MNIST_classifier.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "MNIST classifier.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyOSHivdrLj0zna6odSkKCJ6",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Ananya-Jha-code/MNIST-classifier/blob/main/MNIST_classifier.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Importing Libraries"
      ],
      "metadata": {
        "id": "cgn-J4h1EWqr"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline \n",
        "##to display plot without plot.show\n",
        "import keras\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense, Dropout\n",
        "from sklearn.metrics import confusion_matrix\n",
        "import seaborn as sns\n",
        "np.random.seed(0) \n",
        "##for constant result without being affected by randomness"
      ],
      "metadata": {
        "id": "1y9d0Um2EVZF"
      },
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Load Data"
      ],
      "metadata": {
        "id": "zWZDKdsyFzxW"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from keras.datasets import mnist\n",
        "(x_train, y_train), (x_test, y_test) = mnist.load_data()"
      ],
      "metadata": {
        "id": "2t-IlNT1F4P1"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Shape of the dataset"
      ],
      "metadata": {
        "id": "PQjOc_BsHPMV"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(x_train.shape, y_train.shape)\n",
        "print(x_test.shape, y_test.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "chQoOXkIGERs",
        "outputId": "89aafeb0-921b-4c25-b663-506742ef0df9"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(60000, 28, 28) (60000,)\n",
            "(10000, 28, 28) (10000,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "60000 images for 28X28 dimensions and 10000 labels"
      ],
      "metadata": {
        "id": "cZN22oP-HwZY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "num_classes = 10\n",
        "f, ax = plt.subplots(1, num_classes, figsize=(20,20))      #plot 1 row\n",
        "\n",
        "for i in range(0, num_classes):\n",
        "  ax[i].set_title(\"Label: {}\".format(i), fontsize=15)\n",
        "  sample = x_train[y_train == i][0]\n",
        "  ax[i].imshow(sample, cmap='gray')\n",
        "  "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 160
        },
        "id": "M1lgtcpdG_CF",
        "outputId": "5850e99a-158f-4e54-91cc-0f13677bbc86"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x1440 with 10 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "dataset already balanced "
      ],
      "metadata": {
        "id": "jutwW8QmLRiz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "for i in range(10):\n",
        "  print(y_train[i])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GjAwclqgLVYP",
        "outputId": "d887b38e-8be6-4f58-8b83-6072012162df"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "5\n",
            "0\n",
            "4\n",
            "1\n",
            "9\n",
            "2\n",
            "1\n",
            "3\n",
            "1\n",
            "4\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Labels are a continuous dataset. Convert to discrete to get one digit result "
      ],
      "metadata": {
        "id": "ARZ-iN7KLbf5"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "y_train = keras.utils.all_utils.to_categorical(y_train, num_classes)\n",
        "y_test = keras.utils.all_utils.to_categorical(y_test, num_classes)"
      ],
      "metadata": {
        "id": "ZXOEHw8ELnG2"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for i in range(10):\n",
        "  print(y_train[i])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9DSRJFnfNLJg",
        "outputId": "6eb2d7bc-ea01-45d2-bb94-c1444b3a1fcc"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
            "[1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "[0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
            "[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "[0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
            "[0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
            "[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "[0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n",
            "[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "[0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "forms a vector of 10 elements where 1 represents the number and 0 fills the blanks"
      ],
      "metadata": {
        "id": "CdMWPTJfOI22"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Data Preparation"
      ],
