//fs455xzOOeccTjzxRBYtWjS6PjAwcNhj169fzwc/+MFpr0lEpFClHixOAMuAy4Am4Gdm9gp3b8/dyd1vB24HaGlpOapv8opEkHkDwyMk4tObf42NjWzYsAGAW2+9lTlz5vAnf/Ino68PDQ2RSOT/Vbe0tNDS0jKt9YiITEUxWwTbGX/P1ybG7geb1QqsDe/B+iLwHEEwTLtkNgiGZuaOftdddx0f+MAHOP/88/nYxz7GY489xoUXXsiKFSu46KKLePbZZwH4yU9+wpveFNyT/NZbb+X666/nsssu4+STT+aLX/zijNQqItFWzBbBOmCZmS0lCICrgXdO2OdbBLfq+4qZzSPoKnrhWD70L7+9kad3dOZ9rbt/iIpEjOQUWwTLF9bwyd+d6n3Jg9Naf/GLXxCPx+ns7OShhx4ikUjwwx/+kI9//OP853/+5yHHbNq0iR//+Md0dXXxspe9jBtvvFHXDIhIURUtCNx9yMxuIrjvaxz4srtvNLNPAevdfW342uvN7GlgGPhTd99frJrMjCIMEUzqbW97G/F4HICOjg6uvfZann/+ecyMwcHBvMe88Y1vpLKyksrKShYsWMDu3btpamqauaJFJHKKOkbg7vcD90/Y9omcZQc+Gj6mxeH+ct+8p4uYGSfPnzNdH3dY1dXVo8t/8Rd/weWXX859993Hli1buOyyy/IeU1lZObocj8cZGhoqdpkiEnGRmmsoGY8xODwzYwQTdXR0sGhRcNLUnXfeWZIaRETyiVQQVCRiDAx7Sa4l+NjHPsYtt9zCihUr9Fe+iMwqdrxdYNXS0uITb0zzzDPPcMYZZxzx2P0H+9ne3svpJ9aMnk56PCr05xURyTKzx90977nqx++34VHIvZZAREQC0QqC+MxeSyAicjyIVBBkLyor1YCxiMhsFKkgiJmRjMfUIhARyRGpIICge0hBICIyJnpBkIhpsFhEJEckg2BweISRaTxt9vLLL+fBBx8ct+0LX/gCN954Y979L7vsMrKnwF555ZW0t7cfss+tt97K5z73uWmrUURkMpELguyEc4PT2D20evVq1qxZM27bmjVrWL169RGPvf/++6mrq5u2WkREpipyQVCMawne+ta38t3vfnf0JjRbtmxhx44d3H333bS0tHDmmWfyyU9+Mu+xzc3N7Nu3D4BPf/rTnHbaaVxyySWj01SLiBRbqW9MM/0euBl2PTXpy2l3Th4YpjIRg0Knoz7xFfCGz0z6ckNDA+eddx4PPPAAq1atYs2aNbz97W/n4x//OA0NDQwPD/Oa17yGJ598krPOOivvezz++OOsWbOGDRs2MDQ0xMqVKzn33HMLq09E5BhErkVgBhiMML1Ta+R2D2W7he655x5WrlzJihUr2LhxI08//fSkxz/00EO85S1vIZ1OU1NTw1VXXTWt9YmITKb8WgSH+csdwIDtuzqpSsZZ0lh92H2nYtWqVXzkIx/hiSeeoKenh4aGBj73uc+xbt066uvrue666+jr65u2zxMRmS6RaxFAcaajnjNnDpdffjnXX389q1evprOzk+rqampra9m9ezcPPPDAYY+/9NJL+da3vkVvby9dXV18+9vfntb6REQmU34tggJUJGJ09k7/VNCrV6/mLW95C2vWrOH0009nxYoVnH766WQyGS6++OLDHrty5Ure8Y53cPbZZ7NgwQJe+cpXTnt9IiL5RGoa6qw9XX3s6ujjzIW1xGM23SUWnaahFpGp0jTUE4zOQqorjEVEIhoEiem/qExE5HhVNkEwlS6u4/m+BMdbV56IzH5lEQRVVVXs37+/4C/JeMyImR13XUPuzv79+6mqqip1KSJSRsrirKGmpiZaW1vZu3dvwcfs6+yjPWZ0zKksYmXTr6qqiqamplKXISJlpCyCIJlMsnTp0ikd87d3raO1rZfvffjSIlUlInJ8KIuuoaPRVJ9m24Ee9bmLSOQVNQjM7Aoze9bMNpvZzXlev87M9prZhvDxB8WsJ1emIU33wDBtPYMz9ZEiIrNS0bqGzCwO3Aa8DmgF1pnZWnefOPPa/3P3m4pVx2Qy9SkAth3ooaG6YqY/XkRk1ihmi+A8YLO7v+DuA8AaYFURP29KMg1pALa19ZS4EhGR0ipmECwCtuWst4bbJvp9M3vSzO41s0y+NzKzG8xsvZmtn8qZQYczGgQHeqfl/UREjlelHiz+NtDs7mcBPwDuyreTu9/u7i3u3jJ//vxp+eA5lQnq00m1CEQk8ooZBNuB3L/wm8Jto9x9v7v3h6t3ADN6S65MQ3DmkIhIlBUzCNYBy8xsqZlVAFcDa3N3MLOTclavAp4pYj2HyNQrCEREihYE7j4E3AQ8SPAFf4+7bzSzT5lZ9j6MHzSzjWb2a+CDwHXFqiefpoYU29t7GR7RtQQiEl1FvbLY3e8H7p+w7RM5y7cAtxSzhsPJ1KcZHHZ2d/axsC5VqjJEREqq1IPFJTV25pC6h0QkuqIdBNmLytp0CqmIRFekg2BRfQoztQhEJNoiHQSViTgnzK3StQQiEmmRDgKATEOKVl1dLCIRpiCoT6tFICKRFvkgaGpIs6uzj/6h4VKXIiJSEpEPgkx9CnfY0d5X6lJEREpCQaBrCUQk4hQEui+BiERc5IPgxJoqknHTfQlEJLIiHwTxmLGwLqUWgYhEVuSDAIJTSFs1RiAiEaUgILioTPMNiUhUKQiApvo0B7oH6O4fKnUpIiIzTkGAzhwSkWhTEJAzHbXOHBKRCFIQoIvKRCTaFARAY3UFqWRcXUMiEkkKAsDMgjOH1DUkIhGkIAhl6tO0qkUgIhGkIAhlGtJsO9CDu5e6FBGRGaUgCDXVp+geGKatZ7DUpYiIzCgFQUhnDolIVCkIQpl6XVQmItGkIAhlGnRRmYhEU1GDwMyuMLNnzWyzmd18mP1+38zczFqKWc/hzK1KUpdOqkUgIpFTtCAwszhwG/AGYDmw2syW59lvLvAh4NFi1VKoTH1aYwQiEjnFbBGcB2x29xfcfQBYA6zKs99fAf8XKPnd4zMNKVo1HbWIREwxg2ARsC1nvTXcNsrMVgIZd//u4d7IzG4ws/Vmtn7v3r3TX2koU59me1svIyO6lkBEoqNkg8VmFgM+D/zxkfZ199vdvcXdW+bPn1+0mpoa0gwMj7C7q+SNExGRGVPMINgOZHLWm8JtWXOBlwM/MbMtwAXA2lIOGGs6ahGJomIGwTpgmZktNbMK4GpgbfZFd+9w93nu3uzuzcAjwFXuvr6INR2WLioTkSgqWhC4+xBwE/Ag8Axwj7tvNLNPmdlVxfrcY7GoLmwR6BRSEYmQRDHf3N3vB+6fsO0Tk+x7WTFrKURVMs4JNZXqGhKRSNGVxRNk6tNqEYhIpCgIJsg0pGnVGIGIRIiCYIJMfYqdnX0MDI2UuhQRkRmhIJigqSGNO+xo1ziBiESDgmACTUctIlGjIJhA01GLSNQoCCY4qTZFImZqEYhIZCgIJojHjIV1KV1dLCKRoSDII9OQYpumoxaRiFAQ5JGp17UEIhIdCoI8Mg1p9ncP0N0/VOpSRESKTkGQR1M4HbXuViYiUaAgyEPTUYtIlCgI8tBFZSISJQqCPObNqSCVjOuiMhGJBAVBHmZGU31KLQIRiYSCgsDMqsObzWNmp5nZVWaWLG5ppZVpSGuMQEQiodAWwc+AKjNbBHwfeDdwZ7GKmg0y9Sla23px91KXIiJSVIUGgbl7D/B7wD+5+9uAM4tXVullGtIc7B+ivWew1KWIiBRVwUFgZhcC1wDfDbfFi1PS7NCkM4dEJCIKDYIPA7cA97n7RjM7Gfhx8coqPU1HLSJRkShkJ3f/KfBTgHDQeJ+7f7CYhZXa6EVlahGISJkr9KyhfzezGjOrBn4DPG1mf1rc0kqrpipJbSqpM4dEpOwV2jW03N07gTcDDwBLCc4cKmuajlpEoqDQIEiG1w28GVjr7oNA2Z9XqemoRSQKCg2CfwW2ANXAz8xsCdBZrKJmi0xDmta2XkZGyj7zRCTCCgoCd/+iuy9y9ys9sBW4/EjHmdkVZvasmW02s5vzvP4BM3vKzDaY2cNmtvwofoaiydSnGBgeYU9Xf6lLEREpmkIHi2vN7PNmtj58/C1B6+Bwx8SB24A3AMuB1Xm+6P/d3V/h7ucAnwU+P/UfoXiadOaQiERAoV1DXwa6gLeHj07gK0c45jxgs7u/4O4DwBpgVe4O4QB0VjWzbNxhdDpqjROISBkr6DoC4BR3//2c9b80sw1HOGYRsC1nvRU4f+JOZvaHwEeBCuDV+d7IzG4AbgBYvHhxgSUfu+ydynRRmYiUs0JbBL1mdkl2xcwuBqbl29Hdb3P3U4A/A/7XJPvc7u4t7t4yf/786fjYglQl4yyYW6muIREpa4W2CD4AfNXMasP1NuDaIxyzHcjkrDeF2yazBvjnAuuZMZqOWkTKXaFnDf3a3c8GzgLOcvcVTNKNk2MdsMzMlppZBXA1sDZ3BzNblrP6RuD5giufIdnpqEVEytWU7lDm7p05A7wfPcK+Q8BNwIPAM8A94YR1nzKzq8LdbjKzjeF4w0c5citjxmUa0uzs6GVweKTUpYiIFEWhXUP52JF2cPf7gfsnbPtEzvKHjuHzZ0SmPs2Iw472XpY0HvaMWRGR49Kx3LN4Vp3qWSxNmo5aRMrcYVsEZtZF/i98A1JFqWiWyegGNSJS5g4bBO4+d6YKma1Oqq0iHjOdOSQiZetYuoYiIRGPsbCuStNRi0jZUhAUIFOvawlEpHwpCAqQqU/TqjECESlTCoICZBpS7Ds4QM/AUKlLERGZdgqCAmRvZK8rjEWkHCkICtCk6ahFpIwpCAqQGb2oTEEgIuVHQVCA+XMqqUrGdAqpiJQlBUEBzIwmnUIqImVKQVCgxQ1ptQhEpCwpCAqUqU/ReqAH90jMtSciEaIgKFCmIU1X/xAdvYOlLkVEZFopCAo0dgqpuodEpLwoCAo0egqpppoQkTKjIChQ9upinTkkIuVGQVCgmqoktamkWgQiUnYUBFOQaUhpjEBEyo6CYAoy9Wm1CESk7CgIpiDTkKa1rZeREV1LICLlQ0EwBZn6FANDI+w92F/qUkREpo2CYAqadOaQiJQhBcEUZLIXlWmcQETKSFGDwMyuMLNnzWyzmd2c5/WPmtnTZvakmf3IzJYUs55j1VSfvS+BzhwSkfJRtCAwszhwG/AGYDmw2syWT9jtV0CLu58F3At8tlj1TIeqZJwFcyvVNSQiZaWYLYLzgM3u/oK7DwBrgFW5O7j7j909+636CNBUxHqmRaZBp5CKSHkpZhAsArblrLeG2ybzPuCBItYzLTL1uqhMRMrLrBgsNrN3AS3A30zy+g1mtt7M1u/du/foP2ho4OiPDWUa0uzs6GVweOSY30tEZDYoZhBsBzI5603htnHM7LXAnwNXuXveE/Td/XZ3b3H3lvnz5x9dNU98Df75IuhtP7rjQ5n6NCMOO9v7jul9RERmi2IGwTpgmZktNbMK4Gpgbe4OZrYC+FeCENhTxFpg/sug7UVYexMcw13GmjQdtYiUmaIFgbsPATcBDwLPAPe4+0Yz+5SZXRXu9jfAHOA/zGyDma2d5O2OXeY8eO2t8My34bHbj/5t6nVRmYiUl0Qx39zd7wfun7DtEznLry3m5x/iwptgy8/hwT+HphZYdO6U3+Kk2iriMVOLQETKxqwYLJ4xZvDmf4K5J8J/XHdU4wWJeIyFdVU6c0hEyka0ggAg3QBv/Qp07oD/+sOjGi/QdNQiUk6iFwQAmVfC6z4Fm74Dj/7L1A+vT6tFICJlI5pBAHDB/4SXXQnf/wtofXxKhy5uTLPvYD9P7+gsUnEiIjMnukEwOl5wUjhe0FbwoW9esYgTa6q45o5HFAYictyLbhAApOrhbXdC1074VuHjBYvqUqy54QKqknGFgYgc96IdBABN5wbjBc9+Fx75p4IPa55XPRoG77zjETbu6ChikSIixaMgALjgRjj9TfCDT0Dr+oIPW9IYhEE6GeeaOx5VGIjIcUlBAMF4wap/hJqF8B/vhZ4DBR+6pLGau3PC4DfbFQYicnxREGSl6uGtdwbjBVO8viBoGVyoMBCR45KCIFfTufD6v4Jn74df3jalQxc3pllzw4XMqUwoDETkuKIgmOj8DwTjBT/8JGxbN6VDgzC4QGEgIscVBcFEZrDqtnC84LopjRdAcOOabBi8898e4alWhYGIzG4KgnxSdcH1BQd3w7dunPJ8RNkwmFuV5Jo7HuHJ1mO7GY6ISDEpCCaz6Fz4H5+G574Hv/iHKR+eDYOaVJJ33fGowkBEZi0FweGcdwOc8bvww1vhpUenfHimIc3d71cYiMjspiA4HDO46h+htgnuvX7K4wUwvmVwzR2P8uttCgMRmV0UBEeSqoO33wXde+C+D8DIyJTfoqk+CIO6dJJ3felRNigMRGQWURAUYuEKeP2n4fkH4ZdTHy+AbBhcSF06ybvvUBiIyOyhICjUee+H5avgh38JLz1yVG8RzFp6IfXVFbz7jkf51UuFT30tIlIsCoJCmcFV/wB1mWC8oHv/Ub1Ndgrr+uoK3vOlx3hCYSAiJaYgmIqq2uD6gu698B/XwosPwWDflN9mYU4YXKswEJESMz+Km7eXUktLi69fX/hU0UXx+F3wnY+AD0OiCjLnw9JLYenvBOMJ8URBb7OjvZfV//YIuzv7eEdLhvdevJTmedVFLl5EosjMHnf3lryvKQiOUl8HbP0lvPiz4LH7qWB7xVxYclEYDJfCCS+H2OQNr10dfXz2e5v49pM7GBpxXnP6Aq6/ZCkXntyImc3QDyMi5U5BMBO698GWh+HFnwbBsH9zsD3VAM2XBKFw8mXQeGow3jDBns4+vvbIVr7x6Esc6B7gjJNquP7iZn737IVUJeMz+qOISPlREJRCx3bY8lAQCi/8FDpbg+1zTxprLSy9FOoWjzusb3CY/9qwnS8/vIVnd3cxb04F15y/hHddsIT5cytL8IOISDkoWRCY2RXA3wNx4A53/8yE1y8FvgCcBVzt7vce6T2PmyDI5Q5tL451I734s2DAGaC+GU6+HM5eDZnzRlsL7s7PN+/nyz9/kf/etIeKeIyrzlnI9RcvZfnCmtL9LCJyXCpJEJhZHHgOeB3QCqwDVrv70zn7NAM1wJ8Aa8s2CCZyh72bxkLhtz+GwW6Y9zJY+e4gFKrnje7+270HufPnW7j38VZ6B4e58ORG3nfJUl59+gJiMY0jiMiRlSoILgRudff/Ea7fAuDuf51n3zuB70QmCCbq74KN98ETX4PWxyCWhNOvhBXvgVMuh1gwRtDRM8jd617irl9sYWdHH82Nad578VLeem4T1ZWFnakkItFUqiB4K3CFu/9BuP5u4Hx3vynPvndymCAwsxuAGwAWL1587tatW4tS86ywZxP86mvw67uhZz/UNMGKa+Cca6B+CQCDwyM8uHEXX3r4RX71UjtzqxKsPm8x77lwCU316RL/ACIyGx33QZCrLFsE+QwNBPdOfuKr8Nv/DradfBmsfA+c/kZIBAPHT7zUxld+voX7n9qJu/O65SfwmjNO4FXL5nFSbapk5YvI7HK4IChmf8J2IJOz3hRuk0IkKuDMNweP9m2w4Rvwq6/Dve8NTkk9+2pY8W5WLl7OysX13PKG0/nqL7dy7+OtPLhxNwCnzK/mVcvmc8mp87jglEbmqPtIRPIoZosgQTBY/BqCAFgHvNPdN+bZ907UIjiykWF44SdBK2HTd2FkEBa1BK2El/8eVM7F3dm0q4uHn9/HQ5v38diL++kbHCERM1YurueSZfO4ZNk8zlpUSyKuGUZEoqKUp49eSXB6aBz4srt/2sw+Bax397Vm9krgPqAe6AN2ufuZh3vPSAdBru598OT/C0Jh7yZIVsPL3wJnXQ0LlkO6AczoGxzmia1tPLR5Hw89v5eNOzpxh7lVCS46pZFXLZvPq5bNY0mjprYQKWe6oKycuUPrenjiLvjNN4PTUAEqa4JrFBqWQv3S0eX2qiYe2lPJw5vbeHjzPra39wKQaUhxyalBKFx0SiN16YqS/UgiMv0UBFHR3wVbfg4HXgguYDvwYvDctjXoRsqKJaFuMd6wlM5UE88NNPJoWy0/3J1mU38jA1bJKxbVctGp83jFolrOOKmGJQ1pXbMgchxTEETdyDB07hgfDqPPW6C/Y9zuB5PzeIkT2NxXy1A4U3kiZtRUJahNV1CXSlKTSlKbSpA4zIR648SSwb0cGk4OWigNJ492X4lI8ZXqrCGZLWLx4Eu4LhPMb5TLHXrbxoXDnANbWN72Iqd37mBwaJjB4REGhkYYGHIGD4ww4s5B4KBBMhYjmYhRETcq4sFyPGYc8vU+1A9du4CcPzwqa8a6rhpODpazQTH3pMPO2ioi00dBEHVmwV/m6QZYdO64l2JAZfjIcnda23p5ZmcnT+/s5JmdnTyzs4uXDvSM7lOXTrL8pBrOCB/LT6rh1AVzqPABaN8aBM5o99ULsOsp2PQdGBka+6BEVTCuMS4kwtCoWwzxZDF/KyKRoq4hmRZdfYNs2tUVBMSOICA27eqif2gEgHjMaKpPsaSxmubG9LjnTEOKSvNghtYDL+QExZax9aHe8R+YSEFFNVSkgzOmxi2ng/XscjINFXPyLFdDVU0QLAnN7CrlTV1DUnRzq5K8srmBVzY3jG4bHnFe3NfN0zs7eW5XF1v2d7N1fw+/2tpGV//YX/9msLA2RfO8NEsa59PcuIQlTW+k+exqFjekSSVjQbdStvuqYxsMHISBbhjoCc6UGugJ1nvbwm3h+kB3cCe5w7EY1C2BeafBvGXhc/iobizWr0xk1lAQSNHEY8apC+Zw6oI5cPbYdnenrWcwDIZutuzrCZ739/DAUztp6xkc9z4n1lSxpDFNc+Mclsw7nyUNrybTkKKpPk19Onn4O7m5w/DAWCjkBsRgD/S2w4Hfwr7nYN/zwY2FhnLuQ52qHx8QjeFzfXPBtyQ9IvdgDKW/C/o7g+fBXqicG3x+qj5owcihsmNcPQeC7s1UvU5AOArqGpJZp6NnkK0HgmDYui98DoNi38H+cftWV8Rpqk+PBkNTfWrcem1qimMJIyNBi2Pf82E4PBfcbW7fc3Bw99h+sWQwdjFv2VhI1C0JQqe/a8KjM8+2CdtHBievCYIxk2wopBogVRcsZ7/8RrfXj9+ePA7nm8p+uR/cA917gufR5b3Bf4fscveeQ8eWahZCzaLwsXBsvZYRoWUAAAu9SURBVDbclm6MZFjo9FEpGwf7h9i6v5vWtl5a23rZdqAnXO5h24EeugfGdwPVVCXGBUMmDIqmhhSZ+vTUpu/ubR8LhdGgeD4YxzjcF3kiFfx1P+5Rk2dbzvZkVRAQ2b92e9sOffQcgN4DQfhM+tlVwRdfzUKozUBtUzAmUts0tp6qK/x3cLTcg+68rt3BF/noY0+eL/y9+X+fsSRUz4c5C4JH9YKx5VR98Pvo3B6cKt25I1ju2jk+KADilVBzUjCz7yFBsTA4Yy1RGXQZWjx8Dh+x7PrxFyQKAokEd6e9ZzAIiLaeMBzCkAjDom9wZNwxdekkDekK6tJJ6tMV1KUrqE8nqUsnw+VgvTZ8vT5dQapiwj2khweDi/Y6Xsr/pV/MM5zcg26k3rYgFMaFRHZ5P3S0Bi2djtZDg6OyZnww1GXC5fCU4zknjN4T4xAjw8EX98Hdeb7kJ2wb7Dn0+Fgi/EKfHz6fkLM84Qv/aLp9svV1bg9uH5sNiNGwaIXOnUdukeUzGhATwsJiwanP2eV4RfBIVOY8VwYTS+Z9rpxk/0pYfCHMP23qtaIgEAGCoNh3cGBcMOxo76Wte5C2ngHaewZp7xmgrWeQ3sHJB5grE7EwNJL5AyRVMRokwXoQJJWJSb5MZ9LISPDF2NEaBFdHazC7bUf4aN8Gfe3jj4klwr+YMzD3BOjrGPuS79kHPnLo51TVhl/qOY+5Jxy6LVVf+utFRkaCnyMbFl07g3D3kQmP4SB4c7eNDOfZb8Lrw4Mw3B+MAw0PTHjuD6acz76e+1q+cHrT30HL9Uf1YyoIRKaob3CYjt4gINq6g4Bo7x0LjLbuYD0bHO3h9qGRyf9/SlfEw1AYC43aVLY1EgRIbRgcdekKalNJalIJUsn44QfEp1t/VxgUrdD+0lhLon0bHNwFVXWHfrHPPTHnC37B8Tk2MduMjAShkBsWVbVBK/Mo6PRRkSmqSsapSsY5oaaq4GPcne6BYdq6B+joHQxaGL1jLY1gfXB0/bndB0eXDxcgybhRUxVM61FTlQieU8lwW4KaqmQYGmOv1+a8PuWWSOVcWHBG8JDSicUgVhWMFxWZgkBkmpgZcyoTzKlMjLsj05FkAyQbFtmWSGfvEJ19g3T2Bts6+4ZGl7e39wav9w4yMJynayZHZSJGbSrbAglaIbVhWGS31aWTowFSl0qOvq57VkSDgkCkxHIDpKl+ase6O/1DI3T2DtLZFwZGGCDBcvCc+2ht6+HpHUHrpGfg8BfbzalMjIZC9lFdmSBdEQ8f4XLlhPVxy8FzKhnXDLazlIJA5DhmZqPdWAum0I2VNTA0Qmff4GhLpKM36NbqCLuxssvZEPnt3oP0DAzTPTBEz8AwA0OHb41MlErGR4OjuiLB3KpETjdWMqdbayyAanKeqytmeLwkIhQEIhFWkYgxb04l8+Yc3VxLQ8Mj9AwO0zswTHd/EA69g8Fy78Aw3QPD9A4M0T0wHLwWLvcODHOwf4iuvkG2t/fxTG8Xnb2D46YeySeenQ49JxyyYyOViTgViRgV8RgViRjJ8LkibuPWx7bnrMdjVCSMinjwHqlknFRFsBwFCgIROWqJeIyaeIyaqum5VmJ4xOnqC7q3OsZ1d421SoJth46X9OdMmX6YsfcpScRsfPdWRdCSSYXdX9n17PLE7rFURWI0ZJLxWM5jbD2Yvt1IxILtpWjxKAhEZNaIxyy8/uLYbpU6POLhPTRGGBgOHoPZ5aHx66P32xgeO6Z3cHzrpSfsCusJl9t7BtjRPrbeMzA8OtPuscoNifGhYXz4tafxu2cvnJbPyaUgEJGyE48ZqfCv9JkyPOL0DAyFwTEWEkHYOEPZ0Bl2BoeC5cGRnOXsa8Mj4b4+GljZfevSxblKXUEgIjIN4jFjblWSudPUTTaTojESIiIik1IQiIhEnIJARCTiFAQiIhGnIBARiTgFgYhIxCkIREQiTkEgIhJxx90dysxsL7D1KA+fB+ybxnKmi+qaGtU1dbO1NtU1NcdS1xJ3n5/vheMuCI6Fma2f7FZtpaS6pkZ1Td1srU11TU2x6lLXkIhIxCkIREQiLmpBcHupC5iE6poa1TV1s7U21TU1RakrUmMEIiJyqKi1CEREZAIFgYhIxEUmCMzsCjN71sw2m9nNpa4HwMwyZvZjM3vazDaa2YdKXVMuM4ub2a/M7DulriXLzOrM7F4z22Rmz5jZhaWuCcDMPhL+N/yNmd1tZlUlquPLZrbHzH6Ts63BzH5gZs+Hz/WzpK6/Cf87Pmlm95lZ3WyoK+e1PzYzN7N5s6UuM/uj8He20cw+O12fF4kgMLM4cBvwBmA5sNrMlpe2KgCGgD929+XABcAfzpK6sj4EPFPqIib4e+B77n46cDazoD4zWwR8EGhx95cDceDqEpVzJ3DFhG03Az9y92XAj8L1mXYnh9b1A+Dl7n4W8Bxwy0wXRf66MLMM8HrgpZkuKHQnE+oys8uBVcDZ7n4m8Lnp+rBIBAFwHrDZ3V9w9wFgDcEvtKTcfae7PxEudxF8qS0qbVUBM2sC3gjcUepassysFrgU+BKAuw+4e3tpqxqVAFJmlgDSwI5SFOHuPwMOTNi8CrgrXL4LePOMFkX+utz9++4+FK4+AjTNhrpCfwd8DCjJ2TST1HUj8Bl37w/32TNdnxeVIFgEbMtZb2WWfOFmmVkzsAJ4tLSVjPoCwf8II6UuJMdSYC/wlbDL6g4zqy51Ue6+neCvs5eAnUCHu3+/tFWNc4K77wyXdwEnlLKYSVwPPFDqIgDMbBWw3d1/XepaJjgNeJWZPWpmPzWzV07XG0clCGY1M5sD/CfwYXfvnAX1vAnY4+6Pl7qWCRLASuCf3X0F0E1pujnGCfvcVxEE1UKg2szeVdqq8vPgfPFZdc64mf05QTfpN2ZBLWng48AnSl1LHgmggaAb+U+Be8zMpuONoxIE24FMznpTuK3kzCxJEALfcPdvlrqe0MXAVWa2haAb7dVm9vXSlgQELblWd8+2mu4lCIZSey3worvvdfdB4JvARSWuKdduMzsJIHyeti6FY2Vm1wFvAq7x2XFR0ykEgf7r8N9/E/CEmZ1Y0qoCrcA3PfAYQWt9WgayoxIE64BlZrbUzCoIBvLWlrgmwjT/EvCMu3++1PVkufst7t7k7s0Ev6v/dveS/4Xr7ruAbWb2snDTa4CnS1hS1kvABWaWDv+bvoZZMIidYy1wbbh8LfBfJaxllJldQdD9eJW795S6HgB3f8rdF7h7c/jvvxVYGf7bK7VvAZcDmNlpQAXTNENqJIIgHJC6CXiQ4H/Qe9x9Y2mrAoK/vN9N8Bf3hvBxZamLmuX+CPiGmT0JnAP8nxLXQ9hCuRd4AniK4P+rkkxRYGZ3A78EXmZmrWb2PuAzwOvM7HmC1stnZkld/wjMBX4Q/tv/l1lSV8lNUteXgZPDU0rXANdOVytKU0yIiERcJFoEIiIyOQWBiEjEKQhERCJOQSAiEnEKAhGRiFMQiExgZsM5p/NumM7Zas2sOd9MlyKllCh1ASKzUK+7n1PqIkRmiloEIgUysy1m9lkze8rMHjOzU8PtzWb23+G8+j8ys8Xh9hPCefZ/HT6y007Ezezfwjnlv29mqZL9UCIoCETySU3oGnpHzmsd7v4KgqtivxBu+wfgrnBe/W8AXwy3fxH4qbufTTAnUvZq9mXAbeGc8u3A7xf55xE5LF1ZLDKBmR109zl5tm8BXu3uL4STBe5y90Yz2wec5O6D4fad7j7PzPYCTdn548P3aAZ+EN4kBjP7MyDp7v+7+D+ZSH5qEYhMjU+yPBX9OcvDaKxOSkxBIDI178h5/mW4/AvGbk15DfBQuPwjgrtKZe//XDtTRYpMhf4SETlUysw25Kx/z92zp5DWhzOf9gOrw21/RHDXtD8luIPae8PtHwJuD2eOHCYIhZ2IzDIaIxApUDhG0OLu0zIHvMhsoa4hEZGIU4tARCTi1CIQEYk4BYGISMQpCEREIk5BICIScQoCEZGI+/8CPDITSYT74wAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "36U7SqSiH7uJ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        },
        "outputId": "61fc0054-01ed-4a54-c1dd-4c4096e41717"
      },
      "source": [
        "# evaluating on test set \n",
        "tloss_classifier, tacc_classifier = classifier.evaluate(x=X_test, y=Y_test)\n",
        "\n",
        "print(f\"Classifier Test Loss: {tloss_classifier:.3f}\")\n",
        "print(f\"Classifier Test Accuracy: {tacc_classifier:.3f}\")"
      ],
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "313/313 [==============================] - 1s 2ms/step - loss: 0.0394 - accuracy: 0.9883\n",
            "Classifier Test Loss: 0.039\n",
            "Classifier Test Accuracy: 0.988\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nUd3_bSs9yGL",
        "colab_type": "text"
      },
      "source": [
        "# Adding Noise"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pHcrtLgQJ9jI",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# adding noise \n",
        "X_train_noisy = X_train + np.random.rand(*X_train.shape) * 0.9\n",
        "X_test_noisy = X_test + np.random.rand(*X_test.shape) * 0.9\n",
        "\n",
        "# clipping noisy image pixel values to [0, 1]\n",
        "X_train_noisy = np.clip(X_train_noisy, 0., 1.)\n",
        "X_test_noisy = np.clip(X_test_noisy, 0., 1.)"