      "metadata": {
        "id": "xvrefjBVQCiM"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Normalization: probability of whites needed"
      ],
      "metadata": {
        "id": "L7O8Hk8uQG_2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Normalize Data\n",
        "x_train = x_train / 255.0\n",
        "x_test = x_test / 255.0"
      ],
      "metadata": {
        "id": "0XgYg9o8NpBL"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Reshape Data\n",
        "x_train = x_train.reshape(x_train.shape[0], -1)\n",
        "#-1 makes it one dimensional\n",
        "x_test = x_test.reshape(x_test.shape[0], -1)\n",
        "print(x_train.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cb7UL801Np5P",
        "outputId": "c3de769f-c55e-4d40-b90e-4cd89effe18d"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(60000, 784)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Model"
      ],
      "metadata": {
        "id": "MyDwo80PQz3p"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model = Sequential()\n",
        "#adding chain of layers\n",
        "model.add(Dense(units=128, input_shape=(784,), activation='relu'))\n",
        "model.add(Dense(units=128, activation='relu'))\n",
        "model.add(Dropout(0.25))\n",
        "#25% of the dataset will be deactivated during training which will prevent overfitting\n",
        "model.add(Dense(units=10, activation='softmax'))\n",
        "#softmax assigns probability for each class to be correct\n",
        "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "w23yoVVgNsH9",
        "outputId": "105c7a1f-c7e9-4f5d-a332-486fb5dbc4ac"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense (Dense)               (None, 128)               100480    \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 128)               16512     \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 128)               0         \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 10)                1290      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 118,282\n",
            "Trainable params: 118,282\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Training\n"
      ],
      "metadata": {
        "id": "LSO5zJt-TSHS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "batch_size = 512\n",
        "epochs=10\n",
        "model.fit(x=x_train, y=y_train, batch_size=batch_size, epochs=epochs)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CcE3pXyKNuqK",
        "outputId": "90124ad8-1bdd-4b7b-cac0-b5c1fd9a50f1"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "118/118 [==============================] - 2s 9ms/step - loss: 0.6081 - accuracy: 0.8225\n",
            "Epoch 2/10\n",
            "118/118 [==============================] - 1s 9ms/step - loss: 0.2280 - accuracy: 0.9339\n",
            "Epoch 3/10\n",
            "118/118 [==============================] - 1s 9ms/step - loss: 0.1665 - accuracy: 0.9514\n",
            "Epoch 4/10\n",
            "118/118 [==============================] - 1s 9ms/step - loss: 0.1307 - accuracy: 0.9613\n",
            "Epoch 5/10\n",
            "118/118 [==============================] - 1s 9ms/step - loss: 0.1097 - accuracy: 0.9672\n",
            "Epoch 6/10\n",
            "118/118 [==============================] - 1s 9ms/step - loss: 0.0922 - accuracy: 0.9725\n",
            "Epoch 7/10\n",
            "118/118 [==============================] - 1s 9ms/step - loss: 0.0785 - accuracy: 0.9766\n",
            "Epoch 8/10\n",
            "118/118 [==============================] - 1s 9ms/step - loss: 0.0691 - accuracy: 0.9786\n",
            "Epoch 9/10\n",
            "118/118 [==============================] - 1s 9ms/step - loss: 0.0596 - accuracy: 0.9819\n",
            "Epoch 10/10\n",
            "118/118 [==============================] - 1s 9ms/step - loss: 0.0531 - accuracy: 0.9838\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7f8fc70da210>"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Evaluate"
      ],
      "metadata": {
        "id": "8fAS3hC9X8G0"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "test_loss, test_acc = model.evaluate(x_test, y_test)\n",
        "print(\"Test Loss: {}, Test Accuracy: {}\".format(test_loss, test_acc))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zH86C1avNx7s",
        "outputId": "d17c8e5d-88fe-4c01-f7a9-fcae3a5cbbd5"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 2ms/step - loss: 0.0733 - accuracy: 0.9772\n",
            "Test Loss: 0.07328207790851593, Test Accuracy: 0.9771999716758728\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "y_pred = model.predict(x_test)\n",
        "#an array of all the prediction\n",
        "y_pred_classes = np.argmax(y_pred, axis=1)\n",
        "print(y_pred)\n",
        "print(y_pred_classes)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "u0U_puxVN0_n",
        "outputId": "b5c619e6-7f98-4d2a-cf6e-7af3a5a6f1f3"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[[4.9467707e-07 1.0835869e-05 9.8418452e-05 ... 9.9798489e-01\n",
            "  2.0186415e-06 1.1873513e-04]\n",
            " [6.6183939e-06 5.0784298e-04 9.9948263e-01 ... 8.9993444e-09\n",
            "  6.5282836e-07 4.6634024e-10]\n",
            " [6.5107965e-06 9.9906224e-01 1.1205111e-04 ... 2.3488476e-04\n",
            "  1.8433329e-04 1.5631060e-05]\n",
            " ...\n",
            " [5.9758157e-11 4.3746218e-10 1.4243479e-09 ... 1.7599189e-07\n",
            "  2.9111240e-08 1.9845906e-05]\n",
            " [1.7380618e-08 1.2686619e-08 4.7714083e-11 ... 8.3790921e-09\n",
            "  1.6688189e-05 5.4200386e-09]\n",
            " [3.9815887e-08 5.7376909e-11 2.1925175e-09 ... 7.0925319e-13\n",
            "  2.3430069e-10 7.0776163e-12]]\n",
            "[7 2 1 ... 4 5 6]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Single Example\n",
        "random_idx = np.random.choice(len(x_test))\n",
        "x_sample = x_test[random_idx]\n",
        "y_true = np.argmax(y_test, axis=1)\n",
        "y_sample_true = y_true[random_idx]\n",
        "y_sample_pred_class = y_pred_classes[random_idx]\n",
        "\n",
        "plt.title(\"Predicted: {}, True: {}\".format(y_sample_pred_class, y_sample_true), fontsize=16)\n",
        "plt.imshow(x_sample.reshape(28, 28), cmap='gray')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        },
        "id": "QfeDeCHiN3j7",
        "outputId": "20887c6a-c121-43ef-dcbf-03952e99372f"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f8fc56924d0>"
            ]
          },
          "metadata": {},
          "execution_count": 16
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Confusion Matrix\n",
        "network accuracy for every class"
      ],
      "metadata": {
        "id": "dfKeWEuwcAFv"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "confusion_mtx = confusion_matrix(y_true, y_pred_classes)\n",
        "\n",
        "# Plot\n",
        "fig, ax = plt.subplots(figsize=(15,10))\n",
        "ax = sns.heatmap(confusion_mtx, annot=True, fmt='d', ax=ax, cmap=\"Oranges\")\n",
        "ax.set_xlabel('Predicted Label')\n",
        "ax.set_ylabel('True Label')\n",
        "ax.set_title('Confusion Matrix');"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 621
        },
        "id": "4zHN9boDN63r",
        "outputId": "1346e6fc-93cf-497c-c30d-b015d686e068"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x720 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "most of the classes are accurately classified"
      ],
      "metadata": {
        "id": "sG0ENL9Bc_1h"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "errors = (y_pred_classes - y_true != 0)\n",
        "y_pred_classes_errors = y_pred_classes[errors]\n",
        "y_pred_errors = y_pred[errors]\n",
        "y_true_errors = y_true[errors]\n",
        "x_test_errors = x_test[errors]"
      ],
      "metadata": {
        "id": "ypUCZnkkN9q3"
      },
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "y_pred_errors_probability = np.max(y_pred_errors, axis=1)\n",
        "true_probability_errors = np.diagonal(np.take(y_pred_errors, y_true_errors, axis=1))\n",
        "diff_errors_pred_true = y_pred_errors_probability - true_probability_errors\n",
        "\n",
        "# Get list of indices of sorted differences\n",
        "sorted_idx_diff_errors = np.argsort(diff_errors_pred_true)\n",
        "top_idx_diff_errors = sorted_idx_diff_errors[-5:] # 5 last ones"
      ],
      "metadata": {
        "id": "x50H89DpOAa-"
      },
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Show Top Errors\n",
        "num = len(top_idx_diff_errors)\n",
        "f, ax = plt.subplots(1, num, figsize=(30,30))\n",
        "\n",
        "for i in range(0, num):\n",
        "  idx = top_idx_diff_errors[i]\n",
        "  sample = x_test_errors[idx].reshape(28,28)\n",
        "  y_t = y_true_errors[idx]\n",
        "  y_p = y_pred_classes_errors[idx]\n",
        "  ax[i].imshow(sample, cmap='gray')\n",
        "  ax[i].set_title(\"Predicted label :{}\\nTrue label: {}\".format(y_p, y_t), fontsize=22)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "CmZ09PjEOC7A",