      ],
      "execution_count": 86,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7SGEk7bWJc7x",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 594
        },
        "outputId": "e2121816-37ec-4afd-b6ed-c7bb38aa3796"
      },
      "source": [
        "# visualizing noisy data\n",
        "plot(np.squeeze(X_train_noisy, axis=-1), predictions=to_categorical(Y_train))"
      ],
      "execution_count": 114,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 16 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dlaILtLTLain",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        },
        "outputId": "c56d09a1-6c46-4bbd-9e8a-8512b2331790"
      },
      "source": [
        " # evaluating classifier on noisy test data \n",
        "nloss_classifier, nacc_classifier = classifier.evaluate(x=X_test_noisy, y=Y_test)\n",
        "\n",
        "print(f\"Classifier Test Loss (Noisy Data): {nloss_classifier:.3f}\")\n",
        "print(f\"Classifier Test Accuracy (Noisy Data): {nacc_classifier:.3f}\")"
      ],
      "execution_count": 88,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "313/313 [==============================] - 1s 2ms/step - loss: 2.2603 - accuracy: 0.5162\n",
            "Classifier Test Loss (Noisy Data): 2.260\n",
            "Classifier Test Accuracy (Noisy Data): 0.516\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iP34hWWh9zmL",
        "colab_type": "text"
      },
      "source": [
        "# Building an Autoencoder"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0NAeJsKO93NY",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 629
        },
        "outputId": "b80c6bd6-894b-499e-82e3-206da63b3e58"
      },
      "source": [
        "# defining the model \n",
        "input_img = Input(shape=(28, 28, 1))\n",
        "\n",
        "x = Conv2D(16, (3, 3), activation='relu', padding='same')(input_img)\n",
        "x = MaxPooling2D((2, 2), padding='same')(x)\n",
        "x = Conv2D(8, (3, 3), activation='relu', padding='same')(x)\n",
        "x = MaxPooling2D((2, 2), padding='same')(x)\n",
        "x = Conv2D(8, (3, 3), activation='relu', padding='same')(x)\n",
        "encoded = MaxPooling2D((2, 2), padding='same', name='encoding')(x)   # shape: (4, 4, 8) = 128\n",
        "\n",
        "x = Conv2D(8, (3, 3), activation='relu', padding='same')(encoded)\n",
        "x = UpSampling2D((2, 2))(x)\n",
        "x = Conv2D(8, (3, 3), activation='relu', padding='same')(x)\n",
        "x = UpSampling2D((2, 2))(x)\n",
        "x = Conv2D(16, (3, 3), activation='relu')(x)\n",
        "x = UpSampling2D((2, 2))(x)\n",
        "decoded = Conv2D(1, (3, 3), activation='sigmoid', padding='same')(x)\n",
        "\n",
        "autoencoder = Model(input_img, decoded)\n",
        "\n",
        "# defining loss function and optimizer\n",
        "autoencoder.compile(\n",
        "        optimizer='adam',\n",
        "        loss='binary_crossentropy',\n",
        "        metrics = [\"accuracy\"]\n",
        "    )\n",
        "\n",
        "# displaying model summary\n",
        "autoencoder.summary()"
      ],
      "execution_count": 89,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"functional_5\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_3 (InputLayer)         [(None, 28, 28, 1)]       0         \n",
            "_________________________________________________________________\n",
            "conv2d_24 (Conv2D)           (None, 28, 28, 16)        160       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_15 (MaxPooling (None, 14, 14, 16)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_25 (Conv2D)           (None, 14, 14, 8)         1160      \n",
            "_________________________________________________________________\n",
            "max_pooling2d_16 (MaxPooling (None, 7, 7, 8)           0         \n",
            "_________________________________________________________________\n",
            "conv2d_26 (Conv2D)           (None, 7, 7, 8)           584       \n",
            "_________________________________________________________________\n",
            "encoding (MaxPooling2D)      (None, 4, 4, 8)           0         \n",
            "_________________________________________________________________\n",
            "conv2d_27 (Conv2D)           (None, 4, 4, 8)           584       \n",
            "_________________________________________________________________\n",
            "up_sampling2d_6 (UpSampling2 (None, 8, 8, 8)           0         \n",
            "_________________________________________________________________\n",
            "conv2d_28 (Conv2D)           (None, 8, 8, 8)           584       \n",
            "_________________________________________________________________\n",
            "up_sampling2d_7 (UpSampling2 (None, 16, 16, 8)         0         \n",
            "_________________________________________________________________\n",
            "conv2d_29 (Conv2D)           (None, 14, 14, 16)        1168      \n",
            "_________________________________________________________________\n",
            "up_sampling2d_8 (UpSampling2 (None, 28, 28, 16)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_30 (Conv2D)           (None, 28, 28, 1)         145       \n",
            "=================================================================\n",
            "Total params: 4,385\n",
            "Trainable params: 4,385\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Lm6BMiaEMgev",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "0ab22839-8016-48aa-ba35-5450e2c989ad"
      },
      "source": [
        "# training the model \n",
        "hist_autoencoder = autoencoder.fit(\n",
        "        x=X_train_noisy, y=X_train,\n",
        "        validation_split=0.2,\n",
        "        epochs=200,\n",
        "        batch_size=32, \n",
        "        callbacks=[EarlyStopping(monitor='val_loss', patience=5)]\n",
        "    )"
      ],
      "execution_count": 90,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1939 - accuracy: 0.7985 - val_loss: 0.1555 - val_accuracy: 0.8028\n",
            "Epoch 2/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1454 - accuracy: 0.8042 - val_loss: 0.1422 - val_accuracy: 0.8026\n",
            "Epoch 3/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1334 - accuracy: 0.8067 - val_loss: 0.1293 - val_accuracy: 0.8080\n",
            "Epoch 4/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1257 - accuracy: 0.8081 - val_loss: 0.1233 - val_accuracy: 0.8093\n",
            "Epoch 5/200\n",
            "1500/1500 [==============================] - 5s 4ms/step - loss: 0.1207 - accuracy: 0.8091 - val_loss: 0.1205 - val_accuracy: 0.8112\n",
            "Epoch 6/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1171 - accuracy: 0.8097 - val_loss: 0.1159 - val_accuracy: 0.8113\n",
            "Epoch 7/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1147 - accuracy: 0.8101 - val_loss: 0.1146 - val_accuracy: 0.8107\n",
            "Epoch 8/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1128 - accuracy: 0.8105 - val_loss: 0.1122 - val_accuracy: 0.8114\n",
            "Epoch 9/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1114 - accuracy: 0.8107 - val_loss: 0.1112 - val_accuracy: 0.8121\n",
            "Epoch 10/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1102 - accuracy: 0.8109 - val_loss: 0.1099 - val_accuracy: 0.8123\n",
            "Epoch 11/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1091 - accuracy: 0.8111 - val_loss: 0.1101 - val_accuracy: 0.8109\n",
            "Epoch 12/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1082 - accuracy: 0.8112 - val_loss: 0.1086 - val_accuracy: 0.8115\n",
            "Epoch 13/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1071 - accuracy: 0.8114 - val_loss: 0.1071 - val_accuracy: 0.8123\n",
            "Epoch 14/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1064 - accuracy: 0.8115 - val_loss: 0.1062 - val_accuracy: 0.8123\n",
            "Epoch 15/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1057 - accuracy: 0.8116 - val_loss: 0.1061 - val_accuracy: 0.8120\n",
            "Epoch 16/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1051 - accuracy: 0.8117 - val_loss: 0.1059 - val_accuracy: 0.8119\n",
            "Epoch 17/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1044 - accuracy: 0.8118 - val_loss: 0.1046 - val_accuracy: 0.8130\n",
            "Epoch 18/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1038 - accuracy: 0.8119 - val_loss: 0.1042 - val_accuracy: 0.8130\n",
            "Epoch 19/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1032 - accuracy: 0.8120 - val_loss: 0.1039 - val_accuracy: 0.8131\n",
            "Epoch 20/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1027 - accuracy: 0.8121 - val_loss: 0.1028 - val_accuracy: 0.8130\n",
            "Epoch 21/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1022 - accuracy: 0.8121 - val_loss: 0.1024 - val_accuracy: 0.8130\n",
            "Epoch 22/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1018 - accuracy: 0.8122 - val_loss: 0.1024 - val_accuracy: 0.8129\n",
            "Epoch 23/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1013 - accuracy: 0.8123 - val_loss: 0.1018 - val_accuracy: 0.8129\n",
            "Epoch 24/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1011 - accuracy: 0.8123 - val_loss: 0.1026 - val_accuracy: 0.8127\n",
            "Epoch 25/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1007 - accuracy: 0.8124 - val_loss: 0.1008 - val_accuracy: 0.8132\n",
            "Epoch 26/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.1005 - accuracy: 0.8124 - val_loss: 0.1018 - val_accuracy: 0.8140\n",
            "Epoch 27/200\n",
            "1500/1500 [==============================] - 5s 4ms/step - loss: 0.1001 - accuracy: 0.8125 - val_loss: 0.1003 - val_accuracy: 0.8130\n",
            "Epoch 28/200\n",
            "1500/1500 [==============================] - 5s 4ms/step - loss: 0.0999 - accuracy: 0.8125 - val_loss: 0.1015 - val_accuracy: 0.8124\n",
            "Epoch 29/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0997 - accuracy: 0.8125 - val_loss: 0.1005 - val_accuracy: 0.8136\n",
            "Epoch 30/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0994 - accuracy: 0.8126 - val_loss: 0.0998 - val_accuracy: 0.8133\n",
            "Epoch 31/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0992 - accuracy: 0.8126 - val_loss: 0.1006 - val_accuracy: 0.8126\n",
            "Epoch 32/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0990 - accuracy: 0.8126 - val_loss: 0.0994 - val_accuracy: 0.8132\n",
            "Epoch 33/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0987 - accuracy: 0.8127 - val_loss: 0.0989 - val_accuracy: 0.8135\n",
            "Epoch 34/200\n",
            "1500/1500 [==============================] - 5s 4ms/step - loss: 0.0985 - accuracy: 0.8127 - val_loss: 0.0986 - val_accuracy: 0.8136\n",
            "Epoch 35/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0983 - accuracy: 0.8127 - val_loss: 0.0986 - val_accuracy: 0.8135\n",
            "Epoch 36/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0981 - accuracy: 0.8127 - val_loss: 0.0992 - val_accuracy: 0.8141\n",
            "Epoch 37/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0979 - accuracy: 0.8128 - val_loss: 0.0989 - val_accuracy: 0.8131\n",
            "Epoch 38/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0977 - accuracy: 0.8128 - val_loss: 0.0980 - val_accuracy: 0.8136\n",
            "Epoch 39/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0976 - accuracy: 0.8128 - val_loss: 0.0978 - val_accuracy: 0.8135\n",
            "Epoch 40/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0975 - accuracy: 0.8128 - val_loss: 0.0979 - val_accuracy: 0.8134\n",
            "Epoch 41/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0974 - accuracy: 0.8129 - val_loss: 0.0983 - val_accuracy: 0.8140\n",
            "Epoch 42/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0972 - accuracy: 0.8129 - val_loss: 0.0976 - val_accuracy: 0.8139\n",
            "Epoch 43/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0971 - accuracy: 0.8129 - val_loss: 0.0975 - val_accuracy: 0.8137\n",
            "Epoch 44/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0971 - accuracy: 0.8129 - val_loss: 0.0975 - val_accuracy: 0.8140\n",
            "Epoch 45/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0968 - accuracy: 0.8129 - val_loss: 0.0971 - val_accuracy: 0.8137\n",
            "Epoch 46/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0967 - accuracy: 0.8129 - val_loss: 0.0972 - val_accuracy: 0.8136\n",
            "Epoch 47/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0965 - accuracy: 0.8130 - val_loss: 0.0970 - val_accuracy: 0.8141\n",
            "Epoch 48/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0963 - accuracy: 0.8130 - val_loss: 0.0972 - val_accuracy: 0.8133\n",
            "Epoch 49/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0962 - accuracy: 0.8130 - val_loss: 0.0969 - val_accuracy: 0.8136\n",
            "Epoch 50/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0960 - accuracy: 0.8130 - val_loss: 0.0967 - val_accuracy: 0.8136\n",
            "Epoch 51/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0960 - accuracy: 0.8130 - val_loss: 0.0963 - val_accuracy: 0.8140\n",
            "Epoch 52/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0958 - accuracy: 0.8130 - val_loss: 0.0972 - val_accuracy: 0.8133\n",
            "Epoch 53/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0957 - accuracy: 0.8131 - val_loss: 0.0961 - val_accuracy: 0.8137\n",
            "Epoch 54/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0955 - accuracy: 0.8131 - val_loss: 0.0961 - val_accuracy: 0.8139\n",
            "Epoch 55/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0955 - accuracy: 0.8131 - val_loss: 0.0957 - val_accuracy: 0.8141\n",
            "Epoch 56/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0954 - accuracy: 0.8131 - val_loss: 0.0977 - val_accuracy: 0.8130\n",
            "Epoch 57/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0953 - accuracy: 0.8131 - val_loss: 0.0977 - val_accuracy: 0.8132\n",
            "Epoch 58/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0953 - accuracy: 0.8131 - val_loss: 0.0984 - val_accuracy: 0.8130\n",
            "Epoch 59/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0951 - accuracy: 0.8131 - val_loss: 0.0954 - val_accuracy: 0.8141\n",
            "Epoch 60/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0951 - accuracy: 0.8132 - val_loss: 0.0970 - val_accuracy: 0.8143\n",
            "Epoch 61/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0950 - accuracy: 0.8132 - val_loss: 0.0958 - val_accuracy: 0.8137\n",
            "Epoch 62/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0950 - accuracy: 0.8132 - val_loss: 0.0957 - val_accuracy: 0.8141\n",
            "Epoch 63/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0948 - accuracy: 0.8132 - val_loss: 0.0955 - val_accuracy: 0.8139\n",
            "Epoch 64/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0948 - accuracy: 0.8132 - val_loss: 0.0953 - val_accuracy: 0.8139\n",
            "Epoch 65/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0947 - accuracy: 0.8132 - val_loss: 0.0951 - val_accuracy: 0.8140\n",
            "Epoch 66/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0946 - accuracy: 0.8132 - val_loss: 0.0954 - val_accuracy: 0.8138\n",
            "Epoch 67/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0945 - accuracy: 0.8132 - val_loss: 0.0946 - val_accuracy: 0.8142\n",
            "Epoch 68/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0945 - accuracy: 0.8132 - val_loss: 0.0952 - val_accuracy: 0.8139\n",