        "outputId": "f0dbc4f5-f802-460d-f3f3-f53c422b0155"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABqwAAAFtCAYAAAByR0s+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeZglZXk3/vuGGdkVZHEBXhXXmITIIiqo9IjIojGoAVE0GncFNYrh5+s2g/mRuIsLKqI4KC6IyiIRRWFGxB0MiWBGERwCyKIQFAFxhnneP0619DS9PN19qqv79OdzXXWd7qq76nnOmenv1PR9qk6WUgIAAAAAAAC6skHXEwAAAAAAAGBh07ACAAAAAACgUxpWAAAAAAAAdErDCgAAAAAAgE5pWAEAAAAAANApDSsAAAAAAAA6pWE1DZm5OjPLqOWPmfmrzPx0Zj6q6zkOG57fGOuHn8MDZ39WM5eZy5r5L5vCPi9s9lnepzmsbI431I/jTTLWA5uxVrc8zn0z88ZmrD+0ORZMRM52T8725Zj/JzNfkZmnZ+b/ZOafMvOWzPxJZr4tM+/Zr7FgOmRt92Rta+M4p2VOkLPdk7OtjSNnmRPkbPfkbF+OuWFmHpyZ78zM8zLzd80Yl/RrjPlEw2pmvhERJzXLORGxcUQ8PyJ+nJmHdjmx2TKdUGJOOz4itup6EjCCnJWz89nnIuKjEfHUiLguIr4SEd+PiAdHxNER8V+Z+aDupgd/Jmtl7aBxTstcI2fl7KCRs8w1clbOzmdbRMQXI+KoiFgSEQv6za2Lup7APPeOUsrK4W8yc5OIOCEiDouI4zPznFLKTV1NbhL7RMTiiLim64kwN2TmP0TE0yPiuIg4vOPpwDA5y3x2TUS8LiI+U0q5cXhlZm4bvZPRoYhYHhF7dzE5GEHWMjCc0zJHyVkGhpxljpKzzGdrIuLkiLgoIi6MiHtFxFmdzqhDrrDqo1LK7RHxyoi4NXqd0P26ndH4SimXl1JWlVLWdD0XupeZ94+ID0TEjyLi/R1PB8YlZ5lPSinPLqUcO7JZ1az/TfTe7RcR8cTM3HH2Zwfjk7XMV85pmS/kLPOVnGW+kLPMJ6WUW0spz29+f3BB9P7eLlgaVn1WSrklIn7RfPuAiPXvbZmZizLzDZn5n5l5a2bePHL/zHxMZn4hM6/O3mdd/CYzz8zMx483Zmb+dWaelpk3Ncf8SWa+ZKJ5TnR/1Ow5JDPPzswbmnlck5nnZuarR9SViFjafLt01P1il4065maZeVRm/jgzf5+Zt2fmpc3lqpuPM8fFzWv1s+b+s9dl5mcy8wETPbfpyMxnZeaJzZxubsb7ZWYeV/OLxMxckpnfysz/zcw/ZOYFmfn0CeozMw/NzHMy87eZeUf2Pt/khLH+TGbBxyNi04h4UUTc2cH4UE3Oytl5mrPrKaVcHRG/bb7docu5wFhkraydp1nrnJZ5Q87KWTkL7ZKzcnae5iylFMsUl4hYHRElIobG2f7LZvvrmu8f2Hx/ZUScERF3RMQ3I+LzEfHdEfsdGRHrmuXCiDg1ep91cWezvHSMsfaOiNua469qjrmyqX9fs75M8BweOGr9PZo5lohYGxEXRO8zOM6NiOtHHit6tzG6uKm9uPl+eDloRN0OEXFpU3dD9O4re0ZE/LpZ958RsdWoeWwQEV9ttt8eEV+LiFOafX4bvXvSlohYNoU/txc2+ywfY9vaiPhD9N4l9KVm7Kua+t9ExMPG2Gdls/0Dzev9X81r9d3h1z0iXj/Gfosj4svN9tsi4vzmz/q/m3U3RcTuo/YZ/ju0eop/V1842X4R8Y9NzdtGjfWHrn/WLAt3CTk7XLs85Oy8z9lx9tumeU3u9nfEYpmtJWTtcO3ykLXzPmvDOa1lDi4hZ4drl4eclbMWSwtLyNnh2uUhZ+d9zo6oHWpqL+n6Z6yLpfMJzMclJgjDiHhU84NRImJJs274L/JwID5kjP0OaLZfExGPGbVtr4j4XUT8aeQPZURsEhFXN/v9a0TkiG17R+/ywamG4bHN+p9HxCNGbdswIp4+at2yiUIpIjIivtfUfCgiNhk1/8+MFVAR8epm/dUjX6/ofWjil0a8nmOOO85cXjjWWM22QyJi01HrFkXEvzT7nD3GPitHzOMNo7b9bfTuP7o2InYete0dzT7fjogdRm07otn2y4hYNGJ9K2EYvX+obo7eP0iLR43lpNPS2RJyduS6ZRPlXcjZOZ2zE+w3PMeLuv55syzcRdaut07WzuOsDee0ljm6hJwduW7ZRHkXclbOWizTWELOjly3bKK8Czk7p3N2VO1QaFhZpvSijRGGEbFV9D508opm239ExAbNtuG/yCUinjvOMX/YbD9gnO1vaLa/d8S654/4wdlwjH3eO5UwjIjtovfOgjsj4i8rX4vJwnA45L8//HqM2r5Z9N4VsCZGdPAj4vJmvxeOsc92cdc7FsYcd5y5jBuGk+x3TfOabDFq/XAY/nic/YbfYXDCiHX3buZ+S0RsN85+ZzX7/e2IddMNw2dE710d546z/ezoBfZuY4zlpNPS2SJn1zuOnJ3HOTvOPk9usvfOiNi7Xz83FstUF1m73nFk7TzO2nBOa5mji5xd7zhyVs5aLH1f5Ox6x5Gz8zhnR9UOxQJuWPkMq5lZMXw/0OhdInhGRDwoIn4Svcst142xz2mjV2TmNhGxR0T8PiLOGWesbzePjxuxbu/m8QullLHuHfyZyZ/Cep4UvctNv19KuXSK+47nwObxy2O9HqWUW6N3ae2iiHh0RERm7hARO0XvstvPjbHPDTH+6zRtmfmwzHxNZn6wuVfq8sxc3sxtg4h4yDi7fnac9cOv/9CIdUui966FbzfPYyxj/VlPSynltFLKI0op+4ze1txDd/+IeE8p5aKZjgUtkbOTk7NzNGfHkpl/Hb1bDGwYvdupfHuSXWA2yNrJydo5mrXOaZkn5Ozk5KychZmQs5OTs3M0Z1nfoq4nMM99IyKua76+I3r37vxORKwoTTt0lBtKKbePsf5BzeM9I2JtZk405rYjvh7+kPZfjVO7eqIDjeEBzeOqKe43kZ2ax3dn5rsnqR1+bsPP69ellD+NU7t6phMblpmLIuIjEfGS6F0eO557jrN+std/hxHrhl+Ppzb/iE5k20m2T1vzIYXvjd5lxcvaGgf6QM5OTs7OwZwdS2Y+IiK+FRFbRu/deMfM5vgwAVk7OVk7B7PWOS3ziJydnJyVszATcnZycnYO5ix3p2E1M+8opaycQv1YQRjRe5d1RO8eqKdPcozfTmG8qZrsh3M6hp/bt2PyALuyhfFrvDYiXhq9f8xeH737ud5QSrkjIiIzvxe9TvqE/0pVGn49fh4RP5ik9od9GG88+0Qv3G+MiK+P+gd44+Zxk8xc2Xz9llLKBS3OB8YjZycnZ9c3V3J2PZn5sIg4L3q3TDiulPKG2RobKsjaycna9c2VrHVOy3whZycnZ9cnZ2Fq5Ozk5Oz65krOMoqG1dxwVfO4ppTywinsd03z+MBxto+3fjz/0zw+fIr7TWT4uZ1aSjmucp/h53X/zLzHOB38B854Znc5uHl8eSnlrDG2j3eZ6WRzGV5/zYh1w6/HT6f4Z92WB8Vd7x4ZbYO465LmbWZnOtAaObs+OTuLMvOhEbEiIu4XESdE70NrYRDJ2vXJ2tnjnJaFQs6uT87OHjnLQiFn1ydnmXU+w2oOKKVcExE/jYhtMnNoCrsO30fz0MzccIzth01xKudF74P19szMv6jcZzioxmt+nt08HjzO9rsppVwVvUs4N4iIQ0dvz8xtI2Lf2uNVuHfzeNXoDZm5b0x+2ed4r/Pw+pUj1n0req/xkzNzyynMsa9KKctLKTnWEnedhN46Yv1k7yqBOU3Ork/Ozp7MfHD0mlX3j4hPRe/Eu413y0HnZO36ZG37nNOy0MjZ9cnZ9slZFho5uz45Sxc0rOaOtzaPJ2fmU0ZvzMwNM/NJmfnYEau/FBHXRq+7vCxHXJudmY+PiFdOZQLNB8x9LHp/L77c3L5o9Bz+dtRuw53p8cLz9Ii4KCL2zsyPZea9Rxdk5n0z86WjVn+wefz/M3OnEbUbRcRxEbFpzXOqNHw/2Fdm5p9/JppfMn6sYv9HZ+brRq7IzAMj4nkRcWdEfHh4fSnl+ujNf8uIOLP5PJMYte9mmfnczLzPlJ/J3Y/1jMxclZnnzvRYMADk7PrkbMs5m5kPil6zavuIOCkiXqJZxQIga9cna53TQr/J2fXJWTkL/SZn1ydn5ezsKqVYprhE7z6fJSKGKusf2NSvnqTu9RGxtqn9eUScGRGfi15X/X+b9a8Ytc+Tonff1RIR/93Ur4jeD+H7mvVlgufwwFHrN4qIf2+2rYmI85tjfisirh99rIi4b0Tc2tSfH713j38iIp4+omaHiPivpub30fvQw89FxFci4pKIWBcR14067oYR8bVmn9ubOZ0SvfC9MXq/+CsRsWwKf24vbPZZPmr946L3LoTh1/0LEXFO9D6kcWVEfHesP+9mW4mIDzSv9382z+uC4dc9Iv55jHksbp5Laf68L4yILzbrfhgRf2y2PWKqf4cmeM7V+40Y6w9d/6xZFu4ScnZkvZydpzkbET9ptv0xIj4dEcvHWR4xlTEtln4tIWtH1svaeZq1FX9fndNaOltCzo6sl7Ny1mLp+xJydmS9nJ3HORsRH4neZ2n9ICJ+1tTeNmLdD6L3JtjOf+7aXjqfwHxcoqUwbGr/pgmTXzYhcEtE/CIizoiIl0TEvcfZ54zoBeZtEXFxNKEZUwzDZtsG0es8nxsRNzVBcXVEfDMiXjVG/ZLoBfDNTbDdLaSi92Gch0fvEtnhY17bBMG7I2LPMY67OCL+v+iF/B0RcUNEfD56l50vG2ucSV7b4WBYPs5reFZEXNe87j+LiLdF7x+HlWP9eY9cHxFPjt4/Wr+L3j8O342IgyaZz99G790Nv25ejxuj94/DpyLioIhYPJ2/Q+M85+r9wkmnZQ4scvZu9XJ2HubsiL8Dky1Vf88tln4vIWtH18vaeZi1FX9fndNaOlvk7N3q5ayctVj6usjZu9XL2XmasyOew0RL9Ws8n5dsXhAAAAAAAADohM+wAgAAAAAAoFMaVgAAAAAAAHRKwwoAAAAAAIBOaVgBAAAAAADQKQ0rAAAAAAAAOqVhBQAAAAAAQKc0rOawzCzTWJZ3Pe9ambm6mfMD+3CsoeZYK2c8sbrxljfjvbDPx90hMz+UmT/PzNsz84+ZeVlmfiwzd+rnWECPrJ3SseZ11o6Yf83yf/oxJiBnp3iseZ2zo469SWYelZk/zsybM/O2zPxVZp6amXv1ezxYyOTslI41EDmbmTtm5ocz8/LMvCMzf5uZ38jMp/ZzHOAusnZKxxqIrG2O7Zx2li3qegJM6KQx1t03IvaLiFsj4ktjbL+g1RnRmszcJSLOi4gtI+LqiPhGs2n3iHh5RByWmfuVUr7X0RRhUMnaheO6GPvPe9geEfEXEXF5RFw1KzOChUHOLjCZ+aCIOCciHhIR10bEiohYGxEPiIiDIuI/I+K7nU0QBo+cXUAy89ER8fWIuHdEXBkR/x69P+8nRcRTMvPtpZSlHU4RBpWsXWCc03ZDw2oOK6W8cPS6zByKXhD+dqztzGvHRa9ZdUJEHF5KWRMRkZmLI+JjEfGiiPhoRPxNZzOEASRrF45SyqqIeOF42zPzZ82XJ5ZSyqxMChYAObuwZOZmEfHNiNgpIt4YEe8ppdw5YvvWEbF1R9ODgSRnF47M3Dgivhy9ZtWHIuL1pZS1zbY9o9e8eltmXlBK+WZ3M4XBI2sXFue03XFLQJgDmpPOxzXfLh1uVkVENF+/pfl258zcdLbnBzDoMvNx0bu66s6IWN7tbADmtbdExIMj4rhSyjtH/sc+IqKUcmMp5RfdTA1g3ntGROwYvTsCHDncrIqIaO7Gckzz7ds6mBvAIHFO2xENqwEy8n6dmblzcy/N6zLzzsz8p9E14xxjWbN92TjbH5OZX8jMqzPzT5n5m8w8MzMf38fn8YDM/L+ZuSIzr2rux3xT8/1zK/bfLDPfkZlXNPtelb3PhRq3693c//kDeddnR/0+M7/bvJbZr+c2gTujd0npZG6NiNtbngswAVn75/3nY9ZO5EXN49dLKb/udCawwMnZP+8/73I2M+8RES9tvn1f2+MB0yNn/7z/vMvZiHh08/jtkW90HeGc5nGvzLzvLMwHGIes/fP+8y5rndN2S8NqMO0VET+KiF0jYmX07m1820wPmplHRsT3I+KQ6H0OyBkR8cuIeGpEfDszXzrB7lPx/Ij41+i9a2hVRJwWET+LiCdExGcz84MT7HuPiDg3Io6IiEsi4qsRsXHz/fcz8z6jd8jMJRHx04h4TfR+Jr4eET+MiJ0j4lMx8eed3E3e9SGFL6zdpznRPLf59ujs3QZw+HiLI+Jfmm8/6TZVMGfI2nmWtRMca9OIeHbz7Sdnejygb+Ts/MvZ3aJ3a5RrSim/ysxdM/NfMvP4zHx7P395AvSFnJ1/Obt58/jbcbYPr8/o/bkC3ZO18y9rndN2yGdYDaaXRO8y8LeVUtb144CZeUBEvCcifh0Rzyyl/HDEtr0i4msRcVxmfrsPl0N+IyJOK6VcOmoOD41eyL06Mz87cg4jPC4ifhERDy+lXNPst0X0wnSf6N3j+ZARx7xf9O7/vHn0Ptfk08MNoczcMSLOjIjnZ+Z5pZTlM3xek3lV9EL4pRFxQGZe2Kx/dERsFRHHRsRRLc8BqCdr52fWjuXgiNgiIm6IiLM6GB8Ym5ydfzn7183jNZn5nog4ctT2t2bm6RHxvFLKrS3OA6gjZ+dfzt7QPO40zvYHj/j6QS3OA6gna+df1jqn7ZArrAbTquh9DlJfQrCxrHl8yegAKqV8N3pXAC2OiJfPdKBSyo9Hh2Cz/rK460qjv5/gEEcOh2Cz3y0R8Yro3XbvWU3ADfun6DWD3ltKOWnk1UullKvirss/Xz2Fp3B5RPw8In43hX2ilHJFROwZEWdHxA4RcVCzbB+9dy58Z5xL/oFuyNp5mLXjGL4d4KflLMwpcnb+5ey9m8ddovcf+2Mj4iHN3P4uIq6J3vntR6ZwTKA9cnb+5ex5zeNTM3OHMba/csTX95zCcYH2yNr5l7XOaTvkCqvBdMboD4KbiczcJiL2iIjfx133Qx7t283j4/o05sYRsV/0ri7aNiI2ajbdr3l82Di73lxKudu740spv8zMH0TvMtwnRsRnm00HNo+njnO8iyLiDxHxqMzcuJTyx8nmXkrZZ7KasWTmnhHxlei9zn8XEd9rNu0VEe+NiC9n5tJSytunc3yg72TtKPMha0fLzIdEb64RESf245hA38jZUeZBzg6/IXJxRJxcSnndiG1nZuavo3dLnOdn5ttLKZdPYwygf+TsKHM9Z0sp52Xm+c3cvpmZR0QvV+8TEa+NiOdExJro5XA/fzkOTJ+sHWWuZ204p+2UhtVgurLPxxu+jPyeEbE2J/5su21nOlhmPi4ivhi9q4zGM947hVZPsM/q6AXhyOMOX0b/40meV0Rz79LJiqYjM7eMiNMjYrOI2LO52mrYGZl5aUT8V/QuOf188y4GoFuydvxtczJrxzF8ddX3Syn/PYvjApOTs+Nvm6s5e8uIr08YvbGUcmFmXhQRu0fE3tF7xyvQHTk7/ra5mrMRvdtZfyV6c/zWqG3HRsTjo5ezN7U4B6CerB1/21zNWue0HdKwGky3z2DfsW4TuWHz+LvoNVUmMt4Hf1bJ3gffnxa9dwd9MiI+Gr0PDLyllLIuM58SvXunTppalYaf2ykRMVlX/o4+jTmWp0bvH5HzRjWrIuLP7zz4YUQMNYuGFXRP1tabK1m7nszcMCL+ofn2k7M1LlBNztabKzn7q3G+Hl2ze0Tct8V5AHXkbL25krNRSrkhM58QEU+OiCdF75e210fvKo4Lm3f+R0T8tM15ANVkbb25krXOaTukYbXw/Kl53Hyc7Q8YY91VzeOaUsoL+z6j9T0xeiF4USnlJWNsf8gk+z+wYtvI7vtVzTH/Zaz7sc6i/9M8TnQ/1Zubx3tPUAPMDbJ2bmbtaPtF73MC/xC9E2Jg/pCzczNn/2PE11vHXa/5SNs0j39ofzrADMjZuZmzERHRfK7LN5vlzzLzwdG7RdeNEfGTDqYGTI2snZtZ65y2Q2N1aRlswyHwiNEbMnOT6F29s57mg/F+GhHbZObdtvfZcDNmrCCIiHjuJPtvmZkHjl7ZnLQ9NiJKRJw/YtPZzePBU5lkC4bfAbVbZi4evbFZt1vz7XidfWDukLVzM2tHe3Hz+MVSipNMmF/k7BzM2eY1Hv7g77t9XkBmbhURuzbfXjhb8wKmRc7OwZyt8Ibm8eOllD9NWAnMBbJ2Dmatc9puaVgtPOc2j8/PzIcPr2xC8KNx15U+o721eTy5udxzPZm5YWY+KTMfO8P5rWoen5SZfw7rzNwgM98WvXubTua9mTn8oX+RmZtHxEeid1npaaWU/xlR++7ofUjhmzLz8My821WHmfmXmfnM2ieQmedm5qrMfEbtPtEL5Nui9/q/PzOHP7wwmq8/GBE7RsT/Ru9SW2Buk7VzM2tH7r9NRPxt863bAcL8I2fnbs4e0zy+KTN3H3G8jaP3Z3Ov6H1g9veneFxgdsnZOZqzmfnXmbnZqHWLMvPNEfHy6N2u65gxdwbmGlk7R7M2nNN2xi0BF5hSygWZeVZEPC0ifpKZ34mItdG75+a6iPhURPzjGPudkZlHRsS7IuIbmfmLiPh59C57vG9E7BIRW0bEKyPiBzOY309GzO/izFwRvdvkPTp6If2uiDhqgkN8P3qB94vMPC96l9buHb3Ph7o8Ig4fNd5VmXlQRHwpIj4cEW/OzEsj4obm+fx19BpFp0TvQ01rPDh6l+zeq7J++B7Ur4reL00Pj4hnZObw5fu7Re+S/jsi4kWllIluGwjMAbJ2bmbtKM+PiMURsaqU8r1pHgPoiJyduzlbSvlqZr43Io6MiO9l5g+id2uqPSLi/tF7J/FzmttZAXOUnJ27ORu9fD24+Z3BNRGxSfSuVNguep93/ZRSyq1TPCbQAVk7d7PWOW13XGG1MB0cEe+I3g/7k6J3CeO/N4//M95OpZT3Ra958snohc2+0Xt3+g7Ru3zzpRHxxT7M71kR8cbovStoKHqXXl4aEY+Puy4NHc+fovecjo+InSPi6c264yLisaWU60bvUEpZERF/GRH/Gr3X5LHNHP4yIq6IiP8bEW+e4XOaVCnlpOiF3meaOe/bLLdH7zXftZQy2YcpAnOHrB1hrmTtCMMn/SfO4phAf8nZEeZSzpZS3tCMfUH0frFwYPTuJvC+iNillHLZbMwDmDE5O8IcytnTI2JF9D7/5aDo/fJ3dfRuB7hzKWX1LMwB6B9ZO8IcylrntB1JTUAAAAAAAAC65AorAAAAAAAAOqVhBQAAAAAAQKc0rAAAAAAAAOiUhhUAAAAAAACd0rACAAAAAACgU4tmc7DMLLM5HsAovy2lbNv1JNokZ4GODXzORshaoFullOx6Dm2Ts0DHBv6cVs4CHRs3Z11hBSwkV3Y9AYABJ2cBAJjvnNMCtGvcnJ1Rwyoz98/Mn2fmLzPzjTM5FgBjk7UA7ZKzAO2SswDtk7XAIJh2wyozN4yI4yLigIh4ZEQ8JzMf2a+JASBrAdomZwHaJWcB2idrgUExkyus9oiIX5ZSriil/CkivhARf9efaQHQkLUA7ZKzAO2SswDtk7XAQJhJw2r7iLhqxPdXN+vWk5kvy8wLM/PCGYwFsFBNmrVyFmBGnNMCtEvOArTP7w6AgbCo7QFKKR+PiI9HRGRmaXs8gIVGzgK0T9YCtEvOArRLzgLzwUyusLomInYc8f0OzToA+kfWArRLzgK0S84CtE/WAgNhJg2rH0fEQzPzQZl5j4g4NCLO7M+0AGjIWoB2yVmAdslZgPbJWmAgTPuWgKWUtZl5RER8IyI2jIgTSymX9m1mAMhagJbJWYB2yVmA9slaYFBkKbN3y1L3RwU6dlEpZfeuJ9EmOQt0bOBzNkLWAt0qpWTXc2ibnAU6NvDntHIW6Ni4OTuTWwICAAAAAADAjGlYAQAAAAAA0CkNKwAAAAAAADqlYQUAAAAAAECnNKwAAAAAAADolIYVAAAAAAAAndKwAgAAAAAAoFMaVgAAAAAAAHRKwwoAAAAAAIBOaVgBAAAAAADQKQ0rAAAAAAAAOqVhBQAAAAAAQKc0rAAAAAAAAOiUhhUAAAAAAACd0rACAAAAAACgUxpWAAAAAAAAdErDCgAAAAAAgE5pWAEAAAAAANApDSsAAAAAAAA6pWEFAAAAAABApzSsAAAAAAAA6JSGFQAAAAAAAJ3SsAIAAAAAAKBTGlYAAAAAAAB0SsMKAAAAAACATmlYAQAAAAAA0CkNKwAAAAAAADqlYQUAAAAAAECnNKwAAAAAAADolIYVAAAAAAAAnVrU9QQAAAAAFrqDDjqoqu60006rPuY//MM/VNV95jOfqT4mAPPLAx7wgKq6X/3qVy3PZHyHH354Vd1HP/rRlmdC11xhBQAAAAAAQKc0rAAAAAAAAOiUhhUAAAAAAACd0rACAAAAAACgUxpWAAAAAAAAdErDCgAAAAAAgE5pWAEAAAAAANApDSsAAAAAAAA6pWEFAAAAAABApzSsAAAAAAAA6NSiricAAKzvyCOPrKp761vfWlV3r3vdq6ru6quvrqqLiNh3332r6rbffvuqut1226167H77zGc+U1177bXXtjgTAGAhe9jDHlZVt27duupj7rHHHlV1UzkfAmB+KaX0ta4NXY7N3DKjhlVmro6IWyLizohYW0rZvR+TAuAushagXXIWoF1yFqB9shYYBP24wmpJKeW3fTgOAOOTtQDtkrMA7ZKzAO2TtcC85jOsAAAAAAAA6NRMG1YlIs7JzIsy82X9mBAAdyNrAdolZwHaJWcB2idrgXlvprcEfHwp5ZrM3C4ivpmZq0op548saAJSSAJM34RZK2cBZsw5LUC75CxA+/zuAJj3ZnSFVSnlmubxhog4LSL2GKPm46WU3X3QH8D0TJa1chZgZtA3ElQAACAASURBVJzTArRLzgK0z+8OgEEw7YZVZm6WmVsMfx0RT4mIS/o1MQBkLUDb5CxAu+QsQPtkLTAoZnJLwPtExGmZOXycz5VSvt6XWQEwTNYCtEvOArRLzgK0T9YCA2HaDatSyhUR8Td9nAsAo8hagHbJWYB2yVmA9slaYFDM5AqrBa9518Kk7nGPe7Q8k5lbs2ZNde26detanAmjnXHGGdW1T3/606vq9ttvv6q6c845p3psWKhqM/6kk06qPuYhhxwy3emMqTa373//+1cf84c//GFV3aJFdacaG2+8cfXY/TaV1/sZz3hGVd1VV1013ekAAAtU7fniZZddVn3MX/3qV9OdDgDArJv2Z1gBAAAAAABAP2hYAQAAAAAA0CkNKwAAAAAAADqlYQUAAAAAAECnNKwAAAAAAADolIYVAAAAAAAAndKwAgAAAAAAoFMaVgAAAAAAAHRKwwoAAAAAAIBOLep6AnPNFltsUV17zDHHVNUdccQR053OrDn99NOra9/+9rdX1V1yySVVdWvXrq0ee5AMDQ1V1e27777VxyylVNUtXbq0qu6cc86pHhsWqkWL6v4p3WWXXVqeyezafPPNu55C30zlz2bnnXeuqrvqqqumOx0AYIG6/vrrq+pOO+20lmeycCxZsqSqrvbc90UvelH12K961auq6q699trqYwLAfOcKKwAAAAAAADqlYQUAAAAAAECnNKwAAAAAAADolIYVAAAAAAAAndKwAgAAAAAAoFMaVgAAAAAAAHRKwwoAAAAAAIBOaVgBAAAAAADQKQ0rAAAAAAAAOqVhBQAAAAAAQKcWdT2B2bLFFltU1Z133nnVx9xtt92mO50556CDDup77Xve856quqOOOqp67Plgq622qqr7whe+UFW38cYbz2Q6Y/rSl77U92PCQnXbbbdV1b3//e+vPuYLXvCCqrpddtmlqu7222+vqlu1alVV3VSceeaZVXWrV6+uPuaJJ55YVbfRRhtVHxPoxtDQUFXd0qVL+3q8iIglS5ZU1a1cubL6mAC0Z4cddqiqW758efUxn/CEJ1TVLVrU/1+fnXLKKVV1tb87ABjPMccc0/UUoJorrAAAAAAAAOiUhhUAAAAAAACd0rACAAAAAACgUxpWAAAAAAAAdErDCgAAAAAAgE5pWAEAAAAAANApDSsAAAAAAAA6pWEFAAAAAABApzSsAAAAAAAA6NSiricwW1asWFFVt+uuu1Yfc926dVV1119/ffUxay1aVPdHt+222/Z97Fp/8Rd/0dnY/fbkJz+5uvYrX/lKVd3mm28+3emM69Zbb62q+/CHP9z3sYGJHX/88X2vPeCAA6rqbrnllqq6Cy64oKpuKhYvXlxV95Of/KT6mBtttNF0pzOm73znO9W1F110UV/HhkE0NDRUXVt7jg7A/PPsZz+7qu5d73pXVd3GG29cVbfNNttU1QEsFFtuuWVnY69evbqq7uyzz253IswbrrACAAAAAACgUxpWAAAAAAAAdErDCgAAAAAAgE5pWAEAAAAAANApDSsAAAAAAAA6pWEFAAAAAABApzSsAAAAAAAA6JSGFQAAAAAAAJ3SsAIAAAAAAKBTi7qewGzZbbfdqupKKdXHvPnmm6vqtt9+++pj1rrXve5VVff3f//3fR+71te//vXOxq617777VtWdeuqp1cfcfPPNpzudMd1yyy3VtQcccEBV3Zo1a6Y7HWAOOfvsszsbe4MN6t7zcthhh1XVPfKRj5zJdGbkqKOOqq697rrrWpwJzG3Lli2rqlu6dGm7E5nAypUrW6kFGHSZWVX3rGc9q/qYn/70p6vqFi3q76+m/uM//qO69uijj66q+9GPflRVt2rVquqxL7vssqq6xYsXV9X5fz4sPI94xCP6WteG2t9rXnnllS3PhPnCFVYAAAAAAAB0atKGVWaemJk3ZOYlI9bdOzO/mZmXNY9btTtNgMEmawHaJWcB2iVnAdona4FBV3OF1fKI2H/UujdGxLmllIdGxLnN9wBM3/KQtQBtWh5yFqBNy0POArRtechaYIBN2rAqpZwfETeNWv13EXFS8/VJEXFQn+cFsKDIWoB2yVmAdslZgPbJWmDQTfeTLe9TSrm2+fq6iLjPeIWZ+bKIeNk0xwFYyKqyVs4CTJtzWoB2yVmA9vndATAwptuw+rNSSsnMMsH2j0fExyMiJqoDYHwTZa2cBZg557QA7ZKzAO3zuwNgvqv5DKuxXJ+Z94uIaB5v6N+UAGjIWoB2yVmAdslZgPbJWmBgTLdhdWZEvKD5+gURcUZ/pgPACLIWoF1yFqBdchagfbIWGBiTNqwy8/MR8f2IeHhmXp2ZL46Id0TEvpl5WUQ8ufkegGmStQDtkrMA7ZKzAO2TtcCgm/QzrEopzxln0z59ngvAgiVrAdolZwHaJWcB2idrgUE3acOK8W2wQd0dFR/2sIdV1V1//fXVY//ud7+rqvvkJz9ZfcxBsvnmm1fVHXfccVV197znPWcynRl597vfXV37ve99r8WZANzlH//xH6vqPv7xj7c8k/FdeumlVXXXXHNNyzOBuW3ZsmVVdUuXLm13IhM4+uijq+pqnwv9MzQ01Ne6NqxcubKVWhgkW221VVXdKaec0vJMxvejH/2oqu6pT31q9TFvuummqrra3zFM5XcHtc9nxYoVVXVPfvKTq8cGBsODH/zgqrqddtqp5ZmM78gjj+xsbOan6X6GFQAAAAAAAPSFhhUAAAAAAACd0rACAAAAAACgUxpWAAAAAAAAdErDCgAAAAAAgE5pWAEAAAAAANApDSsAAAAAAAA6pWEFAAAAAABApzSsAAAAAAAA6NSiricwW77yla9U1T3jGc+oPuaWW25ZVbdq1aqqugsvvLB67Nraj3zkI9XHrHXllVdW1d1yyy19HXeLLbaorv3Upz5VVfeQhzxkutOZsc9+9rNVde9617tanglAz7bbbltd+4pXvKLFmUzshBNOqKo74ogjqurWrl07k+nAvLd06dLOxl6yZElV3cqVK9udyAIxNDRUXbtixYr2JjLLpvJ3vPbvWu3fXZgv3vrWt3Y29g9+8IOqutrf19x0000zmc6Yttlmm74fE2CmjjzyyE7Gvfnmm6trb7zxxhZnwiByhRUAAAAAAACd0rACAAAAAACgUxpWAAAAAAAAdErDCgAAAAAAgE5pWAEAAAAAANApDSsAAAAAAAA6pWEFAAAAAABApzSsAAAAAAAA6JSGFQAAAAAAAJ1a1PUEZsub3/zmqrpf//rX1cc85JBDquq23Xbbqrrdd9+9euza2le84hXVx6z1gx/8oKru+uuvr6q74oorququvPLKqrqIiGc+85nVtf124403VtW97nWvq6r705/+NJPpAMT97ne/qrovfvGL1cfcddddpzudMV1yySXVtUuXLq2qW7t27XSnA8ySlStXdj2FgVBK6XoKA2NoaKjrKUBfvfzlL6+qO/zww/s+du3vDmr//37DDTfMZDpjqn3er371q/s+dq2rr766s7GBuW2nnXbqZNyvfe1r1bUXX3xxizNhELnCCgAAAAAAgE5pWAEAAAAAANApDSsAAAAAAAA6pWEFAAAAAABApzSsAAAAAAAA6JSGFQAAAAAAAJ3SsAIAAAAAAKBTGlYAAAAAAAB0SsMKAAAAAACATi3qegKzZdWqVVV1r371q6uPuXz58qq6448/vqpu1113rR67S4997GO7nsKcdvrpp1fV/fa3v215JgA9z3zmM6vq9txzz76PvWbNmqq6/fbbr/qY119//XSnA4xh5cqVVXVDQ0N9H3vFihVVdUuWLOn72PPBsmXLup4CMM+96U1vqqrbcMMN+z725z73uaq6xzzmMX0d90UvelF17f77719Vt3jx4ulOZ1y///3vq+qOPfbYvo8NDIYNNqi7FiUzW54J9I8rrAAAAAAAAOiUhhUAAAAAAACd0rACAAAAAACgUxpWAAAAAAAAdErDCgAAAAAAgE5pWAEAAAAAANApDSsAAAAAAAA6pWEFAAAAAABApzSsAAAAAAAA6JSGFQAAAAAAAJ1a1PUE5rOLLrqoqm6vvfaqqnvnO99ZPfajHvWoqrrNNtusqm633XarHpuJ7b///lV1BxxwQFXdt771reqx16xZU10LzH9/9Vd/VVV3xBFH9H3s2rw55phjququu+66mUwHmIGjjz66qm5oaKjvY9ces5RSVVf7XOaLvffeu+spDISVK1dW1y5ZsqS9icACs3Tp0qq6rbfeuuWZzE1vectbquouvvjilmcCzFfr1q2rqqs9l4a5YNIrrDLzxMy8ITMvGbFuWWZek5kXN8uB7U4TYLDJWoB2yVmAdslZgPbJWmDQ1dwScHlEjHXJyPtLKY9qlq/1d1oAC87ykLUAbVoechagTctDzgK0bXnIWmCATdqwKqWcHxE3zcJcABYsWQvQLjkL0C45C9A+WQsMuporrMZzRGb+V3Mp6lZ9mxEAI8lagHbJWYB2yVmA9slaYCBMt2H10Yh4cEQ8KiKujYj3jleYmS/LzAsz88JpjgWwUFVlrZwFmDbntADtkrMA7fO7A2BgTKthVUq5vpRyZyllXUScEBF7TFD78VLK7qWU3ac7SYCFqDZr5SzA9DinBWiXnAVon98dAINkWg2rzLzfiG+fERGX9Gc6AAyTtQDtkrMA7ZKzAO2TtcAgWTRZQWZ+PiKGImKbzLw6IpZGxFBmPioiSkSsjoiXtzhHgIEnawHaJWcB2iVnAdona4FBN2nDqpTynDFWf7KFuQAsWLIWoF1yFqBdchagfbIWGHRZSpm9wTJnbzAiImKzzTarqttll136Pvb73//+qrrddtut72MPkg984APVtf/8z/9cVbd27drpTme+u2jQ79W8UHN2k002qarbeuutW55Jfzz+8Y+vqjvhhBOq6jbddNOZTGdMt912W1XdmWee2fexB8nnP//56tqzzjqrxZn0zcDnbMTCzdqhoaGquqVLl/b9mCwsRx99dFXdsmXL2p3IHFVKya7n0LaFmrNtOOqoo6rq/u3f/q3lmSwMt9xyS3Vt7b+BF1988TRnwwwM/DmtnB0Mq1evrqrbcccd+zruqaeeWl37vOc9r6puAf+ucqEaN2en9RlWAAAAAAAA0C8aVgAAAAAAAHRKwwoAAAAAAIBOaVgBAAAAAADQKQ0rAAAAAAAAOqVhBQAAAAAAQKc0rAAAAAAAAOiUhhUAAAAAAACd0rACAAAAAACgU4u6ngDtuvXWW6vqLrjggr6PffPNN/f9mP323Oc+t6puiy22qD7mscceW1W3ySabVNW99rWvrR77iiuuqKr70Ic+VH1M6NJee+1VVXfUUUdV1T3taU+byXQYYdNNN62qO/TQQ1ueycytW7euunbNmjV9HfvCCy/s6/GgTStXruxrXUTE0NBQX+umYu+99+5s7H47+uijOxt72bJlnY0NTO4973lPVd2GG25YVbdkyZLqsW+88caquksvvbSq7uSTT66q+9rXvlZVFxHx8Ic/vLq2xle/+tXq2osvvrivYwPMloMPPri6tvZ3kN/97nenOx0GjCusAAAAAAAA6JSGFQAAAAAAAJ3SsAIAAAAAAKBTGlYAAAAAAAB0SsMKAAAAAACATmlYAQAAAAAA0CkNKwAAAAAAADqlYQUAAAAAAECnNKwAAAAAAADolIYVAAAAAAAAncpSyuwNljl7g9G5b37zm1V1++yzT9/Hvuaaa6rq9txzz6q6q666qnrs5z3veVV1J5xwQlXdRhttVD32DTfcUFW34447VtWtWbOmeux54qJSyu5dT6JN8yFna3/uIiK+8Y1vVNVtuumm050OA+zDH/5wVd3FF19cfcxPfepT053OQjHwORsxP7IWGFyllOx6Dm2Ts4xl5513rqpbsWJF9TG33HLLqro77rijqm6PPfaoHvuSSy6prmXWDfw5rZwdDF/96ler6g488MCWZzK+888/v6puyZIlLc+EOWbcnHWFFQAAAAAAAJ3SsAIAAAAAAKBTGlYAAAAAAAB0SsMKAAAAAACATmlYAQAAAAAA0CkNKwAAAAAAADqlYQUAAAAAAECnNKwAAAAAAADolIYVAAAAAAAAnVrU9QQYXB/72Meq6vbZZ5++j/2jH/2oqu6qq67q+9gnn3xyVd1DHvKQqrq3ve1t1WNvt9121bXQlfPOO6+6dvHixS3OZP775S9/WVVXm8ef+MQnZjKdOefWW2+tqlu3bl3LMwEAYKb22GOPqrott9yy72OvWbOmqu6SSy7p+9gA43nxi19cVXfSSSdV1T3lKU+ZyXTGtMsuu1TVHXDAAVV1Z5999kymM6YNNqi7pmeHHXaoPuZznvOcqroPfvCDVXW333579djznSusAAAAAAAA6JSGFQAAAAAAAJ3SsAIAAAAAAKBTGlYAAAAAAAB0SsMKAAAAAACATmlYAQAAAAAA0CkNKwAAAAAAADqlYQUAAAAAAECnNKwAAAAAAADo1KKuJ8DguvTSS6vqbrjhhqq67bbbbibTmXMOPvjgrqcAnfjFL35RXbv55ptX1V144YVVdccff3z12HfeeWdV3bnnnlt9zH571rOeVVV3ySWXtDwTAABo1wte8ILOxj711FM7GxtgPLW/Uz300EOr6k477bSqur333ruqLiJiiy22qKp7zWteU1X3xCc+sXrsM844o6rukEMOqap77WtfWz32qlWrquq+9KUvVdVdfvnl1WPPd66wAgAAAAAAoFOTNqwyc8fMXJGZP8vMSzPztc36e2fmNzPzsuZxq/anCzB45CxA+2QtQLvkLEC75CywENRcYbU2Io4spTwyIh4bEYdn5iMj4o0RcW4p5aERcW7zPQBTJ2cB2idrAdolZwHaJWeBgTdpw6qUcm0p5SfN17dExH9HxPYR8XcRcVJTdlJEHNTWJAEGmZwFaJ+sBWiXnAVol5wFFoJFUynOzAdGxC4R8cOIuE8p5dpm03URcZ9x9nlZRLxs+lMEWDjkLED7ZC1Au+QsQLvkLDCoam4JGBERmbl5RHw5Iv6plPL7kdtKKSUiylj7lVI+XkrZvZSy+4xmCjDg5CxA+2QtQLvkLEC75CwwyKoaVpm5OHpB+NlSylea1ddn5v2a7feLiBvamSLA4JOzAO2TtQDtkrMA7ZKzwKCbtGGVmRkRn4yI/y6lvG/EpjMj4gXN1y+IiDP6Pz2AwSdnAdonawHaJWcB2iVngYWg5jOs9oqI50fETzPz4mbdmyLiHRHxxcx8cURcGRGHtDNFgIEnZwHaJ2sB2iVnAdolZ4GBN2nDqpRyQUTkOJv36e90ABYeOQvQPlkL0C45C9AuOQssBDVXWMG0rFq1qqru5z//eVXddtttN5PpzJpDDz20qm777bdveSYwN+23337VtZtssklV3RVXXFFVd8973rN67A9/+MPVtTVWr15dXfu5z32