            "Epoch 69/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0944 - accuracy: 0.8133 - val_loss: 0.0949 - val_accuracy: 0.8139\n",
            "Epoch 70/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0943 - accuracy: 0.8133 - val_loss: 0.0951 - val_accuracy: 0.8137\n",
            "Epoch 71/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0943 - accuracy: 0.8133 - val_loss: 0.0946 - val_accuracy: 0.8141\n",
            "Epoch 72/200\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.0943 - accuracy: 0.8133 - val_loss: 0.0950 - val_accuracy: 0.8138\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7z0H0bQsM9mH",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        },
        "outputId": "393a5288-4061-48c4-c761-f5d65a0e8bdb"
      },
      "source": [
        "# plotting training curves \n",
        "plot_acc(hist_autoencoder)\n",
        "plot_loss(hist_autoencoder)"
      ],
      "execution_count": 91,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEWCAYAAABxMXBSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXxU1fn48c+Tyb4TEtawRAVZVEAjrrjhgrjgWqFapW61Lu3X5WfVqlWrXdVqq+JerbVS1FrRgiIKrqgEBJQdWQMIIUD2STKZ5/fHuYFJyDJAhgHyvF+vvDJz7zl3zhC9zz27qCrGGGNMuGKiXQBjjDH7FgscxhhjdooFDmOMMTvFAocxxpidYoHDGGPMTrHAYYwxZqdY4DCmGSLSW0RURGLDSDtWRD7bE+UyJtoscJj9goisFJEaEcludPwb7+bfOzolM2b/Y4HD7E9WAGPq34jIoUBy9IqzdwinxmTMzrDAYfYnrwCXh7y/AvhHaAIRyRCRf4hIkYisEpG7RSTGO+cTkYdFZJOILAfOaiLvCyKyXkTWisiDIuILp2Ai8rqI/CAiJSLyiYgMDDmXJCKPeOUpEZHPRCTJO3e8iHwhIltFZI2IjPWOTxeRq0Ou0aCpzKtl3SAiS4Gl3rHHvWuUisgsERkWkt4nIneJyPciUuad7yEiT4rII42+y0QRuTmc7232TxY4zP7kSyBdRPp7N/TRwD8bpfkbkAEcAJyICzQ/9c5dA5wNDAHygYsa5X0JCAAHeWlOB64mPJOBPkAnYDbwasi5h4EjgGOBLOB2ICgivbx8fwNygMHAnDA/D+A84ChggPd+pneNLOBfwOsikuiduwVXWxsJpANXApXAy8CYkOCaDZzq5Tftlaraj/3s8z/AStwN7W7g98AI4AMgFlCgN+ADaoABIfl+Bkz3Xn8EXBdy7nQvbyzQGagGkkLOjwGmea/HAp+FWdZM77oZuIe3KmBQE+nuBN5q5hrTgatD3jf4fO/6p7RSji31nwssBkY1k24hcJr3+kZgUrT/3vYT3R9r+zT7m1eAT4A8GjVTAdlAHLAq5NgqoLv3uhuwptG5er28vOtFpP5YTKP0TfJqPw8BF+NqDsGQ8iQAicD3TWTt0czxcDUom4jcBlyF+56Kq1nUDyZo6bNeBi7DBeLLgMd3o0xmP2BNVWa/oqqrcJ3kI4H/NDq9CajFBYF6PYG13uv1uBto6Ll6a3A1jmxVzfR+0lV1IK37MTAKVyPKwNV+AMQrkx84sIl8a5o5DlBBw47/Lk2k2bb0tdefcTvwI6CDqmYCJV4ZWvusfwKjRGQQ0B/4bzPpTDthgcPsj67CNdNUhB5U1TpgAvCQiKR5fQi3sL0fZALwCxHJFZEOwB0hedcDU4BHRCRdRGJE5EAROTGM8qThgk4x7mb/u5DrBoEXgUdFpJvXSX2MiCTg+kFOFZEfiUisiHQUkcFe1jnABSKSLCIHed+5tTIEgCIgVkTuxdU46j0P/FZE+ohzmIh09MpYiOsfeQV4U1WrwvjOZj9mgcPsd1T1e1UtaOb0Tbin9eXAZ7hO3he9c88B7wNzcR3YjWsslwPxwAJc/8AbQNcwivQPXLPXWi/vl43O3wZ8i7s5bwb+CMSo6mpczelW7/gcYJCX5y+4/poNuKakV2nZ+8B7wBKvLH4aNmU9igucU4BS4AUgKeT8y8ChuOBh2jlRtY2cjDEtE5ETcDWzXmo3jXbPahzGmBaJSBzwS+B5CxoGLHAYY1ogIv2BrbgmuceiXByzl7CmKmOMMTvFahzGGGN2SruYAJidna29e/eOdjGMMWafMmvWrE2qmtP4eLsIHL1796agoLnRmcYYY5oiIquaOm5NVcYYY3aKBQ5jjDE7xQKHMcaYndIu+jiaUltbS2FhIX6/P9pFibjExERyc3OJi4uLdlGMMfuBdhs4CgsLSUtLo3fv3oQsk73fUVWKi4spLCwkLy8v2sUxxuwH2m1Tld/vp2PHjvt10AAQETp27NgualbGmD2j3QYOYL8PGvXay/c0xuwZ7TpwGGPMNqow/7+waWm0S7LXs8ARJcXFxQwePJjBgwfTpUsXunfvvu19TU1Ni3kLCgr4xS9+sYdKakw7UBeAd2+G16+A18dCMNhqlvYsooFDREaIyGIRWSYidzRxvqeITBORb0RknoiM9I539I6Xi8gTzVx7ooh8F8nyR1LHjh2ZM2cOc+bM4brrruPmm2/e9j4+Pp5AINBs3vz8fP7617/uwdIasx+rLoPXLoFZf4cDT4EN38G3r+/ZMqz+Cjav2LOfuRsiFjhExAc8CZwJDADGiMiARsnuBiao6hBgNPCUd9wP3IPbGa2pa18AlEei3NE0duxYrrvuOo466ihuv/12vv76a4455hiGDBnCsccey+LFiwGYPn06Z599NgD33XcfV155JSeddBIHHHCABRQTWf5SKCmMbhmCQfjuTXj7Rqip3L1rlayFF8+E76fBOY/DpW9Cl8Ng2oMQqG6b8rZm+XT4+5kw7liY/Q/XZLYrvnoGPv/rruffCZEcjjsUWKaqywFEZDwwCrd1Zj1l+77HGcA6AG+v6M+8vZQbEJFU3D7R1+K2utxt978znwXrStviUtsM6JbOb84ZuNP5CgsL+eKLL/D5fJSWlvLpp58SGxvL1KlTueuuu3jzzTd3yLNo0SKmTZtGWVkZBx98MD//+c9tzsbe7MMHIKUTHH1d212zLgBrZ0G3IRAb33bXDRWodje48g3wf99CXFLreZpTFwDfLtx+VnwKH9wL62a7931OgwGjdq0MFZvg+VNdjePSCXDQqe74qffBPy+AWS/BUT9rmGflZ/DxH6H3MPe5OQfv2mfX27IKXv8pdDwI0rrAxJtg2YdwzmOQ1CH86yx8Fybf7l5vXQ1n/gliItegFMnA0Z2GexoXAkc1SnMfMEVEbgJSgFPDuO5vgUeA3XzU2DtdfPHF+Hw+AEpKSrjiiitYunQpIkJtbW2Tec466ywSEhJISEigU6dObNiwgdzc3D1Z7Ohb/aW7EfQ/O9oladnSqfDpI5CYAUeMhbjEtrnuZ39xT8lJWXDoRTBojAsiLY2oWzfHBZsjxkKMr/XPmPY714wDMO/fLt/OUnV9CYsnw/UzIDkrvHxFS+CDe2DJe5CeC6OehPfvgiVTdj1wzH8LytbBlVOgZ8it6cBTXGD4+E8w+MeQkOaOr/wcXr0YfHGw4hOY9hDk9Hef3+tY6HJow+/jL3Hpln0IVVvgpDuhU7/t52sq4d+XQrAOxrwGHfLgi8fhowfd32Xkn6H38ds/vzmbV8B/r3d/717HwYwnoKYCzv3brgXnMER7AuAY4CVVfUREjgFeEZFDVLXJnikRGQwcqKo3i0jvli4sItfiaiX07NmzxULsSs0gUlJSUra9vueeezj55JN56623WLlyJSeddFKTeRISEra99vl8LfaP7JfqauHNa6ByE9y2pPX/0aKlphL+dwskZLibyuJJcMgFu3/dys3wxV/dTSO1M8x6Gb5+FnL6uSfmwZdC7Pb/RgjUwCd/gk8fBa2DpR/Ahc9DQmrzn7FqBnz+OBx+hXva/3Kce72zQ70/vN/1JQB8+RSccnfL6QPVrpyfPgJxyXDq/e47xSW5G/KyD1zT1a48XS94G7IPbhg0wH2nU++H50+BGU/CSXfAqi9c0MjIhSvedf9uC9911/j4j7jGE1xQ63Io+LfCmq9duvhUEB8s+h8MuwWOv8X9PSbeBD98Bz+eAB0PdPmPvxnyToA3roLXRgPi/o65R0CPo91DQWhNr9bvOvQFuPglyOwFCekw/XdQUw4XvhCRGmgkO8fXAj1C3ud6x0JdhdfcpKozgEQgu4VrHgPki8hK4DOgr4hMbyqhqj6rqvmqmp+Ts8Ny8vuEkpISunfvDsBLL70U3cLszeZNgJLVUFsJCyZGuzTN+/iPsHUVXPIKpHWDua+1zXU/e9Q1t5z1CFz8dxc8z37M3WDevRkeH+RugDUVsH4uPHcyfPJnOOwSOP0hWPo+/H2Ea+9vSnUZvPUz6NALzvgdHH0DFC2C7z/cuXLOeNLVjI4YC/3PdW3yVVuaT7/qC3j6ePj4DzDwPLhpFhz/f9tvnH1Od81mP8zduXIAlBfBqs9hwLlNn889AvqfA1/8zQWIVy+G9K5wxTuQ1hnSu8FR18JP/wf/73v4yVtw2gPQ6xjYsgJqq+C4X8LYSfCrla7sA893/w08fTz871b47g0XOPue3vCzux8BP//C9becdAdk9oBFk2DijfC3I+CbV10tBWDKr93f9LynoUNvF/RO+pX7Oy2cCOPH7H4/UBMiWeOYCfQRkTxcwBgN/LhRmtXAcOAlb2/jRKCouQuq6jhgHIBX43hXVU9q64LvLW6//XauuOIKHnzwQc4666xoF2fvFKxzN84uh7r/Wef8C4ZcGu1S7WjDfNeEMPgyOOBEGDTaPcGX/eDatndV6Tr4+jl3vU793bGkTMj/qbtBL5/untbfvws+eRiqSyG5I4wZDwef6dLn9HNDUJ8f7ppMug1p+Bnv3Qkla+Cnk12t5JALYepvXK3joHBal4G5/3Zl6H8OnPUobFzobmxfjoOT72qYVtV95lfjILOnu4H2aeJz+pwGiGuualzm1ix6FzTYcjPXKfe6WsK/L4WsA11No6m/VUpH17x14CnNXys1By58DgZd4oJ5wQvus4fd2nT6+GT3neu/t6rrX/ngXnj7evffUt8RMPN5OPYm6DeyYf5jboD4FJjxlHtgiE9u+d9jZ6lqxH6AkcAS4Hvg196xB4BzvdcDgM+BucAc4PSQvCuBzbjRU4XAgEbX7g18F045jjjiCG1swYIFOxzbn+233/fbN1V/k6763VuqnzzsXhcvj3apGqqrU31uuOof81Qrit2xoiWurJ89vmP6imLVSberLpqkWhdo+doTf6l6f0fVzStbTrdqhur4S1Xfvml7GUL98J3qowNVf9tJ9eVRqu//WnXOa6ozX3Dl/OA3DdNP/5M7vnHRjtcq36S6bo7q8k9UF76r+sUTqvdnqf79LNWaqu3pXvux6u96qFZuaZj/Y+/a796iWl3e8vd69hT3s7P+cZ7q44NVg8GW0713l+qTx6iWrN35z2hOdbnqvNdb/25NCQZVv/uPK/tv0lWfP101UNN8+lr/rpdTVYECbeKeKroHhm5FW35+vjbeAXDhwoX0798/SiXa8/bL76vqqv11NXD9V1C2Hv4yEE78FZx8Z+v5l0111fj+5+x8W31TZdnwHSx8x3U6dzwQOh8CXQ5x/QPv/QrOf8bVDOo9N9w1r/38i4af/8ZVrhkDXJv10GthyGWuJhGq+Ht4cijkX+k6UndX2QbXNr7uG9i4COq84aidD4VrPmrYVl6xCR4dAIPHuGGs9Wa/4p6og40GcnQdDFdMdIMC6q2fB88Mc53GJ3nTvBb9D8b/2DWjnf9M63+X6X+E6b+H/7cMUlpq5Q5RuRke7gPH3Ain3d9y2vr74962bE9drfu36j3M1XgiRERmqWp+4+PR7hw3Ztctec/drM972nWOZnSHA06Cuf9ywaOlDtNgHfznWqgsdv/znfXIrg2t3LQUZr/sAsaWlSAxkN3XjaYJVG1Pl3eiuxmGGjzGtXWvnwvdBrtjC952QePEX0GnAa4fYMqv3QieQy6Ewy+H3CPdjWza78AXD8OanO6089I6bw8CdQEoXgZFCyF36I4drCnZrtll7njXpJOU6ZpRZjzh/gZHXu06aRPT3e/MXjuO8Ol6GBx8luskP/rnbn7If651bfzn/DW8m3Xf012wW/qB+/cMx+LJEAyENxprbwsY9Xxxrt8nSixwmH2Tqmuzz+zpRprUG3wp/Odq1/GZN6z5/IUzXdA49GJYOgXGHefahU+83bUNt6auFj5/zA3ZVHU3y+NvgYNHuvbsYJ2rEWz4FjYtc/0ujW9CAy9wbflzX3OBo2ITvHsLdB0EJ/y/7TeH9XPdKKnv3oRvXnEjgfqd5QLM8be4G35b88W6oaOhw0cbO/p6N2Hti8dhwwI3wmnotXDG78MfBnri7fDs/2Da790os/hUuOTV8IcpdxnkRpItnRJ+4FjwNmT03Pl+EbONBQ6zb1rxMawtcB2tvpDJjv3Ock+4c19rOXAseQ9iYl1NI1DjOns/f8xV/6/7rOUb1/q58PYN8MO3bqTMmX+C1E4N08T4IKev+2lOcpYLNPMmwGm/dUN1q0vhvHcafqeug9y8hRF/cHMPZr/iBgQkZsBxUVyzrFN/1yH8+eNuuOlZj8KRV+3cNboNhr5nuo5wX4LrgE/vGn7+mBg46DRY9E54kwr9JbB8mgtwe2ttYifUBZXauiDVgSC1dUFUIUYgRoSYGCFGIDUhts1XyLbAYfY99bWN1C6uhhEqPtk9pX/7pruhNzc3YfF7btJWfZv7eU+5UUb/vsw91Q+9pul8H//ZtamnZLsn492dcDj4x7Dgv26o5YK3YfhvoHPjlXk8CWmuqerwy92EOJGdm10cCSfd6Ya2nvGQGy22K06+09UAR/zeDYPdWX1Phzn/hDVfQe/jWk675H3XJ9a/mWG4O0FVqaqtY2tlLYE6RVGC6o5XVNexqaKa4vIaNpVXU+avRXA3chEhRoS6YJBAUN1PnRIIupt/TcAFg6raOkqrain1Byjz11JeHdiWrs7LF04X9aLfjiAxLowJnjvBAodpW9XeEmItTSbbXQvfgZWfwog/Nl0zGHypa0JZONHdmBvbstK13R9+ecPj/c52k6w++4s7FzppDrzZwg+6WsbZf2mbm/aBw93yI/P+7dr2jw2zBtFSTWZP6jEUfv7Z7l2j6yC4bemuL5FxwMmu9rh0CvQ+DlWluqaG4DevUltTzaYDL6A8mEBFTYCDvppAWmIn3l7fmYrVK6iqCVBZU0dVbR3+2jqqvNdVtUH83rHauu13ZwGCqmytrGVLZQ3VgfBW0fXFuCf+oDa82cf5BF+MEBsTQ5xPiPPFEOeLIT42hoTYGNKT4uiemUR6UhppCbHE+mKIjXF5fDHb08f5hPjYGEQEVaUuuD2IxfnafrqeBY4oOfnkk7njjjs444wzth177LHHWLx4MePGjdsh/UknncTDDz9Mfn4+I0eO5F//+heZmQ1H2dx3332kpqZy221t1FnaGn8pFC2G9XNg7Ww3o7hosbvhHnKhG+3T/Yi2bRKoLof37nAjlo68uuk0PY6CrAPcnI6mAsfi99zvg0c0PC7i2tz/eYHLm//T7efqAjD5V65P5bxxu7dOUyhfrCvjV0+760ZoiYhoCwaVOu+GVuc9YfsDdVTXBvEH3A27pKqWrVW1bK2sYWtlLRXVAXcT927m1YEggbogtXXuidw9lQeprq3jT9qPjM/f5OyPj2ZQ3Xzuj3uJ/jHeikcfPsCUuuFMqDuJyfHTGV93Mve9NX9b2eJ8QmKcj6Q43/bf8T6S4mLISoknNiYGke0DrGIEDsuNo0NyPJnJ8XRIjiPO59KIgCAkxfvITk0gJzWBjqnxpCRs/7uqFzxE9t1N1vbP/0r3AWPGjGH8+PENAsf48eP505/+1GreSZMmRbJozaupdDNff/jWBYjSkFVSU3Kg2+HuabzsB9duP+dVt9LowPNd23LJGrcAW9kG1xzT5zTXPt2hV/hl+PgPULoWLnqx+ZusiLsZf/Sgm2jWqdEw5CWT3cinrAN2zHvgKdA93y1zMeSy7X0NBS/CxgXwo1faLmjUO/nXcNR1O9e2v4cF6oKUVwco8wcoqap1N/nKWrZW1VBaFaDSe3KvrKmjqibAlspaNlfUUFxeTXFF+E/moRJiY0iKdzfypDgf8bHbn65jfTEkx8eSleIjKd7H+q0nkL/hKd7OeZp+W6ZTltCFqQc/Ql1yZw5Z/Q+uW/8/fh77DgCnXngtZxxwPMnxsSTH+yLyRN4SEdnnu1cscETJRRddxN13301NTQ3x8fGsXLmSdevW8dprr3HLLbdQVVXFRRddxP337zjOvHfv3hQUFJCdnc1DDz3Eyy+/TKdOnejRowdHHLELbcThmv+W60DufKhrS8452M067nKYW8Mn9P+G0x6AbyfAzBfd+kQxcS5NZg/ocSQUFrgOanA38Y59AHWPdRp0aU+5u+GicRvmu5mwQ34CPY9uuaz5V8Hnf4Op98OPx28/7i91i9Ud/fOm84m4obD/utgNNT38J1BR7IbD5p3o5ny0tdj4Ng8aqoq/Nrjthu6vdU/tNQHXPl7/9F9dG2SL94S/ubJm29N+/bGtlbWU+muprKlr8fNEICnOR3K8u5FnJceTnRpP385pdEyNJzneR2yM67B1TS0xJMbFkBjrnvIT42LISIojMzmOjKR4Mr2n+LAVpcCTT9Gv9AsYdhtpw27l1G2zpc9xCwF+9TSUriV30CnhLepommWBA2DyHe4pui11ORTO/EOzp7Oyshg6dCiTJ09m1KhRjB8/nh/96EfcddddZGVlUVdXx/Dhw5k3bx6HHXZYk9eYNWsW48ePZ86cOQQCAQ4//PDIBo7FkyC9O1z3aevNT4nprikp/yo34SqpQ8M2bFU3T2DpB24Y59ZV3jW9+v73H7nPO/9pN9Q1GHRDVRMzXFBqTXIWDLsZpt7nlmrofbw7/v1HbnJa/XIbTelzmpuw9unDbpXZaQ+69ZrO/OMeGYlTF1S2VNawpaKGLdtu4jXUBIIE1bWTBxX8tXVsrnDpiitcmlJ/wOtQrW3QNh+O2BjZ1vSSmRxHj6xkDu0eR0ZSHGmJcaQlxpKWGEt6UhyZSXFkJMeRmRRPelIsSXG+6Da75PR1gxU69d++YGCorDz39zNtwgJHFNU3V9UHjhdeeIEJEybw7LPPEggEWL9+PQsWLGg2cHz66aecf/75JCe7J6tzz939kSLNqql0q5EOuWznbp4iTc9sFYHsPu7nmOt3PL9+Lrx5NfzjPLcWT4fesOZLOPeJ8JfiPuo6t47TlHvg6g9d4FryngtiuUNbLvOJt7sZzFN/4/ZlOPKaHZu8wuSvraOorJpN5dVs8kbZbKmsoaSqllKv2WdLRS3F3iiczZU1Ye/FkxzvIyslnqwU197es2MK6d7NPS0xlpT42B2ae3wxblRPrE+I98W4tvqUONIiMGxzj9rbl9Tfj1jggBZrBpE0atQobr75ZmbPnk1lZSVZWVk8/PDDzJw5kw4dOjB27Fj8fn9UyraD5dPdTOjGi6lFStdBcO3HbmG8L7xdDXscvePw25bEJbnmrv/+HBa8BQPOcyNv+pzeeif0wSNdk9yMJ9weF42WMKkO1LGxtJofSv1sLK2muKKaTWXVbKqoYVOZa9ffVO4CQXl108vcx8e65pkM7wn+gOxUjuwdT8fUBDqmxNMhxT39u07YOBLjfMSIILhx+glxMW0+zNKYcFjgiKLU1FROPvlkrrzySsaMGUNpaSkpKSlkZGSwYcMGJk+e3OweHAAnnHACY8eO5c477yQQCPDOO+/ws5/9rNn0u2XR/9w+Er2Oj8z1mxKf7HZCO+hUdwM/+y87P2TzsEvcct5T73cd+JXF0PeMHZIF6oIUlVdTVFbN5grXvp/W5acM33AL/82+hg/eWsGWisVsrqihyAsMjYlAx5R4OqYkkJ0Wz6DcTLK9UTX1o2uyUxPITksgKzmepHi76Zt9kwWOKBszZgznn38+48ePp1+/fgwZMoR+/frRo0cPjjuu5clMhx9+OJdccgmDBg2iU6dOHHnkkZEpZLDONfH0OS1y25K2pP/Zu9wMUatC6bG/puNbY6h6/ToSJJZxa/JYu/RbNpZWs7HMzw8lfjaVVxPcoXmoC7nyOBVru9EhpZSs5HhyOyRzeK8OdElPpEt6Ip0zEumUlkB2agJZKfHbxusbsz+z1XHbid36vqtmuI1+LnrRzc/YiwTqgvxQ6mftlipWb67c9rOquJLCLZUUV9Sgqvwz7ncc75vP53UDuSzwa7KS48lJS6BTeiJd0hPokp5Ip3QXBLK2NRPFk5EUZ8HAtFu2Oq7ZdYvedcNpDzotKh+vqhSVV7NsYznLNpazdIP7vXpzJT+U+qkLqSrECHTLTKJXx2ROG9CZzumJdEpLJD7wIDr1Qg497TKWHHfmHh+7b8z+xALHvihY53b1SkgLf4RToAaeP80tJy0xLp8v3o08am77THDDZhf9z+2DnJjeNuVvRkllLfPXl7BgXSlLN5SzdmsV67ZWsXZrVYMJZGkJsRzYKZWheVl0z0yie4ckcjskkdshme6ZScTHNhUUesLAuaSnd9/1pS2MMUCEA4eIjAAeB3zA86r6h0bnewIvA5lemjtUdZKIdATeAI4EXlLVG730ycDrwIFAHfCOqt6xq+VT1X1r+GGwzi29Xb4BtM7NfA7dGKcZqgr+rW62d48jvbUTFLaugQk/gaE/g9N/u+PaTOD2lt6ywg2JbauvEVRWb65k0Q+lLFxfxsL1pSxYX0rhlu37V2SnxtO9QzL9u6YzvH8numcmcWCnVPp0SqNzesKu/d0ye7bZdzCmPYtY4BARH/AkcBpu69eZIjJRVReEJLsbmKCq40RkADAJtyWsH7gHOMT7CfWwqk4TkXjgQxE5U1Un72z5EhMTKS4upmPHjnt/8AgGodILGMGAq2lUV7hJaa0EDlWleON6Eovnw7E3uvkJ9QI1boLcl0+6lUUvfslNlAq16H/u98G7NgxXVVlf4mfumq3MKdzK3DVb+bawhApvJrII5GWnMKhHJj8+qicDu2UwsFs62alNBDFjzF4hkjWOocAyVV0OICLjgVFAaOBQoL79IwNYB6CqFcBnInJQ6AVVtRKY5r2uEZHZQO6uFC43N5fCwkKKiop2JfueVbXZLe4Xm+gCRWwtlG8F3QxpZa1mT9y6lNx5j8MNXzQ8ERsPI37nlhd/+3p45gQ3K3vQ6O3rMS36n1uoMMwlMcr8tXyz2gWIuYUlzC3cSlGZ24I0zicM6JrOBYfnckj3dPp1Sadv5zQblmrMPiaSgaM7sCbkfSFwVKM09wFTROQmIAU4NdyLi0gmcA6uKWynxcXFkZeX13rCaNu6Bv46zK2ZdPZfth//dLJbA+q2ZW7HueaUb4Q3TnAzvpvbm7j/2W6JlKoeVKIAACAASURBVDevhnf/z1338MvdMuPrZsMp9zR7eVXl+6IKpi/eyEeLNjJz5eZtS10cmJPCsIOyOSw3g8E9O9C/axoJsRYkjNnXRbtzfAyuD+MRETkGeEVEDlHVFpfSFJFY4DXgr/U1mibSXAtcC9Cz5z7ctv25FxePv6Xh8bwT3O+Vn8IhFzSf/+vn3DanR9/Q8ud06AVXTXHrOn39DHzxt+2f3a/hHIrKmgBfLi9m+uIipi8uYvXmSgD6dk7lyuPzGHZQDof1yCA9Ma7xpxhj9gORDBxrgR4h73O9Y6GuAkYAqOoMEUkEsoGNrVz7WWCpqj7WXAJVfdZLR35+/r45WaXsB7ch0eAxblXZUF0HQ3ya21youcBRUwkzn3cL+mUf1HSaUCJuu9W8Ya6mU/CCu0bOwawvqeKDBRv4YMEGvlqxmZpAkKQ4H8ce2JFrhuVxcr9O5HZIbv0zjDH7vEgGjplAHxHJwwWM0UDjXXVWA8OBl0SkP5AItNjpICIP4vpDmtnFZz/y+V9dZ/jxN+94zhfrljZf+Wnz+ee+5vpHjrlxpz86mJ7LwgE3M23RRqY8+TnzCksAOCAnhcuP7sWJB+dwZO8sWyvJmHYoYoFDVQMiciPwPm6o7YuqOl9EHgAKVHUicCvwnIjcjOsoH6veVHYRWYnrOI8XkfOA04FS4NfAImC2NxrqCVV9PlLfI2oqNrnNgw69uOkNh8A1Vy15D0rWQkb3hueCQfjyKeg2xHV+h6GorJpPlxbxyZIiPlu2iU3lbj2mwT0yuX3EwZw+oAsHdYrglrDGmH1CRPs4VHUSboht6LF7Q14vAJpckElVezdz2b187GwbmfEEBPww7Nbm04T2cwwa3fDckvfcfhcXvtDiJMFVxRW8P/8HpszfwKzVW1B1C/Ud3yebYX1yGNYnm87pTezrbYxpt6LdOW6aUrnZdWoPPN9tUNOcTgPdkt8rPmkYOIJB+OTPkNHDLSXeSFFZNf/9Zi1vzi5k0Q9uOO+Arun8cngfTu3fmQFd04mx9ZmMMc2wwLE3+uppqCmHE25rOV1MjNvZbsUnbjZ4fc3i2wluGO15T2/bdyJQF+SjRRt5fVYh0xZtJBBUBvfI5J6zB3D6gM70yLKObWNMeCxw7CmVm90mQvV9Ej/+d9M72VVthS+fdkNgOw9s/bp5J8DCiW5ZkKwD3BpWU+93fRuHXYK/to7XZxXy7Cffs2ZzFdmpCVx1fB4X5+dyUKe0tv+expj9ngWOSFs2FT552C3poUFI6QQVRfDZX9z6UI19OQ6qS+DEX4V3/bwT3e8Vn7rA8flfoWwd5ec+y8sfL+fvn69gU3kNg3tk8uuR/Rnev7OtDGuM2S0WOCLtk0dg01IYdhscPAK6DnHLe3z9rFuZNnQ0VNUWNxKq39nQtel9xneQ3QdSO7vmqoNOJfj5Y3yXcQqjX6mgsmYxJ/bN4ecnHchReVl7/5pcxph9ggWOSFKFjQtcJ/cpv95+/OS74Ls3YfrvYdQT249/OQ6qS+GknVjwVwTNO4GapdOYufxGjqyt45fF53HWoK6MPa43A7u1vnquMcbsDAsckVT2g1vOvNOAhscze8KRV7tO8GNvgpyDvdrGOOh/jls3KgyqyidLNzFvRVdu8m/ieD6ioNeVvHnJGLJSorDFqzGmXbDG7kja6C0E3HnAjueG3QZxKfDhA+79jKdcbePE8GobXy4v5kfPzOCKF7/m4xp3fU3tTP6lD1jQMMZElNU4Iqk+cOQ0sdd3Skc47hcw7SFYMsXVNgaMgi6Ntx9paOmGMn43aSHTFhfROT2B3553CJfk94DJXyF9z3B7dRhjTARZ4IikjQtdx3Vzy5kffb2b6PfvS6GupsWRVMXl1Tw2dSn/+no1yXE+7jizH2OP7b19rahzml3v0Rhj2pQFjkjauAA6NVHbqJeQ6nbkm3Sbm+HdxLyNYFB5ecZKHp2yhMraOi49qie/HN6HjrZDnjEmSixwREowCBsXQf5PW053+BWuE/2IK3Y49X1ROb96Yx4Fq7ZwYt8c7jl7gC0yaIyJOgsckbJlBQSqdhxR1VhsPAxvuMNeoC7I85+t4NEPlpAU5+MvlwzivMHdbR6GMWavYIEjUjYudL9bCxyNlFTVcvXLM5m5cgsjBnbhgfMG0inNVqc1xuw9LHBESn3gyDk47CzF5dX85IWvWbqxjMcuGcx5Q7q3nskYY/YwCxyRsnEBZPZyHeBh+KHEz2UvfMWazZU8d3k+Jx3cKcIFNMaYXRPRCYAiMkJEFovIMhHZYWabiPQUkWki8o2IzBORkd7xjt7xchF5olGeI0TkW++af5W9teF/44LwVrcF1myu5EfPzGD91ipevnKoBQ1jzF4tYoFDRHzAk8CZwABgjIg0bvC/G5igqkNwe5I/5R33A/cATW1IMQ64Bujj/Yxo+9LvpkC1232vpaG4nh9K/PzomRlsrazh1WuO5ugDmpnzYYwxe4lI1jiGAstUdbmq1gDjgVGN0ihuX3GADGAdgKpWqOpnuACyjYh0BdJV9Utvb/J/ADtucRdtxcsgGGi1Y9xfW8e1rxRQWlXLa9cezeAemXuogMYYs+si2cfRHVgT8r4QOKpRmvuAKSJyE5ACnBrGNQsbXXPv60HeNqKq+RqHqnLnf75lXmEJz/7kCFvF1hizz4j2IodjgJdUNRcYCbwiIm1SJhG5VkQKRKSgqKioLS4Zvg3zISYWOvZpNsnzn67grW/WcstpfTl9YJc9WDhjjNk9kQwca4EeIe9zvWOhrgImAKjqDCARyG7lmrmtXBPves+qar6q5ufk5Oxk0XfTxoUuaMQ2vUrtx0uK+P3khYw8tAs3nXLQni2bMcbspkgGjplAHxHJE5F4XOf3xEZpVgPDAUSkPy5wNFs9UNX1QKmIHO2NproceDsShd8tLaxRtaq4gpv+NZu+ndP480WDbDa4MWafE7HAoaoB4EbgfWAhbvTUfBF5QETO9ZLdClwjInOB14CxXqc3IrISeBQYKyKFISOyrgeeB5YB3wOTI/Uddkl1OWxd1WzH+B8mL6IuqDx3eT4pCTaNxhiz74nonUtVJwGTGh27N+T1AuC4ZvL2buZ4AdDyphXRVLTI/W6ixvHd2hImf/cDvxjehx5ZyXu4YMYY0zai3Tm+/2lh179HpiwmIymOq4fl7eFCGWNM27HA0dY2LoTYJMjs3eDwrFWbmba4iJ+deADpiXHRKZsxxrQBCxxtbeMC6NQPYrb/06oqf35/MdmpCYw9tnf0ymaMMW3AAkdbUoUNC3boGP98WTFfLt/MDScfSHK8dYgbY/ZtFjja0qalULERuh+x7ZCq8vCUxXTNSGTM0J5RLJwxxrQNCxxtadlU9/ug7SunfLhwI3PWbOUXw/uQGOeLUsGMMabtWOBoS8umQnZf6NBr26FnPvmenlnJXHREbgsZjTFm32GBo63UVMLKz+Cg07YdWlVcwcyVWxg9tAdxPvunNsbsH+xu1lZWfQ511XDQ8G2H3vpmLSJw3uC9bwFfY4zZVRY42sqyqW7+Ri83EV5V+c/stRx7YEe6ZSZFuXDGGNN2LHC0laUfQO/jIS4RgIJVW1i9uZILhljfhjFm/2KBoy1sXg6bv4c+2/s3/jO7kKQ4HyMOsb02jDH7FwscbWHZh+63NwzXX1vHu/PWc+YhXWwFXGPMfscCR1tYNhU69IasAwCYunADZf4AF9oQXGPMfsgCx+4KVMOKT9wwXG9TpjdnFdI1I5GjD+gY5cIZY0zbs8Cxu1bPgNrKbc1UG8v8fLJ0E+cN6Y4vxnb3M8bsfyIaOERkhIgsFpFlInJHE+d7isg0EflGROaJyMiQc3d6+RaLyBkhx28Wkfki8p2IvCYiiZH8Dq1aNhV88ZA3DICJc9ZRF1QuGGJzN4wx+6eIBQ4R8QFPAmcCA4AxIdu/1rsbt6XsENye5E95eQd47wcCI4CnRMQnIt2BXwD5qnoI4PPSRc/SqdDrWIhPAeA/s9dyWG4GfTqnRbVYxhgTKZGscQwFlqnqclWtAcYDoxqlUSDde50BrPNejwLGq2q1qq7A7S8+1DsXCySJSCyQHJJnzysphKKF25upSv0sWF/KWYd2jVqRjDEm0loNHCJyjojsSoDpDqwJeV/oHQt1H3CZiBTi9ia/qaW8qroWeBhYDawHSlR1yi6UrW1sXOh+5x4JuEl/AEfmZUWrRMYYE3HhBIRLgKUi8icR6dfGnz8GeElVc4GRwCstBSkR6YCrjeQB3YAUEbmsmbTXikiBiBQUFRW1cbE9/hL3O8kFioKVW0iIjeGQbhmR+TxjjNkLtBo4VPUyYAjwPfCSiMzwbsqtNeKvBXqEvM/1joW6Cpjgfc4MIBHIbiHvqcAKVS1S1VrgP8CxzZT7WVXNV9X8nJyc1r7mrvFvdb+TMgEoWLWZwT0yiY+1wWrGmP1XWHc4VS0F3sD1U3QFzgdmi8hNLWSbCfQRkTwRicd1Yk9slGY1MBxARPrjAkeRl260iCSISB7QB/jaS3+0iCSLiHh5F4b1TSOhygscCelUVAeYv66UI3tbM5UxZv/W6noYInIu8FPgIOAfwFBV3SgiycAC4G9N5VPVgIjcCLyPG/30oqrOF5EHgAJVnQjcCjwnIjfjOsrHqqoC80Vkgnf9AHCDqtYBX4nIG8Bs7/g3wLO78f13j78EYhMhLpE5yzZRF1SO6N0hasUxxpg9IZyFlC4E/qKqn4QeVNVKEbmqpYyqOgnX6R167N6Q1wuA45rJ+xDwUBPHfwP8JoxyR56/BBJdf0bByi2IwOE9LXAYY/Zv4QSO+3AjmAAQkSSgs6quVNUPI1WwfUJo4Fi1mYM7p5GRFBflQhljTGSF08fxOhAMeV/nHTNe4AjUBZm9aov1bxhj2oVwAkesN4EPAO91fOSKtA/xAseiH8qoqKkj3/o3jDHtQDiBo8jrIAdAREYBmyJXpH2IvwQSMylYuRmAfKtxGGPagXD6OK4DXhWRJwDBzei+PKKl2lf4t0JiBjNXbaFbRiLdbW9xY0w70GrgUNXvcXMnUr335REv1b5AFfwlaGIGBfM2c1Se7b1hjGkfwtrXVETOwq1UmyjeZkWq+kAEy7X3q62EYIASTWZDaTVHWv+GMaadCGeRw6dx61XdhGuquhjoFeFy7f28dapWlPsA698wxrQf4XSOH6uqlwNbVPV+4Bigb2SLtQ/wAseirTGkJcTS1/bfMMa0E+EEDr/3u1JEugG1uPWq2jcvcMwtgsN7dbBtYo0x7UY4geMdEckE/oxbI2ol8K9IFmqf4C1wuGBLjPVvGGPalRY7x729MT5U1a3AmyLyLpCoqiV7pHR7M6/GUUqyrU9ljGlXWqxxqGoQt294/ftqCxqe+sChKRzYKTXKhTHGmD0nnKaqD0XkQqkfh2scL3DUxqXSKS0hyoUxxpg9J5zA8TPcoobVIlIqImUiUhrhcu39/FuplkS6ZWVgMdUY056EM3Pcxpk2xV9CKSn06pgc7ZIYY8weFc4OgCc0dbzxxk7tjfpL2BJMond2SrSLYowxe1Q4TVX/L+TnHuAd3OZOrRKRESKyWESWicgdTZzvKSLTROQbEZknIiNDzt3p5VssImeEHM8UkTdEZJGILBSRY8IpS1urKd9MiSbTM8tqHMaY9iWcpqpzQt+LSA/gsdbyiYgPNyLrNKAQmCkiE73tYuvdDUxQ1XEiMgC3zWxv7/Vo3PpY3YCpItLX23f8ceA9Vb1IROKBqNy5ayu2Uqop9O5oNQ5jTPsSTo2jsUKgfxjphgLLVHW5t/nTeGBUozQKpHuvM4B13utRwHhv+O8KYBkwVEQygBOAF8BtKuXNMdnz/FspJdn6OIwx7U44fRx/w93gwQWawbgZ5K3pjtu7o14hcFSjNPcBU0TkJiAFODUk75eN8nYHqoAi4O8iMgiYBfxSVSuaKPe1wLUAPXv2DKO4O8dXU0o5KXSzPTiMMe1MODWOAtwNehYwA/iVql7WRp8/BnhJVXOBkcAr3mz15sQChwPjVHUIUAHs0HcCoKrPqmq+qubn5OS0UXG3XZyEQDmSmGFrVBlj2p1w9uN4A/B7/QuIiE9EklW1spV8a4EeIe9zvWOhrgJGAKjqDBFJBLJbyFsIFKrqVyFlazJwRFRNOTEEiU21pUaMMe1PWDPHgdD2mCRgahj5ZgJ9RCTP68QeDUxslGY1MBxARPoDibimqInAaBFJEJE8oA/wtar+AKwRkYO9/MOBBexhWrUFgKQ02/XPGNP+hFPjSAzdLlZVy0Wk1R5hVQ2IyI3A+4APeFFV54vIA0CBqk4EbgWeE5Gbcf0oY1VVgfkiMgEXFALADfU1HtyGUq96wWg58NOwv20b2bplEx2AtEwLHMaY9iecwFEhIoer6mwAETkC10ndKlWdhBtiG3rs3pDXC4Djmsn7EPBQE8fnAPnhfH6kbNy4gQ5AVlYb950YY8w+IJzA8X/A6yKyDrd1bBfcVrLt1ubiIgA65nSKckmMMWbPC2cC4EwR6QfU9yssVtXayBZr77Z1SzEAnSxwGGPaoVY7x0XkBiBFVb9T1e+AVBG5PvJF23tVlrjAkZCaFeWSGGPMnhfOqKprQmdnq+oW4JrIFWnvV13uRlWRkN5yQmOM2Q+FEzh8oZs4eWtQxUeuSHu/YOUW/DHJ4Auni8gYY/Yv4dz53gP+LSLPeO9/BkyOXJH2biVVtSQEyqhNSScx2oUxxpgoCCdw/Aq35tN13vt5uJFV7dLq4krSpRISrZnKGNM+tdpUpapB4CtgJW7F21OAhZEt1t5rZXEF6VTiS86MdlGMMSYqmq1xiEhf3CKEY4BNwL8BVPXkPVO0vdOq4goOkAriU9t+xV1jjNkXtNRUtQj4FDhbVZcBeEuDtGsriyvpEFNFbLItcGiMaZ9aaqq6AFgPTBOR50RkOG7meLu2vY8jI9pFMcaYqGg2cKjqf1V1NNAPmIZbeqSTiIwTkdP3VAH3Nqs2lZGsFRY4jDHtVjid4xWq+i9v7/Fc4BvcSKt2p7ImQFX5VmJQCxzGmHZrp/YcV9Ut3s56wyNVoL3Z6s1eMxVY4DDGtFs7FTjau5WbKknH297cAocxpp2ywLETVhVXhNQ4bB6HMaZ9imjgEJERIrJYRJaJyA57g4tITxGZJiLfiMg8ERkZcu5OL99iETmjUT6fl+fdSJa/sfUlfjrF+d0bq3EYY9qpiAUObzHEJ4EzgQHAGBEZ0CjZ3cAEVR2C25P8KS/vAO/9QGAE8JR3vXq/JAqz10v9tXS2wGGMaeciWeMYCixT1eWqWgOMB0Y1SqNA/aJPGcA67/UoYLyqVqvqCmCZdz1EJBc4C3g+gmVvUpk/QHast2uuBQ5jTDsVycDRHVgT8r7QOxbqPuAyESnE7U1+Uxh5HwNuB4ItfbiIXCsiBSJSUFRUtEtfoLHSqlqyfH5AbC8OY0y7Fe3O8THAS6qaC4wEXhGRZsskImcDG1V1VmsX9oYN56tqfk5OTpsUtswfoENMpQsaMdH+pzPGmOiI5N1vLdAj5H2udyzUVcAEAFWdASQC2S3kPQ44V0RW4pq+ThGRf0ai8E0p9deSEVMJSdZMZYxpvyIZOGYCfUQkT0TicZ3dExulWQ0MBxCR/rjAUeSlGy0iCSKSB/QBvlbVO1U1V1V7e9f7SFUvi+B3aKDMHyANW27EGNO+RWzvU1UNiMiNwPuAD3hRVeeLyANAgapOBG4FnvNW3VVgrKoqMF9EJgALgABwg6rWRaqs4VBVyvy1pKRW2BwOY0y7FtFNs1V1Eq7TO/TYvSGvF+Can5rK+xDwUAvXng5Mb4tyhqOipo6gQnKwDBIb9/EbY0z7YT28YSqtqgUgsa7cmqqMMe2aBY4wlfkDAMQHyixwGGPaNQscYSr11+KjjriA9XEYY9o3CxxhKvPXkorNGjfGGAscYSqtCpAutqS6McZY4AhTmb+WDNuLwxhjLHCEq9QfsN3/jDEGCxxhK/XXkuWzPg5jjLHA0ZJv/gnffwTBIKVVge17cSTZqCpjTPsV0Znj+7RgED59BDYvhw55DPWdTpVvk1vM3Wocxph2zAJHc2Ji4PovYcFEKHiR81c/445LDMSnRrdsxhgTRdZU1ZLYBDjsYrhyMjd1GMeU1FEw9FoQiXbJjDEmaqzGEaYFgW7Udf0lp595RLSLYowxUWU1jjCV+gOkJ8ZFuxjGGBN1FjjCVOavJT3JAocxxljgCENNIIi/NkhagrXsGWNMRAOHiIwQkcUiskxE7mjifE8RmSYi34jIPBEZGXLuTi/fYhE5wzvWw0u/QETmi8gvI1n+emV+txeH1TiMMSaCneMi4gOeBE4DCoGZIjLR2/Wv3t3ABFUdJyIDcLsF9vZejwYGAt2AqSLSF7eN7K2qOltE0oBZIvJBo2u2uVJvL460RKtxGGNMJGscQ4FlqrpcVWuA8cCoRmkUSPdeZwDrvNejgPGqWq2qK4BlwFBVXa+qswFUtQxYCER8H9dtNQ7rHDfGmIgGju7AmpD3hex4k78PuExECnG1jZvCzSsivYEhwFdNfbiIXCsiBSJSUFRUtGvfwFNaZTUOY4ypF+3O8THAS6qaC4wEXhGRVsskIqnAm8D/qWppU2lU9VlVzVfV/JycnN0qpPVxGGPMdpF8hF4L9Ah5n+sdC3UVMAJAVWeISCKQ3VJeEYnDBY1XVfU/kSl6Q6Ve4LAahzHGRLbGMRPoIyJ5IhKP6+ye2CjNamA4gIj0BxKBIi/daBFJEJE8oA/wtYgI8AKwUFUfjWDZGyjzOsetxmGMMRGscahqQERuBN4HfMCLqjpfRB4AClR1InAr8JyI3IzrKB+rqgrMF5EJwALcSKobVLVORI4HfgJ8KyJzvI+6S1UnRep7AJRW1SICqfFW4zDGmIjeCb0b+qRGx+4Neb0AOK6ZvA8BDzU69hmwx1cYLPUHSE2IJSbGFjc0xphod47vE0r9tTYU1xhjPBY4wlDmD1jHuDHGeCxwhKG0ymocxhhTzwJHGMr8AdKTrMZhjDFggSMspf5a0qzGYYwxgAWOsJT5A6RbH4cxxgAWOFqlqraJkzHGhLDA0YqKmjqCasuNGGNMPQscrSitsiXVjTEmlAWOVpRt28TJAocxxoAFjlaVbltS3ZqqjDEGLHC0qmzbkupW4zDGGLDA0ar63f9sOK4xxjgWOFphNQ5jjGnIAkcrSv2237gxxoSywNGK0qpa4mNjSIzzRbsoxhizV4ho4BCRESKyWESWicgdTZzvKSLTROQbEZknIiNDzt3p5VssImeEe822VuoP2BwOY4wJEbHAISI+4EngTGAAMEZEBjRKdjcwQVWH4PYkf8rLO8B7PxAYATwlIr4wr9mm3CZO1kxljDH1IlnjGAosU9XlqloDjAdGNUqjQLr3OgNY570eBYxX1WpVXQEs864XzjXbVJk/QJqtU2WMMdtEMnB0B9aEvC/0joW6D7hMRApxe5Pf1ErecK4JgIhcKyIFIlJQVFS0q9/B28TJahzGGFMv2p3jY4CXVDUXGAm8IiJtUiZVfVZV81U1PycnZ5evU2b7jRtjTAORfJReC/QIeZ/rHQt1Fa4PA1WdISKJQHYreVu7Zpsqtd3/jDGmgUjWOGYCfUQkT0TicZ3dExulWQ0MBxCR/kAiUOSlGy0iCSKSB/QBvg7zmm2qzHb/M8aYBiL2KK2qARG5EXgf8AEvqup8EXkAKFDVicCtwHMicjOuo3ysqiowX0QmAAuAAHCDqtYBNHXNSH2HmkAQf23Q+jiMMSZERO+IqjoJ1+kdeuzekNcLgOOayfsQ8FA414wUW27EGGN2FO3O8b1a/XIj1sdhjDHbWeBowbYaR4LVOIwxpp4FjhZsW1LdJgAaY8w2FjhasL2Pw5qqjDGmngWOFmzfNtZqHMYYU88CRwvKbC8OY4zZgQWOFpRW1SICqfEWOIwxpp4FjhaU+gOkJsQSEyPRLooxxuw1LHC0oNQWODTGmB1Y4GhBmT9g/RvGGNOIBY4WlFbV2ogqY4xpxAJHC8r8AVvg0BhjGrG7YguOPqAj3TITo10MY4zZq1jgaMG95wyIdhGMMWavY01VxhhjdooFDmOMMTslooFDREaIyGIRWSYidzRx/i8iMsf7WSIiW0PO/VFEvvN+Lgk5PlxEZnt5PhORgyL5HYwxxjQUsT4OEfEBTwKnAYXATBGZ6O36B4Cq3hyS/iZgiPf6LOBwYDCQAEwXkcmqWgqMA0ap6kIRuR64Gxgbqe9hjDGmoUjWOIYCy1R1uarWAOOBUS2kHwO85r0eAHyiqgFVrQDmASO8cwqke68zgHVtXnJjjDHNimTg6A6sCXlf6B3bgYj0AvKAj7xDc4ERIpIsItnAyUAP79zVwCQRKQR+AvyhmWteKyIFIlJQVFS021/GGGOMs7d0jo8G3lDVOgBVnQJMAr7A1UJmAHVe2puBkaqaC/wdeLSpC6rqs6qar6r5OTk5kS6/Mca0G5EMHGvZXksAyPWONWU025upAFDVh1R1sKqeBgiwRERygEGq+pWX7N/AsW1bbGOMMS2J5ATAmUAfEcnDBYzRwI8bJxKRfkAHXK2i/pgPyFTVYhE5DDgMmOKdzhCRvqq6BNfxvrC1gsyaNWuTiKzaxe+RDWzaxbx7mpU1MvaVsu4r5QQra6S0dVl7NXUwYoFDVQMiciPwPuADXlTV+SLyAFCgqhO9pKOB8aqqIdnjgE9FBKAUuExVAwAicg3wpogEgS3AlWGUZZfbqkSkQFXzdzX/nmRljYx9paz7SjnByhope6qsEV1yRP9/e/cWKlUVx3H8+yO7mIXXEMHiFIpRpGZRWRFlECXhS4FJDxJCEBIK0UWCKOilXroTdO8hIrpYImHaSSIqksxLXjK7CBnq0cCim5T9e1hrdDeN6YYzs5fx+8Bm9l5zzvA7w9I1e+3Z/xXxNulaRbXtnrbjJqnjVQAABTFJREFUezv83u+kb1Z1es3FwOLBS2lmZnWUcnHczMyOEh44Du+ppgPU4KzdcbRkPVpygrN2S0+y6p+XFszMzP6bzzjMzKwWDxxmZlaLB45DOFxl36ZJek7SgKQNlbZRklZI2pofRzaZMWc6VdJKSZskbZS0oOCsJ0haJWldznpfbj9d0ie5L7wi6bims7ZIOkbSGklL83GRWSVtk/R5rmr9aW4rrg8ASBoh6TVJX0jaLGl6aVklTapUFl8r6SdJC3uV0wNHB5XKvteQvhY8R1JpywG+wMHCjy13Af0RMRHoz8dN+xO4LSLOAi4C5uf3ssSs+4AZETGFVJn5akkXAQ8AD0XEBNK9Q/MazNhuAf+8CbbkrFfkahCt+wxK7AMAjwDLIuJMYArp/S0qa0Rsye/lVOA84FfSbQq9yRkR3to2YDrwTuV4EbCo6VwdcvYBGyrHW4BxeX8csKXpjB0yv0W647/orMCJwGfAhaQ7cYd06hsNZxyf/3OYASwlleYpNes2YExbW3F9gFRx+1vyF4dKzlrJdhXwYS9z+oyjsyOu7FuYsRGxI+/vBMY2GaadpD7SmiufUGjWPPWzFhgAVgBfA3sjVy6grL7wMHAH8Fc+Hk25WQNYLmm1pJtzW4l94HRgN/B8ngJ8RtIwyszaUq3115OcHjj+pyJ95Cjmu9aSTgJeBxZGWpDrgJKyRsT+SKf/40lrypzZcKSOJF0LDETE6qazHKFLI2Iaafp3vqTLqk8W1AeGkBaRezIizgV+oW26p6Cs5GtYs4BX25/rZk4PHJ3Vqexbkl2SxgHkx4GG8wAg6VjSoPFSRLyRm4vM2hIRe4GVpOmeEZJa5XlK6QuXALMkbSMtkjaDNDdfYlYi4vv8OECai7+AMvvAdmB7HKzA/RppICkxK6SB+LOI2JWPe5LTA0dnByr75hH9BmDJYX6nBEuAuXl/Lul6QqOUKlU+C2yOiOraKSVmPUXSiLw/lIPVl1cC1+cfKyJrRCyKiPER0Ufqn+9FxI0UmFXSMEknt/ZJc/IbKLAPRMRO4DtJk3LTlcAmCsyaVVdOhV7lbPrCTqkbMBP4kjTHfXfTeTrkexnYAfxB+pQ0jzTH3Q9sBd4FRhWQ81LS6fJ6YG3eZhaadTKwJmfdANyT288AVgFfkaYEjm86a1vuy4GlpWbNmdblbWPr31OJfSDnmgp8mvvBm6RlH4rLCgwDfgCGV9p6ktMlR8zMrBZPVZmZWS0eOMzMrBYPHGZmVosHDjMzq8UDh5mZ1eKBw2wQSNrfVq100IrLSeqrVkE2a9qQw/+ImR2B3yKVKjH73/MZh1kX5XUoHsxrUaySNCG390l6T9J6Sf2STsvtYyUtzmuCrJN0cX6pYyQ9ndcJWZ7vbDdrhAcOs8ExtG2qanbluR8j4hzgcVJFW4DHgBcjYjLwEvBobn8UeD/SmiDTSHdaA0wEnoiIs4G9wHVd/nvMDsl3jpsNAkk/R8RJHdq3kRaH+iYXe9wZEaMl7SGtm/BHbt8REWMk7QbGR8S+ymv0ASsiLc6DpDuBYyPi/u7/ZWb/5jMOs+6LQ+zXsa+yvx9fn7QGeeAw677ZlceP8/5HpKq2ADcCH+T9fuAWOLCo1PBehTQ7Uv7UYjY4huaVA1uWRUTrK7kjJa0nnTXMyW23klaZu5204txNuX0B8JSkeaQzi1tIVZDNiuFrHGZdlK9xnB8Re5rOYjZYPFVlZma1+IzDzMxq8RmHmZnV4oHDzMxq8cBhZma1eOAwM7NaPHCYmVktfwNTbNJT2tC1sgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cvyjk-SoNcLx",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        },
        "outputId": "ed09ea14-a1d9-4b67-a349-e56db9a02409"
      },
      "source": [
        "# evaluating on test set \n",
        "tloss_autoencoder, tacc_autoencoder = autoencoder.evaluate(x=X_test_noisy, y=X_test)\n",
        "\n",
        "print(f\"Autoencoder Test Loss: {tloss_autoencoder:.3f}\")\n",
        "print(f\"Autoencoder Test Accuracy: {tacc_autoencoder:.3f}\")"
      ],
      "execution_count": 92,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "313/313 [==============================] - 1s 3ms/step - loss: 0.0934 - accuracy: 0.8122\n",
            "Autoencoder Test Loss: 0.093\n",
            "Autoencoder Test Accuracy: 0.812\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yZWnPvm4NqIa",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 594
        },
        "outputId": "6943f4f4-92a9-47e6-aafa-3d934b993c4e"
      },
      "source": [
        "#visualizing denoised images \n",
        "X_denoisy = autoencoder.predict(X_test_noisy)\n",
        "plot(np.squeeze(X_denoisy, axis=-1), predictions=to_categorical(Y_test))"
      ],
      "execution_count": 111,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 16 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QNnNp5cx93zV",
        "colab_type": "text"
      },
      "source": [
        "# Composite Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "x_zDl41C941Z",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# defining the model \n",
        "noisy_input = Input(shape=(28, 28, 1))\n",
        "x = autoencoder(noisy_input)\n",
        "y = classifier(x)\n",
        "\n",
        "denoise_and_classify = Model(noisy_input, y)"
      ],
      "execution_count": 96,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uIB2ADXxRDAM",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "denoise_and_classify.compile(\n",
        "        loss='sparse_categorical_crossentropy',\n",
        "        metrics = [\"accuracy\"]\n",
        "    )"
      ],
      "execution_count": 102,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lj1jiY_EQSnX",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        },
        "outputId": "288f23da-ddcc-40b4-f2e2-c956a73428a7"
      },
      "source": [
        "# evaluating on noisy test set\n",
        "dnloss_classifier, dnacc_classifier = denoise_and_classify.evaluate(x=X_test_noisy, y=Y_test)\n",
        "\n",
        "print(f\"Denoised Classifier Test Loss (Noisy Data): {dnloss_classifier:.3f}\")\n",
        "print(f\"Denoised Classifier Test Accuracy (Noisy Data): {dnacc_classifier:.3f}\")"
      ],
      "execution_count": 103,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "313/313 [==============================] - 1s 2ms/step - loss: 0.1090 - accuracy: 0.9641\n",
            "Denoised Classifier Test Loss (Noisy Data): 0.109\n",
            "Denoised Classifier Test Accuracy (Noisy Data): 0.964\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ek7-vYTqQnYX",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 594
        },
        "outputId": "010107c5-d79b-4cfb-c0a5-ce6afcdf15da"
      },
      "source": [
        "# visualizing composite model precitions \n",
        "Y_denoisy = denoise_and_classify.predict(X_test_noisy)\n",
        "plot(np.squeeze(X_test_noisy, axis=-1), predictions=Y_denoisy)"
      ],
      "execution_count": 113,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjwAAAJBCAYAAABCs2+FAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOydd5QV1dbEqyUMIDkIkpMowYwJRFRUzOkZUAQRTJgDomJGBRVFURExoiJmBHPCCEZAyaBIkCggICDOkPr7g8f9etc+zDDInXlc67eWy1v09Jm+HU6f6VNdO4rjGEIIIYQQmcwOhb0BQgghhBDpRgMeIYQQQmQ8GvAIIYQQIuPRgEcIIYQQGY8GPEIIIYTIeDTgEUIIIUTGowGPEEIIITIeDXg2QxRFu0RRlB1F0eDC3hax/RFFUbsoiqZEUfRXFEW/RlHUqrC3SWxfRFE0OIqiBVEUrYii6Ocois4v7G0S2xdRFF0WRdHoKIpyoigaVNjbU9hECh4ME0XRRwBKApgdx/E5hb09YvshiqIjATwF4EwA3wPYGQDiOJ5XmNslti+iKGoKYHocxzlRFO0G4HMAx8VxPKZwt0xsL0RRdCqADQDaAigZx3Gnwt2iwkVPeAJEUdQOwHIAIwp7W8R2yR0AesZx/G0cxxviOJ6nwY7IL3EcT4rjOGeT/O9/DQpxk8R2RhzHQ+M4Hgbgj8Lelv8FNOAhoigqC6AngGsKe1vE9kcURUUANAdQJYqi6VEUzY2i6NEoikoW9raJ7Y8oih6Lomg1gKkAFgB4r5A3SYjtFg14PHcCeDqO47mFvSFiu6QqgGIATgPQCsBeAPYGcHNhbpTYPonj+BIAZbDxXBoKICf3NYQQm0MDngRRFO0F4AgADxb2tojtlr//+/9H4jheEMfxEgB9ARxbiNsktmPiOF4fx/FIADUBdC3s7RFie6VoYW/A/xiHAqgL4LcoigCgNIAiURQ1ieN4n0LcLrGdEMfxsiiK5mKj3yL1z4W1PSKjKAp5eITYavSEx/IENnYoe/33v8cBvIuNDnchtpRnAVweRdFOURRVAHA1gHcKeZvEdsR/z512URSVjqKoSBRFbQGcBb1IIfJBFEVFoygqAaAINv7xXiKKon/tg45/7RcPEcfxagCrN+koilYByI7jeHHhbZXYDrkTQGUAPwPIBvAqgLsLdYvE9kaMjdNXj2PjH6azAVwVx/FbhbpVYnvjZgC3JfQ52PgW6e2FsjWFjHJ4hBBCCJHxaEpLCCGEEBmPBjxCCCGEyHg04BFCCCFExqMBjxBCCCEyHg14hBBCCJHx5Ou19KJFi8bFihVL6SJFipjlO++8s1vn77//NnqHHXbIdXm5cuVcG4sX27fC169fn+s6lStXznM7Vq9ebXTyewHAvHm+1mO1atWMLl68uNHLli0zetWqVa6NWrVq5doGf7esrCzXxuzZs43m47Bu3TqjGzTwWWXTp09Pfc7OzsbatWsj90NpoGjRonHyO+22225mOX9/AOA3Cf/880+jFyxYYPR/QyMNDRs2NJqPDZ9jfJ4CQIUKFYwuVaqU0UuWLDF6p512cm2ULGlLak2dOtXounXrGr1y5UrXBp/f06ZNM5rPB/6dAFCiRAmj+fvvuuuuRs+ZM8e1kbwe5s2bh2XLlhXIOQT486h8+fJmeZUqVULrGJ2dnW00X/MVK1Z0bSxcuNBoPsZ8rvJ+Bvx5UqdOHfczuW0X4I8xw33V8uXL82yXt3XHHXfM83f+9ddfRvO1x/0Xt8k/M2/ePCxdurRAzqPixYvHye9cu3Zts3zNmjVuHb4e+RyZOXOm0aF7UenSpY3m/oz7nlB/xvcrPv952/l+BwAbNmwwmo8v35t+//131wb/Hu4j+foIbQdfh/zdeH+E2uBravXq1UviOHadQL4GPMWKFTMdMu/kHj16uHUmTZpkNB/s8ePHG33MMce4NgYOHGg0X7zHH3+80eeff75rY+LEiUaPHTvW6KpVqxp9882+9FH37t2NrlGjhtGvvfaa0aNGjXJt9OnTx+h69eoZzYMmvlEDwIUXXmh02bJlc22DtwsATjrppNTnMWPGuOXpIisrC02aNEnpr7/+2iwPdcxr1641+oMPPjC6Z8+e7ncwr776qtHfffed0QMGDDCaBzMAcMoppxi97777Gv3kk08afdVVV7k2mjVrZnTLli1zbeOLL75wbVxwwQVGH3LIIUbzdcm/E/ADmieeeMLoTz/91OhrrvG1dG+44YbU51NPPdUtTydZWVlmsHzyySeb5V27+goMfHP6+eefjeb+68wzz3RtPPDAA0ZfcsklRvO52rRpU9cG92e873nQdNNNN7k2+MbCcF81fPhw9zP8ffmPjwMOOMBo7mcA4NtvvzWar7369esbvd9++7k2kn0cX2PppESJEmZ7HnvsMbP8t99+c+vwddGuXTujO3ToYDT31QBw0EEHGf3ee7YeLA8KeeAB+IEEn/+87Xy/A4CcHFuWjX8vD8QffNBXXOLfc8YZZxjN1yHf7wH/R1/16tWN5nPqxx9/dG0MGjTI6O+//362+yFoSksIIYQQ/wLy9YRn/fr15hHmiy++aJb369fPrcNTOPw4iv9SCU0ltGnTxmj+a/3qq682+rrrrnNtvPOOTfbfY489jL711luNPv30010b/JfYvffeazT/RfT000+7Nvix4KJFi4zmqbfQaJanfXgd/q733XefayP5V8XBBx/slqeLIkWKmCcQ/AQr9NfMWWedZfSBBx5oND/x4b+yAf8UkI/Nl19+afRee+3l2thnH1tOjf9qatvWViD5/vvvXRs1a9Y0ev/99zf6/vvvN5r/QgaAuXPnGn3EEUcYfdpppxkdeqzOfzV16tTJaD5nLrroItfGPffck/rMUz3ppk6dOuYY8r4ObQ9Py/F35Ovq119/dW189tlnRvOx4HUeffRR18bzzz9v9Oeff250pUqVjP7jjz9cG0OGDDF68uTJRvM1wk8vAH8u7rLLLkbzdAUvB/zTiaOPPtroFi1aGM1P0gFg6dKlqc+hKe10UblyZfO0lPv3E044wa3DT0Zfeuklo/lJaGhKk5+s8X7lPjE0pfnVV18Zzf0mt8F9JABceumlRvPT9oceesjo0BQvP+Vu1KiR0bztnTt3dm2cd955RvN34esj9OT1xhtvNHpzTwr1hEcIIYQQGY8GPEIIIYTIeDTgEUIIIUTGky8PT+nSpY1Phd824PlswL+Fwv6CCRMmGH3bbbeBYec3eyl4LpId24B/c4m3lb0UM2bMcG3wG2d777230fxKYmi+ctasWUbzGxk8h81zk6F233zzTaPfessWVOb9A1ivVOi1x3TRsGFDs338ejjPCQPeb/XKK68YzW92tW7d2rXBnh32OHz44YdGDx061LXB/gN+Y4PfNuDXKwF//NkncNdddxnNb18A/u0KPpf5vOTzAwDuuOMOo/m78FtqoWiDww8/PPWZ59nTTXZ2tvGtsA8t9FZK8u1AwL+Gyz6mkSNHujZ4X/JbpfyKcegV2jfeeMPo5557zmj2ffCxAoBzzjnHaPZY8Rs2F198sWuD9we/YbQlb4vxW6bsi2JvHHvUAPumW8jDmS5WrlyJESNGpDTvI36TE/DHju+BvI/4jUnAe7jYO8RvMrE/C/B+Gj7P+Hc88sgjrg2+17BvsUyZMkaH3qDu3bu30R9//LHRSZ8fADRu3Ni18eyzzxp97bXXGs3nWCjuhX2+m0NPeIQQQgiR8WjAI4QQQoiMRwMeIYQQQmQ8GvAIIYQQIuPJby0tE2LGsem9evVy63CZB47EZrNoKOyJzYVvv/220e3btzf6lltucW1wBDob1D755BOjOWYd8KZlNotyLbGQ8Tlp9AR8WYBhw4YZHYo3Z+MXh4pxqFyoTEay7AWbbdPJ/PnzzfZwaCQbbgG/zzh07eWXXzY6VFqCA8G4HAWbJbnODODPETaKcujalVde6dpgIyQbBzlkMVTDio3/zODBg3NdDnhzIZ9nbM4MlY5Imri5TlW6WbFihek72ITOfQTgXyoYN26c0WwWDZl0OViNQ065j+DwNsAfY+4nuFwB/04AuPvuu43efffdjeYARL5mAN9fcX/GJtVQeCEH2nENMzbXhsJpDz300NRnrhOYTkqWLGnKrnCJnR9++MGtw6WBrrjiCqOTJXsA4P3333dtPP7440ZzyY5kECMQfqmEQ2/5Z7hMBJdACa3D2/7TTz8ZHQqw5Rdk+AUjDgnkl4cA3/f279/faL6fP/zww66N0H4OoSc8QgghhMh4NOARQgghRMajAY8QQgghMp58TbxXrFjRzD9z4bjLL7/crcNz6+xZYQ9Px44dXRtHHXWU0Rx29Mwzzxgd8vBcddVVRvP8Jc/FhwKSeA6bvRLs6XjyySddG998843RNWrUMJrnzXluHgCqVatm9EEHHWQ0BzOGPB8///xz6nNB+i9WrFhhQu6aN29ulnMxWcDO8QPeB/PLL78YHQra44KMfM6wR+U///mPa4P9ZX379jX6tddeMzpZmHATHLzGBftCvp90ELpWkyS9DUA4/CzZxvz587fNhm0h3Bddf/31ZjkHAAL+vDjyyCON5gKzXKQYAL744gujOQCQt6Nu3bquDS4oyr49PidCfSL3G8cee6zRyesb8AWGAR/gWbp0aaO5qCUH5AEbfTBJuKjlnnvuaTQXkwYKNmwwSZkyZXDYYYelNPs82dMF+GLPfG+68MILjeZjC3jfHntnOEg1VISY71/spUoW+QZ8ECXg+4C8glNDXjL2aHHoL4dZst8W8MGyXISZ74mhewRfl5tDT3iEEEIIkfFowCOEEEKIjEcDHiGEEEJkPBEXiMuNrKysuGbNminNc7yhtjjvhAvl8Rx4qNAlF9Nbs2aN0Tz3zkXxAF/E7cUXXzSa5wVDxRC5eCTP8bMfacCAAa4N9uRwkUbeP7///rtrgzM1uJgkb2dofjPpFTn66KMxbty4Aqkg2qxZszhZmHPixIlm+cKFC906XBgyr3lhPh8A70HhnAouvhcq/NmtWzejudBpuXLljObzFvB+BS6cx/P324JQFgxfq5zTwZkboVyipOdh1KhR+PPPPwusCm2tWrXiZJFBLn7KWR6Az0xJFkIGfPHEkFeAz4Hs7Gyj+bzinCnA+144yyS07QwXlQ15RfKCfUCcIcQeTc5QC63DmVnsm+LrHbBeoubNm2P06NEFch6VLVs2TnoIuSgl+/wA7+PjewD7czj7CQA+++wzoz/66COjd911V6NbtWrl2mDPFmcG8fXA3iLAe3h69OhhNPej7JUM/VvIo5Oka9eu7t/YL8seN96uc88917XB+2jkyJFj4jhuzj+nJzxCCCGEyHg04BFCCCFExqMBjxBCCCEynnwFsJQpU8bMlXF2yXHHHefWeeCBB4zmGjBcV+Spp55ybZx++ulG8/wl17y55557XBs//vij0eyV4bo606ZNc2088sgjRrP/aPr06UaH8jO4XgtnuYTqlTCc78I5BZyvMWXKFNdGch40NM+cLtauXWtq9vB8baju14oVK4zmWipr1641mjMpAO/f4MyN/fff3+g777zTtcF1Yp5//nmj+dieeOKJrg3OJdkWnh3OZeJaTSE/EvsVuH4PZzuFjkuDBg1Snwu6llaxYsWMp4T7FfbGAf4c4Hpb7Nn58ssvXRtlypQxmjNS2F+4pfkg+WVrPDtM9erVjWY/Dn/XkBeM+9W8PJnc3wFA0tOXDg/b5mjUqJHxurDfiL1VgPdLcm2tZK4PEPas8D2Oc9PY98r+U8Bfj1wL79133zU6VF+wU6dORvP9jfOgQllbobyr3OC+GvD+m8WLFxvNnraQT4j7fM6U2oSe8AghhBAi49GARwghhBAZjwY8QgghhMh4NOARQgghRMaTL6dh1apVTfDWunXrzHI2SwLe+MRFzzgwiU1hoTbYZMuGyz/++MO1wSa/jz/+2Oh33nnHaC7yCHijM5unuXgohxsC3ggXKjCZF1ywrkuXLkbz/goF8bVu3Tr1mYP80kkURShevHhKs+GaQ7kAf6zYhFulShWjQ+FmbC5lUzKbp7l4JuDDCzlALhTMxrBZemtgEyMb/0eNGmX0+eef79rgwLTXX3/daD6XQ8Unk8GayfOpIJg/fz5uu+22lObrio2foX8bPXq00RzExtc74Iv/skl5e4L7XjbTrly50uhevXq5NtgsygGW/NIKG3QBYPjw4anPBWlazs7ONi+n8AsF/FIC4L8fF8PmfcRtAv5aWbBggdF83wi9dMLmaA7v4yK2XKAT8EZmvjdziCIXbQbCQbG5EXopacaMGUZzkW5+IeKyyy5zbYTM0CH0hEcIIYQQGY8GPEIIIYTIeDTgEUIIIUTGk6/ioeXKlYuTgWQc9tWoUSO3DocVcSBQpUqVjA4FdXFxNZ43//nnn3P9+RAcKsY+EPY4AN6zw94SnosMeWd4TrdevXp5bivDAWA8F8sFR0MeneRc6rPPPosFCxYUSMG+Bg0axMlgSN7vXEwTACpXrmw0F33ceeedjU4GmW2CCyHy8f/rr79ybRPw3iAOK6xYsaLR7C3bGnr37u3+rU2bNkaPGDHC6FmzZhkdmr/nwqdcOJO/G1/rgA0v7NGjB2bMmFFgxUObNWsWJ31HXKRzzpw5bh0OgkwGYAK+4GJov/F5ESowub1Su3Zto7m4KPvtAB8Sx8VDk+GUAPD000+7NpJ97ZIlS7BmzZoCOY9q1qwZJ8Mo2aMS8uTxecbnFHt8ksW2N8H989577200+2Lq1q3r2mCfD5+77M/ab7/9XBt8P+P7xBFHHGF06JrKL6H73YEHHmg0exC5YOuiRYtcGxw23KBBAxUPFUIIIcS/Ew14hBBCCJHxaMAjhBBCiIwnXzk8tWrVMoUbeU6PiycCPkOAsw2efPJJoznrBPAZIDyPysVFt4SxY8cazcXXeF4V8N4InifljAn2mgBhb0hucEE3wOcdde7c2egWLVoY/dNPP7k2pk6dmvqcnZ2dr236J8ydOxfdu3dPac6QCRV95H328ssvG83nTCjrI/l9AT9fffvttxs9e/Zs10Yy9yXEtvDs8PUQyhSqWrVqrr+3RIkSRu+6666ujZdeeslozqHi4xDyM7zxxhupz1zgNd0sXLgQ9957b0pzDk/If8O5Ow0bNjSaix+zJwvw/cL/ioenVq1aRm/YsMFozosBgBdeeMFoLh6azFwDvEcR8J4czrviAsp83vHvDWW9pIsyZcqYTJwrr7zSLGfvCOAzc7h46uDBg41mHwzgc9TYg1mnTh2jOe8OAD7//HOjuT+74447cv15YKNfKgnfa7aFZ4c9iKGcPfaGcYHZa665xmjebsD2RbmhJzxCCCGEyHg04BFCCCFExqMBjxBCCCEynnx5eH755RcceeSRKX3xxReb5aFMn5YtWxq92267Gc35ASHYs/PKK68YPW7cuDzb4Pf/OduA8xNC84T8bzw/yXVl2FuzNYRqFHHuEOf98Nwr10niNgrSw9OoUSO8+eabKc2+gJtvvtmtw3kgnP3BfqzQPuvfv7/R7C3gDBKeV08X7GHiY9ekSRO3DnsrOMeC5+u5Fg3g/Ug8T877MFSbKpmZxF6GdFOtWjXjBePaZ1zDCfB+KPbtcRtcrwjYutp3BQH73Nifs3r1arcO1wXj64jrWoXyvDjzijPC2Of41VdfuTaSHp6C7IuKFCmCsmXLpjR7IznbBfA19zhXjv1yIQ/evHnzjOasOj4OyW3cRP369Y3m+yxnc1177bWuDe4XhgwZ4n4mL7iuIde04n3KeWeA76/4/sbeu1Cdwy31xuoJjxBCCCEyHg14hBBCCJHxaMAjhBBCiIxHAx4hhBBCZDz5Kh5au3btOGl+YgNSKKiIixRWqFDBaA4mfOKJJ1wbXGCSTV1bApsquTgoh7eFgro48I0LLDZu3NhoNqdtCRwgFjLPcvgZB2SxyTcUXtijR4/U544dO2LKlCkFUrCvfPnycdIQW6RIEbM8FDx46623Gs0GTC5suX79etfGOeecYzSH8/35559Gv/vuu64NLtiXF3yuAz5UjAP7OHiNzcSAP5c5qKtkyZJGT5s2zbXBhl0OyWSzJYd/AkCXLl1Sn4899liMHz++wIqHlilTJk72LaeeeqpZHupH2FD59ddfc5tG33DDDa4NPtdC5+v/AlzIuWvXru5nOKyRXyBho3ooZJGLzlavXt1o3sf8OwEb8HjIIYdg7NixBXIeFStWLE4WJh42bJhZ/txzz7l12LTOfS+/ZMBtAt78zdcrt8F9FeD7p+TLRACwyy67GB0K9C1fvrzRyUKqW8opp5xi9BlnnGE0jxFCBcbZcMwGay64zeGWANC2bVujr7nmGhUPFUIIIcS/Ew14hBBCCJHxaMAjhBBCiIwnX8GDy5cvx/Dhw1Oai3ZyKBMAfPvtt0Y3b26n1ThAaccdd3Rt5Nezw/N5gJ+/79Wrl9ETJkwwOlQ88pZbbjGai81xCNPWwAGIHNwFAO3atTOa59HZf8HBVoANyArNEaeLnJwcEyzF4ZWhoo/shUkG3gEbC5ImSRa43cRJJ51kNIdXcuHIbUHofPjwww+N5mJ6HGQWCvT75JNPjOb9w/v0vvvuc21wEV8+784//3yjQ0GcSf/C9OnT3fJ0E0X/b/XgfcJeG8Af89dee81oDjwLhWByP8GhaNy/FRb8XTi8FQC+/PJLo88++2yj+/XrZ/SiRYtcG3ztcfAeh9uFiswmr4ltUYB3S8nKykKDBg1S+v333zfLuVAo4K8t9peuWbPGaPaoAv4+wuGrfE4l/ZabYN8PbxcHh4b2e378u5uDQwC5X2nTpo3RoSBV9kU1bdrUaB5nhO6Jv//+e94bCz3hEUIIIcS/AA14hBBCCJHxaMAjhBBCiIwnXx6eXXbZxcy3cu5OqNjaqlWrjGbvDM8B//LLL66NqVOn5rpd7PvhuXnAz/FffvnlRn/00UdGh+YEeW7xiCOOMHrWrFm5bmcIzgvhrB+eEwZsbgXg834OP/xwo0eNGuXaSM7phnJr0kX9+vXx8ssvpzRnn4SKHPL89E033WQ0f1/2JgA+z4bPO25jS4ra5sXJJ5/s/o3n/LloKR8LzpcBwnkoSb744gujW7Vq5X6Gi/px7hTPk4d8dAcddFDqc8jzlk6ysrKM/4/P8VARynXr1hnNOSTJorYAkMxo2QQX5eTz6phjjjGafSEh2I+xww7271D2HwLhfiG/8DHnvuizzz4zmn1egP++Rx99tNHsQXv66addG8kipiGvWLqoXr268dOwl4YLoQK+kOdPP/1kNOfOJQvcboJ9XnfddZfR7Gu98MILXRu8nzgziL2BW1JgOy9CfcD1119vNHvFOHeJ93GoXS6OzL459hcC3iu8OfSERwghhBAZjwY8QgghhMh4NOARQgghRMaTLw9Pdna2yRWoVKmSWX7mmWe6dXhOs0+fPkazH4UzVrYE9v2Eam3wPDjXDeGMiZAP6P777zeat31rYB/Qxx9/bPRtt92W7zb5u77++uvuZ5YvX576HMqbSBcbNmwwvi72VYRqz/A8OR8bnvN+8sknXRtPPfWU0ewt2xaeHfaEcF00ACb3A/BZGOznCOUDcbvsK2HPDte3Abwfg6/LpK8i9DsBm4+xLTwC+WHDhg3Gh8bz/rvttptbh+ulcX2hQw891Oh33nnHtcHHmGsDDRgwwOiQj6thw4ZGP/DAA0a3b9/e6JAfiX+mf//+Rm9NLtJ+++1nNF+bIS/lwIEDjeb+7PHHHzea6yQB9trkGm/ppEiRIihXrlxK832E/UmA9a0B/vpk31PouuD6a5MmTTKaPamhWoi8n/hnatas6dbJLy1btjSa+wgAqFevntF8brMH8brrrnNtsI+TPansvw3VyeN7Md+rN6EnPEIIIYTIeDTgEUIIIUTGowGPEEIIITIeDXiEEEIIkfFE+SkgVqNGjbhr164pzQXqQuYqNphy8BAHOd14441bvD2b2GmnnYzu2bOn+5lk0VPAmxrZXMaF8wBvBk0afwsTNqVedtllRocCo5IGtEWLFmHNmjWR+6E0kJWVFVerVi2lq1evbpafcsopbp3OnTsbvXDhQqPZKMjGSQB47LHHjN5///2N7tixYy5b/b8F76MaNWoYzaGaoaAuDrhkEzMHm3HBQ8CaWJs3b47Ro0cXyDkEAPXq1YuThn42trPpGvD9AodPcnBo6CWMv//+22gOfGNjZ6dOnVwbHBo4c+ZMo5NGWiD8cgT329wXcfjkllCnTh2jOcwxZFp+4403jOY+kr/roEGDXBvJMNrOnTtjypQpBXIelS9fPj744INT+sEHHzTLQy9zcChkVlaW0fzyQ8i0zEZePg85rDFkWudjxS9qbM29ifsNvheHDOX8kgm//DNy5Eijk/t7E9yf8YsARYoUMfq7775zbXCB5HHjxo2J49hV8tUTHiGEEEJkPBrwCCGEECLj0YBHCCGEEBlPvoIHd955Z1O48Y8//jDLTzrpJLdOiRIljGZ/RahAX37heUMuNhkir6J+u++++z/apm1FyNPUunVro2+++eZc2wiFUCXnmkMFKtNF06ZNTXHPZCFRABgyZIhbhz0M7IsoXbq00Q8//LBrg70E25Nnh+Eil3yusrcuFKL5+eefG83+O762OdwQsIFgyRDAgiArK8uEnB122GFmOReYBYBHH33UaPYOsM/v7LPPdm1wGCEXT9xnn32MDhXc5AKyr7zyitEnnnii0RwICPiA0gsuuMBo9viwVzIEF4AdMWKE0aFwzi5duhjNAZX8XQYPHuzaSJ47+fGU/lMqVaqEDh06pDT7UTgAEACKFy9u9NChQ42+8sorjQ55ybiNDRs2GM3hoyE/1rJly4zeGs8OF9zk0Mxbb73VaPYrAT6Ik/1ofH2EAjHzCkrlUNFQgXE+7zYXhKonPEIIIYTIeDTgEUIIIUTGowGPEEIIITKefHl45s6dixtuuCGleT6OMyoA4M477zSaC3/x3GMoy4fzATKJhx56yOjGjRsbHfIiPP/880bzPDH7fkKelmTRO84fSScTJ05EkyZNUrp27dpmeShDiXNmki6Mu7EAACAASURBVB4gAPj111+NZt8Y4H0vmQQXeeRrjH0FAHDXXXcZ3bt3b6OTxwjwc/UA8OKLL6Y+c2HOdFO0aFHj/+M8Jy5iCPgijd27dzea/WScjwL485MzQdj3M3r0aNcG+xgnTpxoNHs8Qvue/ZKcQ8TZWyeccIJr4+2333b/luTcc881OlSUd86cOUavXLnS6LvvvjvX7QKA+fPnb3b9dLJs2TLjweHCoDk5OW4dvl9xMUz2Y4W8Nfxvhx9+uNFctLRYsWKuDfajbQ3vvfee0XzvYW9RyI/EvrdnnnnGaPb2hY7vhRdeaDQXWOactRCc1RYq/AvoCY8QQggh/gVowCOEEEKIjEcDHiGEEEJkPPny8JQpU8ZkwHA+ypgxY9w6PA/Mc4/sxwj5LyZMmGB0soYO4LMQuLYW4LN6ypQpY3SpUqWM/v33310b+WWXXXZx/8Zz7zy3XrSoPSShnKJLL73UaM5LYH9GyAeUzJw46qij3PJ0Ua9ePTz++OMpzbWHDj30ULcOZ8BMmTLF6D333NNorhMFeL8Z7+etIYpsyZ8tyRDhGm6h+kT5hfcH+0xCtXjat29vNM+B77333kaHMliS+zTkd0gn8+fPx+23357SXKOH61EB3tvEOTPff/+90ffcc0+e2/Hhhx8azf0I93+ArxWU9EUCwFNPPWU0ZwwB3tfIXknOLjn55JNdG3l5eJjQd+EsIz5vrr76aqPZbwfYa4B9I+mkVKlSaN78/8stcQ5R6H7GniQ+77jW1sCBA10bnLOT9MIBwP3335/LVm87eF+fdtppRvfo0cPo008/3bXB3rFXX33VaK7jx95AwPebnMPTrFkzoz/44APXBudSbQ494RFCCCFExqMBjxBCCCEyHg14hBBCCJHxaMAjhBBCiIwnX87NUqVKmeJ4bDjdeeed3TpsYlu8eLHRV1xxhdEtWrRwbXAxRA4vuuOOO4w+4IADXBtsWOTCcHXq1DGajb8hODCKt50Ne4Av/HnVVVcZ3b9/f6MXLFjg2uCAtIsuushoNo6Fis+NHDky9XnVqlVuebpYt24dlixZktJcXJHNloA303EA1vr1640uUqSIa4PNkhzmtjXF97am0OG2MCnnBRf+DBl4x48fbzSHnXFYZajA7IEHHpj6/OOPP+Z7O/8JGzZsMEGnXAgxVPiRgwSThSMBb5ivW7eua6NevXq5bhcHXCZD9TZx2WWXGd2rVy+jOeDvhx9+cG20bdvWaA7w5HMgVIQ4v/ALJgCwdOlSo/mFgfvuu89oNtgDtn/fmutwa1m/fr0pwskG7FC/yIWa+VhysczJkye7NjhoL9RfbWtCxZLfffddo7mf+O2334wOFQfngE/WHGrL1xzgTfl8rvL9bu3ata6NpPk8N/SERwghhBAZjwY8QgghhMh4NOARQgghRMaTLw8P+y/Kli1rlp955pluHQ68Yn/Jlng4OLyI50kvuOACo0OFA/PyJHC4XYizzjrLaJ7zP/LII40OzQHzXGOVKlWM3muvvYx+6623XBscrFitWjWjjz32WKN5bhoAFi5cmPocKk6XLlasWGFCojgUkYOsAB/mdt555xnNgXG8TwGYkDrAz61zIT0+5/6XueWWW4zmEMCQF4WL1vK8+QMPPGD0008/7dp4/fXXU5+TXoiCYIcddkBWVlZKc0haMlhzExziud9++xndqlUro3kfbA3jxo1z/xbyMSThY8HHFwAefPBBo9mTNWPGjC3dxM3CwamhIEb2S/7yyy9Gc4heyI+U9NNx6Fw6WbNmjfGYsBcw6VHbBPsy2TvWtGlTozlUENiye80/hT2pocBHDkGtVKmS0dzP8v0e8PciDgU+/vjjjQ4VwuX9wb+Hgzj79u3r2vjkk0/cv4XQEx4hhBBCZDwa8AghhBAi49GARwghhBAZT748PLNmzUKXLl1SmudbBw8e7NZhPwnPv3EhNc45AIBnn33WaM4L4CwfzoYAfIZEKJskL/j38Lw5F2R85plnXBu8zh577GE0FyOsWbOmayOZoQP4+Xqe8xw0aJBr4+uvv059njdvnlueLkqXLo2WLVumNOdDcLYR4LNPihcvbjTnRYQKJfLcMs89c07T/wqhbCvOaipdurTR/N1C3gs+Z/7880+jOduJr2PAFgIuaM/T+vXrsWLFipR+7LHHzPLksk2wr4E9h+zhSeb8bOK5554zmvsz9uyECrfmxZYUMR07dmyuelvAfpxQnhcXHmb92WefGR3yWiQ9TVuTbbW11KtXz3hsuB957bXX3Drs++KfSfraAH9tbiu6du1qNHu6uIgpZz0B/t7C28p9Yv369V0bSS8o4I8fF/7m3CIAqFChgtGff/55rutwthMA9OvXz2gu8roJPeERQgghRMajAY8QQgghMh4NeIQQQgiR8eTLw7PDDjsYzw1nMHTq1Mmtw3VhOA+F83BOOOEE10bjxo2N3nPPPY2eM2eO0TwHDvgaXsksmBCh2jOcocO+h/3339/ogQMHujZ+//13o2vVqmU01wkLZQpxPgJTsWJFo5O+q00k5+MLct58zZo15nhxVhHnqQDef8PHks8H9hEAeddr4bwj/p2A93Q8//zz7mf+KXxsQ7lUXH+N82I45yXkveB6Pjxfz16UUJ5I8hoqSB8YsDETjOtFJWE/BuA9KexJ4DwrruED+FpP7JXgY8N5R1vDqFGj/nEbWwNnyoRy1thzydcN94khkuc41ztLJzk5OZg5c2ZKT5w40SwPbTv7qa699lqjN+cdScL5Ppzlw3B9MgD44IMPjG7Tpo3RI0aMMDp0X33//feNnjZtmtF33nmn0aFMKfbT8L2Hr9Hhw4e7Nnif8XXKWT6hml7sa9wcesIjhBBCiIxHAx4hhBBCZDwa8AghhBAi49GARwghhBAZT5Qfw+rOO+8cJ02VbGLjYnMA8MgjjxjNIYIcrMZBfIAvBtmuXTujk0UEgXAhQza25gUb9gAf1HTNNdcY3aBBA6PZBAcABx10kNFs9vzrr7+MDh0fDmt86aWXjGZjZYsWLVwbSSN4586dMXXq1Mj9UBpo3LhxnAxv4/BKNt8B3mDOhfCaNGlidCgwjMMLTz/9dKPZsBoqcsdmYC7Seeuttxq94447ujbOPfdco/ncZaM7/zzgw744WJIDQEPBk0WKFMm1TT4vBwwY4NpIhmTec889mD17doGcQ4A/j7hIJRcLBnxxXzZtspk0VMiWAxZ5v7GJ+ZtvvnFt/K+SDCMFvLGfDdmAD3TlYEH+/tx3AfbYPfzww5g7d26BnEfVqlWLk9c0h/Mdcsghbh0OyeOXSvh+ty0ItckvWbA5/phjjjE6VPiT79983+R+NxQKyyZuLlTdp08fo0MvenA/yX0gFxcNvawwZswYo++7774xcRw355/TEx4hhBBCZDwa8AghhBAi49GARwghhBAZT76CB8uUKWPmNTkAiOfvAODSSy81mr0D7733ntE8Rwr4op0cTLRkyRKjOZRpa7jgggvcvx166KFGV65c2WgOanr33XddG5MmTTKavUKNGjUyOlQ8lNfp3bt3rr8jNPeenDctyNC4okWLolKlSinN5wwX5wP8nO3FF19sdLKIJRAOSOOij7NmzTKaA8Q47BIAbrnlFqM54JLXCQWGrVmzxug33njDaPZj8bEFgOnTpxvN+yMZpgZ4nwXg/SrFihUzmosR8u8ANhbw3ETIN5dO1qxZY8IFuWBqKASVQ/E4SJJD07jAKgAcffTRRnOwKh8vPs8Af02zr+GVV15x6/xTQkVo+ffydcb9e6iwM5+/yXMC8F65EFy0taCoUaOG8e2w9/HEE09063DRTi4Gzfcv9vUBwA8//GA09xPsC5s8ebJrg4uU8r2mRIkSRrdu3dq1wdc8By3yuc5Bu4C/n5999tlG9+zZ0+jq1au7NtjXyp4d9lOyHxPwRWo3h57wCCGEECLj0YBHCCGEEBmPBjxCCCGEyHjy5eEpVaqU8VhwTsFuu+3m1mE/Ac/x8jxi3bp1XRucf1KnTh2j9957b6NXrFjh2uD5aC5i+s477xjN7/UDwPz5843mrAPOS+C5SQB4+umnjeasA/5u5cqVc22wJ+Occ84x+vLLLzc6NH+bzPuZMmWKW54uli5danwq++67r1n+1FNPuXU4Y4I9XFwYtHjx4q6NDRs2GM1zz5xVE8phYg8X+2A4g4XzowCfEcTnFBc+5dwqwBfbY98Ae3zKlCnj2mB/AhcO5GKiIV9Jcn4+lDmVTrKzs423gYvyTp061a3Dfgv2JLDv5/HHH3dtcG4UexC5sC0XAwZ8P8nFfdnDcckll7g2uPApn0dt27Y1eq+99nJtcN/LGUOcMxbKlTr55JON5uuMs3xC13fSi8Uet3Sydu1a41/k7xe69p555hmj2ZOSVyFfwHvHuN/gfcb9G+Dvq/fee6/R3FeF8o/Y28le2bPOOstoPqcAn4HGfiO+t9x0002uDb4e2rdvbzRfYy1btnRt5FWAdRN6wiOEEEKIjEcDHiGEEEJkPBrwCCGEECLjyZeHJycnx2RT8Hv6PJ8L+DlPrsfB89PXXXedayM5xwt4nw/P+4Y8C9wGa87u+eqrr1wbnNXCc9zfffed0ePGjXNt8JxmsjZZaDnX/wF8baxjjz3W6Pvuu89o9jcAtu7Z2rVr3fJ0sXz5clMHZsiQIWY5ewAAn1PBmROcdRI6dnPnzjWaPScjRowwmvOQAODCCy80mr1i7In4z3/+49o47rjjjOacDp5HZ68G4Out8fHmOjtvv/22a4N9A+wt4zn/V1991bWR3Eec+5FuSpcubfJbfvzxR7N8n332cetcffXVRnPNMfZCXXTRRa6N1atX59om5xWFMsHYk8bnc3Z2ttGffvqpa4P72r59+xrNnjT2GwLes8LHmL1iIX8l19/ifBj2o4T8Fz169Eh93lIvxrZg7ty5uOGGG1K6W7duZvnw4cPdOnxtcf/N/Qy3CQBjx441mv1zXLeOfTKA92RxP8H+wtCx45qUfD589NFHRof8aLw/2NfLmVOhXCb+Luyt4745lI8VuteG0BMeIYQQQmQ8GvAIIYQQIuPRgEcIIYQQGY8GPEIIIYTIePJlWl63bp0p1Hn//feb5T/99JNbhwsssgmXCz2yaRPwRRlr1aplNBcu5OKZgC9qFsex0RwoFgqd4tBE1nfddZfRITMwh04lTXOAD4Pi0CUA+Pzzz41mgyYHDYaKz4WMkAVBrVq1TDhXx44dzXI2AQK+6COHRnbv3t1oLgIIeJMyBzxy0VI+LoA32B988MFG9+vXz2g29QPAsGHDjOYAwAYNGhgdKoLJgWBscuTzkgMxAW9y5WuKC/KGDIsjR45MfQ4VFkwn8+fPN8GHL7/8slkeKtzLhngOwWPjI/cJgH9BIGm4Bfx+DYXG9enTx2gunslmUg6RA3ygHRdpZmMnF88FfDFQNiCz4TYULskGev4ZNmizERaw5yv3y+lk9erV5p7FRStDRXc5fC8nJ8dofglhjz32cG3wNc3maC5Ay0ZowB8bDjnl7xIyLfOLO2x0v/76641mczXgXxDiFyS4fw8F6XKhbja6T5gwwWgOb80PesIjhBBCiIxHAx4hhBBCZDwa8AghhBAi44nyM2caRdFiALPTtzmikKgTx7GvdJkGdA5lLAV2DgE6jzIY9UViWxA8j/I14BFCCCGE2B7RlJYQQgghMh4NeIQQQgiR8WjAI4QQQoiMRwMeIYQQQmQ8GvAIIYQQIuPRgEcIIYQQGY8GPEIIIYTIeDTgIaIoqhtF0XtRFC2LomhhFEWPRlGUryKr4t9NFEWr6L/1URT5Cp5C5EIURY2jKPo0iqI/oyiaHkXRKYW9TWL7Ioqiy6IoGh1FUU4URYMKe3sKGw14PI8BWARgZwB7AWgN4JJc1xAiQRzHpTf9B6AagL8BbH2JX/Gv479/ZA0H8A6AigAuBDA4iqJGhbphYntjPoC7ADxT2Bvyv4AGPJ56AF6N4zg7juOFAD4A0LSQt0lsv/wHGwfQXxX2hojtit0AVAfwYBzH6+M4/hTAKAAdCnezxPZEHMdD4zgeBuCPwt6W/wU04PE8BKBdFEWloiiqAeAYbBz0CLE1nAvg+Vg1XMQ/JwLQrLA3QojtFQ14PF9i4xOdFQDmAhgNYFihbpHYLomiqA42Tok+V9jbIrY7pmHjk8HroigqFkXRUdh4LpUq3M0SYvtFA54EURTtgI1Pc4YC2BFAZQAVANxbmNsltls6ABgZx/HMwt4QsX0Rx/FaACcDOA7AQgDXAngVG/8IE0JsBRrwWCoCqA3g0TiOc+I4/gPAswCOLdzNEtspHaGnO2IrieN4fBzHreM4rhTHcVsA9QF8X9jbJcT2igY8CeI4XgJgJoCuURQVjaKoPDZ6MMYX7paJ7Y0oiloAqAG9nSW2kiiK9oiiqMR//YTdsPHN0UGFvFliO+K/97ESAIoAKPLf8+lfG7OiAY/nVABHA1gMYDqAtQCuLtQtEtsj5wIYGsfxysLeELHd0gHAAmz08rQBcGQcxzmFu0liO+NmbIzFuAHAOf/9fHOhblEhEunlESGEEEJkOnrCI4QQQoiMRwMeIYQQQmQ8GvAIIYQQIuPRgEcIIYQQGY8GPEIIIYTIePL1Pn758uXj6tWrp/T69evN8mLFirl1oigyevHixbmuU6JECddGkSJFjJ4xY4bRzZrZ8jIrVqxwbaxdu9bo8uXLGz1v3jyjd9jBjwV5OypXrmw0f9e///7btTF79uxc2+TtKlrUH6K//vrL6DVr1hi92267Gc3fHQDmz5+f+rxy5UpkZ2dH7ofSQPny5eNq1aptdvmcOXPcv9WpU8fo1atXG837sEyZMq4N3gd8HvI5w78TAP74w9bfq1SpUq7LK1as6NrgbVu50r61zudH6LsUL17caN6ffF3+8ssvro3kdQz486x06dJGL1myxLWRfMPzjz/+wKpVqwrkHAKA4sWLx8m+gvf18uXL3Tp83uy4445G16hRw+iff/7ZtVGlShWjly5dmmsb/DsA3y8kr0XAn1cVKlRwbUybNs3oDRs2GF2/fn2jFy1a5NpYtWqV0XwelSxZ0mg+7wDg999/N5r7TT6/y5Yt69pI9nnz58/H8uXLC+Q8KlmyZJzcnuzsbLM89H35WPB+zcmxqQGh+8iuu+5qNPfnv/32m9GhPoDXady4sdETJ07Mczt22WUXo/mcatSokdHcvwG+r+G3vnfeeWej161b59rgPp/b4O/P3x3w1/batWuXxHFchX8uXwOe6tWr44UXXkhp7lT4Ygf8gObxxx93bSbhkwHwg4AzzjjD6FGjRhn9wQe+1ufChQuNPumkk4y+6aabjA51VHwTOP/8843mC2TSpEmujQsvvNDocuXKGX388ccbXbVqVdfGt99+azQP1kaOHGk0d6gAcMcdd6Q+Dx061C1PF9WqVcMTTzyR0nwhXnPNNW6dgQMHGv3TTz8ZzcfqiCOOcG3wPhowYIDRn376qdH9+/d3bSTPfQDo0MEWrn7++eeNbt++vWvj8MMPN3rEiBFGX3TRRbn+PADUqlXL6BtuuMHoP//80+ijjz7atZE8/oC/yR500EFGP/PMM66N5E22V69ebnk6KVGiBJo3b57SfCzeeOMNt87YsWONPvDAA42+++67jW7Tpo1r4+KLLzb6lVdeMfquu+7K9XcAwLhx44zmY3Huueca/Z///Me10apVK6P5Zj1o0CCjH3nkEdfGV199ZfT1119v9F577WU099UA0LdvX6NLlbKlvrgvCl2byT6vY8eObnm6KFu2LM4+++yUnjJlilleu3Zttw7fex599FGjedAQGuB99tlnRnN/ftlllxndunVr18b339vA7R9++MFo/qOXjwsAvPvuu7n+Hl7+3HM+NJ77Gv7j+9ZbbzU6NGi66qqrjOY/Tnm7+LsD/p4wZ86c2e6HoCktIYQQQvwLyFfwYFZWlpnSGj58uFke+isi+VcY4J/g8GO10FMRHhEfcMABRn/99ddGH3LIIa4NfmT/5JNPGv3qq68azX9BAcCXX35pND/hGjx4sNFXXHGFa4P/oucR79SpU40+88wzXRv8JI33B//Vecstt7g2ko9An3vuOSxYsKBAHiNXrVo1bteuXUq/9957Zjk/4QL8kxI+p/gvgoceesi1ccopp/B2GN2iRQuj+VEs4B9n81+8nTp1Mpqf+AD+SRJPx/J3Cz3O3mOPPYy+5JJLjOap1ssvv9y1wY+8eTqWn3Dts88+ro399tsv9fnRRx/F3LlzC2xKq169evFtt92W0meddZZZzn9ZAkDyvAP81NI777xjdGjf89TZ559/bvQDDzxgdL9+/VwbfI5zP8LX81tvveXamD59utEnnnii0dxHvP32264Nngbr06eP0S+99JLR3GcC/lr79ddfje7Zs6fRoSf4t99+e+pzq1atMHbs2AI5j3baaac4+fTstNNOM8s/+eQTtw4/OeRpwIMPPtho7ncAP1PAT467detmdOipFz9x5X6DZzD4vgL485+fJvN3++6771wbPKvBfSLfM4888kjXBj+N5CfYfL2EZiR46nTcuHFj4jhuzj+nJzxCCCGEyHg04BFCCCFExqMBjxBCCCEynny9pdWwYUMMGTIkpdkrcMwxx7h1+JXh5HwtAPTo0cNonkcG/Fx6y5YtjeY3J3h+O7TOvvvua3TDhg2NZp8E4N8w49dW+fXfUBvshbj6aluInb9L6PVoduSzQ51fNw295ZF8e45fP00nRYsWNecNv0HEc/4A8NRTTxnNr9x27drV6CuvvNK1cd111xn9+uuvG530owDA7rvv7tr46KOPjH755ZeN5tcr2a8DAL179zb6xhtvNJrf0GFPCADzlhvg/RrsNQq9xsnnxL333mt00h8D+OsDsNvOb2ukm5IlSxr/E78tw292At4fyOcRv8nJbxgB3sfyxRdfGM3+m7p167o2eN+yz4P7RL6eAd/XnHzyyUazt4K/GwC8+OKLRjdp0sRo9v6F9im/Ictv6fBbWezXAIDzzjtvs+unk+rVq5u36vit2x9//NGt8/DDDxvN3lD25CXfAtsE3xPZK8Vv8d18sy9uzm9zsv+I32zi18cB4JxzzjGaz1X2Z4XiXg477DCj999/f6P5bbGjjjrKtcF+Se57k2MOIOxpmjlzptG8DzehJzxCCCGEyHg04BFCCCFExqMBjxBCCCEyHg14hBBCCJHx5Mu0HMexqYXB5RjYsAUA1157rdFsWmOjZ6g8BRuO2UyYV52s0O9p0KCB0d27dzc6VFqCo7XZYMrhbRzCBXjzLMdks4ExVHuE48v593AAZCjsKRm8FwpYSxfly5c3oVhsjOPjAPiANC6VwaFaoeBBrvnDJl0OgOR4d8AHxHH9Fg7ICh270aNHG81GaA4QC5WnYOM7B3XtueeeRnN0O+CN/Vzig0PIQsbnZMha6EWBdLJ06VJj9uRAuJCxkY8fnxN8PEMlatj8yeGSbGpmIzAATJgwwWguc8LLQ6Fx3CdecMEFRi9btszoUAgqh8JxaQ0uLRGqlcihrx9//LHRd955p9H8ggFgDfFZWVluebpYsGCBMS1z6RHedsAH5fI5xOfDrFmzXBt8rvL9il9s4EBTwPcLbGxns3CovAwbsPka55InXH8Q8CZz7je2pM4hl9/g85+NzqG+JlT2JISe8AghhBAi49GARwghhBAZjwY8QgghhMh48uXh+e2334y3gefNOHQL8IFXY8aMMTonJ8foKlWquDYmT55sNM+j85wfh/cBwA033GB0vXr1jOZibDzPCgCLFi0ymoPnuNgcF1cF/FwrB5dxwBvPiQJ+Xpi3ledvd9ppJ9dGMpiMvUfpZMGCBWZunD1LHKAG+DBKDsDi8Ma+ffu6Nth/dc899xjNAX8cmAb4gqMc3sfH9umnn3ZtDBw40Gj243DRztq1a7s2ODCPfWFfffWV0WeccYZr48ADDzSa/RlbEl754IMPpj6Hrpd0UrNmTXPM5s6da5aH/IRc+LNXr15GDxs2zOjs7GzXBntn2MfI50jIx/fuu+8azd4RbiMUxscFksePH280+/xCAXjsl+HixxxoygGggPessKeFC2WGiocm7xHsx0wn69atM9499lvNnj3brTNjxgyjeR9xEdtGjRq5NrhQNfcJfOxCvhf2aXLwJvejof0a8rkl4cK4oe/C3kf2dHF4ZcgLyGGFbdu2NZqDCRcsWODa2NLwXD3hEUIIIUTGowGPEEIIITIeDXiEEEIIkfHky8OzZs0ak2fDc3rsgwG874Xn1nnu//3333dtcCZKt27djGYPRyh3hItwskfht99+M5qLi4aYOHFirsuPPfZY92/VqlUzmueAOYMhVEyV/Qqc/8Lz+aeeeqprI5mPUJAZKjvvvLOZ12UfAc9nA8Dff/9tdLt27Yzmc4izMgB/vHmenIsgsi8G8D4X9qw9++yzRnPGEgAceuihuW4He8v4+gF8VhH7BuI4NrpVq1auDb4OuXjqNddcY3Toukx65/h3pps5c+aY656PVyj/ZujQoUZzFhdf8/zzgPdO3H///UZz38M+CMB7sE4//XSjuZjkmWee6dpg7xefJ8mMJAAm+2oTnM/FPg/2oIX8N126dDGaC9tylk3ouyS3I+SbSRdRFJnin+zJ7Ny5s1uHM764f+ZrnotlA754LBelzitTBwDeeustoy+//HKjV65caTSfc4C/5tm3+Nprrxl9wgknuDa43+SsK85yCvl82XPLvk7uq0IFlfO6B25CT3iEEEIIkfFowCOEEEKIjEcDHiGEEEJkPPny8FSuXBnnn39+StetW9csD9X94dpBPNfOOuS/4Hlx/hmeJ9xjjz1cGzznGapxkhflypUzmmvvbEmGBOcS8Zw31wlj/wrga0txLgFndITmPJNz+gU5bz5jxgwzj8+1h0LzU71TfwAAIABJREFUxG+++abRFSpUMDqZKQT4GmehNjiX57PPPjM6VL+Gjz9n2XDGBM+RA76mG+cs5VVXBvDZJu+9957RXL+Gc3kA71dhX9jMmTON5vMWsOduqM5OOqlQoYKZ62evCM/pAz5LiGtUcR/BfjrAX0uckcKZVyEPFl+/XOeI+1HO+gG89409O5wBxrW1AF+jiL1+TZs2NZozdwCfKfPCCy8YzXXdOMcGsFktt956q1ueLrKyssw97JtvvjHLOSMO8P0Ge0M5Y4kzwwBgyZIlRrdp08Zo9k6NGDEizzbYK8SZSXx9A77v6dmzp9HcF4fqXE6aNCnX38temg8//NC1wVljfM3w/b1FixauDR5nbA494RFCCCFExqMBjxBCCCEyHg14hBBCCJHxaMAjhBBCiIwnX6blChUqGGNbMrQJ8GFYAPDpp58aXatWLaN33313o7kIGuDNU2xi/vbbb41u3ry5a+OVV14x+qCDDnI/kxdsKL7xxhuNZgPuHXfc4drgkCkOnitVqpTRDz/8sGuDjd5vv/220VxclQv6ATYgK1TQLV3Url3bHOPy5cub5VwoDvAhWrzfuY1QkUMudMuhmUkz/ub4888/jQ4Z8JLw+bIlcDAbmyQB4LHHHjOaz5Eff/zR6F9++cW1weFeHHjIJuVQG0nTMhtv083ixYvNcebQNC4OC3gz94knnmg0B7ytW7fOtcEhjmwOZiNwKKyPw+k44G1LYIM1hxeyIZtfZACAu+66y2g2v3MgKb+kAnhDPL8cwJrN1oAt7Bwy+aaLVatWmWKXgwYNMstD4aMcTsjnDBvOuVg24I8V6xdffNHo0MsP/OICG+zZTB16MYWNztwHHnfccUZzvwIAl156qdH8ffn4h14m+Pjjj41mk/IVV1xhNIcEA/4Fks2hJzxCCCGEyHg04BFCCCFExqMBjxBCCCEynnx5eKZPn24C67ggGRcxBICbb77ZaA6z4qAmnpsEfMFFDsBiXwz7YAAf1LQ1/PDDD0bz/OUnn3xiNAfkbQn33Xef0exfAbzfgv1IvI9Dc69Jf1FoXj1d5OTkmDlYPv4dOnRw61x//fVG85z2McccY3TIk8Q+J56fXrNmjdENGzZ0bRREkVUuNhkq2sl+GQ4Mq1ixotE8nw8AQ4YMMTqv4MW+ffu6NsaOHZv6HAomTCc1atTA3XffndJcxJOvVQC45JJLjL733nuN5gC4kAePvV+hwo5JQn6DbQH7jxj+bqFiyPPmzTOa+81ff/3V6FAxTT73OICSgzO5MCoAVKlSJfWZfaHpJCsry1zn7J3hPgHwQXocmsf7ObTPOASQAzE5wPb11193bbBXjP1Y/fv3d+sw3Afy/ZuLcrO3BvBBkXz8evXqZTR7egAfisn3TQ6v5MLGgPd+DRgwwP0MoCc8QgghhPgXoAGPEEIIITIeDXiEEEIIkfHka8K0SZMmJleH35/v3r27W4czBrjYHs/xhuYrf/rpJ6N5XvCMM84wOjSvPnDgQPdv+YUzNYYNG2Y0Z1388ccfro3WrVsb/cUXXxjNc7ODBw92bXz33XdGN2nSxOgjjzzS6JCHJ+kVCfmE0sWCBQtMoceaNWua5bx/AODss882ukSJEkZfcMEFRnMeEmB9AoAvMLt8+XKj+TgAPmeKCyduC9ivxL4wwOaWAL7YHhdP3WEH/3cNey/4XG7Xrp3RXPASsBlbBXkOARu9Asm5fi4eevnll7t1+PrlPCfer+zX2RI4qybkSWncuLHRffr0MZqzqLp06eLaCPWTSQ444ACjQz4u9kZwvhlnG4V8Wuw/4cwU9rBkZ2dvZos3Eipymi6KFi1q+gXOkAn131z8tGTJkkZz8Vj2ZAK+T+f+motBc0FWwPdP7Os77bTTjA6dL6tWrTKafXx8DnFuD+A9ascee6zRhx12WK5tAn6fci4V+wdD2XR5nVeb0BMeIYQQQmQ8GvAIIYQQIuPRgEcIIYQQGU8Ux/EW/3CJEiXi5FwwewVCc808L8z1OjgPJJTDw3kA/Hv69etn9N9//+3aYPbff/88f4bhbA+uPcM5DiEvScijkmTSpElGcy4L4PMT2EvCc7GhOjrJNtq1a4dJkyZFuW7YNqJq1apxsl4UzzWHag9xJgzPefNc9EcffeTa4GwHzuphzwrXTAKA0aNHG83z0x988IHRofo1nJexNfA8+OGHH240Z520bdvWtcE5Q1zfifNUkpk3m0jmYb3xxhtYvHhxgZxDAFCzZs045NPZBPv+AF9zj700ocyjvGDPDl+vIe8g/16ujcaentC5yLCPjQnl9nCNQiaZswT46xDYmIeUhGu/cXZV6B6R9Jj9+uuv+PvvvwvkPKpfv36cPK/5+4Z8L1ynkb0z7MlkfyXg/WV8rfE+CtWo5PqI7J0K1d9i2F/E/RX3K7vssotrg3287Dfi+2zIa8P7jO+RvH9C5zpf7z179hwTx7ErqqknPEIIIYTIeDTgEUIIIUTGowGPEEIIITIeDXiEEEIIkfHkK3iwQYMGxmT84YcfmuUPPvigW4dNXhxExYa8UNE/Nkez8ZONg4MGDXJtDB061GguesZBbCE4fK1WrVpGsxkxZFpmOPyMgxlD28VFS/MiFNSUDBXjoKt0UrVqVVP8jQs6cqggAFx33XVGn3feeUazEZ7NwwDw1VdfGc1GQA43XL16tWuDw9vWrVtnNBdbvPjii10bvK187m4JHDz5zDPPGM3nzPDhw10bXBiTi6vm5OTk2ca1116b+jxq1KhctnjbU7x4cWOYfeSRR8xyDmMEvMGUiw4zzZo1c//GRYg5jI3N8PxSBuD7My44yUb+EBxOyIGVHOjJRXq3BA5ejCLvJZ4/f77Rjz76qNHlypUzOhlWuYkWLVqkPocCLtPFX3/9Zcz6HBIZetmDDbR8v+J9xOGOgA8e5HBRfimBQ1EB4NlnnzV65MiRRp966qlG8/0PACZPnmw0f7fk9Q34/g7w5zf3vcWKFTM6tD9++eUXo8uXL280n0Oha4r7q82hJzxCCCGEyHg04BFCCCFExqMBjxBCCCEynnwFD9aoUSPu2rVrSv/8889mOQc3AT54iOfNOawvFDzIYUU8x8ceD/bWAMBtt91mNM9x8vxl06ZNXRvsHapatarR48aNMzq5rzYHz5PzPt0a2CPAPioAyMrKSn2+++67MWvWrAIJ+9pxxx3jZHFX9ob06NHDrcOeHZ7nP/TQQ43ee++9XRtXXnml0RzcxR6IkH+DQzRnzpxpNB9vDpQDgP79+xvNno8tYffddzea5/zZS3TzzTe7Nth7kQwRBLwvqGPHjq6NZKHEjz76CEuXLi2w4MESJUrESf8TF21cu3atW2fEiBFG77jjjvn+vQMGDDCafV1cTPHNN990bUybNi3X3zFr1qw8t4PPcfb+sX8wFHjKxUDZT8ZeCQ6ADa3DfSLrUDHV5LZ16tQJU6ZMKZDzqEqVKnGy+CmH97388stuHfa1cHFs9pKE/Dd8PXJwLgcTTpw40bXB4aotW7Y0OuSnZXjbOcC1cuXKRodCFLkv5vOSQwVZA0C1atWM5iLkvE7IS8bhu2PHjlXwoBBCCCH+nWjAI4QQQoiMRwMeIYQQQmQ8+fLwlClTJk7O0a1Zs8Ysv/DCC906J598stGcf8PFIp9++mnXBufqsL+A5yLfeust1wbncvAcKM9n/y/D2Qa8TzlzI5RbUK9evdTn5s2bY/To0QUyb16hQoU46blZv379ZrdrE5xTwgXqnnjiCaOnTJni2njnnXeM5kwJPv4PPfSQa+P+++83ulu3bkaPGTPGaPbFhLZja2CvGH8X9gQUKVLEtcG+gLlz5xrNBfq4SCRgMzc+++wzLFu2rMA8PHvvvXecLJLLfqqzzjrLrZPMewGAYcOGGb1s2TKjL7roItcGF1HlHB72G4b6lSpVqhjN+56LH3O+F+DPce4T2DsT8qPkRf369Y0O5bCwr40zVTjfiotNAsDBBx+c+nzppZfi559/LpDzaK+99oqT29O7d2+zPHRvrFSpUq5tst8kdB6yH27FihVGs/ePM3YA79k65ZRTjG7fvn2u2wl4rx97VNkbGcq3y8v7x7lLJUuWdG2wD4zPO9Z8PwA2Fi9O0q1bN3l4hBBCCPHvRAMeIYQQQmQ8GvAIIYQQIuPJl4eHsy8474azAACfU8Hv0HPuSsOGDV0bnCmRzE4AfDZGqF4Le4m4ZlMyG6Yw4WyQUA2cn376yWiuq8PHgT0CgM1pKMgMleLFi8dJfwH7i0LZNd98843RfOy4LhZnKgFAhw4djObaSzxPzr8TAD7//PNcNc8j87w64Gt6bQ133nmn0XyN8baHfCScucG1tDgfK9RGMmNm8uTJ+OuvvwrMw8OZYOy3CfkeON+E62JxhkooE6xv375Gr1y50mj2aIXq1E2YMMHoo446ymj2VrBPCAD+/PNPo+fNm2c0+822BvZo8jUEeP8F+zj5WuQaUID18Q0bNgyLFy8ukPOocuXK8YknnpjSxYsXN8tD+VWcZ8SeJO57+boC/PWZ9KIBPlcrVH+Kc6e49h/XY1u6dKlrgzOj2G/Vq1evXH8n4GsDsleW6wlyXg7gfU/sg+N+NcTAgQONHjlypDw8QgghhPh3ogGPEEIIITIeDXiEEEIIkfFowCOEEEKIjMdXcsuFsmXL4rDDDkvpqVOnmuVslAKAihUrGs2BWUOHDjU6FNTExmoujFa2bFmjR48e7drgMEI2eW0L2DwdCnxj0xYb1DiULFQI9d577zX6tddeM/rGG280OlkodBPJsMJQcbp0UbVqVVPIkwvBsnEU8EbRZOAdABx55JFGv/fee66Nb7/91mguFvrDDz8YzWFXgDc683G45557jN4WBuUQc+bMMfrDDz80ms2mvN2ADxLcY489jH7ppZeM5mKjAPD999+nPnPgXLopXbq0Caxj4+cxxxzj1mFzOxt9ufgxG7sBH+B30003Gc2hkGyMBrxJk/c194F9+vRxbYTCNbc1bEDmgDzA76MuXboYzddR6MWWZAHldF0zIXJycoxBnF+YKVWqlFvnjDPOMJqPDb8c8+ijj7o22OjLwalc2Prrr792bfB985BDDjGaiyOH4JdZ+H5erlw5o0P3Ii7Sy/0I9xv8gg3g+2K+V3PB2dD96rvvvjM69NICoCc8QgghhPgXoAGPEEIIITIeDXiEEEIIkfHkK3iwdOnS8V577ZXSPD+ZDALbBM/JsoeFw4x233131waHEz7wwANGcxBdqEDZtoDnXtkbc9555xkdKoTK34Xb5DlO9p4APsyJ53x5n4fmPJPeqn79+mHOnDkFEvZVsWLFuE2bNinNhRJDRQ65WBx7Cw4//HCjQ/6Gxx9/3Giee65Zs6bR06dPd23wsWF/BvuAtgW33HKL+zcuHMhz3BzUlfRIbOLnn382moPJWrVqZXSon0h6UWbPno3s7OwCCx6sVKlSfNxxx6U090WhsM3333/f6GOPPdbodu3aGT1+/Ph/upkFxtVXX2304MGDjQ6FcTLsJ+R9ykGbAHDggQcaPWvWLKM5BJb7LsB6VB544IEC64saNWoU9+/fP6XZfxIqIMxFda+55hqjk+0BwCeffOLaYJ8iX89c/PfMM890bfDx5mPFIZshuDgo+23Zlxfqm/Pyj3KYIxcoBXzYLI8J2rZtazR7jQDvnercubOCB4UQQgjx70QDHiGEEEJkPBrwCCGEECLjyVcOTxRF2GGH/x8jnX766WY5F84DgBEjRhjNWT1cGI/nJgGge/fuRrNn48cff9z8Rm9DnnrqKaP/+usvo3lONJkTsgnOUOE8haOPPtroZs2auTYqVKhgdLdu3YzmAqShPKDmzf9/ejOU05Mu4jjGhg0bUprneNkDAADnnnuu0Vw89u677za6du3arg32/fDcOmdu/PHHH64N9v1wG5xbsS3g7BjA5wx16tTJaPYShQp/8nnG1xjnjXABXwCYNm1a6jN70dLN+vXrTQHNV155xSwPHQv2D7L3iYuyhopH8rXEhXwLCvbtsUeHfY733Xefa4O9fdyvJK9TIJzFxB4d9k927NjR6NA+rVy5cupzOvLRNseyZcvMOcH3L+6rAX+/2nXXXXP9HSFPKnuhuHD1ZZddZjT7UwCfXbM1+41zw5L3dgC47rrrjOb7CuAz0EaNGmV0slA0EC4EytlVfA4NGzbM6BkzZrg2Hn74YfdvIfSERwghhBAZjwY8QgghhMh4NOARQgghRMaTLw/PjjvuaLIa2EtxwgknuHW4Pgl7NrKzs40uVqyYayNZ9wkI+zzygucfOWclVOOD4bnVZ555xuhkvgwQrgvGv5fzQo4//nijeX8Bfq6V80M4TyLURtKjEfJ4pIv69esbv0XPnj3N8mrVqrl1uF4P+xG4jkzIv8G1lthvxW3MnDnTtcGZOJMnT3Y/80/hGjihek6cmcMZHJzdtH79etcG1zjia+qII44wOrRPJ0yYkPrMeUrpJo5jk2fCxy/kJ+T6YckcH8DnoYT8c+wXHDJkiNG77LKL0ZzvBACvv/660exJ4H6Us14An7PCuUrsN9ySenmcI8b911FHHeXWYa/b7Nmzjeb9FfJwJM/fL7/8Ms/t3FaUKFHC+GeqV69ulofuZ+xR4VpqyVqTQNg7xdcSH3++B3Tu3Nm1wfeiK664wv1MEvZBAt4/uv/++xvNfquQhyuKbGRS+/btjebjz3X/AH+t5nVPYM8q4Gsycv+1CT3hEUIIIUTGowGPEEIIITIeDXiEEEIIkfFowCOEEEKIjCdfxUPLli0bJwPr+vbta5a/9dZbbp0FCxYYzcFdbHbkQDTAh6DlBRulAR94yKFLWwKHTHFRNzZXH3PMMa4NDjvjcC8uPnfKKae4Nni/ly5d2mg203KwEwBUqVIl9XnAgAGYN29egRTsq1GjRnzRRRelNBe2/O2339w6bCZkszibOrnoHeD3IxvjkoUwAeC7775zbfTq1cvoBx980P1MfunTp4/RbAxs2bKlW4fPQw735IJ+IUMjX6tcOJCvuRtuuMG18eKLL6Y+n3DCCRg/fnyBFQ/dY4894qTZkU26IaPv3nvvbTS/qMBmUS5aCAC33Xab0Wwo5ZcSQgVlOSSRTeV16tQxOmQ45nA2NgMPHz7caDZKh+CQPC7qyIZ6wJtFuf/iUD02hgPWYNqxY0dMmTKlQM6jSpUqxcljzGbY0IsLbHxPFtMGvOmaQ/MAYMyYMUZzAVbuvx555BHXBhfh5H4k2b9vDg5j5euDX44JFVTm+zVv+6JFi4wOmafPOecco/llqGXLlhkdurb5ft67d28VDxVCCCHEvxMNeIQQQgiR8WjAI4QQQoiMJ1/Bg9nZ2WaOjgvUcegU4IOIeM5v/PjxRp944on52SQA3o8QCoTjufe8CBW544Aknp9kv9LUqVNdG2+++abRTZs2zfd2sM+Di6l27drVaC5yCVh/0fLly3Pdhm1JTk6OOU/4/HjppZfcOhdffLHRPOfL3gteDvigNp4X530a2icc8LgtPDwjR440ms+PUHHKJk2aGM2eHvYRsMcL8AUMr7zySqN53jwUGJYM8wyFW6aTnJwc0xdxIV8uahmCfyYZZAiEPQvslWGvF+/XkAevS5cuubbBfptQf8bhdeyvCRXQzQv2cHDfxCGpgD8/uaglh2CGAl6TYbMcZJdOSpUqhX333Tel2Rezbt06tw572zi8j/fHgAEDXBt8vPk+wV6qUKgn95uDBg1yP5MXHMa70047Gc1+rJCPj/1oXOyY74khXxSHDed1DX3zzTeuDfYG9e7d2/0MoCc8QgghhPgXoAGPEEIIITIeDXiEEEIIkfHky8PTrFkz4w/gOUD2FgBAhw4djOa5Nc4ZCRVK49wV5qGHHjK6UqVK7md4fj4vQvO3XKDwo48+MrpHjx5G8/4BgGuvvdZozuHhQpCcSQD4fAzOcWCvUaiAXXLuemuKsW4tO+20kyl8N3jwYLM85AXh4nGjR482+sknnzSa58AB79fgueesrCyjQ94pnp/Oi1AWBvuAWrdunevvuOqqq1wbnCnFvjcuAslFTwGfzcM+r8qVKxvN3iLA+ll+/fVXtzydLFq0yHgMVq5caZbz8QWAFi1aGP3999/nurxWrVquDfa5dO/e3WjONbv//vtdG7fffrvR7PvgIo777befa4NzV7hgMPsgtgTuizt16mT0888/79Zhf8nDDz9sNPd3XFwTAKpWrZr6HPIJpYs1a9YYPyH72L766iu3znvvvWc0F7/m3B32sADAxx9/bDRnNfFy9jAC3rOyNdcf5/BwZhJnW/F2AcB1111nNN/zuGhzv379XBuNGzc2mj25XFyVc+gAXzx3c+gJjxBCCCEyHg14hBBCCJHxaMAjhBBCiIwnXx6eyZMnmznsNm3amOWh7AfOGeGsHq6lFKrXkhc8f7ct5oFDeUDsL+LMiOOOO85ozn4BgAMOOMBo9gFxbZqvv/7atcH+Iq6b8/bbbxvNngHA5jbwvHQ6ycrKQt26dVOazweuMwP4eXDez3zehdpgLwznyvA5M2nSJNcG+0TyYvHixe7fOC+GM3Q4CyOU5cQ+J84CYc8WZ10B3vd2wgknGM01gkJ+hiFDhqQ+h2p+pZNq1aqZXBD2+fF+Bfz1yllL7KVhLxSw0ceYhHNF2F8WquvHtc+4X+FaQSEfX7IeHbDRj5KEM2XY1xdi7NixRnNdu7POOsutwz6fCRMmGM0ej5C3iH1tBcXatWuxcOHClOZ6cqE8r5o1axp91FFHGc2ZMWeffbZrg/2C7KXh8zDkBV27dq3Ro0aNcj+TJOSNffbZZ43ma4azuHg7Qz/D/huuxxXyV3JfzD4wzmYLXdsvv/yy+7cQesIjhBBCiIxHAx4hhBBCZDwa8AghhBAi49GARwghhBAZT75My1EUoWjR/1+FzaNcTBQASpQoYTSb+tiAy2FgW8LWmJSThkvAG2NDRl4OGeOwLy5qxgY+wBu9n3vuOaPZWBkqprl69WqjOQwtGSoY2k7AFvFbunSpW54uli1bZsK5uPBpyNTGBQfZKMcF+tgYDvhQMTYGcuAfB4ptDWy+A7xRlg3FrVq1MpqPJQD06tXLaD4v2WwZMpvuueeeRnMgKBtJ58yZ49pIHsdQoc10sm7dOixZsiSlOXzzhRdecOtMmzbN6JtuusnoZN8GAN26dXNtsGmZDcVcyDYUGpcsegr4FxP4RY7q1au7Njh8lNvgMNFQUU4OSeSAN74GQibexx9/3Gi+FrlQ6osvvujauOOOO1KfQ+badJGVlWVM5WzK5UBTwG8/n2e8TqgNNi3zSydsQA7d3zgks0iRIu5nkqxfv97927vvvms0m/QPPfRQo0MFlTk0MflCCuCDGO+88848t42L5SaDKQFfHBsAVqxY4f4thJ7wCCGEECLj0YBHCCGEEBmPBjxCCCGEyHjy5eGpV6+embPkwmBlypRx63BBMi42x4UhOdgJ8HPLHJjE4VehcKv3338/19/btm1bo5PzypvguXf2J33wwQdGh+beOQCMPSwcfsX7C/DzyA8++KDR7KcIFTAM+a0KghIlSpjgKA6Meu2119w6HL7H/hsujhkKWeNwQg5uW7VqldGLFi1ybeSXkB+Bg9gmTpxodP369Y3mIp8h+PvyucuFQAF/jnBBPvYzcPFVwAZrnn766Xlu57akbNmyxg/HXhG+3gHvL+D9xp6GUOFH9gp17NjRaPZfzJs3z7XBBWPZx8fbFSq4yf4a9nWdf/75RrNfJ0SoOGQS9nQA3n/CAZ98XoU8acmgxYL0E5YrV854rtgbycWgARjfGOC9f+wTC4X18XnG5+Wll15qNHtYAOCSSy4xmvvE66+/3ugLLrjAtcHeSPZfcVHP0PXA1z37es844wyjQ8HCvN95zMDLQ/5KDvhMFjlPoic8QgghhMh4NOARQgghRMajAY8QQgghMp58eXhmzpyJDh06pDTPT4ZyGni++u677zaaC11yYT3A+xh4ju/NN980mudVAT+nz94ZnnvneVUAqFSpktGlS5c2mvNuQr4GLkr6xRdfGM25LDwnCgDt27fPVf/6669Gh4ppJrMduPBgOsnJyTHbx3PiXAgW8PP65cqVM5pzGUJFbDnrgQuBsn8hlJ/BmSqcQcLnXahoKxcH5aKcDf6PvTOPt2r6//9rN9/mOUmTTIWiASGEQkRIZMo8zwmZQiizZCoiYx8kkjEhSYaSIU1EGiTNaY7avz/o/Pb79V5176Vz77fd6/l4eLivztnr7nP22muvu9drv94NGhjNvjDA+1VuvPFGo7mQJheFBHx/P+aYY4zm3CL2SQG2n4VyerLJzz//bAozsq+Ni3oC3qfHvqT69esb3bdvX9cGF+U84IADjObjlZOT49pgTxp7Y7744otN/g4AaNSokdFcMJY9aqFxNbesqWRxVsD3M8AXsaxUqZLR7GPjHDZul8fDbLJy5UqzP5x/xHkwgPdwcQ5Tx44djeZrFQD079/faD6fOQMrdO7x9euII44w+p577jE6VMi4X79+RnOWDRcEDu0Hn1Nc2PiSSy4xOuRH45yhatWqGc1e2JDPi32sPAZuQHd4hBBCCJF6NOERQgghROrRhEcIIYQQqSdfHp6cnBzjU+H6HXvssYfbhmttsK+lYsWKRnONH8Bnk3CtpORaPuCzTQC/lsprsZxjEcryefDBB43mOipc36ZHjx6ujb322sto9o7wOnIo/4TrfHE9Ml57b9++vWsj6S8J7We2WLduHZYuXZrR7LeZMWOG24Y9DXwc2OfE6+iA9xrw8X3yySeN5jVywHuj2J/Gx4UzSQCga9euRs+cOdNo9vDwZwe8z4l9EpwNE6pfw14x3lf2jbBPDAAGDBiQ+Xlz1B7LDytXrsTXX3+d0c2bNzevczYXAHz++edG77jjjkazLybkFeDxi8eRWrVnSegMAAAgAElEQVRqGc35KIDP2Rk4cKDRXK8r5J3h2lnst3jjjTeMDvmRQtksSTj/JJTDkuwDANChQwej+TiEvBVJL1xB1mT766+/zDFu27ateT1UO4o9KgxvM2TIEPcezho79thjjeZzrXfv3q4NzvthvxFn1yWzjjawfv16o1u2bLnJ/QrVhWNP24oVK4xu166d0aFsI+7L7IVk/xFfZwE/Nm8M3eERQgghROrRhEcIIYQQqUcTHiGEEEKkHk14hBBCCJF68mVarlOnjjG/seH42Wefddu0aNHCaDZTcfgRF+QEvJGXDYpcoI5DtwBfkI0DENkIy+ZiwIe1ceAdGz/ZFAb4ICYuLsht8H4B3ujMBfy4YFuo+NyaNWsyP+elsGC2KFLEzrlDBj0OwWNTMvePkFm4ZMmSRnMAFhdb5FBBwJtJk+ZrwBftDJlCub+zSZ9Dt3i/AeDEE080mkP2+EGBkNmUgzW5GCG/zgGJgDX633DDDe71bFKlShV06dIlo/ncHDRokNuGDZMcetqqVSujJ0+e7NrgkDQuWsjHi4sUA8DEiRON5oKy/PAHh7sB/hzgBzc43JAfZAD8uHHllVcaPXz4cKP5YRHAB05yQWkOheWAPAB46KGHMj+HxsxsUaJECWMy5xBUPs8A4NVXXzWaDeXffvut0Z07d3Zt8EMEHArJ/SP0AA1fa/n6xg8lhIInue8OHjzYaH4YhI3SADB69GijR4wYYTSH84ZCgbt37240h8RyPwzNEbj488bQHR4hhBBCpB5NeIQQQgiRejThEUIIIUTqifLj34iiaD4AnwwntnTqxnFcLfe3/XfUh1JLgfUhQP0oxWgsEpuDYD/K14RHCCGEEGJLREtaQgghhEg9mvAIIYQQIvVowiOEEEKI1KMJjxBCCCFSjyY8QgghhEg9mvAIIYQQIvVowiOEEEKI1KMJT4Aoik6KomhyFEUroij6KYqiVrlvJcTfRFFUL4qit6MoWhxF0dwoih6OoihfhXqFiKLo+SiKfoui6I8oin6Iouic3LcSwhJF0cgoilZHUbT8n/98FdCtBE14iCiK2gC4C8CZAMoBOADAz5vcSAjLowDmAagJYA8ABwK4aJNbCOHpBaBeHMflARwN4PYoipoV8j6JLZNL4jgu+89/Oxf2zhQWmvB4bgVwWxzHn8dxvD6O41/jOP61sHdKbFHUB/ByHMer4zieC+BdALsW8j6JLYw4jifGcbxmg/znvwaFuEtCbNFowpMgiqKiAJoDqBZF0bQoimb/sxyRU9j7JrYoHgRwUhRFpaMoqgXgCPw96REiX0RR9GgURSsBTAHwG4C3C3mXxJZJryiKFkRR9GkURQcV9s4UFprwWGoAKA6gI4BW+Hs5Yk8ANxbmToktjlH4+47OHwBmAxgH4PVC3SOxRRLH8UX4e2m9FYAhANZsegshHNcC2B5ALQD9AQyLomirvFOoCY9l1T//7xvH8W9xHC8AcD+AdoW4T2ILIoqiIvj7bs4QAGUAVAVQCX/7woTIN3Ecr4vjeDSA7QBcWNj7I7Ys4jj+Io7jZXEcr4nj+BkAn2IrvaZpwpMgjuPF+Psv8mQJeZWTF/mhMoA6AB7+Z4BZCOBpbKUDjNisFIM8POK/EwOICnsnCgNNeDxPA7g0iqLqURRVAnAlgDcLeZ/EFsI/dwWnA7gwiqJiURRVBNAFwHeFu2diS+Kf8eekKIrKRlFUNIqiwwB0BvBBYe+b2HKIoqhiFEWHRVFU6p/x6BT8/eTxVukp1ITH0xPAWAA/AJgM4GsAdxTqHoktjeMAHA5gPoBpAP7E3xNnIfJKjL+Xr2YDWAzgXgBXxHH8RqHuldjSKA7gdvw9Fi0AcCmADnEc/1Coe1VIRHGsFRshhBBCpBvd4RFCCCFE6tGERwghhBCpRxMeIYQQQqQeTXiEEEIIkXryVcE5JycnLleu3EZfD722Zo0NBmWT9DbbbGP0zz/7Op3ly5c3evXq1UZXrFgx1zb49yxYsMDonBxbPaJs2bKujSpVqhhdpIidL/71119GL1261LXx+++/G837znrx4sWujWLF7GH7888/jebvOIp85EKyjSVLlmDFihUFkstQpUqVuG7duhnNx4qPE+D3v3Tp0kb/9NNPRleqVMm1wX1o2bJlRleoUMHo0PfOv5ffU6JECaPnz5/v2thxxx2N/u2334wuVaqU24ZZuXKl0dWrVzf6jz/+MDp0/Plc5T7DfWzGjBmuje233z7z86xZs7Bw4cICy/YoV65cXK1atYzmc5G/IwBYu3at0StWrDC6atWqRvP3CAA1a9bcZJv8vZUsWdK1sX79eqP5mE+bNs1twxQtWtRo7t+LFi3a5H4Cvs/zmMfbFC9e3LWxatWqTb6Hx2o+TgBQq1atzM8zZszAggULCqQfVa5cOa5du3ZG8/jN1wjAn0t83vD4HTr+3K/4O+JzM3Rd/eWXX4zmY8efJXT8uQ8xvF+VK1d27+FrIvfd5DkK+PkA4M+Z2bNnG73zzrbWaei48Dk1b968BXEcV+P35WvCU65cOXTs2HGjr7du3dr9G1+M+EBce+21Rp900kmujbZt2xo9adIko4855hijTznlFNfGVVddZfQTTzxh9K672tqOBxxwgGvjtNNOM5oHKh5k3nnnHdfGvffeazTv+3HHHWf04MGDXRt8QecLK0+AQgNVsvM+9thj7vVsUbduXXz88ccZzd8pHyfAT0abNGliNH9nnTp1cm0ceuihRn/44YdGH3nkkUa//rqvBLH77rsb/eqrrxqdnMgB4e/13Xdt/MWdd95pNJ/cIb7++mujL7rIFmJ///33jeaJGOD7Nw8YfPG/4IILXBsvvPBC5uc2bdpsYo83P9WqVcNtt92W0XzBHzt2rNvm119tDeDPPvvM6HPOOcfoESNGuDa6d+9u9MyZM43mySdPcAE/0eJjftRRRxkdujDxBe6www4zOnlsAP/ZAeCII44wev/99zeaLzzcJwA/FvMfLFOmTDGaz2UAuOuu/x9Cvvfee7vXs0Xt2rXNGM0X0oEDB7ptcpvwHH300UbvsMMOrg0+P/k7OvDAAzepAeCss84yer/99jOarwmh488T3nXr1hk9depUo0PXVR6/27dvbzSPTTwfAPz17LrrrjP6jTdsEgNfuwFg+fLlRvft29f/hQYtaQkhhBBiKyBfd3iqVq2Kc889N6O/+eYb83ryL64NJG8ZAn5W2bdvX6Ofeuop10bjxo2NTt5KB4DLLrtsk78D8DNe/uucZ5FnnHFGrvvBf61/+eWXRu+2226ujU8++cToH36w+U98x+eggw5ybfCMmP/q4L9UuE0AeOSRRzI/h27dZ4t58+bh4YcfzugHHnjAvH722We7bZJ9DgC+//57o/k2Kt9FA4DevXsb3aJFC6OXLFliNP9lAvi/TsqUKWP0Rx99ZPQhhxzi2uB9/+ADG5zLfYb/cgOAIUOGGM39gZfeQnereBv+C7BRo0ZG83kM2LuzfDcg25QsWdLcPeH95aUnwP8lzXe5+O7EF1984doYPny40V26dDE6t7vAgL/Nz3/B853h0C18HvO47/E4ynfWAd/3WHPfmzhxomvj22+/NXrQoEFG812jSy+91LWRvNPGd7+yyZo1a8zS0LBhw8zrderUcdvUqFHDaO5nPCbw+A4A/fv3N5rvWCTveAHAjz/+6NrYc889jea7vk8//bTRyTF3A7nd4Zs8ebLRBx98sGuDVzWOPfZYo6+55hqjeewC/HfE+8p9KpQduM8++xjN84oN6A6PEEIIIVKPJjxCCCGESD2a8AghhBAi9eTLwzN79myzbs9rgKF1wltvvdXoli1bGj169GijL7zwQtcGP8XwyiuvGH366acbHXq64sYbbzT6vPPOM/qSSy4xOuSd4SdB+BHzZ5991mh+qgcAPv/8c6P56ZgJEyYYHfIj8Xrt448/bvQzzzxj9B577OHaSK698/pnNilfvjwOP/zwjGZXf+hpoKFDhxrNn/fUU081mp+WAoDnnnvO6Lfeesto/o6ST5JtgNerBwwYYDSv799+++2uDX4M/8wzz9zkfjz00EOuDX4akj/b+PHjje7Ro4drg/vZ8ccfb/T06dONfvLJJ10byc/PT71lm9mzZ+Pqq6/O6FmzZpnX+YkrwI8j7Oti3wufzwDQtGlTo+fMmWN0z549jWZfDOAfXeYnythL8uKLL7o2+PhcfPHFRvOjvuwbAoD//e9/Rr/33ntG89OQ7McA/FNJHAfAT7+GxsSkN5L9Z9mkbNmy2HfffTOa/YR8HgHAd999ZzSPnfy0Z79+/Vwb7Jls0KCB0Xw+h8Znvl7Vq1fPaPYK8rULAF577TWjefxq1aqV0ePGjXNtsIeNPYj8SP19993n2mBvHT9xyH0sFPfB84iNoTs8QgghhEg9mvAIIYQQIvVowiOEEEKI1KMJjxBCCCFST75My1EUmZA7DjQLGZLefvtto9noyebCrl27ujZefvllo7ncAptar7jiCtcGR8Bvu+22RnPQHBtUQ1x++eVGczgbBzMC3rTduXNno9m0GjIb5hbEt9NOOxnNoYqANRPyd5NNpk+fbiLKOcwqtK9shGPDOce3h8K+OLzrxBNPNJoD0diQDgDdunUzmsO/OHSLzdWhfWNTHxvyOagS8H3oq6++2uR+fPrpp64NDkVkwyq3EQodSxquQ/W6skn16tXNMeNaQaEyN1y2gPsEj1+huj9cboBj8HksCvVFNubefPPNRnMAYmgMGDVqlNFs9OTQRDZoA0CfPn2Mzi1Ild8PeMMtByCy+T9U0yk5boZKIGSLRYsWGUM4j6shgzWPCzz2cO2s0HnDwYJcnoO/51CoJ9fSYgM6G585VBDw1w1+GIYN2tznQu+5//77jW7evLnRoXJLHCTMJUz484euEXl9aEJ3eIQQQgiRejThEUIIIUTq0YRHCCGEEKknXx6e1atXmzXpkSNHmte5YBngC7DxenVOTo7RHCAGeF8Dh7U1a9Zsk20CPjitYcOG7j1JOLgL8AX42KPDPoZQ+BkHkXE4HXsA2CMAAB07djSavVRcWJBDqgC7ph8KFMsW9erVMwFvefFwsTdqxowZRnPR2tAaLxf54zVu7jMhDwgHdS1btszoZBFEAHj00UddG6tXrzaag8r4WHFAIuAD9N5//32j2cN2/fXXuzaShTcBoESJEkaz94A/O2D7UEEWoAWApUuXmu+GfQ/Fixd32/A4ctRRRxnN3oqTTz7ZtbFu3TqjOSiT+xF7oQAfgsrjF8OeBgDo3r270bfccovRy5cvN5rDWQHgzjvvNJr9FgsXLjSaQxYB78nhADwOWQx5mnbZZZfMz0WKFNzf4KVKlTJFZ9nHxwWmAeDNN980mv2FrEPnHheY5T4yZswYo9lbBXi/GV9H2OfIwaKAD4Xs1auX0c8//7zRHIgJ+HGCzyEODQz5ojjwkwsbJ8MhAaBTp06uDfZC3n333e49gO7wCCGEEGIrQBMeIYQQQqQeTXiEEEIIkXry5eHZZpttTPHQ3377zbzOxUQBn+9x7733Gs3FRffbbz/XBq+9c/YD5wPw+jbg8wEY9uywXwfw69O8TsiZQ5zrAPjiepyHwmuvIS/R5MmTjeY1YPZ4hDw6ybyIglw3nzNnjvEbsM8lVCiPc5i4j/A6Ma8BA95Lxrkk7L0IeR7YS8R5KlzENFT0ccGCBUbzGj9n9/DvBID69esbzb6Cm266yejQd8o+qO22285ozhzi3wnYfIxQvko2KVu2rCnwy+MK+yQA/11XrlzZaP6Mw4cPd23w5+S+x4VaX3rpJdcGFxDODe4zgPe69e3b1+ipU6cazX0T8Hk/vK/sHzv77LNdG7kVUy1VqpTR/J0DQLt27TI/DxkyxL2eLdasWWO+J/apxXHstmGvCGeezZ0712j24wDeT8jjOec/cRFPwI8tZcqUMZr9WS+88IJrg4twsmfrjjvuMDrkjeVcIs6m4zyoY445xrXB13POheNrxFlnneXa4GvExtAdHiGEEEKkHk14hBBCCJF6NOERQgghROrJl4dn+fLlpi4PP6fPXhvA+xx4m9zqdwDAgQceaPSFF15oNK8ThmAvEcOenVDmwIQJE4w+/PDDjR40aJDRXCME8OuR3CZ7jbjmFwCccMIJRq9fv95orhFUt25d10Zyzbcg/Rfbbrut8Tmw34ZriQG+Pg97MdjDxfVtAJ9DwZ4PPnahDBbmnnvuMZprvHGNKwAYPXq00bw+z3WTQseO675xrSzOYWLPAACTPwIAP//8s9Hs8Ql5QJL5KexDKAiS5yx73XiMAHy9Hf7uOUuIxybA+xQ5N4n9OYsXL3ZtcI4Y92+uk5QXOEOGYX8G4D8f97WkRwoI5wWx34IzsNijGfITFqSHMEmZMmVMfTX2rfG5CgDly5c3+sEHHzQ6LzlMTZo0MZrz28qWLWs018kCfL4Re3bYCxqqyXfDDTcYzWPAY489ZnQom+zQQw81mnP2pkyZYvS8efNcG+zb6tKli9FcO5C9w4CvScmZQhvQHR4hhBBCpB5NeIQQQgiRejThEUIIIUTq0YRHCCGEEKknCoUrbYwdd9wxfuCBBzKaTZpt27Z12+y///5G//rrr0az6Y1DlwBfGHDgwIFGs8Hyzz//dG0cd9xx7t+ScJhR//793Xu4WCoXD504caLR/FkBb9p75JFHjE6GcAHhoKak0Q7wgYhsBGaTJG9z4oknYuLEiZF7UxaoUqVKnPyMXAiVzXiAL0p7xhlnGM0Gtb322su1kQzJA4A999zTaA7m6tChg2ujRYsWRrMxdtq0aW6b/wqb2gFgyZIlRnOoIpuWQ2ZTNhtygb68kDRt9+nTB7NmzSqQPgQApUuXjpOmaS5+ygVVAV9wMRmiCnjDLRdYBYDp06dvss2rr77aaA5iA3wYH4fGffvtt26bbMDBimzA5gdIQiZ8NoIzHMQXMr8nC72edtppmDRpUoH0o1KlSsXJYE8eJzlUD/APSLApm4t4Ll261LXBD8Rw4Vc2vofGAH4wgx8Y+vLLL43mB2wAb2zm/eDjH5or8EMmXEw3OV8AwuM7m/T5msgP7oQ+C5uWy5Ur91Ucx835fbrDI4QQQojUowmPEEIIIVKPJjxCCCGESD35Ch5cv349Vq1aldFcsI7XDQFg3bp1RnPBOl6/5vAvAGjYsKHRXDht+fLlRucWMhiiatWqRnNROMAHc7F3hgPg2OME+LVWXvN/6KGHjA6FkPFaOxcKZA9AKPzsmWeeyfwcCnLKFpUrV0anTp0ymteNufAl4IMn2cPCRVrZwwL4vsnhhVyQdPfdd3dttGnTxv1btgntB/uNnn32WaO5CGQozDPkc9oUIU9Tsijpk08+ma/2/iuNGjXCuHHjMrpnz57m9SOPPNJtw/44DgFlfvjhB/dv7ONiPyEH/IUCAdkrxGGsPBZxQVnAF1xkr9CsWbPcNswpp5xi9F133WU0h+aFQkF5PGN/IY9fobH54IMPzvzMBS2zTdGiRTM/X3nlleY19vkBwDnnnGM0jwn8eQ844ADXBnv/uPAle1ZDx5LH7OSYCnhPW6gfslfms88+M5rDeNnDCni/LPst2RuYvO5sgPvZq6++ajT7L0PXZi6gvDF0h0cIIYQQqUcTHiGEEEKkHk14hBBCCJF68uXhKVKkiPHccCGwUKFLLrbHOTucY8DvB4DbbrvNaPYFsacn6TPKK1z07I477nDv4UwhLtLJ69ecFwP4DAb2jnAhwZD/hv0J7EfhfA3+fgC7lhrKT8oWq1atwvfff5/RvPbMxWYB/51xBsWYMWOM5uKagC/Axz6JFStWGN2+fXvXxv8VXn/9daPZi8H5GuzxyQtc0DBUxLR69eqZn7l4Z7aZNm0ajjrqqIzmPsweF8AWOwV8P+ECjJztAXgfEOdI8dgUyq7hcZKPD2eGsacB8EUsly1b5t6TG/wd8b5zwd1QkeEnnnjCaPYg8nnF5yoA9OvXL/Mze0+ySYUKFXDYYYdl9CGHHGJeD/lvmjZtavTOO+9sNGdxhbxyixYtMprPX/ZK3X333a4NzoDjNrjfvfvuu66N5PcOACeddJLRfH1/5513XBv8b7zvvJ8hOAMvN3g+APj8o42hOzxCCCGESD2a8AghhBAi9WjCI4QQQojUk6+F91WrVpm6HqNHjzav8xooAHzwwQdG8/PyvGb7008/uTaeeuopo9n3wTVv+Ln9vHDQQQcZffTRR7v3cGYQZyFwJkfIB8LZH1x/i70mSb/LBrp06WL0+vXrjWY/UqgeV7KmSUH6LxYvXozXXnstozmTpHlzV/7ErYNzVgdnY3AtMcBnRnEf4fo1/5fh7BP2kuVl3Tw32MPTuHFj957kuRyqGZRN6tWrZ3xonAMUGouuuOIKo9kvWLJkSaMnT57s2mD/AGeAsX+QM6IA77HjNlu1amV0yJ+zOforH+Nq1aoZzeMX518BwJw5c4zm85f3nWstAUCNGjUyP4fqIGaLYsWKoUqVKhnNftIePXq4bdj3wuP50KFDjQ55MDkni78jzh374osvXBucX5b8DgHviwn5+NizxXXh/s04kts2p512mvu35557zmiuacbnYSibjL1kG0N3eIQQQgiRejThEUIIIUTq0YRHCCGEEKlHEx4hhBBCpJ58uVUXL15sCp1tt9125vVBgwa5bbgwGIc5VahQweikiWwDHFbFBcm4uGheYDMhGwV32203tw2HO7FxjI1TodApNrVxUTc2MU+aNMm1waZkLj6YNAUDYeNzMnArVLA1W1StWtUEMnLxVC6CCPiidhz+xWbLWrVquTbYyMwG9M1hAh0/frzRHAgJ+MBDPnb/hs1hUmbYPBoKYTv88MMzP7NxPtssXLjQmB2HDBliXucihgBMsVEA6Natm9Hdu3c3ulmzZq4NNlQuWLDAaO6LPEYAuRss77//fqN5TPg3FC9e3P0bBy+yiZWLH4eCZTkUj0Mv69SpYzSbvAFgxowZmZ+TxTyzTalSpUwY5fnnn29e54dhAH/N48A7/s640C/gxwAOVmWTMgeNAj6McsSIEUYXVKFjNjrzd8gPVPADSABMgCgArF692mguKM0PqQB+rA2d/4Du8AghhBBiK0ATHiGEEEKkHk14hBBCCJF68uXhqVGjBrp27ZrRvN4aWmv89NNPjeZ1Yl6fvuyyy1wbN954o9Ennnii0SNHjtz4Tm8E9q3wWjR7i4C/fQNJeH329NNPNzpUwHDPPfc0esqUKUaz36Zly5auDS5Q16hRI6M5MIw9PYDddy5ml03Ye8HBZckwuQ2wR4ULFLJnJ1QojwMdmzRpkrcdTpD0PQE+RJPXs0NhfWeffbbR7E/7+eef871fudGnTx/3b/x9sKeN+0zv3r1dG8liuqFzP5usW7fOhLpts8025nXWgA+5HDt2rNGVK1c2OlTIuEWLFkbzOMEBpv/G28T96t/An//xxx9375k6darRXLSUQ+JCPi4Oa2R/CgfLckFSwPZFbi+b/PXXX2bsY+9MyPfE39Gpp55q9Mknn2w0FxcFgF133dXouXPnGs0+1g4dOrg2eF/5/OWwWb7OAD6skH/PDz/8YHTPnj1dG3fddZfR7AubOXOm0aFgWd6Ggxe/++47o0OBkKHrRgjd4RFCCCFE6tGERwghhBCpRxMeIYQQQqSeKI7jPL+5atWqcdIfMmDAAPN6vXr13DYPPPCA0bz2mJdCl7yP7Fl56623jP7ll19cG8xHH31kNOcDcDYA4LNa+LNtDngdtVSpUu49/fv3N3qPPfYwmtfN+TgB1pMxatQoLFmyJHJvygLVq1ePk+v4nEMUyi3hooW5FVh98MEHXRtr1641mvtMXmAPw5FHHmk0F5MNZUpxET/OqdgccOFA9kkB/jzjbBgu6McZNgDQrl27zM+HHnoovvnmmwLpQwBQunTpeKeddspo9qyULl3abfPVV18ZnfQjAt5f9+GHH7o22D/I52I2xoRswX4K/g7ZKxgqYso5O9OnTzeaM1NC/T3p2ejduzdmzJhRIP2oefPmcbJfX3XVVeb1L7/80m3DfYR9e+zhCo1FXHSb89rYaxbKlOH8p9x477333L8l89AAoEgRe/+DPalr1qxxbfC4wN7Yr7/+2uhQIdS3337b6Pfff99o7nehAqR8DXjuuee+iuPYGYZ0h0cIIYQQqUcTHiGEEEKkHk14hBBCCJF68pXDU6RIEeTk5GQ051Tw+i3ga1Sxd2b48OFGcy0SAHjppZeM5nVTXtML5QUw11xzjdG8fstr9QUF55lceeWV7j1cj4vzErge1W233ebaSK4LJ2siZZsKFSqYnCD2DXBtLcDW2gGAfv36GV29enWjOSsD8F6xvffe2+jQ2jIzatQoo7nfsW+E+xgArFq1Ktffk184T4P7A+cHAd4HV7NmTaP5s37wwQeujaS3bunSpXnb2c1EhQoVjIeIz9dQzZ4ostaQSy65xOjZs2cbHcryYV8D58bwsfj9999dG0nvEfB3HkwSzmIKZUYlM4gA7x/jWmghuH9yZgxn96xcudK1wR4d9mOwv5A/K2BrSbFfL5vMnTvX5Miw36R27dpuG/ZHnnTSSUZz/a2QB7Nt27ZG8ziSl9p45cqVM5r76iGHHGJ0qM4l+2+4P3BdLNYAzHwA8N4//qyhnKXHHnvMaK5Hx97g0Lm9YsUK928hdIdHCCGEEKlHEx4hhBBCpB5NeIQQQgiRejThEUIIIUTqyZdpuXLlyqZY2ogRI8zrJ5xwgtumb9++RrPhuH79+kbfcMMNrg0OEnz11VeN7tKly8Z3eiOULVvW6M6dO+e6zfbbb280mwt32203o0PmKg6Zyo2Q8ZWDmNg4x+FPL774omsjaXIMGQmzRcmSJc0x51CtUFgfhxGyWZ6PJZt2AW/8vOOOO4zmgn554d577zWaizH+G1q3bm10KMeMorcAACAASURBVECPDdjffvut0Vz4kwPkAB+iyEGEbEZlUzMAvPLKK5mfuc9lmz/++MMYqYsVs0NZmzZt3DZs3j766KONZgPmrFmzXBsc1jZ58mSj+XzlhzYAf3w4rI6DNbfbbjvXBu87G2zzYlrm0MTcAt+uvfZa1waH5PFDKbvvvrvRSaP5BpLHJWSMzRbbbLMNunXrltFsMA4VD+XiqPy951bYGADOPPNMo/kzc0gqP4QC+GPDRaa5SCsXKAW8WbhXr15G87jBD8uE9o3HATat82cDgPPOO89ofoCIA0P5IRXAG843hu7wCCGEECL1aMIjhBBCiNSjCY8QQgghUk++ioeWLVs2TvpU+vTpY15/88033TaVK1fmNoxmT8/YsWNdGyNHjjT6xhtvNJoDxHhN8N/Afhzg77CzJByIxKFUocC/UJjVpmC/CuBD1nhfufApr+cCNgBv6NChWLBgQYEU7Ktdu3acXPc999xzzeu//vqr26ZSpUpG81rzPvvsY/TLL7/s2uBgtjvvvNNo9l6MHz/etZEN7rnnHqN5zT8UfsceAC4Ou++++xodKvq4evVqo9k7xUVOx4wZ49r47rvvMj//+OOPWLlyZYEVD23cuHGcHG94HOOwSsB7wdi3xcci5IPh8D32pLCXKFS0sXHjxkbfdNNNRn/22WdGs8cH8MUSOZx15syZbpvcYD8dh09y2CHgvUTJEEHA+484dBEAFi1alPm5X79++PXXXwuskHHHjh0zmgsK83cM+CK7fA3g6x0HMQLeg5f8/ABwwQUXGF2xYkXXRihIMAkH/IWuzRySyD6f+fPnG73DDju4NtjDxd/hLrvsYvSUKVNcG+z7YU8p938OCAW83+j7779X8VAhhBBCbJ1owiOEEEKI1KMJjxBCCCFST75yeEqVKmXW5Hg9O5Q5wf4BzkPh3JGPP/441/3gQo8TJkzIdZvc6NChg9GhvJyBAwcanVz/BYAyZcoYHVoD5jVfzhhiuAgc4Iv63X///Uazh4fXagHg0Ucfzfy8fPnyTe7D5mT16tUmJ4dzTEK+J84z4oKMnB9y8MEHuzY4c4KLeBaUZ4fPkWQhVcB7IEIFG9kXwgVm999/f6MXLlzo2mCvCRf+5ayYUB/Jj/9vc1OiRAnUqVMnoznv5YUXXnDbcG4Wn8+chxIai/i7Zu9Ts2bNjG7RooVrg/NdODPoxx9/NPr66693bbAvb926dUZzkVP2ZIVgbxwXseQxEgDuvvtuo59//nmjeUwMFQd+8sknMz9zxlo2qVKliincyd8pF7YG/LFiLxx7ZZJ9dAPst+LcMPY1hq4R/HvZ+8o+l1AuFXvHeF/5d3DuHuA9l/wezmUKFfbmfVuwYIHRXAya9wvweX+hay+gOzxCCCGE2ArQhEcIIYQQqUcTHiGEEEKknnx5eP7880/Mmzcvo7nezpAhQ9w2NWrUMJpzG9hvseeee7o2osjGMowbNy5vO5wPOLsk6XHZANcK47opybosAHDXXXe5NnLz7PB6JfuEAKBTp06b1Lx+z/4MwNbf4myJbLJ+/XqzBss+iVBeBO8f1/ThXJpQXSzOPmEfWDYIZS5xv3rwwQeN5pwK9qYAft95vXrvvfc2mrMwAJ+5MmfOHKOLFi1qNHugAOtnya1fb24mTZpk6vSwD4AzdwD/PbD/hj0robpu7MlJZhEBPq8oVKeOs7Wuu+46ozlXLNSPeD/Yx5YXzw7Tr18/o5977jmj+bMCvvYhf2dc4yl03h177LGZn3/66ae87exmYO3ataZeGmdesd8U8J7KU045xejLLrvMaPZXAj7vieEcJq7xBXjfaiibJjf4HGdvEX+WUF/m61Oo3tqm3g/4z8eZadyHeHwDrA8MAN55553g79cdHiGEEEKkHk14hBBCCJF6NOERQgghROrRhEcIIYQQqSdfpuXixYsbYxeHDoWMvhw8yAbL22+/3ejBgwe7Nvj3sMnp1FNP3cRe5w02+bEJCgBGjx5tdNJsBwAXXXTRf94PLq7KYWmAL9jWv39/o9lIFwpQS4aZhcxo2WL16tWYOHFiRrdt29a83qhRI7fNK6+8YvTZZ59tNAezcVFbwAf8hcLcNjdTp051/8bBdFz4lM3BZ5xxhmuDA8I+/fRTo7lwZqgPcRglmyDZkF+uXDnXRvLfihQp2L+dOASVCwyHDMf8UMGFF15odMiUy3BYIY8BbLrlvgr4IE02JfMDFFwIEvDhmxzg+W/o3bu30RwqmDxvN8APmeTk5BjNxmcOIgTs+dqlS5e87exmYN26dVixYkVGN23a1LzO4ywANGjQwGg+VieffLLRoeLPXIR66dKlRrNJd3PQvXt39298vWLTOocEciFYwJ9DbLjPLfAV8MXAeTzjsM9PPvkk1zY2hu7wCCGEECL1aMIjhBBCiNSjCY8QQgghUk+UnwKATZs2jUeNGpXRXbt2Na9zCBXgPRq8ps2hcjvttJNrg9fjOcyQA8O4MCKw8SCigoaLY3LQHnuJOnfu7Nrg8Cte0yxdurTRRx55pGsj6Qu4++67MXPmzMi9KQtUr149ToZPcbgVh24BQMWKFY1mL8FVV11ldDKQbgPsUwqFMf5XDjvsMKND5xYHC3IQGxcX5TYBv17NoXNvv/220SG/EvtI7rjjDqPZi/L555+7Nr7++uvMz4MHD8a8efMKpA8BQIMGDeKk54S9JFwIFPAeBQ4GZe9AXnwx7CdkrwS3CQAnnnii0RxEx8URQ+cEh42GfA25wf4SHpv+/PNPo7nwMWD7AACUL1/eaD5XQ2Pzfffdl/l5+vTpWLVqVYH0o5ycnHiHHXbIaPZPcgAeANx7771GDxgwwGg+v0MBmFxg9csvvzT6jz/+2PhOb4TkdwgA7777rtEc+AsAS5YsMXrnnXc2ukePHkaHPKrsJePPy5/t9ddfd23wNlwwmT8b97nQvrZs2fKrOI6b8/t0h0cIIYQQqUcTHiGEEEKkHk14hBBCCJF68pXDs2rVKrMm+9FHH5nXa9as6bZp166d0dtvv73R7M8YNmyYa4PzEFhfc801Rvft29e10bp1a6N53zcHXLQytG7Kv5e9Ecccc4zRl19+uWuD8124eCb7gn744QfXRrJwHHsGskmNGjWM94v9RbVq1XLbPPPMM0Zzkb+33nrL6NAaL69ph4rY/Vfee+89o+fOnevew+v37OHhc+iBBx5wbaxbt85o7tu5+SoAnwXCWT2c/cJr8YAt4rt8+XL3ejbJyckxxR25D4eOL+d1sZ+QC6DmxcPDv5e9UGvXrnXbJH2QAFC3bt1N7kdoXN0c8PexePFio/nzX3zxxa4N3nf+/NOmTTM6lOWSzOYpyDyn+vXrm7GFvXChfDe+nu23335GV6tWzeg777zTtcGeOv69fBxCBZV5zONxlLd54YUXXBv169c3eujQoUZzLlEoV4y9Mw8//LDR/H2E8oDYO7bXXnsZzbljoc8S8uyF0B0eIYQQQqQeTXiEEEIIkXo04RFCCCFE6smXh6dIkSJmvZXXY5O1bTYwaNAgo8eOHWs01+j58ccfXRvsjeD6New3CNXReeSRR4xu0aKF0SG/BcN1QjiDgHNXQvknvA7KtZLYs8F1pELceuutRvPaO9c7AWymRtLPk22WLl1qMpEaN25sXudsFABYsGCB0fPnzzeaPVvsgQC8z2tzwPWouO/uvffebptKlSoZzVkYvH4d6pdcW45/D+dahOoTsQ+K19bZSzd8+HDXxrnnnpv5mWvmZJtSpUohmaHC5/ecOXPcNpwTxrkinNUV8l+w/4aPRcizw7DvJ+SN2NxwVhXgx1H2HHJtLa6TBXiPDnspeJtrr73WtXH44Ydnfj7++OPd69li1qxZxk/I16+qVau6bfgawMcyimyEUCi/ij1ZnBvGdeG4phngfS98LLmu5XXXXefa4Lp+fD6wNzSU9cTXXh6bORMudE3kHL0RI0YYzZ81dD3jDKVLL73UvQfQHR4hhBBCbAVowiOEEEKI1KMJjxBCCCFSjyY8QgghhEg9+TItlyxZ0phdV61aZV4PFcpjA/HBBx9sNJuNQqa/WbNmGT1kyBCje/bsucnfCQBfffWV0RxEx6bW4sWLuzY4dIqLnHGxOS7yCQA33XST0SeffLLRM2bMMJqLqwLeKMgFV5OBbABQu3Zt10Yy8G7RokXu9WyxaNEiExzFhRP79+/vtuHjPWHCBKO///57o9mQni3YpHzWWWcZHSqEy0FlbFDkIMLXXnvNtcHnCId/nXDCCUaHjO+872x6PO6444zmAn6A7WcvvfSSez2bLFu2zHx3bB799ttv3TZsFr3yyiuNfvHFF43mcxPwIZhsOOYCpaHQuJCp/r9y4IEHGs3jW8gsyuP3zJkzjWYDcSiIkX8PG3DZCD179mzXxnfffbfRfcomOTk5aNiwYUZzoGeoCDF/z2z+ZwMuX+9C23Af4QKkyX3cABdDbt7c1snkBxe46CkAlChRwmi+frPxN/QABRcD5ZBbLnR8wAEHuDZ4/OKCyRyk+vjjj7s2QtfaELrDI4QQQojUowmPEEIIIVKPJjxCCCGESD0Rh+VtigoVKsQtW7bMaA5h4vVLADjooIOM5vCiZcuWGf3cc8+5Nl5++WWj2aPDBcvatm3r2uAgpssuu8xoLtIZ8gHxeuwHH3ywyd87b9481wavrbLvh/fj/vvvd20wnTp1MpoLJ4bCHJOhec888wzmzp0buTdlgVKlSsVccDDJBRdc4P6NC1vuueeeRrN/o1WrVq4NDjjMBuzn4EKwgA9J5LV4Pl+43wI+zIzPOz7HQoX1OBCMAy+5cGKoLyc9bLNmzcLq1asLpA8BQP369ePk2j8XOgwVqeQiqxyCx+NMqOjuPvvsYzSH03EI6PTp010b55133ib3688//zS6Ro0arg32OXC4JHvBJk2a5NrgIpXsF+RCz+xhBPz5yh4XHt9Cnpbk+NW6dWt8/fXXBdKPdttttzjpD+TxnAtyAt5jyJ836UcCgMqVK7s22O921FFHGc37EfJOsX902223NZoDTDm8DwAWLlxodLdu3YzmsWnAgAGuDQ4e5O+Hr+8cZggAN998s9HsKeVzOTRn4ULAe++991dxHDfn9+kOjxBCCCFSjyY8QgghhEg9mvAIIYQQIvXkO4cnWbCP1xq5MCLgc0c4U4ezL1599VXXBj/rz9kOH3/8sdHjxo1zbZx22mlGc/4Pr09/+OGHrg1e0+c8BfbOcOFUAKZYHeDXRblQ3BVXXOHaYA8A+3748/fq1cu1kfTw8HHMNuvXr8/8zHkvfCwBnyPD3yuvm3MODeBzldgn8W/gInecF8GFcQF/vPn48pp4+fLlXRvsP+McqmSBX8CvbwN+39m/wedpyAeUfE8yW6kgWLZsGT766KOMZp9LyLOV7HeA7/fsx1izZo1rg8+lzp07G83Hr0+fPq4N9jFwls3pp59udHLM3cAbb7xhNPtt+PWnnnrKtcH9ZvTo0UZzkUb2zgHeB8R9k/2Fy5cvd20kPRqh17PFmjVrTKYZZ4KFvnfOCcrNoxTy4PH1rFGjRkYn+zXg+wPgc3f4esX+o6VLl7o22F/Gx4r3M1TYlT1KPL5xDg9fuwGfd8W+ML5GhOYIXBh4Y+gOjxBCCCFSjyY8QgghhEg9mvAIIYQQIvXky8NTtGhR4yngHAv2yQA+l4LX+nmNk/0ogM++4Ofwk/W9AOChhx5ybXAeAD/r/8477xjNtbcA77dgnwB/HyFfA/tNuI1mzZoZvWTJEtcG5/3w98H1l/izA8DAgQMzPy9YsMC9ni2KFSuGatWqZTTXDtt7773dNlwXiWvc1KpVy2jOoQGAU045xejk5w9xzjnnuH/jNW3uu7179za6QYMGrg2uJcR+LF7PDmVKcdYF+9GKFLF/x4T8SpzjEkU2+oTrl73//vuujWQWCOeCZJuaNWuaGjwdOnQwryfzwjYwZswYo/k84XMxVB+M84k434TzgNgXBHhv1/jx441mr8ScOXNcG19++aXRXD+Oxx6u4Qd4XxbXSxs2bJjRIZ/E0KFDjeaaXUmvIPD3NYRJ1kQLjVXZYvbs2cZTyZ6lkAePa85xThpn23DuFgDMnz/faPZXsQ8m5Flhv81dd91lNI8z7NcK7SuPPZwZx74wwNe12m233YweNWqU0ZwXBQAXXXSR0dddd53R7JPisRzw9cgee+wx9x5Ad3iEEEIIsRWgCY8QQgghUo8mPEIIIYRIPZrwCCGEECL15Dt4sH79+hnN5kkOQwKANm3aGP3WW28ZzeFHtWvXdm0kw6EAbyhmI/Chhx7q2uB2uSAZmwC5ECTgA/2+//57o3Mrxgf40LGaNWsa3b59e6NDJm42X/7+++9Gs/kuZDZLGtJC4VjZIo5jU5SODckczAh4AxobMPnYrVu3zrXBBrw77rhjk6+HwhrZcMcmTzZo8nEBfIhWw4YN3XuScCgZ4I3s3Ec47PPBBx90bbz33nub3Fc2YLM5EQBOOumkzM9sPs823I/23Xdf8/qECRPcNnfeeafRc+fONZrNohwqCPjCrb/++qvRbJgPFf/l0Ec2OrOhPGQ4Xrt2rdE8FrOJm0MFAWsWBoCzzjrLaO57/DoAXH311ZvcDz4OvN+ADdfkBwOySa1atUw/Z+M9m9oB/yACX1fYtBsKQeV+yEGDHPDI7wf8GMchmitWrDA6FD7KJn0eR3ncDT0IwA+ZcJAwhwjyw0KAP6e4v99+++1Ghx5K4gLiG0N3eIQQQgiRejThEUIIIUTq0YRHCCGEEKkn4tC6Tb45iuYDKNjFelEQ1I3juFrub/vvqA+llgLrQ4D6UYrRWCQ2B8F+lK8JjxBCCCHEloiWtIQQQgiRejThEUIIIUTq0YRHCCGEEKlHEx4hhBBCpB5NeIQQQgiRejThEUIIIUTq0YRHCCGEEKlHEx4hhBBCpB5NeBJEUVQyiqIBURTNiKJoWRRF30RRdERh75fYsoii6JIoisZFUbQmiqKBhb0/YstDY5HYXERRNDKKotVRFC3/57+phb1PhYUmPJZiAGYBOBBABQA3Ang5iqJ6hbhPYstjDoDbATxV2Dsitlg0FonNySVxHJf957+dC3tnCotihb0D/5eI43gFgFsS//RmFEXTATQD8Eth7JPY8ojjeAgARFHUHMB2hbw7YgtEY5EQmx/d4dkEURTVALATgImFvS9CiK0XjUXiP9IriqIFURR9GkXRQYW9M4WFJjwbIYqi4gBeAPBMHMdTCnt/hBBbJxqLxH/kWgDbA6gFoD+AYVEUNSjcXSocNOEJEEVREQDPAVgL4JJC3h0hxFaKxiLxX4nj+Is4jpfFcbwmjuNnAHwKoF1h71dhIA8PEUVRBGAAgBoA2sVx/Gch75IQYitEY5HIEjGAqLB3ojDQHR7PYwAaAmgfx/Gqwt4ZseURRVGxKIpKASgKoGgURaWiKNIfFyK/aCwS/4koiipGUXTYhjEoiqJTABwA4N3C3rfCIIrjuLD34f8MURTVxd9PQKwB8FfipfPjOH6hUHZKbHFEUXQLgB70z7fGcXxLwe+N2BLRWCQ2B1EUVQPwNoBdAKwDMAXATXEcv1+oO1ZIaMIjhBBCiNSjJS0hhBBCpB5NeIQQQgiRejThEUIIIUTq0YRHCCGEEKknX4/KFilSJC5W7P9vUq5cOfN6xYoV3TZFixbdZJtly5Y1ev78+e49S5YsMbpSpUpGly9fnvfTtfHTTz9tso3ly5cbXa1atVz3gw3fye8GAIoXL+7a4H3NyckxetmyZUYvXLjQtcHbLFq0yOiaNWsa/XechyX5Hc2dOxdLly4tkFyGSpUqxbVq1croUqVKmddnzpzptuF+VqZMGaMXL15sdOg72247W9KqdOnSRv/2229Gr1271rXBfYaPFeuGDRu6NvhY8LHj38t9CvDnSJ06dYyeM2eO0dznAN+Hfv/9d6P5+ylZsqRrI3k+rF69GmvXri2wbI+yZcvGlStXNr9/Y/u2Af4MO+20k9F8LFat8k+CV61a1WjuazwGrl+/3rXBfY3PgZUrV+baxl9//WV07dq1jV63bt0m9yv0ntmzZxvNnz/Un/k93L9LlChhNB8nwF4DZsyYgQULFhRIPypatKi5nvG1ql69em6b6dOnG81j05o1a4zmcQfw5zTrqVNtMfPQ+M3t8rWGt5k7d65ro0KFCkZzX+bX+Vod+r38/VSvXt1oHqsB3/9XrFixyTb4Wg34c3vatGkL4jh2F/F8TXiKFStmfvkBBxxgXj/mmGPcNvylcafab7/9jH700UddG2+88YbRHTt2NLpNmzZG8wURAE444QSjjzvuOKPHjBlj9AUXXODaGDJkiNE8ECUHYACoUaOGa+Pwww83etdddzV65MiRRj/77LOujd13393oF16wT6nefPPNRocmXskL2vnnn+9ezxa1atXC4MGDM3rHHXc0r19yiQ+TPfjgg41u2bKl0S+//LLRzzzzjGvj7rvvNrp58+ZG33777UbPmDHDtcF9ZtSoUZvUH374oWuDj8VLL71k9C+//GI0n+wA8NhjjxnN5wwf/7Zt27o2uA/de++9Rrdo0cLounXrujaS5+XYsWPd69mkcuXK6Nq1a0bzRWLYsGFum/r16xv9/vv2ydxBgwYZPXGiL1t17rnnGv3UU08ZzX2EJ8EAcOeddxq98862ePX48eON5oso4Ceo3L95whcam//44w+ju3XrZvSECROM/uijj1wb/B6+8PBE7IcffnBtJK8Be++9t3s9WxQrVsz8cch/0Dz99NNum86dOxvdunVro/n85eMC+D+mq1SpYvRBBx3k9pPp3bu30dtuu63RfJ3l8xvw1yIeN9u3b2/0vvvu69rga9yZZ55p9IUXXmg0X0MBYJdddjH6yy+/NPrSSy81msdZAGjQwFbKaN++vR/AoSUtIYQQQmwF5OsOT/ny5XHIIYdk9KGHHmpeD83e+C9lnt3uv//+RoeWEviuB/+1ybPKI4880rXxySefGP3mm28azX8hNm3a1LXRuHFjo3mm2aRJE6NDf523atXKaF6O4NvqfHsPAPr3728035rm2T7fEQCAm266KfMz33bOJosXLzZ3eHbbbTfz+jfffOO24dvGQ4cONfrxxx83mv/6BfxdMO5TvAx6ww03uDb4Ly2+5c3LHSeffLJrg5cWkscB8HdrQndWuN0ff/zRaL5TEVrOeOedd4xO3i0J/d633nrLtZH8y3zy5Mnu9WwSRZFZluvUqZN5PXT+3nfffUbzWPP2228bHbrbeOqppxrNd+j4L+vQX+c9ethMSl4GSI6xQHhp7X//+5/RCxYsMPr00083mv+KBnyf57vxL774otE8ZgL+ziiPTdyfJ02a5NpIfmeh8S5b1KpVy+w/32kJwdcvvrPAd59D1oj777/faF4p4CXoLl26uDb4rjZfA/haFLpzxnfO+Q5Xz549jeYxAwAeeugho3mc5Tt6fA4CwNVXX200W0V4CeuRRx5xbfBdsY2hOzxCCCGESD2a8AghhBAi9WjCI4QQQojUky8Pz/r16816MnslQk9Ysb+GfT+81sheG8D7a6666iqjTzzxRKM//vhj1wY/6XXrrbcazWvx33//vWuDHwU96qij3HuSvPLKK+7f+NFO/r381EeHDh1cG82aNTOaXezsEeA1Y8B+h/w4ajZZtWoVvvvuu4xmR/6rr77qtuEnEp588kmje/XqZfSsWbNcG/ydTZs2zWhe4w75N/iRWvbS8BMb/NQe4Nf02ePGT63xsQT80zQcuXDPPfcYHTov+VFQ7nesQ0963XbbbZmf+VHqbDN//nzj3UruC+CflgL8uMBegCOOOMLokHeGfT9ffPGF0eyt4CewAOD11183mp+qY0aPHu3+jfsvP4bMTxOFHm3mp3T4iVn2FoUesWbfGvuCuF+wXwOwnyX0RFq2WL16tfGYsD/r3Xd9QfFbbrnFaH7C9dhjjzU66VfcwFlnnWU0e4fYPxi6JnIb/OQfe+769Onj2uCnrPn6xvEX7OkBgAceeMBoHgN5P0aMGOHaYO8Ye4X4msnXA8B7YZ944gn3HkB3eIQQQgixFaAJjxBCCCFSjyY8QgghhEg9mvAIIYQQIvXky7RcqVIlE/DF8fMhY9TFF19sNEfrs+E4ZK4777zzjH744YeN5tCl0H6EQpOSNGrUyGgOP8oLbBTkUDnAGwHPPvtso/v162d0yCj23HPPGc2mNw7z4zAowAYcslktm9SqVcuYjJ9//nnzOocmAt60x+GMvA1/fsAbQ7kcCdeJCdVrYeM7B3dx4GXos3A4J0f+s1k6VBeMDats2LviiiuMDtWW47DJyy67zOg99tjD6FCc++WXX575OVQyIJvUq1cPAwYMyGgOawuZvdlQyucWjxtsDgf8+com9G+//dboUIAlx+3ze9h0zv0O8KUCOIiQgyN5nAX8eMX7zg8HfP75564NNmVff/31RnPAZygAMVk+h8302aRkyZImPJQ/H5tpAT8ucDkGDry89tprXRt8fNmUy78jVBaCHxji8/mwww4zevjw4a4N3ncOkWRTc+jhFi7Pwsf3pJNOMjpUj4vnERxwyQ9dhOothgzmIXSHRwghhBCpRxMeIYQQQqQeTXiEEEIIkXry5eFZsmSJWU9mv00oJI/Di3h9bvvttzea/QmA9yBwsUQu+hcKWuN2Q2ua/xUu2shr5ID3GvD6La/Xh/woSe8C4MOd+HcMHDjQtXHBBRdkfm7Xrp17PVv89ddfptAhh7CFvFZcLJQ9XBw0eNddd7k2OKzqjjvuMJrXokNBe7/99pvRHFZ40UUXGR1aV+bjy74SDrL7888/XRscbte3b1+j2c8R8mhxeCcHlw0bNmyTvxOw3qp58+a517PJtGnTjGfqjDPOMK9PmDDBbcNFGMeMGWM0h5WFAv/4+HHgIbfBRU0BH0jJ/AyIXgAAIABJREFU3pFTTjnF6FAAIgd2sneCA02vu+461wZ7I9mPw/6bUB/gMY+9Xk8//bTRIa9X0ueyePFi93q2KF++vPG6sI8tGZC6AR6vueAmh7zOmTPHtcE+Hz7H2ftZpkwZ1wb7Ns855xyjX3vtNaN3331318aDDz5oNF9X+Njx+QL4awd7uLhwcWhM5DZ4XOWQ4J122sm1waGY7E/agO7wCCGEECL1aMIjhBBCiNSjCY8QQgghUk+Un7yZ0qVLx8l1XvYffPrpp24bXvd97733jObfHyra2aBBA6PZ15LMcQCAzz77zLUxbtw492/Zpn79+u7fOGOBC/jttddeRnMuD+C9CKw5L+Ghhx5ybSTzXUaNGoUlS5b4AKQs0Lhx4zjpuWL/AntYAL/GzZ4uzrWYPHmya4MzVThnho8Le88Av3Z80EEHGc3ejD333NO1wdSsWdNoPh/4swHeJ8Jr4LzWzuv9gF+v5zV+zobhopn8nnnz5mHt2rUF0ocAoGTJkvE222yT0VzEsmvXrm4b9qRwX+O8kxdffNG1wf4LzhnhjKwddtjBtfHHH38Y/fXXXxvNOT0hDx573U444QSj2dfIOWOAz2phL1irVq2M5iLOgM+a4u+QC7LWqlXLtTF27NjMz9OmTcOqVasKbCxKerK++eYb8zoXUwV8BgwXR500aZLRoTwvzhXjPsSeOy4oDADJvg94vw2fz3ysAe+X4kKgfJ0NXUc4i2zJkiVGs7cuVPjz+OOPN5rPD+7LXDgVsH0IAB544IGv4jhuzu/THR4hhBBCpB5NeIQQQgiRejThEUIIIUTqyVcOT/Xq1XHppZdmNHsWuE4QAFxzzTVGc70WXlsMrc/lls3A3qHx48dv8v0h+PdyXhCQ/6wRrpsF+DV93neugVO7dm3XxtFHH200Zz2cfvrpRvOaKGA9D6Gsl2xRrFgx02/YsxKqPbNmzRqjf/rpJ6P52J1//vmuDfaSsUeHfS8hbxtnjHBdJfZv7L///q4NznbhbJ8rr7zSaP6sgPencB4We7rYmwD4tXSucVeuXDmjQ5k0SQ8f575km5ycHOPD4nEl5J9iv9iuu+5qNNdxYv8J4HNFVq9ebTTXOGL/IeDro7Hvgz09IS8Jj7Unn3yy0Vzji70VgM8q4rGpY8eORnO2C+DzqrgeGffFYsX8JSeZ3cK+0GwyY8YM45c69NBDzeshP+mMGTOM5nMveX0EfD4O4HN1+BrA2U6PPfaYayOZZQZ4Txf37VCNSs7E4aw6vvZ89NFHro2nnnrK6GRtMsD7gkJeIh5ruKYXZ6SFvEQrVqxw/xZCd3iEEEIIkXo04RFCCCFE6tGERwghhBCpRxMeIYQQQqSefAUPNmzYME6G/HEwFxv2AG9AYtMaF/3iYELAF2Rjg2Wo4GRuXH755Uaz6al9+/ZumxEjRhjNpraQqSs32MDIZkQupBZ6DxvB2EDKZk0AqFGjRubnkSNHFljwYM2aNeMzzzwzo9lwHDKosxn4kEMOMZoLEoaCB7mgaLJ4KuDNdm3atHFt5OTkGM0FN/n3clHTvMCBiBwQCHhDLodXcnHVc88917XBRkH+vByGFirqm2zjpJNOwsSJEwsseJBDULkYbCjwj8+LChUqGM0PYbzyyiuuDS7sycZPDokLBe21aNHCaDYDc9hoKDiS94NDTjlEMGR85XOPg+aOOuooo3msAvz5yuMoP+jBxSUBG7R35JFH4rvvviuQftSkSZM4efy4eCgH2gLAzTffbDQ/ZDJx4kSj+aEdwH8H9913n9G33HKL0WxQBoAmTZoYzeGUjRs3Njo53m+Axwl+cKNo0aJGL1u2zLXBD7zwgx0cghsycfN+sFmaC7SuW7fOtcHjed26dRU8KIQQQoitE014hBBCCJF6NOERQgghROrJV/BgHMcmBI7XBUPFQ3kdMOnfAHzQ2r777uva4ICskFcoNzhEjMO9uPgghx0BQI8ePYzmIqX/xsPD6/O8jsrF+ADg4osvNvq4444zmgu2cfgXYL0hBRk8WKVKFZx22mkZzcFsBxxwgNuGi4Wyd4b7B4duAX6Nm/sZB8pNmTLFtcHhdlw4MLTGnV+4YCP7TABfXJB9bz179jS6evXqrg320rH3omzZskazfwmwAZehkM1swv1o1apV5nUupgn87Q9Jwt4IPr7XXXeda4PHDS7ayMV/v/zyS9cGexK5UCv3ve7du7s2OJCSPTzs6ypdurRro06dOkbffffdRrO3goM2Ad8Xr7rqKqOvvvpqo7lvAvY7/fnnn93r2eK3334zIX9c6DPkJ2X/CPe7Dh06GB06b3h84iLEN910k9Gvv/66a2PQoEFG8xhZokQJo/k4Ab4fsh+Lrxuh/eBxg8cB9ldysVkA+O6774xm3xf301AfCu1bCN3hEUIIIUTq0YRHCCGEEKlHEx4hhBBCpJ58eXimT59u1vV4/fLDDz9023AuAXt22LMQygxZu3ZtfnYTU6dOdf82fPhwo5NF4wDg1VdfNTqUw8L5CXPnzs3XfoVgr8E222xjNHuLAKB5cxsvwPkh7FUYNmyYayOZfxTKWMkWRYoUMX4C7h+cYwL4jBUuhsqf/5NPPnFtnHfeeUazp4FzHDiHBvC+CPZW8LELFWzMDc6xCBVb5CwL9mxxxg77cQBftJSL/P3vf/8z+sYbb3RtJLM+CtJ7AfzdB5IeC85MCWXXcNFd9gawR4X9KEA40ypJyLPDhNpNwkVnQzksL7zwgtF8HnGmzldffeXaaNmy5Sa34cKQIY8mn2vsOeT8Hy5yCtjMoFDuWLaoWbOmydVh3yYXhgV8NhFnGbG3KlSAdKeddjKav0POfwodf/Z5sX+Q+xhfq4GwNywJe7ryAnuJ1q9fb/SDDz7otmnYsKHRfI7tvvvuRofGd84v2+j+5eldQgghhBBbMJrwCCGEECL1aMIjhBBCiNSTLw9PTk6OyRCYMWOGeT20Pst1MObMmWM0r6uH/CS8XvnII48YPXr0aKO5thTg10V5/Zo9HaHaM507dzZ66dKlRrP/YtGiRa6N3ParZs2aRo8cOdJtw5kKHTt2NJrrqHANIcDW72FPUDaZOnUqWrdundH8+bmuDOC9YewT4LyI888/37XB39nYsWON5qwazmQBgI8//tjoK664wmjOM+IMkrzA+UChHCb2lyWzsQDvgwvVFuPviOvRcf//8ccfXRtlypTJ/FyQfQgAoigyGSdvvvlmrttwfhefn5xlwxk7BUXIs8Gw34j9JzwmhvKt2D/J32HTpk2NHjx4sGuD/WTc59mTd+mll7o2kp6N/NR2/K+sXLnSeJsmTZpkXg95QTmrhnN3uH4e5zYBPluL89vYjxLKFePj3bZtW/eewoA9O3khND4l4X4XqgvGvreNoTs8QgghhEg9mvAIIYQQIvVowiOEEEKI1KMJjxBCCCFST75MyxUqVDBFODlsLGQcHDp0qNFsluQiaKFwq+OPP97oJk2aGM0FzG699VbXRrLQIGCLZwJAr169jE4aMjfARrGFCxcafeCBBxqdNAZvjIMPPtjoUqVKGR0qpsqBdhxCVrduXaO5gBtgi7yx6TWbVKxY0RjV2aT+/PPPu204zIuNgf369TOaTduAN5g3atTI7VcSDqYEgHr16hnNIVqhgrP5hc3C7777rntP//79jea+zMX4OKgR8EZA7u9slGQTLGANqgUZXgn8bTJPHiM2x4bCNn/55RejOayPCxvzQwiAN7Oz0ZUfsAi1wWF1bA7OCxyKxwbkZGFXIFzYmLfZYYcdjObQvNB5xQ9McADkJZdcYjRfDwBrBA4VucwWxYsXN2GhXJSYH1IAfMHNXXbZxehx48YZzaGggA945HDR0AMT/5VQH+PixxxOyA8l5QUeN/laHTr+bMDnAq2PPvqo0fxgD+CDhTeG7vAIIYQQIvVowiOEEEKI1KMJjxBCCCFST748PMWLFzehP3vttZd5vX379m6bU0891Wj243C413XXXefaYH9AtWrVjN52222N5uJsgPf5sK+F2+DfAXjfA2suepYXDw+H6vF3yGFoAPDcc88ZzWFP8+bNM5oD8QC79l6QhR/LlCmDffbZJ6PZw/TNN9+4bdhjxGGWXNCvU6dOrg0ulslFHufPn290KMiKvTPsWWOPw4QJE1wb+YXX+wFg5513NpqPPxfCDe3HkCFDjOa19VCRP+a2227L/NytW7dc3785Wbp0qQlLZB9bKGiPCxtyQVku5Mv9CvAeDQ5s5HGFxzvA+8W42G/v3r2N5uKJeYGLDoeK8nJBUR57ObCUPT+AD5sdOHCg0Rz2eNFFF7k2kp4Vfn+2KVq0aObnESNGmNduuOEG9372X3Gx03vuucfo0LWIvTP8nmx4eEJFWdkf2K5dO6MbNGhgdMmSJV0bfA18++23jWbfXKgP8XXyySefNJqDGkNzBPb+XnbZZe49gO7wCCGEEGIrQBMeIYQQQqQeTXiEEEIIkXry5eGZO3euWaPkYpohv0nPnj2NfuKJJ4xmb01o7Y19PrwOzGutn3zyiWvj4osvNvrwww83mn0Qt9xyi2uD8wB4Pfeuu+5y2+QG5xawLyrkaUlmIQG+sCN/P6EsiKR36Prrr8/Tvm4O4jjG2rVrM/qzzz4zr4fWr2fOnGk0F/7kz899CvBFSY855hijOU8kmVO0AS6QyRkryUyPzUWo2CJnTuSW/RH6LMnCm4DP7mHvXdKvs4Hk9/xvigb+FypVqmSOWbJPAeFMMPY+5ZZnFSrc+vjjjxvNvj3+HZyxAvhMI/YshHJG8suLL75oNBfpBYAbb7zRaB6r+dxkrxFgPTCAH784D4hzawDr4+Nss2yycOFCkxtUokQJ83rZsmXdNjzWsGeF+9DNN9/s2uDzMXR+5gb7vNjXwgVI2a8GAK+//rrRPG5ydk+ogDB7MHlsvvbaa41mryTgs/g4D2q77bYzOnS9ChWdDqE7PEIIIYRIPZrwCCGEECL1aMIjhBBCiNQTxXGc5zc3a9YsTq5Zs3cgVMOK17QvuOACoz/44AOjOR8F8P6B9957z2jOHdkcNY0OOugg928jR440unv37kZzrsHXX3+d6+/hdXFeNw7lVnCNE649wt8pr6MCMHlKbdq0wTfffBPlurObgV133TVO5nu8/PLL5vVTTjnFbTNmzBijuaYVa65nBACHHXaY0VxvrHXr1kaHfGDsx+B+yZ6P8ePHuzZyg2upde3a1b2HfRKcmXPccccZzfkigM9qYs8aZ2OwVwOwnqXHHnsMv/76a4H0IQBo0qRJnPQy8ZjAWVWA92D16NHDaPZLhfwk3Nd4DGQfTJUqVVwbXF+Kjzl/lrzAeU2coRKql8e+Rs6Z4tpSl19+uWuDM2XYF8W/g71hgB2/DjnkkAIbi3bfffc4mUfF141Qhg77mnhM4DZC9ccmT55sdF7rQG1u+NrMteU4l4rHFcB7H7kWINenC42rbdq0MZrzsliHcta471arVu2rOI6b8/t0h0cIIYQQqUcTHiGEEEKkHk14hBBCCJF6NOERQgghROrJV/DgjBkzcP7552f0ggULzOvbb7+924bDCNnUx8X1QobTUKBhklCBxf8KG5RD9OrV6z//nlmzZhnNxfhChuPVq1cbzUbw5DECwuGFTZo0yfwcKi6aLX755RcT2MghUhzECHiTLhc95ICsMmXKuDb4O+DgPTaS7rrrrq4NNodziOa/MSkzvF9scAS8iZv3nY3vffr0cW2wiZsLsvK5HDI+J83xixYtcq9nk7lz55qgz7yEj/Lxeeutt4xmUzOHuwE+nK5Zs2ZGcwDivyn8+W+48sorjeagwVCQKofRcfjksGHDjF61apVrg4NC+cGV2rVrG33uuee6Np566qnMz2xQzSYzZswwpuo//vjDvB4y6XIoHptl2SzPxnjAPzDBAaZ8XeECtSE49HTu3Lm5bjN16lSjTzvtNKM5vJILowK+KCkfP374p3jx4q4NfoDiwAMPNJrN07yfgDc+bwzd4RFCCCFE6tGERwghhBCpRxMeIYQQQqSefAUP7rHHHnFy3Z7X+NhbAPg17l9//dXofffd1+hu3bq5NjhoL83suOOORrMfB/BrwOyLOvroo40uWbKkayMZuDVlyhSsXLmyQMK+qlSpEif9IxzuxUGEAHDNNdcY/cMPPxjNniYuNgd4Hwt/Rw0bNjQ6VAiWQxG33XZbo0OFPvMLFxtcvnx5rtvwfnBBy1DBRvYWcCFcLgaaDKrcQPv27TM/n3/++Zg6dWqBBQ9uu+228dlnn53RHIrXokULtw0HlHLwGvulJkyY4NrgsYiD1thLwb6YrZ1atWq5f0sGtBZkCOqOO+4Y33///eZ3J3nkkUfcNlz8kgv1cv/gIp6AD4Fs166d0RxWeMABB7g29txzz03uFxe6zosPiL2P7J0JjQGVK1c2mr18derUMXrw4MGuDR6becxj31QyuHYDf/31l9FXXnmlggeFEEIIsXWiCY8QQgghUo8mPEIIIYRIPfnK4ZkzZw5uuummjOaCgqECdYcccojRnFPBnoWtya8T4scffzT66quvznWbF154wWjOixk9erTbJuk94HXobFKqVCmTm3TwwQeb10OFExcvXmw05/Kw74uznQCfqcE5FVwoMpT1wLktoXyj/wrvx9tvv+3ew0VKc8tlCuVjDR061OixY8cazfkZoVyipL+soM/bYsWKoXr16hnNvqWqVau6bdgr0K9fP6OTYxsARJG3kuTmheBMkbTD/jDOUXvooYeM5usBAOy3336Zn2fOnLkZ927T5OTkGL8MF3KeM2eO24aLoXIOEechHXHEEa4N9iAeeuihRvO48swzz7g2+PycPXu20Xnx7DCcs8QeplCmFI9FXECbr02hnD0uGM7ZPpwZFiowHRqfQugOjxBCCCFSjyY8QgghhEg9mvAIIYQQIvXky8MTRRFKlSqV0RUqVDCvh2pasZ+gZcuWRg8fPjw/u5AnQjVwatasaXSXLl02++8tLHhNk+tvXXTRRW6bp59+OvNzQdZB+uOPP8w6L689hzKD2CvG9Ws4D6pixYquDa7hw54TzlxJZrxsgLMvNoeHh2szcU2o0PnB2UVc54rPw2Tm0gY6depk9OOPP240r6tzfS7A5kH99ttv7vVsUqVKFZx66qkZzblJodo67ONijwJnkrEfobBo3bq1+zfON+P8qocfftjoZNbNBpYuXWo01x9jPyGPoYD3qHCuWrVq1Yw+6qijXBtJH9Dvv//uXs8W8+bNMx4j9uwUK+Yvj+xZSeb4ALZGIWDrhG2McuXKGc05PJxdB/jMr3+T98TX5ttuu81oziZjvxLg86/Y48Sepp9++sm1cdVVVxl95plnGs3HIZQdyJlaG0N3eIQQQgiRejThEUIIIUTq0YRHCCGEEKlHEx4hhBBCpJ58FQ8tXbp0nDQHsRm2Xr16bhs2kCZNz4A3aG1JcJFKNlt9+umnbhsO3mKT6uagY8eORocM2snAv/333x/jx48vkIJ9e+yxR/z+++9nNBdH5QKOADBgwACjx48fb/S0adOM7tq1q2uDi+uxgZVD50JUqVLF6GOPPTbXbXIjaR4HgI8//thoLnIKeKM3h+y98sorRnPAHuBNr2xIZaNkKOwrWTz0nHPOwZQpUwqseGjVqlXj5HfDRTz//PNPtw2HXHJRQjY+vvnmm66NcePGbXK/eEzgcxEAmjZtajSbhdmUv2LFilzb4DBONguHwkd5X7lYJL/ORYsBb1K+/fbbjf7oo482+Tpgw2eHDRuGBQsWFEg/2n777eM777wzo2+99VbzOgfgAcADDzxgNIcGckjq9OnT871fF154odEh0zqf4xw+OmzYsFx/T6tWrYzmcYLHhA8//NC1wcZ3NrrzuMtFPoG/H2RJwseBx6JQYef77rvP6KZNm6p4qBBCCCG2TjThEUIIIUTq0YRHCCGEEKknX8GDtWrVMuFEV1xxhXn9pZdectucd955Ri9btiw/v/L/FFywjf1Hbdu2NXrUqFGujXPOOWeTbXCxPV7PBHxxUA4d4/CnUAG7bbbZJvMzF7PLJpMmTTL+Aw67ChXsK1++vNHPPvus0RzE9/PPP7s2uFhmkSJ2rj9x4kSj77nnHtdGo0aN3L/lF/69XDz37rvvNjq05j1mzBij2YfHXpSk12YDHObIRUrLlCljdCgAMRk2GPLMZJNVq1aZY7bjjjua14sXLx7cJgkXQ2Q/FZ/vAHDjjTcaPWjQIKP53ONQOcCH1bH/iH0gHCIIeF9e0hcHeA9PyINVp04do7mvse8nOWZsgEMvp06dajR78kJ+pGRQLPvzskmlSpWMl4+/91B4JZ9LHLh5+OGHG80eJ8AX1GS4T4VCI7lYbm6wxwXwBUZvueUWo5944gmjQ5427hOfffaZ0Tx29+jRw7XBfZPHrwYNGhgdKlLOnqaNoTs8QgghhEg9mvAIIYQQIvVowiOEEEKI1JMvD8/atWvNmiU/D798+XK3Da9Xli5d2uhQdk9BwB4FXvPu27ev24azWjiX4JhjjjGaM3cAX7SRMzfYw8HruQAwduxYoxcuXLjJ/TrrrLNcG8n1efZzZJOdd94Zr7/+ekYnfwaAlStXum3YB3b11Vcbfccddxj9yy+/uDa4iB3ng3AeUGjtnf1VvF6fF9avX2/0k08+aXT37t2NXrJkiWuD95ULFHJuB2dlAL5/s9eECwuG8oB69uyZ+bmgi4fWrl3b+BLYx/TVV1+5bfbZZx+j2ffDYxVniAD+fOTvjcc3zlQBgDPOOMNo9vGddtppRufk5Lg2eKyZMmWK0ew/Wbt2rWuDc3UGDhxoNPvrXnvtNdcGF/tdsGCB0ddff73RIY9O0m8VuoZkix9++MH4Lt977z3zOo8rgM9y2mOPPYxmr9tll13m2uDvlf1Y7J3Kr18H8DljoT7EfjT2PvI2oeLgJ554otHsY+Vsn1DuH/tv2JPJxY9DfsFu3boZ3atXL/ceQHd4hBBCCLEVoAmPEEIIIVKPJjxCCCGESD358vDMnz8fjz/+eEazNyCUfcHr4pw5cckllxgdypzIBlwTpUSJEkaH8k8474a9E+zP4FwHAOjQoYPRc+fONTqZSQEAu+yyi2uD/RLsr+D6XOwZ4Pece+657vVsEcex8RN07tzZvB7yE/Xp08do7kO87l+pUiXXBr8nWUMnBK9NA8DgwYM3uU1eKFu2rNHXXnut0V988YXREyZMcG38v/bOJLSqdIvCK1iKolFRoyRRS9TYg6BlG1sUgj22iIggggNB1IHiQHSgE0HEiYKIA5uZLXY4iMFuIELZi2KD2Ia0SIw95tagynD22lt9gXfzeKfWN8riJDv33vOf//zcf521O3ToYDSPIX5v0X4275NzPhD3J+OxD9jcLe7DlG0aGhpM/yDOd4n62LE3JulBArzX7cWLF64GX9MVFRVGs78sGs88PrkPVkFBwU+PAz7Pi/sa3rp1y+go+4XHHveg455t7FcBfEbK4sWLjZ4wYYLR9+7dczWSY7w5vWD5+fnGY8RZRpwpBPjrl69Pzu+Ksum6du1qNM+/3BfrV/3bIk6dOmV05I1i7xD3dONzu3r1aleD5x7u4cWZYVGuHI8Znps4H4vnph/VjdA3PEIIIYRIPVrwCCGEECL1aMEjhBBCiNSjBY8QQgghUk9OFAT0I4qKijJJA2kUzseUlJQYzcF6HH7FDcwAYP/+/UaPGTPG6D179hg9d+5cV4PNVNz0bNmyZUZz+BPgzbNsWj5w4IDRHA4GeMMtG+W4+Ro3eAO86Y0NtmyU5IaGgA34e/XqFT5//pzjfikL5OXlZZLG7enTp5vjUWggm+MLCwuNTjYjBf4OpWPY5MkhW+fPnzeaQyYBby4sLi52v/Mr+HWw6ZHNwWxiBvyY4eA6Dp2LDMcrVqwwms22HACZk+OHR35+fuPPR48eRWVlZbOMoX/+dyb5HniMcLAoAAwdOtTo3r17G83zGRu5AW/cZHMo12TDJeCbJR46dMhoDlG7c+eOq8HzAp9zngM4WBXwhtrXr18bzc1VeY4E/EMnHIrKc1MUgJc04ZeWlqK2trZZxlGrVq0y3bp1a9Rs7o/ub2z+/u03+9wPm8ejB1f4gRhuysnnKvrcOfT00qVLRrPRmY8D/v3yNcT3ZjbTA35uYRPzhQsXjOZ7JODv1/wAAv8NBxEC/p5QVlb2ZyaT+YN/T9/wCCGEECL1aMEjhBBCiNSjBY8QQgghUk+TggfLy8tNqBkHdSX39L/DDdlmzZplNAftRUFN8+fPN5r3hbnp3bFjx1wN9s5MnTrVaA5h4uaSUV0O4lqwYIHR7du3dzU4EIw9K7yPumvXLleD95Y5qIqDB7mBH2D3Wnfu3OmOZ4t3796Z/eQZM2aY41FQW/fu3V2NJB8/fjQ6arY3ePBgo3kfnYPzon1zDiLjIELeN+dzDQAPHjwwmr0z3IyRmxUC/nzzGPlPwt742uUxsG/fPqMjr1/SB8fBndmmvr7eBKU9fvzYHI9C0pJ+DQC4ffu20TyuouBB9nHx58JzEY87AOjcubPR27ZtM5r9ZEuXLnU12OvHTUtLS0uNjubE58+fG82NL7lJLftzAD+fd+zY0Wj2uEResGTQHIffZZPCwkLTyJMb97J3FADOnj1rNHun+DrgYFXAe06fPn1qNIfvjhgxwtVgnws37eT7HYcIAt5/xo2K2eMzc+ZMV4M9TfxaeS7iezngP+dJkyYZzdfct2/fXI01a9YYzZ6e7+gbHiGEEEKkHi14hBBCCJF6tOARQgghROppUg7PgAEDMsm9/S1btpjjGzdudH/DGROckcIeFm6+Bnh/Ae/PsU+IfSEA8OjRI6M5h2XlypVGc34A4Pc8OU+D96ujTBlu5Mk1R48ebTQ3G43+D+/5Dxo0yOgokyTZSPHatWuoq6trluyLfv36ZZJ71LtLo0toAAAE9ElEQVR37zbHI88Kn0/+3C9evGg07+cCwOTJk41mv8K8efOMjpq28mvlrJ6DBw8azVkZAHDy5Emj2TfCrzNqYMj74OwjGDJkiNGRj2DdunVGs++JX2eUJ5Pcv58zZw7u3r3bbDk8ffr0yWzfvr1Rs3+Os7oA73Ph98Tz2d69e10NzlFi3yL7+saOHetqJBswA34ssu9n0aJFrkb//v2NnjJlitGcGRblOfGY5/9bXV1tdOThYa9Ir169jH727Jn7GyY5P125cgVv375tlnHUo0ePTLJhKp9LnosBYMOGDUbX1tYaPW7cOKM5ywbw/tCamhqjuanruXPnXI3Nmzcbzd6/vn37/rQmAEybNs3oHTt2GM33lVGjRrka3ByU/YScAcevCwCWL19uNPsHOQ+LPXCAn+MWLlyoHB4hhBBC/DvRgkcIIYQQqUcLHiGEEEKknibl8LRt2xYjR45s1PxcfrRfG+37JeH8jGjPm3sFXb582WjeW2YfBOD3FtkrxPk/vJ8Z1WX/zf37943mXAfAZ3vwnjB7J9auXetqcLYB+6IePnxoNOcrALavUNSrJ1u8f/8e169fb9S8b15VVeX+hnMXysrKjE7uwwPxfjV7PDgvgnsTRfvE7AN78uSJ0fzalyxZ4mqw3yZ5PQE+L6Nnz56uxqZNm4xm/xnvvd+8edPVYO8F5w6x14zzggDrI3jz5o07nk1yc3MxceLERs2ZIREnTpwwmscRw+cG8J8l+7pyc3ONjuZE9hIVFRX9VEeZYJwjxT2t2DvJ5xvwmVA8brjGkSNHXA32sLC/kL0VkS8q6eHguSubtGjRwmSlDRw40ByPctTY88rzOfvAxo8f72p06dLFaM5e47FcXl7uanAWHc9FrKMcHv4/PF/xnMC91gB/3+TMHM7DevnypavRsmVLo9u1a2d0Q0OD0VGuGHunfoS+4RFCCCFE6tGCRwghhBCpRwseIYQQQqQeLXiEEEIIkXqaZFquqakxYWNJ4ysQBw9yYzw20A4bNsxoNqAC3kw1e/Zso5PNKAHfwA/wQU2rVq0ymt9L1DySTW0ciNemTRujv3z54mqwyY1D8zp16mR0ZHrkoMVkQ1fAm2sLCgpcjeR7ad26tTueLaqrq03gGxvSOBAP8E04+Xe4YWHUKJF/5+vXr0Zz6BqbAgEf3nb69GmjOQzu8OHDrsbVq1eNZkMqBwDW1dW5GpWVlUb/qtleFMLHJmMOSeQgM75eANvElptZZptPnz4Z021FRYU5zuFlAFBcXGx0fX290dzImM2kgDe382e/fv16ozkEFPDzFxvZmWRI6He4ieWHDx+MHj58uNFRCCY/yME1uLnq1q1bXQ021fMDJmzkZ7M1YOevyKCbLaqqqoyZm8Ma2YAM+HPHDVaPHz9udF5enqvBjT5LSkqM5mbI/FBK9No4EPPGjRtGczNZwF8j3JQ4uhczbEI+c+aM0fzgTjSW2YTM90B+8CkyPkfBqBH6hkcIIYQQqUcLHiGEEEKkHi14hBBCCJF6mtQ8NCcnpwrA8+y9HPE/4vdMJuM3m7OAxlBqabYxBGgcpRjNReK/QTiOmrTgEUIIIYT4f0RbWkIIIYRIPVrwCCGEECL1aMEjhBBCiNSjBY8QQgghUo8WPEIIIYRIPVrwCCGEECL1aMEjhBBCiNSjBY8QQgghUo8WPEIIIYRIPX8BtZpBPL3BuZgAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 720x720 with 16 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}