uqu4Xv/jFNGcDAABcdtllVXXLli1rdyIALfrd735XVXfggQdW1X31q1+tHvtJT3pSVd1TnvKUqrr999+/euzXvOY1VXUbb7xx9TFrnXLKKVV1l19+ed/Hnu8m/QwrAAAAAAAAaJOGFQAAAAAAAJ3SsAIAAAAAAKBTGlYAAAAAAAB0SsMKAAAAAACATmlYAQAAAAAA0CkNKwAAAAAAADqlYQUAAAAAAECnNKwAAAAAAADo1KKuJwAXXnhhVd0TnvCE6mMeeOCBVXXXXntt9TFrbbPNNlV1G264Yd/HhvmgjZ+7WkNDQ9W1hx12WF/HPvHEE6trjznmmL6ODQAA893FF19cVbfnnntWH3O77barqnva055WVXf++edXj/2zn/2suhZgNvzxj3+sqnvuc59bfcwlS5ZU1b3pTW+qqtt5552rx954442ra2ucfPLJ1bXvfOc7+zr2QuIKKwAAAAAAADqlYQUAAAAAAECnNKwAAAAAAADolIYVAAAAAAAAndKwAgAAAAAAoFMaVgAAAAAAAHRKwwoAAAAAAIBOaVgBAAAAAADQKQ0rAAAAAAAAOrWo6wnAG9/4xqq6e9zjHtXHPPzww6vq7nOf+1Qfcz447LDDqurWrl3b8kxgbnr9619fXXvnnXdW1X3iE5+oqjvrrLOqxwYAANb30Y9+tKru2c9+dvUxt95666q6Y489tqpuaGioemyA+eo3v/lNde0Xv/jFvta98pWvrB57p512qqrbcsstq+pqf/8TEXHHHXdU17I+V1gBAAAAAADQKQ0rAAAAAAAAOqVhBQAAAAAAQKc0rAAAAAAAAOiUhhUAAAAAAACd0rACAAAAAACgUxpWAAAAAAAAdErDCgAAAAAAgE5pWAEAAAAAANApDSsAAAAAAAA6laWU2Rssc/YGY+AsWrSouvaII46oqnvzm99cVbf11ltXj13ruOOOq6o777zzqo95+umnV9XN5s/9HHNRKWX3rifRJjk7sZUrV1bXbr/99lV1D33oQ6c5GxhIA5+zEbIW6FYpJbueQ9vkLNCxgT+nlbNAx8bNWVdYAQAAAAAA0CkNKwAAAAAAADqlYQUAAAAAAECnNKwAAAAAAADolIYVAAAAAAAAndKwAgAAAAAAoFMaVgAAAAAAwP9r7/5BLL3KOI7/HszaGIuIICFGo2IXlijBRhEbRdPENGIK0coUCgoWio1pBBGjdgHFgIJ/EPyLCGohqI2YDcHEBP8gWTSsCa6FWhn1WMwVhmVn52Zn7j3v+8znA2Fn37nZe5775nyzcLh3YCoHVgAAAAAAAEzlwAoAAAAAAICpaoyxvyer2t+TwRbOnTu31ePOnz+/9Z958eLFrR53+fLlrR63zz16BlwYY9w5exG7pLPAZO07m2gtMNcYo2avYdd0Fpis/d9pdRaY7MjOHvsOq6q6tap+WlVPVNVvqupDm+v3V9XTVfXo5p+7TnvVAGeBzgLsntYC7JbOAuyWzgJnwQ1bPObfST4yxnikql6c5EJV/WTzvc+NMT6zu+UBnAk6C7B7WguwWzoLsFs6C7R37IHVGONSkkubr/9RVU8muWXXCwM4K3QWYPe0FmC3dBZgt3QWOAuO/UjAw6rqtiSvS/LLzaUPVtWvq+qhqrrplNcGcOboLMDuaS3AbukswG7pLNDV1gdWVXVjkm8l+fAY4+9JHkzymiR35OB0/4Ej/r33V9XDVfXwKawXoC2dBdg9rQXYLZ0F2C2dBTqrMcbxD6o6l+QHSX40xvjsVb5/W5IfjDFuP+bPOf7JYI/OnTu31ePOnz+/9Z958eLFrR53+fLlrR63zR5laxfGGHfOXsTV6CzQxGI7m2gt0MMYo2av4Sg6CzSx2L/T6izQxJGdPfYdVlVVSb6U5MnDIayqmw897J4kj590lQBnkc4C7J7WAuyWzgLsls4CZ8ENWzzmjUnek+Sxqnp0c+3jSe6tqjuSjCRPJblvJysE6E9nAXZPawF2S2cBdktngfaOPbAaY/wiydU+cuCHp78cgLNHZwF2T2sBdktnAXZLZ4GzYJt3WEFbzz333FaPu3Dhwo5XAgAAAAAAZ9exP8MKAAAAAAAAdsmBFQAAAAAAAFM5sAIAAAAAAGAqB1YAAAAAAABM5cAKAAAAAACAqRxYAQAAAAAAMJUDKwAAAAAAAKZyYAUAAAAAAMBUDqwAAAAAAACYyoEVAAAAAAAAUzmwAgAAAAAAYCoHVgAAAAAAAEzlwAoAAAAAAICpHFgBAAAAAAAwlQMrAAAAAAAApnJgBQAAAAAAwFQOrAAAAAAAAJjKgRUAAAAAAABTObACAAAAAABgKgdWAAAAAAAATHXDnp/vr0kuXnHtpZvrXXSap9MsSa95Os2S7G+eV+7hOWbT2fXpNE+nWZJe8+js6ere2k6zJL3m6TRL0msenT1d3Tub9Jqn0yxJr3k6zZJo7WnS2fXpNE+nWZJe80zvbI0x9vD8R6uqh8cYd05dxCnqNE+nWZJe83SaJek3z9J0e33Ns1ydZkl6zdNplqXq9Bp3miXpNU+nWZJe83SaZam6vcad5uk0S9Jrnk6zJP3mWZpur695lqvTLEmveZYwi48EBAAAAAAAYCoHVgAAAAAAAEy1hAOrL8xewCnrNE+nWZJe83SaJek3z9J0e33Ns1ydZkl6zdNplqXq9Bp3miXpNU+nWZJe83SaZam6vcad5uk0S9Jrnk6zJP3mWZpur695lqvTLEmveabPMv1nWAEAAAAAAHC2LeEdVgAAAAAAAJxhUw+squrtVfXbqvpDVX1s5lpOqqqeqqrHqurRqnp49nqer6p6qKqerarHD117SVX9pKp+v/n1pplr3NYRs9xfVU9v7s+jVXXXzDU+H1V1a1X9tKqeqKrfVNWHNtdXd3+uMctq78/Sdepssu7Wdups0qu1nTqbaO2+6eyydGqtzi6Xzu5fp9bq7LJo7TLp7P516myy7tbq7HJ16myy3NZO+0jAqnpBkt8leWuSPyf5VZJ7xxhPTFnQCVXVU0nuHGP8dfZarkdVvTnJP5N8ZYxx++bap5P8bYzxqc3/rG4aY3x05jq3ccQs9yf55xjjMzPXdj2q6uYkN48xHqmqFye5kOSdSd6Xld2fa8zyrqz0/ixZt84m625tp84mvVrbqbOJ1u6Tzi5Pp9bq7HLp7H51a63OLovWLpPO7le3zibrbq3OLlenzibLbe3Md1i9Ickfxhh/HGP8K8k3ktw9cT1n2hjjZ0n+dsXlu5N8efP1l3PwH+ziHTHLao0xLo0xHtl8/Y8kTya5JSu8P9eYhd3Q2QXp1NmkV2s7dTbR2j3T2YXp1FqdXS6d3TutXZBOnU20dql0du90dkF0drk6dTZZbmtnHljdkuRPh37/5yzgBTmBkeTHVXWhqt4/ezGn5GVjjEubr/+S5GUzF3MKPlhVv968FXUVb828UlXdluR1SX6Zld+fK2ZJGtyfBerW2aRfa1e9j4+w6r3cqbOJ1u6Bzq7D6vfyEc8BfAAAAthJREFUFVa9j3WW69CttTq7Dqvey51aq7N70a2zSb/WrnofH2HVe7lTZ5NltXbqz7Bq5k1jjNcneUeSD2ze7tjGOPjsyDmfH3k6HkzymiR3JLmU5IG5y3n+qurGJN9K8uExxt8Pf29t9+cqs6z+/rA3bVu7tn18hFXv5U6dTbSW69a2s8k69/IVVr2PdRaS6OwarHovd2qtznICbVu7tn18hFXv5U6dTZbX2pkHVk8nufXQ71++ubZKY4ynN78+m+Q7OXg77do9s/ksy/9/puWzk9dz3cYYz4wx/jPG+G+SL2Zl96eqzuUgHF8dY3x7c3mV9+dqs6z9/ixYq84mLVu7yn18lDXv5U6dTbR2j3R2HVa7l6+05n2ss5xAq9bq7PKteS93aq3O7lWrziYtW7vKfXyUNe/lTp1NltnamQdWv0ry2qp6VVW9MMm7k3x/4nquW1W9aPODyVJVL0rytiSPz13Vqfh+kvduvn5vku9NXMuJ/D8aG/dkRfenqirJl5I8Ocb47KFvre7+HDXLmu/PwrXpbNK2tavbx9ey1r3cqbOJ1u6Zzq7DKvfy1ax1H+ssJ9SmtTq7Dmvdy51aq7N716azSdvWrm4fX8ta93KnzibLbW0dvEttjqq6K8nnk7wgyUNjjE9OW8wJVNWrc3BanyQ3JPna2mapqq8neUuSlyZ5Jsknknw3yTeTvCLJxSTvGmMs/ofkHTHLW3LwNsaR5Kkk9x36bNFFq6o3Jfl5kseS/Hdz+eM5+EzRVd2fa8xyb1Z6f5auS2eT9be2U2eTXq3t1NlEa/dNZ5elU2t1drl0dv+6tFZnl0drl0ln969LZ5P1t1Znl6tTZ5PltnbqgRUAAAAAAADM/EhAAAAAAAAAcGAFAAAAAADAXA6sAAAAAAAAmMqBFQAAAAAAAFM5sAIAAAAAAGAqB1YAAAAAAABM5cAKAAAAAACAqRxYAQAAAAAAMNX/AEeDTFTt2XmvAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 2160x2160 with 5 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}